Insect Behavior 2ed PDF

Insect Behavior
Robert W. Matthews · Janice R. Matthews
Insect Behavior
Second Edition
123
Robert W. Matthews Janice R. Matthews
University of Georgia University of Georgia
Dept. Entomology Dept. Entomology
Athens GA 30602 Athens GA 30602
USA USA
[email protected] [email protected]
ISBN 978-90-481-2388-9 e-ISBN 978-90-481-2389-6

DOI 10.1007/978-90-481-2389-6
Springer Dordrecht Heidelberg London New York
Library of Congress Control Number: 2009926821
© Springer Science+Business Media B.V. 2010

No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by
any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written
permission from the Publisher, with the exception of any material supplied specifically for the purpose
of being entered and executed on a computer system, for exclusive use by the purchaser of the work.
Cover design: Boekhorst Design BV
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)

Preface to the Second Edition
This book is for all who are interested in the biological sciences. Like the course
that originally inspired it, the text is designed for use at senior undergraduate level
for college and university students, so we trust that it will find readership among
those who have had some basic introduction to entomology and animal behavior.
However, we also hope it will prove useful to newcomers who may be approaching
behavioral study from other perspectives.
This is our second edition, born anew after thirty years. Much has changed during
that time, especially in the breadth and depth of a field that (like us) was fairly young
back in 1978. New technologies are allowing scientists to shape—and answer—
questions in ways that once could not even have been envisioned. Insect behavior
research now has wings, and is poised to take off. However, at this juncture, we all
must also take care not lose an awareness of our roots. Thus, as writers introducing
this exciting field of study to the next generation of scientists and insect enthusiasts,
we have tried to strike a balance between new ideas and old, and between modern
developments and historical insights.
Our objectives in writing this edition remain the same as they have always been.
The first of these has been to help readers understand how a number of major
behavioral systems function. Thus, this is not an encyclopedia, but an introduc-
tion to fundamental concepts and processes as seen from a comparative evolutionary
viewpoint. We have not documented numerous strings of examples merely for ‘com-
pleteness of coverage’ but instead have tried to give a flavor of the diversity of ways
in which insects approach similar life tasks. Because the Internet and excellent
search engines have made access to information sources easy and nearly instan-
taneous, we have not burdened readers with a cumbersome citations in the text;
searching on key terms, aided when necessary by references associated with fig-
ure credits at the book’s end, will provide entry into additional literature for those
interested in further pursuing subjects we can but introduce.
Our second objective has been to help readers gain insights into accessible
ways in which behavioral research can be conducted. Whenever possible, we have
included discussions of important experiments and investigations, rather than pre-
senting a rhetoric of conclusions. Selected principles are interwoven with case stud-
ies of specific situations, presenting actual examples in a manner compatible with
the dynamic, open-ended field and laboratory experiences in which they have arisen.
v
vi Preface to the Second Edition
Like any writers of a general textbook, we recognize a deep obligation to many

others—to those of whose work we write, to other authors whose ideas we use,
to our own teachers who have shaped our perspectives and interests, and to our stu-
dents, friends and colleagues with their many stimulating and invaluable suggestions
and criticisms.
We also thank the many scientists and journal editors who have freely granted
permission for the use of published material. Many colleagues have generously
provided us with photographs, which are acknowledged in the credits list; special
thanks are due to Douglas W. Whitman for contributing numerous new images and
to Robert E. Silberglied and Carl W. Rettenmeyer, both now deceased, whose pho-
tographic talent continues to enhance this edition. We also are especially grateful to
Joan W. Krispyn and Paul H. Matthews for numerous original drawings.
Athens, Georgia Robert W. Matthews

Janice R. Matthews
Contents
1 The History and Scope of Insect Behavior . . . . . . . . . . . . . . 1

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.1 What Is Insect Behavior? . . . . . . . . . . . . . . . . 1
1.1.2 Insect Behavior’s Biological Context . . . . . . . . . . 3
1.1.3 Historical Foundations . . . . . . . . . . . . . . . . . 6
1.1.4 The Watershed Years . . . . . . . . . . . . . . . . . . 9
1.1.5 The Rise of Ethology . . . . . . . . . . . . . . . . . . 13
1.2 Conceptual Frameworks . . . . . . . . . . . . . . . . . . . . . 15
1.2.1 Evolution by Natural Selection . . . . . . . . . . . . . 15
1.2.2 Genetics and Behavior . . . . . . . . . . . . . . . . . 17
1.2.3 The Comparative Approach . . . . . . . . . . . . . . . 24
1.2.4 Conceptual Pitfalls . . . . . . . . . . . . . . . . . . . 25
1.3 Phylogeny’s Role . . . . . . . . . . . . . . . . . . . . . . . . . 28
1.3.1 Microevolution and Macroevolution . . . . . . . . . . 28
1.3.2 Phylogenetic Systematics and Cladistics . . . . . . . . 33
1.3.3 Behavior and Speciation . . . . . . . . . . . . . . . . 36
1.4 Questions and Perspectives . . . . . . . . . . . . . . . . . . . . 38
1.4.1 Proximate and Ultimate Analyses . . . . . . . . . . . . 40
1.4.2 Types of Approach . . . . . . . . . . . . . . . . . . . 41
2 Programming and Integrating Behavior . . . . . . . . . . . . . . . 45
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.2 Nerve-Based Coordination . . . . . . . . . . . . . . . . . . . . 46
2.2.1 The Insect Nervous System . . . . . . . . . . . . . . . 47
2.2.2 Simple Reflexes and Repeated Motor Patterns . . . . . 50
2.2.3 Ethological Explanations . . . . . . . . . . . . . . . . 56
2.3 Life in a Stimulus-Rich World . . . . . . . . . . . . . . . . . . 59
2.3.1 Sensory Tuning and Filtering . . . . . . . . . . . . . . 60
2.3.2 Memory and Learning . . . . . . . . . . . . . . . . . . 68
2.3.3 Insect Intelligence . . . . . . . . . . . . . . . . . . . . 81
2.4 Hormone-Based Coordination . . . . . . . . . . . . . . . . . . 83
2.4.1 Clocks and Reiterative Rhythms . . . . . . . . . . . . 86
2.4.2 Gated Rhythms . . . . . . . . . . . . . . . . . . . . . 88
vii
viii Contents
3 Spatial Adjustment . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
3.2 Locomotion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
3.2.1 Terrestrial and Aquatic Locomotion . . . . . . . . . . 96
3.2.2 Aerial Locomotion . . . . . . . . . . . . . . . . . . . 98
3.3 Orientation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
3.3.1 Locomotory Responses . . . . . . . . . . . . . . . . . 101
3.3.2 Posture and Position . . . . . . . . . . . . . . . . . . . 103
3.3.3 Orientation to Radiant Energy . . . . . . . . . . . . . 105
3.3.4 Magnetic Field Orientation . . . . . . . . . . . . . . . 108
3.3.5 Orientation to the Evidence of Others’ Presence . . . . 109
3.4 Thermoregulation . . . . . . . . . . . . . . . . . . . . . . . . . 110
3.4.1 Dormancy and Thermotolerance . . . . . . . . . . . . 110
3.4.2 Regulation of Heat Gain . . . . . . . . . . . . . . . . 113
3.4.3 Heat Production . . . . . . . . . . . . . . . . . . . . . 114
3.5 Migration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
3.5.1 Seasonal Migration . . . . . . . . . . . . . . . . . . . 117
3.5.2 Migration Under Ephemeral Conditions . . . . . . . . 122
3.5.3 Dispersal and Navigation . . . . . . . . . . . . . . . . 124
4 Foraging and Feeding . . . . . . . . . . . . . . . . . . . . . . . . . 131
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
4.1.1 Food Recognition and Acceptance . . . . . . . . . . . 134
4.1.2 Regulation of Feeding . . . . . . . . . . . . . . . . . . 137
4.2 Foraging Strategies . . . . . . . . . . . . . . . . . . . . . . . . 140
4.2.1 Herbivory . . . . . . . . . . . . . . . . . . . . . . . . 141
4.2.2 Active Search . . . . . . . . . . . . . . . . . . . . . . 141
4.2.3 Trapping and Ambush . . . . . . . . . . . . . . . . . . 146
4.2.4 Parasites and Parasitoids . . . . . . . . . . . . . . . . 148
4.2.5 Theft and Kleptoparasitism . . . . . . . . . . . . . . . 152
4.2.6 Insect Agriculture . . . . . . . . . . . . . . . . . . . . 154
4.2.7 Nest Symbionts: Becoming a House Pet . . . . . . . . 157
4.3 Coevolution and the Arms Race . . . . . . . . . . . . . . . . . 164
4.3.1 Attack, Defense, and Counterattack . . . . . . . . . . . 165
4.3.2 Employing Mercenaries for Protection . . . . . . . . . 170
4.3.3 The Tommy Tucker Syndrome: Food in Return
for Services . . . . . . . . . . . . . . . . . . . . . . . 173
4.4 Feeding as a Communal Activity . . . . . . . . . . . . . . . . . 177
4.4.1 Simple Groups and Feeding Aggregations . . . . . . . 177
4.4.2 Social Feeding Behaviors . . . . . . . . . . . . . . . . 181
5 Defense: A Survival Catalogue . . . . . . . . . . . . . . . . . . . . 185
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
5.2 Defense Messages . . . . . . . . . . . . . . . . . . . . . . . . 186
5.3 Passive Messages . . . . . . . . . . . . . . . . . . . . . . . . . 187
5.3.1 Crypsis: ‘I’m Not Here!’ . . . . . . . . . . . . . . . . 187
Contents ix
5.3.2 Systemic Defenses: ‘I’m Noxious!’ . . . . . . . . . . . 194

5.3.3 Mimicry: ‘I’m Someone Else!’ . . . . . . . . . . . . . 197
5.3.4 Aposematic Defenses: ‘I’m Dangerous!’ . . . . . . . . 202
5.4 Active Messages . . . . . . . . . . . . . . . . . . . . . . . . . 204
5.4.1 Attack: ‘I’m Turning the Tables!’ . . . . . . . . . . . . 204
5.4.2 Startle: ‘I’m Not What You Thought!’ . . . . . . . . . 209
5.4.3 Group Actions: ‘We’re in This Together!’ . . . . . . . 213
6 Chemical Communication . . . . . . . . . . . . . . . . . . . . . . . 217
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
6.2 Mechanisms of Chemical Communication . . . . . . . . . . . . 217
6.2.1 Odor Creation and Reception . . . . . . . . . . . . . . 218
6.2.2 Communication Through Chemistry . . . . . . . . . . 223
6.3 The Functions of Chemical Communication . . . . . . . . . . . 227
6.3.1 Finding and Choosing Mates . . . . . . . . . . . . . . 228
6.3.2 Assembly, Aggregation, and Recruitment . . . . . . . 231
6.3.3 Alarm and Alert . . . . . . . . . . . . . . . . . . . . . 240
6.3.4 Host-Marking . . . . . . . . . . . . . . . . . . . . . . 242
6.3.5 Recognition . . . . . . . . . . . . . . . . . . . . . . . 246
6.4 The Information Content of Pheromones . . . . . . . . . . . . 249
6.4.1 Physiological Adjustments: The Q/K Ratio . . . . . . . 249
6.4.2 Pheromones as Language: Syntax and Lexicon . . . . . 251
6.4.3 Exploitation and Code-Breaking . . . . . . . . . . . . 255
6.4.4 The Chemical Channel and Other Signal Modes . . . . 256
6.5 Chemical Communication and Insect Control . . . . . . . . . . 258
7 Visual Communication . . . . . . . . . . . . . . . . . . . . . . . . . 261
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
7.2 Bioluminescence . . . . . . . . . . . . . . . . . . . . . . . . . 262
7.2.1 The Physiology of Insect Light Production . . . . . . . 263
7.2.2 Bioluminescence as a Communication Method . . . . . 264
7.3 Light Reception . . . . . . . . . . . . . . . . . . . . . . . . . . 268
7.3.1 Receptors and Form Perception . . . . . . . . . . . . . 268
7.3.2 Visual Acuity and Flicker Vision . . . . . . . . . . . . 272
7.3.3 Polarized Light Perception . . . . . . . . . . . . . . . 275
7.3.4 Color Vision . . . . . . . . . . . . . . . . . . . . . . . 277
7.4 Functions of Visual Communication . . . . . . . . . . . . . . . 279
7.4.1 Aggregation and Dispersion . . . . . . . . . . . . . . 281
7.4.2 Alarm . . . . . . . . . . . . . . . . . . . . . . . . . . 283
7.4.3 Sexual Signals . . . . . . . . . . . . . . . . . . . . . . 284
7.4.4 Multimodal Signaling . . . . . . . . . . . . . . . . . . 288
8 Mechanocommunication . . . . . . . . . . . . . . . . . . . . . . . . 291
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
8.2 Producing and Sending Signals . . . . . . . . . . . . . . . . . 292
8.2.1 Sound Creation . . . . . . . . . . . . . . . . . . . . . 293
x Contents
8.2.2 Distance and Substrate . . . . . . . . . . . . . . . . . 295

8.3 Receiving Signals . . . . . . . . . . . . . . . . . . . . . . . . . 296
8.3.1 Vibration . . . . . . . . . . . . . . . . . . . . . . . . 298
8.3.2 Hearing . . . . . . . . . . . . . . . . . . . . . . . . . 303
8.3.3 Communication by Touch . . . . . . . . . . . . . . . . 305
8.4 The Acoustic Channel . . . . . . . . . . . . . . . . . . . . . . 307
8.4.1 Parameters of Insect Song . . . . . . . . . . . . . . . . 307
8.4.2 Song Synchronies . . . . . . . . . . . . . . . . . . . . 308
8.4.3 Active Acoustics . . . . . . . . . . . . . . . . . . . . 311
8.4.4 Sound as a Communication Method . . . . . . . . . . 313
8.5 Functions of Insect Communicative Sounds . . . . . . . . . . . 314
8.5.1 Protest, Alarm, and Aggression . . . . . . . . . . . . . 314
8.5.2 Aposematic Sounds and Acoustic Mimicry . . . . . . . 319
8.5.3 Sexual Signals . . . . . . . . . . . . . . . . . . . . . . 321
8.5.4 Social Sounds . . . . . . . . . . . . . . . . . . . . . . 327
9 Reproductive Behavior . . . . . . . . . . . . . . . . . . . . . . . . . 341
9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
9.2 Courtship and Mating . . . . . . . . . . . . . . . . . . . . . . 342
9.2.1 The Physiology of Mating Behavior . . . . . . . . . . 344
9.2.2 Reproduction Modes . . . . . . . . . . . . . . . . . . 346
9.2.3 Complexity and Plasticity . . . . . . . . . . . . . . . . 349
9.2.4 Pollination and Male Reproductive Behavior . . . . . . 354
9.3 Courtship and Conflict . . . . . . . . . . . . . . . . . . . . . . 359
9.3.1 Dimorphism, Sexual Selection, and Mate Choice . . . 361
9.3.2 Intrasexual Competition . . . . . . . . . . . . . . . . . 369
9.3.3 Territoriality and Dominance . . . . . . . . . . . . . . 371
9.3.4 Nuptial Gifts . . . . . . . . . . . . . . . . . . . . . . . 372
9.4 Mating Systems and Parental Investment . . . . . . . . . . . . 380
9.5 Oviposition Behavior . . . . . . . . . . . . . . . . . . . . . . . 382
9.5.1 Selecting a Site or Host . . . . . . . . . . . . . . . . . 383
9.5.2 Reproductive Rates and Energy Allocation . . . . . . . 387
10 Parental Behaviors and Social Life . . . . . . . . . . . . . . . . . . 389
10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 389
10.2 Social Organization . . . . . . . . . . . . . . . . . . . . . . . . 390
10.2.1 Aggregations and Simple Groups . . . . . . . . . . . . 390
10.2.2 Parent-Offspring Interactions . . . . . . . . . . . . . . 394
10.2.3 Solitary and Communal Nesters . . . . . . . . . . . . 400
10.3 The Insect Social Register . . . . . . . . . . . . . . . . . . . . 407
10.3.1 The Ants . . . . . . . . . . . . . . . . . . . . . . . . . 410
10.3.2 The Eusocial Wasps . . . . . . . . . . . . . . . . . . . 416
10.3.3 The Bees . . . . . . . . . . . . . . . . . . . . . . . . . 420
10.3.4 The Termites . . . . . . . . . . . . . . . . . . . . . . 423
10.3.5 Lesser Known Candidates . . . . . . . . . . . . . . . . 427
10.4 Implications and Correlates of Social Life . . . . . . . . . . . . 428
Contents xi
10.4.1 The Ecology of Parental Care . . . . . . . . . . . . . . 430

10.4.2 Paradoxes of Insect Sociality . . . . . . . . . . . . . . 434
10.4.3 Interspecific Social Interactions . . . . . . . . . . . . . 440
Credits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445
Plates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 503
Case Studies
1.1 Foraging Onset in the Honey Bee, Apis mellifera . . . . . . . . . . . . . . . . . . 42

2.1 Escape Behavior in the Cockroach, Periplaneta . . . . . . . . . . . . . . . . . . . 53
2.2 How Noctuid Moths Hear Bats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
2.3 Learning in the Parasitoid Wasp, Hyssopus . . . . . . . . . . . . . . . . . . . . . . . 77
2.4 Initiation of Adult Behavior in Saturniid Moths . . . . . . . . . . . . . . . . . . . 89
3.1 Migration in the Milkweed Bug, Oncopeltus . . . . . . . . . . . . . . . . . . . . . . 118
4.1 Myrmecophily in the Rove Beetle, Atemeles . . . . . . . . . . . . . . . . . . . . . . 159
4.2 Coevolution of Acacias and Their Ants, Pseudomyrmex . . . . . . . . . . . . 171
4.3 Mutualism Between Fig Trees and the Wasp, Blastophaga . . . . . . . . . 175
4.4 Group Feeding and Cannibalism in the Willow Leaf Beetle,
Plagiodera . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
5.1 Melanism in the Peppered Moth, Biston betularia . . . . . . . . . . . . . . . . . 191
6.1 Identification of the Mate Attractant of the Silkworm Moth,
Bombyx mori . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
6.2 Aggregation in a Bark Beetle, Dendroctonus . . . . . . . . . . . . . . . . . . . . . . 232
6.3 Host-searching by an Ichneumonid Wasp, Pleolophus . . . . . . . . . . . . . 243
7.1 Visual Communication in the Silver-Washed Fritillary,
Argynnis paphia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
8.1 Vibrotaxis in the Thorn Bug, Umbonia crassicornis . . . . . . . . . . . . . . . 300
8.2 Aggressive Singing in Two Crickets, Acheta and Gryllus . . . . . . . . . . . 317
8.3 Sexual Attraction in the Mosquito, Aedes aegypti . . . . . . . . . . . . . . . . . . 324
8.4 Communicatory Interaction in the Honey Bee, Apis mellifera . . . . . . . 328
9.1 Courtship in the Crane Fly, Tipula oleracea . . . . . . . . . . . . . . . . . . . . . . . 351
9.2 Female Choices in the Stalk-Eyed Fly, Cyrtodiopsis . . . . . . . . . . . . . . . 365
9.3 Nuptial Gifting in the Arctiid Moth, Cosmosoma . . . . . . . . . . . . . . . . . . 374
10.1 Reversed Sex Roles in the Giant Water Bug, Abedus . . . . . . . . . . . . . . . 398
10.2 Cyclical Foraging in Army Ants, Eciton . . . . . . . . . . . . . . . . . . . . . . . . . . 413
10.3 Egg-Carrying in the Golden Egg Bug, Phyllomorpha . . . . . . . . . . . . . . 428
10.4 Parental Care in the Stink Bug, Antiteuchus tripterus . . . . . . . . . . . . . . 432
xiii
Chapter 1
The History and Scope of Insect Behavior
1.1 Introduction
An overview of the insect world reveals two paradoxical characteristics: great diver-
sity and equally great constancy. On the one hand, there are over one million named
insect species, with estimates ranging up to three million. How can such a great
diversity be explained? Study of this basic question has become the domain of evo-
lutionary biology. On the other hand, each kind of organism tends to reoccur in
virtually the same form with the same basic features for generation after generation.
Why do they tend to show such constancy, such resistance to change? The study of
this question, in turn, is largely the domain of genetics. Together, these two great
branches of biology—evolution and genetics—form a powerful tool for the inves-
tigation of nearly every aspect of life. This introductory chapter deals briefly with
their application to the study of behavior and then turns to an overview of behavior
as a field of study to provide a perspective for the chapters that follow.
1.1.1 What Is Insect Behavior?
A flashing firefly flits through the evening shadows. In a tree, a caterpillar pauses
in its feeding, stiffens, and sways back and forth. Behind a stone, a cricket chirps,
while nearby ants scurry along in precise single file.
Behavior can be simply defined as what animals do. More precisely, it is the
ways in which an organism adjusts to and interacts with its total environment. As
such, insect behavior encompasses the relationships an insect has with members of
its own species, with members of other species, and with the physical environment.
A species must behave in the ‘right’ ways in order to survive, and its members must
survive (at least long enough to successfully reproduce) if it to be evolutionarily
successful.
Admittedly, the term ‘behavior’ covers a very wide range of activities, and it
can be helpful to recognize some subcategories. General locomotion, grooming,
and feeding, for example, are essentially individual matters. These maintenance
activities keep an insect in good shape but usually have little influence on others
R.W. Matthews, J.R. Matthews, Insect Behavior, 2nd ed., 1

DOI 10.1007/978-90-481-2389-6_1, C Springer Science+Business Media B.V. 2010
2 1 The History and Scope of Insect Behavior
of its kind. On the other hand, a broad range of communication activities are con-
cerned with conveying information to, and influencing the activities of others. Often
such actions are conspicuous and stereotyped, and not surprisingly they have been
a favorite study material for behaviorists. The firefly’s flash and the cricket’s chirp
may have the same function, to gain a mate. The caterpillar in essence sends the
predator world the message that it is not food. The ants share their message that
travel along this particular trail is apt to be rewarding. But while the firefly, cricket,
and ants are communicating with their own kind, the caterpillar obviously is not.
Thus, insect communication signals come in two broad, somewhat overlapping cat-
egories; even when the method of signaling is the same, the results of intraspecific
and interspecific communication are quite different.
This division is broadly reflected in the book’s organization. The first chapters
concern behavior of the individual insect—how it moves, orients, disperses, and
feeds, including the role of the nervous and endocrine systems in integrating behav-
ioral responses. The chapters on communication in a sense form the core of the book
and logically relate to defensive, reproductive, and social behaviors, all of which are
mediated by communicative codes.
The question ‘what is insect behavior?’ also can be answered another way. Insect
behavior is, of course, a discipline nested within the larger field of animal behav-
ior study. For many thousands of years, humans and their ancestors have keenly
observed animal habits and characteristics for entirely practical reasons, ranging
from the need to hunt game animals to the desirability of avoiding biting insects
and encouraging useful ones. Early Greek and Roman scholars such as Aristotle
and Pliny often wrote at some length about the natural world, including its insect
inhabitants. However, the rigorous scientific study of animal behavior only began
in the latter part of the nineteenth century, with the convergence of three major
developments—publication of the theory of evolution by natural selection, develop-
ment of a systematic comparative method, and studies in genetics and inheritance.
In the 1930s, Darwinian natural selection was combined with Mendelian inheritance
to form what is called ‘the modern evolutionary synthesis’, an explanation of ways
in which natural selection and genes interconnect (see Section 1.2.2).
Over time, from these beginnings, different major research areas arose within
the general field of animal behavior. One approach concentrated on the mechanisms
controlling behavior; this became the fields of comparative animal psychology and
physiology. A second approach concentrated on the functional significance and evo-
lution of behavioral traits, especially in natural settings; this became the field of
ethology. Somewhat later, a third area of study concentrated upon the biological
relationships between an organism and its environment, especially from an ecolog-
ical and evolutionary viewpoint; this became the field of behavioral ecology. Most
recently, significant technological genetic advances have been enabling and support-
ing the emergence of behavioral genetics and behavioral genomics as new research
concentrations.
In this way, four great disciplines have contributed to the study of behavior—
physiology (particularly neurophysiology), ecology, ethology, and psychology.
There are no distinct boundaries, yet each has had its own developmental history
1.1 Introduction 3
and tradition of established methods and brings its own viewpoint to the subject.
Somewhat facetiously, they once were distinguished from one another by behav-
ioral scientist Kenneth Roeder as follows: The ethologist, attempting to leave the
animal as unrestricted as possible in order to study its ‘normal’ behavior, tolerates
any necessary discomforts while enclosing himself in a blind. The psychologist,
attempting to reduce external variables, places the blind around the animal, thereby
making it uncomfortable. The physiologist, attempting to learn what makes the ani-
mal behave, removes it from the blind and probes directly into its nervous and motor
systems. The behavioral ecologist, we might add, spends his time studying how the
blind itself affects the animal’s behavior. Finally, the behavioral geneticist snips off
tiny samples of the animal and takes them behind the blind for a closer look.
As will become evident in the examples in this book, in recent years these
approaches have increasingly melded back toward a single discipline that contains
elements of all of these approaches. This does not mean that these fields have lost
their distinct character; in this chapter we’ll touch on what that means for behavioral
insights. However, it does mean that no matter what they may call themselves and
what direction they may approach a problem from, insect behaviorists today pursue
quite comparable goals and rely upon shared theoretical frameworks.
1.1.2 Insect Behavior’s Biological Context
Insects make up a significant proportion of the world’s biota (Fig. 1.1). The approxi-
mately 920,000 species that have been described represent almost 85% of all known
animal species; many more have yet to be been given scientific names.
Insects belong to the phylum Arthropoda, a very large assemblage of animals
with jointed legs and a hard outer skeleton. One major group in this phylum, the
Chelicerata, have sickle-shaped jaws and lack antennae; they include the Arachnida
(spiders, mites, scorpions, etc.) and two smaller marine groups. The other major
group, the Mandibulata, possess antennae and have mandibles (mouthparts) that
work against each other. Besides the insects (including the entognathous hexapods),
other major groups in the Mandibulata are the crustaceans (a predominantly aquatic
group), the centipedes, the millipedes, and two smaller classes, Symphyla and
Paurapoda. Of all the land arthropods, insects are by far the most abundant, followed
by mites and spiders.
The class Insecta is divided into a number of orders. The exact lines along which
these divisions should be made remain a matter of dispute, but in general, ordi-
nal divisions reflect the present understanding of the evolutionary history of the
class (Fig. 1.2). More than one-third of the named species of insects are beetles, the
Coleoptera. The next largest orders, in descending numbers, are Lepidoptera (but-
terflies, moths, and skippers), Hymenoptera (wasps and bees, and Diptera (flies).
Together, these four orders include more than 80% of the named species of insects.
Broadly speaking, four important stages are distinguished. First was the appear-
ance of primitively wingless insects (including the three orders of entognathous
Fig. 1.1 Tallying the numbers. Insects comprise a major proportion of the world’s biota, and
beetles make up a major proportion of the Insecta
hexapods) probably in the late Silurian Period. Bristletails and silverfish are living
representatives of these earliest insect forms. Second was the development of wings,
hypothesized to have occurred during the late Devonian or early Carboniferous.
These early winged insects had a wing-hinging mechanism that did not permit
the wings to fold, so they had to be held out from the body. The Ephemeroptera
(mayflies) and the Odonata (dragonflies and damselflies) are surviving remnants of
these ancient groups.
The third stage, the Neoptera, developed a different wing flexion mechanism, that
had evolved by the late Carboniferous. Now able to fold their wings down tightly
over their abdomens, insects could more easily run and hide from predators and
move into a wide variety of previously inaccessible niches. Among contemporary
insects, roughly 97% have flexing wings, and this mechanism is one reason for the
dominance of insects today.
The fourth important stage was the development of complete metamorphosis
(holometaboly), which also seems to have arisen by the late Carboniferous. The
earliest insects remained essentially similar in their wingless body form through-
out their entire lives. More advanced groups developed the simple metamorphosis
exhibited by insects such as grasshoppers today, where immature stages resemble
miniature adults but wings are lacking (although external wing buds are plainly
visible) until the last molt, when the insect becomes sexually mature. The most
highly advanced groups, however, evolved the complete metamorphosis illustrated
by the familiar life cycle of a butterfly. The immature stages, the larvae, bear no
resemblance to adults, and wing buds are developed internally, becoming visible
only when the larva transforms into the pupal stage, from which the winged adult
emerges.
1.1 Introduction 5
Fig. 1.2 Looking for connections. Probable evolutionary relationships among the living insect
orders
1.1.3 Historical Foundations

The earliest observers of animal behavior were almost certainly driven by the need
for animals as a source of food, clothing and raw materials for tools, but curios-
ity about the natural world has been part of the human experience since earliest
times, and ancient perceptions were often surprisingly astute. As ubiquitous as they
are, insects have been common elements in human folklore, superstitious tales,
and mythology from earliest times. These stories serve a variety of purposes, from
reenacting significant events such as the creation of the earth to instilling moral
lessons or providing instructions for living one’s life in a particular way, to simply
entertaining the listeners. However, often they were an attempt to explain natural
phenomena in a way that made sense to nonscientific minds. Explanations for com-
monly observed insect behavior often still take the form of folk tales and rhymes,
such as the story from the Tahltan of British Colombia in which a beetle larva and
a mosquito lived together. The envious larva asked the mosquito where it was able
to find food so regularly, and not wanting to divulge its secret, the mosquito replied
that it sucked its meals out of trees. The larva began boring into wood looking for
food, an activity that continues to this day.
The many volumes of observations about animals compiled by scholars such as
Aristotle (384–322 B.C.E.) and Pliny (C.E. 23–79) show that animal natural history
has been a matter of extensive interest for some time. Some of these observations
were quite astute. For example, a famous gold pendant from the Minoan culture
some 2000 years B.C.E. very accurately depicts the essentials of the life cycle of a
common social wasp (Fig. 1.3) and reveals a rather sophisticated understanding of
these insects.
However, in other cases observations were misconstrued by incomplete knowl-
edge about what was occurring or biased by beliefs based on religion or philosophy.
As a case in point, consider spontaneous generation, the idea that living things could
arise from non-living sources (Fig. 1.4). One of the most persistent of such miscon-
strued beliefs, it held sway for two millennia. Aristotle is credited with synthesizing
this theory, compiling and expanding the work of earlier natural philosophers and
various ancient explanations of the appearance of organisms; his examples included
the generation of maggots from rotting meat, mice from dirty hay, birds from trees,
and lice from sweat.
In the days before microscopes and careful scientific investigations, the ideas
seemed quite reasonable to most people, but how could spontaneous generation be
reconciled with Christian beliefs? Pope Innocent V in the thirteenth century had
declared that belief in spontaneous generation went against Church teachings, since
all life was created in the first days of Creation chronicled in Genesis. Saint Thomas
Aquinas concluded that spontaneous generation of insects was the work of the Devil.
The Greek tradition prevailed in the scientific community.
As the seventeenth century turned to the eighteenth, experiments and natural his-
tory observations began to rebut the prevailing opinions. Studies of insects led the
way. One of the first experiments in the modern sense (with the use of controls)
was the classic study of Francesco Redi (1668) who showed that maggots did not
1.1 Introduction 7
Fig. 1.3 Depicting nature. An ancient pendant, considered one of the most outstanding examples
of early Minoan goldworking, shows two paper wasps, Polistes. Their legs embrace a granulated
disc thought to represent the paper nest; a droplet of food or wood pulp is held in their jaws. The
cage-like fixture atop the heads may be an attempt by the artist to depict the vigorous antennal beat-
ing characteristic of all encounters between individuals on a nest. The opposition of the abdominal
tips may have been intended to suggest mating, but is biologically inaccurate; more probably it
reflects the heraldic symmetry common in Minoan art. Most classics texts incorrectly refer to the
insects as bees, sometimes with elaborate explanations about the symbolism inherent in the ‘honey
droplet’ between their jaws
arise spontaneously from rotten meat. Maria Sibylla Merian (Fig. 1.5) took the bold
step of actually studying butterflies and moths throughout their entire life cycles,
taking detailed notes and making careful illustrations. However, when Merian pub-
lished her lavishly illustrated works, the scientific community largely ignored them
although both Linnaeus and Fabricius’s subsequent insect descriptions were appar-
ently influenced by her work. The invention of the microscope in the 1600s might
have resolved the matter of spontaneous generation but in fact it only served to
enhance the belief by revealing a whole new world of microorganisms that appeared
to arise spontaneously. The controversy around spontaneous generation continued
until it was ultimately settled by the experiments of Louis Pasteur in the nineteenth
century. Finally, well into the twentieth century, the work of Maria Merian was redis-
covered and recognized. Her portrait has been printed on German postage stamps
and was on the 500 DM note before Germany converted to the euro; in addition,
many schools (and a modern research vessel) have been named after her.
The first ‘major player’ to introduce elements of scientific discipline to insect
study was Rene Antoine Ferchault de Reaumur. Reaumur was a widely trained
scientist who made major contributions to areas as disparate as geometry, met-
allurgy, and meteorology, but many consider his greatest work to have been in
Fig. 1.4 Explaining insect life. A sixteenth-century woodcut supposedly shows honey bees being
generated spontaneously from a dead animal. Some medieval texts contained detailed directions
for creating insects from substances as varied as dead animals, logs, and dirt
entomology. Published between 1734 and 1742, his six large volumes of Memoirs
pour Servir a l’Histoire des Insectes applied precise observation, detailed experi-
mentation, and accurate recording to phenomena as varied as social life, parasitic
habits, and leaf-mining.
The historical development of the scientific field of animal behavior (and insect
behavior as a sub-discipline within it) accelerated from about 1750 onward. At this
time, all of European society was becoming ‘scientific’, and each important expedi-
tion began to include at least one professional naturalist. As a result so many exotic
plant and animal specimens were being brought back to Europe from around the
world that chaos loomed for the naturalists who were trying to identify, classify, and
communicate what they gathered. Linnaeus’ method of classification was gaining
increasing exposure and acceptance as a way to organize all these data. When the
tenth edition (1758–1759) of his Systema Naturae was chosen as the starting point
for zoological nomenclature, it marked a major milestone in biology.
Thus, by the time the 1700s drew to a close, three kinds of entomological texts
were prevalent. First, there were beautifully colored illustrative works such as those
of Maria Merian. Second were descriptive classificatory works, such as Linneaus’
Systema Naturae with its system of binomial nomenclature. Third were detailed
works such as those of Reaumur’s Memoirs, detailing specific aspects of insects
such as their development, physiology, or internal anatomy.
1.1 Introduction 9
Fig. 1.5 Illustrating the truth. Rejecting spontaneous generation, Maria Merian accurately
unraveled the mysteries of metamorphosis in many Lepidoptera in the late 1600s, and published
her findings in beautifully illustrated detail. (left, a portrait by her son-in-law, Georg Gsell; right,
her drawing of the silkworm life cycle.)
However, for the next half-century it would still be a time when general observa-
tion predominated over specialization. Enthusiasm and subjectivity often affected
the accuracy of behavioral observations, and interpretations were often slanted
to embrace a particular philosophical creed. Most scientists still subscribed to
Aristotle’s Scala Naturae, a theory that all living beings could be classified on an
ideal pyramid. The simplest animals occupied the base, and complexity rose pro-
gressively to the top, which was occupied by human beings. Animal species were
generally thought to be eternal and immutable, created with a specific purpose (by
which most people meant ‘created by God to serve mankind’). It seemed like the
most logical, and perhaps only possible, explanation for the incredible variety of
living things and their surprising adaptations to their environment.
1.1.4 The Watershed Years

During the latter half of the 1800s, three conceptual advances occurred that were
so revolutionary they deserve independent mention: evolution by natural selection
(Section 1.2.1), the discovery of the genetic bases of biology (Section 1.2.2) and the
development of a comparative approach to biological study (Section 1.2.3).
Keeping those developments aside to consider separately, consider for a moment
the scientific mileau into which those three developments emerged. It has been said
that upon seeing an animal, the first question of both layman and scientist alike is
always, ‘What is it?’ During the 1800s, a major driving force in science was the need
simply to identify, name and classify the diversity that was present. The problem
was particularly acute for places that were less well studied than the Old World that
most scientists called home. With a system of binomial nomenclature having been
established only relatively recently, and an entire world beckoning, the adventure of
collecting and classifying organisms called out to many natural historians.
One who answered that call was John Lawrence LeConte. Like many scien-
tists of his day, LeConte had the money and connections to indulge his passion.
A few months after he was born in New York City, his mother died; raised by his
father, a well-known naturalist, LeConte trained as a medical doctor. He never offi-
cially worked as an entomologist—the field was still too new for that. During the
American Civil War he worked as a surgeon with the California volunteers, reaching
the rank of lieutenant colonel. In 1878 he became the chief clerk (assistant direc-
tor) of the United States Mint in Philadelphia, and retained that position until his
death. Throughout his life, he traveled abroad extensively, explored the Colorado
River, and accompanied the teams that built the Honduras Interoceanic Railway and
Kansas Pacific Railroad.
Over the course of his life, LeConte almost single-handedly put American ento-
mology on the map. He published his first scientific work in 1844, while he was
still in medical school. Everywhere he went, he sent back insect specimens, some-
times by the tens of thousands. In all, he was responsible for naming and describing
approximately half of the insect taxa known in the United States during his lifetime,
including some 5,000 species of beetles. (Not surprisingly, he is often described
as ‘the father of American beetle study’.) Le Conte was probably the best known
entomologist of his century, but it was a time of rapid growth and expansion in
the sciences, and LeConte was part of an entire roster of insect afficionados whose
names are still familiar to working scientists today. His influence was considerable,
for LeConte was also very active in the scientific societies of his time; in addition
to serving as vice-president of the American Philosophical Society and president of
the American Association for the Advancement of Science, he was a founder of the
American Entomological Society, and a charter member of the National Academy
of Sciences.
Meanwhile, in France, the late nineteenth century had introduced the works of
Jean-Henri Fabré. Born a pauper, his education and opportunities were limited.
He was an avid naturalist, not a scientist; a moody and morose soul, his personal
life reads like a tragic drama. However, the publication of his romantically titled
Souvenirs Entomologiques, a set of ten volumes describing in detail the lives of
insects in his own backyard, popularized insect behavior in writings that achieved
the heights of literary excellence. (In both 1904 and 1911, he was nominated for
a Nobel prize in literature.) Modern scientific advances have rendered much of
his work obsolete, and for a time the scientific community had difficulty accept-
ing him, particularly because he vigorously rejected the theory of evolution and was
absolutely convinced of the ‘fixity of instinct’. However, Fabré was one of the great-
est popularizers of entomology the world has ever known, and his work served to
1.1 Introduction 11
establish standards of observational patience and accuracy that subsequent workers

were to continue. In addition, in the process of countering his ideas, his critics fre-
quently made valuable contributions of their own, particularly in helping to focus
on the variability of behavior and emphasizing the necessity for a firm taxonomic
foundation for behavioral studies.
In emphasizing the growth of entomology as a field of study within zoology,
it is important also to acknowledge that the study of insects has a unique twist
because insects and humans have always had an uneasy relationship. In the struggle
for human survival and well-being, insects have also been relentless pests, com-
petitors for food and fiber, and threats to human health and comfort. When human
agriculture arose, a mere 10,000 years ago, primitive hunter-gatherers entered into
a new partnership with plants that intensified this antagonism. There are records of
the Sumerians using sulfur compounds as insecticides over 4,500 years ago; the Rig
Veda, which is about 4,000 years old, also mentions the use of poisonous plants for
pest control. Ancient Chinese and Egyptian cultures are known to have used chem-
ical pest controls. With the industrialization and mechanization of agriculture in
the eighteenth and nineteenth century, chemical pest control became widespread,
a development that accelerated with the twentieth century discovery of several
synthetic insecticides, such as DDT.
Thus from the beginning, there has been a basic duality in entomology’s aims,
and it continues to this day. One branch of the field studies insects for the basic
love of them; the other, with a desire to control or kill at least a certain subset of
them. The former is often called ‘basic’ entomology; the latter, ‘applied’ or ‘eco-
nomic’ entomology. Economic entomology involves the study of insects that are
of benefit or those that cause harm to humans, domestic animals, and crops, often
with intent of either eliminating, reducing or augmenting their numbers. Insects that
cause losses are termed as pests; those that cause indirect damage by spreading dis-
eases, vectors. Those termed ‘beneficials’ include insects valued for food products
such as honey, for substances such as lac or pigments, for pollination, or as natural
enemies for pest control.
Because control efforts offered immediately obvious benefits (at least theoreti-
cally and often as a practical outcome), economic entomology was a great driving
force behind the establishment of insect study as a separate field. Across the United
States, one state after another appointed state entomologists, beginning in 1854 with
the appointment of Asa Fitch as the first professional Entomologist of the New York
State Agricultural Society. The field grew even more rapidly in the late nineteenth
and into the twentieth century, with the appearance of an influential body of reports
written by Charles Valentine Riley and the establishment of the US Department of
Agriculture.
Concurrently, similar government-sponsored growth in entomology occurred
through military service, setting the stage for a new field of medical entomology.
One of its early successes occurred when Walter Reed, a United States Army major,
was appointed president of a board ‘to study infectious diseases in Cuba paying
particular attention to yellow fever’ (Fig. 1.6). (Although Reed received much of
the credit in history books, Reed himself correctly credited the Cuban doctor Carlos
Juan Finlay with identifying Aedes mosquitoes as the yellow fever vector and thus
determining how yellow fever could be controlled.) Not long afterward, Carlos
Chagas observed the peculiar infestation of rural houses in Brazil with Triatoma,
demonstrated that it was the vector of Trypanosoma cruzi, and proved experimen-
tally that it could be transmitted to marmoset monkeys that were bitten by the
infected bug. His description of the new disease was to become a classic in medicine
and brought him domestic and international distinction.
Fig. 1.6 Discovering a killer’s cause. Yellow fever was one of the most dreaded diseases in port
cities of the Americas. Poorly understood and feared as cancer is today, and known to have its ori-
gins in the Caribbean region, the disease was thought to spread directly and through contaminated
clothing and bedding. When yellow fever became a problem for the Army during the Spanish
American War, felling thousands of soldiers in Cuba, Major Walter Reed, M.D. (left) headed a
team that proved that yellow fever is transmitted only by Aedes mosquitoes (center), confirming
a theory first set forth in 1881 by Cuban doctor/scientist Carlos Finlay (right). The insight gave
impetus to the new fields of epidemiology and biomedicine, and allowed the United States to build
the Panama Canal, something that had confounded the French attempts only 30 years earlier
The ‘basic’ side of entomology also grew rapidly during this time. In 1901,
the first widely used textbook, A. D. Imms’ General Textbook of Entomology was
published; going through numerous editions, for close to a century it remained
one of the most widely used of all insect texts. At Columbia University, Thomas
Hunt Morgan became one the first to conduct genetic research with the fruit fly
Drosophila melanogaster, a development that would ultimately underpin the field
of behavioral genetics.
In a similar way, all of animal behavior study was growing and subdividing into
a multidisciplinary subject. The nineteenth century was a time when the nature of
scientific enquiry was changing in ways that would have major impact on the field of
insect behavior. In addition to general natural history studies, an increasing number
of specialized disciplines began to characterize the natural sciences. Psychologists,
anthropologists, ecologists, geneticists, and many others who contribute to the study
of animal behavior can trace early common roots back to this period. It was during
this time, born in the principles of evolution delineated by Charles Darwin and nur-
tured by this convergence of entomology and more objective scientific analysis, that
modern insect behavior arose. However, as a field of study it went on to show its
1.1 Introduction 13
greatest development after the turn of the century, with the emergence of the largely
European school of ethology.
1.1.5 The Rise of Ethology
The word ethology has a long pedigree. Based on the Greek ethos, which has a
variety of meanings, the term has been applied to everything from stage actors
who portray human characters to people who study ethics. However, as it is used
today, ethology means the study of the behavior of animals in their natural habitat.
The term was first popularized in English by the American myrmecologist William
Morton Wheeler in 1902, but it was hardly a new idea. Over 3,000 years ago,
King Solomon recommended the study of insects in their natural habitat with those
famous words in Proverbs 6.6: ‘Go to the ant, thou sluggard; consider her ways, and
be wise.’
In the early years of the twentieth century, two different major conceptual
approaches to animal behavior study split off from one another. For most of
the twentieth century, the discipline of comparative psychology developed most
strongly in North America, whereas ethology was stronger in Europe. This led to dif-
ferent emphases, different philosophical underpinnings, and different experimental
approaches. Comparative psychology came to pay particular attention to the psy-
chological nature of human beings in comparison with other animals. A practical
focus on laboratory studies made a logical base for experimental studies on human
and animal brain function, learning, and motivation. Well-known studies include
those of Ivan Pavlov on conditioning in dogs, Harry Harlow on the effects of social
deprivation in monkeys, and those of various researchers on language abilities in
apes.
In the 1920s and 1930s, when most behavioral research was laboratory based
emphasizing the role of learning and hormones in the modification of behavior,
two Europeans, Konrad Lorenz and Nikolaas (‘Niko’) Tinbergen, began publishing
intensive and extensive natural behavior studies that quickly caught the attention
of scientists and the general public. As their carefully detailed work unfolded, so
did the foundations of ethological theory. As in most cases, their scientific work
owed a debt to others. In a sense, Charles Darwin might be called the first modern
ethologist; his book, The Expression of the Emotions in Man and Animals, continues
to influence ethologists today. Some scholars would acknowledge Darwin’s protégé
George Romanes (see Section 1.2.3) as another ‘father’ to the field, despite his odd
methods. Other clear leaders were Oskar Heinroth and Julian Huxley, who both
concentrated on behaviors they called instinctive, or natural. Their focus was upon
building an objective cumulative base of data about behavior. Thus, their first step
in studying the behavior of any new species was to construct an ethogram, a kind
of glossary describing each type of natural behavior shown by the animal and the
frequency of that behavior’s occurrence.
Later chapters in this book will return to the ideas of Lorenz and Tinbergen;
for now it enough to realize that their research rested on a different practical and
philosophical foundation. Whereas the comparative psychologists focused on labo-

ratory work in order to have more control over environmental parameters and over
knowledge of an animal’s history, these European behaviorists emphasized the study
of animals in their natural environment, so that behaviors could be more easily
placed in context and their evolutionary relationships could be more easily dis-
cerned. They also typically focused on a behavioral process rather than a particular
animal group, and they often studied one type of behavior (such as aggression or
sociality) in a number of unrelated animals. In this they also differed from compar-
ative psychologists, who despite the name of their discipline, focused their efforts
almost exclusively on the white rat.
The work of Lorenz and Tinbergen set in motion a whole new approach to behav-
ior study. First, it provided hints of a general order and logic in the behavior of
animals. Second, it demonstrated that certain behavior patterns are just as charac-
teristic of species as certain morphological features are. Third, because it regarded
behavior as part of every organism’s equipment for survival and the product of adap-
tive evolution, its focus was the objective study of whole patterns of animal behavior
under natural conditions. Thus, it emphasized the functions and evolutionary history
of behavior patterns. Behavior, they said, always has a cause, and it always has one
or more functions.
Through the efforts of Lorenz,Tinbergen, and their colleagues, ethology devel-
oped robustly in continental Europe in the years before World War II. After the
war, Tinbergen moved to Great Britain, and ethology became much stronger in the
United Kingdom. It was a time of increasing transcontinental study, and at this time
ethology also began to take hold in North America. Similar travel in the other direc-
tion carried comparative psychology to Europe. As the years have passed, there has
been an increasing realization that both approaches have their strengths and limi-
tations, although it took a while before this was grudgingly acknowledged on both
sides of the Atlantic.
Lorenz, Tinbergen and their early followers dealt mostly with birds, and included
insects (and fish) only in passing. However, as the field has developed, it has come
to encompass all animals and even human beings. Modern ethology combines labo-
ratory and field science. It has strong ties to other disciplines such as neuroanatomy,
ecology, and evolution. Since the 1970s, the field has subdivided once again, this
time into comparative ethology and social ethology. The latter concentrates on
the behavior of social groups and the social structure within them; jointly with
evolutionary psychology, it lays claim to the study of sociobiology.
The appearance of insect behavior as a separate sub-discipline is a relatively
recent phenomenon, but it shares a conceptual base with the entire field of ethology.
Like all of behavioral research, insect behavior is still relatively young, however, and
as such it lacks the strong theoretical framework that has come to typify more mature
sciences such as chemistry. Additionally, since insects comprise over three fourths
of all animals, insect behavior is a potpourri of knowledge and ignorance. Insects
provide some of the most outstanding and best-developed examples of behavioral
phenomena, notably in the areas of communication, courtship, mimicry, and the
1.2 Conceptual Frameworks 15
development of sociality. But at the same time, even according to very liberal esti-
mates, fragmentary behavioral data (host records, food plants, etc.) are recorded for
no more than 5% of the described insect species; the number of species subjected to
intensive investigation is far smaller yet.
1.2 Conceptual Frameworks

The rigorous study of animal behavior, and insect behavior as a subfield within it,
began with three major developments in the last half of the 1800s: publication of
the theory of evolution by natural selection, studies in genetics and inheritance, and
development of a systematic comparative method. To understand the conceptual
frameworks that underlie biological study in general and insect behavior study in
particular, let us briefly consider each of these developments in turn.
1.2.1 Evolution by Natural Selection
Few people have never heard of the theory of evolution by natural selection and
Charles Darwin’s book, The Origin of Species. Darwin formulated his idea of natural
selection in 1838 and was still developing his theory in 1858 when Alfred Russell
Wallace sent him a similar theory that he had independently developed from his own
data set. The works of both men were presented to the Linnean Society of London
in separate papers.
Though these publications set a major paradigm shift in motion throughout the
entire scientific community, their theory of evolution did not develop in a vacuum. It
drew its ideas from the contributions and suggestions of many other people, from the
selective breeding experiments of English farmers to the ideas of Thomas Malthus
about human competition for survival and the fossil observations by geologist Sir
Charles Lyell that suggested a process of continuous change in living material
through time.
Jean-Baptiste Lamark, not Charles Darwin, actually was the first prominent biol-
ogist to set forth a complex theory of evolution. His theory proposed two major new
ideas. One was that animal organs and behavior were not fixed; they could change
according to the way they were used, and these changes could be transmitted from
one generation to the next. The second idea was that every living organism, human
beings included, tends to reach a greater level of perfection over time. Darwin was
fully aware of these theories at the time of his journey on the H.M.S. Beagle, and
he was deeply influenced by them. As an interesting historical side-note, whereas
much debate ensued in England over Darwin’s publication, it was a non-event in
France. In the eyes of the French, Darwin was just attempting to reformulate theories
already put forward by two of their own scientists, Lamarck and Etienne Geoffroy
Saint-Hilaire, a prominent colleague of Jean-Baptiste who expanded and defended

Lamarck’s evolutionary theories.
In a sense, one could say that the theory of evolution by natural selection gained
its immediate prominence because it arrived at a point in history that provided a
‘perfect storm’ for its acceptance. As such, it set the stage for modern biological
science and remains the foundation of every behavioral explanation couched in ulti-
mate analysis terms. However, because the theory is often misunderstood, it is worth
briefly reviewing what the theory is and is not.
Evolution refers to the natural or artificially induced process by which new and
different organisms develop as a result of changes in genetic material. Although
Darwin and Wallace did not have the benefit of modern genetic knowledge, they
astutely recognized that evolutionary change is inevitable if just three conditions
are met. First there must variation; some members of a species must differ in some
of their characteristics. Second, the variation must be due to heredity; parents must
be able to pass some of their distinct characteristics on to their offspring. Third,
there must be differences in reproductive success; thanks to their distinctive charac-
teristics, some individuals must have more surviving offspring than others in their
population. When (as almost always happens) there is hereditary variation within
a species and some variants consistently reproduce more successfully than others,
eventually the increased numbers of living descendents of the more successful types
will change the makeup of the species. Over many generations, adaptations (trait
differences that make an animal better able to survive and reproduce) occur through
a combination of successive, mostly small, random changes in traits and natural
selection of the better-suited variants. Evolution occurs when these heritable dif-
ferences become more common or rare in a population. Because the process that
causes the change is natural, Darwin called it natural selection.
Evolution is a powerful explanatory and predictive idea set that directs research
by constantly raising new questions. In the years since Darwin and Wallace first
presented its tenets, evolution has become the central organizing principle of modern
biology, providing a unifying explanation for the diversity of life on Earth. It is
unfortunate that Darwin’s choice of words in calling his ideas a ‘theory’ has caused
problems for some non-scientists. When a layperson uses the word theory, he or she
often means an idea or belief about something that is arrived at through speculation
or conjecture—basically, a guess unsupported by evidence. However, theory means
something entirely different in a scientific context. When a scientist uses the word,
he or she means the body of rules, ideas, principles, and techniques that applies to
a particular subject. In essence a scientific theory is a set of facts, propositions, or
principles analyzed in their relation to one another and used to explain observed
phenomena. Thus, a scientist may speak not only of the theory of evolution but of
Einstein’s theory of general relativity or Newton’s theory of gravitational attraction.
Nowadays, the fact that organisms evolve is uncontested in the scientific lit-
erature and the modern evolutionary synthesis is widely accepted by scientists.
However, evolution remains a contentious concept for some segments of society. As
Darwin himself recognized, the most controversial aspect of evolutionary biology is
its implications for human origins.
1.2.2 Genetics and Behavior

Nowadays even most schoolchildren know that a great deal of the variation among
individuals in a species is due to differences in their genes (DNA segments that
encode the information needed to make various proteins). Because genes can be
copied and transmitted to offspring, parents can pass on the information needed to
carry on these relevant distinctive characteristics; genetic variation within a species
occurs when a gene exists in two or more forms, or alleles. Thus, it is sometimes dif-
ficult to remember that only about 150 years ago, when Darwin, Wallace, Lamarck,
and others were developing their theories, scientists knew so little about the nature
of heredity that they could not account for how various characters might be passed
down from generation to generation.
In 1865 Gregor Mendel found that traits were inherited in a predictable man-
ner, but at the time his work went largely unnoticed, and Darwin and Wallace were
unaware of it. Then, in the early 1900s, Hugo de Vries and others rediscovered
Mendel’s work. Although the rediscovery provided the impetus for scientists to bet-
ter understand how natural variation arises, disagreements over the rate of evolution
predicted by early geneticists and biometricians led to a rift between the Mendelian
and Darwinian models of evolution. In the early years of the twentieth century, a
combination of specialization and poor communication among biologists resulted
in much confusion and difficulty.
The rift did not begin to heal until the 1920s and 1930s, when evolutionary biol-
ogists such as J.B.S. Haldane, Sewall Wright, and Ronald Fisher established the
foundations for the field of population genetics. In the 1940s, Oswald Avery and
colleagues identified DNA as the genetic material, and in 1953 James Watson and
Francis Crick published its structure, demonstrating the physical basis for inher-
itance. In 1956, Margaret Bastock published the first demonstration that a single
gene could change a behavior pattern.
Beginning in the 1930s and 1940s, a number of leading biologists worked
together to reconcile their various fields resulting in a more comprehensive view
of evolution that combined natural selection and Mendelian inheritance called the
modern evolutionary synthesis. It drew together ideas from several branches of biol-
ogy that had become separated and provided compelling evidence that field-based
population studies were crucial to evolutionary theory.
Today, genetics and molecular biology have become core parts of evolutionary
biology. A newer field, genomics, has also developed. While the terms genetics and
genomics are sometimes used interchangeably, genetics usually refers to the study
of individual genes, whereas genomics refers to the study of the compete set of
genes in an organism. The latter field of study is moving rapidly to sequence insect
genomes.
It is now also clear that evolution is driven by not just one but two major mecha-
nisms. The first is natural selection, the process just outlined, acting on the existing
genetic variation in a population. The second is genetic drift, an independent pro-
cess that produces random changes in the frequency of traits in a population due to
chance rather than selection or gene flow.
Natural selection and genetic drift can only work on what is available, and for
this reason it is important to recognize that the traits and alleles that are present in
a population can vary over time. Various catastrophes, expansions into new habitats
(causing a founder effect), or events that divide a population can lead to a popula-
tion bottleneck, in which the number of breeding individuals in a population shrinks
temporarily and therefore the population loses genetic variation. On the other hand,
new traits can also arise, either from mutations in genes or from the transfer of
genes between populations and between species. Mutations are the ultimate source
of variation, but in species that reproduce sexually, new combinations of existing
alleles are also produced by genetic recombination (by exchange of chromosome
sections during meiosis). Gene flow is the exchange of genes between populations,
usually of the same species. For example, some grasshoppers migrate widely before
breeding; this migration brings previously isolated populations together and facili-
tates gene flow within the species, slowing the process of speciation that might have
occurred had the populations remained segregated. Migration not only provides an
opportunity for mixing; it also may add or remove genetic material from the gene
pool of a population.
Genetic drift is random change in the frequency of alleles, caused by the random
sampling of a generation’s genes during reproduction. It results from the role prob-
ability plays in whether a given trait will be passed on as individuals survive and
reproduce. Genetic drift is analogous to the way that flipping a coin over and over
may give a proportion of ‘heads’ and ‘tails’ that differs from the equal numbers one
would expect. Because the relative importance of natural selection and genetic drift
in a population varies depending on the strength of the selection and the effective
population size (the number of individuals capable of breeding), the size of a pop-
ulation can greatly affect its evolution. Natural selection usually predominates in
large populations, while genetic drift dominates in small populations. For example,
a population greatly reduced by a bottleneck will often show increased genetic drift.
The dominance of genetic drift in small populations can even lead to the fixation of
slightly deleterious mutations.
The changes produced in any one generation by selection, drift, mutation, and
gene flow are small, but these differences tend to accumulate with each subse-
quent generation and over time they can cause substantial changes in the organisms.
Consider a hypothetical case in which all of the beetles of one species living in the
trees on an isolated mountaintop represent a population. A single gene in this pop-
ulation exists as two alleles that account for variations in the physical appearance
(phenotype) of the organisms, in this case brown or green body color. A gene pool
is the complete set of alleles in a single population, so each allele occurs a cer-
tain number of times in a gene pool. The fraction of genes within the gene pool
that are a particular allele is called the allele frequency. Evolution occurs when
there are changes in the frequencies of alleles within a population of interbreeding
organisms. Suppose birds more commonly notice the brown beetles and eat them
before they can reproduce whereas the green beetles, successfully camouflaged by
the tree foliage, continue to breed. In each generation, the allele for green body
color in the population would become more common and the brown allele, less so.
Over time, unless some other factor favored brown beetles enough to overcome this
difference, a population of green beetles would evolve.
Under what conditions would a population not evolve? One way to examine
whether a population is undergoing evolutionary change is through the use of the
Hardy-Weinberg equilibrium. The Hardy-Weinberg principle is an example of a null
model. Just as a null hypothesis is a hypothesis of no difference, this null model
tells us what to expect when no evolutionary forces are at work. Based on the fre-
quency of alleles, Hardy-Weinberg predicts the frequency of (diploid) genotypes in
the absence of evolution. If the genotype frequencies we observe in our study pop-
ulation are the same as those determined by the principle, our population is said to
be in Hardy-Weinberg equilibrium; it is not evolving. If we observe different fre-
quencies in our study population, we can conclude that some evolutionary force is
at work.
Like the fine print in an insurance policy, the Hardy-Weinberg equilibrium has
a great many restrictions. The equilibrium is relevant only for diploid organisms
with sexual reproduction. Mating must be random. The gene under consideration
has only two alleles, and allele frequencies are identical in males and females.
Population size is very large, so that genetic drift is essential imperceptible. Gene
flow and mutation are so small as to be negligible. Finally, natural selection is
not acting on the alleles under consideration. Of what possible use can such a
restrictive formula be? The answer lies in some of its permutations, which allow a
researcher to sample a population at one point in time and to determine whether it
is at equilibrium for a set of alleles without having to take repeated samples across
several generations.
A major condition under which evolution would slow or halt would be when there
is little or no genetic variability for natural selection to act upon. Clearly, for some
species, genetic variability is very important, as illustrated by those insect species
that seem to have enough persistent allelic diversity to be able to evolve resistance
to one insecticide after another. Genetic variability is also important in the ability
of both invaders and insects imported for biological control to establish themselves,
as well as in their potential to attack new and non-target hosts. On the other hand,
genetic diversity seems fairly unimportant for some species. For example, many par-
asitoid wasps such as Melittobia (see Fig. 9.13), typically undergo one generation
after another of brother-sister matings, and are assumed to be highly inbred. In the
case of some invasive insects, lack of variability may actually contribute to their suc-
cess. Research has shown that invasive populations of the Argentine ant Linepithea
humile have lost the genetic ability to distinguish one colony from another; as a
result, they escape from population control imposed by deadly conflicts between
colonies. So how important is genetic diversity itself to the persistence of insect
populations? Surprisingly, this question is still an open one, and difficult to answer
in a general way.
Ultimately, one of the goals of a behaviorist is to pair a particular behavioral com-
ponent with a particular gene, determining the actual site at which the gene wields
its influence and how it does so. In fruit flies (Drosophila), probably the single best
genetically studied insect genus, a host of intriguing behavioral abnormalities have
been traced to mutations of single genes. Intensive inbreeding of laboratory stocks

has produced populations identical genetically in every respect except for a single
mutant gene induced by radiation or certain chemicals. Such populations have been
demonstrated to differ behaviorally, often in several traits.
However, in general the task of pairing behaviors and genes is complicated by
the fact that the behavioral impact of a mutant gene may not be its primary effect.
A functional nervous or muscular system must be constructed in a precise way for
‘typical’ behavior to occur. Since genes code for proteins such as enzymes, even
a single gene mutation may have broad ramifications on the construction of entire
systems. A detailed look at the field of behavioral genetics is outside the scope
of this book. However, throughout the chapters that follow we repeatedly assume
the genetic determination of behaviors. For this reason, it is worth noting two useful
methods by which such genetic assumptions might be verified: crossing experiments
and selection experiments.
In order for natural selection to act on a trait, that trait must be passed down from
one generation to the next and there must be genetic variation in it. However, when
we see differences in a trait (such as different behaviors, for example) we cannot just
assume they are genetically based. Perhaps the differences occurred because indi-
viduals were raised on different diets, had different experiences, or were exposed to
different environments. How can we determine what portion of these differences is
genetic and thus is accessible to natural selection? A crossing experiment depends
upon mating individuals that differ in a particular kind of behavior and then exam-
ining the behavior of their offspring. For such experiments to be meaningful, of
course, the environment must be constant or very well controlled.
The simplest type of crossing experiment would be to use individuals that exhibit
certain behaviors that are known by previous crossing experiments to differ at
only a single gene locus. Many complex behavior changes have been produced by
artificially inducing single-gene mutations in the laboratory, but such simple sys-
tems have been difficult to identify in nature. A famous early example was Walter
Rothenbuhler’s 1960s study of two strains of honey bees that differ in their reac-
tion to a serious bacterial infection called American foulbrood. One strain is very
susceptible to the infection; the other is resistant. This difference is largely due to a
single behavioral difference: worker bees of the resistant strain consistently remove
dead larvae before the infection can spread (Fig. 1.7).
When Rothenbuhler crossed this ‘hygienic’ strain and the susceptible ‘nonhy-
gienic’ strain, the resultant offspring were nonhygienic, indicating that the genes
conferring resistance were recessive. Going further, Rothenbuhler did a series of
backcrosses to the homozygous recessive resistant line; roughly a quarter of the
resultant colonies were hygienic. Among the nonhygienic colonies produced, a third
uncapped the cells of dead larvae but would not remove the larvae. Another third
would remove the larvae, but only if the researcher removed the caps for them.
The remaining third would do neither. Thus, the behavioral and genetic results were
consistent in indicating that two loci—one controlling uncapping, the other affect-
ing removal—were involved in the difference in hygienic behavior between the two
strains (Fig. 1.8).
Fig. 1.7 Testing for hygienic behavior. Honey bees of some strains will remove diseased and
dead larvae and pupae from brood combs infected with American foulbrood so rapidly and thor-
oughly that they effectively control the disease. (left) To test his or her breeding stock, a beekeeper
can pour 250–300 ml of liquid nitrogen through an open-ended soup can, freeze-killing the brood
in this spot, and then after 48 hours, count the number of dead brood the bees have removed. (cen-
ter, right) The response shown in the center frame indicates a lower level of hygienic behavior in
this colony than that of the colony at the right
Parents F1 hybrids Backcrosses (n=29)
Females showing Non-hygienic generation F1 males X original hygienic females = ?

hygienic behavior
Offspring prediction:
If 1/2 hygienic: 1 gene
If 1/4 hygienic: 2 genes

X
If 1/8 hygienic: 3 genes
Males that do not
Offspring results:
6 hygienic
9 uncapped
14 non-hygienic (6 of these would

remove the dead, but can’t because they
don’t uncap them first)
(Conclusion: (Conclusion: Hygienic behavior due to (Conclusion: Two genes are involved,
Difference in the one or more recessive genes) one for uncapping and one for removal)
two lines)
Fig. 1.8 Determining hygienic behavior genetics. Experimental crosses and backcrosses done
by Walter Rothenbuhler in the 1960s suggested that two recessive genes were responsible for
hygienic behavior; later studies have indicated that additional genes are also involved
Rothenbuhler’s model has been widely cited as an instance where a natural

behavior is controlled by a few Mendelian genes of large effect, but if it holds up
it will be the only known example of naturally occurring variation in a complex
behavior of a higher organism that is under such strict genetic control. When Marla
Spivak and two Australian colleagues reexamined hygienic behavior in 2002 using
molecular techniques, they found evidence that the actual genetic basis is probably
more complex, and that many genes probably contribute to the behavior.
More commonly in nature a trait will appear to vary along a continuum; these
are often called quantitative traits. The influence of a single gene is usually unde-
tectable because genes at many loci have small additive effects, and the environment
may have a large influence upon the trait’s expression. With such polygenic systems,
statistical approaches are necessary to estimate the roles of inheritance and environ-
ment; the principal one used is analysis of variance, a powerful tool of behavioral
genetics.
A second important way to measure the degree of genetic involvement is through
selection experiments, which involve choosing only a certain part of each gener-
ation to become parents of the next generation. By such ‘artificial selection’ the
population becomes altered through time for the quantitative character under con-
sideration, in a direction that depends upon which part of the population is chosen
for succeeding generations (Fig. 1.9).
Fig. 1.9 Choosing one’s (insect) parents. Three different modes of selection yield quite different
characteristic means and frequency distributions in the subsequent generations. Ordinate indicates
the frequency of individuals in the population. Abscissa indicates variation for the quantitative
character being considered, as expressed in some linear metric measure. Directional selection is
the most common mode
With artificial selection of the type called a ‘truncation selection experiment’ a

three-step approach is used. In the first step, the trait is measured in all the individ-
uals of the population, and the numbers are averaged together to obtain a mean. In
the second step, the variation in measurements for that trait is truncated, or cut off,
at some value, and a new mean is calculated for just those individuals that measure
below that value. In the third step, only those selected individuals in the truncated
group are allowed to breed, and the mean score for their offspring (called the ‘filial
generation’) is determined.
These measurements are all we need to determine what proportion of our
observed trait variation can be attributed to genetic variance. As a result of our
truncated selection experiments, two measures can easily be determined empir-
ically (Fig. 1.10). One is the magnitude of the response (R)—how much the
mean values differ between the parent and offspring generations. The other is
the selection differential (S)—how much the chosen parents deviate from the
whole population in their mean value for the trait under selection. Dividing
R/S gives a very useful measure called heritability (h2 ). It is defined as the
degree of genetic determination of the variability that is present in the sam-
ple population, for the kind of behavior observed and for the precise method of
observation used.
Fig. 1.10 Determining heritability. See text for further explanation
What would happen if artificial selection were attempted on an inbred homozy-

gous strain? In this case, none of the variation present has a genetic basis. Thus,
the population mean will not significantly change, regardless of the sort of selection
being practiced. The mean of the filial population will not differ significantly from
that of the parental population, regardless of the selection differential. Therefore, h2
would be 0.
What if the filial population turned out to have a mean that was more or less iden-
tical to that of the selected sample? This would indicate that the trait was completely
under genetic determination, and h2 would be equal to 1. Most cases, of course,
lie somewhere between these two extremes. Studies within evolutionary biology,
population genetics, and animal behavior have found that many traits show low
(0.1–0.1) to moderate (0.1–0.4) heritability.
As a general rule, most arthropod behavior is highly stereotyped and will tend to
yield rather high heritability measures relative to those of vertebrates. For example,
selection experiments on milkweed bugs under laboratory conditions shifted the
age at first oviposition from 63.5 days to 16.7 days in seven generations. Average
heritability from all generations was 1.18, which indicates that heritability for this
trait must be very close to 1.0, since additive genetic variance cannot be greater than
total variance.
One caution should be noted before concluding this section. After several
generations in the laboratory, artificial selection may alter the cultured insects’
behavior in ways that adapt the insect for the more simplified environment.
Therefore, behavioral data based on laboratory-reared insects may not accurately
reflect the behavior of natural populations of the same species. In fact, different
behavioral results obtained by investigators working on the same insect in separate
laboratories may be explained as strain differences arising in cultures. Few behav-
iorists maintain careful checks on the ‘quality’ of the laboratory-reared insects that
serve as their research subjects, but all should consider it.
1.2.3 The Comparative Approach
Biologists have long used the technique of comparing different species in order to
understand a trait better. If the first important development for the rigorous scien-
tific study of animal behavior was publication of the theory of evolution by natural
selection, and the second was the rise of genetics, the third almost certainly was
the development of a systematic comparative method, a development that is usually
credited to two Englishmen, George John Romanes and C. Lloyd Morgan.
The name of George John Romanes is much less well known than the names
Darwin, Wallace, and Mendel, probably because although he is generally credited
as being the first to formally use comparative methods to study animal behavior, his
motives and methods were a bit unorthodox. Romanes’ interest lay in studying ani-
mals to gain insights into the behavior of humans. Specifically, he wanted to show
that mental processes evolve in a continuous way from lower to higher forms, with
humans at the top of the list. Without much direct evidence, Romanes constructed
an elaborate scale for emotional states. Worms were lowest because he judged them
capable of feeling only surprise and fear. Insects were higher, being capable of var-
ious social feelings and curiosity; fish showed play, jealousy, and anger; reptiles
displayed affection; birds showed pride and terror; finally, various mammals were
credited with hate, cruelty, and shame.
A British psychologist, C. Lloyd Morgan was also interested in this field, which
he called ‘mental evolution’. Dismayed by the ways in which people like Romanes
relied on inferences and anecdotes rather than recorded facts or direct behavioral
observations, Morgan developed and promoted the use of the observational method:
only data gathered by direct experimentation and observation could be used to

make generalizations and develop theories. He also developed an approach called
‘Morgan’s canon’. As Morgan envisioned it applying to his ‘mental evolution’, his
canon stated that an action should never be interpreted as the outcome of the exer-
cise of a higher mental faculty if it could be interpreted as the exercise of one that
stands lower in the psychological scale. For example, an animal’s behavior was only
to be considered ‘conscious’ if there was no simpler explanation.
Morgan’s canon was a special case of a common principle in medieval philoso-
phy called Occam’s (or Ockham’s) razor or the principle of parsimony. These days
in many different areas of study it is interpreted to mean something like ‘the sim-
pler the explanation, the better’ or ‘don’t multiply hypotheses unnecessarily.’ At its
broadest interpretation, the principle of parsimony is sometimes interpreted to mean
that one should always seek out the simplest explanation for observed facts, and
whenever possible one should reduce complex hypotheses to their simplest terms.
Doing so can make a problem easier to study and easier to explain to other peo-
ple. However, although this ‘law’ has become almost axiomatic in animal behavior
studies, there also are times when it can lead to interpretations that are not just
simple but simplistic and misleading because they do not account for all the facts.
Likewise, when faced with two competing theories, we cannot just assume that the
simpler theory is correct and the more complex one is false, even though we know
from experience that, more often than not, the theory that requires more complicated
machinations is more likely to be wrong. It is important to remember that a more
complex theory competing with a simpler explanation can be temporarily put aside,
but it cannot be thrown onto the trash heap of history until it is actually proven to
be false. As Albert Einstein is famously paraphrased, ‘Explanations should be as
simple as possible, but no simpler.’
As used today, the idea and logic of the comparative approach is to look at
the occurrence of a particular behavioral trait in both closely related and distantly
related species. If hypotheses and predictions about the adaptive value of a behav-
ior for one species are supported, then convergent evolution could result in similar
behavior in quite unrelated species subject to similar selective pressures (see wasp
prey carriage in Section 1.3.1). The converse is that behavior traits expected due
to shared ancestry may in some cases be overridden due to novel selection pres-
sures arising, resulting in divergent evolution. As used today, application of the
comparative method relies on well-supported phylogenies (see Section 1.3.2), but,
historically, phylogenetic rigor was not so important.
1.2.4 Conceptual Pitfalls
When discussing the conceptual frameworks that inform insect behavior, it is impor-
tant to acknowledge some other hazards that need to be watched for and avoided.
One set of traps arises from the nature of language itself.
All behavioral study rests on the bedrock of detailed description. However,
human language arises from human experience and thus we lack the words to
describe the behaviors of another species in a way that is fully divorced from our
own experiences. For one thing, almost all descriptive terms possess inherent human
connotations that often imply human purpose and/or motivation. Attributing human
characteristics to other species of animals is termed anthropomorphism (Fig. 1.11).
Even though this practice is generally unwarranted, anthropomorphic terms catch
people’s attention better than blander text does, so they are often used to describe
animal behaviors, particularly in items written for the general public. However,
when a scientist uses or encounters an anthropomorphic term, it should be with
full understanding that an interpretation is being made of the motivations of the
animal and that any such interpretation is highly suspect. Descriptions of a cricket
‘happily’ chirping on the hearth or a ‘fearful’ cockroach scurrying across the floor
may be commonplace in nature books for children but have no rightful place in the
scientific study of behavior. We simply don’t know what the sensation of happiness
is to a cricket or fear to a cockroach.
Fig. 1.11 Casting insects in human terms. Anthropomorphism has been common since the
dawn of human experience. A familiar example to most people is the story of the grasshopper and
the ants, which goes back at least to Aesop (620–560 B.C.), a slave and storyteller who lived in
Ancient Greece
A second difficulty with language use, more insidious and surprisingly pervasive
in the field of animal behavior, is the problem of teleology. A form of anthropomor-
phism, teleology is the doctrine that the processes of nature are directed toward some
discernible ‘goal’. In effect, teleology endows animals with motivations similar to
those that humans might show under similar circumstances. Much human behavior
is goal directed, or purposive; for example, when hungry, one goes to the table or
refrigerator in anticipation of food. However, most animal behavior is directive, that
is, the outcome of the response is not foreseen by the organism. Thus, there is a
confounding of ends and means in statements such as ‘bees visit flowers because
they want nectar.’ The bee is brought to the nectar by its response to the stimulus
provided by the flower, not because it consciously anticipates the outcome of its
actions.
A third language-related problem is that, because most behavior is described in

terms of its end result, the very act of labeling a particular observed behavior tends
to color subsequent interpretations of it. For example, labeling a chemical secreted
by a female moth as a ‘sex attractant’ is done because the end result of her secre-
tion is that males arrive and attempt to mate with her. This label suggests that such
chemicals provide directional clues of use to an approaching insect orienting over
some distance, but such cues may have not been demonstrated. Furthermore, use of
the label ‘sex attractant’ potentially masks appreciation of a whole congregation of
other behaviors by the respondent such as orientation to wind and light.
In a larger context, it is customary to label communication signals by apparent
function such as sexual, aggressive, alarm, etc. However, it is often difficult to dis-
tinguish the responses that such categories of signals will actually elicit because
the environment or ‘context’ of a signal may alter its message (see Chapter 6). For
example, the same chemical in harvester ants may elicit alarm under one set of cir-
cumstances but elicit approach in another situation. A second problem is that the
respondent’s first detectable behavioral response to many signals is just to change
positions; as it moves, it enters new stimulus situations. In analyzing the insect’s
ultimate behavior, it is difficult to separate the influence of these new situations
from that of the initial signals.
Even labeling behavior in order to catalog it can color one’s observations. A
widespread system for classifying communication sets its categories as visual,
acoustic, chemical, tactile, or electrical, for example; for convenience, a similar sys-
tem is used in this book. However, for insects this is admittedly simplistic. An insect
may be receiving information simultaneously in a number of sensory modes, and in
many cases the total message and its specificity may depend upon receipt of all these
different channels together.
A different set of conceptual pitfalls concerns the theory of evolution by natural
selection. Although this theory remains the single best unifying explanation for the
ultimate causes of the traits that organisms possess, it is important not to assume
automatically that the behaviors one sees represent some sort of optimized state
and are the best of all possible choices, even though simplistically it might seem
that over time, evolution by natural selection would result in every organism being
homozygous for the most favorable alleles.
Clearly, this is not what happens. The real picture is much more complex, for at
least five reasons. First, not all traits are heritable. Many behaviors may be caused
entirely by actions of the environment; this can be particularly true in animals that
are capable of learning.
Second, natural selection is not the only evolutionary force as work. For exam-
ple, as noted above, genetic drift and gene flow both can influence the evolution of
traits. Another driving force in the evolution of animal behavior is the concept of
an evolutionary stable strategy, a term derived from game theory; it recognizes that
animal behavior is governed not only by what is optimal, but also by what other
strategies are found in the population, and at what frequency.
Third, selection pressures can change as a result of many factors. Environments
change over time; what might be good behavior today, for example, might not be
the best behavior in 10,000 years or even in 10 years. A previously beneficial trait
may become deleterious, or vice versa, under new conditions such as a changed or
different environment.
Fourth, not every trait can be optimized by natural selection, because traits are
linked together in many ways. In some situations a heterozygote has greater fitness
than either homozygote. In other situations, epistasis—where genes at different loci
interact to produce the phenotype—results in selection acting differently on genes
in different genetic backgrounds. Many genes affect more than one trait; one allele
may produce a positive effect on one trait but a negative effect on another. Thus, by
necessity, tradeoffs must occur. Furthermore, genes can also be physically so close
together on the same chromosome that they tend to be inherited together, making it
difficult for natural selection to optimize each trait independently.
Fifth, evolution does not have a predetermined goal for any species. There often
may be cases where an organism simply cannot evolve a particular trait that would
suit it well, because it does not have the relevant genes in its gene pool to do so.
Mutations are random; they do not arise to fulfill a need.
If these precautions can be kept in mind, adaptation and natural selection can
provide a powerful framework within which to generate sound, testable hypotheses.
1.3 Phylogeny’s Role
Phylogeny, the presumed history of ancestry of a group of taxa, provides a strong

evidence line to help decipher the evolution of a trait. In a sense, phylogenetics can
be thought of as ‘evolutionary geneology’. Its tools include the twin fields of tax-
onomy and systematics, the sciences involved in finding, describing, and naming
organisms by an agreed-upon set of rules, and then classifying them into increas-
ingly broader categories that are based on shared features that presumably reflect
evolutionary relationships.
Incidentally, we have these fields to thank for introducing the useful general term
taxon (plural taxa), meaning any taxonomic unit without specifying its rank. Thus,
a species is a taxon, but so is a group of species or an entire order. A related term,
clade, is a group of organisms that includes an ancestor and all its descendents (see
Section 1.3.2).
As we have seen, much of behavior has a genetic basis, and traits with a genetic
basis are capable of evolving over time. But what is the evidence for behavioral
evolution? To answer this, using the comparative approach is a necessity, but in
addition, one must consider phylogeny, which involves both microevolution and
macroevolution.
1.3.1 Microevolution and Macroevolution
Long before it was ever given its name, people have recognized the reality of
microevolution (genetic changes within populations or species) because of two
1.3 Phylogeny’s Role 29
easily recognized phenomena: domestication, and observations of natural selec-

tion in the field. Dog behavior is probably the best-known example of evolution
under domestication, but even in Darwin’s day any farmer worth his salt recognized
that the animals under his care changed from one generation to the next in ways
that could be guided by selective breeding. Likewise, although long-term studies
were not always possible, differences often were quite obvious when populations of
the same animal species from different environmental conditions were compared.
Knowing the examples would be familiar to his readers, Darwin went to some length
to address such domestications and natural changes as background for his new ideas.
While microevolutionary patterns can often be directly seen, macroevolution
(change of evolutionary patterns recognizable above the species level) usually oper-
ates over such a long time frame that it is more apt to depend on indirect evidence
such as the fossil record and/or a reconstruction of presumed relationships between
similar-seeming species. Although one might think that the fossil record would not
be much help for behavioral studies, given that behavior itself does not leave fos-
sil remains, sometimes a great deal can be inferred from morphological structures,
tracks, and artifacts such as preserved nests and burrows. However, in general the
fossil record does not usually shed much light on how a behavior pattern evolved;
rather, it is more apt to demonstrate the antiquity of the particular behavior that
is preserved. Some examples include an entire colony of weaver ants from the
Oligocene in Africa, fossil dung balls fashioned by scarab bettles in the Lower
Oligocene, and fossil lepidopterous leaf mines from the early Eocene of Wyoming.
The fossil ant nest demonstrates that distinct morphological castes in ants evolved
very early and that social life in ants has apparently changed but little over the past
60 million years. The fossil dung balls similarly demonstrate scarab beetle nesting
behavior (see Chapter 10) to have been at a fairly advanced level very early in the
history of the group.
Species that share a recent common ancestor tend to resemble one another in
many respects simply because they share a common ancestor. Behavioral differ-
ences and similarities within a group of species generally reflect the phylogenetic
relationships as determined morphologically or genetically.
Perhaps one of the most elegant examples supporting this assertion is found
among the cockroaches, an ancient but highly successful group. Cockroaches
reached their highest development during the late Carboniferous, and have shown
relatively little morphological change in the intervening 250 million years; a fossil
roach generally resembles a recently swatted roach except that early forms tended
to have visible exerted ovipositors much like those of present-day katydids.
If you have ever tried to swat a cockroach, you know the creature’s main defense
is extreme proficiency at quick getaways (see Case Study 2.1). What you may not
appreciate is that the common household varieties of roach represent only a minis-
cule proportion of some 5,000 living blattid species that have invaded nearly every
conceivable ecological niche from desert to aquatic habitats. Despite a compara-
tively good fossil record, the classification and phylogeny of these thousands of
species were subject to dispute by various authorities. As a result, three classifica-
tions arose, based on different sets of morphological characteristics. Rehn’s system,
based primarily on wing venation (thereby excluding the many wingless forms),
classified cockroaches into five families and 17 subfamilies. Two other cockroach
classification systems, Princis’ system and Bey-Bienko’s system were based on
diverse and apparently unrelated structural differences. Princis divided the group
into four suborders and 28 families, and Bey-Bienko recognized just three families
with a total of 15 subfamilies.
Behavioral characteristics and internal morphology were given little attention
in any of these early schemes, yet cockroach reproductive biology is exceedingly
diverse. Although it is unusual to find more than one mode of oviposition behav-
ior in a particular insect group, cockroaches display an almost complete spectrum.
Like most insects and birds, females of many cockroach species lay eggs in packets
(oothecae). Such reproduction, where fertilized eggs develop outside the female’s
body, is termed oviparity. Other cockroach species exhibit ovoviviparity—the eggs
are first extruded and then retracted into a brood sac where the embryos absorb
water from their mother until they mature; then the ootheca is again extruded and
the nymphs hatch and drop free from their mother. A third type of reproduction,
viviparity, is found in still other roach species. This oviposition method is super-
ficially similar to the previous type, but when first formed the eggs lack sufficient
yolk to allow complete development, so both nutrients and water must be absorbed
during embryonic development in the brood pouch.
When the species having these three types of reproduction are superimposed on
the three older classifications they intermingle throughout the suborders, families,
and subfamilies rather than sorting out into any consistent pattern. In 1964, this
puzzling fact prompted Frances McKittrick to undertake a new look at cockroach
classification. For comparative study, she chose four character systems: female gen-
italia, male genitalia, the proventriculus (a portion of the intestinal tract specialized
for grinding food, also referred to as the gizzard), and oviposition behavior. Upon
analyzing her results, McKittrick was able to recognize two phyletic lineages based
primarily on reproductive behavior. In one lineage all species remained oviparous,
undoubtedly the ancestral form of reproduction. This group (Blattidae) is exempli-
fied by the notorious American roach, Periplaneta americana. The other lineage
encompassed the ovoviviparous and viviparous species, including the equally noto-
rious German roach, Blattella germanica. In the years since her work was published,
McKittrick’s conclusions have gained widespread acceptance and support from a
variety of other studies (Fig. 1.12), and she is still recognized for being the first
researcher to stress reproductive behavior as an important trait in the evolution of
cockroaches.
The most recent systematic studies of the group, including both extensive mor-
phological and molecular analyses, still leave some phylogenetic details unresolved
and uncertainties remain. The one striking and fundamental difference between the
earlier and current phylogenies is that the increasing weight of available evidence
has revealed that the termites (Isoptera) properly fall within the cockroach lineage
as a sister group to the oviparous Cryptocercidae, a group of uncertain affinity in the
earlier classifications.
When comparative behavioral studies uncover a seeming progression in one
or more characters within a set of related species, it is tempting to conclude that
Fig. 1.12 Classifying cockroaches. (left) Three older competing schemes (see text) based solely
on external features observed in museum specimens gave way to one in 1964 based primarily on
female reproductive behavior. (right) The currently accepted higher classification and phylogeny of
the Blattaria based on a much more extensive data set, including molecular data, also incorporates
the termites
evolution has actually proceeded through such a series. However, such a conclusion
should be tempered with caution. The fact that intermediate forms are possible and
do exist indicates only that evolution might have proceeded through a similar (but
probably not identical) series. In the absence of a fossil record or some compelling
logical argument, it does not even indicate the direction of the evolution of the trait
in question.
Prey transport, for example, varies considerably among predatory solitary wasps
(Fig. 1.13). Many species carry paralyzed prey back to the nest in their jaws; others
always hold the prey with their middle legs, and still others use their hind legs. Some
fly hunters even carry the prey impaled on their stinger, but perhaps the most remark-
able prey-carrying adaptation is found in a crabronid wasp, Clypeadon laticinctus,
that transports paralyzed ants on a specialized ‘ant clamp’ formed by the modi-
fied apical abdominal segments. While prey carriage seems to be constant among
all members of a particular morphologically defined genus, there appears to be no
correlation of prey carriage type with wasp phylogeny beyond this, nor with nest
type or phylogenetic position of the prey. This appears to be a case where similar
selective pressures have molded the evolution of prey carriage mechanisms.
What are these selective factors? In 1962, Howard E. Evans suggested some
answers. In the soil-nesting species, mandibular prey carriage tends both to obstruct
the major sense organs and to impede the wasps’ ability to dig into their nest
entrance. Furthermore, species that use their mouthparts in prey transport either tend
to leave the nest open between foraging trips or use such large prey that they must
drag them back to the nest; if they were to close the nest between prey trips, they
would be forced to lay the prey down temporarily while they reopened it, thereby
leaving the prey vulnerable to attack by predators and parasites. Shifting the prey
beneath their body and back from the head or even clamped or impaled on the stinger
Fig. 1.13 Carrying prey. (Top) A female of the Australian eumenid wasp, Pseudabispa
clasps her caterpillar prey in her jaws, supported by her middle and hind legs as she flies to
her nest
would increase their own efficiency and better protect their offspring from intruders.
The relative rarity of abdominal clamp prey carriage mechanisms suggests that the
advantage of having all three pairs of legs free may be offset by the relatively greater
risk of exposure of the prey to parasites, especially to attacks by parasitic satellite
flies (Fig. 1.14).
Another important line of evidence is provided by detailed studies of adap-
tive radiation (the rapid evolution of new lineages). Isolated areas such as oceanic
islands are particularly good places to look for examples of adaptive radiation; only
a few species may manage to invade such remote areas, and in the relative absence
of interspecific competition, new forms may rapidly evolve. The famous case of the
adaptive radiation of Darwin’s finches on the Galapagos Islands is almost certainly
the world’s best-known example. The Bembix sand wasps of Australia provide a
behavioral example of adaptive radiation. Although flies are the prey of nearly all of
the 300-plus species of Bembix throughout the world, 11 of the 28 Australian species
for which biological data exist have switched to other prey (Fig. 1.15). This unique
radiation in prey preferences in Australia is presumably related to the absence of
related genera of sand wasps that occupy these food niches elsewhere in the world.
Other factors may also be involved, such as scarcity of usual prey in some habitats,
especially arid ones. Species such as B. variabilis may be in the earliest stages of
splitting into two distinct species because different populations seem to have spe-
cialized on quite different prey. Throughout most of Australia B. variabilis use flies
Fig. 1.14 Looting its host. A satellite fly, Senotainia trilineata (Sarcophagidae), rests on a nest-
marker nail. Such flies typically follow just behind prey-laden wasps like a satellite, and attempt to
swiftly dart in and deposit their live larvae on the wasp’s prey as the wasp enters its nest. Smaller
prey slung under the wasp’s body may be less vulnerable to such parasites
as prey, but a population discovered in the far northwest of the continent seem to
exclusively use damselflies as prey.
1.3.2 Phylogenetic Systematics and Cladistics
To deduce the evolutionary direction of apparent behavioral progressions, a useful

procedure is to plot the behavioral evidence against a ‘family tree’ (see Figs. 1.2 and
1.12). Recognizing that closely related species are more likely to share traits than
are distantly related species, Darwin himself introduced the metaphor of a tree to
describe the relationships among taxa.
Many of these trees have their roots in traditional taxonomy and systematics, and
owe an intellectual debt to Carolus (Carl) Linneaus, the Swedish botanist, physician
and zoologist who laid the foundations for the modern scheme of binomial nomen-
clature. Linnaeus also developed what became known as the Linnaean taxonomy.
This system of scientific classification, still widely used in the biological sciences,
assigned every organism its position within a nested hierarchy based on observ-
able characteristics. Thus, above the basic level of genus and species, Linnaean
classification placed groups of organisms into families, orders, classes, phyla, and
kingdoms.
The underlying details concerning what are considered to be scientifically valid
‘observable characteristics’ have changed with expanding knowledge (for example,
DNA sequencing was unavailable in Linnaeus’ time), but the fundamental principle
remains sound. Some taxonomists point out, however, that Linnaean classification
can be misleading because it implies that different groupings with the same rank
Fig. 1.15 Undergoing adaptive radiation. Known prey preferences of 28 Australian Bembix
species. Most species hunt only flies (Diptera), the prey used exclusively by studied species of this
genus elsewhere in the world. However, seven species are specialists on three other orders while
three species (variabilis, kamulla, and moma) hunt both flies and species from a different order,
and one species (allunga) takes prey from three very different insect orders
level are equivalent. For example, although cats (Felidae) and orchids (Orchidaceae)
are both ranked as family level groups in Linnaean classification, the two groups are
not comparable. One has a longer history than the other and the extent of diversity
within each group differs dramatically.
In response to these concerns, scientists have developed other ways to show evo-
lutionary relationships in a tree-like or bush-like form. The best-known alternative
is based on an approach called cladistics. Cladistics began in the 1950s with the
work of the German entomologist, Willi Hennig, who referred to it as phylogenetic
systematics.
Cladistics differs from other taxonomic systems in its strong focus on evolution
rather than similarities between species, and in its heavy emphasis on objective,
quantitative analysis, particularly using DNA and RNA sequencing data. The cladis-
tic approach relies on identifying monophyletic groups, clusters of taxa that share a
more recent ancestor with each other than they do with other groups. Polyphyletic
groups, by contrast, are ‘mixed bags’ that may seem to share similarities but actually
arose from two or more different ancestors. This approach has two main differences
from the Linnaean system. First, phylogenetic classification tells you something
important about the organism: its evolutionary history. Second, phylogenetic clas-
sification does not even attempt to rank organisms. In contrast to the traditional
Linnaean system of classification, phylogenetic classification names only clades, or
groups that include both a single ancestor and all its descendents.
The most important assumption in cladistics is that characteristics of organisms
change over time, because it is only when characteristics change that different lin-
eages or groups can be recognized. The original state of the characteristic before
it changed is called plesiomorphic; the new state after the change, apomorphic.
Though some people use the term ‘primitive’ instead of plesiomorphic and ‘derived’
instead of apomorphic, biologists generally avoid using these words because they
have inaccurate connotations. It is all too easy to think of primitive things as being
simpler and inferior, but in many cases the original plesiomorphic state of a charac-
ter is more complex than the changed, apomorphic state. For example, as they have
evolved, many cave insects have lost effective vision, and many island-dwelling
species have reduced wings.
Instead of a Linnean-based evolutionary tree, cladists construct (or in the words
of some, ‘reconstruct’) a phylogenetic tree. This is a diagram intended to repre-
sent the evolutionary history of modern taxa. The trunk, or earliest ancestor, gives
rise to limbs that give rise to branches that terminate in twigs. The tips of the
twigs represent species that are alive today. Sometimes branches fall off, represent-
ing extinction. (Some scientists prefer to think of a phylogenetic bush with many
branches, rather than a tree with a single trunk; both terms are used.)
How does one go about constructing such a phylogenetic tree? No diagram can
take every trait into consideration, so the first thing one must do is identify the par-
ticular characters (inherited traits) one will consider, and then describe the ways they
vary, i.e. their character states. Traits can be almost anything that has an assumed
genetic base and that can consistently be measured. Historically, trait measurements
were morphological or anatomical, and many were gathered from fossil evidence.
Sometimes traits corresponded to a single structure (such as hind leg lengths in
various grasshopper species). In other cases, traits were expressed as ratios that
described the measurements of a set of related structures (such as the tibia: femur
ratio for various grasshopper species). These ratios were helpful to account for envi-
ronmental influences on factors such as absolute body size. Behavioral (see Section
1.3.3 below) or physiological traits, such as reproductive modes (see Fig. 1.12), can
be useful as well. With the advent of molecular genetics, many traits are measured
as DNA sequences, and molecular genetic maps are used to help build phylogenies
by comparing and contrasting sequences across different species. Hopefully, one’s
wise choices will lead to trees that approximate reality. However, there is always
the possibility that choosing different characters might have led to different results.
Cladistics can produce profoundly complex analyses, so it is important to remem-
ber that however fancy an analysis may look or how many molecular data have gone
into it, the result is still a tree or bush that represents a hypothesis. New and better
data could change the outcome and support a different hypothesis about the way that
the organisms are evolutionarily related.
Next, one must judge which species are most closely related to each other and
thus are descended from a more recent common ancestor. This involves decisions
concerning homologies—characteristics assumed to be shared by species through
descent from a common ancestor rather than being a product of a similar environ-
ment. As theories go, homology is relatively simple, but since it involves judgments
about past events that can never be known with absolute certainty, in actual practice
it sometimes can be quite controversial. For whatever character set one has chosen,
the most similar pairs are considered the most closely related; more differences are
taken to mean a more ancient split in ancestry, based on the fact that it takes more
time for multiple mutations to occur. Unfortunately, all this is assumptive. History
is not something we can see. It happens only once, and only leaves clues behind for
those who attempt to reconstruct evolutionary history.
Whereas evidence for homology can help one make sound phylogenetic deci-
sions, a related possibility can sabotage it. Evolutionary convergence is the tendency
of different clearly unrelated species to develop similar characteristics in response
to a set of environmental conditions. For example, sociality appears in both ants
and termites, though they are only distantly related (see Chapter 10). Although the
selection pressures under which sociality evolved may have been similar, the trait
was not shared with a common ancestor.
After determining the characters and examining each taxon to establish its char-
acter states, some types of cladistic analysis require that one decide the polarity, or
direction of evolution, for each character; this can take some work. Then the taxa
must be grouped by synapomorphies—‘changed’ character states that they share. If
conflicts arise, they must resolved by some clearly stated method; the usual one is
parsimony (see Section 1.2.3).
Finally, when it is time to actually construct the phylogenetic tree, ideally it
should follow two rules. First, all taxa go on the endpoints of the tree, never ‘lower
down’ at nodes as Linnaean trees would. Instead, all nodes must have a list of
synapomorphies that are common to all taxa above the node (unless the charac-
ter is later modified). Second, all synapomorphies must appear on the tree only once
(unless the character state was derived separately by parallel evolution). In practice,
‘hybrid’ trees often appear.
Once constructed, a phylogenetic tree is a very helpful tool that can be employed
to make predictions about fossils or about poorly studied species. It also can help sci-
entists learn about the order of the evolution of a particular trait, a complex feature,
or observed diversity.
1.3.3 Behavior and Speciation
Can behavioral patterns be used like morphological or genetic characters in con-

structing a phylogeny? At first the idea was hard to sell. Many scientists said
no; behavior is too variable, and is under too much environmental influence.
However, one of the cornerstone beliefs of ethological research has always been
that behavioral patterns can be treated like morphological ones, and this included
showing homologies. Behaviors also could act as barriers that interrupt gene flow,
thus starting in motion the process of speciation, the division of a species into two
or more new biological species that are genetically unique.
Differences in mating behavior often constitute the strongest sort of species-
isolating devices. For example, consider the fireflies of eastern North America. This
is a confusing group in which the males show almost no differences in structure
or body coloring but much variation in flash pattern used to attract females dur-
ing courtship. This problem attracted the attention of Harry S. Barber; after much
study, in 1951 he published a description of 18 species of Photuris fireflies classified
mainly on the basis of male flash patterns. Ten of these were named as new since
they had not been previously recognized by morphology alone. Similarly, when
studying the common smaller fireflies of the United States, Jim Lloyd found several
such ‘hidden species’ in the related genus Photinus. First recognized by consis-
tent differences in flash signals (Fig. 1.16), these fireflies later were found to differ
Fig. 1.16 Sorting species by flash. Using behavior to inform decisions about speciation, Lloyd
compared flash patterns of Photinus fireflies. Differences among several North American species
in the male flash signals (left) and female response flashes (right; note different time scales) are
shown at typical field temperatures (signal timing is temperature dependent). Female responses
are timed from initiation of the last pulse in the male flash pattern (open bars indicate optional
responses)
in minor morphological details. In many places, two or more species of Photinus

fly together but are prevented from interbreeding by their specific courtship pat-
terns. The males fly at different heights and in different flight patterns (see Fig. 7.2);
their flashes differ in length, number of pulses per flash, and sometimes in color or
intensity of the light.
One cannot construct a tree using behavioral characters and then use it to test
hypotheses about those characters, of course; that would be circular reasoning.
However, a very useful approach can be to construct a phylogenetic tree using mor-
phological and/or genetic traits, and then map behavioral characters onto it and see
if they match. By examining the tree, it is also sometimes possible to determine the
evolutionary direction of a behavioral trait.
One example of this practice involved Herbert Ross’s study of the aquatic insect
order Trichoptera. The caddisflies make up a very large insect order, with about
10,000 described extant species distributed among 45 families. Their aquatic lar-
vae are found in almost all freshwater communities, where they occupy a diversity
of microhabitats and trophic niches. Larvae of certain caddisfly families construct
a fixed retreat; those of others construct a portable case; and still others are free
living. In 1967, Herbert H. Ross provided the first modern phylogenetic hypothe-
ses of subordinal and superfamily relationships in the Trichoptera. Superimposing
larval behavior upon a morphologically based phylogeny indicated to him that
fixed retreat construction was the ancestral behavior and portable case building
the derived state, with a series of steps leading from one to the other. However,
through the years after his work, his view was challenged by at least four very differ-
ent morphologically-based hypotheses of the relationships between the Trichoptera
suborders (Fig. 1.17 top). In 2001, a research team of cladistically-oriented entomol-
ogists began their own analyses of the Trichoptera that included both a reevaluation
of others’ morphological data sets and newly collected molecular sequence data.
The first of its kind for the Trichoptera, the project is large and ongoing, but satis-
fyingly, the phylogenetic relationships among the suborders so far seem to closely
resemble those proposed by Ross over 40 years ago on the basis of larval behavior
(Fig. 1.17 bottom).
1.4 Questions and Perspectives
In various versions of an old tale from India, a group of blind men (or men in the
dark) each touch one different part of an elephant to learn what it is like. When they
compare notes on what they felt, they find they are in complete disagreement. The
story is used to indicate that reality may be viewed differently depending upon one’s
perspective, suggesting that what seems an absolute truth may be relative due to the
deceptive nature of half-truths.
Over the years, behaviorists and non-scientists alike have asked questions about
almost every aspect of animal behavior, while trying to avoid being like these fabled
blind men. Even today, such attempts are sometimes successful, sometimes not.
Different ways of looking at a behavior clearly lead to different answers.
1.4 Questions and Perspectives 39
Fig. 1.17 Evaluating Trichoptera relationships. (top) The original hypothesis of Ross about
Trichoptera (caddisflies) phylogeny compared to subsequent worker’s hypotheses of subordinal
relationships based on differing character sets that show a lack of consensus. (below) A recent
composite phylogeny of family group taxa of Trichoptera proposed by K. M. Kjer and colleagues
incorporating data from molecular analyses strongly supports five monophyletic clades. With
respect to the relationships among the three suborders it closely resembles that proposed by Ross
on the basis of caddisfly behavior
Table 1.1 Tinbergen’s four questions: Basic approaches to the study of animal behavior
Types of Types of explanation

questions
Proximate—how Causational—mechanistic Developmental—explanations based on

organisms explanations of individual ontogeny starting with DNA and
work responses to stimuli, both progressing through life stages from
internal and external immatures to adult
Ultimate—why Functional—explanations of Evolutionary—explanations that attempt
organisms are adaptive significance based on to reconstruct the phylogenetic history
the way they presumed selective advantage of a lineage often using a comparative
are approach
In a classic paper the pioneering ethologist Niko Tinbergen pointed out that
essentially all research questions address one of four major concerns about a behav-
ior: the nature of the immediate stimuli that evoke it, the way in which it develops,
the survival function it has, and how it has evolved (Table 1.1).
1.4.1 Proximate and Ultimate Analyses
While Tinbergen’s Four Questions have become axiomatic in behavior study, others
viewing the list realized that they could be grouped into just two kinds of analyses:
proximate analysis and ultimate analysis. Proximate analysis focuses on the imme-
diate causes of a behavior—those that occur during a given organism’s life. Ultimate
analysis is defined in terms of the evolutionary forces that have shaped a trait over
time. For example, a group of scientists might ask why house flies are drawn to
sugar. Those favoring a proximate explanation might consider factors such as the
way sugar acts on the fly’s sensory system (Tinbergen’s ‘immediate stimuli’) or per-
haps the physiological events that occur in the fly’s brain (‘development’). Other
scientists in the group might approach the same question from a different analyti-
cal viewpoint, an ultimate analysis in terms of the ways that sugar attraction might
increase flies’ chances of surviving long enough to reproduce (‘survival function’)
or the selective forces in the evolutionary past that favored sugar-preferring individ-
uals (‘evolutionary history’) and whether relatives of the house fly display similar
responses to sugar.
It is sometimes said that proximate questions generally ask ‘how’ and ‘what’
whereas ultimate questions tend to ask ‘why’ but this simplistic division can be con-
fusing. It may also seem to imply that proximate analyses are less important than
ultimate analyses, but that is not the case. There are fundamental links between the
two analyses. Proximate analysis, by shedding light on neurobiology, endocrinol-
ogy, molecular genetics, and so on, illuminates the variation available for natural
selection to act on. At the same time, understanding something about the natural
selection pressures that have acted on a behavior can help in designing the best way
to do a proximate analysis. While each explanation of the fly’s attraction to sugar
is perfectly reasonable, none is inherently superior to the others. However, taken
together, the different explanations and approaches to studying the question provide
a fairly complete understanding of the behavior.
1.4.2 Types of Approach
In addition to the differing viewpoints of proximate and ultimate analyses, scientists

have different intellectual approaches to a given set of observations and experi-
ments. A conceptual approach typically seeks to join together formerly unconnected
ideas in new cohesive ways. Natural history and experimentation may be taken
into account, but the broad-based concepts that result are usually not tied directly
to any specific observation or experiment. For this reason, this approach is some-
times (somewhat disparagingly) called ‘arm-chair science’. However, new concepts
tend to generate a great deal of new experimental work by others, and a new
conceptual advance can be a powerful tool that entirely reshapes the way that a dis-
cipline views itself. A clear example would be William D. Hamilton’s ideas about
kin selection and its role in the evolution of social behavior in the Hymenoptera
(see Chapter 10).
Theoreticians come at behavioral study from a different intellectual direction.
The theoretical approach entails the generation of some sort of mathematical model
to explain observations. The first major studies applying this approach to animal
behavior involved foraging. Out of all the foods available to an animal, which ones
should it choose to eat? Under what conditions? To address such questions, a math-
ematical tool called optimality theory was developed (see Chapter 4). Critics of the
mathematical theory approach are wont to point out that such theories often do not
match the details of any given system. However, that is not the theoreticians’ aim.
They are not trying to mimic the natural world with their model. Instead, they are
trying to strip away the specific details to uncover the barest ingredients that will
allow them to make general predictions that might apply to many systems.
The empirical approach essentially has two bases, observation and experimen-
tation. It basically equates to the ‘scientific method’ taught in introductory biology
classes. One begins with a question, devises a possible answer with an explicit pre-
diction or hypothesis about what one expects to observe, and then uses this potential
explanation to design and then conduct relevant experiments or observations. After
collecting data in a way that minimizes the influence of other variables, one evalu-
ates the results in an appropriate way (which generally includes the use of statistical
analyses). If the actual results do not correspond to the expected ones, the underlying
hypothesis is probably false; if the evidence does match the prediction, the hypoth-
esis is tentatively accepted as true, though it can never be absolutely proved so.
In practice, of course, insect behaviorists can and do combine perspectives and
approaches in countless ways. Many aspects of behavior are examined from both
proximate and ultimate perspectives; others lend themselves more readily to one
perspective than the other. Various approaches are also used (though the empirical
approach remains the most common).
The most powerful investigations always have been the multi-faceted ones in
which research shines new light on a phenomenon from every possible direction.
A premier example is the honey bee, Apis mellifera. Because of its importance to
agriculture and its advanced social system, it has become one of the most-studied
insects in the world. In Chapter 8, a case study considers its foraging from the aspect
of communication. Case Study 1.1 considers factors that start the bee out of the hive
door in the first place, and out into the world outside.
Case Study 1.1: Foraging Onset in the Honey Bee, Apis

mellifera
Proximate causation can be considered to have four components—hormonal,

neurobiological, molecular genetic, and developmental. These areas interact
in complex ways and boundaries between them can be difficult to define, as
illustrated by a single species, the honey bee.
For centuries, honey bees have enthralled scientists and the general pub-
lic alike. Their development and colony life have been subjects of special
fascination, but perhaps most of all, observers have been entranced by their
foraging. When a new bee-pollinated flowering plant begins to bloom, some
time may pass before a single bee arrives. Soon after this initial visit, how-
ever, the plant is buzzing with visitors. Clearly, some sort of communication
is involved; deducing its nature has been one of the major accomplishments in
insect ethology (see Chapter 8). The effects of neural structures, genes, devel-
opment, and hormones have been a lesser-told story. Much of this proximate
analysis comes from the work of Gene Robinson and his colleagues at the
University of Illinois.
Honey bees have an age-based task allocation system. Younger bees gener-
ally stay within the hive, doing housekeeping tasks such as cleaning the hive
and feeding the larvae. Older bees give up these tasks and switch to search-
ing outside for pollen, nectar, water and propolis (gummy plant resins used
in nest repair). Like other insects that search for food away from their nests,
these foraging honey bees must remember and retrieve information from the
environment.
In most invertebrates where it has been studied, such spatial navigation is
linked to cell clusters called mushroom bodies (see Plate 1) located in front
of the brain. Do honey bees also rely on their mushroom bodies? Robinson’s
research team dissected bees with known ages and jobs. The researchers found
that the mushroom bodies of foragers were almost 15% larger than those of
bees that remained in the colony, but the relative volume of other nerve clus-
ters in the bees’ brains remained relatively unchanged as a function of age and
task allocation.
Was foraging the sole reason? By removing a significant number of the
current foragers from a hive, researchers induced week-old bees from the hive
to take on the forager role immediately rather than waiting until the normal
age of about 20 days old. When they examined these precocious foragers,
their mushroom bodies resembled those of the normal-aged foragers. Thus,

it seemed that foraging behavior actually changed the brain, rather than vice
versa.
How might this developmental shift be influenced at the molecular level?
A gene called period (per) affects circadian rhythms and development time
in fruit flies (see Chapter 2). Alberto Toma and his colleagues asked how per
messenger RNA (mRNA) levels in bees might influence the developmental
changes associated with becoming foragers. They measured mRNA levels of
groups of laboratory-reared bees of different ages, marked a group of their
bees that were only one day old, and added them to a natural bee colony
in the field. At days seven and 24, they recaptured these marked bees and
measured their mRNA as well. In both laboratory-raised and outdoor-living
bees, per mRNA was significantly greater in older individuals that foraged
than in younger bees that stayed home (see Fig. 1.18).
Was per mRNA higher simply because of age? To answer this, Toma
used precocious week-old foragers. When he measured their mRNA, it did
not differ from those of typical older foragers. This showed that there was
a fundamental link between per mRNA and foraging, not just a more gen-
eral correlation between per mRNA and development. Of course mRNA does
not directly cause the behavioral change; rather quantity of mRNA can be
assumed to reliably indicate the extent of gene expression, in this case the per
protein translated from the mRNA.
Fig. 1.18 Switching jobs. As they grow older, worker bees shift from hive work to foraging.
How might genes affect this behavior? Toma and colleagues measured mRNA. (left) Foraging
bees have substantially higher levels of per mRNA than younger, nonforaging hive bees. This
difference could be due to various factors such as age, behavior (forager vs. nonforager), or both.
(right) Tricking bees into searching for food much earlier than usual, Toma’s team produced 10-
day-old precocious foragers; when they were compared to normal 22-day-old foragers, there were
no statistical differences in per mRNA levels
The ways in which genes affect the switch from hive work to foraging have
received a great deal of attention (Fig. 1.18). Rapid advances in molecular genetic
technology are leading to ever more detailed glimpses into honeybee foraging. In
a large-scale 2003 study of 5,500 genes, changes in mRNA levels were found to
be associated with about 2,145 genes involved in the hive work/foraging transition.
As such large-scale genomic approaches to behavior become more common, the
challenge will be to figure out how to handle the massive amounts of data they are
capable of gathering. One way to facilitate the molecular analysis of a complex
behavior such as honey bee foraging is to dissect it into simpler behavioral com-
ponents. This is currently a very active area of research. A foraging (for) gene has
been found to influence the phototactic elements of foraging, and a malvolio (mvl)
gene influences the responsiveness of honey bee workers to sucrose and through that
response, the age at onset of foraging and the tendency to forage for either pollen
or nectar. Presumably these are only two of many genes that will be found to play a
causal role in honey bee behavioral maturation.
Chapter 2
Programming and Integrating Behavior
2.1 Introduction
Lacewing males, Chrysoperla downesi, sit among the branches of an evergreen tree,
softly drumming their abdomens against terminal twigs and needles in a long, com-
plex pattern of volleys that attracts females. Moths of the black cutworm, Agrotis
ipsilon, fly by at recorded ground speeds of between 97 and113 km/h (60–70 mph).
Foraging Cataglyphis ants scurry around upon the floor in the Sahara Desert at sur-
face temperatures of up to 70◦ C (158◦ F), scavenging upon the corpses of insects and
other arthropods that have succumbed to heat stress in this extreme environment.
Insect behavior—hidden and obtrusive, incredible and yet somehow familiar,
often admirable and at other times repulsive—has always been a source of amaze-
ment and curiosity. Even with their extra legs and inexpressive faces, insects look
and act enough like little people to make one wonder how and what they think, or
whether they think at all. What regulatory and integrative mechanisms coordinate
the expression of their behavior?
Instead of answering the question, a bit of observation only leads to many more.
A headless male preying mantis can complete a sequence of mating behavior (see
Fig. 9.16). A decapitated cricket will sing several of its song patterns when its neck
connectives are stimulated. Headless cockroaches can learn to keep a leg flexed to
avoid an electrical shock. How necessary is an insect brain?
Having followed a tail-wagging dance for the first time, novice honey bees can
fly in the indicated direction and distance. The first time it strikes, a hand-raised
preying mantis will react to exactly the same stimuli as experienced mantises do.
How much of insect behavior is controlled by instinct?
Honey bees learn to avoid a food source from which they have received an elec-
tric shock, and can remember the previous location of a displaced hive even 12 days
later. Female sand wasps return to their temporarily closed nest and, on the basis of
a quick inspection, decide whether many or few caterpillars must be provided to the
young larva inside. How much reasoning can insects do?
Such questions have a long history and are still far from fully answered. Yet
behavior fairly begs for some degree of functional explanation, because ultimately
behavior means making physiological decisions—when to move and when to

46 2 Programming and Integrating Behavior
remain still, what objects to approach and what to avoid, which muscles to contract
and which to relax. For any organism, such decisions require some sort of inter-
nal communication system that can organize, manage, and synchronize incoming
information and outgoing responses. Like most other complex organisms, insects
have two such internal communication systems: a nervous system and an endocrine
system. The nervous system is electrochemical, and acts through impulses traveling
over the surface of specialized cells called neurons. The specificity of the messages
in this system resides in the nature of the pathways over which they travel, not in
the differing nature of the messages, because all are due to the same type of physio-
chemical event. The endocrine system, on the other hand, is more purely chemical.
Messages are encoded in specific substances produced and released by specific body
cells. As these chemicals are transported through the body, many cells are exposed
to them, but only certain uniquely sensitive cells respond.
There are real advantages in this duality, two systems existing side by side.
Nerves send electrical messages rapidly, but a nervous system that reached every
cell would take an enormous amount of space. Hormones, on the other hand, can
bathe each cell; however, rapid responses such as escape would be poorly served
by the relatively slower speed of the hormonal bath. In effect, by evolving a dual
system, organisms have made a reasonable trade-off between response speed and
system complexity.
In another sense, however, one might also argue that the insect system is not dual
at all but a continuum broken only by imposition of the human need to classify.
Neurosecretory cells occur throughout the central nervous system and their secre-
tory products are important functional parts of it. Neurosecretory cells emit products
that serve as hormones, neuromodulators, neurotransmitters, regulators of hormonal
secretion, and a variety of other functions. Almost all are peptides or small proteins
that control physiological or biochemical processes in which sustained stimulation is
needed. Although over 100 have been identified in insects, clearly defined functions
have been demonstrated for very few.
2.2 Nerve-Based Coordination
‘What is the difference between a vertebrate and an invertebrate?’ ‘One goes squish-
crunch; the other goes crunch-squish.’ It’s an old riddle, but it sums up a vital
difference. The characteristic feature of arthropods, including the insects, is a hard,
jointed exoskeleton or cuticle made up of a series of hard plates, the sclerites.
For an insect, having its support frame on the outside is a mixed blessing,
because the same cuticle that provides a defense against assault and dehydration
also separates the insect from its environment. If the insect is to make behaviorally
appropriate responses, the nervous system inside this shell must be able to receive
information across the barrier and that requires a moist receptor surface, but for
such a small organism, water loss absolutely must be minimized. The compromise
for insects has been to expose receptor surface only through pores that, while open
2.2 Nerve-Based Coordination 47
continuously, are so extremely small that they minimize the potential for water loss.
The receptors also have been concentrated on a very few body areas, such as the
mouthparts, antennae, legs and cerci. Most of the body has remained waterproof
but also insensitive to external stimuli. Where sensory reception occurs, the basic
unit is the sensillum, which usually looks like a tiny hair in a socket (see Fig. 6.2).
The sensory neurons inside are bipolar—one end extends into the cuticular portion
to receive stimuli and the other end sends messages to the central nervous system.
Sensilla have many structures and functions, including not only those one might
expect such as taste perception, but also unexpected ones such as the infra-red irra-
diation detectors in species of buprestid beetles that breed only in trees recently
killed by fire.
All receptors code their information in units called action potentials, which are
self-regenerating standard signals that travel through other nerve cells along the
length of long cytoplasmic cell projections (axons) through small gaps (synapses),
the switchboards of the nervous system, where they are transmitted by the release of
chemicals having specific effects upon the neuron or muscle cell across the synapse.
Stimulation usually leads to the production of not one but many nerve impulses all
of the same amplitude. Information about the stimulus is coded in the number and
frequency with which they follow each other, within limits of the system.
2.2.1 The Insect Nervous System
The insect nervous system (Fig. 2.1) is made up of two highly structured, inter-
twined systems. The first is the visceral (also called stomatogastric or sympathetic)
system that controls alimentary canal movements and is closely concerned with the
process of neurosecretion. The second is the central nervous system (CNS), which
coordinates the peripheral sense organs and muscles. As concerns insect behavior,
the latter is of the most direct relevance. (A third term, peripheral nervous system, is
also sometimes used; it refers to all the nerves that radiate from the CNS to innervate
muscles, stretch receptors, sensory receptors, etc.)
Because insects evolved from the segmented system of annelid worms, their
organization still reflects that past. In the earliest insects, each body segment
probably contained a knot-like cluster of nerve cells, called a ganglion (plural,
ganglia) that was responsible for the activities of that segment. However, in all liv-
ing insects today, different parts of the system have fused together in various ways.
A ganglion typically contains a mass of neuron cell bodies at its outer edges and a
central region, the neuropil, where synapses occur. The largest and most complex
neuropil occurs in the brain, but all ganglia contain neuropil regions. The neuropil
appears to keep growing, even after rest of the nervous system stops; this suggests
the importance of new information processing and integration throughout an
insect’s life.
Nerve cells, like highways, have many shapes, and no single shape can be called
characteristic. Like superhighway systems connecting cities, nerves radiate from
Fig. 2.1 Visualizing the insect nervous system. (a) Side view of a representative grasshopper
head. A chain of ganglia trail from the brain, dipping below the gut to lie just inside the ven-
tral cuticle; nerves from the ganglia branch out to the sense organs and muscle systems, while
neurosecretory cells in the ganglia chemically link the nervous and endocrine systems. (b) Front
view shows the three main regions of the brain—the protocerebrum, the deutocerebrum, and the
small tritocerebrum—and the nerves that connect the tritocerebrum with the frontal ganglion and
labrum; circumesophageal connectives at the back of the tritocerebrum connect the brain to the
subesophageal ganglion, a complex structure supplying nerves to mouthparts, neck, and salivary
glands
the ganglia to organs and muscles, generally within the segment where the ganglion
resides. Also, like most superhighways, they generally (but not always), are two-way
streets, with sensory (or afferent) neurons delivering signals to the central nervous
system, and motor (or efferent) neurons carrying output back to muscles, glands,
and organs. The few that act like one-way streets (such as the purely sensory ocellar
nerve) do not directly synapse with one another; instead they connect through one
or many interneurons.
In most insects, the brain is a large group of neurons that lies above the
esophagus. For that reason it is sometimes called the supraesophageal ganglion.
(Surprisingly, while the brain is in the head of most insects, in dipterous maggots
it is several body segments back.) There is still some disagreement among insect
physiologists about the brain’s precise origins and number of primitive segments
that comprise it. However, three parts are generally recognized (Fig. 2.1b). The
most anterior section, called the protocerebrum, is the most complex part of the
insect brain. It directs neural traffic at the crossroads between sensory input and
motor output. At each side, optic lobes extend to the compound eyes; at its center, a
pair of large mushroom bodies (corpora pedunculata) process olfactory information
and control tasks that require visual coordination of locomotor activity and spatial
orientation (see also Plate 1). The mushroom bodies and associated cells provide a
structure for elaborate interconnections that allow learning and memory to occur.
The second brain section, called the deutocerebrum, connects to the antennae.
Its neurons are of two types—one type processes chemosensory information; the
other, mechanosensory. The former have received a great deal of research atten-
tion because of the importance of the chemosensory neuropil in mediating feeding,
sexual, and social behaviors.
The third and smallest part of the brain, the tritocerebrum, connects the central
nervous system to the ventral nerve cord through the circumesophageal connec-
tives. It also innervates the labrum (upper lip), pharynx (region between mouth and
digestive system), and the rest of the visceral nervous system.
Beyond the brain, ganglia line up along a nerve cord that runs along the under-
side of the insect’s thorax and abdomen. The first ganglion of this ventral nerve
cord (and the only one still in the head) is the subesophageal ganglion. Another
nerve supercenter, it controls the rest of the mouthparts, salivary glands, and neck
muscles; axons from its neurons travel both forward to the brain and posterior to the
thoracic ganglion. The subesophageal ganglion influences motor patterns involved
with walking, flying, and breathing, although those motor patterns actually begin in
other ganglia.
The rest of the ventral nerve cord consists of a further series of paired ganglia
joined by lateral connectives. Typically, the first three are in the thorax and control
the wings and the legs. The number of abdominal ganglia varies because they have
fused in various ways in different insect groups.
The brain and central nervous system almost certainly are not merely simple
relay stations; rather, they function as integrative machines. Part of the evidence for
this comes from ablation experiments, in which anatomically distinct parts of the
system are surgically removed and their role is deduced by comparing behaviors
before and after excision. The crudest of these is simply to cut off the insect’s head,
an ‘experiment’ that has been performed since far antiquity.
Ablation experiments have confirmed that for many stereotyped behaviors the
insect brain actually is unnecessary—an observation which led many early biol-
ogists to believe that insects had none. The confusion was understandable. A
decapitated insect may even show complex behavior for many days, because (in
contrast to the vertebrates) insects have not concentrated their life support systems
within the brain. Crucial controls of functions such as respiration are decentralized;
so are many postural, locomotory, sexual, and grooming mechanisms. At the same
time, however, certain crucial differences can be noted when an insect is deprived of
various parts of its brain. Most notable is that the relative incidence of various types
of actions changes markedly. Often, there is an uncontrolled release of competing
behavior modes.
In a well-studied example, a preying mantis deprived of its protocerebrum simul-
taneously and continuously performs two opposed behaviors—grasping, which
holds it back, and walking, which pulls it forward. This is behavioral nonsense,
resulting in a hopelessly entangled and exhausted mantis. Intact mantises, of course,
spend most of their time doing neither; rather, they wait motionless, often for hours,
to ambush unsuspecting prey.
When the subesophageal ganglion is removed along with the protocerebrum,
however, the mantis becomes permanently immobile. Does the neural program for
leg movement reside in the subesophageal ganglion? No, because electrical stimu-
lation of a disconnected thoracic ganglion produces vigorous and complete limb
movements. Rather, neural inhibition occurs—the capacity of a neuron to exert
a blocking action on cells connected to it. In an intact mantis, the protocerebral
lobes apparently send out inhibitory messages that differentially block parts of the
excitatory activity being generated by the subesophageal ganglion. The excitatory
messages that are allowed to pass are transmitted to the thoracic ganglia where
grasping or locomotion is initiated.
How, then, does one explain the copulatory behavior of the decapitated male
mantis? When his head is destroyed, as often occurs during courtship with the preda-
tory female (see Fig. 9.16), the mantis walks in a circle while vigorously performing
continuous copulatory movements. Here, the ventral ganglia themselves possess an
endogenous activity that is usually inhibited by the subesophageal ganglion. Such
inhibition is a fundamental property of nervous systems. If a nervous system were
unable to inhibit those circuits competing with the one responsible for a desired
behavior, the result would be behavioral chaos.
2.2.2 Simple Reflexes and Repeated Motor Patterns

Many rapid autonomous behavior patterns in insects depend upon relatively simple
neural circuitry. One set of such patterns includes reflexes. These are familiar to us; a
gentle tap on one’s knee elicits a knee jerk reflex, for example. Movement is the most
common form of response to a stimulus. It is also the least complex, because the
input and output have a one-to-one relationship. A knee jerk, for example, does not
repeat itself unless the knee is tapped repeatedly. Because of their relative simplic-
ity, reflexes provide some of the clearest examples of the ways in which behavioral
stereotypy depends on properties of nervous systems. Other examples include kine-
ses and taxes, two locomotory responses that will be covered in more detail in the
next chapter.
A related set of simple behaviors are those involved in various rhythmically
repeated motor patterns such as insect songs, flight, and walking. How are these
physiologically generated, maintained, and coordinated? Historically, one explana-
tion (called the cyclic-reflex hypothesis) stated that feedback from the act itself was
sufficient to cause the act to be repeated. Thus, such rhythmic actions would con-
tinue in a repetitive circular loop (like perpetual motion) until inhibited by other
reflex paths or by the brain. Since the 1960s, however, accumulating evidence
has supported the concept that many patterns are generated in the central nervous
system—the central pattern generator (CPG) hypothesis.
In the 1960s, Donald Wilson conducted a series of now-classic studies supporting
the CPG hypothesis. Technical advances in electrophysiology had made extracellu-
lar recording with cathode ray oscilloscopes into a powerful new tool for studying
the neural basis of behavior when Wilson began collaborating with Torkel Weis
Fogh, who had developed an experimental system in which a tethered locust could
be induced to fly normally in a wind tunnel.
Working together, the two men built a detailed picture of neural output to the
locust’s flight muscles, then compiled a catalog of sensory inputs to the thoracic
ganglia. Next, in a series of skillful experiments, Wilson successively eliminated
sensory input to the thoracic ganglia while recording motor output. First, he
removed the locust’s head and subesophageal ganglion; lacking its wind-sensitive
sensory hairs, the locust flew slowly but otherwise normally. Step by step, further
operations severed one source of sensory input after another, until Wilson was able
to show that even total removal of all sources of periodic input did not abolish pat-
terned motor output (Fig. 2.2). When the breezes in the wind tunnel were kept still,
the basic pattern of flight motor nerve discharge remained unaltered. When Wilson
severed the nerve fibers from the stretch receptors that register wing movement,
the pattern persisted. When he dispensed with timing cues from other moving body
parts, nothing changed. Clearly, the locust was not timing its wing beat through
sensory cues created by wing movement. The motor pattern must have originated
within the thoracic ganglia.
The concept of central coordination does not exclude peripheral influences,
however. Though the sensory and motor systems were not linked in the locust in
terms of timing, the average frequency of wing beat was correlated with the dis-
charge rate in the receptor nerves and it slowed when receptor nerves were cut.
Sensory feedback did not cue sequences, Wilson postulated, but instead modu-
lated genetically determined motor programs. Subsequent studies have added a
layer of complexity to Wilson’s conclusions, however, with the recognition of neu-
rohormonal modulations of these central circuits. Even though locomotion (and
Fig. 2.2 Demonstrating the central pattern generator. Experiments with tethered locusts,
which will perform normal flight movements in a wind tunnel, demonstrated that the insect cen-
tral nervous system, without sensory input from wing sense organs, generates a pattern of motor
neuron output that closely resembles the pattern produced in normal flight. Sensory discharges in
nerves from the wing and wing hinge were recorded by manipulating wires into a locust’s largely
eviscerated thoracic cavity. Down-stroke muscle potentials repeat at the wing beat frequency, while
stretch receptors fire two to three times per wing beat
a great many other rhythmic behaviors) may depend on genetically determined

central pattern generators, they usually can be modified by sensory feedback that
serves to adjust them to various real-world factors such as turbulence, injury,
and aging.
Partly for the simple reason that so much is known about it already, cockroach
escape behavior continues to be an active research area today (see Case Study 2.1).
It has been shown that the behavior has two distinct parts, an initial directional
turn to orient away from the threat, followed by a run of more random direction.
Variants in the initial turn have been described, but leg kinematic motions appear
quite stereotyped.
Case Study 2.1: Escape Behavior in the Cockroach,

Periplaneta
Perhaps the most striking characteristic of many simple reflex-like actions of

insects is their speed. To an escaping insect, a few thousandths of a second
may be the difference between the quick and the dead. How are cockroaches,
for example, so proficient at reacting to our approach, even from behind?
One of the advantages of studying neuroethology in invertebrates is the
possibility of working at the level of individually identified neurons. Escape
reactions are one of the easier behaviors for such work because the rapid
nerve impulses necessary to evade a predator are conducted along axons of
especially large diameter. Unlike the numerous small interneurons that char-
acterize many other circuits, these giant interneurons (also called giant axons)
run for relatively long intervals before synapsing. Periplaneta americana, the
American cockroach, has 14 of them, each of which can be easily identified
and distinguished with a microscope.
In pioneering research that began in the 1940s with simple eqiupment
(Fig. 2.3), Kenneth Roeder showed that although touching a cockroach will
make it flee, actually the major (and often only) sensory cue a cockroach
needs to detect a predator’s approach is the tiny wind gust that precedes the
predator’s moving body. Two cerci, thin projecting appendages at the end of
its abdomen, detect the air puff, and in response, the cockroach turns and runs
away, legs scurrying along in classic CPG manner. It happens so quickly a
cockroach can successfully detect and flee the displacement wave preceding
the tongue of a striking toad. A slowly walking cockroach can take as little as
11 ms to respond to an air puff of particularly high acceleration. The situation
is considered a model example of mechanoreception of ultra-low frequency
signals.
Both high wind acceleration and a certain critical velocity are important
in causing the insects to run. On the underside of cockroach cerci, Roeder
observed delicate hairs that bend in their sockets, each exciting a single sen-
sory neuron at its base. Because stimulation of the cerci excites many of these
sense cells at once, a volley of action potentials arrives at the synapse more or
less simultaneously. This source-related additive effect, called spatial summa-
tion, generates an impulse in a giant fiber. These in turn excite motor neurons
of the cockroach leg. Spatial summation is the reason why cockroaches do
not dash madly off in response to minor stimuli such as kitchen drafts, which
would cause only a few cercal setae to be deflected and fire their sense cells.
Likewise, impulses also would not be propagated along a motor axon,
Roeder discovered, until at least two (often three or four) volleys of impulses
Fig. 2.3 Signaling cockroach escape. (above) Apparatus used by Kenneth Roeder to study startle
response in the American cockroach, Periplaneta americana, one of the commonest cockroaches
throughout the world. Glued to a stick, the cockroach is allowed to grasp an unattached cork ball
that turns as the roach runs. Near the cockroach’s cerci are an air nozzle and a small paper flag. As a
puff of air is delivered to the cerci the flag is deflected and an impulse registers on the oscilloscope
screen. As the cockroach reacts a second deflection is registered and the difference between the
two is the startle response time. (Below) Air puffs stimulate receptor hairs on the cerci that relay
impulses to giant interneurons running up the central nerve cord to the metathoracic ganglia, from
which signals activate motor neurons in the legs. Though the sensory hairs hang below the cerci,
they can detect wind from all directions
had arrived at this synapse in a time-related additive effect (temporal summa-

tion). When he continued to send more volleys of impulses along the giant
axons, they quickly stopped responding. On behavioral grounds, this was
logical—in a natural situation, continual directives to begin running are redun-
dant information to a cockroach that is already fleeing! Roeder also found that
when impulses in the motor fibers were generated, they continued to repeat for
several seconds after all input from the giant fibers had ceased. This illustrated
an important adaptive aspect of evasive behavior—movement continues for
some time after contact with the stimulus, thereby taking the animal out of
harm’s way.
Since Roeder’s studies, others have confirmed that for most invertebrates,
the receptor system associated with detecting danger is associated with a
rapidly conducting nervous system that can bring about action through the
shortest/fastest neuronal route. In most cases, this is via rapidly conduct-
ing giant fibers in the ventral nerve cord. For a number of species, a single
interneuron is enough to trigger normal escape. In many cases, all the neural
circuitry needed for the basic oriented response occurs in the ventral nerve
cord and thoracic ganglia.
While the nerve circuits guiding escape reactions superficially seem
quite simple, detailed studies over the past forty years have shown that
cockroach escape behavior is actually quite complex, both behaviorally and
physiologically. Jeffrey Camhi and colleagues have been major contributors
to this new understanding. They began with a few simple questions. Could
cockroaches actually determine a threat’s direction from a puff of air, or
were they simply scurrying off in a random direction after being startled?
Plotting many trials showed that the insects usually did turn away from the
source of the wind before running away, so they must be able to determine
wind direction. How? In a series of trials, various areas on the cerci were
covered with wax. When the cerci’s upper side was covered, the cockroach’s
directional response was unaltered. When the underside of either appendage
was covered, most of the turns were away from the uncovered one, regardless
of the wind direction. When the underside of both cerci was covered, the
cockroaches no longer responded to an air puff at all.
Close examination of the bottom side of the cerci shows that the many tiny
pivoting hairs run in columns from the front to the rear. In the late 1960s,
R. Nicklaus, a German researcher, had shown that in any given column, all
the hairs deflect most easily in the same two opposite directions, and that the
directions of maximum pliancy differ from one column to the next. By record-
ing sensory output from a number of individual hairs, Nicklaus showed that
deflection of a hair in one direction maximized the number of action poten-
tials fired by its sensory cell. Deflecting it in the other direction inhibited even
spontaneously occurring action potentials.
Researchers in Camhi’s laboratory set out to determine the best excitatory
direction for each of the nine columns of cercal hairs of P. americana, and
to show the range of wind directions to which each column responded. Their
results showed that each sensory axon responds maximally to a fairly narrow
range of wind directions. However, as a group, the hairs provide 360-degree
coverage. (Thus, even if you approach a cockroach head-on, the cerci at its
tail end sense you coming!)
What happens to this directional information as it transfers up the neural
chain of command? Camhi and his colleagues made intracellular recordings
from individual giant interneurons, then marked each with various colors of
injected dyes. They learned that each giant interneuron responds characteristi-
cally to air puffs from different directions, and each also differs in its breadth
of response. Moreover, each giant interneuron has a characteristic firing rate
for a particular wind velocity. Comparing the firing rates of two or more giant
interneurons reveals differences uniquely related to wind direction. But how
do the interneurons ‘decide’ where the wind is coming from and which direc-
tion to turn in response? Various mechanistic theories have been proposed;
in recent years, Camhi and colleagues have been addressing these by stimu-
lating cells of identified giant neurons with artificial spikes added during the
sensory-induced behavior, then analyzing the resulting directional leg move-
ments. Though the jury is still out, it appears that directional determination is
based on collaborative calculations of direction by the giant interneurons as a
group, an approach called distributed neural processing.
Currently, a generalized picture of insect escape reactions sees a system in

which all the neural circuitry for a basic oriented response is found in the ventral
nerve cord and thoracic ganglia. A threat activates sensory neurons associated with
peripheral sensory structures such as abdominal cerci. The sensory neurons synapse
onto interganglionic interneurons (such as the giant interneurons) or touch-sensitive
interneurons in the abdominal nerve cord. In each of the thoracic ganglia, various
pathways converge onto interganglionic interneurons that are thought to be the locus
of control of the escape system. They determine whether an escape turn is to be ini-
tiated and in what direction the insect is to turn. They also connect to the leg motor
neurons in their own and adjacent ganglia, both directly and indirectly by way of
local interneurons.
Initially, researchers such as Wilson and Roeder believed that the ‘higher
centers’ of the insect nervous system acted mostly to start or stop lower pattern
generators. Evidence now suggests that whereas fast-acting behaviors of short
latency such as being startled (often called ‘ballistic’ responses) probably do
usually rely upon circuits outside of central control, higher centers interact to
initiate, maintain, and modulate ongoing rhythmic patterns in ways that produce
advantageous behavior in a complex environment. Interestingly, research engineers
and the US military establishment currently are showing interest in this central
pattern generator/peripheral feedback model as a basis for the design of controllers
for autonomous hexapod robots.
2.2.3 Ethological Explanations

Sand wasps, with no contact with members of a previous generation, perform com-
plex nest-burrowing behaviors. Caddisflies reared in isolation spin perfect cases.
Male crickets reared in solitary confinement still sing species-typical aggressive and
rivalry songs when confronted with another male.
In many ways, the behavior of insects gives the impression of acting from a
prewritten script. Faced with a given situation, even the novice behaves appropri-
ately. Perhaps for this reason, the concept of instinct, so emotionally controversial
in vertebrate behavioral studies, has been more easily accepted when applied to
insect behavior. As numerous studies have shown, even among insects that have
never been in contact with conspecifics, all individuals of the species still exhibit
many species-specific motor patterns in exactly the same form.
While their counterparts in physiology and neurology were examining insect
behavior from a mechanistic viewpoint, the early ethologists (see Chapter 1) were
more interested in a functional view of behavior. Where the physiologists and
neurologists saw pre-wired endogenous motor programs, the ethologists saw fixed
action patterns (FAPs). First described and elucidated by the pioneering ethologists
Konrad Lorenz and Niko Tinbergen in the 1930s, fixed action patterns generally
have several common characteristics. First, once initiated, fixed action patterns
require few or no external stimuli or additional sensory cues for their maintenance
or completion. For example, even after certain body parts are removed, many insects
will proceed to clean the missing appendages as though they were still there; FAPs
have been likened to a piece of memorized music played by a pianist without
hearing, seeing, or feeling tactile sensation from his hands.
Second, fixed action patterns can be evoked by a variety of stimuli, some-
times from different sensory modes, and they may show variability in orientation.
However, once evoked, their basic structure is stereotyped. Third, they are found
in all individuals of the species that displays them. Like morphological characters,
FAPs are subject to selective pressures and have a genetic basis. Fourth, once they
are initiated, fixed action patterns occur in the absence of positive feedback. Nothing
additional is necessary to keep them going to completion.
Finally, fixed action patterns generally involve the coordination of several dif-
ferent muscle groups, and can be quite complex. In crickets, 48 separate FAPs
are involved with molting; the system, coordinated by a hormonal cascade, is
multifaceted and intricate. Why such complexity? The general explanation is that
selection has been intense. With behaviors as fundamental as eclosion, there is no
room for error; mistakes are deadly.
What triggers the release of a fixed action pattern? Tinbergen, Lorenz, and other
ethologists studied dozens of instincts seeking an inclusive theory. They found that
first, an organism’s internal physiological state has to be such that it is in a condition
of readiness to respond, and second, the sensory-neural mechanism involved has to
be exposed to very specific forms of stimulation to set a given FAP into motion.
They called these cues releasers or sign stimuli. In some situations, the releaser
appeared to be the total stimulus configuration, or Gestalt. However, in most cases,
only certain simple, specific aspects of the stimulus were needed. For example, the
attraction of a male mosquito to conspecific females may be evoked by a particular
sound frequency, which we recognize as the females’ distinctive hum; artificially
produced sounds will produce exactly the same behaviors in the complete absence
of females.
A releaser need not in itself have any inherent relevance to survival. Rather, it
may be some property of a particular situation that has biological importance and a
property that is shared by all other similar situations. For example, skatol and ammo-
nia release feeding behavior in both dung beetles and blowflies; these substances are
produced by decaying matter, but neither are by themselves of any nutritive value
nor do they owe their existence to dung beetles or blowflies. This type of releaser
has been termed a token stimulus (see also Chapter 4).
In passing, it should be noted that the naturally occurring releaser for a behavior
is not necessarily the optimally effective one. Male Argynnis butterflies are more
strongly attracted to solid orange than to the orange-black pattern of the female.
Such a supernormal stimulus—a signal that is even more effective in eliciting a
given behavior than the natural signal is—usually represents an evolutionary com-
promise. For example, in the sexual approach of Heliconius butterfly males, use of
experimental models has shown that pure red wings upon the female would be the
optimal attractant. However, since the coloration also has mimicry functions, natural
selection has dictated limiting the red signal to a stripe.
A releaser that is simple can maximize an animal’s recognition of biologically
important stimuli while minimizing the amount of neural circuitry required, but it
also increases the risk of exploitation by other species that use the FAP for their own
purposes, a tactic known as code breaking. A well known example of this is the way
in which many tropical orchid flowers provide the visual and olfactory releasers that
trick male wasps into attempted copulation (see Fig. 9.8).
These classic ethological observations interface with modern insect physiology
through the useful concept that for each FAP, there exists a sensory-neural release-
controlling mechanism called the innate releasing mechanism (IRM) that evaluates
incoming signals from one or several sense organs and triggers (or fails to trig-
ger) the signals that go out to initiate a reaction. Complex behaviors resolve into
a sequence of stereotyped motor patterns; each step depends on the presence of
appropriate external stimuli and internal physiological state before the IRM will
act. If any link in the chain is broken, the preprogrammed sequence of behaviors
(reaction chain) does not continue and hence cannot run to completion.
By itself, the IRM is a ‘black box’ sort of concept, because it considers only
input–output without considering the physiological components behind the control
of fixed action patterns. At each step, the IRM activation threshold may be highly
variable and may respond to both internal and external influences. In common expe-
rience, we say an organism shows varied motivation to perform the act. A highly
motivated individual may be so ready to act that it will perform a fixed action pat-
tern when confronted by a stimulus bearing but slight resemblance to the typical
releaser. For example, if the sand wasp Ammophila is thwarted too long from pulling
a paralyzed caterpillar into her nest, she will begin to retrieve substitutes. Most moti-
vation initially generates quite unspecific behavior, sometimes simply locomotion.
The behaviors that follow become increasingly specific until they ultimately culmi-
nate. For example, in the chain-like reciprocal signal exchange that occurs between
the two sexes during courtship (see Fig. 9.4), the culminating act of copulation can
occur only at the end of a hierarchical sequence of several releasers and FAPs. The
2.3 Life in a Stimulus-Rich World 59
prey-catching behavior of the bee wolf (see Chapter 4) is another example of such
hierarchial sequences.
What happens when the environment contains sign stimuli for quite different or
even contradictory forms of behavior? Or when motivations conflict? Under normal
circumstances, behavior outcome depends on the relative strengths and effective-
ness of the different motivational factors and sign stimuli involved, and the behavior
that is activated partially or wholly suppresses other non-compatible behaviors. In
most insects, escape behavior is inhibited during copulation. In aphids, flight and
settling behaviors inhibit each other. If honey bees are exposed to releasers for
activation and inhibition of their communicative dances simultaneously, trembling
dances may take place. When stimuli for completely contradictory behaviors com-
pete for attention, apparently nonfunctional stereotyped actions called displacement
activities often occur. Most recognized cases have involved cleaning or preening
behaviors. Faced with the opposing behavioral motivations to stay at a food source
and to leave it, bees will begin grooming, independent of actual grime. Dusting the
bee with pollen or flour does not change the frequency of this behavior; rather, it
merely directs the preening movements toward the dustier body parts.
Most behaviors are continuously influenced through feedback from both their
internal environment and the external one that surrounds them. A classic exam-
ple of internal feedback is provided by the feeding behavior of the blowfly (see
Fig. 4.3). When stimuli that release feeding are continuously present, more or less
continuous food uptake ends only when inhibitory stimuli arise; these come from
stretch receptors in the gut wall that serve as internal monitors, firing as the gut fills.
Action potentials carried to the brain via the recurrent nerve trigger the eventual
motor response, namely, retraction of the fly’s proboscis from the food. Cutting the
recurrent nerve eliminates the negative feedback from the stretch receptors, with the
result that the fly literally explodes from overeating!
2.3 Life in a Stimulus-Rich World

A rabbit possesses over 100 olfactory receptor cells. A representative caterpillar
has only 48. Both herbivores successfully find and discriminate between potential
food plants. The caterpillar’s receptor system seems comparatively limited in com-
parison to the rabbit’s array of receptors. How do insects maximize the amount of
biologically significant information sent to higher levels in their nervous systems?
At any given moment the environment contains a kaleidoscopic array of stimuli,
only a fraction of which contain biologically relevant information. Our rabbit and
larva are surrounded by myriad smells arising from the earth, sky, competing plants,
other animals—the list could go on and on. In what ways might they most efficiently
ignore irrelevant elements and non-appropriate stimuli?
One way to treat different kinds of stimuli differentially is to evolve a receptor
system that is designed to screen out some kinds of stimulation while reacting to
other kinds. Every animal species has done this to one degree or another for the
stimuli that are requisite to their survival and reproduction. The simplest receptor
takes the form of one cell that both receives the environmental stimuli and relays it.
All sensory receptors of insects are primary sense cells. The same cell produces both
a receptor potential and an action potential, and there is no need for a second neu-
ron. Vertebrates have secondary sense cells—cells of non-neural origin that receive
a stimulus and link to a neuron to send the message along. For example, touch recep-
tors in the vertebrate skin are modified epidermal cells. More complex sense organs
may contain thousands of tiny sensory devices comprised of specialized nerve cells.
Uniquely sensitive to particular kinds of environmental energy, they simultaneously
screen out all stimuli outside a selected signal range.
2.3.1 Sensory Tuning and Filtering
Honey bees learn landmarks around a nectar feeder, but ignore these same land-
marks when presented at a novel location. Studies with mutant Drosophila suggest
that mushroom bodies (see Fig. 2.1) are important for odor learning, but not for
visual or tactile learning. Several species of butterflies land preferentially on leaves
of particular shapes, with further discrimination occurring only after landing.
Ever since ethologists first recognized releasers and sign stimuli, it has been
clear that certain types of receptors are tuned to very specific aspects of a gen-
eral stimulus (see Fig. 6.4). One could envision neural information processing as a
hierarchical system through which the various stimulus properties are filtered until a
particular behavioral response is ultimately released. Ethologists suggested the con-
cept of two sensory filters, one peripheral and the other central. Peripheral filters
functioned at the level of the sensory receptors. Central filters occurred within the
nervous system, sorting out incoming information, selecting relevant stimuli for fur-
ther action, and eventually producing a particular response. Central filters thus were
components of what they called innate releasing mechanisms, or IRMs (see Section
2.2.3).
Chemosensory systems provide many examples. Partially because insects are
known to live in a chemical sensory world and partially for the practical rea-
sons of control of agricultural pests, chemoreception has received a great deal of
research attention in the context of stimulus filtering. There are an inestimable
number of odors in the world. How can the sensory and nervous system of any
given insect species decipher them all? The answer, as numerous studies have
shown, is that among both taste receptors and olfactory receptors, some are rela-
tive specialists. For example, each feathery antenna of a male polyphemus moth
has about 150,000 sensory neurons; 60–70% of these are specialized for detect-
ing female-produced sex odors. Other receptors are generalists that respond to
several or many kinds of chemicals but varying in their pattern of sensitivity. By
summing the responses of the different chemoreceptors, the CNS could obtain
characteristic total response profiles, each uniquely representative of a particular
chemical compound. An additional dimension of information would be acquired if
instead of being silent in the non-stimulated (resting) state, chemoreceptor cells
maintained some constant spontaneous baseline level of firing activity. Then a
particular stimulus could be recorded as either an increase or decrease in firing

over the spontaneous rate. The resulting code of neural activity from a field of
generalist receptors, known as across-fiber patterning, allows a large number of
different stimuli to create different patterns that can be coded by the central
nervous system without requiring large numbers of specific receptors for each
molecule.
The olfactory sense of many caterpillars seems to operate in just this fash-
ion, resulting in different response profiles for each receptor (Fig. 2.4). Since the
caterpillar receives information from all chemoreceptors simultaneously, discrimi-
nations between numerous natural plant odors, even in excess of those likely to be
encountered in nature, are easily within the caterpillar’s sensory powers using only
Fig. 2.4 Discriminating

between plants. Activity
spectra of each of four
olfactory receptors of the
tobacco hornworm (Manduca
sexta) in response to 11
different odors show that a lot
of information can be
encoded with only a few
chemoreceptors. Each
receptor’s sensitivity takes the
form of an increase or
decrease over its unique
spontaneous firing rate
a few receptors. It is likely that the relative sensitivity of different receptors also
regularly fluctuates, since it is known, for example, that many insects are rhythmic
in their degree of susceptibility to insecticides.
A supreme example of sensory tuning occurs in predatory bats and the moths
on which they prey. As a simple experiment, suddenly jingle a bunch of keys
near a group of moths flying around a streetlight or window screen. The response
of the moths will at first seem chaotic. Some nearly fall to the ground, while
the flight of others becomes quickened and more erratic. Some that were flut-
tering may become motionless, while others, previously motionless, may take
flight.
What pattern can there possibly be to all of this? To answer this question, one
must appreciate the importance of a principal predator of moths, namely, bats. Fifty
years ago, precise and ingenious experimentation by Donald Griffin revealed a bat
echolocation system (Fig. 2.5) capable of indicating size, distance, location, and
considerable detail about its surroundings, down to items smaller than midges.
However, taking the story to the next level took the work of several researchers,
including the same Kenneth Roeder that unlocked the basis of the cockroach escape
reaction. Case Study 2.2 presents some of the story.
Fig. 2.5 Listening for bats. The relative acoustic sensitivity of a noctuid moth at various frequen-
cies (open circles), contrasted with summed intensities of all natural environmental noise recorded
in the moth’s environment at night (solid line). Sounds below 15 kHz were mostly of insect origin.
Those between 23 and 50 kHz came from passing bats. It is apparent that moth ears are maximally
sensitive to sounds in the latter range and relatively insensitive to other sounds
Case Study 2.2: How Noctuid Moths Hear Bats

Bats and moths have been playing the proverbial cat-and-mouse game for
millions of years. Moths are one of the main sources of food for certain fam-
ilies of bats. Yet the game is clearly a balanced one. While all bats in these
families probably locate and capture some moths, some moths recognize and
evade all bats. How does all this work?
Except in the case of certain ‘singing’ insects, one does not usually think
of insects as possessing ears, but members of certain moth families, especially
the Noctuidae, possess true tympanic organs, located in the thorax just below
the second pair of wings. From the physiologist’s standpoint, a simpler system
for experimental analysis would be difficult to find. Each of the moth’s tym-
panic organs consists of only two acoustic sense cells, each with a different
threshold, coupled to a thin tympanic membrane. The more sensitive acoustic
cell is known as A1, the less sensitive, as A2. Although unable to discrimi-
nate between tones of different pitch, the tympanic organs are tuned for the
reception of ultrasound.
The first verification that moth tympanic nerves respond to bat chirps
was obtained quite by accident at Tufts University, when a hibernating bat
was brought out of a refrigerator into the laboratory near a moth tympanic
nerve preparation linked to an oscilloscope and microphone. Unexpectedly
the bat recovered from the cold enough to shriek, bite the investigator’s hand,
and escape into the room, where it flew ‘silently’ around near the ceiling.
Throughout this flight, the tympanic nerve proceeded to deliver a rapid series
of short spike bursts.
Excited and encouraged at this, Kenneth Roeder and his students eagerly
lugged some 300 pounds of electronic gear out onto a grassy hillside where
bats were known to feed in the evening and set up a moth for tympanic record-
ing. Very soon, they found themselves able to decode the movements of bats
from the nature of the moth’s tympanic nerve responses. Rigging up a flood-
light so they could actually see the bats (since they were unable to hear them
independent of the moth’s ears), they quickly learned that the range of the
moth ear was much greater than they had expected. Later experiments, in fact,
showed it to be in the nature of 100–200 ft when bats were flying toward the
moth at an altitude of about 20 ft. In contrast, a bat is unable to track a moth at
ranges greater than about 10 ft. The bat, however, has the advantage of much
greater speed.
When Roeder turned his attention to the moth’s flight behavior, the
complexity of the natural situation seemed almost to defy study—‘a dizzy
dogfight’ was his impression, ‘extrapolation of a string of acoustic dots in time
. . . pitted against unpredictability; power and speed against maneuverability.’
It might be compared to combat between a fighter plane and a helicopter, the
bat flying a time intercept course while the moth undertakes selective evasive
tactics on the basis of the bat’s range and speed of approach.
To reduce this complexity, Roeder replaced the bat with an artificial one:
a stationary multidirectional electronic transmitter of ultrasonic pulses, linked
to a camera system and perched 16 ft in the air. When exposed to these elec-
tronic bat cries, moths showed the same bewildering variety of reactions as
they did to real bats. After many hours of observation and over a thousand
photographically recorded moth tracks, Roeder determined that the nature of
the evasive tactics of the naturally flying moths was related to the distance
between the moth and loudspeaker. Moths that flew close to the loudspeaker,
encountering ultrasound that pulsed at 10–30 times a second, responded with
a variety of maneuvers that usually ended in a dive. Those cruising at greater
distances characteristically turned and flew directly away from the sound
source.
The adaptive value of this response difference seems obvious. There would
be little survival advantage to the moth in attempting to flee a bat at close range
or in making erratic turns and twists while the predator was still distant, but
every advantage in erratic behavior when the bat was close. Not surprisingly,
straight directed fleeing occurred only in response to low-intensity sounds;
when the loudspeaker signals were made progressively weaker, the distance
at which directed fleeing was released was correspondingly reduced.
Interestingly, moths cruising at about the same height as the loudspeaker
turned and flew away in the horizontal plane; those above it were observed to
fly directly upward, or redirected their flight by making a sharp turn before
flying straight away. How can a moth equipped with only four sense cells ori-
ent and steer itself with respect to a sound that comes from various angles? As
background for an answer, one must remember that free-flying moths typically
flap their wings many times per second. Unlike an airplane, they rarely fly for
long in the same direction on an even keel. Because the moth ear is located
below its wing, its wing movement has profound effects on the acoustic sensi-
tivity and directionality; on the downstroke, most sound will be screened from
reaching the ear, while at full upstroke the ear is accessible to sound from all
directions.
To investigate the effects of wing position more fully, Roeder and his
associates devised an elaborate apparatus that would measure the acoustic
sensitivity of one ear to sound coming from all points in an imaginary sphere
surrounding a moth fixed in different flight positions. Obtaining these record-
ings was a long and arduous task, and Roeder was fully aware that the surgical
insult of implanting electrodes, plus the restriction of wing movements that
was necessary, might affect the results. However, recordings of the A fiber
responses revealed two types of acoustic asymmetry. One was a left–right
asymmetry. The right ear detected sounds most effectively when they origi-
nated on the right half of the sphere, and best of all when the sound source
was roughly at right angles to the body axis with the wings at the top of
their upstroke; at the same time, hearing in the left ear was at its minimum.
Alternating with this was a dorsal–ventral asymmetry. The flapping of the
wings had the effect of damping the sound when the wings were in the down
position. Thus, the left–right differences in sound intensity would alternate
with the temporary disappearance and reoccurrence of sound occurring some
10–40 times per second in synchrony with the rate of wing flapping.
How do these asymmetries relate to a moth’s ability to escape a hungry bat?
A left–right difference in A-fiber discharge when the wings are up probably
provides the moth with a rough horizontal bearing on the position of a bat
with respect to its own line of flight. The absence of a left–right differential
discharge and the presence of similar levels of on and off from both ears might
inform the moth that the bat was above it. If neither variation occurred at the
regular wing beat frequency, it would mean that the bat was below or behind
the moth (Fig. 2.6).
It would seem that a single sense cell in each ear could transmit enough
information to inform the moth of a cruising bat’s bearings. Why does each
ear have two acoustic sense cells? Roeder suggested that the answer might
lie in their differential sensitivities. His comparative measurements showed
that A1 responded first, firing over a range of low to moderate intensities.
Cell A2 began to fire only at moderate intensities and fired even faster at
high sound intensity, a range within which A1 was saturated and therefore
incapable of further increase. Thus, by operating in piggyback fashion, A1
and A2 might provide a combined signal from which intensity differences
could be discriminated over a range wider than either alone could accomplish.
In effect, their combined signal would inform the moth how far away the bat
was, which ultimately would decide the form of the moth’s evasive behavior.
He suggested testing this theory by observing the anti-bat behavior of another
group of moths (Notodontidae), because their ears each contain only the A1
cell.
A decade after Roeder’s research raised the puzzle about the A2 receptors, a
report appeared that European notodontids do, in fact, show a two-part response
to approaching bats. However, years passed before the A2 question could begin to
be addressed directly rather than inferentially. The problem lay in the difficulty of
neurally recording moth auditory responses not just to one set of constant acoustic
parameters, but to the complex sounds of real attacking bats. As a bat approaches its
prey, its echolocation calls shift in duration, intensity, rate and frequency. The ideal
situation would be to record the moth’s responses during an actual bat attack, but
the recording equipment itself made it impossible to set up this scenario.
Fig. 2.6 Assessing bat direction. Roeder showed that the activity in the A1 receptors of a moth’s
ears differs upon detecting bat cries from different points in space. (A) When the bat is to one side
of the moth, the receptor on the closer side fires more rapidly than the shielded receptor. (B) When
the bat is directly behind the moth, both A1 fibers fire at the same time and rate. (C) When the bat
is above the moth, activity in the A1 receptors pulses in synchrony with its wing beat. The role of
the A2 receptors, which are missing in some moths, is less clear
In 1994, James Fullard and colleagues found a way around the problem by train-
ing a bat in the laboratory to attack a microphone that it expected would be an
edible target. These recordings (and a noise-reduced, computer-generated digital
replicate) were then broadcast to five species of noctuid-like moths to observe their
ears’ responses. The results appear to confirm that the A1 cell encodes the approach
calls of an attacking bat up to about 100–200 ms before the bat would capture the
moth. Then, at this last moment of the moth’s life, the A1 firing drops off to a level
similar to that detecting a faraway bat. Surprisingly, the A2 cell responses were
fundamentally the same, but Fullard found variations among the moth species he
tested. In some, the A2 cell response was vigorous; in others it dropped off early,
was sporadic, or did not appear at all. Fullard and colleagues have suggested that this
may mean that the A2 cell is vestigial, and no longer used in the flight responses of
moths.
A sound also has several other features that could convey information to the
moth’s nervous system, including its intensity, pulse duration, the interval between
pulses, and so on. Furthermore, the receptor is but one point along the chain of neu-
rons leading to the brain where selective decoding of stimuli might occur. Thus, in
the case of the ultrasonic bat cry, moth ear receptor cells might selectively respond
to a particular range of frequencies at one point in the filtering, but after the recep-
tor cells transmit an impulse, subsequent interneurons along the ventral nerve cord
would no longer respond to frequency; rather, other sound properties such as pulse
intensity, pulse duration, or interval between successive pulses would now be mon-
itored. For example, Kenneth Roeder identified a ‘pulse marker’ interneuron in the
moth that responds to three or four sensory impulses separated by short intervals by
firing just once, but the interneuron made no distinction as to duration of a pulse; to
it, 0.5 and 500 μs were the same.
The dramatic conflict between moths and bats has caught the attention and imagi-
nation of scientists in many disciplines. A number of other insects quite distant from
moths have been found to respond evasively to bat ultrasonic cries, from mantises
that loop and plummet erratically downward to lacewings that take a power dive.
The moth-bat story also has come to include some interesting side plots. One is the
production of anti-bat sounds—some arctiid and ctenuchid moths produce clicking
sounds at frequencies close to those used by bats, and bats respond by veering away.
The click may disorient the bat, jam its echolocation system, and/or act as a warning,
signaling the bat that the moth is distasteful (see Chapter 8 and Plate 28).
Another side plot involves mites, common ectoparasitoids that often attack insect
tracheal systems. The mite Dicrocheles phalaenodectes infects only one ear of its
noctuid moth host, destroying the typanum, but leaves the other ear intact; a related
species infests both ears, but leaves the tympanum intact. Presumably, both strate-
gies leave the moth host with some chance of still avoiding bats and thus living on
to support its parasites.
Since many moths have evolved ways to hear bats, why don’t bats evolve a coun-
termeasure such as calls that are too soft for a moth to detect or outside the frequency
range of moth hearing? Mostly likely such drastic changes would severely restrict
the range and acuity of the bat’s sonar system and would only result in related
adaptive changes in the moths. Such an evolutionary arms race may in fact be under-
way now. Studies in Canada indicate that where bat species-diversity is high with a
wide range of sonar frequencies, moth sensitivity covers this frequency range, but in
areas where the range of bat species is small and the sonar range is more restricted,
moths have a similarly reduced frequency sensitivity range. Similar relationships
have been shown in other areas, including between moths of Panama and those of
Hawaii.
2.3.2 Memory and Learning
Dragonfly larvae of the genus Anax raised in aquaria soon come to associate the
sight of their caretaker with food; when he or she appears, they may even snap in
anticipation before food actually appears. A mantis, although it originally would
attack, will learn not to strike at an object after it has received an electric shock
or bad taste from this object. Upon emerging from their pupal cases, new adults of
many ant species notice and retain the scent of their colony.
In the past, it was repeatedly suggested that insects show little or no learning,
for ‘obvious’ reasons such as their small brain and short life span. However, even
little Drosophila larvae apparently can easily learn to avoid odors linked to an elec-
tric shock, and parasitoid wasps rapidly learn cues associated with host availability.
In fact, current knowledge indicates that many insects learn extensively during
all major life activities, and furthermore that individuals within a species show
genetically based variation in learning abilities.
Learning may be defined as any relatively permanent, usually adaptive change
in behavior that occurs as a result of experience or practice. This change usually
progresses gradually with continued experience to some asymptote. It often can be
modified by novel experiences, and the effects of experience eventually wane if not
reinforced. Memory—the capacity to store information—is a prerequisite, resulting
in a linkage between stimulus and response that would not have occurred without
the previous experience with that stimulus. In physiological terms, one could define
learning and memory as the acquisition and retention of neuronal representations of
new information, respectively.
One of the stranger things about studying learning is that while learning undoubt-
edly involves neuronal modifications, it still can be assessed only indirectly through
its potential effect on behavior; there is no way to directly quantify it. To infer its
presence or absence takes a series of very unusually well controlled experiments
to rule out both observer bias and the presence of feasible alternatives. Likewise,
claims for a lack of learning ability in a certain species could be due to low moti-
vation or be an artifact of the experimental set-up being used, rather than a genuine
inability to learn. Thus, one must be cautious, both in conducting experiments and
in interpreting the published literature.
Not surprisingly, something as widespread and far-reaching as learning eludes
any easy, satisfying categorization, but it can be useful to think of it as coming in
two forms, non-associative and associative. Non-associative learning includes both
habituation and sensitization. Habituation, probably the commonest form of insect

learning is a gradual waning of responsiveness to a stimulus that occurs as an insect,
through experience, finds it to be harmless or at least unavoidable. For example, a
variety of insects such as ants and mantises can be ‘tamed’; if handled frequently
enough, they will come to respond calmly to being picked up and moved about.
Sensitization, its opposite, occurs when repeated presentation of a stimulus enhances
a response to that stimulus and often to related ones.
Associative learning occurs when an organism comes to connect one stimulus
with another (or with a motor pattern) such that its response to the first stim-
ulus changes as a result. Associative learning is typically evaluated in one of
two ways: by classical (Pavlovian) conditioning or by instrumental conditioning
(roughly equivalent to the term, operant conditioning). The former can be thought
of as ‘learning to recognize’ and the latter as ‘learning to do’.
In classical conditioning, a previously meaningless stimulus (the ‘conditioned
stimulus’ or CS) is paired in time and space with an ‘unconditioned’ stimulus that
already elicits a response, often to some form of reinforcement such as reward or
punishment. Afterward, when the CS elicits the same response that the US for-
merly did, one can say that the two have become ‘associated.’ In insects, as in most
other animals, associative learning is undoubtedly an important part of individual
accommodation to a changing environment, a method by which an IRM can be
fitted more exactly to the environmental situation. Associative learning studies are
now receiving new impetus with advances both in the genetics of learning and mem-
ory and in physiological studies of identified neurons and neuronal circuits. Honey
bees, Drosophila flies, and cockroaches (particularly Periplaneta americana) hold
particular places of honor in such studies. Recently, associative learning has taken a
decidedly practical turn with the development of ‘bomb-sniffing bugs’ (Fig. 2.7).
Classical conditioning has been widely exploited in experiments designed to
delimit the sensory capacities of insect; most classical conditioning studies in insects
have used the proboscis extension reflex as the response. A goal of behavioral
neurologists has always been to trace specific behaviors all the way back through
the neuronal level to their intracellular and genetic roots, and learning has been
no exception. However, although Drosophila have provided valuable insights into
the genetics of learning and memory, they have been extremely difficult to use for
physiological studies of identified neurons and neuronal circuits. Cockroaches, on
the other hand, have been useful for cellular-level studies of identified neurons and
circuits, but because the behavior models in use were designed for free-moving ani-
mals, they have not been suitable for studies at the cellular level. However, a 2004
study with cockroaches showed a simple, imaginative approach to bridge this gap.
To study conditioned feeding responses, David Lent and Hyung-Wook Kwon
kept adult male cockroaches hungry for 24 hours, then placed them in small plas-
tic tubes that held their heads in place but allowed their antennae to move freely.
Each tube holding a restrained cockroach was placed horizontally in the middle of
an arena with a red light bulb overhead for warmth and cockroach-invisible illumi-
nation. On the arena wall, a green light-emitting diode (LED) provided stationary
light flashes. A single red LED was used as a control test to determine whether
Fig. 2.7 Bloodhound wasps. Small parasitic wasps such as Microplitus croceipes (above, inspect-
ing host frass) track their prey from the larva’s scent and/or odors in its frass, a mixture of feces and
undigested plant materials. This is a learned response; odors differ when the caterpillar feeds on
different host plants. As a result, by classical conditioning techniques, the wasps also can be taught
to respond to tiny amounts of novel substances that are not biologically relevant. This suggests
practical applications such as the detection of explosives, narcotics, or contraband. After even one
conditioning trial, these wasps learn to associate a novel odor with successful host-finding, feeding,
and oviposition. When exposed to the odor again, they actively antennae and energetically search
the vicinity (bottom left, a wasp responding to odor wafting through a hole). When the odor is not
present or is different, they remain relatively passive or recoil (bottom, right)
sounds from the light switches were being detected. To evaluate their unconditioned
arousal responses, the cockroaches were treated to an odor cue (a puff of air passed
over peanut butter), a light cue (a flash from the green LED), a mechanosensory cue
(a high-current air puff), or auditory cues at a frequency of 1.8 kHz; during each
trial, antennal movements were videorecorded so that antennal angles could later be
measured from the digitized images. Then cockroaches were trained to project their
right antenna toward the green light as the peanut butter air puff was presented.
The cockroaches were surprisingly bright (Fig. 2.8), and quickly learned to asso-
ciate the visual cue through classical conditioning with no other reward than an
attractive odor, which could be explained as intrinsically rewarding in itself. After
just five training trials, 90% showed significant learning, and it persisted unex-
pectedly long. Thirty minutes after training, 60% still showed antennal projection
responses to the CS, and almost half responded even after 72 hours. Optimal learn-
ing performance was achieved when the odor and light cue were simultaneous or
nearly so, which one might expect because in nature, foraging animals detect salient
cues just before or concurrent with the reward.
Fig. 2.8 Remembering

associations. Periplaneta
americana cockroaches held
in the center of an arena
learned to associate a green
LED light with the smell of
peanut butter by one of three
classical conditioning
procedures (i, ii, iii). Then,
after five training trials, at
intervals their learning
performances were assessed
by percentage of antennal
projection responses. The
roaches showed no significant
decline in learning for up to
30 minutes afterward, and
retained a considerable
degree of memory for up to
72 hours afterward
In contrast to classical conditioning, instrumental conditioning involves a rein-

forcing stimulus that is contingent upon the insect’s own motor patterns. An example
would be the way that a bee or butterfly might learn, through trial and error, how
to extract nectar from a newly encountered type of flower. The most widely known
laboratory-based example requires a tethered grasshopper or cricket to move its leg
in response to an electric shock, heat, or access to food. The fact that a headless
cockroach will still learn such a task demonstrates that this type of conditioning can
occur at the level of ganglia.
Scientists who specialize on insect learning recognize various other basic cate-
gories as well. Food aversion learning is particularly interesting because an aversion
can form even when many hours pass between ingestion of a food and the resulting
indigestion.
Spatial learning is an area of particular interest to behaviorists. Bees and wasps,
for example, learn the location of their nests through landmarks recognized and
remembered from a previous orientation flight (Fig. 2.9). Honey bees appear to pos-
sess a topographically organized landscape memory that allows them to navigate
along a previously untraveled route.
The area over which an animal normally wanders in search of food, shelter,
and/or mates is called its home range; within it, others of its own species may be
tolerated and the space thus shared. Among insects, many species that are long-
lived as adults and/or social have a well-defined home range, such as the foraging
area of an ant or termite colony. Tropical Heliconius butterflies have learned home
ranges within which they memorize the location of their food flowers and larval host
plants; adults tend to return nightly to a particular communal roosting place for up to
six months. Similarly, tropical euglossine bees fly along regular trap-lines to which
they return day after day to collect pollen and nectar from newly opening flowers
(Fig. 2.10).
It is sometimes said that observational learning, in which a subject imitates the
motor actions of a demonstrator, is characteristic only of vertebrates. However, this
may be splitting hairs, because social interactions clearly influence what insects
learn. Young foragers learn from experienced ones how to locate profitable flowers
through the well-known honey bee waggle dance (see Figs. 8.13, 8.14). Experienced
Temnothorax ants teach inexperienced nestmates to locate food sources by a tech-
nique known as tandem running, adjusting their speed to that of their pupil. After
watching a model land on a flower, inexperienced observer bumblebees are more
likely to seek out that flower type than an unvisited alternative. Such behaviors are
usually called social learning, defined as learning from other individuals. Social
learning is distinctive because it allows new behaviors to spread quickly within and
between generations. It is faster than individual learning, and saves the fitness costs
of errors associated with inexperience.
Social learning has been more thoroughly studied in mammals, birds, and fish
than in insects, and until fairly recently, insect social learners were thought to
be restricted to social Hymenoptera. However, this view is changing with greater
research effort in this area. In fact, social learners may not even be restricted to
colonial species. For example, a 2005 study of wood crickets, Nemobius sylvestris,
Fig. 2.9 Learning landmarks. After the nest entrance of Philanthus triangulum is ringed with fir
cones, the wasp learns to associate her nest with this distinctive landmark through an orientation
flight (a). While the bee wolf is away hunting, the ring of cones is displaced. Upon return the
prey-laden female flies to the center of the fir cones (b)
Fig. 2.10 Memorizing a trapline. The tropical bee Euplusia surinamensis constructs her nest (1)
130 ft above the ground under loose tree bark. Each day she flies this foraging route through the
Costa Rican forest, visiting ground plants (2, 11, 13), shrubs (6, 7, 9), understory trees (10, 12),
vines (3, 5, 8), and an epiphytic shrub (4) to collect nectar (n) and pollen (p) from newly open
flowers. It is a trapline worth remembering because although each plant produces only one to a few
new flowers each day, flowering may continue for up to six months. Marked females returned to
the plants daily even when flowers were artificially removed prior to their visit, showing that they
actually memorize the trapline
in France showed that these crickets adapt their predator-avoidance behavior after
having observed the behavior of knowledgeable others, and they maintain these
behavioral changes in a lasting way after the demonstrators are gone. Moreover,
evidence suggests that Nemobius are actually more likely to acquire social infor-
mation about the presence of predatory wolf spiders through encounters with other
crickets than to acquire personal information through direct encounters with the
spiders.
Many vertebrates have been shown to learn various tasks and concepts most
easily during a critical period in development; if learning does not occur during
that window of opportunity, it is difficult to nearly impossible. Does a similar phe-
nomenon occur with insects? A few studies suggest it is possible. For example,
Cotesia congregata are braconid wasps that hunt for hosts and mates among plant
foliage. If a young adult female wasp is exposed to wild cherry trees, an inherently
unattractive plant, during the first four hours after she ecloses, she will continue
to search among wild cherry trees rather than cabbage, a normally attractive plant.
Such a general increase in responsiveness is referred to as ‘priming.’
Certainly, for many or perhaps most insects, searching behavior relies upon spe-
cific host-derived stimuli that the adults recognize innately, often during a critical
phase such as the time around adult emergence. However, sometimes cues appear
to be learned by the immature insect, a process called ‘preimaginal conditioning’
and these are subsequently manifested in their responses as adults. Many cases
of preimaginal conditioning have been documented. Early in the twentieth cen-
tury, researchers observed that phytophagous beetles selected oviposition sites on
plant species similar to the ones on which they had been reared. This phenomenon,
sometimes called the Hopkins’ host-selection principle, has been used to explain
host preference or selection in various groups. For example, in one of the earliest
detailed studies on parasitoid learning, when the ichneumonid parasitoid Venturia
(Nemeritis) canescens was reared on wax moths, a species they do not attack under
natural conditions, as adults they preferred the odor of wax moth larvae over that of
their natural meal moth host. Likewise in a much later study, when tobacco horn-
worm caterpillars were trained to avoid the odor of ethyl acetate by pairing it with a
mild electric shock, they retained this response as adult moths.
Does actual learning persist through complete metamorphosis? This is still a
debated topic, and although some research suggests it might be so, the idea has
its skeptics. Metamorphosis is an extreme event; not too long ago, it was gener-
ally thought that a larva’s body essentially turned back into a sort of soup during
the pupal stage, only to be completely restructured into an adult insect. Yet despite
the drastic nature of the metamorphosis process, neurophysiological evidence from
Drosophila and various Lepidoptera that indicates that parts of the brain involved in
learning do remain intact. An interesting one involves a parasitoid that attacks the
codling moth, the insect responsible for the ‘worm’ that is (hopefully not) in your
apple (Fig. 2.11).
The study of Hyssopus pallidus outlined in Case Study 2.3 was one of the first to
study preimaginal learning with such careful attention to detail and control of vari-
ables. For example, a major criticism applied to other preimaginal learning studies
Fig. 2.11 Learning as a youngster. The parasitoid wasp, Hyssopus pallidus, finds the concealed
caterpillar of its codling moth (Cydia pomenella) host by chemical cues in the host’s frass. (above)
Exposure to either apple fruit extract (a) or apple fruit (b) during the parasitoid’s larval stage
resulted in significantly higher response to moth larval frass as an adult female compared to con-
trols (c). (center) Addition of apple fruit extract either to the host moth larval diet (a) or to the diet
of the developing parasitoid (b) also resulted in significantly higher adult female responses to moth
larval frass compared to controls (c) lacking apple fruit extracts. These findings suggest the adult
parasitoid ‘remembers’ its larval experiences
was the difficulty of excluding a ‘chemical legacy’ of contamination from the larval
to the adult environment, especially when learned cues are directly present in the
insect food or in the host cocoon. For this reason, these researchers did not offer
their chemical cue as a liquid or solid substrate, but as a volatile substance that was
removed long before pupation and adult emergence. To counter the possibility that
test organisms might have become contaminated by the relevant chemical cues as
they eclosed into adulthood, in each case the researchers placed their treated filter
papers 4 cm away from the host caterpillar. Because the caterpillar is paralyzed and
cannot move during parasitoid development, and the parasitoids do not leave the
host, the wasps were never in direct contact with the cue-laden filter papers during
the exposure period.
Case Study 2.3: Learning in the Parasitoid Wasp, Hyssopus

A parasitoid’s host-searching ability depends on responses to chemical cues,
and these responses have both genetic and learned components. Is there a
critical phase at which a parasitoid is more likely to learn? And does such
learning carry through from one life stage to another?
To apple growers, the larvae of the codling moth Cydia pomonella are
major pests that burrow into fruits, rendering them unfit to sell. However,
the little parasitic wasp Hyssopus pallidus is a promising candidate for bio-
logical control, because females of this parasitoid have the ability to enter
infested fruit, irreversibly paralyze the feeding caterpillar within its burrow,
and deposit eggs on its body. The hatched parasitoid larvae feed from the
outside of the host caterpillar, develop gregariously in large numbers, then
disperse as new adults throughout the orchard to find new hosts in other
apples.
To find these hosts, H. pallidus relies on its response to traces of frass (feces
and silk) left behind on the surface and in the channels of the infested fruit.
Thus, it was with some dismay that Michela Gandolfi, Letizia Mattiaci, and
Silvia Dorne found strong deterioration of the response to frass in their labo-
ratory culture of H. pallidus after 30 generations on codling moth caterpillars
that were being fed an artificial diet based on wheat germ instead of apples.
Was this behavioral deterioration the result of an inadvertent selection pro-
cess? Apparently not. When the Swiss researchers reared the parasitoids in
the presence of fruit cues, their original level of response was completely
restored within a single generation. This suggested that perhaps the para-
sitoids’ responses had two components. Perhaps the host-derived chemical
components of the frass were recognized innately, but learning was required
before the wasps could recognize the fruit-derived components.
Did the parasitoids need to learn the fruit-related chemicals from fruit-
fed caterpillars and their feces? Or was exposure to fruit itself sufficient?
Gandolfi, Mattiaci, and Dorne infested apples with first instar codling moth
caterpillars, waited until they reached the fifth instar, then cut apart worm-
laden apples to extract them. Rearing the larvae individually in plastic boxes,
the researchers fed some with the wheat-germ based diet alone. These would
serve as their control group. For a second group of larvae, the artificial diet was
treated with a solvent extract of grated apple skin, pulp, and seeds; after the
solvent evaporated, the mixture simulated the apple components of frass. In
each of a third set of boxes, a piece of apple was suspended; in each of a fourth,
a piece of filter paper treated with their apple extract. Then they introduced a
mated female parasitoid in with each host so that it could lay eggs.
When the new generation of parasitoids emerged, they were transferred to
new cages without ever being exposed to apple cues. As a bioassay, when
the mated parasitoid females were four days old, the researchers prepared
filter paper discs by treating them with apple extract or host frass extract,
transferred these second-generation wasps into individual dishes, and placed
the open side of the dishes on top of the filter paper. For ten minutes, they
watched each wasp and recorded the amount of time it spent in ‘searching’,
which they defined as intensive antennal examination of the filter paper.
Comparing the groups, it was clear that rearing Hyssopus larvae on hosts
that were fed either apples or an artificial diet treated with apple extract had a
similar effect—the parasitoids showed a significantly greater adult response to
host frass than parasitoids reared without apple cues (see Fig. 2.11). Did a high
response to host frass require exposure to both the taste and smell of apple?
To find out, the researchers compared the responses of parasitoids reared on
caterpillars fed with apple-supplemented diet to those reared on caterpillars
exposed only to the odor of apple fruit extract. Both of these treatments sig-
nificantly enhanced female responsiveness to host frass cues, indicating that
olfaction alone was as effective as taste and smell together. This was surprising
because studies with other insects had suggested taste to be more important
than smell in the chemical learning process.
Was there a sensitive window in the parasitoid’s life history for learning the
relevant stimuli? Gandolfi, Mattiaci, and Dorne reared another set of wasps on
caterpillars fed on plain artificial diet, but this time they exposed the wasps to
apple fruit extracts at different developmental stages. For comparison, they
used both parasitoids exposed to apple extract during their entire develop-
ment and parasitoids that were never exposed at all. As an extra precaution,
a parallel series of experiments was conducted in which female wasps were
introduced into untreated vials and into vials from which extract-treated fil-
ter papers had just been removed after having been in the vial for 12 hours.
Then the searching activity of the wasps in these two groups was bioassayed.
The lack of any statistically significant difference between them satisfied the
researchers that no contamination had been left on the inner surfaces of the
vials after the treated filter papers were removed.
As they emerged, the new adults were either immediately transferred to

new cages or allowed to spend one additional day in their rearing situation
before transfer. Then, when they were four days old, the mated females were
tested for their ‘searching’ response to extracts prepared from natural frass
produced by codling moth caterpillars feeding on apples. When the results
were analyzed, it was clear that exposing Hyssopus to apple extract at differ-
ent developmental stages affected the response of adults to host frass cues
in different ways (see Fig. 2.12). As before, parasitoids never exposed to
the apple extract at all during development showed no significant response
to apple extract as adults, but those exposed to extract during development
responded strongly to it as adults. The surprise came when the exposure times
were compared. Exposure at the beginning of development was the most
effective—as effective, in fact, as exposing the parasitoids to extract through-
out their entire development. However, exposing them either at the pupal
Fig. 2.12 Remembering through metamorphosis. Codling moth parasitoids, Hyssopus pal-
lidus, exposed to odors from apple fruit extract at different developmental stages differ in their
response as adult females to frass extracts of their apple-fed moth larvae hosts. Those exposed at
their earliest stages of development (egg to late larva-pupa) gave significantly stronger responses
as adult females to larval frass than those exposed only after pupation or as adults; the latter
responding no differently than controls with no exposure to apple fruit extracts at any time in
their lives
stage, during emergence, or at the adult stage one day before the bioassay
was ineffective, even though in these cases the interval between removal of
the bioactive cues and the bioassay was much shorter.
Some other studies had found cases in which an insect showed preimaginal
learning of some cues and not others; the presumption was that learning had
to be biologically relevant if it was to occur. Gandolfi, Mattiacci, and Dorne
wondered if this would be true for Hyssopus. Was preimaginal learning con-
fined to fruit cues, or was it a more general phenomenon? They repeated their
studies, this time exposing parasitoids from egg to larval stages to menthol, a
compound not detectable in apple extracts or codling moth caterpillars. As
a control, they used only the solvent. Parasitoids that were never exposed
to menthol during development were repelled by it as adults, and actually
responded more to the control than to the menthol. However, when parasitoids
had been exposed to its odor from egg to larval stages, as adults they showed a
significantly higher tolerance to menthol than did parasitoids without menthol
exposure, and responded to it in essentially the same way they responded to
the filter papers with the solvent control.
Thus, it appears that the parasitic wasp Hyssopus pallidus learns to locate
its concealed host by zeroing in on its frass, and this recognition is a two-
stage process. A high response requires both host-derived chemicals in the
frass that are recognized innately, and apple-derived chemicals that must be
learned through exposure early in the parasitoid’s life cycle. The preimaginal
learning is a general phenomenon in which the wasps are capable of learn-
ing a variety of cues. While chemical cues are indispensable, for this wasp
species at least, physical stimuli such as visual or tactile cues seem to be
of only minor importance. The parasitoids did not need to contact the fruit
directly; learning occurred just as well when they were reared on hosts fed
an apple-supplemented artificial diet. In contrast to general expectations, the
sensitive period for learning these olfactory cues was not during eclosion or
early adulthood, but in the earliest stages when the wasp was developing from
egg to young larva.
The codling moth parasitoids exposed to odors from apple fruit extract at differ-
ent developmental stages seem to show that some sort of learning persists through
metamorphosis. Or is this not ‘really’ learning? The phenomenon is undoubtedly
real; the definitions used to describe the effects of experience on insect responses
may be the cause of confusion. One could consider it to be a case of priming, in
which certain experiences make the parasitoids more responsive to foraging cues.
Alternatively, one might call it preference (and avoidance) learning, a case of asso-
ciative learning where the increase in responsiveness is specific for the cues that
the insects encounter during the experience. Then again, one might say it is at least
analogous to sensitization to chemical stimuli, a mechanism well known among
adult insects in which the response to a stimulus gradually increases with repeated
exposure to that stimulus.
Such studies offer tantalizing glimpses into the complexity of neuroethology,
where questions outnumber answers. Studies that couple behavioral research with
detailed neurophysiology and genetics are still embarrassingly meager, and have
almost entirely concentrated upon a mere handful of insects from four of the 28
or so insect orders. These animals—mostly cockroaches, grain beetles, fruit flies,
blowflies, and a few of the many ants, wasps, and bees—have been chosen not
because of their representative nature but primarily for convenience, availability,
ease of rearing, and sometimes, perceived importance to society.
2.3.3 Insect Intelligence

Honey bee workers can learn signals quickly and with apparent ease in every known
sensory modality. Multiple tasks pose few problems; they can link up to five differ-
ent visual signals with correct turns in a maze. Nor do sequential tasks pose great
difficulty, for bees may learn to visit a single location up to six different feeding
times during the day or four different places at four different times. Some ants, such
as Formica pallidifulva workers, can learn a six-point maze at a rate only two to
three times slower than that of a relatively advanced vertebrate, the laboratory rat.
Female Ammophila wasps that have made an inspection visit to a temporarily closed
nest burrow can still remember and properly perform the action determined in that
single visit, even 15 hours later.
Faced with a myriad of such examples, it is easy to succumb to a sense of wonder
and conclude that some insects are comparable to vertebrates in intelligence. (It is
small wonder that many fiction writers have done so.) However, these are but frag-
ments of information from the insect’s behavioral repertoire, and intelligence in this
sense of the word is a meaningless measure. In fact, in this sense animals do not
even possess varying degrees of intelligence; rather, they possess specific abilities
to learn selected things. For any organism there are certain types of tasks that can
be mastered and others that cannot. This is no less true with insects. Honey bees,
which can learn quickly to orient with respect to attractive odors, cannot do so at
all with respect to repellents. Thus, one is often struck by an apparent paradox—
the contiguous association of a seemingly startling propensity for learning and very
inflexible behaviors. For example, the predatory wasp Liris nigra is no less able than
others of its kind to utilize latent learning to relocate nests after foraging trips. Yet
in prey capture, which normally involves stinging a cricket in four ganglia ending
with the subesophageal ganglion, the wasp appears intensely frustrated when con-
fronted with a decapitated prey. She will sometimes spend more than an hour vainly
searching the cricket for the site of the missing ganglion.
Intelligence can be defined in various ways, but it generally includes the ability
to learn facts or skills and apply them in a rational way, that is, to generalize learned
information by transferring it from one set of circumstances to another. Insects do
not measure up very well in this regard. Their ability to transfer memories to assist
in the learning of new situations is nearly or totally absent.
Consider Ammophila pubescens, a European wasp that raises its young in bur-
rows that it digs in the soil. Unlike most other digger wasps, it attends to up to 15
nests at a time, remembering the locations of each. Because the nests are begun
at different times, the young inside may range from a newly hatched larva capable
of consuming but a single caterpillar a day to an older offspring requiring three to
seven caterpillars, or a full grown larva does not need more food but should have
its burrow sealed so it can safely mature underground. The ability to do so many
complex things at once would seem to indicate a fairly high degree of intelligence,
but rationality is not this wasp’s strong suit.
Watching Ammophila closely in the early 1940s entomologist G. P. Baerends
noted that the female begins each day with inspection visits, flying from nest to
nest and inspecting the contents of each before she begins to bring prey to any
of them. Might this behavior provide a clue to her ability to multitask? Baerends
located a series of nests being attended by a single female and carefully replaced
the nests with plaster of Paris casts he could open and close. In the following days
he spent many hours substituting larvae of various sizes for each other and adding
and removing caterpillars the female had brought. Very quickly it became apparent
that the single morning inspection visit set the wasp’s behavior for the rest of the
day. Substituting a larger larva after the inspection visit made absolutely no differ-
ence, nor did it matter whether he added prey or removed some that she had already
brought. If her inspection visit indicated permanent closure, this she would do, even
if the larva had subsequently been removed so she now was sealing a completely
empty nest! Thus, provisioning behavior in Ammophila is paced to meet the needs
of each growing larva in a way that is both sophisticated and restricted. Each step
in the unfolding behavior patterns of the day is guided by the sign stimuli present in
the single brief examination of nests, and after this each appropriate motor pattern is
performed in a genetically determined sequence. In nature, in the absence of med-
dling entomologists, the information obtained through a single daily visit would be,
and clearly is, sufficient.
Insects also appear almost entirely incapable of insight, that is, reorganizing their
memories to construct a new response in the face of a novel problem. The French
naturalist Fabré (see Chapter 1) demonstrated this long ago with another hunting
wasp that briefly drops her cricket near the hole to enter her burrow before reap-
pearing to drag it inside. While she was inside the burrow, Fabré moved the cricket
a short distance away. With insight, a wasp might be expected to recognize that
her prey had merely been moved and that her burrow, having been inspected, was
ready for stocking. Instead, when the wasp reappeared, she returned the cricket to its
proper place at the edge of the hole once more, then descended again, alone. Fabré
moved the cricket; again, the wasp reappeared to reposition it and reenter her hole
alone. Fabré re-elicited this response 40 times before he lost patience!
For a long time during the twentieth century, the ‘nature–nurture’ controversy
was vigorously debated; instinct and learning were pictured as diametric oppo-
sites in command of the behavior of different kinds of animals. However, the
controversy really is a spurious one, and few behaviorists nowadays seek to deter-
mine whether a particular behavior response is learned or instinctive. Rather, it is
2.4 Hormone-Based Coordination 83
generally agreed that behavior is rarely determined either solely by the type of out-
side events impinging on the individual or by inborn heredity alone, but rather by
interaction of the two. A high reliance upon learning ability may be adaptive when
an animal is relatively long-lived and/or faced with a good deal of uncertainty or
variability regarding aspects of its environment that are biologically significant.
In contrast, the advantage of a high reliance upon innate behavior patterns lies in
their reliability. When a particular environmental cue can be linked dependably to
a biologically appropriate response, innate fixed action patterns are certain to be
successful responses. Selection also will favor innate behavior when the cost of an
initial mistake is high, such as would be true in the case of the cockroach and moth
escape reactions. Furthermore, innate behavior presumably permits economy in the
nervous system.
2.4 Hormone-Based Coordination
Reproductively active adults of the lacewing Chrysopa carnea are green, but they
turn brown when they enter diapause in the autumn; in the spring, when they become
active, they again turn green. A grasshopper typically goes through a series of five
to seven molts over several months between hatching from its egg and becoming a
sexually mature adult; if its prothoracic gland is removed when it is still an early
nymph, its development halts, and its behavior never matures. A female cockroach
produces chemical sex attractants; if the corpora allata in her brain are surgically
removed, she is incapable of attracting males and will not mate.
It is well known that insects rely heavily upon circulating chemistry. They pos-
sess a traditional endocrine system that produces blood-borne chemicals that act
on receptor-bearing target tissues elsewhere in the body. They have nerve cells that
rely upon chemical neurotransmitters to propagate their messages (and thus could
be considered a hormone that is acting locally within the synapse). And they have
functional hybrids of neurons and endocrine glands called neurosecretory cells that
cluster in their brains and throughout their central nervous systems.
Hormones, by classical definition, are substances secreted by endocrine glands
and transported by the circulatory system to other body parts, where even tiny
quantities evoke physiological responses in target tissues. A broader definition
recognizes that in addition to endocrine glands, single cells can produce hor-
mones. Various body organs, especially those associated with reproduction (such
as ovaries, testes, spermatheca, etc.), also are known to have a secondary endocrine
function.
Hormones connected with growth, development, sexual maturation and repro-
duction (see Chapter 9) have received the most study, but internal hormone
secretions have been implicated in nearly every aspect of insect life history, includ-
ing important controls on migration, orientation, and periodic behaviors as well as
activation of adult behavior.
Behaviorally, hormones often act as primers or modifiers that start the internal
motivation of an insect to later perform a particular behavioral act. Strong evidence
is obtained when removal of a particular endocrine center abolishes a specific behav-

ior that can then be restored through gland implantation or hormone application.
For example, the corpora allata are a pair of endocrine glands behind the brain.
Their removal (an operation termed allatectomy) from last instar or newly emerged
females of Gomphocerippus grasshoppers completely prevents the onset of normal
sexual receptivity. Females so treated respond to courting males by kicking defen-
sively and trying to escape. However, implantation of actively secreting corpora
allata into such females changes their behavior toward males to a state of copula-
tory readiness. Allatectomy of previously sexually receptive females leads to loss
of this receptivity and reappearance of defensive kicking within six days. Thus, one
can state with fair confidence that for these grasshoppers the internal impetus toward
mating is directly mediated by secretions from the corpora allata.
A useful conceptual overview places genes at the beginning of all behavior, pro-
ducing their effects by establishing the parameters that will govern (1) the nature
of particular neural pathways that are responsible for the manner in which stimuli
are processed, and (2) the production of hormones that trigger the expression of
the fixed action patterns present in the nervous system. These hormones can inter-
act with the nervous system and the insect’s physiological state immediately (in
which case they are called releasers) or belatedly (when they are called modifiers);
in either case, the outcome is that they activate appropriate behaviors. Thus, for
example, while she is developing her eggs, a female mosquito releases one hor-
mone that temporarily keeps her from seeking new hosts until after they are laid;
a second hormone increases sensory sensitivities that help her find a site to lay
them.
Many functions of an insect’s life are under endocrine regulation. Here, we
will concentrate on molting and metamorphosis, two developmental behaviors well
suited to physiological and biochemical control by the hormonal system, which
tends to be slower and more sustained than primarily nerve-mediated responses.
At the same time, remember that the nervous system still exerts an influence on
both events through nervous feedback and by secreting neurohormones.
It is an appropriate place to start, because all insects share a fundamental prob-
lem. They have an external skeleton, and as they grow, it becomes too small.
Consequently, they must grow a new, larger, more flexible exoskeleton and then
shed (molt, ecdyse) the old one. Most insects also metamorphose into an adult form
at the last molt. For some groups, this represents no great change beyond adding or
maturing functional sex organs; for others, it is a drastic revolution. New structures,
musculature, and nervous innervations appear; food choices differ, new sensory
receptors arise, and entire behavioral repertoires change. Many larval neurons die;
others are remodeled; new adult neurons proliferate from neuroblasts. All of this
alteration is linked to endocrine changes and hormones.
Hormones control not only the molt, but also the development or degeneration
of many systems that affect behavior. One was discovered when pioneering insect
physiologist V. B. Wigglesworth implanted an active corpus allatum into last instar
larvae of a bloodsucking bug, Rhodnius, and found that they molted into additional
larval instars rather than producing adults. At first Wigglesworth thought the new
hormone was an inhibitor, but as soon as he realized its true effect he named it
juvenile hormone (JH) to reflect that fact. Juvenile hormone is now recognized to
be not just a single hormone, but a group of sesquiterpenes having at least six major
members. These mediate a wide variety of functions in addition to metamorphosis.
Another, ecdysone, was the first insect hormone to be structurally identified.
As new techniques have allowed others of its type to be isolated, ecdysteroid has
become the preferred name for the group. Ecdysteroids’ primary claim to fame is
that they directly affect gene expression and thus trigger the separation of the epider-
mis from the overlying cuticle, which begins the molt. Like JH, they now have been
shown to have wide-ranging effects at every developmental stage but particularly in
embryogenesis and other reproductive and developmental tasks. The primary site of
ecdysteroid synthesis is the prothoracic gland, but despite its nervous connections,
the primary mode of gland activation is hormonal. The classic understanding is that
when a winged (pterygote) insect becomes a full adult, it no longer molts, so this
gland is no longer needed. Exposure to ecdysone in the absence of JH then leads to
apoptosis (programmed cell death) that causes the gland to degenerate. Apterygote
insects, on the other hand, continue to molt as adults, and they retain their active
prothoracic glands. Recent discoveries add complexity to this picture but do not
change its general outlines.
A third important hormone goes by the acronym PTTH (short for a rather
unwieldy term, prothoracicotropic hormone). It actually was the first insect hor-
mone to be discovered, but the last major hormone to be structurally identified. In
1917 Stefan Kopec reported that the brain of gypsy moth caterpillars was neces-
sary for successful pupation, based on his well designed experiments in which he
surgically removed brains from some larvae while controls had sham surgery (an
incision made, but the brain left in place). He also tied silk string tightly around var-
ious parts of gypsy moth larvae, then watched their subsequent development. When
a caterpillar was tied around its midsection early during its last larval instar, only
the head half later pupated; when tied late during that instar, both halves pupated.
Based on this, Kopec proposed the concept of a critical period, a time period when
the brain was necessary for its hormonal influence to be exerted. Removing the brain
before the critical period prevented development but afterwards, it had no effect. He
postulated a brain hormone was responsible.
Unfortunately, Kopec was ahead of his time, and his work was poorly accepted.
The prevailing wisdom was that insects had no hormones. Furthermore, everyone
‘knew’ the nervous system and the endocrine system were functionally distinct;
brains and nervous tissue certainly did not produce hormones. Neurons secreting
chemicals at their synapses had not even been envisioned. When later work by others
showed that the brain produced not just one but many different hormones, even his
terminology fell out of favor.
Today, however, Kopec would be vindicated. His ‘brain hormone’ has been
demonstrated to be PTTH, and it shows a satisfying complexity. The PTTHs from
only a few insects have been identified, but they appear to fall into two size groups.
The ‘small’ PTTHs were renamed bombyxin; their exact role is unknown, but they
appear insulin-like and may be involved in ovarian development and the utilization
of carbohydrate during egg maturation. The ‘big’ PTTHs show the activity for
which they were named, acting on the prothoracic gland to regulate the synthesis
of ecdysteroids. When an insect receives the proper stimuli from the environment
(such as photoperiod or temperature) or from its nerves (such as stretch receptors
firing after an assassin bug’s blood meal), it releases PTTH. This in turn activates
the prothoracic glands to produce ecdysteroid, triggering the molt. Interestingly, for
many insects that undergo pupal diapause, PTTH release is regulated by photoperiod
during a circadian window.
Ecdysteroids, juvenile hormones, and PTTH are still recognized as the three
major hormone groups regulating insect development and reproduction, but other
important hormones that play more specialized roles are continually being dis-
covered. Though their functions have yet to be fully elucidated, many hormones
previously isolated from vertebrates have been found. Insects produce insulin and
melatonin. They have counterparts to vertebrate adrenaline and noradrenaline. Their
reproductive hormones are close enough that female rabbit fleas depend on the hor-
mones circulating in the blood of the pregnant rabbit host in order to reproduce.
Considering that a hundred years ago, insects were thought to lack both hormones
and brains, it is somewhat humbling to learn that they have counterparts to so many
vertebrate systems—as it is to recognize that, because they are the more ancient
group, it would be more correct to salute the insect-type life processes that are found
in us.
2.4.1 Clocks and Reiterative Rhythms
After 17 years of silence underground as nymphs, great numbers of periodical

cicadas emerge to fill the air with their raucous noise. The snowy tree cricket (see
Fig. 8.2) sings in such a rhythmic tempo that one can ascertain the temperature on a
summer’s eve by listening to its chirp with a watch in hand. Silverfish scurry about
each night, resting in cracks and crevices during the light of day. Much of what
insects do is rhythmic, from heartbeats and songs to cycles of sleep and wakefulness,
to reproductive cycles measured in weeks, months, or even years.
Intuitively, one can see the adaptive advantages of most cycles, but how are they
initiated and maintained? In discussions of biological rhythms, the term clock is
widely used as a convenient analogy. It does not refer to a specific physiological
mechanism; instead, it is a convenient catch-all term to describe the largely unknown
biochemical systems driving the rhythmic cycles. The discovery of a great many
‘clock mutants’ in Drosophila melanogaster that show different circadian cycle
lengths (or no rhythm at all) confirms that the periodicity is innate and genetically
coded.
Although further research will probably show many variations in the clocks of
different species, investigators are closing in on a general description of the system.
Working with vertebrates as varied as hamsters and humans, biologists have iden-
tified an area in the hypothalamus of the brain called the suprachiasmatic nucleus
(SCN) that appears to contain a master clock. The SCN clock operates via rhythmic
changes in the activity of three genes. The first two, called per and tau, respectively,
code for a protein (PER) and an enzyme that degrades PER. Production of PER
varies over a 24-hour schedule, gradually building up inside and outside the cell
nucleus. When PER is in peak abundance in the cell, tau turns on the enzyme’s pro-
duction, slowing PER’s rate of accumulation. At the same time, extra PER bonds
with a protein coded by a third gene, tim. In this bonded form, PER is carried back to
the nucleus, where it blocks the activity of per, the gene that produced it. However,
this blockage is only temporary, and soon a new cycle begins. Surprisingly, these key
genes of vertebrates have also been found in both Drosophila and honey bees, where
they play an equally critical role in enabling circadian rhythms. Thus it appears
likely that they arose early in evolutionary development, perhaps from an ancient
animal that lived about 550 million years ago.
Reiterative rhythms occur with a regular repeated periodicity in the life of a sin-
gle individual. Some reiterative cycles cover a relatively long period, such as the
lunar periodicity of nesting behavior in the nocturnal sweat bee, Sphecodogastra tex-
ana (Fig. 2.13). In many temperate insects, seasonal rhythms are based on changes
in photoperiod, involving an assessment of the duration and accumulation of daily
changes in day or night length over a period of time. At the other extreme, reiterative
cycles of feeding and locomotion may be very short. Typically, however, reiterative
Fig. 2.13 Responding to moonlight. The sweat bee Sphecodogastra texana shows lunar peri-
odicity. In good weather adult female activity always begins about sunset. Concentrated pollen
collection and nest cell construction occur only during that part of the lunar cycle where moonlight
is continuous with the twilight, thus permitting extended nocturnal foraging. When the moon rises
after the close of twilight (ca. 9:30 pm), the bees close their nests at the end of twilight and make
very few pollen collections from their host plant (the evening primrose, Oenothera rhombipetala)
even when pollen is abundant. Brood development in excavated nests correlates with the observed
foraging activity
behavioral rhythms have a periodicity of about 24 hours and thus are called cir-
cadian rhythms, derived from the Latin circa, about, and diem, day. For example,
anyone who has cohabited with cockroaches has noticed their characteristic pattern
of activity that begins shortly after dusk. The timing of moths’ sex odor release is
also on a circadian cycle.
Circadian rhythms all share some important characteristics. First, although these
cycles average about 24 hours, they are not exactly so, and will drift out of phase or
free run for some time unless they are entrained (kept in line, reset) by certain exter-
nal environmental cues. Crickets are a well-studied example. Each evening under
natural conditions, males begin singing at about the same time relative to nightfall.
In a room with constant bright light, although they still sing, each day their call-
ing starts about 25–26 hours later than it did on the previous day. The very fact
that the length of a free-running cycle is not exactly matched to the earth’s many
24-hour environmental cycles provides strong evidence that internal components
drive a given rhythm. However, it does not prove it, because one cannot always
exclude other environmental cues with certainty.
The cyclical environmental cue that entrains a circadian rhythm is called a
Zeitgeber. Most circadian rhythms studied to date have relied upon light/dark
transitions as their Zeitgeber, but other cues may also prove important. Some insects
held in continuous darkness, for example, distinguish between long and short days
in a temperature cycle. However, in general, circadian rhythms are temperature
compensated. Unlike most other physiological processes, within normal biologi-
cal limits the periodicity remains stable under changing temperature conditions.
Consider the midge, Clunio marinus, that spends its larval and pupal life among
red algae between the tidemarks of certain European beaches. Adults emerge only
at low tide; they live for just 2 hours and must mate and oviposit before the tide
advances back over their breeding ground. A number of genetically distinct geo-
graphical races exist, each with its own emergence rhythm precisely adapted to its
local tidal conditions. In southern races the daily periodicity is circadian and free
runs in constant light. In the Arctic, however, emergence is strictly tidally con-
trolled and stops in constant light or dark; it is as though an internal alarm clock
were measuring 10 hours from the time of first exposure to the previous ebb tide.
Subsequent research has shown that the receptors for circadian rhythms and for
seasonal activity are located in the brain of some insects, in the compound eyes
of others, and in other cases, in both. Cryptochrome, a light-sensitive pigment that
contains the vitamin riboflavin and a protein, is a major player in sensing circadian
light information. Interestingly, in Drosophila, cryptochrome in the brain regulates
the morning activity rhythm and cryptochrome in the compound eyes controls the
evening activity peak.
2.4.2 Gated Rhythms
Many behavioral events, such as the periodical cicada’s emergence (see Plate 44) or
the mass emergence of mayflies from a pond, look like a rhythm when one views the
whole synchronous population, but occur only once in any individual’s life. These
one-time events are called gated rhythms, because at the appropriate stage of the
life cycle it is almost as though some mechanism were opening a behavioral ‘gate’
at the appointed time. All those individuals streaming through are allowed to begin
a given behavior, but an individual that misses a given gate opening time must wait
for the next one.
Ideally, one attempting to investigate gated rhythms should seek an organism
with a relatively simple nervous system, choose neurons important to a clear-cut
behavioral alteration, and then follow the changes occurring in those neurons during
the course of hormone action. In the 1970s, two researchers found their model sys-
tem in giant silk moths (family Saturniidae), not to be confused with the silkworm
moth, Bombyx mori.
Case Study 2.4: Initiation of Adult Behavior

in Saturniid Moths
An adult silk moth acts very different than the pupa it previously was. How do
hormones and the nervous system interact to produce this change?
In moths, as in most insects, emergence of an adult from its pupal skin
occurs only during a specific period of the day; individuals that do not emerge
during that particular time cannot eclose until the proper time on the following
day. However, simple escape from the pupal skin does not herald the arrival
of a fully functional adult. A newly eclosed individual displays little in the
way of adult behavior and will make only immature rotary movements of
its abdomen. The problem is not lack of appropriate neural machinery; it
possesses a fully developed adult nervous system. Nor is the lack of adult
behavior merely due to the insect’s restraint—peeling the pupal cuticle away
from the mature moth within still does not mature the behavioral repertoire
before its normal emergence gate. Complex motor patterns such as flight or
walking, on the rare occasions when they can be elicited at all, are uncoordi-
nated and abortive. Even simple reflexes, such as the ‘righting reflex’ when
overturned, are missing.
At the arrival of the eclosion gate, however, the behavior of a ‘prematurely
peeled’ moth changes strikingly. In perfect pantomime, the moth sheds its
phantom pupal skin and escapes from its nonexistent cocoon. At the end of
the performance, it inflates its wings and assumes the full repertoire of adult
behavior. This behavior suggested that adult actions must be controlled in a
central manner that was closely linked to the timing of eclosion.
Hoping to locate the site of the eclosion clock, James Truman and Lynn
Riddiford performed a series of surgical ablation studies on various parts of
the pupal nervous system. When the brain itself was removed shortly after the
onset of adult development, development proceeded normally and the result-
ing moths went on to shed their pupal skins. However, their emergence was
quite abnormal; some usual behaviors were entirely omitted and others were
out of sequence. Even more striking, the eclosion was no longer gated. Moths
emerged at all sorts of odd hours randomly distributed throughout the day and
night.
Was all this behavioral confusion simply due to removal of important neu-
ral centers? The researchers tackled this suggestion by implanting loose brains
into the abdomens of debrained pupae. In these ‘loose brain’ moths, neural
connections between the brain and nervous system were never established. Yet
the resultant moths emerged at the proper gating time and displayed proper
emergence behaviors! Clearly, the appropriate messages were being carried
through chemical channels rather than along nerve fibers. An eclosion hor-
mone appeared likely, and the case was strengthened when injecting brain
homogenates into moths prior to their normal eclosion time resulted in moths
emerging early.
But what was the actual role of the brain? For example, did the brain
include the photoreceptor, the clock measuring the time after lights-on
or lights-off, or both? Knowing that different species of giant silk moths
have quite different eclosion gating times, the researchers performed brain
transplants between species. The results (Fig. 2.14) clearly confirmed their
suspicions. The moth brain contained the gating clock. By interchanging the
brain, one could interchange the time of emergence.
Fig. 2.14 Interchanging brains. Hyalophora cecropia silk moths eclose in the morning, whereas
Antheraea pernyi ecloses just before dark. When their brains are removed, eclosion rhythmicity
is abolished in both species. When their brains are reimplanted in their abdomens, rhythmicity
returns even though there are no nervous connections between brain and CNS. If the brains are
removed and implanted into individuals of the opposite species, the restored rhythmicity is that of
the donated brain rather than the recipient body. The hormonal action is not species-specific, only
the timing of its release
Were there also photoreceptors in the brain? Removing a moth’s compound

eyes did not halt its response to light–dark cycles. Truman implanted brains
into the head region of ten previously debrained pupae, and into the abdomen
of another ten. Then he plugged the pupae into holes in a partition and placed
them into a chamber so as to expose their anterior and posterior halves to
different photoperiod regimes. In each case the subsequent eclosion of the
host depended only upon the photoperiod to which the brain was exposed.
The research outlined in Case Study 2.4 concentrated on one critical point of
development. However, molts between larval instars, from larva to pupa, and from
pupa to adult differ in the degree of internal morphological change that occurs. The
system usually runs smoothly, but errors do occur. As noted for the assassin bug
Rhodnius given supplemental juvenile hormone (Section 2.4), insects sometimes
molt more times than usual, becoming ‘supernumerary larvae’ rather than changing
into adults as expected.
How is a proper degree of change regulated? The brain is in charge, as Truman
and Riddiford demonstrated. Upon receiving proper stimuli, it secretes PTTH that in
turn stimulates the prothoracic glands to synthesize and secrete ecdysteroids. These
combine with a receptor protein in the cell nucleus, bind with DNA, and induce tran-
scription of a few master genes. These transcripts, in turn, start a flurrying cascade
of gene activity that ultimately results in everything from changes in morphology
and physiology of internal organs to secretion of a new cuticle. However, it is
juvenile hormone—through its timing and quantity at target cells—that modulates
ecdysteroid-induced gene switching so that the proper amount of change occurs.
More exactly how it does so at the gene level is a matter to be elucidated by current
and future research.
Chapter 3
Spatial Adjustment
3.1 Introduction
The burrows of most wood-boring cerambycid beetle larvae are very irregularly ori-
ented; why? At temperatures so low that most other insects are inactive, bumblebees
still fly about; how? Migrating locusts appear to swarm single-mindedly toward a
fixed goal. Do they?
Though the scale of movement varies widely, the ability to change position
within the environment is essential to the survival of nearly every animal, including
most insects. Escaping predators, gathering food, locating a mate, adjusting to envi-
ronmental variables such as temperature and humidity—these and other important
behaviors all depend upon an insect’s ability to adjust its spatial relations.
One of the most generally known facts about insects is that they possess three
pairs of legs. This is, in fact, the fundamental ground plan of insects, and one that is
amply represented in the fossil record. It is derived from an ancestral arrangement
in which serially uniform legs occurred on the majority of body segments. Over
time, some legs became modified into various appendages such as mouthparts, tho-
racic legs, genitalia and cerci, while others on the abdominal segments typically
were lost. With the passage of further evolutionary time, insect thoracic legs have
developed an enormous diversity of structure and function (Fig. 3.1). In addition
to differences between taxa, variation can be found within an individual, between
larvae and adults, and between males and females.
The acquisition of wings was a second major development, the importance of
which can hardly be overstated. Insects were not only the first organisms to develop
the capacity for powered flight, they remain the only group of invertebrates to pos-
sess this ability. This development opened the third dimension to insects, setting the
stage for improvements in such diverse but crucial behaviors as dispersal, escape,
thermoregulation, feeding, and mate location. It probably also led to an expansion
in neural capabilities. It has been noted that some of the most ‘intelligent’ insects
(i.e. those that are most capable of learning), as well as those with the most acute
vision and olfaction, are found among actively flying predators and pollinators.
Together, diversification in the structure of legs and wings undoubtedly has been
a key factor in the overwhelming success of insects worldwide. Through variations

94 3 Spatial Adjustment
Fig. 3.1 Showing a bit of leg. Diversity of insect legs and some of the purposes of their modifica-
tions. (A) Drosophila cracens foreleg—courtship, (B) thrips foreleg—crawling, (C) grasshopper
hind leg—jumping, (D) Corixa water boatman hind leg—swimming, (E) Cicindela tiger bee-
tle hind leg—running, (F) cicada nymph fore leg—digging, (G) Rhagovelia obesa water strider
middle tarsus—walking on water, (H) Ranatra fusca water scorpion fore leg—prey capture, (I)
caterpillar thoracic leg—crawling, (J) honey bee hind leg—pollen transport
on their basic body plan, insects have surmounted the barriers to inhabit nearly every
terrain but deep saltwater. Moreover they have done this so elegantly that they have
attracted the attention of engineers who study insect locomotion as inspiration for
moving robotic devices (Fig. 3.2).
3.2 Locomotion
A tiny flea’s jump may be 13 inches long. A blood-sucking bug, Rhodnius, may
move about with a meal 10–12 times its own body weight, corresponding to a human
drinking 200 gallons and subsequently weighing nearly a ton. Click beetles can
suddenly flip into the air to a height of four times their body length. Cockroaches
have been clocked at speeds of nearly 3 miles per hour—a remarkably high speed
in relation to their body size.
3.2 Locomotion 95
Fig. 3.2 Creating a useful mimic. Hexapodal robots have potential for performing many tasks
such as exploring terrain of distant planets. Around the world, robots such as this one are being
developed that mimic the analogous working morphology of common insects. The wireless camera
mounted atop the robot’s rear monitors the environment and gives the robot the appearance of
having a stinger like a scorpion
Many insect locomotory activities appear extraordinarily impressive by human

standards. Are insects endowed with comparatively tremendous muscular power or
a different sort of muscle from those we possess? Not really. Physiological studies
have shown that their muscles are quite similar in almost all respects to our own,
although the insect may possess many more individual muscles than a human does.
Instead, many of the strange powers insects appear to have (as well as many of the
problems they face) are the consequence of a simple physical relationship between
surface and mass. As the size or mass of any object diminishes, the relative amount
of its surface increases. (The volume of a sphere is 4/3π r3 , where r is the radius of
the sphere; the surface of a sphere, however, is 4π r2 .)
For an organism as small as an insect, this surface to volume relationship has a
marked effect on muscle power. The power of a muscle is proportional to the area
of its cross section, whereas the mass it has to move is proportional to volume. We
are amazed by the long jump of the flea, which proportionately carried out by a man
would cover 800 ft. However, we must realize that a flea the size of a man would
have relatively much more mass per unit cross-sectional area of muscle than does a
normal-sized flea.
3.2.1 Terrestrial and Aquatic Locomotion

Houseflies easily walk upside down across the kitchen ceiling. Mole crickets burrow
through the soil with their powerful, spade-like forelegs. Many aquatic insects such
as dytiscid and gyrinid whirligig beetles use their flattened, fringed legs as oars for
paddling or swimming. Although a few insect species have secondarily lost most
or all of their use of their legs as locomotory appendages, walking and running are
common behaviors for the adults of nearly all flying and non-flying insects and for
many immature forms as well.
The power for most terrestrial locomotion comes from the thoracic legs, which
move in various sequences at different speeds so that stability is always maintained.
Coordination of these patterns, understandably crucial, is mediated both through
central mechanisms (often termed central pattern generators) and through segmen-
tal reflexes. Two general principles appear to underlie the walking sequence. First,
no leg is raised until the leg behind it is in a supporting position. Second, the move-
ments of the two legs of a segment alternate. A pattern of alternating triangles of
support is commonly observed; with never fewer than three legs on the ground, an
insect can stop at any point without losing stability. Stability is also enhanced by the
fact that the insect body is slung between the legs in such a way that the center of
gravity is low.
Walking is done by moving the three legs contacting the ground backwards while
the other three legs are raised and moved forward. This propels the insect’s body
forward, and when the raised legs are all the way forward they lower and make
contact while the legs that were down are raised and the whole pattern repeated.
Many immature insects move in a manner similar to adults. However, because an
insect’s thoracic legs can function only when the external skeleton is relatively rigid,
soft-bodied larval forms generally employ somewhat different methods. Many lar-
vae crawl, moving about by changing the shape of their body as a result of muscles
acting against the body fluids. Caterpillars and sawfly larvae have thoracic legs, but
also have hollow cylindrical outgrowths of their abdominal wall. Waves of contrac-
tions pass along their body from back to front, progressively lifting body segments
and these fleshy prolegs. At each point in time, at least three segments are in different
stages of contraction, a process that calls for a high degree of nervous coordination.
Although most insect legs are adapted for walking, climbing, or running, some
are modified in ways that aid other forms of locomotion. The ability to leap or
jump appears to have repeatedly and independently evolved in insects of all sizes,
particularly as an escape reaction. Hind leg modifications are the most common but
only one of a variety of specializations, most of which are based upon the sudden
release of stored tension.
3.2 Locomotion 97
In groups such as Orthoptera and fleas, jumping has become a pronounced spe-
cialization. Efficient long-distance jumping presents a special challenge—in most
cases it requires a powerful and rapidly accelerating movement of the jumping legs,
and this in turn requires morphological modifications. Even the large femur extensor
muscles of grasshoppers’ jumping legs cannot generate the quick extension needed
for an efficient jump without some mechanical modification within the leg struc-
ture for storing energy. Locusts meet this challenge with a tendon of the tibial flexor
muscle that moves over a stop; the stop allows the extensor to contract without mov-
ing the leg when the muscles are coactivated. This stores energy in the mechanical
distortion of the femur, tibia, and extensor tendon in a manner analogous to the bow
and arrow of an archer.
Among fleas, a rubbery protein called resilin in the cuticle stores and subse-
quently releases energy for the jump (see Plate 2). The material displays a 97%
recovery after stress is applied, exceeding that of elastin, the human elastic protein.
In 2005, an Australian research team produced resilin protein in purified form by
cloning a portion of the ‘resilin gene’ in Drosophila, with the hopes developing
ways to use the material for human spinal disc implants.
As far as is known, all biological pre-launch amplifiers depend on the same
mechanism—energy storage in deformed elastic materials. One of the most spec-
tacular jumping strategies does not even involve legs. Click beetles shoot upward
as much as four times their body length by rapidly accelerating the joint between
two thoracic segments; here again a mechanical stop prevents movement until large
isometric force has been achieved.
Aquatic insects have evolved two general sorts of locomotory adaptations—
those enabling them to propel themselves upon or up to the top of the water and
those by which they ‘swim’ beneath the water surface. Diving beetles in the family
Dytiscidae are Olympic-quality insect swimmers; their body shape is so similar to
a small-span wing profile that it is believed to create dynamic lift during fast swim-
ming. Many bottom-dwelling insects such as larval Odonata and Trichoptera walk
over the substrate just as terrestrial insects do (even though the larval case of some
caddisflies can be quite a hindrance to movement).
Insects that live in lakes and other slow-moving waters generally swim well. The
trunks of their bodies are streamlined and well adapted to flow. They generate thrust
by synchronous power strokes, and often have adaptations such as flattened rear
legs. Most free-swimming insects paddle with their hind legs, sometimes together
with the middle legs. Efficiency is often increased by devices such as hairs or cutic-
ular blades and/or modification in the morphology and relative size of the legs. In
contrast to the general rule in terrestrial locomotion, in swimming the two legs of a
segment sometimes work together like oars.
Surface dwellers take great advantage of the relationship between their body size
and the physical properties of water at temperatures and pressures characteristic of
their environment. Specifically, under these conditions water tends to have a rela-
tively high surface tension, so that the water-repellent surface of the insect cuticle is
sufficient to support many small surface dwellers as though upon a thin elastic mem-
brane. Many insects also secrete additional waxy material upon their tarsi, allowing
them to walk or row across the water film without breaking its surface. Some, such
as water striders in the family Gerridae, have hydrophobic tarsal hairs and special-
ized claws on their long legs that allow them to skate on the water’s surface. One
of the more spectacular surface dweller adaptations occurs in Stenus, a genus of
staphylinid beetles that live on grasses along mountain streams. If they accidentally
tumble into the water, as they often do, Stenus can walk upon the water’s surface,
but only slowly. In response to apparent danger, however, they release an anal gland
secretion that lowers the surface tension of the water behind them. Drawn forward
by the higher surface tension in front, the beetles propel themselves along at speeds
of 45–70 cm/s, moving their abdomens from side to side to direct their movements.
Insects with gills or other aquatic respiratory adaptations can live permanently
submerged. Locomotion methods among insects that live beneath the water sur-
face vary greatly. Most live predominantly at or in the surface layer of the stream
bottom, but an unusual behavior occurs among a few species in which normally
terrestrial adults dive to reach submerged areas to oviposit. Female black flies dive
through shallow moving water to reach the surface of rocks where they affix their
eggs. Similarly, some female caddisflies dive vertically and swim to oviposition sites
below inclined submerged stones. Although adult stream insects rarely swim, larvae
of many insect groups can swim by body undulations. Some mayfly larvae escape
from predators or aggressive conspecifics with such strong, active swimming that
they can travel against a current. Dragonfly larvae force jets of water rapidly out of
the rectal chamber so that the body is driven forward. Still other species use claws,
silk, suckers and other devices to help themselves maneuver.
The drift of insects downstream with the current, a behavior that typically occurs
at night, is perhaps one of the most frequently studied topics in stream ecology,
but because of the diversity of stream insects and the diversity of running water
conditions, researchers have found it difficult to identify clear patterns, much less
to develop predictive models. It is clear, however, that drift is more than a passive
activity. When a habitat patch is overcrowded and resources are low, as much as
10–30% of the insect population of a stream may drift in a single night, traveling
between 2 and 20 m during one drift movement. Black fly larvae drift by first fixing
a silk thread to a rock on the stream bottom, then prolonging the thread by spin-
ning and in essence rappelling themselves several centimeters downstream before
resettling at the bottom.
3.2.2 Aerial Locomotion

Insects alone among the invertebrates possess the ability to fly, and flight is one of
the most important reasons for their success. How and when did this remarkable
ability arise? The acquisition of wings in vertebrates is a familiar story, but few peo-
ple realize that insects were actually the first organisms to develop powered flight,
and they did so at least 90 million years (and perhaps even 170 million years) before
the earliest winged vertebrates.
3.2 Locomotion 99
Insect wings appear to have arisen upon rather large active insects sometime prior
to the Late Carboniferous Epoch. They were not modified limbs, but two or three
pairs of sideways expansions of the upper part of the thorax, and presumably, at
first these expansions only allowed an insect to glide. Flapping and steering would
come later.
On one point, scientists agree: Despite a stunning amount of structural diversity
in insect wings today, they evolved only once. Throughout the fossil story, wing
venation has remained relatively consistent, and the changes that have occurred can
be homologized across insect orders, as can many other important morphological
aspects. Beyond this, the simple question, ‘What is the origin of insect wings?’ has
been the subject of competing theories for over one hundred years. For one thing,
the question is actually two queries. One centers on homology, asking what mor-
phological elements gave rise to wings. The other concerns behavior and evolution,
asking what purposes early wing-like structures served and what conditions favored
their origin.
Currently, the best accepted theories are that insect wings may have arisen either
from pronotal lobes on the thorax or from modified gills. As a result, unlike the
wings of birds or bats, the wings of insects contain no intrinsic muscles. Instead,
they attach to the thorax by a complicated hinge structure that amplifies the tiny
strains of the flight musculature into the large sweeping motions of the wing. To
transmit force to the wings, the flight muscles are attached to the thorax by two
different systems. In one system, direct flight muscles connect directly to the wing
sclerites. In the other, the flight muscles insert within the thorax at some distance
from the wing base, and deform the overall shape of the entire thorax so that parts
push on the wing base and move it up and down (Fig. 3.3).
Odonata still possess only direct flight muscles, but most other insects possess
some combination of direct and indirect muscles so that while the muscles altering
wing inclination remain attached to the wings themselves, the muscles responsi-
ble for wing flapping are attached to the thoracic walls. Most present-day insects
also have developed a musculature that allows the wings to fold backward over
the abdomen. Thus while the wings are flapping, wing inclination is synchronously
changing, so that the overall wing flight pattern becomes much like that of a pair of
small propellers directing an air stream downward and backward. In typical forward
flight, each wing traces a pattern that resembles the numeral 8 relative to the body
at its base, and many insects can hover or loop by changing the inclinations of this
‘figure 8’ relative to their body.
By necessity, as insects became smaller their wing movement rate increased.
While a house fly may have a rate of about 200 beats per second, mosquitoes
have a rate of up to 600 beats. Tiny ceratopogonid midges have been clocked at
a wing vibration speed of over 1,000 beats per second. How can this be possible?
No known animal nerves are physically capable of transmitting stimuli fast enough
to cause contraction and relaxation at these high speeds. The elastic nature of the
insect thorax and the action of resonating flight muscles hold the key to this para-
dox. In many insects, especially certain Diptera and Coleoptera, the wings have
two stable positions—completely elevated and completely depressed. As the wings
move downward, normal thoracic elasticity resists this motion until a certain point
Fig. 3.3 Flapping their wings. Rather than flapping up and down, most insects wings move in
a back and forth motion. This is a result of the arrangement of two groups of large antagonistic
flight muscles that attach to the inner thorax walls rather than directly to the wings (thus they
are called indirect flight muscles). One group (DLM = dorsal-longitudinal muscles) is inserted
at right angles to the other group (DVM = dorsal-ventral muscles). Alternating self-sustaining
contractions of these two muscle groups changes the shape of the thorax driving the wings forward
and backward. Tiny steering muscles attached to wing base apodemes alter the wing attack angle,
with the result that with each wing stroke the wings move propeller-like around their base before
flipping over and sweeping back in the opposite direction, following a figure-8 pattern
is reached. At this ‘click point,’ three things happen simultaneously. First, the resis-
tance vanishes and the wings click into a new position below the thorax. Second,
their inclination automatically changes in readiness for the upstroke. Finally, the
muscles that have been contracting are suddenly released. As they relax, the oppos-
ing muscles are suddenly stretched, which causes them to contract instantly. In this
remarkable oscillating process, these insects have developed a system that does not
require the synchronous nervous control for every contraction that is characteristic
of dragonflies, locusts, and butterflies. Once initiated, this ‘improved model’ can be
operated at almost any speed, depending on thoracic elasticity, and can be modified
by secondary controls as circumstance dictates.
3.3 Orientation
Having briefly viewed how insects move, let us turn our attention to why and where
they move, first as individuals and then as populations. The subject of spatial adjust-
ment is a critical one touching many facets of the life of an organism. A major part
of an insect’s behavior is in fact orientation to factors such as food, mate, prey, host,
3.3 Orientation 101
etc. Thus, it is unsurprising to find that the study of orientation and navigation is a
dynamic part of modern biology, with a rapidly growing literature. We can only be
concerned here with some of its more general tenets.
Orientation is the self-controlled maintenance or change of an organism’s body
position in relation to external cues. It occurs when certain stimuli in the envi-
ronment elicit a responsive sequence of behaviors that results in a non-random
pattern of locomotion, direction of body axis, or both. The fact that orientation is
self-controlled in this way distinguishes it from passive transport. That it includes
position maintenance means that orientation also may be taken to include postural
adjustments such as response to gravity.
3.3.1 Locomotory Responses
When too far from its host plant to receive directional cues, the red cotton bug
Dysdercus congregates in humid areas; it does not directionally follow a humid-
ity gradient, but simply moves about randomly more actively when in drier areas
and more slowly in more humid areas. To escape predators, a male grayling but-
terfly will fly upward toward the sun; if blinded in one eye, he will ‘escape’ in
circles. Caterpillars move down the stems of their food plant when they are about to
pupate in the ground. Sexually mature female crickets turn to face and approach the
recorded song of a male cricket.
In the early years of the twentieth century, Jacques Loeb theorized that orienta-
tion occurred because of asymmetrical stimulation of an animal’s sensory organs.
For any bilaterally symmetrical animal, Loeb argued, a stimulus that registered
unequally on the animal’s two sides would simply cause the animal to turn until the
stimulus was equalized. Loeb called such a directed movement a tropism. However,
since then, the term tropism has come to refer primarily to movement in plants, and
the term taxis (plural, taxes) is used when referring to this behavior in animals. Some
also broaden the definition of taxis to describe any oriented heading of an animal,
whether moving or stationary.
While Loeb’s explanation was useful in stimulating research on orientation,
investigators soon found many examples like those above that did not fit Loeb’s
simple scheme. In 1940, Gottfried Fraenkel and Donald L. Gunn proposed a more
comprehensive classification of orientation behavior according to the reaction mech-
anisms involved. A response like that of the red cotton bug would be called a kinesis
(plural, kineses). This is perhaps the simplest type of locomotory response that an
animal can make to a stimulus—moving in a way that is related only to the intensity
of that stimulus while disregarding any spatial properties that the stimulus might
possess.
A response like that of the grayling butterfly continued to be called a taxis, now
defined as a directed reaction in which the organism’s long body axis is aligned with
the stimulus and movement is more or less directed toward or away from the stim-
ulus. Movements such as these would seem to be among the most straightforward
types of orientation to study because the insect’s track appears obviously related in
direction to some physical or chemical polarization of the environment. In fact, such

movements were one of the earliest types of orientation response to attract serious
biological interest, and they have continued to do so.
With Fraenkel and Gunn’s system, it became common to create terms that com-
bined the supposed environmental cue with the form of taxis or kinesis, and an
unwieldy terminology soon arose. A number of specific taxes were described, based
on whether an animal moved toward or away from certain environmental factors, the
complexity of the organism’s sensory apparatus, and the manner in which the ani-
mal moved. The classifications that resulted are still in use, but they can be slightly
overwhelming. For example, an insect that reacts to light is said to show a photo-
taxis. If, as with a maggot, it reacts by bending more vigorously on the lighted side
of its body, an action that will move the creature into darker areas, the phototactic
orientation is called klinotaxis. If instead it can be demonstrated that the organism’s
eyes mediate the movement so that removal of one eye causes it to move in circles
in the light but not in the dark, the behavior is called tropotaxis or, more correctly,
phototropotaxis (Fig. 3.4). Here, like the two reins of a horse, each receptor has
a one-way turning action; pulled equally, the horse goes straight but if just one is
pulled, the animal turns. In still other cases where orientation with form vision is
clearly demonstrable so that some degree of orientation to the light is still possi-
ble after unilateral blinding, the taxis is now termed telotaxis or, more correctly,
phototelotaxis. In addition to all this, movements toward a stimulus are generally
prefaced by the word ‘positive’ whereas those oriented away from the stimulus are
termed ‘negative’.
Fig. 3.4 Learning to adjust. When a honey bee’s left eye is blackened, initially it circles toward
the intact eye, but after 20+ repeated trials the bee learns to compensate and again moves directly
toward the light. Only a sample of trials is shown; the light was directly above the starting point in
each case
3.3 Orientation 103
At first, Loeb’s theory had such appeal that taxes were considered to be forced
movements over which the animal had little or no control. But gradually the accu-
mulating data began to give biologists a new appreciation of the complexities of
animal orientation, and Loeb’s simple theory was discredited on several counts.
For one, taxes are obviously as variable as the rest of behavior. A given taxis may
depend on environment, context, experience, and/or the organism’s internal state
(nutritional, sexual, developmental, etc.). For example, although when preparing to
pupate some caterpillars move downward, their first responses as emerging moths
may be to climb upward as high as possible. In addition, a tactic response may
change in type or sign at short notice. Thus, the blinded circling grayling butterfly
will immediately follow in a straight line should a female grayling pass by.
Tactic responses are often far less simple than they appear, and tactic interactions
are common. On the one side, taxes grade into kineses, from which they differ in
being directed responses of the insect relative to the stimulus source. In another
direction, they overlap with such longer-range phenomena as migration, discussed
later in this chapter; at times the two are difficult to distinguish. Taxes also grade into
still another quite sophisticated set of responses—the whole subject of positional
orientations in general and transverse orientations in particular—that may or may
not be called a subcategory of them.
Not surprisingly, other systems for classifying orientation behaviors have arisen.
One considers that orientation basically involves the positioning of an organism
in response to various stimulus fields such as heat, magnetism, light, gravity,
pressure, and chemicals. Thus, one may consider chemical orientation, gravity
orientation, astronomical orientation, orientation to polarized light, or any of a host
of other orientation subdivisions. (For discussion of a variety of sense organs used
in perception of chemical, visual, and mechanical stimuli, refer to Chapters 6, 7,
and 8, respectively.) Another system is based on the observation that orientation
occupies an interface between behavior and ecology. For any organism, the envi-
ronment contains both positive and negative factors—not only resources needed
for sustaining life or their absence but stress sources such as intense sunlight,
which can be rapidly debilitating in the absence of compensating behaviors. A
maximally fit organism is expected to behave in a manner that consistently works
to minimize its body distance from resources (food, shelter, etc.) and maximize
its distance from sources of stress. Viewing orientation from the perspective of its
adaptive significance has led to a more complete classification of orientation than
past attempts. However, for most scientists, taxes and kineses remain a major con-
ceptual scheme for understanding and investigating the actions of insects and other
organisms.
3.3.2 Posture and Position
Locusts in flight maintain an even body keel partly through visual reactions to incli-
nation of the horizon. Flies mechanically sense angular acceleration and angular
motion by rapid oscillations of their gyroscope-like halteres (modified knoblike
vestiges of their hind wings). Through continual compensatory reactions to a

directed light source, a dragonfly banking its wings while flying a curve still keeps
its head in a perfectly upright position.
All such examples involve positional orientation—compensatory maneuvers for
body stabilization against displacement by wind, water, etc. The most widespread
forms of positional orientation in insects are the transverse orientations, those in
which the body is positioned at a fixed angle relative to the stimulus.
Among the best known transverse orientations are the dorsal and ventral light
reactions. In this type of light orientation, well illustrated by dragonflies, moths,
and butterflies, both the long and transverse axes of the body are kept perpendicular
to a directed source of light at all times. (Thus, these light reactions contrast with
phototaxis, where orientation is parallel to the light rays.) For example, the dragonfly
shows a dorsal light response, mediated by the upper ommatidia of its eyes, that
assures that the upper part of its head remains turned toward the light. If blinded
in one eye and illuminated equally from all sides, the insect will roll continuously
toward its seeing side. These reactions are particularly common among both flying
and swimming insects (Fig. 3.5).
Light reactions are particularly important in relation to gravity perception, and
can involve polarized light (see Chapter 7). Insects lack specialized gravity recep-
tors that function like the inner ear in vertebrates or the statocysts found in various
crustaceans. Instead, insects have relatively unspecialized sensory hairs (proprio-
ceptors) that are usually clumped into plates in positions where they can measure
the relationship between body parts differentially affected by gravity. For example,
Fig. 3.5 Heeding unnatural cues. The larva of the diving beetle Acilius normally swims to the
surface for air, cued by its dorsal light reaction; placed in an aquarium lit from below, it will
descend with its back down, attempting to get air at the bottom as though it were the surface.
Arrows indicate direction of swimming movement
3.3 Orientation 105
in ants, gravity receptors are located at points of body articulation (neck, antennal
joints, thorax and petiole, petiole and gaster, and joints between thorax and coxae);
stimulation of any one point alone is enough for gravity orientation, although the
different joints are not equally reliable.
Visual input sometimes serves as an effective substitute for gravitational forces.
In fact, phototaxis and geotaxis, two of the best-studied taxes, have several aspects
in common. For most insects the sun is upward, and positive phototaxes and neg-
ative geotaxes are the norm. One of the most remarkable features of honey bee
communication is the two-way transfer from the angle between sun and food source
to the angle between vertical and direction of the straight part of the waggle dance,
then back again (see Chapter 8). For the dancing honey bee, positive phototaxis is
coupled with negative geotaxis; if the food is in the same direction as the sun, the
straight run of the waggle dance is directed upward, opposite the direction of grav-
ity. When information about gravity is experimentally altered, the dance performed
by the bee reflects the changed input (Fig. 3.6).
Probably the single most striking aspect of postural control among insects is
the manner in which it depends upon input from a great number of sources acting
in concert to the point of redundancy. For example, an ant can correctly orient to
gravity using any one of its five proprioceptive joint systems alone if the others are
fixed in position with wax.
3.3.3 Orientation to Radiant Energy
When an ant on its way back to the nest is placed in a dark box for a period of
time, how will it orient when released? Lasius niger proceeds in the same course
as before, relative to the sun, but because the sun has shifted position in the sky
during the ant’s incarceration, the released ant’s orientation is incorrect in terms
of its nest.
An orientation like this one—locomotion at a fixed angle relative to light rays—
is termed a light-compass reaction. It has been demonstrated in a wide variety
of insects, including caterpillars, bees, and certain beetles and bugs. Light reac-
tions also have been well studied in the context of navigation during migration
(see Section 3.5.3). The polarization of light rays often serves as an orienting
cue. In one of the more unusual examples, certain African dung beetles apparently
use polarized moonlight to whisk balls of dung in a straight line away from the
dung pile.
Have you ever noticed that various crawling and flying insects nearly always
travel in a straight line across roads at right angles to their direction? This appears to
occur irrespective of compass directions or other external stimuli. One hypothesized
explanation is that the insects orient by balancing their reception of a symmetrical
source of shortwave (infrared) radiant energy. Although roads are relative newcom-
ers to the environment of insects, the adaptive significance of such behavior may
be a survival advantage conferred in crossing large bare spots of earth or bodies of
water in the shortest possible time with minimal energy expenditure and exposure
to the elements or predators.
Fig. 3.6 Turning gravity on its head. Proper orientation during the honey bee waggle dance
depends on gravity perception. Because of the way it is connected at the neck, the lower half of
a honey bee’s head weighs almost twice as much as the upper half. (A) When an unaltered bee is
dancing on the vertical surface of a comb inside the hive, gravity causes this lower portion to swing
downward, tilting against the sensitive proprioceptive hairs on the neck, and stimulating nerves at
their base. Severing these nerves causes total disorientation. (B) A tiny weight glued to the top of
the bee’s head alters its balance. (C) The proprioceptive information received by the bee’s brain is
reversed, resulting in a dance that is the opposite of normal
3.3 Orientation 107
Insect reactions to infrared radiation are rarely reported, but they have been
demonstrated across diverse taxonomic groups, and with further study they may
prove to be more common than expected. Certain wasps that parasitize beetle larvae
found in dead timber can detect their hosts by means of infrared receptors on their
antennae. The beetle Melanophila acuminata, which lays its eggs in freshly killed
conifer trees, uses a pair of specialized infrared sensory pits to sense forest fires,
even over long distances (Fig. 3.7). Thermoreceptors located in the wings and anten-
nae of species of darkly pigmented butterflies have been shown protect them from
heat damage while sun basking. Research has shown that some blood-sucking bugs
are able to perceive the radiant heat emitted by their warm-blooded prey, even at a
distance, and others have the ability to locate blood vessels under their host’s skin by
sensing temperature gradients. Likewise, Atta leaf-cutter ants can learn the location
Fig. 3.7 Finding fires. The jewel beetle, Melanophila acuminata, possesses a pair of infrared
detector organs in pits next to the junction of its middle legs and body. Each organ contains a cluster
of about 70 individual dome-shaped sensilla (below, highly magnified), which are exposed during
flight. Electrophysiological recordings reveal that these receptors respond to infrared radiation,
characteristic of a forest fire, enabling the beetle to find smoldering wood in which to lay its eggs.
Biomimetic engineers have developed an infrared sensor based on this organ. The beetle is about
10 mm long
of a food reward by using thermal information as an orientation cue; shielding their

heads and thoraxes did not weaken this ability, suggesting that only the head and
antennae are involved in assessing temperature.
3.3.4 Magnetic Field Orientation

Some Australian termite species build large ‘magnetic’ mounds oriented perfectly
north–south (Fig. 3.8). Resting termites adopt positions aligned with the cardinal
axes of a magnetic field. Weaver ants following a foraging trail in dim, diffuse light
will reverse their heading when exposed to an artificially induced magnetic field
with polarity opposite that of the geomagnetic field.
The ability to sense the earth’s magnetic field has long been suspected for some
insects, but until fairly recently, experimental evidence has been tenuous, indirect,
and often confounded by other factors, particularly the skill with which insects use
visual cues such as the sun, polarized light, the moon and landmarks for spatial ori-
entation. In fact, when salient cues such as light are present, an insect’s competence
in magnetic field orientation may be entirely hidden.
With wax secreted from specialized abdominal glands, the honey bee builds par-
allel vertical combs that have hexagonal cells on both sides separated by a thin
middle wall. The cells are tilted slightly upward at about 13◦ to the horizontal, an
adaptation which serves to keep honey in. When a swarm of bees takes possession
of a new home such as a hollow tree, they will reconstruct, literally overnight, a
complete series of combs having the regular parallel construction. How do the bees
‘decide’ how their new combs will be oriented?
About fifty years ago, Martin Lindauer and colleagues decided to attack this
question experimentally. They transferred bees from a conventional hive into a plain
empty cardboard cylinder with the entrance hole centered in the bottom, and found
that the bees produced new combs with an orientation that exactly corresponded
Fig. 3.8 Orienting magnetically. Mounds of the magnetic termite, Amitermes meridionalis, in a
seasonally flooded field in northern Australia. The long axes of all these nests run north–south
3.3 Orientation 109
to that of the combs in the original parent colony. Since all directional landmarks
had been removed in the round cylinders, the researchers hypothesized that the bees
were somehow able to orient to the earth’s magnetic field. To test this hypothesis,
they placed a powerful magnet outside the experimental nest cylinder so that the
natural magnetic field was deflected. Invariably, the same bees that had previously
reconstructed faithful new combs in experimental cylinders now built combs that
differed from those in the previous nest by 40◦ —the exact angle of the artificial
magnetic deflection.
The magnetic sense of insects and its adaptive importance have been most thor-
oughly investigated in social insects, especially ants, bees, and termites. There is
now growing experimental evidence for magnetic field sensitivity in some Diptera,
Coleoptera, and Lepidoptera as well. Contexts for its use have included home
range orientation, homing, long-distance migration, and nest and body alignment.
In experiments with artificially induced magnetic fields, the insects have always
responded to changes in the magnetic field’s declination, which implies sensing of
magnetic polarity.
Although the sensory system that mediates magnetic signals has not yet been
definitively identified, evidence favors the use of intracellular, submicroscopic mag-
netite particles that have been found in honey bees, monarch butterflies, and some
ants. An alternative hypothesis draws on a system that has been demonstrated in
some birds, in which certain photochemical reactions are magnetically sensitive; if
such reactions are linked to light reception in the eye, then changing the wavelength
of ambient visible light could alter the directional orientation to the geomagnetic
field. Such effects have been obtained in male Drosophila melanogaster.
3.3.5 Orientation to the Evidence of Others’ Presence
Both orientation and navigation depend upon perception of environmental features

that exist quite independent of the insect itself. One another’s actions and adjust
their own movements accordingly. Gallery orientation by the larvae of long-horned
wood-boring beetles (Cerambycidae) is an example. Larvae are responsive to
sounds emanating from other larvae. Experimentally, they will change direction
when exposed to simulated gnawing sounds originating ahead of themselves or
slightly to one side. As a practical result, examination of gallery complexes reveals
that the amount of turning and irregularity of burrows is directly proportional to
the degree of infestation, and highly contorted burrows are a characteristic of heavy
infestations.
In another example, groups of whirligig beetles (Gyrinus) spin crazily about over
the surface of most freshwater ponds. As long as they all keep moving, they seem to
avoid contact; but if one stops, another beetle may bump into it. Through impulses
registered by receptors on a modified second antennal segment, each beetle is aware
of the location and direction of the others by the vibrating waves they set up as they
ripple over the surface.
3.4 Thermoregulation
The sphinx moth Celerio lineata stabilizes its thoracic temperature during flight
over a range of ambient temperatures. A wide variety of insects, from butterflies and
beetles to cicadas and arctic flies, bask in sunny spots. Male tettigoniid grasshoppers
elevate their thoracic temperature prior to singing.
Most people think of insects as being purely ‘cold-blooded’ (ectothermic) crea-
tures that passively reflect the temperatures that surround them. However, this view
is overly simplistic. In most habitats, seasonal and diurnal temperature oscillations
are considerable.
Heterothermic is probably a better term to describe the insect’s life with fluctu-
ating body temperatures. Because of their very high ratio of surface area to mass,
small organisms readily lose and gain heat; insect body temperatures can change
abruptly by 10◦ C (50◦ F) or more when direct sunlight is obscured by the shade of a
passing cloud. This can be a serious matter—for insects, as for other animals, body
temperature is closely attuned to activity and energy supplies, and thus affects all
aspects of life from the rate at which food can be located and harvested to the ease
with which predators can be avoided. Thus, one should perhaps not be surprised to
find that insects have ways to control their body temperatures within a far narrower
range than that of their surroundings.
Adaptations to temperature fluctuations can (and do) take many forms. Some
adaptations are primarily behavioral, whereas others rely more heavily on physi-
ology, but any attempt to strictly divide them soon shows the two systems to be
intertwined.
3.4.1 Dormancy and Thermotolerance
Most insects face periods of adverse conditions, in which temperatures are outside
their physiological limits and/or food is wanting. However, being heterothermic,
they have an advantage that is denied to their homeothermic brethren. As long
as the temperature does not reach lethal extremes, insects are capable of ceasing
development and reproduction temporarily, then resuming these activities when con-
ditions again become favorable. This ability is undoubtedly one more reason behind
the enormous success of Class Insecta, because it allows them to exploit seasonal
resources while successfully bridging the harsh periods between.
The external conditions that temperate-zone insects must tolerate differ markedly
in summer and winter, so despite some similarities the insects’ responses have come
to be known under different names. We speak of winter hibernation (although it
is not directly comparable to hibernation in mammals) and summer aestivation.
Both actually jointly serve two different purposes. One is to promote survival
by depressing metabolism and energy utilization during adverse times. The other
is to synchronize periods of feeding, growth and reproductive activities with
those times of the year when food is available and environmental conditions are
suitable.
3.4 Thermoregulation 111
Short- and long-term shut-downs follow different strategies. To cope with rela-
tively short periods of non-lethal but unfavorable circumstances, insects commonly
go into a state of easily reversible, directly temperature-dependent developmen-
tal arrest called quiescence. Depending upon the temperature, the insect’s growth
slows, gonads mature more slowly, and feeding and other behaviors lose speed.
Quiescence is a rapid response that can occur at any life history stage.
Severely hostile conditions that last several months, such as a temperate-zone
winter or a hot, dry desert summer, require a different approach because an insect in
the active state simply could not survive. A programmed developmental arrest called
diapause occurs. Unlike quiescence, diapause happens only during a specific devel-
opmental stage, usually (but not always) before adulthood, and its timing depends
on the insect species. For example, the European corn borer diapauses only as a fifth
instar larva; the cecropia moth, only as a pupa; the Colorado potato beetle, only as
an adult.
Diapause can be facultative, occurring in response to environmental cues, or it
may be obligatory, occurring during each generation. A number of economically
important species (such as the gypsy moth) have an obligatory diapause, but fac-
ultative diapause is far more common. For an insect with facultative diapause, a
relevant environmental cue must be received during a sensitive period in develop-
ment for diapause to occur; otherwise, development simply continues to proceed.
This system can allow exquisite fine-tuning in the insect-plant arms race, as many
home gardeners can attest. Throughout the warm months, these insects can pro-
duce multiple non-diapausing generations, building up high populations. Then, in
late summer or autumn they produce a generation that goes into diapause and thus
escapes the winter months when their plant hosts are unavailable.
While the diapause itself is a physiological event, behavior is intimately and
intricately tied to it. Usually, the insect moves to a site (or in the case of
cocoons, constructs one!) that offers additional protection from the full onslaught
of inclement conditions. Specific behavioral changes are often associated with
movement to overwintering sites called hibernacula. This may involve short- or
long-range migration, a subject covered later in this chapter. It also includes local
movements such as are displayed by certain caterpillars that have been feeding up
in trees but now climb down and begin burrowing into the soil.
Feeding usually stops during diapause, so it no surprise to find that many species
‘fatten up’ in preparation. Of course, their exoskeleton does not allow this in a lit-
eral sense, but insects that are about to enter diapause often sequester twice as much
lipid reserves as their counterparts that are not so programmed. Because metabolism
slows with dropping temperatures, energy reserves may be less a problem, however,
than water loss. With no access to free water and possession of a large surface-to-
volume ratio, diapausing insects are particularly vulnerable to loss of water across
their cuticle. Insects that diapause as adults (such as many beetles, bugs, and butter-
flies) often gather into large aggregations (see Plate 5). For those that are distasteful,
this may increase their protection during this inactive life phase. Equally impor-
tant but perhaps less immediately obvious, an aggregation provides a more stable
microenvironment for its members and increases local humidity.
Whereas quiescence occurs in direct response to adversity, diapause begins

before adverse conditions begin. How can this be? Most commonly, the answer
seems to involve changes in two important aspects of the environment—photoperiod
and temperature. Other seasonal changes may vary with latitude, altitude, and
habitat, but changes in the length of day versus the length of night are a reli-
able signal that unsuitable conditions are coming. Photoperiod changes are so
predictable that in most cases, the developmental period that is sensitive to photope-
riod actually occurs far in advance of the actual diapause stage. This is critically
important, because such early programming offers the insect the time it needs
to prepare for diapause, both physiologically and behaviorally. It is important to
remember, however, that photoperiod is not the only trigger for diapause. Daily
fluctuations make temperature less reliable as a seasonal cue, but they do act
on this basic signal, and can shift the photoperiod-based diapause response over
a broad range or in extreme cases even eliminate it. Other environmental cues
that have been identified include moisture conditions and changes in host plant
quality.
The assumption has been that the programming of diapause has the same genetic
basis as other biological rhythms (see Chapter 2). Thus, the basic requirements for
a photosensitive diapause system would be a mechanism to distinguish long days
from short days and a way to keep track of them—in other words, the insect must
have a clock and a counter. A great deal of genetic research in the past few years
has centered upon finding, identifying, and characterizing critical clock genes. Only
the best known of these—the period (per) gene—has been studied in connection
with diapause. For the drosophilid fruit flies that have been studied, the somewhat
surprising findings indicated that null mutants that fail to express per enter diapause
just as readily as the wild type.
Diapause does not end immediately after the diapause-inducing conditions disap-
pear. Some time is required before neurohormonal systems return to configurations
that support development and reproduction. Interestingly, overwintering insects
appear to terminate diapause in early winter and after this point, their continued
development is halted only by a direct effect of low temperature. In essence, their
diapause turns into quiescence.
It is clear that like all organisms, insects of a given species are adapted to a set
temperature range within which they can effectively reproduce and develop, and out-
side this range, they clearly do so with varying levels of success. What about those
insects that do not diapause? Through what behavioral and physiological responses
do they counter the effects of high- and low-temperature stress?
Survival at extreme temperatures—called thermotolerance—can be increased
several ways. One way is through genetic adaptation; differences in thermotoler-
ance can be detected both in diverse geographic populations and in laboratory lines
selected for this trait. Another way is through long-term acclimation; rearing indi-
viduals for long durations at high temperatures can result in striking increases in
thermotolerance. Acclimation is a complex adjustment that involves major body
changes; in some insects it actually includes the synthesis of antifreeze proteins. A
great many behavioral changes support these physiological alterations. These can
range from forming thicker cocoons to voiding the gut to remove food particles that
would be powerful ice nucleators.
A third way to increase thermotolerance is through a response called rapid heat
hardening, in which brief exposure to an intermediately high temperature provides
protection from injury at a more severe temperature. For example, Sarcophaga flesh
flies live only a brief time at 45◦ C (113◦ F) when moved there directly from a 25◦ C
(77◦ F) chamber. However, if they are first exposed to 40◦ C (104◦ F) for two hours,
they survive much longer at 45◦ C (113◦ F). This protection develops within minutes,
reaches a maximum within a few hours, and then decays rather slowly over several
days. Does diapause represent a shutdown in gene expression or does it represent
the expression of a unique set of genes? Perhaps both. Studying Sarcophaga flesh
flies, David Denlinger has shown that far fewer proteins are synthesized in the flies’
brains during diapause. However, in addition, their brains synthesize a set of pro-
teins that are not observed in brains of non-diapausing flies, and various classes of
diapauseupregulated genes can be distinguished based on their expression patterns.
Other research suggests that insects respond to heat stress by suppressing the nor-
mal pattern of protein synthesis and concurrently synthesizing several new proteins
called heat-shock proteins. Several types of heat-shock proteins have been found
in organisms ranging from bacteria to plants to insects, and behavioral geneticists
suspect their upregulation may be common to many different types of dormancies.
The most highly expressed heat-shock proteins are highly conserved; the gene that
encodes one group called Hsp70 is over 50% identical in bacteria and Drosophila
melanogaster.
3.4.2 Regulation of Heat Gain
Certain Pheidole ants forage above ground all day during cool, cloudy periods but
when the weather is hot and sunny they restrict their foraging to the period from
late evening to early morning. The adult form of a flightless midge called Diamesa
walks on glacier ice even when its body temperature is chilled to –16◦ C (3.2◦ F). It
is so sensitive to heat that if you were to pick one up in your hand, the warmth of
your skin would kill it.
Cryptic insects that rest by day on a matching background, predators that use
a lie-in-wait strategy, and many other arthropods that are exposed to sunshine all
have at least a potential problem with heat gain. Not surprisingly, various behavioral
adaptations have evolved that involve postural adjustments to minimize the body
surface area exposed to the heat source or that lead to avoidance of solar radiation
totally during certain periods. Structural features also help. For example, the long
legs of many ants and beetles living on sand in direct sunlight lift their bodies above
the substrate, while light body pigmentation reduces heat input from above.
Especially when the weather is warm, flying insects face an additional problem—
dissipating their own body heat. Small insects have much lower body temperatures
in flight than large insects do. However, this is not because they produce less heat—
instead, they actually produce more. Rather, it is because smaller insects have
more surface area relative to their mass, so that internally generated heat is lost
more rapidly by convection. Insect flight is actually one of the most energetically
demanding activities known. Most insects produce more heat per unit muscle mass
when they fly than almost any organism on earth. Almost 94% of the energy
expended by their contracting flight muscles is degraded to heat, while only about
6% appears as mechanical force on the wings.
Most insects apparently do not cool themselves as we do with active evapora-
tive mechanisms, but some of the exceptions are striking. The tsetse fly, Glossina
morsitans, is able to feed while standing on the hot hide of a mammal’s back under
the equatorial sun; to lower its body temperature, it opens its spiracles, allowing the
water drawn from its blood meal to evaporate. Diceroprocta cicadas of the Sonoran
desert overheat when they sing in hot weather; to cool down, they actually sweat
through large pores on their back, using liquid they suck from deep-rooted shrubs.
Social insects exhibit a variety of architectural and behavioral devices that main-
tain nest temperature and humidity within carefully controlled tolerance ranges, irre-
spective of season or outside temperature (see Chapter 10). While such insects are
individually heterothermic, socially they are nearly as homeothermic as birds and
mammals. Honey bee workers maintain the hive interior at temperatures between
34.5 and 35.5◦ C (94–96◦ F) by fanning with their wings to promote air circulation
and cooling by water evaporation. (In winter, the bees cluster loosely at warmer
temperatures and very tightly during extreme cold, keeping that same range.)
3.4.3 Heat Production
Male Neoconocephalus katydids raise their flight-muscle temperatures above 30◦ C

(86◦ F) before starting their ear-shattering mating concerts. Founding queens of
Vespula wasp colonies incubate their brood clump by perching upon it and pressing
their abdomen over it much as a hen incubates her eggs. Prior to flight, nearly all
but the smallest temperature zone butterflies bask in various distinct postures that
simultaneously maximize solar input and minimize convective heat loss.
If avoiding too much heat is one important biological task, its flip side is staying
warm enough to function. Endothermy—the ability to increase body temperature
beyond that of the environment—has been long known in insects, but scientists
once thought such regulation was accomplished solely by behavioral means such
as the basking frequently observed in butterflies. An increasing number of stud-
ies now show that some insects use physiological means as well. On the basis of
body weight, most flying insects have higher rates of metabolism and hence of heat
production than other animals.
In insects, essentially all endothermic increases of body temperature have the
same causal mechanism—heat produced by the active flight muscles. Because these
are the most metabolically active tissues known, endothermy in flight is largely
an obligatory phenomenon. In addition, many species use high-energy fuels and
intense metabolic rates to produce body heat at rates sufficient to increase their
body temperature beyond that of their environment.
High muscle temperature is, however, not just a consequence of muscle activity.
In many situations, especially flight, it is also a prerequisite. Until the temperature
of the muscles is sufficiently high, there is little overlap in the contractions of the
antagonistic muscles, the wing-beat frequency is very low, and the insect remains
grounded. Different relative wing sizes and power requirements determine the mus-
cle temperature and wing-beat frequency necessary for a given insect to become
airborne; this varies with body size, between different muscles used for different
activities, and between different species.
Large wings are one way around the problem—they allow an insect to fly with
a low wing-beat frequency. As a result, some butterflies are able to initiate flight
without prior endothermic warm-up and to continue flight by gliding; this also
reduces the energy expenditure of locomotion. Another response is to evolve a dense
pile coat; this can cut the rate of convective heat loss from the insect’s body by
half. Bumblebees (Bombus) provide an elegant example; their combination of high
metabolic rate, relatively large body size, and good insulation helps them not only to
elevate thoracic temperatures passively during free flight but also to maintain a suf-
ficiently high thoracic temperature to fly at very low ambient temperatures. Due to
their rapid metabolism, their body temperature while flying usually exceeds ambient
temperatures by 5–10◦ C (9–18◦ F) and sometimes b as much as 20–30◦ C (36–54◦ F).
However, thermoregulation is still a costly behavior, especially when an insect is not
physically active. When temperatures are very low, it takes nearly as much energy
for a stationary bumblebee to maintain its body heat as it does to fly. Thus, it is all
the more impressive that queen bumblebees use thermoregulation behaviors to raise
the temperature of their brood.
The requirement for a high thoracic temperature to start flight poses a real behav-
ioral problem. When an insect comes to rest in the shade, its body temperature
rapidly becomes practically the same as the ambient temperature. Small flies such
as midges and fruit flies have rapid heat loss and little buildup of body heat during
flight; their wing-beat frequency and flight speed varies nearly directly with ambi-
ent thermal conditions. Some of the larger insects, such as bumblebees and some
moths, however, must warm their flight muscles to about 40◦ C (104◦ F) before they
can attain sufficient wing-beat frequency and lift to support themselves in free flight.
Without some means of increasing muscle temperature, the insect could remain
permanently grounded.
Shivering is a widespread mechanism for increasing thoracic temperatures before
flight. It involves many patterns of flight muscle activation and can work flight mus-
cles harder than flight itself does, but it is well suited for variable rates of heat
production because it can occur at a wide range of activation frequencies. Among
many Lepidoptera, the rates at which wings vibrate during such shivering have been
shown to be directly correlated with muscle temperature. Interestingly, it took scien-
tists a long time to recognize that insects shiver, because it is almost never externally
visible even if one looks very closely.
The evolution of shivering is clearly related to the evolution of flight, but it is
unrelated to an insect’s place on the phylogenetic tree. A physiological warm-up
is found in all large, active flyers among the dragonflies, moths and butterflies,
katydids, cicadas, flies, beetles, and wasps and bees. It is missing in small (and
therefore non-endothermic) members of the same groups.
In addition, an insect may regulate its body temperature through discontinuous
activities such as intermittent flight, intermittent shivering, or some combination of
the two. However, some insects are behaviorally better suited to make use of this
option than others. For example, a hovering sphinx moth or dragonfly in continuous
flight is in a less advantageous position in this regard than is a bee that lands on
flowers at frequent intervals while foraging.
3.5 Migration
Foraging army ants, Eciton hamatum (see Chapter 10) commonly move out of the
nest in columns along branching trails to seize and carry back to the nest all small
prey in their way. However, as new brood matures within the colony, instead of
simply returning to the nest, workers reverse to lead a mass exodus that carries
the whole colony away along one of the day’s trails. During their march the ants
neither react to prey nor branch off. Meanwhile, in California large numbers of
convergent lady beetles, Hippodamia convergens, spend the winter at high altitudes
in the Sierra Nevada mountains, then move down to agricultural areas in the Central
Valley in March to lay their eggs; beginning in June, their offspring fly back to
intermediate altitudes, gather into aggregations, then move back to higher altitudes
to overwinter. On a much smaller scale, two species of pierid butterflies regularly
maintain a directional flight across Gatun Lake in the Panama Canal.
What triggers the mass exodus of army ants? How are the movements of the
lady bugs coordinated and maintained? Why do the butterflies go back and forth?
Up to this point, we have been viewing insect spatial adjustment primarily as a
phenomenon involving a single individual within its immediate environment. Now
it is time to step back and view insect movement on a larger scale, considering these
behaviors in terms of the population or species.
A fundamental concern for most insects is finding the optimal habitat in which to
live and reproduce. For dispersing individuals, the problems are particularly com-
plex and acute. At what point should an individual stop expending energy on the
search and settle for whatever situation is available? When should selection favor the
evolution of ways for individuals to return to specific sites after displacement from
them? How should areas of fluctuating or disparate resources be best exploited?
When times are hard, does it make more sense to move, or to just hunker down and
ride it out? Such theoretical questions have been receiving much attention in recent
years, and some interesting answers are beginning to appear.
In general, insect movements are of two main kinds. The first kind, concerned
with resources and home ranges, has the ecological outcome of keeping an insect
in the habitat where it can obtain the resources it needs for given stages of its life
cycle. The second kind is quite different. Migration, one of the most important forms
of insect dispersal, has been succinctly defined as ‘adaptive traveling.’ It involves
3.5 Migration 117
leaving the current home range and habitat. (As a fine point of terminology, although
migration acts as an active mass movement functioning to displace populations,
those who study it usually distinguish between migration as an individual behavior
and dispersal and aggregation as population processes; see Section 3.5.3. However,
in popular usage, the distinctions are drawn less finely.)
Behaviorally, migration is characterized by persistent, enhanced locomotion in
a straightened-out manner, and migrating individuals do not typically respond to
stimuli for ‘vegetative functions’ such as feeding, reproduction, etc. For example,
the flight of certain scolytid bark beetles cannot be arrested by their host plant’s odor
until after they have been flying for many minutes. After long-distance flight, the
thresholds for vegetative activities are lowered, and further migration is inhibited.
Insects move around quite a bit, so it is important to distinguish migration from
those local movements of varying length and orientation that are concerned with
food and mate finding, escape from potential enemies, location of suitable oviposi-
tion sites, territorial defense, and other such activities. Such ‘appetitive’ behaviors
may lead to some dispersal, but often no effective displacement occurs at all, despite
a good deal of activity. Migration, in contrast, involves a phase in adult life during
which directed movement (usually flight activity) dominates over all other forms of
behavior. In many insects, such activity is restricted to a short period, after which
only appetitive movements occur; in some species, the flight muscles may break
down after migration so that no further flight is even possible.
Why do insects migrate? Such a question is more properly two questions. First,
the proximate one: what triggers the migratory ‘urge’ and how is it maintained?
Second, the ultimate one: what ecological and/or evolutionary conditions might
favor development of a migratory mode of life?
Migration clearly involves more than simply responding to the onset of adverse
conditions by escaping to a new habitat. Endocrine changes occur in correlation
with particular environmental effects (crowding, food deficiency, short days, etc.),
and in turn cause physiological and developmental changes. The insect’s behavior,
ecology, and genetics all come into play.
3.5.1 Seasonal Migration
Large concentrations of bean aphids appear in central France in early to midsum-

mer every year; over succeeding generations they spread westward and northward,
so that by late summer bean aphids have reached high densities as far north and west
as Scotland. Bugs in the genus Dysdercus are seed predators on many subtropical
and tropical plants, including commercial cotton. At the beginning of the tropical
dry season, as fruiting begins, the bugs migrate into the plants. As the females begin
to feed, they histolyze their wing muscles, develop large numbers of eggs, and repro-
duce explosively. Then as the dry season progresses and fruiting ends, the absence
of moisture and food causes the new generation of bugs to retain their wing muscles,
depart and go into diapause until their next dry-season migration begins.
Many examples of insect migration have been described around the world. One
set involves the many insects that leave the breeding site, oviposit elsewhere, and
die, all in a single season. A second group includes short-lived adults that emigrate
and return. For example, many dragonflies depart from ponds to terrestrial feeding
sites some distance from water; after the eggs mature, females return to the vicinity
of their original breeding site and oviposit. Another category includes longer-lived
adults that hibernate or aestivate away from the original breeding site, then return to
it the following season.
How can one tell if a flight is truly migratory? The definition of migration should
offer an objective and experimental test: during the flight, the insect should not
be responsive to stimuli triggering vegetative behavior. The test has seldom been
applied, but where it has, there does appear to be a reciprocal interaction between
migratory flight and vegetative activities. Stimuli that evoke flight inhibit settling,
and stimuli that evoke settling inhibit flight. One example where such migratory–
vegetative interaction has been approached experimentally is that of the milkweed
bug, Oncopeltus (Case Study 3.1).
Case Study 3.1: Migration in the Milkweed Bug, Oncopeltus

How can you tell if a flight is truly migratory? The definition of migration
should offer an objective and experimental test: during the flight the insect
should not be responsive to stimuli triggering vegetative behavior.
The milkweed bug (see Plate 11) is a wide-ranging species occurring from
Canada to Central America. It arrives in the northern reaches of its range
between spring and early summer. Here, females settle on patches of milk-
weed, mate, and lays their eggs close to the developing seedpods. The young
bugs grow quickly and after a few weeks, they become breeding adults.
Throughout the summer the population increases rapidly, but with the shorten-
ing days of early autumn, numbers decline as maturing adults begin to leave.
Flying south on the prevailing winds, these adults are able to avoid the oncom-
ing winter, while those unable to complete an adult molt before the first severe
frost are killed.
Are these bugs truly migrants? In the early 1970s, Hugh Dingle and his col-
leagues sought to answer this question through detailed field and laboratory
investigations. Rearing studies soon confirmed the first criterion—a separa-
tion of flight and vegetative activities. Flight activity peaked eight to ten days
after the adult molt, whereas oviposition began only after 13–15 days. Flights
were also well separated from feeding. A few hours after the adult molts,
females entered into a high rate of feeding that, until day seven, lacked period-
icity. By day eight, however, a fully developed circadian rhythm was evident.
Thereafter, the peak feeding activity for females occurred at the end of the
day and proceeded simultaneously with mating. Interestingly, for males the
3.5 Migration 119
feeding rate fell markedly after day six to persist at a very low rate for the
remainder of life.
Temperature appeared to have a very direct effect upon migration. Raising
the temperature from 23◦ to 27◦ C, a warmth about optimal for population
growth, Dingle found that a lower proportion of the population now exhib-
ited tethered flights of 30 minutes or longer (his operational criterion for
migration). This suggested that once the bugs reached a thermally favorable
environment, they would tend to settle there.
Laboratory studies clearly indicated that peak flight preceded reproduc-
tion. Would prolonging the pre-reproductive period also prolong migration?
By altering day length in the laboratory, Dingle knew that he could delay
oviposition from 15 to 45 days after the adult molt. Raising bugs under both
regimes, he tested comparable groups of females for duration of tethered
flight 25 days after adult molt. The early-ovipositing females generally flew
for only a few minutes or less. The delayed females, which showed no signs
yet of reproductive development, performed like typical migrants. Under
field conditions, Dingle reasoned, this phenomenon would have important
consequences. One would be that in the autumn females would be capable of
migrating for much longer periods, thus improving their chances for escaping
the oncoming winter.
In what other ways might a short photoperiod affect migratory behavior?
In tethered flights, Dingle compared sets of short-day and long-day bugs of
both sexes repeatedly between 8 and 30 days after adult molt. Invariably,
a greater proportion of the short-day bugs flew for long periods (at least
30 minutes, usually 2–3 hours). Significantly, the results held true for both
sexes. Evidently, in addition to its indirect effect via ovarian development,
photoperiod was having a direct effect upon migration as well.
Environmental factors obviously had great importance as determinants of
migratory behavior. What about hereditary influence? Under strong selection,
Dingle found that he could increase the proportion of migrants of a population
of milkweed bugs from 25% to over 60% in one generation. Clearly, migratory
capability in Oncopeltus could be altered rapidly.
Not just in milkweed bugs, but in a great many insects, migration is associated
with an additional photoperiodic response—diapause, discussed earlier in this chap-
ter. Short days in autumn cue an inactivation of the reproductive system and thus
permit long flight using energy that otherwise would be channeled to reproduction.
Perhaps the best known of these insects is another milkweed-feeder, the monarch
butterfly, Danaus plexippus (see Plate 3); these long-distance fliers seek out and con-
gregate in sheltered cool locations. Through this behavioral thermoregulation, they
conserve the energy reserves they need to tide themselves over the winter diapause
portion of their migratory cycles.
Earlier biologists thought migration was probably just a way to escape from one
habitat to another more suitable one. On first examination such a simple answer
seemed sufficient. Faced with a temporarily unfavorable period or untenable habi-
tat, an insect species could adopt one of two evolutionary strategies. It could migrate
first or it could go into immediate diapause where it was already living. If the
change in habitat were reversible, as with seasonal changes, diapause in place
would be favored. When habitat changes were irregular, migration would have a
clear advantage over diapause. A recrystallization of ideas drawing from the fields
of physiology, behavior, and population ecology now suggests that migration is
not only a means of escape from unfavorable environments, but a positive act of
dispersal over all available habitats. Under such a view, insect migrants are more
accurately to be viewed as colonizers than as refugees, and migration as an evolved
adaptation, not a reaction to current adversity.
Entomologists and government agencies have paid particular attention to some
major agricultural pests—particularly several noctuid moth species and various
leafhoppers and plant-hoppers—that make round-trip seasonal migrations spanning
several generations. Many of these seasonal migrants may have originally found
migration adaptive for locating ephemeral habitats induced by rainfall in more
tropical parts of their range, but with extensive acreage now under agricultural
cultivation, they build up enormous populations that spread far beyond the insects’
ancestral range over the temperate zone spring and summer. Late-season migrants
are trapped by cold weather and die, because none of these species can diapause,
either as adults or juveniles. How can this be adaptive? At first, some researchers
speculated that the scenario was a ‘Pied Piper’ phenomenon, but others argued that
such massive losses would result in severe selection against migration. Current
evidence suggests the phenomenon of late-season deaths may be overstated.
Though many of these migrants do die, weather systems safely bear many other
individuals southward to winter breeding areas.
Early studies on insect migration understandably concentrated on the long-
distance flight of spectacular insects such as milkweed bugs and butterflies.
However, it has become increasingly obvious that migration is a far more
widespread phenomenon than previously suspected, including many small species
whose movements, relatively speaking, are neither far nor spectacular. In all
instances, migration is a distinct behavioral and physiological syndrome closely
intertwined with reproductive timing and strategy.
First, migratory flights are limited to a specific stage in the life history, after
the cuticle has hardened but before reproduction begins. Second, whether or not the
males are in accompaniment, migration always involves the female sex, and migrant
females are generally sexually immature and thus have a high reproductive value or
expected contribution to population growth (a high r, see Chapter 9). Most migration
correlates strongly with age (Fig. 3.9). That is, it occurs prior to egg development,
and while the development of the flight system is maximized, that of the reproduc-
tive system is minimized, a phenomenon that results in migration occurring chiefly
in young female adults. In the 1960s, C. G. Johnson, a leader in the development of
migration theory, termed this the oogenesis-flight syndrome. Third, migratory flight
3.5 Migration 121
Fig. 3.9 Making trade-offs. Duration of tethered flight as a function of age in three insects: (top)
the frit fly, Oscinella frit, (middle) the fruit fly, Drosophila funebris, and (bottom) the milkweed
bug, Oncopeltus fasciatus. The decline in flight occurs concurrently with an increase in reproduc-
tion (the oogenesis-flight syndrome). Similar effects have been demonstrated in a wide range of
insects, including bugs, flies, mosquitoes, aphids, moths, grasshoppers, and beetles
is segregated by time of day (or night) from various non-migratory activities. As a

result, circadian cycles, migration, and ontogeny interact. The conditions that pro-
duce delayed reproduction and adult diapause, such as short days, may also induce
migration. Migrants also are often similar to diapausing adults in having imma-
ture ovaries and hypertrophied fat bodies. In fact, in many species, migratory flight
appears to actually accelerate subsequent reproductive activity.
As with everything, additional studies of the oogenesis-flight syndrome have
shown the situation to be more complex than initially envisioned. Migration and
reproduction are not totally mutually exclusive, and many permutations of the syn-
drome have come to light. However, the oogenesis-flight relationship has guided
a whole field of studies into the physiological activity that underlies these behav-
iors, a syndrome that involves hormones, metabolic pathways, and the actions of
the nervous system.
3.5.2 Migration Under Ephemeral Conditions

In the arid northern subtropics and tropics of Australia, Heliothis moths breed fol-
lowing rainfall; if their productivity is high on the new vegetation, they soon become
large populations that move southward on the winds to agricultural areas.
Environments change, whether viewed over time or over space. Ecological suc-
cession occurs, local climate varies, and the surroundings that once were favorable
are suddenly no longer so. The impermanence of a habitat may result also from
seasonal or irregular climatic changes or simply from the ephemeral nature of the
habitat (such as flowers, fungi, carrion, etc.). In other cases, the insect’s require-
ments may change at different life stages, so different habitats would be needed;
acridid grasshoppers, for example, have different requirements for feeding and for
oviposition.
There is clearly a strong relationship between ephemeral habitats and an
increased incidence of migration, and evolutionary biologists have studied this rela-
tionship extensively from many different theoretical and mathematical perspectives.
The central thesis of their models is simply that migration evolves to keep pace
with changing habitat structure. Models based on statistical probabilities have been
expanded in various ways to incorporate factors such as the costs or risks of migra-
tion, carrying capacities, and population growth rates. In general, one can say that
developing such theories has proven easier than gathering extensive empirical data
to test them.
Seasonality places such a major imprint upon the environment that it is no sur-
prise to find examples of insect migration that are strongly tied to seasonal cues.
Likewise, habitats such as plant communities in early successional stages and small
or shallow bodies of water are relatively temporary, but their changes can be antici-
pated with some reliability. In other cases, environmental change is less predictable,
requiring more flexibility in the physiology and behavior of organisms that would
exploit it. For example, temperate and tropical forests are a very ‘patchy’ and
ephemeral environment for those bark beetles and weevils that require trees of a
particular species and in just the right condition.
Patches of new plant growth in the world’s arid regions are even more ephemeral
and unpredictable. Although the rainfall upon which they depend is seasonally
correlated, it is also scarce and unpredictable both in its timing and in its geogra-
phy. As theory would predict, migration is common among desert and dry-country
organisms. However, these migratory behaviors look quite different from seasonal
movements of insects such as the milkweed beetle and the monarch butterfly.
Particularly in tropical and subtropical arid regions, it is common to find cases of
migration that appear nomadic and opportunistic. Sometimes, at first glance it is not
even clear whether such movements represent true migration or simply extended
foraging in search of scattered food. One of the paradoxes of migration, however,
is that for an organism to respond appropriately to environmental change, in most
cases it must alter its physiology. This cannot be done instantaneously, so it would
be highly advantageous to have some way to anticipate changes that will occur.
When the environment is unpredictable, how can this be done?
3.5 Migration 123
The complexities of migratory behavior are well illustrated by the group of

grasshoppers collectively known as locusts. Five species cause the most crop
damage, and these have been the focus of most studies of locust biology. To
varying degrees, all show an extreme response to increased population densities
called gregarization. Each species occurs in two different races (phases) that are
biologically and morphologically distinct (see Plate 4). The sedentary or solitary
phase, colored inconspicuously green, gray, or reddish, is choosy about its food. The
swarm-forming migratory phase, marked with contrasting colors and longer wings,
devours any and all plants. In the desert locust Schistocerca gregaria, the differences
in the two forms are so great that they were considered separate species until 1921
when Boris Uvarov, a major pioneer in acridoid biology, demonstrated that the two
forms were a single species and developed a theory that explained the phenomenon,
now referred to as phase polyphenism. The term means a change in morphology
and behavior caused by density effects. (Whereas a ‘polymorphism’ involves
genetic differences, a ‘polyphenism’ involves differences that are environmental
in origin.)
The most conspicuous differences between Schistocerca locust phases occur
in their behavior. Which phase is present in the locust population depends upon
how crowded they, and their parents, have been. Previously uncrowded locusts live
quietly, tend to repel one another and tend to remain upon clumps of vegetation.
Hoppers that have been continuously crowded, however, are very active, strongly
attract one another, and dodge clumps of vegetation, being arrested only briefly even
by food plants. In its early stages, either phase can be formed from the other without
intervening generations. For example, insects of the solitary phase may transform
into migratory individuals if placed with swarm-forming companions.
Differences begin to show up as early as the second instar, when the nymphs
begin to exhibit mutual attraction and form up into bands that can contain many
thousands to millions of individuals. Then the still wingless nymphs begin to march;
the mass pattering of their feet is clearly audible as they travel, often downwind but
also in response to the direction of the sun. Before eclosing to adulthood, the band
may move tens of kilometers, stopping only at intervals when hunger demands.
With the winged adult swarms, locust migratory behavior reaches its peak. At
first sight, a migratory swarm appears to be a vast army flying single-mindedly
toward a fixed goal. However, photographic analysis has shown that in the swarm
as a whole the locusts are randomly oriented toward one another. Such randomness,
in combination with disruption from air turbulence, would be expected to lead to
dispersal of the swarm, were it not for a striking phenomenon. All the locusts at the
edges of the swarm orient toward the body of the swarm, and those at the back and
sides actively rejoin the swarm, probably responding to both visual, auditory, and
chemical cues.
Feeding by large swarms can strip vegetation down to bare earth, and during
years when major outbreaks occur, an area of approximately 20% of the earth’s
land surface can be affected. Because of this extreme economic importance, the
relationship between meterologic factors and locust swarming has been the object
of a great deal of research. About 60 years ago, R.C. Rainey developed a pioneering
model that related locust outbreaks and weather patterns (particularly those around
the Intertropical Convergence Zone) for the desert locust, Schistocerca gregaria.
During the locust plague years of 1954–1955, Rainey’s theories were put to the test
in a large international effort that monitored locust populations on the ground and
followed swarms by aircraft over much of the northern two-thirds of the African
continent. Results appeared to confirm the theory’s predictions, and for decades this
theory was established dogma.
Later reexamination of the data, combined with studies of other species, con-
siderably complicated the picture and cast some doubt whether the Inter-Tropical
Convergence Zone was involved at all. Rainey’s theory relied heavily on obser-
vations of gregarious swarms carried passively along during the day by winds at
different heights. A major discovery, confirmed by radar studies, was that soli-
tary locusts fly about at night, often for longer distances than gregarious locusts
do. Radar also showed that even when the wind was blowing strongly enough to
carry them along, these night-fliers exhibited considerable mutual alignment and
collective orientation. In other words, they were navigating (see below).
These discoveries changed the entire scientific view of locust migration. It now
appears that most large-scale migration takes place by single individuals migrating
at night, not by swarms moving across the countryside by day. In all locusts and
grasshoppers that have been studied, swarms have been found to function more as
extended foraging groups than as migratory groups.. The rolling swarm so visible
to an outside observer actually occurs from the movement of feeding individuals.
At the leading edge of the swarm, locusts settle and feed. This reduces the vegeta-
tion available to those locusts that are behind them, so these latter individuals pass
over this area, settle into a new leading edge, and feed. Most recently, it also has
been found that some locust species help this movement along by nipping at the
abdomens of the locusts in front of them, and escaping from others approaching
their own rear end.
Although the details of the swarming-foraging and individually migratory
aspects of locust life cycles will continue to be debated, most scientists agree that the
system provides a very effective way to exploit ephemeral habitats in arid to semi-
arid regions. Not surprisingly, a number of other insects take the same approach
to the problem, in the process making themselves major agricultural pests in drier
regions around the world. This includes several moths in the family Noctuidae,
particularly in the genera Heliothis, Helicoverpa, and Spodoptera. Interestingly,
large-scale mark-recapture, radar tracking, and detailed laboratory studies have con-
firmed that Spodoptera armyworms do disperse in a pattern that is closely associated
with rainfall occurring with the passage of the Inter-Tropical Convergence Zone.
3.5.3 Dispersal and Navigation
Scelionid wasps ride upon the backs of female grasshoppers; ultimately, those tiny
parasites will oviposit upon the grasshopper’s eggs. Tropical leafcutter ants are the
3.5 Migration 125
original suburban commuters; day after day they travel hundreds of meters back and
forth along trails that extend both horizontally along the forest floor and vertically
into the canopy. Every spring, vast numbers of insects are transported by winds up
through the Mississippi Valley into upper midwestern agricultural regions of the
United States and Canada.
Dispersal—movement away from one’s place of birth or from centers of popula-
tion density—is part of the life cycle of almost every living thing, and unsurprisingly
it occurs by a wide variety of methods. Dispersal may be random or directed, pas-
sive or active. Young individuals quite commonly leave the area in which they were
born. Usually such dispersal occurs either passively or under the juvenile’s own
volition, but in some cases dispersal may be encouraged by the indifferent or even
hostile behavior of their parents or nest mates. As a result, through time all pop-
ulations have a tendency to spread out spatially. Many different mechanisms may
be involved, from relatively simple responses to gradients of certain environmental
factors, to various active or passive dispersal mechanisms associated with the search
for a mate or food.
The distinction between passive and active dispersal is made mainly for
convenience; in reality a continuum often exists between them. For example, a
Pemphigus aphid, which lives on the roots of the sea aster growing in salt marshes,
is photonegative for most of its life. However, first instar nymphs are photopositive,
climbing up the sea asters until they set themselves adrift on the rising tide. Sea
breezes send them scudding across open water to be deposited at low tide on another
mud bank where they seek out and colonize new plants. Upon arrival, as a result
of their waterborne dispersal experiences they reverse their reaction to light and
become photonegative. In another example, female gypsy moths are unable to fly,
so natural dispersal of this well-known pest occurs primarily through young larvae
being blown on their silken threads by wind. (A similar mechanism, ballooning,
is employed by many newly hatched spiderlings.) However, gypsy moth larvae
vary in their dispersal propensity, and the behavior of first instar larvae depends on
several variables such larval density, food availability, and their mother’s nutritional
status.
Many flightless insects are transported by others, a phenomenon known as
phoresy. Human bot flies will attach their eggs to the legs and body of mosquitoes,
in this way transporting them to a human host for hatching and larval development.
Some Trichoptera larvae undergo their development within gelatinous capsules
upon the bodies of chironomid midges. Tiny wingless Mallophaga attach themselves
to the bodies of the hippoboscid flies which parasitize their bird hosts, in this man-
ner being carried from one host to another (Fig. 3.10). A great variety of mites ride
upon beetles, ants, and other insects; the insects are probably not injured unless the
numbers of mites become excessive. In several groups of wasps, a unique type of
phoretic dispersal has independently evolved on more than one occasion. All cases
involve parasitic females that have become so highly modified for tunneling into
the soil or food that they have permanently lost their wings. During copulation the
genitalia in these species lock together, so that the males carry the smaller females
about, suspended in this way, for considerable periods of time. This phenomenon,
Fig. 3.10 Hitching a ride. Two examples of insect phoresy. (left) Feather lice firmly attach
their mandibles to the abdomen of hippoboscid flies, obtaining transport to new bird hosts; it is
not unusual to find 20 lice (arrow) upon a single fly. (right) Phoretic copulation in a parasitic
wasp, Dimorphothynnus haemorrhoidalis (Tiphiidae). The male both inseminates and disperses the
smaller, short-legged wingless female (arrow), often carrying her to sources of nectar or honeydew
and in some cases actually feeding her by regurgitation
termed phoretic copulation, allows adaptation for a burrowing life combined with
effective dispersal of inseminated females into areas where new populations of hosts
may be discovered.
When wind, currents, or other organisms can move an insect to a suitable habi-
tat, little active orientation is necessary. However, many migrants cover very long
distances and/or must find very precise locations for breeding, feeding, or diapause
(Fig. 3.11). For these species, accurate navigation, the act of moving through a
place or along a route, becomes a major life task. As with so many other major life
tasks, navigation generally uses multiple cues and these cues serve as backups to
each other. It is postulated that each species recognizes some sort of cue hierarchy.
(However, even for such well-studied organisms as birds, there is no consensus for
any organism as to exactly what the hierarchy might be.)
Complex navigation is often divided into three general categories. The first is
piloting, which is considered to be the ability to use fixed known reference points
(‘landmarks’) to orient or navigate. The second is compass orientation, directional
ability without reference to a particular origin or destination; in essence, the insect
does not know where it is, but only what direction it is heading. The third is true
navigation, the complex ability to move toward a particular goal in completely unfa-
miliar territory without sensory contact with that goal. It is important to recognize,
however, that these categories are for human convenience; they do not necessarily
reflect increasing levels of complexity or the order in which the capabilities evolved.
Even though a number of common themes seem to apply across migration systems
and taxa, there are simply too few data to make such sweeping generalizations.
Navigation, particularly across long distances and/or lengthy time frames, often
appears so impressive that humans have been reluctant to ascribe it to simple pro-
cesses. However, a few detailed studies have shown that quite simple mechanisms
such as wind transport can result in surprisingly precise results.
To many researchers, true navigation implies that an organism possesses an inter-
nal ‘cognitive map’ that represents the geometric relations among points in the
3.5 Migration 127
Fig. 3.11 Migrating monarchs. (above) The fall, spring and summer migratory routes of North
American populations of the monarch, Danaus plexippus. Fall migrations take the butterflies to
congregation sites in coastal California and central Mexico; marked individuals have been docu-
mented to fly over 2000 miles during the fall migration to Mexico. (below) The exact destination
in Mexico was unknown prior to 1976; since then, at least 22 overwintering aggregation sites have
been reported in stands of mature oyamel (Abies religiosa) trees in Mexico’s trans-volcanic belt
at altitudes between 2700 and 3600 m. Since 1986 several of these forest areas (shaded) are now
legally protected in the Monarch Butterfly Biosphere Reserve
environment. To humans, ‘map’ implies visual cues. However, given the importance
of chemistry in other aspects of insects’ lives (see Chapter 6), such a map for many
migrating insects might well be chemically based instead, operating analogously to
the ways that homing salmon use the odor structure of water masses to move from
the ocean into river mouths.
Some migrants travel by day, others by night, and the sky provides potential guid-
ance at both times. For insects, day travel is by far better studied. At its simplest use,
a day-flying insect could just orient to the sun’s brightness, and move toward it with-
out correcting for the sun’s movement during the day. Even the elegant navigation
of the monarch butterfly has a component of this type; because the butterflies do
not start flying until they have warmed up, following the sun’s path takes them in a
southwestward orientation during their fall migration. However, additional precision
may rely upon other factors (Fig. 3.12).
In 1911 while studying desert ants in Tunisia, Felix Santschi demonstrated insect
sun-compass orientation for the first time. Using mirrors, he altered the sun’s appar-
ent direction as viewed by ants and showed that they would change direction at the
same angle as the displacement of the sun. Others have since shown that insects can
compensate for the movement of the sun across the sky. Time-compensation adds
more accuracy to sun-compass orientation; this requires that the insect have an inter-
nal timing method or biological clock. Such a system also has been demonstrated
for monarchs.
Finally, in addition to information provided by its position and movement, the
sun provides information via the plane of polarization of light or the e-vector. Some
insects clearly use both the position of the sun and patterns of polarized light, as has
been shown in elegant detail for the honey bee. Mechanisms for detecting e-vector
Fig. 3.12 Shifting direction. The rotational orientation hypothesis holds that monarch butterfly
orientation shifts clockwise at a rate of 1 degree per day throughout the year for all generations
of the annual migratory cycle. The spring equinox at overwintering sites apparently triggers north-
ward migration from Mexico. As time goes on, the orientation of subsequent generations shifts
as shown. Once the last generation reaches the overwintering sites, migration is repressed, but the
internal migratory clock keeps running
3.5 Migration 129
Fig. 3.13 Polarizing cues. Two representations of the e-vector sky or plane of polarization pattern
◦ ◦
of light with the sun in different positions above the horizon (24 left and 60 right) represented by
the black dot on the axis. Orientation of each black bar gives the e-vector direction for that bar, and
bar width indicates degree of polarization. Below each is the pattern of an optic interneuron firing
response obtained from a cricket exposed to changes in the polarization of incoming light. Arrows
indicate onset and end of the light stimulus
information have been well established in ants, bees, crickets and other insects.
Scanning the sky, the insect matches a small patch of polarization detectors in a
specialized portion of the retina to the distribution of the e-vector pattern in the sky
(Fig. 3.13); the firing of receptor interneurons speeds up or slows down, depending
on the degree of polarization.
Although they operate at different scales, both navigational cues and positional
cues are intertwined elements essential to spatial adjustment. Every information
source that is available—from infrared and magnetic waves and planes of light
polarization to visible landmarks of the earth, sea, and sky—is used somewhere.
We’ll be returning to these in following chapters in different contexts.
Chapter 4
Foraging and Feeding
4.1 Introduction
Gypsy moth larvae defoliate vast areas of forest across Europe, northern Africa,
and North America, feeding on over 50 tree species and reaching densities of tens
of thousands of individuals per hectare. South American phorid flies lay their eggs
on fire ants, and the grub that hatches invades the ant’s head, consuming the brain;
ultimately the ant’s head falls off and a new fly emerges. Bed bugs live only with
humans, and do poorly when fed other kinds of blood.
Much of the financial and social support enjoyed by entomologists derives from
the negative impact that some insects cause by feeding upon human food plants,
fiber sources, stored foods, waste products, farm animals, and even upon humans
themselves. In turn, the propensity of other insects to feed upon man’s pests has
made them noticeable allies. For the insects themselves, finding and consuming
food while simultaneously avoiding becoming food for other organisms are two of
life’s most essential behaviors. The former is the focus of this chapter; the latter will
be the subject of Chapter 5.
Let’s begin with an overview of foraging and feeding to put things in context.
First, a note on vocabulary: Purists would say that foraging is the preferred collective
term for all the behaviors involved in obtaining food, whereas feeding should be used
more narrowly for the actual act of consuming that food. As such, feeding research
generally is directed toward proximate factors, the ‘how’ questions that examine
ways in which behaviors are directly produced and regulated. Foraging research
tends to concentrate more on questions of ultimate causation. Simplistically, one
could call these ‘why’ questions. In practice, the terms feeding and foraging are
often used interchangeably, and in the end of course a complete picture involves
understanding both proximate and ultimate factors whenever possible.
One popular way to sort feeding habits is by principal nutrition source, e.g.,
plant, animal, carrion, dung, etc. It provides a convenient way to describe indi-
viduals or species, but the larger the taxonomic group one attempts to apply it
to, the less useful such a simple scheme becomes. Important exceptions occur in
most taxa. Furthermore, many insects that appear to feed on one type of food upon
closer scrutiny may be found to be entirely dependent upon another. For example

132 4 Foraging and Feeding
Fig. 4.1 Defying easy classification. Mosquito feeding habits differ with life stage and sex. Here,
female Aedes mosquitoes feed on an adult mantis; most of their feeding punctures are through
intersegmental membranes. The extent of such invertebrate blood feeding by mosquitoes in nature
remains to be documented, but these laboratory females were able to develop fertile eggs in the
same manner as vertebrate blood-fed individuals
although most people associate Drosophila fruit flies with decaying organic matter
such as overripe bananas, their larvae are actually eating, not the fruit itself, but
microorganisms associated with the decay.
Feeding behaviors also can be partitioned in various ways throughout a species’
life; many species feed at different trophic levels at different stages in their life
cycles. For example, mosquito larvae feed upon plankton and suspended organic
matter. Additional partitioning arises from sexual differences; adult females suck
vertebrate and invertebrate blood (Fig. 4.1) but males feed only on nectar. For a
surprisingly large proportion of insects, partitioning can be so extreme that entire
life stages do not feed at all—e.g., the pupa stage of holometabolous species. Insects
such as the silk moths, bot flies, and mayflies do not feed at all as adults. In such
cases, the larva acts as a rapidly growing ‘feeding machine,’ storing huge quantities
of reserves for adult life.
Another broad way to classify insect feeding habits is by selectivity. How choosy
are insects about their diet? Some insects accept a wide variety of foods (polyphagy),
a fact that often makes them significant pests. Migratory locusts, carpet beetles,
and cockroaches are familiar examples. Generalist feeders are especially likely to
adapt rapidly to new crop varieties, including transgenic crops, and to insecticides.
It is said that over 500 insect species targeted by crop protection strategies are now
resistant to a variety of insecticides.
4.1 Introduction 133
However, most insect species accept only a limited range of foods and usually
prefer one or two (oligophagy). Often this is reflected in the common name they
are given. Tobacco hornworm moth larvae, for example, feed on various solana-
ceous plants but prefer tobacco or tomato. Just as most insects are herbivores, so
also most herbivorous insects are food specialists, feeding on closely related plant
species (or sometimes just a single species). It is generally assumed that chemical
or physical differences among plants select for insect host-specificity. The suppo-
sition is that physiological costs impose trade-offs among adaptations to different
plant characteristics, but only a few genetic and physiological studies provide strong
explicit support for this hypothesis. Host specificity also might be favored by other
advantages such as use of specific plants as mating/aggregation sites, more efficient
host-finding, or better defense through such means as sequestering host-produced
toxic compounds.
Some insects take their choosiness to the extreme and exhibit strict specificity
to one food, often a single host (monophagy). Monophagous specialists occur in
almost every insect group. For example, certain leaf-mining caterpillars and gall
insects can develop successfully on only one species of host plant. Parasitic species
also tend to show a high degree of host specificity.
Like other organisms, insects need the appropriate balance of proteins and car-
bohydrates. Can insects ‘choose’ to balance their diet? Under laboratory conditions,
both locusts and caterpillars have been able to select foods that provide the appropri-
ate balance, but the composition of these foods was artificially extreme. It remains
to be shown whether insects can and do fine-tune their choices when feeding on
natural food with much smaller deficiencies.
Regardless of an insect’s food choice and the way it is classified, the ability to for-
age successfully resolves into a remarkably constant chain of behaviors, each link of
which facilitates the next: (1) food habitat location; (2) food finding; (3) food recog-
nition; (4) food acceptance; and (5) food suitability. (For some parasitic species, a
sixth link—host regulation—may operate as the parasite accelerates, retards, or oth-
erwise modifies the host’s physiological development.) This chain can involve any
or all of the senses, but the chemical sensory systems predominate. Consider the
bee wolf, Philanthus triangulum, which stalks honey bees that it captures, stings,
paralyzes, and carries back to its nest. As the pioneering ethologist Niko Tinbergen
described it:
A hunting female of this species flies from flower to flower in search of a bee. In this phase
she is entirely indifferent to the scent of bees; a concealed bee, or even a score of them put
out of sight into an open tube so that the odor escaping from this is clearly discernible even
for the human nose, fails to attract her attention. Any visual stimulus supplied by a moving
object of approximately the right size, whether it be a small fly, a large bumblebee, or a
honey bee, releases the first reaction. The wasp at once turns her head to the quarry and
takes a position at about 10–15 cm to leeward of it, hovering in the air like a syrphid fly.
Experiments with dummies show that from now on the wasp is very susceptible to bee scent.
Dummies that do not have bee-odor are at once abandoned, but those dummies that have
the right scent release the second reaction of the chain. This second reaction is a flash-like
leap to seize the bee. The third reaction, the actual delivery of the sting, cannot be released
by these simple dummies and is probably dependent on new stimuli, probably of a tactile
nature.
4.1.1 Food Recognition and Acceptance

Larvae of potato beetles react to the presence or absence of alkaloids that occur
in close relatives of the potato, but if their palps and antennae, which carry the
olfactory receptors, are removed the larvae will eat substances positively harmful
to them. Adult tsetse flies are drawn to their host by odors associated with animal
breath, urine, or skin secretions; final orientation is visual, but flies that have seen
a potential host and landed on it will begin to probe and feed only in response
to the host’s body heat, followed by formation of a blood pool at the bite. The
polyphagous chalcid wasp, Trichogramma evanescens, parasitizes the eggs of more
than 180 insect species; oviposition requires only that the object be firm enough to
walk upon, protrude from the surface, fall within certain size limits, and have no
dimension greater than four times any other. Odor, color, and surface texture are
irrelevant, but one important distinction is made. The little wasp will not oviposit in
eggs that have already been parasitized by another wasp, unless they are washed to
remove the smell the previous female has left behind.
On what specific basis does an insect recognize that an item is food? Or decide
to accept it? These questions have immense practical interest, and since the 1940s
a great deal of research has been directed toward answering them. In the pro-
cess, two schools of thought arose. Some scientists proposed that feeding was
a simple matter of the presence or absence of certain plant chemicals. Roughly
equivalent to ‘flavors,’ these compounds had no obvious direct nutritional value
but appeared to be acting as classical behavioral releasers (see Chapter 2), so they
were called token stimuli. Other scientists disagreed with this interpretation. Token
stimuli might be important in the initial discrimination of different plant species,
they argued, but factors more closely related to nutrition must mediate finer deci-
sions such as the discrimination of leaves of different ages. Thus, food preference
was a two-step or ‘dual discrimination’ process. When an insect showed food pref-
erences, it relied upon stimuli that were more closely related to nutrition than
to flavor.
Parts of both of these two theories now seem correct, as will become apparent
in more detail in Section 4.3. As with the potato beetle, nutritionally nonessential
(or even toxic) compounds may exercise an important influence as feeding stimu-
lants. The most nearly universal is certainly sucrose, which is an effective stimulus
for most insects. Nutritionally, many monophagous feeders have become physio-
logically incapable of surviving away from their usual host, and will respond only
to a stimulus that is species specific to their hosts; for some plant-eaters, this may
even be a substance originally evolved to repel herbivores. Polyphagous insects, on
the other hand, show such a broad responsiveness that they may respond directly to
nutrients without the intervention of a mediating chemical stimulus; any apparent
choice they exhibit is likely to be based on selective rejection due to repellent sub-
stances. One useful experimental technique for sorting out the relative importance
of plant chemical constituents is the ‘feeding preference test’ in which the insect is
given a choice of different food combinations; the amount of feeding is used as the
measure (Fig. 4.2).
Fig. 4.2 Determining insect food preferences. A simple smorgasbord of potential foods (clock-
wise from 12 o’clock are fresh pieces of Brassica rapa leaf, apple, lettuce leaf, white cabbage,
purple cabbage, and carrot) was offered in a petri dish with two Pieris rapae caterpillars released
at the center. Two hours later relative feeding damage, readily apparent visually, may be quantified
by comparing the food weights before and after exposure to the caterpillars. With modifica-
tion a similar technique can also be used to assay the relative importance of various plant
products by incorporating chemicals to be tested in standard agar discs made with powdered
leaves
It should be noted that well-designed laboratory experiments that concern feed-

ing behavior have many factors to take into consideration, including the age of the
insect, its previous experience, and its nutritional needs. For example, phytophagous
insects generally eat more in the middle of a developmental stage, and they generally
eat more in the light than in the dark.
Some (and probably most) insects also become less selective if they experience
a long period without food, and for some species, ‘long’ may be shorter than some
laboratory researchers realize. In one study of the red locust Nomadacris septem-
fasciata, gut samples were obtained at short intervals all through the locusts’ adult
lives; when the temperature was 30◦ C or above, their foreguts became more than
half empty within an hour.
Many insects that are deprived of food will sample items they would otherwise
reject. In general, grasshoppers were once considered to be unselective in their
choice of foods. In newer observations, several grasshopper species have shown
an increasingly wider acceptance of a range of host plants with increasing length of
food deprivation. Sometimes this may signal a simple need for water rather than for
nutrients. A well-hydrated locust will move away from a piece of wet filter paper,
but a dehydrated one will attempt to eat it.
Despite voluminous records of individual insect–plant and parasite–host rela-
tionships, a deeper understanding of the evolution of insect diet selection has
been difficult to gain. For example, why are some insects monophagous and
others polyphagous? A feasible explanation is found in the congruency hypothe-
sis advanced by Vincent Dethier. Both feeder and food are continually evolving
against a background of multiple pressures—the plant, prey or host by synthe-
sizing different chemicals, the insect by developing different sensory and central
decision-making capabilities. Specific changes occur in both organisms as a result
of random mutations. Congruency, or match between food source on one hand and
insect as feeder on the other, occurs whenever these two independently mutating
systems interact in such a way that formerly non-attracting chemicals now stimu-
late feeding. Sometimes, the change may be in the insect, resulting in the addition,
subtraction, or substitution of capabilities. For example, insect neural changes may
cause formerly neutral chemicals to become attractive or repellents to be no longer
detected. At other times, various chemicals may arise by mutation in the plant, prey,
or host.
For phytophagous insects, little information exists upon which to test the con-
gruency hypothesis, although a number of documented cases of sudden irreversible
shifts in insect feeding habits suggest that it is not unreasonable. It provides explana-
tions for a number of previously puzzling phenomena, such as why some plants such
as ferns have feeding deterrents even though they evolved long before phytophagous
insects did. The theory also explains why shifts in diet may occur in any direction,
including from polyphagy to monophagy, and why some apparently suitable plants
are not eaten.
At the same time, like any model, the congruency hypothesis presents an over-
simplified view. The model assumes reproductive isolation of the new mutant and
that mutations represent quantum jumps rather than intermediate states. It ignores
ecological factors. For example, the host plant or animal is not only a source of
food; often it equally may be a place to live. Thus, for a new feeding habit to be
established, the host need not necessarily be better nutritionally. For many insects,
plant selection is selection of a whole community—a microclimate, a shelter, a set
of predators and diseases. The picture at any given moment must be one of dynamic
equilibrium, and feeding must be viewed as a compromise between nutritional and
ecological optima.
Feeding habits also reflect only a subset of the choices that could potentially be
made, because for a plant and insect to interact they must be in contact with one
other. This has been graphically illustrated by introduced pest species throughout
the world, enthusiastically munching upon certain plants not formerly available to
them. Nor does accessibility require that it be the pest that is introduced; a striking
example is the Colorado potato beetle. This native American insect fed upon weedy
scrub until humans introduced the potato into its world; with this new opportunity,
the beetle population evolved to prefer potatoes over native hosts.
4.1.2 Regulation of Feeding
Preying mantises will consume a relatively constant number of house flies each day
when flies are continually available. Blowflies will maintain a relatively constant
daily food intake at a constant sugar concentration; if the concentration is decreased,
daily intake increases, and vice versa. In fact, most insects will feed to a point and
then stop. What causes a feeding insect to finally stop eating? What determines the
timing—how long and how often—of feeding?
One of the most thoroughly studied cases of feeding regulation involves Phormia
regina, a blowfly investigated by Vincent Dethier and his associates. The adult
blowfly needs only water, carbohydrates, and oxygen for maintenance, receiving
all other necessary materials during its larval stage; adult feeding occurs only to
provide locomotive energy. Thus, the blowfly offers a simple system for studying
two essential aspects of feeding behavior: the nature of the ‘on/off’ mechanism and
the nature of quality control.
The ‘on/off’ part is relatively straightforward. Blowfly feeding starts when a
fly steps on potential food, stimulating taste receptors on its tarsi. These lead to
the extension of its tubular mouthparts (proboscis), which thus contact the food
solution. At the proboscis tip, a lobe (labellum) has mechano- and chemorecep-
tive hairs that experience the consistency and taste of the food. The food quality,
along with peripheral and central adaptive processes, determines the sucking reac-
tion. Cessation of feeding behavior involves a homeostatic mechanism (Fig. 4.3) so
that, in the presence of excess food, a constant amount is ingested per day. As the
gut is filled, stretch receptors in the foregut are activated, and their firing inhibits
brain input from the external chemoreceptors that elicited feeding. At this stage in
feeding regulation, nonnutritive or metabolically useless foods are not distinguished
from nutritious ones. A blowfly will, for example, take up the useless sugar fructose
and regulate the amount until it dies.
Timing of insect feeding is more complex. In general terms, the length of the
period between meals is usually related to the quantity and nature of the previous
meal and to the amount of energy expended in the interim. For most insects, feeding
occurs at relatively short intervals of minutes or hours. However, some insects, such
as certain filter feeders, eat almost continuously. Others, such as some parasites,
may feed only at wide intervals of many hours, days, or weeks.
Even long-term changes in feeding behavior such as seasonality may rely on the
same basic physiological mechanisms as short-term feeding regulation, however.
A good example is provided by the face fly, Musca autumnalis, studied by John
Stoffolano. During the summer, face flies, which feed upon cattle blood and lay their
Fig. 4.3 Stopping when full. Metabolic homeostasis in the blowfly, Phormia regina: a model
system illustrating the interactions of external and internal stimuli and resultant feedback which
regulate the physiological aspects of feeding. The actual cessation of food intake is probably
mediated through negative feedback from internal receptors. The internal sensing upon which this
depends may be based on any of quite a number of variables, such as food bulk and/or gut capacity,
level of sensory stimulation, and length of time spent at continuous feeding
eggs in fresh, undisturbed cow manure, go through several generations. Although

adult males, which apparently feed on nectar, show their greatest feeding response
to glucose, adult female face flies exhibit a cyclical pattern of protein ingestion that
increases dramatically when eggs are being matured; they apparently cannot live
solely on protein, however. Female face flies respond to both blood and glucose as
long as their ovaries are undeveloped. However, as their ovaries begin to swell and
act on abdominal stretch receptors, female feeding begins to decline. Fully gravid
females do not feed at all. With responsiveness to blood and glucose effectively
removed, the lower level of responsiveness to manure they have maintained all along
now asserts itself. In the fall, under the influence of short days and low temperature,
the ovaries of the last generation of flies of the season fail to develop. Now face
flies of both sexes ignore cattle and feed only on nectar. On this diet and under
changed hormonal influences, a face fly’s fat body swells and its abdomen distends
until abdominal stretch receptors apparently respond with impulses to the central
nervous system, nullifying the sensory input from tarsal receptors. Feeding stops.
But gradually through the winter, fat is utilized; the fat bodies shrink and the crop
gradually empties. Spring marks the end of this developmental arrest. Flies then
feed on nectar until cattle are put to pasture.
Among plant-eating insects, feeding periodicity is also common. One important
evolutionary reason undoubtedly is that a plant is not a homogeneous chemical
entity. Rather, it is a heterogeneous, ever-changing microchemical environment.
Important plant constituents such as carbohydrates, fats, proteins, minerals, alka-
loids and essential oils vary with a number of factors, such as time of day and
season, plant growth stage and tissue, climatic and soil conditions, etc. For example,
consider the seasonal changes that occur in oak leaves. New leaves contain rela-
tively more protein, water, and sucrose and less tannin than older leaves. As tannin
increases, protein availability is reduced until its decline becomes a limiting factor
for most herbivores. Not surprisingly, most lepidopteran larvae attacking oaks feed
early in the season on new foliage; by so doing, they make the best use of available
protein and avoid most of the toxic tannins that slow growth rates and reduce fecun-
dity. Plants also show shorter-term cycles, the most familiar example of which is
the diurnal rhythm of nectar production and flower opening in many plants. Honey
bees, apparently able to remember not only the location where food is available but
the time of day at which this occurs, continue to gather only during the previously
learned hour.
Environmental variables are also important in the establishment and maintenance
of feeding periodicity, and it would be a mistake to underestimate their effect. For
example, most ants are more active at some times of day (or temperature and light
ranges) than at others. Many ant species in temperate climates forage throughout the
day, but in hotter lands some regularly stop feeding for a midday break. Many others
forage only at night; 54 out of 58 species of ants living in the Sahara are nocturnal.
The effect of such foraging rhythms, of course, is to confine the outside activity of
the ants to times when temperature and humidity are least harmful and/or to when
their food is most easily obtained. It is sometimes difficult to determine the degree
to which an insect’s foraging rhythm depends upon internal biological clocks, how-
ever. For example, leaf-cutting ants, Atta cephalotes, normally collect leaves all day,
beginning early in the morning. Shading their nest entrance between 5:30 a.m. and
6:00 a.m. will delay the time of appearance of the first workers, but no amount of
light before 5:30 a.m. will make foraging start. Some internal clock is apparently
responsible for bringing the workers to the nest entrance to inspect for light. At the
same time, many studies have shown that when ants are fed regularly at the same
time for three to five days, they learn to search for food at that time on succeeding
days. This implies that ants in the wild would learn to forage when food was most
plentiful. Thus generalizations must be made quite cautiously. Probably the daily
foraging activity of most insect species is the complex result of minute-to-minute

conditions, an internal rhythm itself partly dependent on past conditions, and, in the
case of social species, the individual past experience of the colony members.
4.2 Foraging Strategies
In olfactometer tests the wasp Diaretiella, a parasitoid of cabbage aphids, is attracted

to collard leaves; wind tunnel experiments show that it reacts to sinigrin by walking
or flying upwind. The ichneumonid wasp Pimpla ruficollis is a parasitoid of the
European pine shoot moth, but for the first few weeks of adult life it is repelled by
the odor of pine; only when a female is ready to oviposit does she find pine odors
attractive and return to the forest where her host is located. The aquatic larva of
many chironomid flies constructs silken tubes with a conical catch-net spun across
the lumen of the tube; periodically the larva devours its catch-net with adhering
debris that has been swept into the tube by the water currents.
Procuring and recognizing food are items of overwhelming importance on the
life agenda of all insects, whether they live by scavenging, chewing or sucking upon
plants, or capturing and/or parasitizing prey. Might it be possible not just to cata-
logue foraging and other behaviors but actually to predict them? In 1966, Robert
MacArthur and Eric Pianka hypothesized a concept that came to be called optimal
foraging theory. First, because of natural selection animals should evolve ways of
maximizing their food intake (benefit) while minimizing time and energy expended
(cost). Second, whenever several roughly equivalent patterns of behavior are avail-
able to an animal population, the pattern with the most favorable cost-benefit ratio
should eventually defeat other less efficient patterns and come to prevail. For con-
venience, MacArthur and Pianka viewed these patterns of behavior as though each
pattern was equivalent to using a different ‘strategy’; the term does not imply con-
scious intent, but something like a general set of ‘rules for behaving’ that in the
animal’s evolutionary history has usually resulted in a particular outcome.
From its inception, optimal foraging theory showed promise of being a powerful
theory, in that its predictions could be tested quantitatively. Factors presumed to
underlie particular foraging decisions could be identified, and one could measure
cost-benefit ratios in the currency of calories of food taken in and energy spent.
MacArthur and Pianka, and others following them, identified a cost-benefit ratio
for many foraging behaviors. Their studies showed that animals are not perfectly
adapted in terms of cost-benefit ratio, but that they come reasonably close, given
certain imperfections in their sensory and information-processing systems.
As this theory came to be invoked to cover virtually all behavior, its name
changed to simply optimality theory. The idea that animal behaviors can be viewed
in terms of a cost-benefit ratio has been an extremely important breakthrough in
evolutionary theory. Today, optimality theory plays a major role in many research
studies on feeding and foraging (see Section 4.2.2).
Feeding does not occur in a behavioral vacuum, however; animals balance many
different needs for overall efficiency in satisfying fitness requirements. For example,
4.2 Foraging Strategies 141
a generally favorable cost-benefit ratio has to include not only the need for food but
also the need to escape from predators and the need to reproduce effectively. And
what about situations where insects compete with other insects for the same lim-
ited resource? In the real world, a decision-making animal often must consider the
behavior of others. In 1974, John Maynard Smith elaborated upon optimality theory
to predict optimal outcomes when individuals interact with one another. Borrowing
from economic theory, he applied game theory to explain how animals develop a
strategy that considers competitors’ behaviors. Equivalent to coaching strategies
in a competitive sport, game theory predicts that an individual will choose differ-
ent strategies in proportion to their relative effectiveness against choices made by
opponents.
4.2.1 Herbivory
Plant eaters comprise almost half of the world’s insect species today; they occur in
about one-third of the major insect groups, including some of the most numerically
successful ones such as the Lepidoptera. Fossil evidence indicates that terrestrial
arthropods have been consuming plants for at least 400 million years, and studies
of a range of ecosystems demonstrate that insects consume more plant mass than
all vertebrate herbivores combined. Even the vast ungulate herds of the African
savannas are estimated to be responsible for only 15–35% of the herbivory there,
with insects causing 10–75%.
Herbivory (also called phytophagy or simply plant-feeding) most commonly
involves feeding upon foliage, but one can also find sap feeders, root feeders, seed
feeders, and, in the case of woody plants, bark feeders. (In fact, with the excep-
tion of coal there is hardly a source of organic carbon not used to some extent by
some insects.) To these directly phytophagous species may be added the complex of
insects that feed on the rotting debris below the plants and even the opportunists that
partake of the sugary honeydew excreted by the sap feeders. Each of these species
in turn serves as a potential host for one or more parasitic insects that may in turn be
attacked by their own specific enemies, called hyperparasites. Thus, even a single
plant species can form the base of a complex food web (Fig. 4.4).
Herbivory is such an important part of insect behavior that it keeps cropping up in
different contexts. Thus, for example, in Section 4.2.6, we’ll consider those special
cases in which insects can be said to be farmers, rather than hunter-gatherers. In
Section 4.3, we’ll be returning to the subject of herbivory from the perspective of
insect-plant coevolution.
4.2.2 Active Search
Blood-sucking Diptera can be attracted with CO2 . Phytophagous aphids come to

light of longer wavelengths, especially yellows and greens. Dragonfly larvae often
stalk, then snap at, nearby moving objects within certain size limits; hungry preying
mantises behave in a similar fashion.
142
Fig. 4.4 Life is seldom simple! Food web associated with cabbage plants in Minnesota. Herbivore species include 11 leaf feeders, 10 sap feeders, and
4 root feeders. Other trophic levels were less thoroughly studied, but included at least 21 detritivores, 79 honeydew feeders, and 85 predatory, parasitic, or
4 Foraging and Feeding
hyperparasitic carnivores
‘If you want dinner, go find it!’ For by far the majority of insects, food location
involves some manner of active search. Plant-feeding insects (if not hatched on their
host) may find their hosts by seeking appropriate habitats, by increased activity that
maximizes their chances of encountering a plant, by completely random activity
halted by strong arrestant properties of the plant, or by attraction to a plant from a
distance by smell or vision or both. Predatory insects, on the other hand, are usu-
ally tuned to movement, which is probably the most widely shared characteristic of
potential prey; parasites often cue in to by-products of their hosts’ feeding activity
such as frass (see Chapter 2).
Because active search for food requires energy expenditure and time involve-
ment, one would expect selection to favor behaviors that increase its efficiency.
Optimality theory predicts that when food is mobile, it is advantageous to catch
more food in less time, to catch larger food items in the same time required to take
smaller ones, and to avoid chasing food or hosts likely to escape or prove unsatis-
factory for some other reason. The variety of means by which such ends have been
approached is nearly limitless.
What of those species like gall midges, in which the mother oviposits directly
on or in the plant, or like parasitic wasps that oviposit on an animal host? Locating
food would appear to be no crisis for these species, because the young are literally
surrounded by it, but in actuality the burden of food location has simply been shifted
to another life stage.
When food is unevenly distributed, organisms must develop ways to locate and
identify the richest clumps, or patches, of that food. Studies first done with birds
but probably applicable to most searching foragers show that optimal foraging has
two phases. When they first locate a food source, many animals spent the first few
minutes in a ‘sampling phase’ that looks like random wandering though several
patches. Then, in what is called the ‘exploitation phase’ they spent most of their
time foraging in the richest patch.
Under natural conditions, however, food density and condition are often in a state
of flux. Consequently, a useful strategy is to sample food availability and condition
periodically to keep up with current conditions. In the case of nectar feeders, the
amount of food that can be obtained from any one flower is rather limited, and
for maximal efficiency one might expect bees, for example, to be able to exploit
new food resources as they become available and to differentiate between more
and less rewarding flowers in bloom at any one time. Studying bumblebees in the
1970s, Berndt Heinrich performed some of the first direct analyses of insect forag-
ing behaviors in terms of caloric costs and benefits. He confirmed that bumblebees
generally collect nectar from several different plants during a foraging trip, spending
most of their time on the richest flowers while simultaneously hedging their bets by
investing some energy in less productive species. By experimentally enriching some
of these ‘minor’ flowers with sugar syrup, Heinrich demonstrated that an individual
forager immediately will switch from its previously most preferred flower and adopt
the fortified flower for as long as the sugar syrup is added. (Nectar is far more than
just sugar water, however. It contains considerable quantities of various amino acids
that are nutritionally essential to adult insects unable to synthesize them.)
Exploiting a food uses it up, of course. Imagine a ladybug beetle on an aphid-

infested plant. At first it can eat voraciously, but as it feeds, fewer and fewer aphids
remain and they become progressively more scattered. The rate at which the beetle
takes in food slows down progressively. By comparison other patches begin to look
more attractive, but there are potential costs to moving on, such as the loss of time
and energy and the possibility of predation during travel. Might it be possible to
predict how long a forager should stay in the dwindling food patch versus striking
out to find a new one? This question led Eric Charnov to develop an optimality
model that generated a result known as the marginal value theorem. He based his
model on a concept called the marginal rate of food intake, defined as the rate of
food intake associated with the next food item in the forager’s current patch.
The marginal value theorem makes clear testable predictions that—as for other
optimality models—can be presented in mathematical, algebraic, and graphical
forms (Fig. 4.5). In non-mathematical terms, one prediction is that a forager should
stay in a patch until its marginal rate of food intake there is equal to that of the
average rate of food intake across all available patches, i.e. keep feeding until you
could, on average, do just as well somewhere else (after accounting for the costs
you must pay to get there). A second, related prediction is that the greater the time
or energy it would take to get to another patch, the longer a forager should stay in
the current one, i.e. travel costs increase with longer distance, shifting the balance.
A third prediction is perhaps less obvious. If the patches in an area are already of
Fig. 4.5 Foraging optimally. The marginal value theorem allows one to predict how long an
animal should continue to feed at a particular location (food patch) based on cumulative food
intake. For example, it might be applied to bees visiting patches of flowers for nectar. A simple
graphical solution is depicted here. First, generate a curve that represents the cumulative food
obtained over time from an average food patch. Initially when food is abundant the curve rises
rapidly; later as food is depleted it levels off. Next, quantify the travel time between food patches
and determine an average travel time (t). Starting at point t, draw a line to the right that is tangent
to the cumulative food gain curve. From the point of tangency, drop a line to the x-axis. The
intersection of that line with the x-axis gives the predicted optimal time (T) for the forager to
remain in the patch, optimizing benefit/cost ratio
generally poor quality when the forager enters one, it actually will stay longer than
if it were foraging in an environment full of more profitable patches. This is because
in a poor patch, it takes longer to accumulate the amount of energy needed to make
up for the travel costs associated with a move.
In some cases, researchers have been able to gain insights into the nature of the
selective forces acting upon a selected insect species by building models for various
alternative foraging decisions and comparing the predictions of each model with
its observed behavior. For both practical and theoretical reasons, a major focus of
insect research on marginal value theorem concepts has been with parasitoid wasps,
which lay their eggs on insect hosts that are distributed in discrete patches in the
environment. Many of these parasitoid-host systems lend themselves easily to labo-
ratory manipulations of pertinent factors such as host quality and host and parasite
density. Applied ecologists have hoped to gain information that could inform bio-
logical control measures; population ecologists have seen these insects as a way to
explore the dynamics of victim-enemy relationships.
Models such as those provided by optimality theory and the marginal value the-
orem should be seen for what they are: not some sort of inviolate truths, but instead,
valuable research tools. Rarely has there been a perfect quantitative fit between
model predictions and empirical test results. Still, their use has yielded valuable
insights and sometimes has led in unexpected directions. For example, optimal
host selection models predict that when an insect has its choice of two foods or
hosts of different potential values, the choices it makes between them depend on its
encounter rate with the more profitable species. If the insect’s encounter rate with
the more profitable species is above some threshold value, the less profitable species
should always be ignored; if encounters with the more profitable species are below
this value, the less profitable one should always be attacked.
Studies with a braconid parasitoid Asobara tabida that attacks Drosophila show
both the ways such studies work, and the potential pitfalls involved with laboratory
experiments that purport to represent natural field behavior. At the high population
densities typical of most laboratory experiments, Asobara is selective when offered
the choice between two host species that differ in survival probability for its off-
spring; it also avoids superparasitism. However, as is probably true for most insects,
encounter rates under natural conditions are much lower. When Asobara encoun-
ters only one host or less per hour, the wasp always accepts either host and also
superparasitizes.
Prey-catching wasps show a relatively constant host stinging sequence that
appears to be genetically programmed rather than learned; in all known cases, the
first sting is directed at ganglia involved in locomotion and defense, in particular
paralyzing the prey’s legs, which might otherwise be used as powerful defensive
weapons. Then the prey is transported back to the nest in various ways, often with
considerable effort. However, the wasp Ampulex compressa has evolved an effi-
cient alternative, zombification; its cockroach prey (usually Periplaneta americana)
follows it home like a submissive dog on a leash.
To accomplish this feat, Ampulex relies upon two injections of venom. The first,
delivered into the thorax, transiently paralyzes the roach’s prothoracic legs. The
second, given very precisely into the neck, targets the cerebral ganglia. In response,
rather than fleeing, the roach stays put, frantically grooming itself, while the wasp
cuts the roach’s antennae with its mandibles and feeds on its hemolymph. Then
Ampulex grabs one of the roach’s antennal stumps and, walking backward, leads the
unprotesting roach back to a pre-selected burrow for oviposition. The stung roach
follows the wasp inside and lethargically waits while the wasp glues an egg to its
cuticle, then exits and closes the burrow entrance with small pebbles. For days, while
the egg hatches and the larva feeds, the roach remains alive but passive; it only dies
after the pupa finally becomes ensconced in its abdomen.
Researchers have become quite interested in the proximate mechanisms involved
in zombification. To resolve whether the venom is injected directly into the central
nervous system or simply diffuses in from the hemolymph, Frederic Libersat and
colleagues injected Ampulex with a mixture of carbon-14-radiolabeled amino acids;
they found that the radioactive signal was positioned in the central nervous system
so precisely that they compared it to the most advanced stereotactic administration
of drugs.
How can a small change in one particular brain center, protein, or gene lead
to such a dramatic effect? In most cases the mechanisms that underly such a dra-
matic behavioral change are unknown. In a first step toward an answer, D. G. Biron
and colleagues have been examining the observed behavioral changes in the insect
host using proteomics. A rather young discipline, proteomics can be defined as the
qualitative and quantitative comparison of proteomes (genome operating systems)
under different conditions to unravel biological processes. Because it allows one to
study the host-parasite interaction at the genetic level precisely in the midst of the
manipulative process, the technique appears very promising as a tool for exploring
the proximate mechanisms responsible for host manipulation.
4.2.3 Trapping and Ambush

The bola spider twirls a sticky ball of silk from a silken thread. Moths approach the
swinging ball, drawn by an impregnated attractant, and become stuck to it, where-
upon the spider retrieves its catch and consumes it. Little zorapterans, members of
one of the smallest and most poorly known insect orders, live in decayed logs and
sawdust piles; they feed primarily on fungal hyphae and spores but are not adverse
to an occasional meal of nematodes, mites, or springtails should one of these crea-
tures pass by. Some preying mantid species are exquisitely camouflaged as leaves;
an unsuspecting herbivore that approaches is quickly snatched and consumed.
Remaining relatively motionless to trap incoming food is less common than
moving about to search for it, but a sit-and-wait strategy has many advantages,
both reducing searching energy expenses and allowing an insect to engage in other
concurrent activities such as environmental monitoring for mates and/or territorial
intruders. Ambush is the most common lay-in-wait strategy, and it can be extremely
effective when employed by an agile predator. The grand masters of lay-in-wait
are undoubtedly the spiders, whose evolutionary repertoire includes not only web
Fig. 4.6 Trapping supper. Ambush by the worm-lion larva, Vermileo comstocki. In frames (1)
and (2), the larva makes a pit; then it hides at the base of the pit (3). An approaching ant (4) slides
into the pit, where it is seized and paralyzed (5). Hauled under the surface of the sand, the ant’s
body fluids are sucked out by the worm lion (6)
weaving but also quite a wide range of other ambush methods. In a similar vein,
many predatory insects such as tiger beetles, ant lions, and worm lions construct
devices such as pits or snares that increase the probability of locating and captur-
ing prey (Fig. 4.6). Considerable time and energy are often invested in the building
and maintenance of such devices. However, they can result in a good return in food
which otherwise would be so widely dispersed as to be uneconomical to wait for
and perhaps difficult to catch as well.
Lures can improve the odds that dinner will appear. Certain assassin bugs have
forelegs covered with hairs that exude small droplets of a highly viscous substance
that looks like dew; when the assassin bug spots small, fast-moving prey such as fruit
flies, it raises its forelegs and holds them parallel to the ground. Flies attracted to
the ‘dew’ become entrapped and quickly are consumed. The Javan bug, Ptilocerus,
has a tuft of bright-red hairs on its body, marking the spot where a gland opens
beneath the abdomen; secretions from this gland are very attractive to ants. However,
after partaking of these secretions, an ant collapses, apparently from narcotic action,
whereupon the bug pierces the ant through the neck and promptly sucks it dry.
Attracting prey often involves some sort of deceit or false sensory message.
Certain Malaysian preying mantids greatly resemble flowers; unsuspecting prey
attracted visually to these ‘blossoms’ become a quick meal. Similarly, female
Photuris fireflies flash false coded messages to sexually lure unrelated males,
which are then eaten. The predatory larvae of many green lacewings have the
habit of covering themselves with camouflaging debris of one type or another; in
the case of Chrysopa slossonae, this disguise allows access to the woolly aphids
that are its sole diet without alerting the aphids’ ant guards. These behaviors are
described as aggressive mimicry; despite the superficial similarity in terminology,
however, they are quite different from the classical types of mimicry treated in
Chapter 5.
Obviously, a lie-in-wait strategy requires that the food itself be mobile. Usually
this implies a diet upon animal prey. However, when wind or water currents exist,
small plants and microorganisms may become an important diet item for lie-in-wait
food gatherers. Trichoptera larvae of the family Hydropsychidae use nets spun with
their silk glands to capture drifting food particles in streams. These feeding nets
sometimes have an exceedingly fine mesh, allowing the caddisfly larvae to graze
upon fine particulate organic matter, phytoplankton, and bacteria strained out of
the moving water. The evolution of hydropsychid larvae reveals a tendency toward
more complicated larval feeding structures and smaller capture-net mesh size. In
some species, plant detritus may comprise more than half of the diet of younger
instar individuals.
4.2.4 Parasites and Parasitoids
Little wasps in the genus Copidosoma, which lay their eggs in the eggs and young
larvae of noctuid moths, regularly produce brood of over 2,000 young. Tsetse flies
have been dubbed the world’s least specific bloodsucker; they will apparently feed
indiscriminately upon any vertebrate they encounter. The wasp Ichneumon eumerus
attacks the caterpillars of Maculinea rebeli that live inside the brood chambers of
ant nests; to reach them, the wasp releases chemicals that cause the ants to lock into
combat, leaving the caterpillars defenseless and available for oviposition.
Organisms that feed upon a host that they do not kill are termed parasites.
Examples include mosquitoes, lice, bed bugs, fleas, and the many other insects
that feed on humans and other animals that often have significant health and med-
ical impacts. Haematophagy or the utilization of blood as food occurs in six insect
orders. The order Diptera has the largest number of haematophagous species and
includes both ectoparasitic forms that feed from the animal’s body surface and
endoparasitic forms that feed from within. Distinctions can be drawn even more
finely, as for example between ectoparasites that stay permanently on their hosts
and those that are off their hosts except when feeding, or between ectoparasites that
are the most to the least host-specific with regard to both their host location and host
preference patterns.
Parasitism is a major life style, occurring in so many of the world’s organisms that
it deserves, and has, coverage in textbooks of its own. Here, we will restrict mention
to some of the relatively unusual insects that are external and internal parasites of
other insects. Their parasitism can sometimes have enormous effects on the behavior
of their hosts.
First, however, a distinction must be made between the true parasites and another
much larger group of insects that live at the expense of another insect (host) that
is always consumed and dies as a result. Members of this group (which includes
the so-called ‘parasitic’ Hymenoptera and some Diptera and Coleoptera) are termed
parasitoids, and like true parasites they may feed either externally (ectoparasitoids)
(Fig. 4.7) or internally (endoparasitoids). Predators also consume their prey, of
course, but parasitoids require only a single host individual to complete their
development; predators generally require more than one prey. Most parasitoids are
parasitic only in the larval stage, being free-living as an adult, but in the real world
exceptions to strict categorization are not unusual.
Parasitoids may be solitary or gregarious in their development. Some paralyze
their host and develop rapidly, a situation that differs only slightly from predation.
Others manipulate host physiology in ways that permit the host to continue to grow
and develop after the parasitoid lays its eggs; meanwhile parasitoid development
may be concurrently delayed until the host becomes sufficiently large enough to
sustain it.
Studies of parasitoid-host interactions are fundamental for implementing biolog-
ical control of pest insects. A large body of published literature addresses aspects
of parasitoid biology for a wide range of species. Some of these are discussed
elsewhere in this book, in contexts as varied as enemies of solitary wasps, host
finding and learning host odors, foraging strategies, host marking, courtship, and as
a selective force in parental care.
In considering parasites and parasitoids, one once again is faced with a choice of
viewpoints. Whereas many researchers study the proximate mechanisms involved
in such behaviors as predation and parasitism (such as the zombification already
Fig. 4.7 Growing up together. Melittobia digitata, a gregariously developing ectoparasitoid,

attacks the prepupa of various solitary bees and wasps. The female wasp oviposits up to several
hundred eggs; upon hatching, they develop rapidly and completely consume the host
mentioned), others are particularly interested in parasitism’s ultimate evolution-

ary context. What is the best way to interpret changes—particularly behavioral
changes—observed in a parasitized host?
It would probably be rare to find a parasite that had absolutely no effect on the
behavior of its host. However, simple changes such as a host becoming less active
are generally just considered to be straightforward byproducts of the infection.
Complex aberrant behaviors are a different story. They seem to beg for an evolu-
tionary explanation. Are these changes in behavior simply a side effect of infection?
Alternatively, might they be host adaptations that reduce pathological consequences
but have little (or only accidental) adaptive value for the parasite? Or perhaps they
represent manipulation by the parasite for its own purposes?
One who has searched most intently for an answer to these questions is Laura
Beani of the University of Florence. Her work involves a social paper wasp Polistes
dominulus, a strepsipteran parasite Xenos vesparum, and the concept of adaptive
manipulation. Adaptive manipulation includes any alterations in host behavior that
increase a parasite’s ‘inclusive fitness’ (see Chapter 10); thus, it includes not only
direct changes like those that improve the probability of the parasite being transmit-
ted but also indirect changes that improve the host’s fitness and thus in the process,
its own.
The normal behavior of paper wasps is particularly well-studied. In brief, mated
females emerge each spring to start new nests. At first, they must feed and care
for their young by themselves, but as their offspring become adults, they join their
mother in feeding and caring for subsequent young and in expanding and defending
the nest.
Fewer people know of the Strepsiptera, an order of insects related to the beetles
with only about 600 species. The early stage larvae and the short-lived adult males
are free-living, but most of a strepsipteran’s life is spent as an endoparasite in other
insects (see Plate 6). The first-instar larva (called a triungulin) enters a host, usually
in its abdominal region. Once inside, they induce the host’s tissue to produce an
internal pouch-like structure that protects them from the host’s immune defenses
as they feed and grow. Though the parasites do not feed directly on their host’s
body organs, they do greatly reduce their nutrient intake. Sometimes over 30 of
these parasites may be found on one host, though three or four is far more common;
when several are present, they may occupy up to 90% of the abdominal volume of
their hosts.
For most of the period of infection, no external signs of the parasites are visi-
ble. However, after pupation, adult males leave the host body, and their sclerotized
cephalothecae remain behind as a visible sign of this ‘stylopization’. (Stylopidae
is the best-known strepsipteran family.) Females, however, have no wings or legs.
They remain inside, with their heads barely protruding from the host abdomen. Their
odor attracts males that mate by rupturing the female’s brood canal opening, which
lies between the head and prothorax. Each mated female later produces many thou-
sands of triungulin larvae that emerge from the brood opening on her head and
disperse to continue the cycle.
Strepsiptera attack many kinds of insects in at least seven different orders, with
hymenopterans being one of their more preferred hosts. It should be noted in pass-
ing that communal and social insect colonies must present a particularly appealing
target for parasites, given their high density of potential hosts of a similar genotype.
However, successfully breaking through these social defenses is generally so diffi-
cult that colonies have been called ‘factory fortresses’. Studies of social wasps have
suggested that nest aspects such as multiple comb construction and nest envelopes
may have evolved in part to provide some defense against parasitoids and preda-
tors. Behavior is also important; for example, in Chapter 1 we discussed honey bee
defense against brood diseases.
Every aspect of Strepsiptera that has been studied has proven to be highly
unusual in some aspect, from their bizarre and complex life cycles to their genome,
which is one of the smallest known for any insect, although the size of the
entire 18S ribosomal DNA sequence is among the biggest, containing a number
of unique expansion segments. No aspect, however, is more unusual than the many
ways in which these parasites alter the morphology, physiology, and behavior of
their hosts.
Striking changes in host morphology are often evident even before adults become
visible. The parasites’ presence may change the color and shape of the host’s
abdomen. In many host species, both male and female become sterile, having
been effectively castrated by their parasites. Females of solitary bees often have
their pollen collecting baskets greatly reduced, and the sting is also often reduced
in size. In males, the copulatory apparatus is often greatly reduced. Male and
female hosts often appear to have acquired some appearances of the opposite
sex.
Behavioral changes also occur. In ants, unhealthy workers become diurnal, aban-
don their nests, and climb high on grasses and bushes. When a Stylops female is
releasing larvae, her normally slow-flying Andrena bee host will suddenly race
from flower to flower, dragging her abdomen through the stamens and spreading
triungulins that other foraging bees will take into their crops, becoming vectors for
infection.
It was with this background in mind that Laura Beani and William Hamilton
watched an aggregation of Polistes behaving strangely in the surroundings of her
house in Italy. It looked like an overwintering aggregation (see Plate 5) but it was too
early in the season. At Hamilton’s urging, Beani began to investigate their unusual
behavior. From her own inspection and previous observations by others, she was
able to piece together a remarkable story. The wasps in the aggregation were para-
sitized by the strepsipteran Xenos vesparum, and their strange behavior was part of
an even bigger picture of aberrant acts that spanned the seasons.
After overwintering, affected Polistes females did not form a stable association
with healthy nest foundresses; they showed no inclination to nest or to act as helpers.
However, some made unusual short-range ‘migrations’ from one nest to another.
Resting on the nest combs, they were occasionally attacked by the nest’s owners,
but usually they were just ignored.
In the summer, wasps hung around listlessly in unusual aggregations like those
Hamilton and Beani had observed; these were always away from the nest, often
at lekking or pre-hibernant sites. The groups were initially and mostly formed by
workers, not sexuals, and though aggregations might persist for days, individual
turnover rates were high.
When Beani artificially infected wasps with strepsipterans and moved the wasps
into large cages in the laboratory, these infected workers deserted the colony early.
Though they were of no help at the nest, they were not evicted by healthy wasps,
but left to their own volition. When she caged naturally parasitized and unparasitized
wasps together, the stylopized wasps hung out sluggishly in the corners, ignored by
the healthy ones, and they lived a long time, unusual since workers normally die and
do not overwinter.
What does all this mean? It appears that the Xenos/Polistes interaction represents
a case of adaptive manipulation. How might such diverse alterations enhance par-
asite success? By castrating its hosts, Xenos shifts the host’s resource allocations
from reproductive organs to energy reserves that can directly benefit the parasite.
By causing infected female wasps to cluster in groups while healthy gynes scatter
to occupy good nesting sites in early spring, triungulin-laden overwintered females
essentially wait within an aposematically defended herd until healthy wasp larvae
are available to be parasitized. Likewise, the high turnover rate in the strange sum-
mer aggregations can be explained by the fact that these are mate-encounter sites
for the parasite, not for the castrated hosts.
If only for curiosity, one must, however, return to a proximate-mechanism ques-
tion. How do the parasites do this? In Polistes, normal worker behavior is known
to depend on juvenile hormone. The stylopization syndrome resembles the pre-
hibernation physiological state of overwintering wasps, when they are in a diapause
ovarian state that can be broken by applying juvenile hormone. Studies have shown
that the strepsipterans alter juvenile hormone production, not the wasps’ sensitivity
to the hormone, by their infection. Strepsipterans are the only parasitoids that cause
a host’s life span to be lengthened to allow for the parasitoid’s life cycle completion.
What about looking at the wasps’ behavior from a colony perspective? Some
earlier observers, noting how infected workers leave their nest to aggregate, sug-
gested that this nest desertion could be interpreted as an ‘altruistic act’ (see Chapter
10) because it reduces the infection of kin. However, only wasps carrying Xenos
females with triungulins are vectors for the next generation; these have never been
found in these aggregations.
4.2.5 Theft and Kleptoparasitism

When the fly-hunting solitary wasp, Stictia heros, returns to her burrow with prey,
another female Stictia often knocks her to the ground and retrieves the dropped
fly for herself. Among honeybees, the large honey store is a magnet to forag-
ing conspecifics; weakly defended colonies are highly susceptible to raids. Many
dung beetles, known as rollers, are noted for collecting feces into spheres that
are used as a food source or placed in brood chambers; these dung balls must be
rapidly moved away from their source, because other dung beetles will abscond with
them.
Individuals of the same species usually feed on the same type of food in much
the same way, and even the diets of other closely related species may not differ by
much. This may not cause much problem in situations where food is abundant and
evenly distributed. (Why put out the energy to fight when there is nothing additional
to be gained by winning?) However, animals that live on patchily distributed foods
can experience considerable competition and even overt aggression. When compe-
tition is intraspecific, it sometimes leads to the establishment of social hierarchies
and/or territoriality, systems that reduce the expression of aggression between group
members.
When competition is between members of different species, extreme conflicts are
expected. These conflicts exemplify the well-known ecological competitive exclu-
sion principle: no two species can coexist in the same locale and utilize the same
resources at the same time. The less successful competitor will either be driven to
extinction or be forced to modify its feeding pattern. Modification is probably the
more common outcome. An insect species may switch to a slightly different type of
food, feed at a different time of day or night, or forage in a slightly different area.
As a result, such competition promotes evolutionary divergence in food choice and
feeding patterns.
Cuckoo bees and cuckoo wasps deposit their eggs in the nests of other bees and
wasps. These insects are normally referred to as kleptoparasites, sometimes spelled
cleptoparasites. (Their counterparts the cuckoo birds are a special case of kleptopar-
asitism known as ‘brood parasites’ because the immature parasite is fed directly by
the adult of the host and raised as the host’s offspring; this situation often occurs in
birds but is almost unknown among insects.)
In the broadest sense, both parasites and kleptoparasites can be thought of as
predatory thieves. Parasitism is a type of thievery in which one organism attacks
another, stealing its energy stores and sometimes its life. In contrast, a kleptoparasite
attacks the host’s young not primarily as food but rather to free up the host’s food
store for its own offspring. In many cases kleptoparasitism probably began with
simple prey stealing, such as the brigandage reported often among solitary wasps.
Many kleptoparasites are closely related to their hosts and appear to have shared
a recent common ancestry. Several species of a spider wasp, Ceropales, enter the
nest burrow of close relatives after the rightful owner has prepared and stocked the
hole with paralyzed spiders and laid an egg; inside the burrow, they lay a cleverly
concealed egg of their own upon the stored prey (Fig. 4.8). Some Ceropales species
substitute their own egg after eating the host egg themselves; others leave the host
egg for their own young to eat.
Theft can be a great strategy when it works. Small flies that look superficially
like house flies (see Fig. 1.14) are among the most abundant parasites of solitary
wasps. Each of the several genera involved has its particular method of attack, but
all of them produce eggs that hatch just before they are laid. These young maggots
attack and destroy the wasp’s egg or larva, but the maggots develop primarily upon
Fig. 4.8 Hiding the baby. The kleptoparasitic Australian spider wasp Ceropales ligea (left) hides
its eggs in the book lungs of the spider prey used by its host, the wasp Elaphrosyron socius (right).
Oviposition occurs very rapidly when the wasp briefly drops the spider in order to clear the entrance
to her nest burrow
the prey in the cell. Being ‘food thieves,’ the flies are not at all host-specific and
will attack species in several solitary wasp families indiscriminately if they are in
the proper ecological location.
There is no free ride, however, and anticipated victims often fight against such
theft. Many kleptoparasites have become armored in various ways, apparently as an
adaptation for deflecting the bites and stings of their intended hosts. For example,
the exoskeleton of cuckoo wasps (Chrysididae) is very hard and coarsely punctured
(see Plate 7); when disturbed or threatened, these wasps curl into a small tight ball
that is nearly impenetrable and difficult to grasp.
4.2.6 Insect Agriculture
Leaf-cutting attine ants of the New World have entered into partnership with various
species of fungi that do not occur outside of their nests. A nest-founding queen has
a blind pouch off her pharynx in which to carry a small fungal mass to inoculate the
garden in her newly established nest. The fungus will serve as the colony’s primary
and probably sole food source; these ants are unable to feed directly upon cellulose.
Under the ants’ care the fungus flourishes in the underground nest chambers, but if
ants are denied access to this garden, it rapidly deteriorates. The leaf cutters file out
along well-defined trails to cut leaves, flowers, and stems to be transported to the
nest, where each will be cleaned and scraped, then chewed into a small pulpy mash
with salivary secretions. As the ants insert it into the fungus garden, they cover the
mash with a transplant of several tufts of fungal mycelium and one or more drops of
fecal ‘fertilizer.’ This fecal material contains significant quantities of all 21 natural
amino acids, a nutrient supplement of considerable importance, as well as providing
the fungus with a proteolytic enzyme that breaks polypeptides in the leafy mass
into forms that are usable by the fungus. The ants have even evolved a means to
protect their food source from destruction, transporting and cultivating a crust of
Streptomyces bacteria that produce antibiotics against serious bacterial parasites of
the fungus.
This example is classic, but not alone—mutualistic feeding arrangements such
as this are relatively common. Termites, ants, and beetles appear to have become
fungiculturists by different evolutionary routes (Fig. 4.9). In termites it is believed
that fungi were an important food source before cultivation, and fungiculture arose
when the termites developed an ability to manipulate fungal growth in their nests.
In contrast, ancestral beetles may have begun as vectors for fungal spores, as
many non-fungus-feeding relatives of ambrosia beetles still are now. It is not
clear which of these evolutionary pathways might have given rise to fungiculture
in ants.
The parallels with farming have intrigued observers for centuries. Other exam-
ples involve plants and ‘livestock’ such as aphids. Cornfield ants (Lasius spp.) gather
eggs of the corn root aphid, Anuraphis, in the fall and take them home for the winter.
All winter the ants solicitously move the eggs about the nest, keeping them at appro-
priate temperature and humidity. When the aphid eggs hatch in the spring, the ants
place the young aphids near weed roots to feed. Later in the season, these solicitous
caretakers move their aphid herd to corn roots, dispersing them around the cornfield
where their feeding can cause considerable damage.
Many insects have a symbiotic relationship with microorganisms. Sometimes,
as in this leaf-cutting ant example, the microsymbionts live outside the insect’s
body, though they may be temporarily stored in special organs of ectodermal origin
for purposes of dissemination. A special type of ectosymbiotic relationship occurs
among some of the wood-inhabiting insects—the ambrosia beetles, wood wasps
of the families Siricidae and Xiphydriidae, and some bark-feeding bark beetles.
Sometimes, the fungus is eaten together with wood particles; in the most impressive
cases the insects feed upon their fungus alone.
Ambrosia beetles generally have an extremely wide host range but tend to occur
as secondary insects on diseased trees or felled logs. Each species is symbiotically
associated with one or more specific fungi indispensable for the development of
its brood. Identification of fungal symbionts is a difficult matter because they do
not produce fruiting bodies; however, it appears that the true ambrosia fungi are
highly specialized forms that cannot grow in the host plant in the absence of the
symbiotic insect. In return, the fungus provides a rich and available food containing
important vitamins for the insect and its larvae. How is fungus transmitted from
an old tree or log to a new one? Early workers thought spores were simply carried
in the beetle’s gut or upon its integument. It now appears, however, that certain
specialized organs of variable location and structure but usually confined to one sex
are involved. They serve to protect the fungi from desiccation, provide secretions
Fig. 4.9 Farming fungus. Comparison of the patterns of evolutionary diversification in three
major groups of fungus-farming insects—termites, ants, and ambrosia beetles—and their cultivated
fungi food. In the left cladograms, farmer lineages are shown in black and non-farmer lineages in
gray; in the right cladograms, fungal cultivar lineages are shown in black and non-cultivated feral
fungi in gray. Photos depict typical examples of each group: (d) Macrotermes bellicosus fungal
garden grown on walls of termite fecal pellets; (e) Atta cephalotes queen resting on her incipient
fungal garden; (f) Trypodendron lineatum gallery with brood niches showing the black ambrosia
fungus lining the walls
necessary for germination and arthrospore formation, and ensure the mechanical
dissemination of the fungus on the tunnel walls.
Wood wasps (Siricidae) also attack weakened trees or freshly cut logs. However,
their symbiosis with wood-inhabiting fungi is entirely different from ambrosia cul-
tivation. Only the larvae bore in the wood, but larval development may last two or
three years. During their adult stage, wood wasps are on the wing. Female wasps
possess intersegmental pouches at the base of the ventral parts of their stingers;
during oviposition, symbionts are pushed out with the egg.
Whereas some insects live mutualistically with ectosymbionts, more often the
mutualism takes the form of an internal symbiosis, with the microorganism living
within one of several organs inside the insect’s body. Most of these endosymbionts
of insects are microscopic. Their discovery dates back to the nineteenth century, but
initially their existence as independent living beings was difficult to comprehend,
and they were variously described as sporozoa, yeasts, metabolic products, and yolk
spheres.
Partnering with endosymbionts has enabled insects to feed on a wide range of
nutritionally incomplete and/or indigestible matter. The first, and the most numer-
ous, category of endosymbiotic hosts includes insects that suck sap, which is rich
in carbohydrates but poor in protein—scale insects, leaf lice, and aleurodid and
psyllid ‘flies.’ Among the bugs (Hemiptera), predaceous forms possess no sym-
bionts, but those such as stink bugs (Pentatomidae) that have switched over to
plant juice for nourishment have symbionts. Blood-suckers are a second important
host category. All insects that suck vertebrate blood for the whole of their lives
have symbionts; this includes the bed bugs and their relatives. Vertebrate blood is
deficient in certain vitamins that the symbionts can synthesize. Mosquitoes are a
familiar exception; they need no symbionts because they suck blood only as adults
and have a bacteria-rich nourishment at their disposal during earlier developmen-
tal stages. A third group of insects with symbionts includes species that feed on
cellulose-rich substrates: most famously termites. Symbionts also occur among that
small group of insects whose whole life is spent on keratin-containing food, the
feather lice.
Endosymbionts pose an interesting problem for their hosts—they must be rigidly
retained, but they also must be passed on to offspring. Varied arrangements have
arisen for housing them; these are species specific in both origin and development.
Depending upon the location of these guests within their host, these include oral
uptake of symbionts by young brood, smearing of eggs with symbionts, or infection
of eggs or embryos before laying.
4.2.7 Nest Symbionts: Becoming a House Pet
The nests of social insects provide living quarters for a diversity of creatures besides
those building them. Like our own homes despite our efforts to keep them clean,
insect nests are besieged by everything from cockroaches to flies. In addition to
these uninvited guests, some social insects purposely keep guests of other species in
their nests, much as we keep various pets.
Depending on their hosts’ identity, such symbionts or symphiles are referred to
as termitophiles, mellittophiles, etc. A particularly impressive range of vertebrate
and invertebrate species can be associated with long-lived termite and ant nests,
in particular, including organisms as diverse as microorganisms, plants, insects,
amphibians, reptiles, and mammals. In the case of species that live in the nests of
others of their own kind, the term ‘inquiline’ is used (see Chapter 10).
Fig. 4.10 Begging like an ant. Larva of Atemeles beetle being fed by regurgitation from a worker
Formica ant. When the beetle larva secretes a chemical from a row of paired glands along the sides
of its abdomen, ants respond by intensively grooming the larva. In response, the beetle larva rears
up and attempts to make mouth-to-mouth contact with the ant
Nest guests are particularly frequent in the ants, where all manner of organ-
isms have specialized as myrmecophiles, literally, ‘ant lovers.’ Many bristletails
(Thysanura), for example, live with ants. Feasting on debris, food scraps, and even
ant corpses within their common dwelling, these bristletails reward the ants with a
‘tidy house’ in turn for being inadvertently fed by their ant hosts. When army ants
emigrate, bristletails may be seen running in the column or riding along upon ant
larvae and booty.
Beetles are also common in the nests of social insects. Sometimes, acting as
scavengers they appear to be ignored by the nest owners. In other cases, the ants not
only tolerate the interlopers but feed, groom, and rear their guests, as happens with
Atemeles rove beetles (Fig. 4.10; see Case Study 4.1).
In many cases, even when their intruders eat the host ants’ young, they are treated
cordially. For example, Formica sanguinea ants avidly seek out beetle larvae of
Lomechusa strumosa, even though the latter usually destroy the whole population
of the ant hill. William Morton Wheeler expressed the puzzle well when he stated:
Were we to behave in an analogous manner we should live in a truly Alice-in-Wonderland
society. We should delight in keeping porcupines, alligators, lobsters, etc., in our homes,
insist on their sitting down to the table with us and feed them so solicitously with spoon
victuals that our children would either perish of neglect or grow up as hopeless rhachitics.
Case Study 4.1: Myrmecophily in the Rove Beetle, Atemeles
How do nest symbionts gain such acceptance? Clues are apparent in the case
of one well-studied example.
Among the beetle family Staphylinidae a number of myrmecophilous
relationships have evolved. Some of these beetles only live along ant food-
gathering trails or at ant garbage dumps outside the nest. Others live within the
nest’s outer chambers, but a select number have penetrated all the way inside
the brood chambers. Of this last group, one well-known example is Atemeles
pubicollis, a European rove beetle that spends its larval stage within the nest of
the mound-making wood ants, Formica polyctena. In examining their behav-
ior in some detail, Bert Hölldobler unraveled much of the communicative
behavior between host and symbiont.
Watching the behavior of ants encountering Atemeles larvae, Hölldobler
noticed that brood-tending ants respond to Atemeles by grooming them
intensely. Might the ants be imbibing something from the larval cuticle?
Hölldobler ran experiments with radioactive tracers, and confirmed that some
sort of chemical signal was passing between the ants and the larvae. To deter-
mine its action, he variably coated large numbers of beetle larvae with shellac,
then placed them at the nest entrance. As long as at least one body segment
was left unpainted, the ants carried larvae into the nest and adopted them.
When the entire larva was covered, however, the ants either ignored it or
deposited it in the garbage dump. Hölldobler washed some beetle larvae with
acetone; ants dumped most of these ‘deodorized’ larvae as well. Then he made
filter paper dummies soaked with the acetone ‘bath water’ and left them at the
nest entrance; like real larvae, many were carried into the nest. Paper dum-
mies soaked with fresh acetone, however, were either ignored or carried to the
dump.
Hölldobler turned his attention to a second puzzling behavior. Within the
brood chambers, adult ants scurry about feeding their young. When an adult
ant’s antennae or mouthparts touch an Atemeles larva, the latter rears up and
attempts to contact the ant’s head. When the attempt is successful, the ant
regurgitates a droplet of food (see Fig. 4.10). It appeared that the beetle larvae
were imitating the behavior of ant larvae. Wondering how the beetle’s success
rate compared with that of the ant larvae, Hölldobler gave the ants food labeled
with radioactive sodium phosphate. In a mixed brood of beetle and ant larvae,
the beetles obtained significantly more than their proportionate share of food.
Apparently, they were more intense as beggars than were the ant larvae. As
if this were not enough, the predacious beetle larvae actually ate small ant
larvae, as well. How could the ant colony survive such intense competition
and predation? Observation provided the answer—the beetle larvae were not
only predacious but cannibalistic. Unable to distinguish their fellow larvae
from ant larvae, they cut down their own population ruthlessly, whereas ant
larvae did not. Thus, the brood chambers soon contained clusters of ant larvae
but only a few scattered beetle larvae.
Further studies revealed even more behavioral complexity. Atemeles bee-
tles, it turns out, have not one but two ant homes: a summer woodland
domicile with Formica and a winter grassland one with Myrmica. After the
beetle larvae have pupated and eclosed in their original home, they beg for
one final, ample food supply. Drumming rapidly with their antennae upon
an ant to attract its attention, they touch the ant’s mouthparts with their own
maxillae and forelegs; since this mechanical signal is quite similar to that used
by ants among themselves, the ant responds by regurgitating food. Now fully
fed, the adult staphylinids begin to migrate out of the nest. Guided primarily
by light and odor, the beetles move into open grasslands and find Myrmica
nests. When a staphylinid encounters one, it wanders around until it encoun-
ters an ant worker. Going through a brief ritual, the beetle moves in, carried
by its host right into the brood chamber. Although the beetles are now adult,
they are still sexually immature. Within these latter nests, they continue to be
fed until sexual maturity the following spring, at which time they return to
Formica nests to mate and lay eggs.
To penetrate Myrmica nests, Atemeles must adopt a new set of skills and
a second language. Hölldobler again suspected chemical cues, because as the
guest and host first encounter one another, the beetle antennates the ant lightly,
then raises its abdomen toward its host. In response, the ant licks the tip of the
beetle’s abdomen, seeming to grow calmer in the process, then moves on to
the side of the abdomen. Finally, the beetle lowers its abdomen. The ant then
grasps bristles around the beetle’s sides and carries its tightly curled guest
inside. True to prediction, Hölldobler found two types of glands and secretions
(see Fig. 4.11). At the tip of the abdomen, ‘appeasement glands’ produce a
partially proteinaceous secretion that apparently suppresses aggressive behav-
ior in the ant. Along the sides, a series of ‘adoption glands’ produces a
chemical necessary if the ants are to welcome the beetle in. Apparently, this
odor mimics the odor of members of the ant species.
Chemistry plays a large role in the acceptance of adult Atemeles (see Case
Study 4.1), but this is not unusual. For example, ‘appeasement glands’ (Fig. 4.11)
turn out to be fairly widespread among the better-integrated myrmecophiles includ-
ing many of the ant nest beetles and lycaenid butterfly caterpillars. They are of a
number of novel forms (including glandular hairs) and exist in different locations.
Many types of glands aid in mediating myrmecophily, and appeasement devices are
just one of a syndrome of changes in morphology and behavior that social symbionts
belonging to various insect groups have independently undergone. Three are worth
special mention.
Fig. 4.11 Getting adopted. Adult Atemeles beetles’ entry into nests of Myrmica ants is largely
chemically mediated. Encountering a potential host and worker, the beetle presents the tip of its
abdomen and taps the ant lightly with its antennae (1). The ant responds by licking the abdomen,
ingesting an ‘appeasement’ chemical from the ‘appeasement gland’ that apparently serves to sup-
press the ant’s normally aggressive behavior toward intruders (2). Next, the ant licks the sides of
the beetle’s abdomen where it obtains an ‘adoption’ secretion (3, 4) that releases its brood-carrying
behavior. It picks up the tightly curled beetle and carries it into the nest brood chamber (5) where
the beetle now has access to the ant’s brood
First, many symphiles look very much like their hosts. Commensal staphylinids,
in particular, strikingly resemble ants, with a slender body form, antlike ‘petiole,’
and even body sculpturing and color (Fig. 4.12); this antlike appearance is found
almost nowhere else in this large family of beetles. Early investigators working from
Fig. 4.12 The habits of ant nest beetles (Carabidae: Paussinae) range from free-living and preda-
tory to obligatory myrmecophily, and these differences are reflected in their morphology. All these
beetles are displayed here with the right legs and antennae at their broadest view and the left
extremities at their narrowest view. Scale bar, 1 mm
museum specimens believed that the antlike body form of the symphiles was tactile
mimicry to deceive their host ants and that color was visual mimicry to deceive birds
and other predators that attack marching ant columns. However, for ants chemical
identification is paramount; if an ant’s surface odor is disturbed even slightly, other
workers will immediately attack it, even though its morphology has not changed.
Therefore, it seems unlikely that such mimicry is tactile and directed toward ants.
Current thought is that predators watching ant columns for edible morsels have been
the selection agents for both the color and morphological mimicry.
A second adaptation characteristic of many unrelated symphiles is physogastry,
a condition in which the abdomen is greatly enlarged, particularly in its membra-
nous parts. Physogastry appears to be one of the major ways in which termitophiles,
particularly among Coleoptera and Diptera, mimic their hosts. To human eyes,
such swelling often makes them superficially resemble termites, but the question
of whether physogastry is a primarily visual or primarily chemical adaptation is dif-
ficult to answer because direct study of termites and termitophiles poses problems.
Termitophiles are extremely delicate, and it is usually necessary to crack open the
termite nest to reach them and then make one’s observations under quite artificial
laboratory conditions. In at least some cases, there appears to be a mutual exchange
of exudates between termites and termitophiles, and histological studies of some
termitophiles demonstrate the existence of various abdominal glands and pores.
Whether termitophiles and their exudates are beneficial to termites is unknown, but
it is clear that termitophiles depend on their hosts. Most physogastric species have
such rudimentary mouthparts that they could no longer feed themselves and such a
heavy dependence on the controlled temperature and humidity of the termite nest
that they do not live long if they are removed from it.
A third morphological adaptation of many termitophiles and myrmecophiles is
a limuloid (teardrop-shaped) body. Generally, it is assumed that this body form
has arisen for defense because it would be difficult for an ant or termite to grasp
the smoothly streamlined dorsum. However, actual field observations of limuloid
symphiles are even scarcer than those of physogastric species.
Termites, ants and other social Hymenoptera all care for their larvae progres-
sively and exhibit fundamentally similar ways of communicating; no really great
differences exist in colony size or in length or stability of colony life. However,
specialized symphiles are almost unknown in nests of the social bees and wasps.
The few mites, beetles, and flies that live as scavengers and brood commensals are
quite generalized in form and behavior by comparison with symbionts of ants and
termites. Most are probably either attacked by their hosts or treated indifferently.
Why this difference? Probably the most plausible explanation centers around nest
structure, location, and feeding habits. Ants and termites live in relatively open
systems rich in refuse. Many chambers and galleries go unguarded from time to
time, and the nest interiors are generally made of material not too different from
the immediate environment—soil and rotting vegetable matter. Furthermore, young
are reared clustered in groups. Social wasps and bees, in contrast, ‘run a tight ship’
and construct compact (and often tightly sealed) nests, typically in arboreal loca-
tions. Thick envelopes of carton or wax pose formidable obstacles; nest entrances,
often narrow and tightly guarded, may also be lined with sticky substances and/or
repellants. Young are reared individually in specially constructed cells that make it
more difficult for a symbiont to conceal itself. Furthermore, rubbish inside is usu-
ally sparse; even when detritus occurs, workers simply heave most of it out the nest
entrance.
4.3 Coevolution and the Arms Race

Caterpillars of tropical heliconiine butterflies feed on particular species of toxic
Passiflora vines; some Passiflora produce tiny structures that mimic Heliconius
eggs and thus fool visiting butterflies into avoiding ‘further’ oviposition. Oak trees
produce high levels of tannins in their leaves that chelate nitrogenous compounds;
gypsy moth caterpillars maintain a gut pH so high that it inhibits chelation, allowing
them to feed on the leaves with impunity.
How did such systems come about? Paleoentomologists say the earliest insects
were small, wingless creatures that lived and fed chiefly in moist forest floor litter,
a mode of life still continued by the bristletails and springtails today. However, var-
ious insects also grazed on the nutritious spores from sporangia on the tips of plant
branches. Over evolutionary time, as plants became vascularized and grew to greater
heights, insects followed, as crawling, jumping, and finally flying forms arose. Seed
plants evolved; seed-eating insects followed. Hard seed coats developed; insects
found ways to tunnel into them. When the first flowering plants appeared early in
the Mesozoic Period, an almost limitless new food source became available. Insects
were quick to take advantage of it, and a rapid evolutionary radiation occurred.
When two or more populations, species, or genomes interact in this way, imping-
ing so closely upon each other that each acts as a strong selective force on
the other’s evolution and causes direct, genetically based changes, the recipro-
cal stepwise adjustments that occur are called coadaptation or, more commonly,
coevolution. These interactions cover a broad range of scenarios—including compe-
tition between species, mutualism, mimicry, or such ‘consumer/victim’ exchanges
as predator/prey, herbivore/plant, or parasite/host relationships. Thus, even though
it can be difficult to prove that true coevolution has occurred in a particular case
under observation, the concept has broad explanatory power and intellectual appeal.
Detailed coverage of coevolution is beyond the scope of this book, but a brief intro-
duction to the topic is justified because insect feeding has figured so prominently in
coevolution research.
Evidence such as Australia’s dramatic control of adventive prickly pear cactus
by an introduced cactus-feeding moth demonstrates that insect herbivory at least has
the potential to act as a powerful selective agent on plants. Likewise, outbreaks of
agricultural and forest pests emphasize the selective potential of insect feeding. That
the challenge of herbivores should be regarded as a major selective pressure in plant
evolution was first explicitly emphasized in the mid-1960s by Paul Ehrlich and Peter
Raven. In the years since, the plant/herbivore aspect of coevolution has emerged as
a major interdisciplinary focus. Plant species have been likened to islands defended
by chemical, mechanical, and biological barriers, penetration of which may allow an
insect species to exploit a new ‘adaptive zone.’ Presumably, the metabolic costs of
maintaining these barriers and the depredations of those species that have overcome
and even exploited them are more than offset by the general degree of protection
these defenses provide against other potential herbivores and pathogens. (One’s
meat is another’s poison.)
4.3 Coevolution and the Arms Race 165
According to coevolutionary theory, a succession of defenses against herbivores

is produced in a stepwise fashion by a particular plant species. At each step, some
herbivores go extinct, while others circumvent the botanical innovations, forcing the
plant to evolve further deterrents. Through time, the process may be regarded as a
series of filters or doors through which the plant and remaining successful herbivore
taxa are alternately released for flurries of adaptive radiation. This process is often
called ‘escape and radiate’ coevolution.
Cospeciation as a result of reciprocal interactions between two interacting
species, a process called pairwise coevolution, has received a great deal of research
attention. However, in coevolution as in war, battles are often waged simultaneously
on several fronts. Diffuse or guild speciation occurs when the genetic change in at
least one species affects its interaction with two or more other species. For example,
early season attack by flea beetles makes sumac plants more susceptible to stem-
boring cerambycid beetles; developing resistance to the former would also reduce
the impact of the latter. On the other hand, some cucumber genotypes have evolved
high levels of the chemical cucurbitacin. This chemical helps them resist attacks by
mites, but concurrently the plants become more attractive to cucumber beetles.
4.3.1 Attack, Defense, and Counterattack
Potato plants are members of the Solanaceae, a family of plants whose species
are often rich in alkaloids with effects that range from mildly irritating to fatal in
small quantities. Yet consider the common potato beetle, Leptinotarsa decemlin-
eata. Although the mother beetle actively chooses a site for egg laying, the larvae do
not usually begin feeding near the place of hatching. Instead, they begin searching,
guided mainly by vision, until they encounter a potato plant; at this point, they begin
randomly biting the substrate over which they are walking. Close-range olfactory
stimuli from the plant are attractive to the larvae. If the proper smells are present,
they clip off any hairs on the plant surface and pierce the epidermis. Tasting the
potato plant, they finally decide whether to reject it or begin feeding.
Despite a common misconception, coevolution does not necessarily enhance
mutual harmony or promote stable coexistence of species. Obtaining, maintaining
and protecting an energy pool is the major preoccupation of every individual in
any species. Sharing this energy with organisms of other species is evolutionarily
maladaptive unless one gains, or expects to gain, at least indirect benefit in return.
Therefore, an insect should protect its own energy pool against other would-be con-
sumers and at the same time try to break through the defenses of other species,
particularly those at a lower trophic level, to take advantage of a lucrative poten-
tial energy source. For each species, this results in a more or less continual battle
of attack and defense or counterattack, waged over the ultimate trophy of enough
energy for successful reproduction.
Plants likewise seek to protect themselves, often by chemical means, but many
insects turn this around and use it to their own advantage. Odor normally mediates
the first steps in herbivory, determining whether the insect detects the plant as a
potential host for sampling and whether it takes the first bite. Removing or block-
ing olfactory organs in a variety of insects leads to acceptance of non-host plants
otherwise treated as inedible.
Considerable experimental and observational evidence confirms that a particu-
lar subset of plant chemicals are utilized by insects as distinctive cues to identify
and discriminate among their host plants for feeding and breeding. They include
alkaloids, terpenoids, essential oils, and quinones. Terpenoids are particularly
widely represented. The lower molecular weight representatives include highly
volatile compounds that function as floral scents that attract pollinators. Those with
higher molecular weights include plant resins, cardiac glycosides, and saponines
that are typically distasteful or toxic to herbivores.
Because these chemicals, which are of almost universal occurrence in plants,
included a diverse array that were not known to have any function in plant growth
or metabolism, early researchers called them ‘secondary substances’. One view
held that they were either metabolic by-products (waste products) or metabolic pre-
cursors needed for some as yet unknown physiological functions. Another school
contended that such metabolically expensive ‘curiosities,’ which are often produced
in large quantities, have a primary significance to the plant as defensive substances,
and to label them as ‘secondary’ was in a sense an ironic injustice.
From the 1950s on, evidence has rapidly accumulated in favor of the latter inter-
pretation. Some authorities prefer the term ‘plant natural products’ for this reason.
However, the term ‘natural product’ also has come to mean a chemical compound
or substance produced in nature that has a pharmacological or biological activity of
potential use in pharmaceutical drug discovery and drug design. Many of the sec-
ondary substances do in fact fit this definition, but others do not (at least yet). At
present, secondary plant metabolites (SPMs) seems to be the preferred name for
these chemicals. The term ‘secondary’ distinguishes them from ‘primary’ metabo-
lites, the removal of which results in immediate death to the organism. Sometimes,
elimination of a secondary plant metabolite may have no apparent effect at all, but
more often closer study reveals that removal can impair long-term survivability or
fecundity.
Whatever one chooses to call them, to an insect these chemicals function widely
as signals of edibility, either by their presence or by their absence. Even though they
may differ widely in other respects, the food plants of a particular insect species,
genus, or even family often share similar secondary plant metabolites. The con-
centration of these chemicals often differs in various parts of the plant in ways
that seem to correlate with the need to protect them from attack. For example,
the natural insecticide pyrethrin is concentrated in flower heads, which seems log-
ical because protecting future seeds would seem more important to a plant than
protecting individual leaves.
Some of the best SPM examples occur in the cabbage family Cruciferae, which
includes many garden vegetables from turnips and radishes to broccoli and kale.
The pungent smell and tastes of cruciferous plants are due to mustard oil glucosides,
particularly sinigrin, that cause most insects to consider them inedible. However, for
some species, the reverse is true; experimental studies show that sinigrin acts as a
feeding stimulant for the turnip aphid and the cabbage aphid, and these species con-
sider plants inedible that lack the chemical. (Interestingly, inedible does not always
equate to non-nutritious. When leaves of 10 non-host plant species were treated
with sinigrin, aphids readily fed upon them all; five of the plant species turned out
to be nutritionally adequate to support the growth and development of at least one
generation of aphids.)
Producing strong chemicals is a knife that cuts two ways, however. Plant defen-
sive compounds can substantially reduce herbivore feeding, but these compounds
may be energetically expensive to produce, and their production competes with the
production of other necessary compounds and tissues. Moreover, insects also iden-
tify plants by their chemical fingerprints, and many have come to use plant volatiles
for their own purposes. One way in which plants can balance the tradeoff is to pro-
duce two groups of chemical defenses. One is a chemical fingerprint made up of
constitutive defenses that are always present; it tends to consist primarily of rel-
atively unspecific but generally effective compounds. If and when actual damage
occurs, however, these general compounds are joined by inducible defenses that
are more specific to particular types of injury. Bark beetles on pine trees provide a
classic example. During the initial beetle assault, the trees produce pitch, a gummy
terpenoid-containing mix that prevents bark penetration and seals wounds. However,
if the beetles successfully break through the bark, the tree produces more complex
phenols that can kill the beetle and cause lesions to form in the tissues surrounding
the wound.
Other herbivores steal the plant’s chemical defenses for protection against their
own enemies. The relatively inert exoskeleton provides an ideal site for passive stor-
age of toxic compounds. A famous example is the monarch butterfly whose larvae
sequester vertebrate heart poisons from their milkweed host plant; later, stored in
the butterflies’ wings, the chemicals confer protection from many birds. Some go
well beyond passive defense. Some sawfly larvae that feed upon foliage of conif-
erous trees store the resinous defensive chemicals in diverticular pouches in their
foreguts and regurgitate the fluid to repel predators. Likewise, some coccinellid
beetles store plant toxins derived from their prey in their hemolymph, and can
bleed the toxic fluid reflexively from their leg joints. We’ll return to this subject in
Chapters 5 and 6.
Still other insects use secondary plant metabolites for sexual communication.
Some male nymphalid butterflies congregate to feed at the dead shoots of plants
containing dehydropyrrolizine alkaloids and then biochemically modify the ingested
alkaloids to produce chemicals used during courtship. Many moths similarly have
an intricate relationship with plant chemistry. Likewise, virgin female polyphemus
moths will not begin their sexual call unless stimulated by trans-2-hexenal which
emanates from leaves of their host plant, oak. (In other tree species, the activity of
this chemical is apparently masked by other odors.)
Insect responses to plant chemicals take many forms (Table 4.1), but generally
can be divided into two categories: immediate and delayed reactions. So far, we
have concentrated on immediate responses, which are basically behavioral. Now,
Table 4.1 Some ways that plant-produced chemicals may affect plant-eating insects
Chemical actions Insect’s response
Repellents Oriented movements away from source

Attractants Oriented movements toward the source
Arrestants Aggregation at source
Stimulants Enhanced feeding, mating, or oviposition
Deterrents, suppressants Inhibition of feeding, mating, or oviposition after
attraction to source
Phytohormones Developmental anomalies; failure to molt or
metamorphose properly
Growth regulators Changes in growth and/or development
Toxins Sickness or sometimes death
Nutrients Reduced fecundity and longevity
Sterilants Incapacitated reproductive system (none presently
reported)
let’s briefly consider delayed reactions, which are largely physiological and include
developmental anomalies, toxicity effects, and hormonal changes.
Insect development and reproduction are governed primarily by two hormones—
molting hormone (ecdysone) and juvenile hormone. A large number of plants
produce their own version of these hormones. Some of these analogues are
extremely potent; plant-produced ecdysones have been identified that are as much as
20 times more active than the ecdysones produced by the insects themselves. These
hormone mimics are highly disruptive to insect development, typically preventing
maturation or producing imperfect and sterile adults. They can function as potent
deterrents and anti-feeding agents, sometimes even at dosages strikingly lower than
expected. (As might be expected, however, some insects have evolved enzymes
capable of detoxifying such compounds.)
Research is finding that many ferns and gymnosperms produce ecdysone ana-
logues. There is a common myth that ginkgo trees and ferns are free from insect
attack; as survivors of relatively ancient plant groups, they are said to have outlived
all their enemies. As with all myths, there is a kernel of truth. Chemical analysis of
ginkgo leaves has confirmed that the leaves are highly acidic and inhibitory to insect
feeding. The ‘bullet-proof’ reputation of ferns has not held up as well. A number of
common and widespread ferns are almost insect-free because they produce high
phytoecdysone concentrations, but other fern species quite commonly suffer attacks
from insects that range from bark beetles to Lepidoptera larvae (Fig. 4.13).
Certain plants also have mimicked that other major insect hormone class, the
juvenile hormones. A well-known example is the so-called ‘paper factor’ story.
The setting for this episode was the Harvard University laboratory of Carol M.
Williams in 1964, when Karl Slàma came from Czechoslovakia to spend a year
and brought along his favorite laboratory insect, the native European fire bug,
Pyrrhocoris apterus. Very soon it became clear that cultures of Pyrrhocoris were
not faring well in William’s laboratory. Instead of metamorphosing into adults at the
end of the fifth instar, the fire bugs molted to an extra larval stage to form giant sixth
Fig. 4.13 Feeding on ferns. Christmas fern frond, knotted (corner insert) at the tip as a result of
feeding by a Herpetogramma aeglealis (Pyralidae) moth larva (arrow). The larva notches the stem
so that it coils, but leaves the vascular bundles of the rhachis intact. No one yet knows whether this
is an adaptation to avoid plant-produced ecdysones
instar nymphs, which ultimately died without becoming sexually mature. Every pre-
caution had been taken to eliminate all sources of hormone contamination from
the culture containers, yet the problem seemed to be clearly of a hormonal nature.
Baffled, Slàma and Williams undertook a systematic analysis of every item in the
bug’s culture environment, searching for evidence of suspected juvenile hormone.
By a process of elimination the culprit was at last found—the paper toweling used to
line the cage floors. When the researchers substituted the filter paper that Sláma had
always used in Prague, the bugs again developed normally. Further detective work
traced the origin of the hormone to paper pulp derived from the Balsam fir (Abies
balsamea), a principal pulp tree indigenous to North America. This bit of serendip-
ity was the first indication of the existence of juvenile hormonal materials occurring
naturally in plants. As with ecdysone, juvenile hormone analogs have since turned
up in a diverse array of plants.
Whereas ferns and gymnosperms reply primarily on insect hormone analogues
and terpenoids for defense, angiosperms more commonly produce alkaloids, phe-
nols, and many other compounds. In fact, just as it is often said that ‘no two
snowflakes are alike,’ many scientists postulate that each plant species has a unique
‘chemical fingerprint’ conferred by its own mix of secondary plant metabolites.
Plants also produce analogues of essential amino acids; over 300 of these are
known, primarily from legume seeds. These are potent defenses against insect feed-
ing, because proteins constructed with these analogues are toxic, interfering with
many physiological processes. A particularly potent group of these analogues are
the alkaloids, which have been identified from about 20 percent of higher plant
families. Used with extreme medical care, many of these (such as atropine, bel-
ladonna, digitalis, strychnine) are familiar pharmaceuticals, but even at relatively
low concentrations they are highly toxic and teratogenic.
Not all plant defenses are chemical, of course. An arsenal of spines, thorns,
pubescences, and tough cuticles characterizes many plants from cactus to acacias.
Such structural features of plant surfaces have long been assumed to confer a cer-
tain measure of resistance to herbivore attack, and their prevalence seems to bear
testimony to the intensity of herbivore pressure. (However, some may have arisen
in response to other selective pressures; for example, the waxy texture of succulent
leaves undoubtedly helps to prevent desiccation.)
Physical barriers can be very effective. Female cereal leaf beetles lay signifi-
cantly more eggs on smooth-leaved strains of wheat than on strains with leaves
covered in dense, fine hairs; furthermore, larval survival is much lower on densely
pubescent plants. Various morphological traits of the bean family (Leguminosae)—
including such devices as flaky-surfaced pods that scale off eggs, gummy sap, and
pods that explode when penetrated—lower the success of bruchid weevils attempt-
ing to oviposit in or on the seed pods. The foliage of many species of Neotropical
passion flower vines is eaten by larvae of flashy heliconiine butterflies. However,
when vines of one species, Passiflora adenopoda, were exposed to Heliconius
attack in the laboratory, no feeding damage occurred. In fact, the larvae that had
been placed on the plants were found dead and desiccated by the next day, even
though other Passiflora species in the laboratory were heavily eaten. The cause of
the immunity of P. adenopoda was traced to the cloak of hooked trichomes (hairs)
that cover the plant’s surface. Scanning electron micrographs verified their defensive
function—the trichomes made numerous puncture wounds in the larval integument,
immobilizing the larvae and causing them to starve to death.
When they occur together, mechanical deterrents and chemical ones provide
considerable plant defense against feeding, but as always, some insects have
turned the tables. Camphorweed (Heterotheca psammophila), a common road-
side plant of the American southwest, has leaves, stems and buds covered with
tiny glandular hairs which produce droplets of aromatic resin that effectively
deter most insect feeding. However, the female of a reduviid bug Apiomerus
spends hours harvesting these droplets, which she then uses to fasten her eggs
into a resin-covered cluster. The resin, a complex mixture of terpenes, deters ants
and other egg-eating insects and may protect the eggs from microbial pathogens
as well.
4.3.2 Employing Mercenaries for Protection
One of the stranger types of defense employed by plants against herbivores uses
other organisms as a front guard or standing army. These mercenaries are paid, in
turn, through benefits they receive from the plant. Mutualistic relationships between
organisms are widespread. Some of the best illustrations may be found among the
ants, many of which have become specifically adapted to live upon certain plants.
One of the best known is that between ants and a diverse assortment of sap-sucking
plant bugs that excrete sugary drips of semi-processed phloem sap called ‘honey-
dew’ as they feed. Various ants exploit the concentrated food bonanza in such a
systematic fashion that some early naturalists likened it to cattle farming. In return
for the aphids’ honeydew—produced not only passively during feeding but also
on demand in response to stroking by the ant—the aphids receive protection from
their predators and other services (see Plate 10). These ant-aphid protection systems
are generally very well integrated. For example, when ants are exposed to artifi-
cial sources of the green peach aphid’s alarm odor, they respond aggressively. Ants
guarding Prociphilus will even readily attack a human finger that approaches their
wooly aphid flock.
In a very real sense, plant-ants may be regarded as analogous to the secondary
compounds found in the foliage of most plants, the primary defense of their plant
against herbivores, as was illustrated by Daniel Janzen’s studies of acacia bushes
and the ants that guard them (see Plate 8 and Case Study 4.2). Many other ant–
plant relationships have been noted both before and since Janzen’s work. Each will
require the same sort of detailed investigation in order to understand the dynamics
and evolutionary adaptations on the part of the participants before they can be truly
termed mutualisms. One of these is the interaction between Cecropia plants and
certain species of Azteca ants throughout Central America; like the acacia, Cecropia
produces glycogen-rich food bodies (trichilia or Müllerian bodies) that are harvested
by its ants. Another is a ‘parasite’, Pseudomyrmex nigropilosa, that Janzen later
discovered; this ant harvests the resources of the swollen-thorn acacias but it does
not protect the acacia, attack foreign objects, or clean debris from the foliage.
Case Study 4.2: Coevolution of Acacias and Their Ants,

Pseudomyrmex
The mutualism between plant ants and ant plants is highly coevolved and costs
each contributor a substantial energy outlay, but each participant also derives
considerable benefits.
In disturbed areas in the lowlands of Mexico and Central America, a com-
mon shrubby tree is the bull’s horn acacia, Acacia cornigera, so named for the
pair of swollen, hornlike thorns that occur at the base of most of its leaves. A
rapidly growing woody plant that cannot tolerate shading, it quickly springs
forth as sucker growth from old rootstocks in pastures, along roadsides and
in natural disturbance sites such as riverbanks and arroyos. Close examina-
tion of the plants reveals that they crawl with small Pseudomyrmex ferruginea
ants, which pour forth from holes in the thorns whenever the acacia is touched
(see Plate 8). In the 1800s the naturalist Thomas Belt had discovered a similar
relationship between an ant and acacia in Nicaragua and noted that the ends
of each leaf segment were modified into peculiar little oval structures. These
‘Beltian bodies’ seldom survived long, because the ants cut them off and either
ate them or fed them to their young. Subsequent studies have shown that they
have an unusually high food value for foliar tissue, on the order of yeast in
quantity and quality of nutrients.
Following Belt’s discovery there was a great deal of armchair speculation
on the exact nature of the relationship between ant acacias and acacia ants. The
ants clearly depended upon the acacias because most acacia ant species had
been recorded only from living ant acacias. But did the acacias benefit from
the presence of the ants? Two viewpoints persisted: one stated that the ants
were merely exploiting the acacia, while the other regarded the relationship as
a true symbiosis. This question became the focus of intensive eastern Mexican
field studies by Daniel Janzen.
One of Janzen’s early observations was that these acacias were commonly
left untouched in pastures with cows and other domestic animals. From a local
farmer, he borrowed a pet native deer that was thought not to have had any
previous experience with A. cornigera. When he offered it foliage that had
been cleaned of all ants, the deer ate the foliage readily, including some of
the thorns. After several days of feeding the deer unoccupied foliage, Janzen
offered a branch complete with some very agitated ant workers. As the deer
began to eat the foliage, ants ran onto its face and stung; immediately the deer
stopped feeding and withdrew to clean them off. Similar simple experiments
with cows and burros—animals that may have had previous experience with
A. cornigera—gave results that did not differ materially but suggested addi-
tionally that these herbivores may have learned to recognize the alarm odor of
the ants and learned to avoid contact with the plant.
The survival of the bull’s horn acacia is dependent upon rapid growth so
as to remain unshaded. Thus, Janzen began to examine the relative growth
and development of occupied versus unoccupied shoots of A. cornigera by
treating selected shoots with an insecticide. A number of striking differences
emerged. First, the frequency and extent of phytophagous insect damage was
greatly increased in shoots from which ants were removed; this resulted in
a great lowering of the growth rate of the shoots. In contrast, ant-occupied
shoots remained virtually free of phytophagous insects because the latter were
quickly attacked and removed from the shoot by worker ants. At least 40
species of insects fed on unoccupied shoots of A. cornigera, but only eight
species attacked occupied acacia shoots, and even the cumulative sum of their
feeding was not serious. During the rainy season, phytophagous insect activity
increased greatly, and the differential growth rates of occupied versus unoccu-
pied shoots became even more striking. Unoccupied shoots showed almost no
growth, but occupied shoots grew vigorously. Because of their slow growth
rate, surrounding vegetation quickly shaded the unoccupied acacias, and this
further slowed their growth. Many unoccupied shoots simply died.
Data from experimental plots also confirmed another observation that
Janzen had made, namely that occupied shoots were almost always free of
living vines even though the acacia plant appeared ideal for the support of
vine growth. Whenever foreign vegetation contacted the acacia plant, the
ants would maul it and chew off the growing tips. In this way they pre-
vented vines and lateral branches of neighboring plants from growing into
the canopy of occupied A. cornigera. Unoccupied acacias, in contrast, accu-
mulated heavy masses of vines during the rainy season. Vegetation-mauling
activity of worker ants also resulted in a bare circle on the ground around
the base of the shoot. Janzen felt that this behavior was of significance in
protecting the shoot from fires as well as in lowering the incidence of phy-
tophagous insects reaching the shoot. Even when the shoot was killed by fire,
suckers from protected rootstock in the bare basal circle quickly sprouted and
were immediately colonized by the mature ant colony, giving the new shoot
an immediate competitive advantage in the post-fire succession.
Based on these data and observation of occasional naturally unoccupied
acacias, Janzen concluded that a shoot of A. cornigera must be occupied by
a colony of P. ferruginea for a substantial part of its life to produce seeds
and become a part of the reproductive population. Because of the interdepen-
dence of the ant and acacia for normal population development, the interaction
between them could properly be called obligatory mutualism.
4.3.3 The Tommy Tucker Syndrome: Food in Return

for Services
‘Little Tommy Tucker sang for his supper . . .’ Mother Goose may have been
on to something. Many insects transport pollen and/or seeds in return for food
or other rewards. Violets, for example, reward ants for collecting their seeds
by incorporating lipid-rich compounds (elaiosomes) into them that are highly
attractive. In the previous section we discussed coevolution in the context of plant
defense. However, this is only one part of a much broader coevolutionary picture
that has been particularly well studied in the context of pollination.
Pollination in its most basic sense is simply the transfer of pollen from the male
sex organ (anther) to the receptive portion of the female sex organ (stigma) in flow-
ering plants. If the transfer is successful, it leads to fertilization and seed production,
and the plant reproduces. This process often involves some external vector, usually
wind, water, or animals. Authorities are in general agreement that the first insects to
become anthophilous, or pollen-loving, were the Coleoptera; even today, most sim-
ply rummage around in flowers, destroying some flowers by feeding on them but in
the process managing to pollinate others. Pollination apparently appeared next in the
Diptera, but fly pollination is understudied and probably also underrated. With the
development of pollination habits in Lepidoptera, highly mutualistic associations
began to become quite common, reaching their culmination in the Hymenoptera,
especially among the bees. Today, while many insect taxa visit flowers and thus can-
not be dismissed as pollination agents, members of these four largest insect orders
still are pollination biology’s major players.
Many flowering plants provide primary rewards—including nectar, pollen, lipid
secretions, food bodies, scents, resins, and nest building materials—that attract pol-
linating insects and keep them returning. In addition, most insect-pollinated flowers
also produce sensory cues—including odors, colors, shapes, textures, and tastes—
that distinguish their own flowers from other species and make it easier for an insect
to find more of their own flower type. For most flowering plants, this step is crucial
for pollination success because they are self-incompatible.
Have plants tailored these cues to attract certain pollinators and exclude oth-
ers? The concept remains controversial. Patterns certainly exist. A typical flower
pollinated by a night-flying hawk moth (see Plate 42) is white and has a strong
sweet fragrance; a typical flower pollinated by day-flying butterflies is red and
has very little odor. Particularly among tropical flowering plants, these signals
are often combined into patterns that some scientists have recognized as polli-
nation syndromes related to the type of pollen vector. However, it is also true
that rather than a single pollinator type, many flowers are visited by ‘guilds’ that
include diverse insect taxa, and the composition of these guilds may change over
time. Furthermore, just because an insect is an apparent guild member and visits a
Fig. 4.14 Pollinating figs. (left) Caprification in an orchard in Greece involves hanging bunches
of ’goat figs’ from tree branches. Pollination by the wasps that develop from these inedible fruits is
necessary to produce a mature edible fig crop. (right) Not all figs require fig wasps for pollination,
but among those that do, the fruits come in two types; compare the internal flower structure of the
‘goat figs’ at the left and the edible fig on the right. Fig wasps develop in the ovule of the short-
necked female flowers that pack the inside of goat figs. Before dispersing, newly emerged winged
females (middle bottom) mate with wingless males (middle top) and collect pollen from the male
flowers that cluster around the upper end of the goat fig. When the females find the new crop of
fruit, they enter the figs and thus pollinate them, but inside these edible figs the flower necks are too
long for their ovipositors to reach the flower ovaries. Thus, they cannot lay eggs or rear offspring
in the figs’ edible generation
flower does not always mean it is actually an effective pollinator of that plant. For
these reasons, some pollination biologists see little conceptual value in pollination
syndromes.
Because pollination is so closely tied to reproductive behavior, not just of the
plants but of the insects as well, this subject will come up again in Chapter 9. For
now, let us consider an example of the extreme forms that pollination mutualism
can take—the relationship between fig wasps (Agaonidae) and their fig plant hosts
(Fig. 4.14).
Although the Smyrna fig story has received the most attention (see Case Study
4.3), figs actually include about 850 species of woody trees, shrubs, vines, and
epiphytes, most of which are native throughout the tropics. The fruits of many
species are edible and serve as important food resources for wildlife. Most tropi-
cal Ficus species bear several crops of syconia through the year, with short- and
long-necked female flowers and male flowers in the same syconium. Because the
female flowers are receptive several months before the male flowers release pollen,
they require wasp pollination between synchronized syconia of the same or different
species.
Case Study 4.3: Mutualism Between Fig Trees

and the Wasp, Blastophaga
The most delicious figs require tiny symbiotic wasps to pollinate their minute
flowers. The relationship is so perfectly timed and intricate that if one partner
should die out, the other would inevitably follow.
Although European honey bees were valued co-travelers to the New World
for their honey production, one of the first insects introduced into North
America specifically to pollinate a crop came to southern California in 1899
to help with a problem—whereas common figs had been grown successfully
in the American Southeast and Mexico for two to three hundred years, the
choice cultivar called the Smyrna fig consistently failed to produce a crop.
The trees grew well, but immature figs simply dropped off rather than ripen-
ing. Why? Farmers were puzzled until they realized that the trees were missing
an essential pollinator that was not native to California. However, they could
not merely import the wasp, because the mutualism was very complex.
What we call a fig is actually a syconium, a specially adapted involuted
group of multiple flowers, rather like an almost-closed jug with flowers lin-
ing the inside. The flowers cannot be seen unless the fig is cut open; doing so
reveals the flowers as fleshy ‘threads’ that each bear a single seed inside. There
are many kinds of figs, but only one, Ficus carica, is cultivated on a commer-
cial scale. Its ancestors arose long ago in Asia Minor; some archeologists say
fig trees were cultivated before cereal grains were domesticated. Wild pop-
ulations of this tree include both male trees and female trees. Early authors
thought they were different species; a key to understanding this complex sys-
tem came with the realization that they are actually two natural sexual forms
of the same species, and that the two flower sexes develop asynchronously. A
second breakthrough was the realization that these Old World fig trees house
a tiny wasp, Blastophaga psenes, that is vital to their reproduction.
Male Old World fig trees actually bear both staminate and pistillate flowers,
but they produce fruits that contain only wasps and pollen. Generally not eaten
by humans, the fruits were named ‘caprifigs’ because they were commonly
fed to goats. Male trees produce three crops of inedible seedless figs per year
that maintain the pollinating wasps—profichi that produce abundant pollen
and ripen in early summer; mammoni that have short-style flowers and ripen
in the fall; and mamme with long-style flowers that overwinter on the tree
and ripen in spring. Within the profichi, female flowers develop first; their
shape is modified in a way that makes egg-laying easy for dispersing female
Blastophaga wasps that push their way through a tiny pore into them. Soon
afterwards, the pore closes, and the syconium becomes virtually impervious
to insect entry. In response to oviposition, the flowers swell like tiny galls,
each nourishing a developing wasp larva that feeds safely inside.
As the profichi reach maturity, so do the young male and female wasps
within. Wingless males chew holes in the sides of their nursery chambers,
escape, find females still imprisoned in adjoining chambers, chew into these,
and quickly mate. Pollen-bearing male flowers in the profichi are now also
mature; to leave the fig where they were born, females wasps must pass
through a ring of these male flowers surrounding the fig’s entrance, becoming
well dusted with pollen.
Female trees, by contrast, produce only one or two crops: a parthenocarpic
(seedless) breba crop that ripens in early summer and a main crop that ripens
from late summer to fall. Though the figs on female trees also have flow-
ers of both sexes, the male flower parts fail to develop. Breba figs are never
pollinated because they develop before male profichi ripen on nearby male
trees. Main crop figs may become pollinated or not, depending whether they
receive pollen via pollen-loaded B. psenes wasps from nearby mature profichi.
(Commercially cultivated fig tree types differ in their production of breba figs
and their need for pollination.)
As the female wasps emerge from caprifigs and begin to search for
new immature syconia in which to oviposit, they encounter an evolutionary
surprise—the fig trees are now producing not profichi, but edible figs that have
only large numbers of female flowers with long-necked styles. Ovipositing in
these is all but impossible, but in their repeated vain attempts to do so, the
female fig wasps successfully pollinate what will become edible figs. Some
fig wasps find still another type of small, inedible figs on the tree’s uppermost
branches. These ‘mamme’ or mother figs contain only short-styled female
4.4 Feeding as a Communal Activity 177
flowers in which the female wasps can oviposit successfully. Young male and
female wasp larvae hibernate within, to begin the profichi–fichi–mamme cycle
anew the following spring. The delicious Smyrna and Calimyrna figs are a
dead-end for the pollinating fig wasps, however.
As a result of this understanding of the fig/fig wasp mutualism (see
Fig. 4.14), Calimyrna figs are now grown successfully throughout California’s
hot San Joaquin Valley. Every June, visitors see a most unusual sight—
thousands of acres of fig orchards ‘decorated’ with small brown or white
paper bags stapled to the lower limbs. Inside each bag is a caprifig branch
with pollen-bearing profichi, picked prior to emergence of its associated
fig wasps. Although wasp-bearing caprifig trees have become naturalized in
moist riverbeds and creeks in California, they are kept at some distance from
orchards. Controlling the pollination process using only the bagged caprifigs
is important, for if too many wasps pollinate an edible fig, it splits open and
has no commercial value.
Typically, only one wasp species is capable of fertilizing the flowers of each
species of fig. Therefore, plantings of fig species outside their native range results
in effectively sterile individuals. It should be noted, however, that the common fig
cultivars grown in home gardens in southern and western North America require
no pollinator. A single dominant mutant gene is responsible for this change; rather
than prematurely dropping unpollinated fruits as the Smryna figs do, the common fig
trees expressing this gene retain unpollinated developing figs to maturity. The ability
to produce sweet fig fruits without the need for male trees that carry symbiotic fig
wasps within their syconia is very advantageous to farmers in regions where wild
caprifigs and natural pollinator wasps do not occur. However, these fig trees must be
propagated by cuttings because they do not produce seeds.
4.4 Feeding as a Communal Activity

Gregarious feeding is characteristic of some herbivores, such as tentworm caterpil-
lars and various sawflies (Fig. 4.15). Courtship feeding, another type of cooperative
social feeding between conspecific individuals (see Chapter 9), is also common. It
is among the truly social insects, however, that cooperative feeding behaviors are
exhibited in their widest scope and variety. In fact, certain social feeding behaviors
are more or less unique to the social insects.
4.4.1 Simple Groups and Feeding Aggregations

Larval sawflies, caterpillars, and chrysomelids all have converged on a lifestyle
that includes a syndrome of maternal care, larval gregariousness, or both, sug-
gesting to some evolutionary biologists that there might even be something about
Fig. 4.15 Gregarious feeding. ‘Escalator style’ feeding by larvae of the Brazilian sawfly, Themos
olfersii, whose host plant possesses extremely tough leaves. Aligning themselves in two convergent
rows, the young larvae use their enormous heads and jaws to gain access to the thick edges of the
leaves (a). As the larvae feed, the leading individuals are pushed forward until they collide (b).
Upon collision, the leading larvae retreat to the ends of the rows (c) where they cannot readily
feed, while others are moved forward for a turn at eating. This process continues for many hours
the nutritional and defensive ecology of exposed leaf feeding that selects for this
syndrome.
Gregariousness, the tendency to gather in groups, begins with laying eggs in
batches. Often, even the eggs of gregarious species are chemically defended; the
tip-off is their bright coloration. It might seem puzzling for vulnerable eggs to be
advertised, but advertising is the point. Such eggs are, in fact, packed with unpalat-
able or even toxic compounds, often in such high doses that a single egg is enough
to kill an ant.
All gregarious leaf-feeding species that have been studied are chemically
defended as well. There have been relatively few empirical tests of the efficacy of
these chemicals against an insect species’ known natural enemies, but some suggest
that predators are more likely to be deterred than permanently stopped. This may be
enough, however, to give prey the statistical advantage necessary to select for the
strategy.
Because of these chemical defenses and bright coloration, it is generally assumed
that almost all simple feeding groups and feeding aggregations function to improve
defensive capabilities. However, increased feeding efficacy may be equally or even
more common. One well-studied example involves the imported willow leaf beetle
(see Case Study 4.4).
Case Study 4.4: Group Feeding and Cannibalism in the

Willow Leaf Beetle, Plagiodera
Among some group-feeding larvae, the Dr. Jekyll of better defense and feeding
efficiency has a flip side, the Mr. Hyde of cannibalism. What balance of costs
and benefits maintains such patently antisocial behavior?
Across eastern North America each spring, females of the willow leaf bee-
tle, Plagiodera versicolora, (see Plate 12) emerge from overwintering sites
under loose bark and begin laying the first of their two or three clutches of
15–20 eggs. The larvae of this little metallic blue-green chrysomelid complete
their development in about two weeks and soon become reproductively active
themselves, so that even on a single tree, broods of different generations and
ages can be found as the summer progresses. The larvae stay in tight aggrega-
tions throughout the first two instars, then gradually become less gregarious as
they continue to grow. Like many larvae with a social childhood, Plagiodera
produce copious chemicals, in this case through paired dorsal glands. They
also often assume a ‘cycloalexis’ formation with their heads facing out, like
pioneers circling their wagon train.
Since 1985, Michael Wade and Felix Breden have been thoroughly
researching Plagiodera group behavior. Their early field studies yielded puz-
zling results; the first year, larval survivorship showed a significant positive
relationship to initial group size. The next year, they could only say that the
group size-survivorship relationship varied considerably in space and time.
What was going on? Breden and Wade designed a clever way to address
the question. First they located natural P. versicolora clusters in the field, and
randomly assigned them to group size and duration classes. Then they ‘down-
sized’ the aggregations to create replicate groups of various sizes that were
followed and collected at various intervals over a week’s time. When they
analyzed their data, once again survivorship was unaffected by initial group
size, but an interesting new relationship surfaced. Larvae gained significantly
more weight in larger groups. Thus, Breden and Wade concluded that for this
species, the benefit of group living related more to increased feeding effi-
ciency than to increased defense. Careful observation provided the reason. In
1994, Wade described feeding facilitation in this species. Each larva attempts
to break through the cuticle of a willow leaf by rocking from side to side as
it bears down with its mandibles. When one larva succeeds, others stop their
own rocking and move to the break to feed. Larger groups presumably allow
more and earlier breaches of the plant’s defense.
Gregariousness begins with laying eggs in batches. Imagine the surprise
of these observers upon finding that newly emerging Plagiodera larvae are
intensely cannibalistic for about 24 hours before they switch to become strict
vegetarians. At first glance such behavior looked very antisocial and selfish;
if it happened frequently enough, it seemed as though it could completely
destroy the group. There must be some serious trade-offs in costs and benefits
for this cannibalism to persist, they reasoned.
Feeding efficiency might be one benefit. At the end of 24 hours, cannibal-
istic individuals weighed a very noticeable 14% more than non-cannibalistic
larvae in the cluster. Presumably these larger larvae would have an increased
chance of early successful leaf penetration, and through their success, other
group members’ feeding would be facilitated, as well. However, the positive
effect of group size on larval survivorship was weakened in groups with can-
nibals simply because cannibalistic feeding has the obvious effect of reducing
group size.
At present, perhaps all that can be conclusively said is that for a non-
cannibalistic P. versicolora larva, being a group member is clearly advan-
tageous in terms of survival and growth, but that the presence of cannibals
in the group can completely undermine these advantages. Interestingly, for a
cannibal, the work of Wade and Breden indicates that fitness is not dependent
on group size or the percent of cannibals in the group. Rather, it depends on
the average larval fitness of everyone in the group, cannibal and non-cannibal
alike, plus a personal fitness gain that the cannibal experiences from dining
on its siblings.
Like water from a bottomless well, complex and fascinating new questions
about P. versicolor keep arising. For example, if average survivorship declines
in the presence of cannibals but the cannibals themselves do well, is this
an evolutionary equilibrium? To assess selection within and between groups,

Breden and Wade have performed a great many field-based experiments in
which they manipulated group size and proportion of resident cannibals. Is
there a genetic basis for cannibalism? Very possibly; without it, the behavior
cannot be subject to selection. Certainly, the species appears to be polymor-
phic for it; some 41–77% of the colonies exhibit cannibalistic individuals.
Are group members closely related? With colleagues Wade and Breden have
examined kinship structure, and found that it varies widely among localities
and years, due both to multiple mating and to colony merging. Is cannibalism
lower in groups consisting of closer kin? This needs further study; if so, it
would be consistent with kin selection (see Chapter 10). Do cannibals repro-
duce more than non-cannibals? This is another key question that needs an
answer, because if so, it could provide a strong group selection example (see
Chapter 10). Another question that has been raised concerns whether the can-
nibalism is a simply a carryover of an ancestral trait; the type of comparative
study of solitary relatives that would be needed to determine this has yet to be
done.
4.4.2 Social Feeding Behaviors
When a social wasp larva is fed by an adult, the larva almost always secretes a
droplet of salivary fluid, which the adult imbibes. In many social insects, chemi-
cal communication signals are spread through the colony by an exchange of liquids
between nest mates. The reciprocal exchange of liquid foods between colony mem-
bers, a behavior called trophallaxis, occurs both between adults and between adults
and larvae.
Trophallaxis is highly developed in social insects. It appears to occur generally
throughout the eusocial wasps. Among bees and ants, on the other hand, its occur-
rence is highly variable, apparently determined both by phylogenetic position and
ecological constraints. In termites, trophallaxis has multiple functions. Lower ter-
mites share both ‘stomodeal food’ from the salivary glands and crop and ‘proctodeal
food’ from the hindgut. The former is the principal nutrient source for the royal pair
and nymphs; the latter, a milky material quite different from feces, contains sym-
biotic flagellates. These break down the cellulose that the termites are otherwise
unable to use as a source of nutrition. With each molting these hindgut symbionts are
lost, and the nymphs must acquire new ones. Among the higher termites, the habit
of proctodeal trophallaxis has been lost along with dependence on symbiotic flagel-
lates for cellulose digestion. However, the nymphs have become entirely dependent
on stomodeal exchanges and no longer even possess functional mandibles.
Since 1918 when entomologist William Morton Wheeler first coined the term
for this unilateral or bilateral liquid food exchange, there has been strong dispute
over the signal value of these liquids. Some entomologists have even suggested
that the word trophallaxis should be used synonymously with communication

because it provides a mechanism for maintaining a colony-specific odor. However,
trophallaxis serves more functions than simply communication. The larvae of some
vespid wasps, for example, take over certain metabolic functions for the colony
through trophallaxis. Receiving proteinaceous food from the adults, they transform
it into carbohydrates that are stored and, if required, later returned to the adults.
Because the adults are apparently unable to synthesize these essential nutrients, such
behavior is critical to colony survival during periods of food shortage, such as rainy
periods.
Both prolonged direct observations and radioactive tracer studies have shown
that in the majority of social insects trophallaxis allows a material to be distributed
throughout a colony with striking rapidity (Fig. 4.16). Trophallaxis is an open
system—each individual shares with an unlimited number of nestmates. Although
trophallaxis requires both a giver and a receiver, which individual performs which
role largely depends upon the state of their crop contents at the moment, and roles
can easily be switched. For example, a set of Formica ant workers showing predom-
inantly begging behavior can be shifted to predominantly donor behavior simply by
feeding them to satiation. Because of this rapid exchange, it has been said that ants
have a ‘communal stomach’; each worker shares virtually the same diet and is kept
informed of the nutritional status of the colony as a whole. Therefore, when an indi-
vidual worker reacts to feeding stimuli in terms of its own hunger or satiation, in the
majority of cases it is also acting in a way that is appropriate for the colony as well.
What are the releasers for a trophallactic exchange? In the honey bee, they appear
to be the combination of tactile and olfactory cues provided by a honey bee head
with antennae intact. A freshly severed head will elicit either begging or offering; if
its antennae are removed, the head is less favored until insertion of imitation wire
antennae of the proper length and diameter restores its effectiveness. In addition,
Fig. 4.16 Sharing a ‘communal stomach.’ Formica fusca is an ant species that engages in rapid
oral trophallactic exchanges. Within a day after a single worker was fed small amount of honey
mixed with radioactive iodide, evidence of the radioactive food was present in varying amounts in
every colony member, including the 2 queens (stippled part). Over time, the frequency distribution
of individual shares of the radioactive food became progressively more normalized
both ants and bees appear to favor larger individuals, with the result that queens,
males, and larger workers tend to receive more than they give.
In social wasps, the release of trophallactic behavior is more structured and
complex, because dominance relations place severe constraints on food exchange,
slowing the rate of food distribution and increasing the variance in crop contents
among workers. For example, an inert severed head does not release trophallaxis in
Vespula; rather, the pair must engage in a very definite pattern of continuous recip-
rocal antennal signaling. Dominant workers receive more food than they give. The
mother queen always receives and seldom ever gives, and virgin queens dominate
over their worker sisters. Males, possessing a quite different antennal form, are quite
inept at begging and must rely primarily upon surreptitious sips of the regurgitated
liquids being passed between others or upon stimulating larvae to give forth salivary
secretions.
In the previous section, we considered cannibalism in the context of simple
feeding groups; now let’s revisit the subject in the context of eusocial insects. All
termites studied to date eat their own dead and injured on at least some occasions,
a degree of cannibalistic behavior that is far more intense than in any other social
insect group. Some Reticulitermes termite workers even eat apparently healthy nest-
mates when grooming is carried too far. If the cuticle of a leg is broken, for example,
the leg is eaten, and then the whole termite is consumed. Winged Coptotermes repro-
ductives that are unable to leave on a normal nuptial flight are finally killed and
eaten by workers. Alien conspecific workers that enter a termite nest are generally
disabled, then consumed. Because termite diets under natural conditions are usually
low in protein, it is generally believed that this cannibalism functions as a protein-
conserving device. Colonies of Zooptermopsis angusticollis, for example, become
intensely cannibalistic when reared on a laboratory diet of pure cellulose, but adding
sufficient casein to the diet reduces the cannibalism to almost zero.
Among the social Hymenoptera, cannibalism of adults is rare in some groups,
unknown in most. However, a related phenomenon is common—eating immature
stages. In ant colonies, for example, injured eggs, larvae, and pupae are quickly
consumed. When colonies are starved, workers begin attacking healthy brood as
well. Hunger and the degree of such brood cannibalism are so precisely related as
to suggest that the colony’s store of immature stages functions normally as a last
ditch emergency food supply to keep the queen and workers alive.
A related phenomenon, widespread in the social Hymenoptera, is egg canni-
balism, oophagy. In the more primitive social groups, the dominant queen eats
the eggs laid by subordinates, thus ensuring that her own progeny will predomi-
nate. This exploitive character has been transformed into something quite different
among the higher social Hymenoptera, however. In these groups it has become
an important form of food exchange between cooperating members of the same
colony. Sometimes, oophagy functions like brood cannibalism; once colonies are
well under way, if workers are starved the eggs are the first brood stage to be
eaten.
In other cases, workers lay eggs that seem designed only to be eaten. Most
workers of Atta leafcutter ants have rudimentary ovaries that never produce eggs.
However, a few small and medium workers form a retinue around the queen and lay
large, flaccid eggs that lack yolk. The eggs are not viable; they serve only to feed
the queen. Such trophic eggs have been reported in a wide diversity of ants and are
apparently laid in enormously varying frequencies among the various ant species.
As a rule, however, the more frequent the exchange of trophic eggs, the less fre-
quent the liquid exchange through trophallaxis. Other cases of trophic egg-laying,
primarily in which mothers provide non-developing eggs for offspring to eat, have
been found not only among social and subsocial insects, but throughout the animal
kingdom from amphibians to fish and marine gastropods.
Chapter 5
Defense: A Survival Catalogue
A predator can be a predator more than once, but an individual

that becomes a prey item necessarily finds itself in this position
once only. . .
5.1 Introduction
A great variety of moths spend their days resting safely upon tree trunks, so perfectly
matching the mottled bark that their very invisibility forms their defense. To repel
honey thieves, some stingless social bees erect walls of sticky resin in front of or
around the nest entrance, while other species smear a repugnant liquid there. Taking
a more active and direct approach, pentatomid bugs earn their common name of
‘stink bug’ from the odorous, distasteful chemical they discharge when disturbed.
Because herbivory is such a major insect lifestyle, Chapter 4 concentrated upon
the evolutionary arms race between herbivorous insects and plants. This chapter will
consider the analogous evolutionary battle of wits between insect predators and par-
asites and their insect prey. Almost all insects face a nearly constant threat of death.
A familiar estimate states that a single pair of houseflies is capable of producing
125 billion great-great grandchildren were all their offspring to survive. Obviously
this does not occur. When averaged over a span of many years, the hundreds or
thousands of eggs a female insect may lay usually result in only a few adults that
reproduce again. This high percentage of losses is due to a variety of factors, but a
major one is predatory attack.
Defined broadly, defense includes all the ways that organisms respond to per-
ceived threats by potential predators, parasites, and microorganisms. Whereas the
insect immune system generally provides the ultimate defense against bacteria,
viruses, protozoans, and fungi (and backup defenses after parasite attack), behav-
ior and life history strategies provide front-line defenses against assault from
macroscopic predators and parasites.
Most past attention focused upon insects as prey of vertebrates, with defense
against invertebrates tending to be overlooked or its importance discounted.
However, insects are often their own worst enemies. Aphids are preyed upon by
a wide variety of insect enemies from adult pemphredonine wasps to coccinellid

186 5 Defense: A Survival Catalogue
beetle larvae, syrphid flies, and lacewings, yet these small, soft-bodied creatures are
far from helpless. A formicine ant stumbling into a foreign ant colony is quickly
hit with venom spray; the victim responds with much grooming and rubbing of its
mouthparts on the ground. The nests of certain tropical social wasps are suspended
on a single stalk that is continually treated with secretions from the wasps’ abdom-
inal glands; ants are repelled from walking upon treated stalks but otherwise enter
and destroy developing brood within the nest. Likewise, relatively little research has
accounted for the fact that insects are also continuously under siege by less conspic-
uous but nonetheless significant microorganisms, such as bacteria, fungi, viruses,
nematodes, and parasites. In these contexts, such chemicals may have long-term
antagonistic or even antibiotic effects.
The potential predator as an evolutionary force has important implications in
every form of defense. For example, in considering color crypsis, the optic system
of the predator must be considered. Insects and vertebrates inhabit what amount
to two very different spectral and focal worlds, a subject that will receive more
attention in Chapter 7. Moreover our own sensory world often differs from that of
an insect’s vertebrate predator, a fact that can lead human investigators astray. For
example, Chapter 8 raises the possibility that many insect ‘protest sounds’ might
actually be vibrational signals intended for conspecifics.
5.2 Defense Messages
Eleodes beetles react to threats by performing a headstand and firing a stinking

irritant spray from their abdomens into the faces of their small vertebrate assailants.
Some mice have evolved a way of thwarting this defense—quickly grabbing a beetle
they stuff its abdomen into the sand and proceed to calmly devour their prey from
the head down.
In order to live long enough to reproduce, all animals must eat, and their food
choices are limited to materials derived from other life forms. With animal flesh
being among the richest food sources in energy and nutrients, not surprisingly a
fierce competition has arisen, as organisms strive to eat others without themselves
being eaten. Under such severe pressure, selection greatly favors those individuals
which are able to reduce the risk of attack or injury with the most effectiveness
while simultaneously interfering the least severely with their performance of other
necessary activities.
Defense is often complex and multi-modal for the simple reason that no given
set of defensive maneuvers is equally effective against all enemies nor equally func-
tional at different life stages. The puss moth caterpillar, Cerula vinula, is one of the
most elaborate in its defense. When it rests upon a branch it resembles a curled
poplar leaf with a blackened margin, even to holding the two prongs of its tail
together like a leaf stalk. When a threat appears, however, the caterpillar quickly
throws off its passive protection, rearing its head to display the startling crimson
front of its prothorax with its two eyespots. Simultaneously, the caterpillar spreads
its forked tail, everting red whip-like threads from its ends and flourishing them in
5.3 Passive Messages 187
the air. Finally, when the attack persists, the caterpillar forcibly ejects from its pro-
thoracic gland a burning, colorless fluid containing 40% formic acid. The adult puss
moth has none of these defense options open to it.
At the same time, certain defense strategies have widespread utility, so evolution-
ary convergence commonly occurs in various structural and behavioral adaptations.
Thus, defensive behaviors have been classified in various ways. One way is to exam-
ine each sensory modality in turn; for example, Table 5.1 presents basic types of
visually perceived anti-predator adaptations. Another approach to classifying insect
defenses is to consider the prey-to-predator message that is implicit in a particular
behavior or structural adaptation. Approximately half a dozen such messages will
handle the bulk of the variety of insect defenses. Lets examine these and see how,
when, and why these defenses have survival benefit.
5.3 Passive Messages

Broadly speaking, insects sort into two non-exclusive categories—passive or
active—with respect to adaptations for defense. For example, body coloration is
a passive defense that can make an insect either conspicuous like a monarch but-
terfly or difficult to see like many moths that blend into their background. On the
other hand, when an insect aims a shot of acrid chemical upon a potential predator,
thwarting the capture attempt, it is active indeed.
Some defenses are completely obvious, always apparent, and carried with an
insect as it goes about its daily life. They are sometimes called primary defenses.
Most of these require no particular special action, and by virtue of always being
present, can be considered to be passive messages. Other defenses come into play
only after actual contact with a potential enemy. The latter are termed secondary
defenses; they typically require some type of active response or reaction on the part
of the insect, and so can be considered as active messages. However, like nearly all
attempts to categorize behaviors, such a dichotomy is not absolute. For example, a
passive defense such as camouflage is enhanced by the active behavior of becoming
motionless. On the other hand, a noxious chemical may be passively incorporated
into body tissues but the predator must actively contact it in order to learn to avoid it.
In the first stages (search and detection) of the predation cycle, a prey’s primary
defenses are of critical importance. Later in the predation cycle (attack and han-
dling), as a predator actually contacts a potential prey and attempts to ingest it, the
stakes are raised for a potential prey. At this point, one or more new forms of active
defensive behavior generally are immediately elicited.
5.3.1 Crypsis: ‘I’m Not Here!’

A whole series of insect larvae including geometrid moth ‘loopers,’ noctuid cater-
pillars, and sawfly larvae have come to resemble pine needles, with alternate stripes
of pale and dark green and a resting behavior in a straight line parallel to the nee-
dles themselves. Stick insects often resemble the substrate upon which they rest with
Table 5.1 Visually mediated systems of passive defense and their correlates
188
Adaptation category Correlated abilities Predator response
Morphological Behavioral Populational
I. Crypsis Countershading; disruptive Prolonged immobility; night Genetic plasticity and/or Fails to discriminate prey
(a) Imitate background coloration; homochromy; feeding; correctly orient; polymorphism for color from substrate
dorsoventral flattening; select proper background; forms; low population
body flanges; often special poses; dash and density; dispersed
secondary startle freeze or move slowly distribution
mechanisms
(b) Resemble inedible Correct shape, color; finely As above; rocks, teeters, Dispersed distribution; low Confuses prey with inedible
objects detailed patterns favored; mimetic poses population density objects
often have startle
mechanisms
II. Bizarre forms Unusual appearance, style, Unexpected, unusual and/or Dispersed distribution; low Fails to see prey as food;
Gestalt; may possess startle startling behaviors population density startled, lets prey escape
mechanisms
III. Simple aposematism Weapons (stings, bites, Usually diurnal; may Clumped distribution; high Learns to recognize prey as
poisons); bold vivid aggregate; conspicuous population density distasteful and/or
pattern, often red, yellow, behavior; warning displays dangerous
orange, black; conspicuous
structure
IV. Mimicry Same as for III; superficially Conspicuous behavior similar Clumped distribution; high Learns Gestalt is distasteful
(a) Müllerian similar to others in complex to others in complex; may population density and/or dangerous
aggregate
(b) Batesian Conspicuous coloration Similar to model(s) Mimetic polymorphisms; low Confuses prey with one it has
and/or structure similar to population density; learned to avoid
model(s) dispersed
(c) Wasmannian Similar to predatory host, at Similar to predatory host that Low population density; Allows prey to approach; is
least in releasers; chemical, serves as both model and linked to social insects with exploited, may become
auditory, tactile mimicry selective agent large colonies at maturity prey itself
favored
5 Defense: A Survival Catalogue
Fig. 5.1 Disrupting the outline. (left) A threadlike reduviid bug (Emesiinae) from Panama is so
thin that it ‘disappears’ except when viewed from the side. Its banded legs are an example of disrup-
tive coloration that enhances its ability to merge into the background. (right) The banding pattern
on the antennae and hind legs of this resting Neotropical long-horned beetle (Cerambycidae) create
the additional illusion that the beetle is going in the opposite direction
striking accuracy. Katydids may imitate leaves to such an extent as to include copies
of blemishes, fungal spots, or bird droppings in addition to reproducing the proper
leaf tint and venation.
By far the commonest method by which insects evade potential predation is
through camouflage or crypsis, that is, by imitating certain environmental back-
ground features; a wide variety of diverse unrelated insect groups often have evolved
to simulate the same inedible (to a carnivore) object. Crypsis involves at least shape,
color, and color pattern (see Plate 14). In some instances it probably also involves
scent and sound matching although human sensory apparatus may not be equal to
the task of discerning it. To be maximally camouflaged, a cryptic individual also
must solve the major problem of body contour. Probably the most widely used solu-
tion is disruptive coloration, a visual breaking of the insect’s outline so that parts of it
appear to fade separately into the background (Fig. 5.1). A second way of minimiz-
ing contour cues involves actually or apparently reducing any telltale shadows. This
may be accomplished through a dorsoventral flattening, as occurs in many aradid
bugs, often in combination with lateral flaps or various irregular body protuberances
that bridge the gap between body and substrate.
Of course, even the most exact reproduction will be of little value in concealment
without appropriate behavior, such as resting on the proper background in the right
orientation and attitude, moving seldom and/or slowly and in such a manner as to
attract as little attention as possible. Thus, properly considered, crypsis is not solely
a matter of coloration or morphology but a type of behavior as well. For exam-
ple, many moths have wing stripings that resemble grooves in bark; those moths
with vertical stripes rest on tree trunks with their heads pointed up or down, but
horizontally striped moths orient themselves at right angles to the trunk.
Another widely encountered method for eliminating shadows is countershad-
ing, that is, a compensatory deepening or lightening of body color to counteract
Fig. 5.2 Countershading. Nearly every green caterpillar that rests in the open has color shad-
ing that compensates for the shadow on its lower parts, producing an apparent flattening. (above)
A hornworm caterpillar (Manduca) in normal resting position. (below) Inverting the caterpillar
destroys the illusion, and the addition of countershading to natural shadow makes the caterpillar
highly conspicuous
for apparent color changes due to light intensity. Insects that have countershading
invariably rest with the darker surface directed toward the light, a behavior critical to
the success of countershading (Fig. 5.2). Shadow also may be minimized by proper
body alignment. (It has been suggested that many cryptic butterflies perch relative
to the sun so that their wings throw the least shadow, but the observation is compli-
cated by the fact that butterflies also control body heat by soaking up or avoiding
sunlight.)
Some insect species have the chameleon-like ability to change their color to
match temporary backgrounds. By varying the quantities of orange, yellow, and
black pigments they form, many locusts can adapt their color from dirty white to
yellow, brown, or black; kept for a few days on burnt ground, even adult locusts can
darken. Only when the hoppers are kept amid green vegetation and fed with abun-
dant moist food in a very humid atmosphere do bright-green locust hoppers appear.
Likewise, certain caterpillars are able to change color and markings to match the
background of their varied diets. When they feed upon birch, the larvae of peppered
moths are smooth and purplish brown with darker ‘lenticels’ resembling a birch
twig; on oak, however, they are brownish green with markings suitable to these
trees. On oak covered with lichen, they become mottled in such a way as to mimic
a lichen-covered twig!
Genetic variations in cryptic coloration also occur. One of the most thoroughly
studied cases of rapid directional selection in progress involves crypsis in this com-
mon English insect, the peppered moth (also sometimes called the salt-and-pepper
moth), the subject of Case Study 5.1.
Case Study 5.1: Melanism in the Peppered Moth,

Biston betularia
What happened to change the color of a little cryptic moth? The research that
answered this question has become a classic story in biology textbooks.
Peppered moths, like many others, fly at night and rest quietly by day on
the sides of tree trunks. Naturalists of the nineteenth century knew them well
and noted that they were usually found on lichen-covered trees and rocks,
against which their pale mottled coloring made them practically invisible.
Prior to 1845, all museum records of Biston betularia were light-colored or
‘typical’ specimens, but in that year, near the growing industrial center of
Manchester, England, one black or ‘melanic’ moth of that species was cap-
tured, so distinctive that it was given the name B. betularia f. carbonaria (see
Fig. 5.3). As the years passed, black individuals turned up with increasing
frequency, mainly from the vicinities of industrial towns. By the mid-20th
century, the peppered moth population around Manchester consisted of nearly
99% carbonaria.
Fig. 5.3 Attracting attention—or not. The peppered moth, Biston betularia, and its black form
carbonaria at rest on a lichen-covered tree trunk in a non-industrial area of England. The typical
light form (upper center) illustrates the principle of disruptive coloration, but the dark form (lower
center) shows conspicuous contrast against the same background (see also Plate 13)
Why were melanic moths becoming predominant? In 1952, H. B. D.

(Bernard) Kettlewell undertook the first field experiments on the phenomenon.
Smoke from the expanding industrialization of England had polluted the sur-
rounding countryside, he noted, killing the epiphytic communities on the trees
and blackening tree trunks and the surrounding rocks and ground. It seemed
plausible that upon this darkened background the dark moths were better pro-
tected from predators, particularly birds, but Kettlewell had little support for
this hypothesis. Many entomologists and ornithologists vigorously objected
that no one had ever seen any bird eating a B. betularia of any color, nor was
there any published support for such predation on any of a variety of related
cryptic moth species.
Thus, Kettlewell first needed to determine whether in fact birds actually
ate peppered moths and whether they did so selectively. Into a large outdoor
aviary containing a nesting pair of birds with young, he placed both light
and dark tree trunks and boughs of equal surface area. Into the aviary he
released ten moths—five melanic and five typical B. betularia. Based upon
their appearance against these backgrounds at a distance of about two meters,
resting moths were classed as ‘conspicuous’ or ‘inconspicuous’.
For the first two hours of the first day, the birds failed to recognize the
moths of either form as food. But in the next hour, all five of the ‘conspic-
uous’ forms and two of the ‘inconspicuous’ moths were taken. In a repeat
experiment the next day, Kettlewell released 18 moths; 16 were eaten in the
first half hour, leaving only two ‘inconspicuous’ survivors. Subsequent repli-
cations followed the same pattern, suggesting that the birds required a period
of contact with the prey before recognition but that after recognition was
established, all the moths were under risk of predation. Satisfyingly, the data
seemed to indicate that the most conspicuous were preyed upon first.
Encouraged, Kettlewell turned to field experiments in two study sites: a
polluted area and an uncontaminated woodland like that which must have
prevailed 200 years before. Setting up a camp that included sheds to house
some 3,000 B. betularia pupae to raise for release, Kettlewell began round-
the-clock studies. During the day, he released melanic and typical moths in
known numbers, carefully marking each with a paint spot beneath the wings
where it could not be seen by a predator. Then he watched wild birds’ behavior
toward the moths. During the night he operated light traps, capturing the pep-
pered moths of both forms that were attracted to the ultraviolet bulbs. (The
use of such a method was essential because of the moths’ crypticity. When
Kettlewell attempted to visually discover and count his released moths again
immediately after their release in an area where his continual presence pre-
cluded predation, over a third of the cryptic forms were already ‘missing,’
that is, so well hidden that they eluded even Kettlewell’s experienced eye.)
With help from the prominent ethologist Niko Tinbergen, Kettlewell
watched bird predation on the moths, both with binoculars and with a movie
camera from behind camouflaged blinds. In this way he was able to document
at least five bird species actually selecting and eating the moths, which they
did with such alacrity that it is surprising they were not previously observed.
More important to his hypothesis, on the majority of occasions all the bird
species at both locations took all of the more conspicuous moths before any
of the inconspicuous phenotypes. Further proof of the selective advantage of
having coloration appropriate to one’s background was obtained by recap-
ture of marked moths at night in the light traps. In unpolluted woodland,
the survival rate of light moths was about twice that of dark ones, while in
soot-darkened woodland the ratio was reversed.
Do moths discriminate among a choice of possible resting substrates? As
a pilot study, Kettlewell outfitted a barrel with a lining of alternate black
and white stripes of identical texture to eliminate tactile cues. Each night he
released three moths of each form into the barrel; in the morning, he recorded
their resting positions. Of 110 moths so tested, 65% chose the ‘correct’ back-
ground. However, the artificiality of such tests disturbed Kettlewell, so he
selected a number of lichen-covered tree trunks and then carefully removed
the bryophytes from one side. Over this denuded half he painted a soot sus-
pension and then placed the trunk upright and covered it with a muslin tent.
At the midline of the two sides, he released equal numbers of females of each
form and recorded their final resting position. Every one of 31 carbonaria
individuals chose the black background; of the light forms, 75% selected the
‘correct’ background. These highly significant results provide quite conclu-
sive evidence that female Biston betularia are able to selectively choose their
background with respect to their own coloration, probably through visual cues
(see Plate 13).
Several important additional implications arose from Kettlewell’s studies.
One major revelation was the speed with which evolution can occur at the level
of the single gene. Geneticists had previously considered a selective advantage
on the order of 0.1% as being ‘normal’—Kettlewell illustrated an advantage
on the order of 50%! This selection, intense enough to cause a significant
change in gene frequency within a single generation, when extended over tens
of generations could result in a nearly complete reversal in the favored gene.
In the years since Bernard Kettlewell’s study, others have continued to research
what is now called ‘industrial melanism’ in peppered moths, and there has been
some evidence that other differences in fitness among the phenotypes that are not
directly related to the visual differences in color pattern may also be involved.
However, studies continue to support the fundamental finding of a relatively rapid
switch in survival of the pale and dark phenotypes in response to corresponding
changes in crypsis arising from an altered environment (Fig. 5.3). Interestingly,
as Great Britain and other countries in northern Europe have reduced levels of
air pollution from soot and gases such as sulfur dioxide, the tree trunks are once
more showing their pale bark’s true color; in response, the relative fitnesses of the
two moth phenotypes are changing again. The fully black form is again becoming
more conspicuous and vulnerable to birds, and the now-better-camouflaged light
phenotype of the peppered moth is on the increase.
5.3.2 Systemic Defenses: ‘I’m Noxious!’

Certain brightly colored tiger moths that fly at night may be captured by bats but
rejected by experienced bats because the moth’s bodies contain toxic alkaloids
obtained as larvae. Aphids that feed on oleander and milkweed similarly gain protec-
tion by accumulating host-produced steroids that are toxic to predators. Distasteful
lycid beetles are particularly prone to bleed from the wings, which possess swollen,
easily ruptured veins.
Although hiding from predators is a common insect defense, another approach is
to become noxious, that is, harmful to a predator’s life or health. Chemical defenses
are common; an insect may smell or taste so bad that predators spit it out, get sick
and learn to avoid it, or even die and thus no longer be around to bother it. Many
interactions of insects with potential predators involve allomones (see Chapter 6)—
chemical agents that benefit their producer while (often detrimentally) affecting
another species. Allelochemic defenses are widespread throughout the biological

world, but it is among the invertebrates, especially the arthropods, that chemical
defense seems to have reached its peak of diversification.
One way to acquire defensive chemicals is to ‘borrow’ deterrents already present
in one’s prey or host plants. A classic example is the caterpillars of the monarch
butterfly that can feed harmlessly on milkweed plants whose sap contains cardiac
glycosides that are potent emetic agents to birds; when they become adult butterflies,
the monarchs retain these toxic chemicals in their wing veins. However, depend-
ing on a plant’s toxic principle can be an evolutionary constraint in its own right.
Monarchs raised on cabbage instead of milkweed have no chemical protection.
An alternative is to synthesize one’s own anti-predator chemicals. For exam-
ple, aquatic Hemiptera differ from their terrestrial relatives in producing aromatic
instead of straight-chain hydrocarbon defensive compounds. The apparent signifi-
cance of this difference is that the aromatic compounds are much more repellent
to fish.
Like plant chemical defenses, some insect defenses are strictly passive: parts of
the insect must be eaten before an effect is felt. However, an insect normally can
ill afford to allow body parts to be ingested before a predator learns of its mistake.
An individual facing predation would have an obvious advantage were it to have
some ‘disposable’ way to be sufficiently bad tasting or bad feeling for the predator
to reject it while leaving the rest of the insect relatively unharmed. For example,
a v-shaped notch in the wing of a butterfly is convincing evidence of an avian
‘educational experience’ allowing the butterfly to survive another day. Like monar-
chs, many butterfly species incorporate systemic noxious chemicals in their wing
tissues.
Some distasteful insects have cuticular outgrowths such as hollow, brittle spines
that are easily broken by predators. Many caterpillars have developed a coat of long
hairs that render them unacceptable to most predatory birds, but some species gain
still further protection by the inclusion of urticating hairs, delicate hollow organs
each containing a minute quantity of irritating venom secreted by a gland at their
base (Fig. 5.4). These occur in at least eleven lepidopteran families. When brushed
against, these fragile hairs break and act like tiny poisoned barbs.
The evolutionary and ecological implications of various systemic chemical
defenses can be problematic to explain. Consider the case of the blood and tissues
of meloid blister beetles, which contain cantharadin. Long known as Spanish Fly,
cantharadin was traditionally but incorrectly reported to have aphrodisiac proper-
ties, but the chemical actually is capable of inducing severe systemic effects when
ingested. It is present in such quantities in some meloid species that swallowing a
single beetle could kill some vertebrates. How can a systemic poison like Spanish
Fly be adaptive? It would seem as though no learning could occur if the potential
learner is now dead, and the beetle is no better off, being dead as well. And what
about the many other cases where noxious effects are delayed—how can a predator
be expected to associate a particular feeding event with ill effects that occur so much
later? In fact, since toxins require energy in their production or incorporation, their
development actually would seem to be disadvantageous.
Fig. 5.4 Stinging caterpillars. (top) Feeding group of pine processionary caterpillars
(Thaumetopoea pityocampa). The short hairs along the dorsal surface contain skin-irritating chem-
icals. The processionary behavior of this species was subject of a study by J. H. Fabré over a
hundred years ago. Fabré attributed the circling to blind instinct, and it is one of his best known
insect stories, perhaps because it came to be viewed as a metaphor for mindless living or blindly
following a leader. (below) The buck moth, Hemileuca mala, is another stinging caterpillar. The
impressive crowns of urticating spines that adorn each segment (inset closeup) can cause severe
itching,
Often such questions cannot be answered, if for no other reason than incom-
plete knowledge of the particular system. Lethal dosages almost certainly depend
on the predator involved, a fact often unknown. In still other cases, even if predatory
organisms neither die nor discriminate against the causative agent, systemic chemi-
cals may impair vigor or fecundity; any substance that does this to the predators of
an area is advantageous to the species that produces the poison.
How can the sacrifice of ingested individuals be justified in adaptive terms?

Sometimes the answer may lie within the concept of altruism (see Chapter 10), in
which the surviving insects are such close kin to those sacrificed that their survival
is nearly equivalent genetically to the survival of the sacrificed individual itself.
Support may be found in the observation that many poisonous arthropods develop
dense, highly localized, and often quite static aggregations, conditions that would be
expected to favor a high degree of relatedness. Clonal aphid clusters are an example.
Behaviors such as egg placement in one or a few concentrated masses (a common
oviposition pattern in butterflies) also probably increase the likelihood of genetic
relatedness of adjacent individuals.
Chemical defenses have received so much research attention that before leav-
ing this subject, it is important to reiterate that being noxious is a broader concept
than merely being foul-tasting or unpalatable. It can also involve defenses that have
no direct connection to chemistry. For example, most people avoid bumblebees
because of their stings, and assume that stings are the reason that predators do,
also. However, it appears that the tough chitin and hairy bodies of these insects are
at least as important, making it necessary for birds to expend great efforts to sub-
due, tear apart, and swallow them. Naïve young birds have been observed taking as
long as 18 minutes to kill, dismember, and eat a single bumblebee; the equivalent
handling time for houseflies, mealworms or beetles was a fraction of a minute, even
for young birds.
It is also important to note that when it comes to noxiousness, defense and offense
are particularly closely linked. Behaviorally, the use of systemic chemical defenses
covers a range that overlaps at the active end with forms of assault, the subject of
Section 5.4.
5.3.3 Mimicry: ‘I’m Someone Else!’

Cerambycid beetles and a host of different flies often resemble ants, bees, or wasps.
A larva of the moth Hemeroplanes responds to threat by inflating its anterior end to
form an excellent representation of a snake’s head, then waving it about. The viceroy
butterfly, Limenitis archippus, sports a color pattern that is completely unlike any of
the species to which it is closely related but is very similar to that of monarch but-
terflies; furthermore, the viceroy has three color forms, each resembling a different
species of monarch.
When insects imitate environmental objects such as twigs, leaves, and thorns,
scientists call it crypsis or camouflage, but when insects imitate another living
organism, scientists call it mimicry. Mimicry is the resemblance of one organism
(the mimic), usually in color, pattern, form or behavior, to another organism (the
model). Butterfly mimics and models have received what is probably more than
their fair share of the attention, because numerically Hymenoptera form the most
important group of models, and Diptera are said to form the largest group of mimics
(particularly due to almost 300 species of mimics in the hoverfly family, Syrphidae).
Mimicry can be likened to a stage performance with a minimum of three actors:

two signal senders and one receiver, with the receiver judging the resemblance of
the signals from the two senders. In most cases, the resemblance of one organism
to another can be considered mimicry only if both organisms are found together.
(Migration and long-term associative memory might provide exceptions.) When
similar forms are found in different areas, it is more likely to be an example of
evolutionary convergence. Although visual mimicry has received the most attention
and study, mimicry also may be olfactory, tactile, or auditory. Mimetic signaling is
often multimodal, and the particular evolutionary direction it takes in a particular
case undoubtedly has been influenced by receivers’ sensory ecology. Appropriate
behavior is necessary for an effectively functioning system, with mimics acting in
ways that enhance their deception.
There are several types of mimicry. Two of the most common types, called
Batesian and Müllerian, are related through the important evolutionary dynamics
that arise from whether receivers are expected to try to discriminate or generalize
on the senders’ signals. In other words, are the senders sending dishonest signals or
honest ones?
Dishonest signals characterize Batesian mimicry, the situation that occurs when a
palatable species evolves an appearance similar to that of an unpalatable, venomous,
or otherwise protected one (see Plate 15). The phenomenon is named after Henry
Walter Bates, a British entomologist who first observed it while traveling in the
Amazon in the mid-1800s and proposed that by imitating a model that is unpalat-
able or dangerous, individuals can gain an increased protection from predators that
confuse them with their aposematic model.
For a Batesian mimic, being rare is advantageous; becoming more common
increases the probability that a predator will discover that some insects (the mim-
ics) in the complex are palatable and therefore all (mimic and model alike) are
worth sampling. Thus, the fitness of Batesian mimics is inversely related to their
proportion in the prey community, a situation called negative frequency depen-
dence. It is no surprise that in natural situations Batesian mimic species are as a
rule far less common than their models, but when a mimic species is numerous,
an interesting phenomenon occurs. Consider, for example, the African swallow-
tail butterfly, Papilio dardanus, the female of which exists in half a dozen different
color forms. In different regions of its range, different color assortments appear,
and the balance of the percentage of the forms varies as well. In addition, other
unrelated butterflies in some of these areas produce mimics with the same col-
oration. The example is not an isolated one; a great many Batesian mimics are
polymorphic.
The establishment of such mimetic polymorphisms is relatively easy to explain
on theoretical grounds. Because palatable mimics must be less common than their
models, a mimetic form can increase in abundance only up to a certain point before
losing its advantage. Strong pressure would exist toward development, within part of
the population, of mutants resembling other distasteful species. These mimics, too,
would show negative frequency dependence, so any new (or rare) mutant resembling
still another protected model would be favored. In this way, a complex ‘balanced
polymorphism’ could be established. By presenting varied appearances, polymor-

phism also theoretically would reduce the rate at which predators develop a strong
search image that improves their ability to capture that particular species. This could
be quite advantageous because over time, search image formation results in apo-
static selection, a situation in which the relative rate of predation rises more rapidly
than does the relative rate of encounter.
Some of the best-studied cases of Batesian polymorphism are found among but-
terflies in the families Papilionidae and Pieridae. In some of these species only the
female is mimetic. Negative frequency dependence is given as the usual explanation;
by restricting mimicry to one sex, the effective mimetic population size is halved.
Why should females be the chosen sex? One suggestion has been that mimicry is
more beneficial to them. Female butterflies have less agile flight because of their egg
load, and more predictable patterns when searching for oviposition sites. Another is
that male wing patterns may be constrained by sexual selection; males might not be
able to evolve Batesian mimicry without losing mating opportunities and/or being at
a competitive disadvantage with other males. In experiments with North American
swallowtail butterflies that have female mimics, painting males mimetically lowered
their mating success; it also caused them to lose more fights with other males and
therefore to hold lower quality territories.
Different dynamics occur when the mimetic signal is an honest one. Consider the
various bees and wasps with their bands of yellow and black, or the tiger stripe but-
terfly complex of the Neotropics. Situations like these—when two or more relatively
unrelated insects that are all distasteful or otherwise protected and that share one or
more common predators mimic each other’s warning signals—are named Müllerian
mimicry. The advantages of such a system are obvious. When aposematic species
converge in appearance, all of them benefit, sharing the number of predatory attacks
incurred in the predator’s learning process and reducing its confusion and learning
time.
Müllerian mimicry runs counter to Batesian mimicry on three important points.
First, with Müllerian mimicry there is no bluffing; each warning of unsuitability as
prey is absolutely true. Thus, while Batesian mimicry has been likened to a par-
asitic relationship in which the mimic benefits at the model’s expense, Müllerian
mimicry is a true mutualism. When German naturalist Fritz Müller first proposed
this eponymous concept in 1878, he backed it up by demonstrating mathemati-
cally that two unpalatable prey could each benefit from mutual resemblance, and
showing that this benefit was biased in favor of the rarer of the two species, to a
factor equal to the square of the ratio of the species’ abundance. Therefore, unequal
population sizes translate into even more unequal benefits. However, even at the
extremes, the mimicry is still mutually beneficial. Thus one could say that for
a predator the existence of Batesian mimicry complexes is detrimental, because
they result in avoidance of ‘good’ food, but Müllerian mimicry complexes are
advantageous to the predator, since they result in pursuit of fewer injurious or
unpalatable items.
Second, whereas polymorphism is favored in Batesian mimicry, polymorphism
runs contrary to the advantages of the Müllerian system. Because both the warning
prey and the learning predator benefit from a correct interpretation of the sig-
nal, rare or new variants within a prey population are at a disadvantage; they
will not be recognized as distasteful and thus will suffer higher predation. This
selection against rare forms translates into positive frequency-dependent selection.
Thus, not surprisingly, most distasteful Müllerian mimics are monomorphic in local
populations, with polymorphic forms that bridge narrow hybrid zones between
color-pattern races.
Third, as more Müllerian mimics use the same warning signal, the protection
given by the signal becomes stronger. Predators generalize more, and the selection
for close resemblance appears to be somewhat relaxed. Ultimately it would seem
as though there should come to be just one giant mimetic cluster in the entire habi-
tat, but nature seems to behave otherwise. Particularly in the tropics, aposematic
insects of similar size and shape usually group into not just one, but several distinct
complexes. A ‘mosaic mimetic environment’ is usually invoked as an explanation—
if different mimicry rings occupy different microhabitats and predators do not move
between them, then each ‘subpopulation’ might have different fitness peaks and
also be small enough to be particularly sensitive to genetic drift into different color
patterns in different patches.
Still, despite these theoretical differences, Batesian and Müllerian mimicry both
depend on stimulus generalization by their vertebrate predators, and both are prob-
ably labels for extreme cases in what is essentially a mimicry continuum. Within a
mimicry complex, for example, some members may be more palatable than others.
Should these less protected individuals be considered Müllerian mimics for which
selection for bad taste has been relaxed? Or are they Batesian mimics that have sub-
sequently evolved (or are in the process of evolving) a bad taste? Such questions
are nearly impossible to resolve. In another example, certain mimetic cerambycid
beetles often feed upon the distasteful lycid beetles that serve as their models. Such
mimics might be alternatively Batesian or Müllerian, depending upon how recently
they had fed upon a lycid.
Although Batesian and Müllerian mimicry receive most of the attention, other
types of mimicry are also documented and classified by a number of different names.
One of the better known is aggressive mimicry. James E. Lloyd s investigation of
female fireflies of the genus Photuris revealed they emit the same light signals that
females of the genus Photinus use as a mating signal; male fireflies from several dif-
ferent genera are attracted to these ‘femmes fatales’, and are subsequently captured
and eaten (see Chapters 1 and 7 for more on fireflies).
Whereas in this example the fireflies are the predators, in other cases the reverse
is true and an insect may mimic its major predator and thereby reduce the risk of
being eaten by it (Fig. 5.5).
A slightly different matter is automimicry, or mimicry that occurs within a single
species when a population includes both mimics and models because of a palatabil-
ity dimorphism. For example, monarch butterflies have been mentioned as a classic
Batesian model, but not all monarchs are equally disagreeable. Their caterpillars eat
various milkweed plants in different parts of the monarch’s range and these vary in
the amount of cardenolides they contain, some having little or none. The butterflies’
Fig. 5.5 Sending false signals. The tethritid fly Zonosemata vittigera has a leg-like pattern on its
wings (left) and waves these patterns such that these ‘legs’ mimic the agonistic territorial displays
of its primary predators, jumping spiders. In response, the spiders do not attack, but either return
the display or retreat. Both the pattern and the behavior are necessary for effective mimicry. In
transplant experiments (right), spiders attacked tephritids with house fly wings and house flies
with tephritid wings. Each treatment used 20 flies, and responses were measured as the highest
level of aggression attained during 5 minutes of interaction in a test arena
palatability depends on the amount of cardenolide eaten as larvae; this means that
individuals in a population vary in their palatability to predators. Those lacking
chemical protection are effectively mimics of their fellow species mates that are
distasteful.
Many insects change shape, size, and behavior during growth and metamorpho-
sis; when different instars imitate entirely different models it is called transfor-
mational mimicry (Fig. 5.6). Beetles and other guests of ants sometimes greatly
resemble their hosts (Chapter 4); mimetic resemblance that facilitates cohabitation
with a mimic’s host, its model, has been termed Wasmannian mimicry.
The mimicry catalog could go on and on. However, as in other areas of insect
behavior, studies of mimicry have steadily been stepping back from a natural history
of individual examples to a broader synthetic view—one that includes modeling,
predator behavior (including the psychological processes of learning and forget-
ting), and evolutionary dynamics. A major driving force has been the inadequacy
of older theories to explain the phenomenon of imperfect mimicry. As more and
more mimicry systems have been studied, it has become clear that most Batesian
mimicry, in particular, is of rather poor quality, at least to human eyes looking at
its visual aspects. This poses a theoretical puzzle. Conventional evolutionary theory
Fig. 5.6 Changing tactics. Transformational mimicry in larvae of the eastern tiger swallowtail
butterfly, Papilio glaucus. (left) a first-instar larva resembles amorphous bird droppings, and rests
in full view at the center of a leaf. Notice the nibbled spot at the leaf tip. These larvae feed only at
night, when their diurnally active predators have stopped hunting. (right) Older larvae have grown
too large to effectively resemble droppings. Colored bright green like the leaves they feed upon,
they also possess eyespots and perform a striking snake-like display (see Fig. 5.12) when disturbed
would predict that because individuals with closer resemblance to the model gain
more protection, Batesian mimics should be under constant selection to improve
their mimetic resemblance; hence eventually only very good mimics should exist. A
variety of (often not mutually exclusive) new explanations for this seeming puzzle
have arisen; testing, proving, and disproving them should keep mimicry researchers
busy for some time. One key to the puzzle may lie in the fact that many if not
most vertebrates tend to avoid stimuli similar to, but not identical with, an origi-
nal conditioning stimulus, a learning phenomenon called stimulus generalization.
Experimental studies have revealed that under certain conditions, stimulus general-
ization provides advantage to a wide gamut of prey even in cases where they may
appear only vaguely alike to our perception.
5.3.4 Aposematic Defenses: ‘I’m Dangerous!’
The Carolina locust, Dissoteira carolina, combines vivid yellow and black coloring
with a loud crackling sound. Brightly colored bees and wasps sound a warning buzz
sufficiently effective to be mimicked by a wide variety of non-relatives. Caterpillars
that possess stinging hairs often undulate conspicuously as they crawl about. Other
insects swing rhythmically, produce rattling or rustling noises, or adopt various

bizarre display postures.
With bright, contrasting colors and shapes and conspicuous behaviors, a wide
variety of insects literally advertise their presence to the world. They are described
as aposematic, an adjective that is used most often in the context of visual cues but
can be applied to any conspicuous cue or behavior that indicates special capacities
for defense. Classifications always involve arbitrary choices, but with aposematic
defenses, the division between passive ‘primary’ defenses and active ‘secondary’
defenses breaks down almost completely. The general maxim for an aposematic
insect is almost always, ‘If you’ve got it, flaunt it!’
Aposematic coloration in insects tends to follow definite patterns of rather limited
variety—most frequently combinations of red, yellow, or orange on a contrasting
background, often black. The color pattern of a bumblebee is an example. Such pat-
terns contrast sharply against natural backgrounds that are generally green, brown,
gray, or blue.
The advantages of the end product seem obvious. When one is poisonous, nox-
ious, or otherwise disagreeable or dangerous, it pays to advertise. Standing out
brightly differentiates oneself clearly from cryptic palatable prey. It also may jolt
potential predators out of any non-discriminating ‘state of mind.’ Limiting the num-
ber of patterns enhances the efficiency of the warning by limiting the amount of
information that a predator must learn for accurate feeding judgments. It is even
possible that vertebrate predators have innate biases against yellow and red that
aposematic insects have been able to exploit. All these ideas are supported by
experimental data to some extent, though less than overwhelmingly. Because prey
signaling and predators’ cognition probably coevolved, it is generally difficult to
establish independent experimental evidence for any of these ideas. Happily, they
are not mutually exclusive.
Why and how did aposematic defense arise? Ever since Charles Darwin first
wrestled with an explanation for brightly colored caterpillars and Alfred R. Wallace
suggested that their bright colors might advertise unpalatability, the evolution of
warning coloration has intrigued and puzzled people. How could warning colors
have arisen? If most of a population is cryptic but a few colorful mutants arise,
wouldn’t predators immediately remove them from the gene pool? Maybe not. There
is evidence that the mere fact of being bright and/or different can offer protection.
It seems most likely that aposematic colors evolved in response to vertebrate preda-
tors, particularly birds, rather than insect predators. Insects perceive light at the red
end of the spectrum poorly, if at all (see Chapter 7); red presumably appears gray
or black to most and would therefore be of little significance as a signal to other
insects. On the other hand, birds readily perceive bold patterns of red or orange.
Faced with something unusual, vertebrate predators often hesitate. Many birds, in
particular, will actually avoid novel food items.
At first, scientists had their hands full just cataloguing aposematic examples,
which seemed to occur almost everywhere they looked, from brightly colored poi-
sonous tropical butterflies to bitter-tasting ladybug (ladybird) beetles (see Plate 5).
However, over the years, research on aposematism has become more sophisticated
and complex, shifting to a search for underlying cognitive, behavioral, frequency-

dependent and coevolutionary mechanisms. For example, among saturniid moths,
striking life history adaptations exist which correlate to their apparent defensive
strategies. For example, unpalatable aposematic species tend to have a longer life
span following reproduction than do cryptic, presumably palatable species. The
apparent explanation is that even after reproduction distasteful moths can confer
some protection on their siblings by serving as a predator’s learning experience. In
contrast, the greatest benefit a cryptic species could confer to its relatives would be
to remove itself from the population soon after breeding so as not to chance that a
predator might develop a ‘search image’ for its species.
5.4 Active Messages

When disturbed, swallowtail and parnassian butterfly caterpillars suddenly evert a
snake-like ‘forked tongue’ (osmeterium) secreting intensely odorous butyric acid
derivatives. Solitary wasps and their prey play games of cat and mouse, and the prey
does not always lose. Blister beetles often ooze irritating hemolymph from leg joints
when handled.
Of course, the simplest response to attempted predation is to simply run or fly
away, if you can. Most insects do this if you attempt to grab them, and cockroaches
are masters at it (see Chapter 2). For insects that inhabit burrows or make portable
cases simply withdrawing or retracting their bodies is usually sufficient to avoid
being eaten. Tiger beetle larvae that live in burrows, case-making caterpillars and
case-bearing beetle and caddisfly larvae all exemplify this type of disappearing act.
However, secondary defenses exist in astonishing variety; we will consider them
under three broad groups of active insect defense messages—attack, startle, and
group actions.
5.4.1 Attack: ‘I’m Turning the Tables!’

When threatened, Solenopsis fire ants both sting and bite; clamping down with
their mandibles secures the hold they need to jab their stinger into the enemy.
Although they do not have operative mouthparts, certain insect pupae such as
mealworms have evolved ‘gin traps’ that give them the ability to ‘bite’; along
their abdomen segments, clefts with sharp hardened borders can be clamped shut,
trapping attackers.
The insect arsenal includes many morphological adaptations—powerful mouth-
parts, leg spines, raptorial forelegs, and a whole host of other such features—that
can be used highly effectively for protection, and the behavioral line between
defense and offense can be very thin and finely drawn. However, when it comes
to actual attack, chemistry reigns supreme. The allomones involved in offense and
defense tend to be blends of compounds. The secretions of some pentatomid bugs,
5.4 Active Messages 205
for example, include as many as 18 components. The functional significance of

such mixtures appears to lie in their increased chance of repelling diverse sorts of
predators, combined with the ability of one chemical to alter the physical and/or
chemical properties of another.
Consider the use of venom, for example. The ‘aculeate’ Hymenoptera are so-
named for their weaponry—glands associated with the egg-laying apparatus secrete
venom and the ovipositor itself has been transformed into a sting. Solitary wasps
(Pompilidae, Crabronidae, and Eumeninae) primarily use their venom to paralyze
prey, but among the social bees, wasps, and ants the sting and venom have come
to be retained solely for defense/offense. Occasionally such stings can cause violent
and severe reactions in humans, an example of anaphylaxis, a generalized reaction to
foreign proteins to which the body has become sensitized. Males hymenopterans are
incapable of stinging, although some have a well-developed spine (pseudo-sting) at
the end of the abdomen which is realistically thrust at molesters and may sometimes
even draw blood.
Many non-stinging insects also use secretions and excretions in a manner that
ranges from defense to active attack. Most of them discharge their chemical defenses
only in response to direct contact stimulation, but both intermediates and exceptions
occur. Hemiptera such as Rhopalus collect secretions in depressions on their body
walls; when bothered, they dip their legs into the accumulated chemical and then
wipe their legs on the enemy target.
Some insects’ hemolymph contains noxious poisons as a passive defense.
However, to minimize the chance of being eaten and maximize the chance of escape
many species provide relatively non-harmful ways for predators to taste one’s toxic
hemolymph. One way is to have active control over the release of hemolymph when
under attack (Fig. 5.7). Such reflex bleeding, often occurring from leg joints, is
Fig. 5.7 Bleeding reflexively. When gripped with forceps, an ithomiine day-flying moth from
the Neotropics foams released hemolymph with air from its spiracles
highly developed in the Mexican bean beetle, which may even rotate its leg around
to bring the oozing knee in closest contact to the point of stimulation.
When handled or disturbed, a wide variety of arthropods respond by regur-
gitating or defecating. Often very effective as predator repellents, such enteric
discharges might almost be considered a form of short-range attack. The familiar
frothy ‘tobacco juice’ regurgitate of some grasshoppers, for example, is quite toxic
to mammals. It is a topical irritant to eyes, may induce vomiting when swallowed,
and may cause severe symptoms and even death when injected. When tethered
grasshoppers placed beside ant colonies are induced to regurgitate, the fluid causes
their assailants instantaneously to disperse and begin intense cleaning movements.
When pieces of cut-up grasshopper, some of which have been treated with regur-
gitate, are placed along ant trails, foragers carry away only the untreated pieces.
Such protection can be extremely important, particularly in defense against massed
predators such as ants. Various morphological and behavioral adaptations prolong
the effectiveness of such residual secretions (Fig. 5.8).
Spraying is a common way to apply non-injected secretions, and it is often
accomplished with a high degree of accuracy and sometimes over a considerable
distance. In general, these sprayed chemicals are much stronger than most enteric
discharges. The predaceous reduviid bug Platymeris rhadamantus sprays its saliva
directionally several feet in response to predator attack; the toxic fluid, ordinarily
used to kill prey, is said to be similar to cobra venom. The large two-striped walk-
ing stick Anisomorpha buprestoides of the southeastern United States is exceptional
in that it will spray approaching birds from a distance, whereas most non-stinging
Fig. 5.8 Prolonging repellency. (left) The pupal stage of a Chrysomela leaf beetle hangs fully
exposed to predators on the undersides of the host plant. Persistent odor of the larval defensive
secretion impregnates the shed skin, which remains attached (arrow) and acts as a potent repel-
lent to foraging ants. A second larva preparing to pupate is also present. (right) A larva of the
chrysomelid beetle Cassida rubiginosa forms a shield of cast skins and feces that is held on a fork
projecting from its hind end. When branched spines on its body detect a predator’s probing, it inter-
poses the shield between itself and the enemy, often in such a way that fecal material is smeared
upon the offender. After even a brief contact with the pasty material, ants immediately flee and
clean themselves
insects require direct contact stimulation before discharging their chemical defenses.
Some of the commonest ant species use formic acid as a poison spray and thus have
been used as a natural source of this acid from Roman times until quite recently.
Formic acid is a potent irritant and effectively repels a variety of potential ant
predators. It also serves as an alarm odor (see Chapter 6) that alerts nestmates to the
source of the disturbance. The ants employ stereotyped postures to control spraying
direction (Fig. 5.9a).
Probably the best known spraying insects are the 500+ species of carabid ground
beetles known as bombardiers (Fig. 5.9b). They occur through much of the world,
and vary in size but all have essentially the same defensive mechanism. The
chemicals discharged are 1,4-benzoquinones, irritant chemicals that many other
arthropods also have independently evolved the ability to produce. What is remark-
able about the bombardiers is that they do not store their benzoquinones as such.
Instead, they produce them by explosive synthesis at the moment of ejection, mix-
ing the contents of two internal chambers and catalyzing the reaction with two kinds
of enzymes. In the process, enough heat is generated to bring the resultant discharge
to the temperature of boiling water. Excellent marksmen, bombardiers can spray in
virtually any direction to accurately target any part of their body that is subjected to
assault. The combination of a burning and stinging spray and the audible popping
sound that accompanies the emissions is understandably effective against a wide
range of potential predators, including humans.
The active principles of most defensive secretions are highly volatile substances
of low molecular weight, usually strongly odorous and irritating and in some cases
even painful to inhale. Often, they are present in very high concentrations. How do
these arthropods withstand their own discharges? In many cases, external immunity
appears to be gained through possession of an especially impervious integument.
The cuticle of Hemiptera with defensive glands, for example, is generally imperme-
able to hemipteran secretion unless abrasion takes place. Internally, secretions do
Fig. 5.9 Taking aim. (left) A carpenter ant, Formica integer, responds to an alien in the nest by
bending its abdomen under its body and directionally spraying formic acid that both repels the
intruder and alerts nestmates. (right) A bombardier beetle aims its abdominal tip so accurately that
it can spray its hot quinone secretion in any direction
not come into contact with living tissue because the storage glands are membranous
saclike invaginations of the body wall, and thus are lined with cuticle as well.
But how can the glands produce the poisons without poisoning themselves?
Some arthropod gland cells possess certain more or less elaborate cuticular
chambers and ducts; toxicant synthesis occurs within the lumen of these cuticular
organelles, not in the cytoplasm of the living gland cells associated with them. Such
a situation has been shown in certain tenebrionid beetles. Other arthropods (like the
bombardiers mentioned above) have ‘reactor glands’ constructed in such a way that
chemical precursors of the secretion are mixed only at the moment of discharge.
Not all defensive/offensive secretions are purely chemical in effect. Many arthro-
pods manufacture sticky or slimy materials that mechanically hinder predators.
From their pointed cephalic nozzles Nasutitermes soldiers (Fig. 5.10) eject a sticky,
resinous terpene-containing secretion that dries quickly in the air. Mechanically
incapacitating or even killing insect predators, it also acts as an alarm substance
inducing other termites to converge upon the site. The nymphs of some cockroach
species have rear ends coated with a sticky secretion; when attacked by ants, they
sling slimy droplets over some distance. Although analytical studies have failed to
reveal any chemically active substances in this material, ants targeted by the fluid
are instantly incapacitated. As a final example of secretion taken to the extreme, the
workers of a tropical formicine ant protect their colony by literally blowing them-
selves apart. The large mandibular glands of these ants take up not only the head,
but much of the abdomen. When an ant is under assault, it compresses its abdomen
so rapidly that it bursts, casting the sticky contents of its mandibular glands in all
directions, effectively trapping attackers in the glue.
Fig. 5.10 Squirting string. Soldiers of Nasutitermes orienting to a predator. These termites pos-
sess a hollow nozzle at the front of their heads, from which they shoot strands of a sticky secretion
produced in their frontal glands. Attacking arthropods find the tacky filaments almost impossible
to remove; the material is also a potent irritant
5.4.2 Startle: ‘I’m Not What You Thought!’

When disturbed while at rest, the giant Io moth spreads its forewings, abruptly
exposing a pair of huge, white-centered black spots on the upper surface of its hind
wings. The sluggish Australian katydid Acripeza reticulata lives in leaf litter and
probably escapes predation because of its grayish, bumpy exterior; when disturbed,
females raise their wings to reveal a vivid abdominal color pattern of reds, blues,
and black. Crane flies drop off their own legs when attacked.
Passive ‘primary’ defenses such as disguise, aposematic coloration, and mimicry
are useful against enemies only to a point. Systemic chemical defenses can provide
a ramped up but still essentially passive second line of defense. In this section and
the next we consider some more active approaches based on the concept of unex-
pectedly confronting a potential predator with novel morphology or behavior so
unexpectedly that it causes the animal to hesitate or retreat. For example, some ful-
gorid bugs (Fig. 5.11) have lizard-like false heads; if a predator is still undeterred,
they suddenly expose large colored wings. Similarly, when pecked by a bird, the
large distasteful lubber grasshopper typically flips up its front wings while unfolding
and erecting its ruby-red hind wings.
To startle is to disconcert or frighten by a sudden shock or surprise. One com-
mon method of achieving a startle effect is through such flash coloration, the
phenomenon in which the prey is cryptic when at rest, but upon being threatened,
reveals previously hidden brilliant colors and/or patterns while escaping. Many star-
tle displays include conspicuous structures that are combined with other behaviors
that heighten the startle effect (Fig. 5.12). Flash markings are most common on the
wings of Lepidoptera but are also found occasionally among many other insects.
Early on, these circular marks were named eye spots, another example of the way in
which the label that is placed on something can affect the research questions that are
asked about it. Traditionally, eye spots have been thought to startle predators simply
because they look like the eyes of a vertebrate, perhaps one of the predator’s own
enemies. In this sense, they could be considered a special case of mimicry. Research
showed that the more motion-like the markings appear (by arrangement of colors in
concentric rings and/or by rhythmic shaking of the wings), the greater the startle
effect upon birds.
Some biologists have suggested that eye spots are mimicking owl eyes because
insectivorous birds may have an innate fear of these stimuli, owls being their preda-
tors. Three species of Caligo, the so-called owl butterfly, in Trinidad have enormous
eye spots (15–20 mm in diameter) on the under side of their hind wings; another
butterfly Eryphanis in Trinidad has smaller (6–7 mm), more irregular markings. At
rest, all four species hold their wings vertically so that only one side is visible to
an observer at a time, but the eye spots are continuously displayed. It may be that
Caligo are mimicking Hyla tree frogs (Fig. 5.13), while Eryphanis are mimicking
Anolis lizards, two vertebrate predators common in the resting microhabitat of the
butterflies. To birds viewing the butterflies from a distance, the eye spots do not
interfere unduly with the overall cryptic pattern. However, the lizards, which have
limited binocular vision, maintain territories based on body size; seeing Eryphanis
Fig. 5.11 Mimicking a predator? (center) Lanternaria, a large fulgorid bug from Panama, has
a hollow sham ‘head’ and lizard-like markings that cause it to resemble a small carnivore such as
a gecko (top). If this first line of defense fails, it flashes open its wings to reveal vivid eye spots
(bottom)
Fig. 5.12 Startling display. The mature larva of the tiger swallowtail, Papilio glaucus, responds
to danger by abruptly everting its forked defensive gland, the osmeterium, from beneath its neck
integument. The side nearest the danger tends to be extruded furthest; in some cases the caterpillar
may even arch its body and wipe the horns directly upon the offender. The osmeterium, which
contains butyric acid, is commonly colored bright yellow or red. Thus, the defense is probably a
combination of a visual threat and a repellent odor
Fig. 5.13 Imagining a resemblance? Apparent mimicry of arboreal carnivores by a large but-
terfly, Caligo, that rests by day on tree trunks. (a) The mimetic pattern on the butterfly’s wings is
said to include not only the eyes but the head profile, pectoral region, and amphibian typanum. (b)
The supposed model is the tree frog Hyla crepitans, a widespread predator that also spends the day
resting on tree trunks
at close range as larger rivals of their own species, the lizards avoid them. At the
same time, the larger Caligo resemble full-grown tree frogs that are quite capable of
eating lizards, and thus the lizards avoid them as well.
Another hypothesis, not mutually exclusive, is that eye spots may distract or
redirect a predator’s attention from more vulnerable body parts. Many of the small
lycaenid butterflies commonly known as ‘blues’ have eye spots on the underside
of their hind wings and, close to them, long wing ‘tails’ that somewhat resemble
antennae. It has been suggested that these ‘tails’, which are kept in motion by the
resting butterfly, draw attack away from the true head where a wound would be fatal.
Attacking the eye spot instead, a predator gains only a piece of wing for its efforts.
It is certainly true that for most vertebrates, ourselves included, an eye spot has an
inherent conspicuousness that is difficult to ignore (Fig. 5.14).
Many cryptic insects use flash coloration in reverse, making normal camouflage
more effective through sudden contrasts. Thus, for example, many grasshoppers that
are colored to match the stony or sandy soil when they settle have hind wings of bril-
liant red, blue, or yellow. The grasshoppers are highly conspicuous in flight, but they
conceal their hind wings so abruptly upon landing that they appear to vanish. A wide
variety of other startling display behaviors have also evolved in otherwise cryptic
insects. Some orthopterans perform stereotyped defensive displays when discovered
(Fig. 5.15), raising their long spiny legs in the air over their heads. Various moths
react like whirling dervishes, dramatically flapping their wings about and rocking
from side to side, thus warming up their flight muscles while temporarily startling
their predator. In this way they can often escape before the predator is able to react.
Fig. 5.14 No ignoring it! This diagram illustrates the inherent conspicuousness of an eye spot,
which attracts attention to itself in preference to a variety of other, and even larger, objects in the
visual field
Fig. 5.15 Flailing in defense. Display in two normally cryptic orthopterans. (a) The weta
Deinacrida heteracantha from New Zealand. (b) The katydid Neobarettia spinosa from Mexico
The strategy of startling a potential predator includes another very dramatic and
widespread behavior. In the same way that a fleeing escapee might slip free from
a jacket grasped by his assailant, many insects elude potential predators by leav-
ing behind various dispensable body coverings, a phenomenon known as autotomy.
Perhaps the most familiar of these are the easily shed scales which cover the wings
of butterflies and moths; the loose hairs of adult caddisflies and the scales of sil-
verfish and bristletails may be similar protections against entrapment. Legs of crane
flies represent a more extreme example of dispensable body parts. The larvae of cer-
tain carpet beetles use interlocking hairs for defense; on their abdomens, they bear
prominent tufts of barbed setae that easily detach and hopelessly entangle small
predators such as ants.
Detachable coverings need not be strictly morphological. Gathering on their back
a variety of debris from feces to old cuticle to the sucked-out remains of prey, larvae
of groups as diverse as assassin bugs and lacewings construct shields for protection
and camouflage. Other species are able to maneuver their shield with considerable
agility against predators, in actions that sometimes cross over from defense to attack
(see Fig. 5.8).
5.4.3 Group Actions: ‘We’re in This Together!’

Some Nasutitermes termites make daylight forays above ground in true military
fashion; columns of workers are guarded against arthropod enemy trespass by
flanking rows of soldiers oriented with their spraying snouts pointed outward (see
Fig. 5.10). If its nest is tapped, the mound-building ant, Formica rufa, rushes out
and hundreds of workers eject their formic acid spray toward the sound of the
disturbance.
One of the most-often cited selective advantages of living in a group is protection

against predators. There are several ways that an individual might benefit from not
being alone. One is the encounter effect, which affects the probability that a predator
will locate a group. As group size increases, the rate that groups encounter predators
also increases, but not as quickly as expected, probably because the apparent group
size rises more slowly than actual size. (That is, a group with 1,000 individuals does
not look twice as big as one with 500 individuals, because some are hidden behind
others.) An individual’s predation risk is also lowered by a second factor, called the
dilution effect. This is basically the idea of ‘safety in numbers’; if a predator is going
to eat only one prey, an individual is better in a large group than alone.
An individual’s position within the group also affects its predation risk, of course.
In 1971, a compelling explanation was offered by W. D. Hamilton, one of the pio-
neers in thinking about evolution as a game between competing phenotypes. Some
early ethologists had felt that natural selection at the population level was neces-
sary for aggregations and social groups to form; cooperative behavior had to evolve
‘for the good of the group’ rather than for the good of the individual. Hamilton
argued that rather than cooperating with one another to form a group, under some
conditions a population of prey animals was simply acting as a selfish herd, in
which all the individuals were trying to use each other as shields against preda-
tors. Because individuals at the outside of the group were most likely to be picked
off, individuals—looking out for their own self-interest—were competing with one
another for the safer positions toward the middle. Thus, while the herd looked like a
unit moving together, it actually emerged from the uncoordinated behavior of self-
serving individuals. However, even though being in the center may reduce the risk
of predation, there often is a trade-off in the form of a lowered food supply.
Defense through group action is particularly well developed in gregarious and
social insects. The use of chemical communicating substances for warning pur-
poses (such as inter- and intraspecific alarm odors) is widespread, but is especially
well developed among ants. When such a substance is of sufficient quantity, it trig-
gers rapid dispersal of nearby individuals. Highly volatile, it fades rapidly unless
reinforced, thereby minimizing overreaction.
Young offspring are often especially vulnerable to predators, so it is probably
no surprise to find that in many species, ‘larvel herds’ (see Plate 37) and group
living appear to be an adaptation to protect them. (We’ll be returning to this sub-
ject in Chapters 9 and 10.) For example, females of the bug Elasmucha grisea
defend their eggs and nymphs by shielding them with their body and lunging aggres-
sively at intruders. Often two or more females will guard clutches side by side
with one another on a leaf. In experimental pairings, jointly guarding females lost
significantly fewer eggs to predatory ants than single females did.
When disturbed, swarms of a wide variety of insects seem to ‘explode’ into a
mass of erratically swirling individuals. For species that regularly aggregate for
various purposes, such group ‘panic’ can be a good defense strategy. The unpre-
dictable random flight of fleeing animals may stymie a predator more effectively
than an orderly retreat would. Many predators must fixate upon one individual to
successfully attack, and this is difficult to do amid erratic movements undertaken by
a number of closely spaced individuals. Such defensive behavior probably also has
elements of a mass startle effect.
Mobbing, a type of group defense in which a predator is harassed by a number
of maneuverable prey individuals, has long been known among colonially nest-
ing birds. Many social and/or communal insects have also developed this effective
behavior. Bembix sand wasps, for example, nest in aggregations; when a potential
predator approaches, a score of male wasps may fly up at it, buzzing ominously
and loudly but harmlessly. While mobbing, prey animals may approach, threaten,
or occasionally physically attack a predator. Often, even though the predator is not
actually injured, it moves off. This is probably not only due to the harassment but
also to the fact that its probability of successfully capturing a prey is very low amidst
all the commotion.
Chapter 6
Chemical Communication
6.1 Introduction
Over a century ago, the French naturalist Jean-Henri Fabré confined an unmated
female giant peacock moth in a container hidden on his desk; the following evening,
over 40 male moths arrived, in Fabré’s words, ‘eager to pay their respects to their
marriageable bride born that morning.’ The container was not airtight, and because
even the empty container was of intense interest to the attracted males, Fabré sug-
gested that the attraction must be chemical, but he was puzzled because he could
not smell anything emanating from his female moth.
Communication through chemistry has been found to be a persistent theme
in almost all animal taxa. Chemical communication is both the oldest and the
dominant method of communication among organisms; early metazoans relied
extensively on chemical communication for timing gamete release and to mediate
the union of gametes.
Biologically significant chemicals have been found in both sexes and all life
stages of many species, not just of insects but spanning the zoological universe.
These are chemical hotlines, not polite conversational perfumes. They communi-
cate matters vital to survival, such as the presence of danger, the identification of
friends and foes, the call to arms or emigration, the availability of food, and the urge
to reproduce.
Today the role of odors in animal communication is so well established that it
is difficult to appreciate that it has been only relatively recently that advances in
chemistry and technology have permitted humans to eavesdrop on the private lives
of insects and other organisms where chemical communication is often the dominant
mode. To date, the largest number of communication chemicals in the animal world
has been found among the members of the largest class of animals, the insects. It
is here that they have been studied longest and here that science knows most about
their modes of action.
6.2 Mechanisms of Chemical Communication

It is a hazy summer morning just before dawn, and a man follows a footpath to a
clearing in the middle of an apple orchard. Reaching up, he removes the tray that

218 6 Chemical Communication
hangs from a wooden post and smiles with satisfaction: the sticky tray top is covered
with small rust-and-yellow moths. As he starts back toward his pickup truck, the
man pauses to watch a flying moth. It zigzags, comes closer, then hovers over the
tray, extending its genital claspers. As its wingtips brush the cardboard and are
caught by the gummy surface, its abdomen curves upward, throbbing convulsively
as the moth tries to mate with a polyethylene disc in the center of the tray.
What has happened here? The scene was a carefully chosen test site, the man
a field entomologist; the moths were male red-banded leaf rollers (Argyrotaenia
velutinana). The plastic disc was dipped in a compound that smelled like a secretion
produced by a virgin female red-banded leaf roller. Male moths were lured by the
smell to the trap where, instead of mating, these destructive apple orchard pests
were captured.
The observation that odors secreted by some insects function as stimulants, incit-
ing those of the opposite sex to mate seems, rather obvious today. In fact, the
German medical doctor and botanist Philipp Franz von Siebold proposed such an
idea about 170 years ago. But during the 1800s the phenomenon of chemical com-
munication was largely ignored or disbelieved. Even Fabré, who easily accepted the
idea that insects could detect the odors of other insects, could not believe that such
odors could operate over such long distances; it simply strained credibility to think
that a single female moth could inject into the vast atmosphere enough of a sub-
stance to be perceived by males kilometers away. This would be, he thought, like
tinting an entire lake with a single drop of dye.
6.2.1 Odor Creation and Reception
Where do insect communication chemicals originate? How are they received? Do

insects smell their chemical messages or taste them? In 1900 an American biolo-
gist, A. G. Mayer, undertook some simple experiments to establish the mechanism
by which male promethea moths (Hyalophora promethea) are attracted to females.
When he placed five female moths in a glass jar topped with mosquito netting,
Mayer noted that male promethea moths over 100 ft away were immediately
attracted. Next, he inverted the jar and packed sand around its mouth so that air could
not escape; now males were no longer attracted. So perhaps smell was involved; was
sight also involved? Mayer wrapped female moths loosely in cotton to make them
invisible; males not only came, they grasped the cotton in their abdominal claspers in
typical copulatory attempts. On another set of moths, Mayer replaced female wings
with male wings; males still came and mated without hesitation. Having ruled out
the role of sight in male attraction, he performed one more experiment to establish
that only chemical cues were involved. He set up a small wooden box containing
females, so that air blown into the box came out through a small chimney. True
to prediction, the males were attracted not to the box but to the chimney top, even
when the vicinity contained fumes of carbon disulfide and diethyl sulfide.
Where were the moth’s odors being emitted? Mayer severed female moth bod-
ies and placed the pieces in various locations. Males released five feet away flew
6.2 Mechanisms of Chemical Communication 219
to the abdomens and ignored the remainder. How were the odors being received?
Males with their abdomens cut off still responded; so did males whose spiracles
were covered with glue. However, males whose antennae were gummed up did not
seek females. In fact, they showed no excitement even when held within 2.5 cm of
virgin females.
Insects produce their odors in many ways; this variety and complexity is par-
ticularly striking among the social species (Fig. 6.1). Chemicals come not just
from the abdomen but often from the head or thorax, produced by exocrine glands
(clusters of secretory cells whose products are discharged to the outside of the
body). Sometimes, single chemicals are produced, but more commonly they are
complex blends.
Once produced, the chemicals are sent forth as liquid streams, droplets, thin
films, aerosols, or gases. Emission rates and concentrations may be controlled
through adjustable nozzles, retracting applicators, evaporation pads, or other equally
elaborate devices. Among the more spectacular are the male arctiid moths that pos-
sess feather-like coremata that can be pumped in and out to waft their chemicals
into the evening air (see Plate 16).
Producing chemicals is only part of the picture, of course; receiving and reacting
to them is the other part. Just as with vertebrate animals, one can say that insects
both taste and smell. Also like in most vertebrates, the chemosensory system of
insects seems to rely far more heavily upon smell than on taste. This is probably
no surprise. Taste involves the reception of waterborne compounds with a limited
range of qualities, often in fairly high concentrations, at close range. Odor, on the
other hand, involves reception of chemicals over a greater range and often at much
lower concentrations; though these chemicals are usually in gaseous form, they may
also be either airborne or waterborne. Because the two senses are so closely allied,
rather than considering them separately, many biologists simply speak of insect
chemoproduction and chemoreception.
How are tastes and smells received? The external surface of insects is cov-
ered in cuticle, so it is the place to begin. On various structures, particularly the
antennae, mouthparts, legs and ovipositor, chemoreceptor cells provide ‘windows’
to the outside world. These are surrounded by the characteristic cuticular sensilla
that comprise the most obvious external parts of every chemosensory organ. They
occur in at least four morphological forms: bristles or hairs, pegs, plates, and pits
(Fig. 6.2).
All chemosensory sensilla have one or more pores through which chemicals can
pass. Gustatory and olfactory sensilla are quite similar structurally, but the former
have only a single pore and the latter have many. Once an odor molecule passes
through the pore(s) it binds with specific proteins, triggering a cascade of reac-
tions that ferry it through the fluid media of the sensillum lumen to the dendrite
of a nerve where it activates receptors and is changed into electrical action poten-
tials that travel along axons to the brain. Axons from gustatory sensilla on the head
lead to the subesophagael ganglion; those from olfactory sensilla terminate in the
deutocerebrum. Sensory filtering occurs at multiple levels in the nervous system,
but initial specificity and sensitivity is achieved by the odorant binding proteins.
Fig. 6.1 Producing

communication odors. The
highly developed exocrine
gland system of wasps, bees,
ants, and termites showing
locations of glands most
commonly used in behavioral
interactions. All the glands
shown secrete chemicals to
the outside. Some, such as the
mandibular gland, have
storage reservoirs and release
their chemicals in bursts as
needed; others secrete
chemicals more or less
continuously
Fig. 6.2 Perceiving odors. Scanning electron micrograph (3,900×), showing part of the surface
of segment 10 of a worker honey bee’s antennal flagellum. Four types of sensilla are evident: sf =
sensilla companiformia, sp = sensilla placodea, sb = sensilla basiconica, II = sensilla trichodea.
The arrow points to the pore at the tip of one sensillum. A single antenna may have more than
8,000 sensilla trichodea alone. The sensillae companiformia are likely contact chemoreceptors
To function, the system also requires a rapid enzymatic breakdown of the odorant
and other chemical flotsam that may impact it over time. The process by which the
receptor’s active site(s) are cleared, renewing its sensitivity to incoming molecules,
must happen very quickly (on the order of milliseconds) to permit neural integra-
tion of the rapid directional changes that are observed during an insect’s movements
along an odor gradient. This striking behavior depends on antennae, the best-studied
chemosensory receptors of insects. Use of two identical sets of receptors proba-
bly serves to maximize the sensitivity and efficiency of the system. Moths flying
upwind in an odor plume and ants following an odor trail characteristically trace a
zigzag route in which each antenna appears to be moved alternately in and out of the
odor field. The ‘out’ antenna suddenly stops sending signals to the brain; as though
seeking to restore a balance of input from the two sides, the brain causes a steering
change which soon results in overcompensation toward the opposite side, and so on.
If both antennae remained continuously in the odor field, habituation might quickly
occur. By passing in and out of the threshold concentration level, maximal sensitiv-
ity to the stimulus would be likely to persist. In addition, such a system imparts a
clear orientation axis to the signal path. Orientation experiments using ants or bees
with amputated or crossed antennae lend support to these ideas (Fig. 6.3).
The sensitivity of many insects to odor molecules is further enhanced by antenna
design, as is evident from a brief glance at one well-studied pair of antennae, those
of the male silkworm moth (Fig. 6.4). It appears that the spacing and arrangement
of the antennal hairs allows them to act like molecular sieves; the width of the mesh
they form is so small that chemical molecules, because of their fast thermal move-
ments, cannot pass through without contacting the hairs and being preferentially
absorbed. The physics of airflow over structures this small obviously also must fac-
tor into their design. The sensitivity of male gypsy moths is likewise so great that
scientists estimate that the 30 g of disparlure, a synthetic gypsy moth sex attrac-
tant, already on hand will be enough to bait some 60,000 traps per year for the next
50,000 years.
Fig. 6.3 Following a chemical trail. Trail following by Lasius worker ants with normal and
altered antennae after trail odor was applied along the straight line. (a) Normal trail following, a
tropotaxis in which the ant zigzags evenly first to one side then to the other as its paired antennal
receptors alternately move in and out of the trail’s vapor space and perceive a diffusion gradient.
Each time an antenna leaves the odor tunnel, the ant swings back to the opposite side. (b) With
its left antenna amputated, the ant repeatedly overcorrects to the right side. (c) With its antennae
crossed and glued, the ant is disoriented and relocates the trail with difficulty; its overall progress
in the proper direction is probably mediated by visual cues
Responses of a single olfactory sensillum can be measured electrophysiologically

using microelectrode probes. Studies of whole antenna responses employ an elec-
troantennogram (EAG) that measures the summed receptor potentials in the whole
antenna when an odor puff is passed over it. Signals from the antenna are amplified
and displayed on a screen. Like most sensory neurons, antennal chemoreceptors
are spontaneously active and display a constant low level of firing or background
Fig. 6.4 Finding females. In response to the odor of a female, the male silkworm moth Bombyx
mori begins fluttering its wings at 40-50Hz. Males of the commercial strain, having lost the abil-
ity to fly, walk toward the odor source while continually vibrating their wings, an action that
increases the flow of air over the antennae about 15 times more than just walking. The male’s
plumose antennae each contain 60–70 branches. Each branch is in turn covered with about 17,000
olfactory hairs that average 100 μ in length. About half of these receptors are tuned to detect
the female’s sex pheromone, whereas the antennae of the female do not react to this scent at
all. Electrophysiological recordings indicate that when a single biologically relevant molecule
strikes a single receptor, it elicits a detectable response. However, about 200 molecules must arrive
simultaneously before the male begins an upwind searching movement
noise. For reliable efficiency, some of the better man-made electronic detection sys-
tems currently require input about twice the background noise (a signal-to-noise
ratio of about 2:1). In contrast, the male silkworm moth shows a positive behavioral
response at the incredibly low signal-to-noise ratio of 0.125:1.
While these techniques are useful for screening potential attractant and deter-
rent chemicals, full behavior-based laboratory and field trials are needed before a
chemical can be conclusively said to be communicatory (Fig. 6.5), because physio-
logical state, ecological context, and nervous integration are all intimately involved
in behavioral activation.
6.2.2 Communication Through Chemistry
The story of Fabré and his peacock moths is generally accepted as the first defini-
tive evidence of communication by chemistry, but almost a century passed before
microanalytical techniques reached the point in development that would permit
identification of the minute quantities of chemical that were involved. Beginning in
the period after World War II, however, the field of insect chemical communication
grew quickly. As knowledge about chemical communication systems has grown,
the subject has rapidly become more complex, and so has the terminology to deal
with it.
Fig. 6.5 Proving chemical communication. Example of the procedure leading to identification of
the aggregation attractant of the bark beetle Ips pini. The stages involve the collection, purification,
identification, synthesis, and field testing. At nearly every stage appropriately designed behavioral
bioassays are essential
In 1959, the German chemists Peter Karlson and Adolf Butenandt and the Swiss
zoologist Martin Lüscher proposed a new name for some of these messenger chem-
icals to replace the contradictory term ‘ectohormone’ then in use. From the Greek
pherein, to carry, and horman, to excite, they derived the word pheromone—a sub-
stance secreted by an animal that affects the behavior of other animals of the same
species. Initially it was thought that a pheromone would be a single chemical, but
as detection and bioassay methods became more sophisticated, most pheromones
proved to be blends of two or more components (see below) that are typically active
in very low concentrations. That same year, the first pheromone was chemically iso-
lated and identified—the sex attractant of the commercial silkworm, Bombyx mori
(Case Study 6.1).
Case Study 6.1: Identification of the Mate Attractant of the

Silkworm Moth, Bombyx mori
Today powerful new devices are routine in the chemist’s toolbox, revolutioniz-
ing the study of animal and plant odors, and a single insect may be sufficient
to identify and characterize an emitted odorant molecule. However, the first
time took 30 years and over 500,00 moths.
Bombyx mori is the most famous caterpillar in the world—the silkworm
whose cocoon, spun from a single strand 500 to 1,300 yards long, furnishes
the material for a thread and cloth not yet duplicated by any synthetic fiber.
The only truly domesticated insect, the silkworm moth can no longer maintain
itself in a natural environment but survives only under cultivation, where most
individuals live only until the pupal stage, then are plunged into boiling water
and their cocoons unraveled. However, in each generation a few moths are
allowed to survive to furry, greenish-white adulthood to furnish eggs for new
progeny. Neither sex can fly, and therefore the male cannot easily scout the
terrain to find a mate. Yet it had long been known that female Bombyx in
some manner attract males from extraordinary distances (see Fig. 6.4).
In 1939, the German chemist Adolf Butenandt, who had previously gained
fame for his discovery of the identity of human sex hormones, reasoned that
a biochemical lure in the bodies of virgin females was probably responsi-
ble. Although his available tools would now be regarded as primitive—he
lacked such now-standard equipment as the gas chromatograph—he began to
work on the isolation and identification of the substance. By 1959, Butenandt
and his associates had processed half a million female silkworm moths and
had extracted from the abdomen tips just twelve-thousandths of a gram of a
derivative of the active compound.
Next, the researchers used a combination of chromatography, infrared and
ultraviolet spectroscopy, and chemical structural analysis coupled with bio-
logical assays to see which fractions of this derivative elicited male wing
fanning. They obtained a substance which, in minute quantities, was as attrac-
tive to male silkworm moths as the most seductive virgin female. It was a
primary alcohol with the formula E-10,Z12-hexadecadien-1-ol, which they

named ‘bombykol’. It was the first sex attractant identified in the Lepidoptera
and the first to be chemically characterized.
Within a few years after Butenandt’s remarkable feat, Edward O. Wilson and
William H. Bossert proposed dividing pheromones into two functional groups
according to their mode of influence: releasers and primers. Releaser pheromones
stimulate an immediate and reversible behavioral response mediated wholly by
the nervous system; these pheromones are thus by definition chemical ‘releasers’
in the terminology of the ethologist (see Chapter 2). They are widespread in
insects and serve a great many functions, sex attraction and alarm being especially
important ones.
Primer pheromones, on the other hand, act to physiologically alter the endocrine
and reproductive systems of the receptor animal, re-programming it for an altered
response pattern. In a sense, the receptor’s body is ‘primed’ for new biological
activity, although such activity may not appear until some future time and may
require triggering by another releaser pheromone. Primers are best known in social
insects where they mediate a range of social interactions; in a highly populous
colony where it would be impossible for the queen to physically control the work-
ers, they have obvious adaptive significance. Primer pheromones are also important
in desert locusts, mediating maturation and phase change that facilitates swarm
synchrony.
Other attention turned to substances that transmit external chemical messages
that affect individuals or populations of a species different from their source. These
were termed allelochemicals (Greek: allelon, of one another). Allelochemicals may
affect the growth, health, behavior, or population biology of other species. Two
categories of allelochemicals have been recognized: allomones and kairomones.
Allomones are chemical agents of adaptive advantage to the organism sending them;
kairomones (Greek: kairos, opportunistic), on the other hand, are of adaptive value
to the organism receiving them.
All pheromones fall under the broader umbrella classification of semiochem-
icals or infochemicals. In short, an infochemical is any chemical produced by
one organism that incites a response in another organism. Figure 6.6 puts some
semiochemicals into their behavioral context.
Even with so much terminology, some chemicals that affect behavior do not fit
neatly; for example, compounds that are released by fermenting and decaying fruit
attract parasitoid wasps that attack the larvae of certain fruit flies. A number of
other categories have since been proposed, but classifying the diverse responses of
organisms to chemicals does not necessarily enhance understanding. Furthermore,
the same compound may sometimes have multiple roles depending upon context,
and, as we have noted before, the very labeling of behavioral phenomena tends to
color subsequent interpretations of it.
6.3 The Functions of Chemical Communication 227
Fig. 6.6 Classifying semiochemicals. Pheromone releaser effects are classical stimulus-response
reactions mediated wholly by the central nervous system. With primer effects, behavior is usually
induced not by the pheromone but by later external stimuli for which the pheromone has ‘primed’
the animal’s physiology
Nevertheless, these basic categories are broadly useful. Allomones tend to

include compounds having roles in aggressive relationships, such as defensive secre-
tions and repellents (see Chapter 5) and odors mediating mutualistic relationships
such as flower scents that attract pollinators (see Chapters 4 and 9). Kairomones
include host location cues such as natural plant products that attract herbivorous
insects (see Chapter 4) and prey scents that attract predators and/or parasites. Note
that from the standpoint of selection, kairomones must be viewed as evolutionary
‘backfires’ for the emitting organism; they were originally selected to serve either
pheromonal or allomonal functions of benefit to the emitter, but secondarily they
have come to be used by other organisms (herbivores, predators, parasitoids) to its
detriment.
6.3 The Functions of Chemical Communication
A pair of bright orange patterned butterflies cavort in the air above a patch of gold-
enrod flowers, performing loops and tight rolls, one occasionally hovering briefly
above the other, rising and falling on invisible air currents. Below them on the
ground several ants scurry along an invisible line weaving among the plants. Close
examination of the goldenrod reveals a group of yellow aphids along some of the
stems, some of whom have glistening droplets of liquid on the tips of their corni-
cles; near them a black spotted lady bug is contently munching on one less fortunate
group member. What do these different observations have in common? All are
examples of chemical communication that serve different purposes. In the following

sections, we’ll be examining some of these purposes in more detail.
Any time an organism produces a signal, there is at least potentially an intended
receiver. However, keep in mind that signals also can be intercepted for use by a
variety of illegitimate receivers, for example, predators, parasites, and competitors.
Furthermore, even for the intended receiver, the message that the communicator
encodes may differ somewhat from the meaning that the event might have for the
recipient. Finally, in interpreting communicative emissions one must be careful to
distinguish functions from incidental effects. Functions must increase the fitness of
the sender in order to evolve; the same does not necessarily hold true for incidental
effects. For example, pregnant female rabbits release hormones into the circulatory
system that, when picked up by feeding fleas, stimulate flea reproduction. The func-
tion of the hormone—rabbit reproduction—is positive for the rabbit; the effect—an
increased flea population—surely is not.
6.3.1 Finding and Choosing Mates

Dance flies begin a tap-dancing courtship that includes a chemical exchange. The
phantom midge Chaeborus grooms elaborate plumed antennae as he searches for
cues that a female is nearby. Using pressure from its internal body fluids, a moth
everts its abdominal scent brushes (see Plate 16).
Charles Darwin was well aware of the importance of chemical sexual signals,
and in proposing his theory of sexual selection, he suggested that sexually selected
signals would have many features in common, including the elaboration or expres-
sion of the signals in only one sex, development only in adults, often only in the
breeding seasons, and use primarily or exclusively in mating. Chemical signals, and
the scent glands that create them, commonly show all these characteristics.
Sexual selection can take many forms. The most fully studied are mate choice,
in which members of one sex display to be chosen by the other sex, and ‘contests’
in which members of the same sex compete with one another to be the ‘dominant’
individual that will have exclusive or preferential mating choices. Both of these
aspects are recurring themes in the area of chemical communication, as throughout
all of biology.
Species recognition, mate recognition, and mate assessment are all non-exclusive
functions of sex pheromones. Most often, however, the scenario begins with a
female insect producing pheromones to which males respond, often over extraor-
dinary distances (see Case Study 6.1). Such pheromones exist in a concentration
gradient, and sometimes the same compound calls in the opposite sex from a
distance at low concentrations and elicits courtship behavior at the higher concentra-
tions encountered at close range. Alternatively different components of a pheromone
blend come into play at different points in the attraction process. During pheromone
release, certain stereotyped postures called ‘calling behaviors’ are common among
a wide variety of unrelated insects (Fig. 6.7).
Fig. 6.7 Calling males from a distance. (a) Position adopted by a female Indian meal moth,
Plodia interpunctella, during sex pheromone release. During calling the abdominal tip is elevated
and glands in the intersegmental membrane between the eighth and ninth abdominal segments
are extruded. A similar stance is taken by most calling moths as well as by a number of other
unrelated insects. (b) A wingless thynnid wasp female releases her sex attractant in a characteristic
‘calling’ position atop a grass blade. However, competing for the winged male’s attention is the
flower of an orchid, Drakea (at left) that deceptively signals with the same odor and ‘feel’ as
the female wasp. Male wasps are poor discriminators visiting several orchid flowers that they
pollinate
A well-known prediction from sexual selection theory is that the limiting sex
(almost always, the female) should put less effort or risk into display and attraction,
not more, than the less-limited sex. In other chapters, we have noted that males are
the producers of most visual or acoustic calling signals, as this theory would pre-
dict. So why should it be the female insect that produces these pheromone signals?
There are some reasonable non-exclusive ‘arm-chair theories’ but few experimental
studies to address this question. Signals in the chemical sphere may be energetically
less expensive than producing sounds or active visual displays, and they may involve
less risk of predators homing in on their source. It is also quite likely that males are
actually carrying the larger burden of risk because their flight is energetically costly
and exposes them to a higher risk of predation than the generally hidden, calling
female.
It would be overly simplistic, however, to think of insect sexual attraction as
simply ‘females signal, males respond.’ In many cases, odors also come into play
once the sexes have been drawn within range of each other by other cues. They may
be produced by either sex, but are usually emitted by the male and often as only part
of a complex pattern of courtship behavior.
These male-emitted chemicals are delivered in several ways. The male monarch
butterfly and its relatives display one type of system (Fig. 6.8); hairpencils are
extrusible organs that function as tiny scent-filled brushes, wafting pheromone-rich
clouds at close range over the female’s antennae. Similar structures in male moths
are called coremata. In some butterflies, special glandular scales called androconia
occur on the male’s wings. A courting male of the grayling butterfly (Hipparchia
Fig. 6.8 Delivering perfumes to nearby females. Paired extrusible scent organs (hairpencils) at
the tip of the abdomen of a hand-held male Lycorea ceres (Nymphalidae, Danainae), fully splayed
open
semele) will actually clasp the female’s antennae between his wings, bringing them
into contact with these scent scales.
From tergal glands on their abdomens, several species of male cockroaches
produce secretions on which the females feed prior to copulation (Fig. 6.9); the
substance has been given the whimsical name ‘seducin’. Other male-produced
Fig. 6.9 Maneuvering with lures. A courting male cockroach, Nauphoeta cinerea, raises his
wings to offer secretions from his tergal gland secretion to a potential mate. When a female crawls
upon his back to feed she is in the proper position for the male to connect his genitalia with hers,
and if successful they will copulate in an end-to-end position
chemically diverse substances even include secretions passed to the female with
sperm, which have the potential to affect many aspects of female behavior and
physiology.
Only a few insects reverse roles from the very beginning, with the males pro-
ducing long distance attractants. Most (but by no means all) of these are beetles.
Male callers tend to occur in one of two situations. The first is with insects that
gather in courting associations called leks (see Chapter 9); these males often use
pheromones to attract females, especially among tephritid fruit flies, anthophorid
bees, and various ants and social and solitary wasps.
The other situation is when males locate patchy resources needed by females
and call, as occurs with male papaya fruit flies that call from fruits that will be
oviposition sites. Sometimes, especially among beetles, both sexes may be attracted
by the same scent; at an appropriate location such as a suitable host plant or other
resource, they congregate and mate. Even in cases in which the original function of
the pheromones involved is thought to have been sexual attraction, when both sexes
are attracted these chemicals are generally called aggregation pheromones.
One or both sexes of some insect species use cuticular hydrocarbons as sex
pheromones. Cuticular hydrocarbons are relatively involatile molecules that func-
tion at very close range or require contact for detection. They often play fundamental
roles in species identification. In Drosophila fruit flies they are important in
maintaining reproductive isolation between sibling species. When a courting D.
melanogaster male taps a prospective mate with his front legs he receives con-
tact pheromone signals via chemoreceptors on his tarsi; if it is the correct cuticular
hydrocarbon for females of his species, he vibrates his wings and courtship proceeds
to the next stage.
Unsurprisingly, the majority of the research on insect sex pheromones has been
economically motivated and practically oriented. The specificity of such com-
pounds makes it possible in theory to single out one particular pest species for
detection, monitoring, and/or control. Manipulation of insect behavior by means
of sex pheromones is a promising alternative to the use of conventional chemical
insecticides, although in actual practice it often has been fraught with technical
problems.
6.3.2 Assembly, Aggregation, and Recruitment

Quietly and quickly lift a pile of logs and you may see a large cluster of crickets
beneath it. Look at garden plants and notice growing tips completely obscured by
massed aphids or plant lice. Peer into the attic of a drafty home in late fall and
observe clusters of lady bugs in the corners. Encounter a swarm of mosquitoes, a
disconcerting experience; likewise for a honeybee swarm.
A wide variety of arthropods assemble for various purposes. In the previous sec-
tion we considered sex pheromones, which attract only the opposite sex. Now, let
us consider pheromones that attract both sexes and/or possibly even larvae.
In its purest sense, assembly refers to the process by which both sexes of a species
come together prior to some activity, such as feeding, mating, or hibernation; almost
always, the signals that call them together are unrelated in any direct manner to the
subsequent activity. Aggregation, the crowding together of individuals, is assembly’s
result. Usually, except in the social insects, aggregations are temporary groupings
and pheromones are only one of several ways in which they are promoted and/or
maintained. Among those insects possessing aposematic coloration (see Chapter
5), aggregations induced by pheromones are particularly prevalent. For example,
a male-produced pheromone attracts brightly colored lycid beetles of both sexes
to form prominent clusters, sending potential predators very conspicuous adver-
tisements of their distasteful nature, as do the large hibernation aggregations of
distasteful ladybird beetles (see Plate 5). Aggregations and other nonsexual asso-
ciations are classified in Table 10.1, and the costs and benefits of such behavior are
discussed in Chapter 10.
In conservation biology, aggregation size assumes particular importance in the
context of threatened and endangered species because under-population has a num-
ber of serious consequences. Collectively these consequences are broadly referred
to as Allee effects, after the pioneering animal behaviorist, Warner Clyde Allee,
who first elucidated them in detail. Allee effects are (broadly) defined as a decline
in individual fitness at low population size or density. If group size falls too low,
the reduction or loss of behavioral interactions between individuals can constitute a
severe threat to a species’ survival or even doom its very existence. (The extinction
of the passenger pigeon is a noteworthy example.)
One recurrent benefit of aggregations is to permit more effective exploitation
of the environment than would be possible for single individuals. For example,
the caterpillar-like larvae of the sawfly, Neodiprion pratti banksianae, feed in tight
groups upon jack-pine trees. When young larvae were experimentally isolated from
their companions they suffered 80% mortality, whereas among those allowed to
remain in groups, only 53% died. Why? Newly hatched larvae have considerable
difficulty chewing holes into the tough cuticle of the jack-pine needles. Each larva,
even in groups, individually attempts to establish its own feeding site. When finally
one cuts through into the inner tissues successfully, other larvae are quickly attracted
to the cut, where their feeding widens the breach still further until soon all the larvae
are able to feed. Another benefit, also illustrated well by Neodiprion, is enhanced
defense (see Plate 17).
In still other cases, gregarious behavior clearly functions to bring the sexes
together for mating, sometimes in combination with attraction to suitable host
resources and oviposition sites. Bark beetles are a good example (see Case
Study 6.2).
Case Study 6.2: Aggregation in a Bark Beetle, Dendroctonus
Small cylindrical scolytid bark beetles destroy millions of board feet of stand-
ing timber each year. Exactly how do they manage recruitment for their mass
attacks?
Some of the most destructive of pests to coniferous forest in North America

are the scolytids (now considered a subfamily of the Curculionidae), Their
attack pattern is quite predictable. First, a small pioneer group selects a tree,
usually one damaged in some way. They initiate an invasion and construct
nuptial chambers beneath the bark. Shortly thereafter, an army of attackers of
both sexes rapidly converges upon the site to continue the attack. By the time
mating has taken place, the number of arriving beetles rapidly declines. But
as a result of the mass attack, the tree usually dies.
Because of their economic importance, many studies have been conducted
on a great many aspects of bark beetle behavior. In 1969, Julius Rudinsky
puzzled over one component of the typical attack pattern of some scolytid
species—after several hours of mass invasion of the host tree, the beetles’
attraction stops very abruptly. It seemed unlikely that this behavioral change
was due only to a cessation of pheromone production. It was too sudden, and
Rudinsky knew that female frass may remain attractive to male beetles for
days. He also knew that test logs containing only male-female pairs in the
bark were not attractive to flying beetles.
Looking for an explanation, Rudinsky was struck by the well-known ability
of many scolytids to make squeaky sounds (stridulation, see Chapter 8). In
general, this sound-producing ability is more characteristic of one sex, the
one that is not the original invader. Might this yield an answer?
Through a series of intriguing field experiments, Rudinsky set out to exam-
ine the exact role of stridulation in the Douglas fir beetle, Dendroctonus
pseudotsugae. First, he artificially infested a log section with 30 virgin female
beetles. Carefully, he covered their entrance holes with screen so that males
could not enter and yet the females could still expel their boring dust. Then
he set the test log inside a screen cage. As the females began to bore and pro-
duce pheromone-laden frass, beetles flew in to land on the cage, two males
being attracted for every female that came. However, he saw no males stridu-
lating. Then Rudinsky took 30 male beetles from an ice chest and placed
a single male on each screened entry to a virgin female. As they warmed
up, the males began to stridulate, and suddenly the arrival of flying beetles
halted. He removed the stridulating males; flying males again began land-
ing on the cage. Back and forth, on and off, at about 10-minute intervals, he
could mask the attraction or renew it at will simply by alternately removing
and returning the stridulating males to the screened entries of the attractive
females.
Was the tiny chirping sound of the male really the signal responsible for the
population-regulating phenomenon Rudinsky was observing? Not satisfied
with this simple explanation, he attempted to establish more firmly that the
stimulus was solely auditory. Again, he placed a freshly killed male on each
screened female entry hole and found that, consistent with his theory, flying
males continued to arrive. However, this time he ran a parallel experiment—he

cut the wing covers of 30 live beetles so they could not stridulate, and placed
them in the same position. When these silent males rested on the screened
holes, they had no effect upon the stream of flying males that continued to
arrive. But as soon as the screen was removed so the silent males could enter
and join the female, again the flight aggregation stopped. Because the stream
of beetles could be halted without stridulation, the auditory signals could not
be the only ones involved!
Were the females involved in the attraction? By whacking a hammer on the
log bark just above each entry hole, Rudinsky killed the females inside their
galleries. Beetles kept arriving, apparently due to residual attraction. Then
he added males. Again they stridulated—but flying beetles kept on arriv-
ing! Rudinsky removed the females entirely from some logs before adding
extra males, but again the males’ stridulation was incapable of halting the
immigration.
The puzzle was solved. The arrival of large numbers of beetles quickly
leads to the discovery of the infesting females by males that stridulate as they
dig their way into the gallery to join the female. Responding to the chirps, the
female begins releasing a chemical that acts as a ‘mask’ or anti-aggregative
pheromone that camouflages the normally attractive odors she has produced
in her frass. This in turn serves to inhibit flight response; therefore the number
of arriving beetles drops off sharply and suddenly.
Later, Rudinsky demonstrated that electronic playback of recorded male
sounds has the same effect, thus definitely confirming the novel interac-
tion of sound and chemistry that mediates the beetle’ aggregation. A major
fraction of the anti-aggregative pheromone has been identified as the ketone 3-
methylcyclohex-2-en-1-one (MCH), a multi-functional pheromone that plays
many concentration-dependent roles in insect behavior. It was later discovered
that both this and another isomer of MCH are also released by the entering
male.
Typically, beetles of only one sex initially select the individual host tree.
(In monogamous Dendroctonus beetles the female excavates the entrance tunnel,
whereas in polygamous Ips the male performs this function.) Then as the pio-
neer beetles begin to attack, they discharge a long-range pheromone derived from
ingested host monoterpenes. This simultaneously serves two purposes: attracting
additional individuals of both sexes and inducing the opposite sex to enter the nuptial
chambers and mate (Fig. 6.10). In low concentrations, in synergistic combination
with other host- and beetle-produced chemicals, the pheromone attracts both sexes
and causes males to stop and stridulate. At higher concentrations, it prevents flight
aggregation of both sexes.
The communication system of bark beetles would seem to have high adaptive
value, promoting an even distribution of available mates, preventing overcrowd-
ing with subsequent brood mortality, and allowing the cooperative mass attack
Fig. 6.10 Starting and stopping an aggregation. Summary of the interaction of olfactory and
auditory stimuli in the aggregation response of the Douglas fir beetle and its subsequent inhibition
necessary to overcome the host tree’s resistance. However, such an explanation is

somewhat difficult to reconcile with natural selection theory that reasons that to be
selectively advantageous benefits must accrue to individuals. In addition, aggrega-
tion poses a problem in that the bark is a finite resource and each individual uses
it at a cost, so additional responders can be viewed as competitors whose effect
upon the beetles already present would be to lower their individual fitness. Are there
alternative explanations?
While both sexes are attracted, it does not necessarily follow that their inter-
actions denote cooperation (although there may be Allee effect benefits to the
signaler). Responders may be exploiting the senders’ message as a shortcut to host
location without assuming the risks of attacking resisting trees—in essence, cheat-
ing. Responding males may be eavesdropping, hoping to gain mating opportunities
by intercepting some of the females that are attracted. Context is also important.
Low densities of beetles would be unlikely to be able to subdue a living tree and
thus would achieve little or no per capita reproduction, but an already-dead host
would present little challenge.
If initial colonizers of live hosts experience high mortality, might their behav-
ior be regarded as altruism (see Chapter 10)? Or are they perhaps manipulating
responders to help them defeat the host? Depending on context, there are likely to
be some evolutionary trade-offs; the trick is to maximize the odds of successful host
colonization while minimizing the impact of competitors. Also, it should be noted
that conspecifics are not the only players. The same pheromones that serve as aggre-
gation attractants also are exploited as kairomones by a suite of generalist and spe-
cialist predators and parasites of the beetles and these predators and parasites may
disproportionately affect late-arriving beetles, increasing their costs relative to their
benefits. Finally, recall from Chapter 4 that plants are not passive in their interac-
tions with insects. They might be expected to counter by evolving ways to interfere
with the beetle’s communication system, possibly masking or inhibiting signals.
Many other examples of insect aggregations occur, many of which are
pheromone mediated. Cockroaches often congregate inside of kitchen cabinets.
When experimentally offered simple roosts made of filter paper, groups of Blatella
germanica cockroach nymphs invariably preferred to aggregate on those kept in the
stock cockroach culture for several days rather than roosts of fresh filter paper of
the same size (Fig. 6.11). When their antennae were removed, however, the cock-
roaches aggregated about equally on both roost types, thus implicating a role for
olfactory cues in their preferred choices. In order to determine the source of the puta-
tive pheromone, various cockroach body parts were isolated and washed in ether,
and fresh filter paper impregnated with the extracts. Abdominal extracts caused the
strongest aggregation, and eventually the pheromone source was determined to be
in the feces, presumably originating from rectal pad cells.
The adaptive significance of cockroach aggregations is still not entirely clear.
However, it has been shown experimentally that German cockroaches reared in iso-
lation grow at a slower rate and suffer greater mortality compared to those raised
in groups. Amelioration of the physical environment may also be involved, since
individuals in groups would be less prone to desiccation. In nature such aggrega-
tion could also benefit individual survival in the face of predation, a selfish herd
effect. Interestingly, cockroach aggregation pheromones do not appear to be species
specific; nymphs of several unrelated species in three different cockroach families
aggregate in response to each others’ pheromones. However, nymphs appeared more
responsive to their own species’ pheromones, and a few species appeared to be
repelled by fouling produced by other species. In addition, at least two species pro-
duce their aggregation pheromone in their mandibular glands, so one cannot simply
Fig. 6.11 Investigating roach aggregation. Sixty first instar nymphs of the German cockroach,
Blattella germanica, were introduced into a glass arena with three folded filter paper roosts: two
made of fresh paper, the other (at upper left) of paper left in a stock roach culture for 24 hours
prior to the experiment. Progressively greater preference for the conditioned roost occurred through
time: (1) start of experiment; (2) after 1 minute; (3) after 25 minutes; (4) after 45 minutes
assume that all species of cockroach produce a fecal aggregation pheromone simply
because other species collect on paper contaminated by them.
Although most aggregations are temporary, as we have mentioned, some are per-
sistent, most notably those forming social insect colonies. Those colonies whose
social lives are most highly developed exhibit extensive direct food sharing (trophal-
laxis, see Chapter 4). Also nearly universal among the social insects is clustering
behavior. Workers removed from a nest and placed in bare surroundings will quickly
gather in one or several little groups. If the mother queen and/or some larvae are with
them, this occurs even more rapidly and the grouping is even tighter.
Of all the types of pheromonally induced insect assembly, perhaps the most dra-
matic is exhibited by a fertilized social insect queen, continuously surrounded by
a retinue of crowding, licking, food-offering attendants (Fig. 6.12). The effect is
least marked in small colonies or where there are multiple laying queens. However,
tightly massed workers may number into the hundreds upon the distended bodies of
the ‘physogastric’ queens of army ants, fire ants, and termites. Several investigators
Fig. 6.12 Regulating the workforce. (left) A honey bee queen is surrounded by a retinue of
workers attracted to her by queen mandibular pheromone. (right) As the schematics show, the
pheromone component called homovanillyl alcohol (HV) bears a striking structural resemblance
to dopamine, a biogenic amine that plays a central role in insect behavioral regulation and motor
control
have shown that pheromones are involved; with army ants, in fact, merely allowing
the queen to sit on untreated balsawood strips will transfer enough scent to make
the strips highly attractive to workers.
Whenever joint efforts are needed—whether for exploiting a food source, repair-
ing a breach in a nest, or moving to a desirable new nest site—individuals of a great
many insects, particularly among social species, are able to chemically summon
others of their kind. Although the category is admittedly a loose one, this special
case of assembly has been termed recruitment, that is, communication that brings
conspecific individuals, often nest-mates, to some point in space where work is
required.
The simplest cases of recruitment appear to border on the unintentional. Honey
bee workers, for example, can recognize food odors both from smells adhering to the
bodies of successful foragers and by the scent of nectar regurgitated by them. Even
in the absence of communicative dancing, workers that have previously encountered
a similar odor will search the site for it again. In a slightly more advanced form of
recruitment, some social insects leave chemical ‘footprints’ to attract others, such
as the odor trails laid by both walking honey bees and some wasps in the vicinity of
their nests. While honey bee recruitment to pollen and nectar resources is mediated
by the famous dance language (see Chapter 8), there is also a unique scent to the
dance. Waggle-dancing bees also release small amounts of four hydrocarbons into
the air space surrounding them, and these odors stimulate the dance followers to exit
the hive, thereby increasing foraging activity for the colony.
Recruitment undoubtedly has reached its highest development among the ants,
bees, wasps, and termites, though it certainly is not restricted to them. Although
trail following has long fascinated biologists, the mechanism behind this behavior
was unclear until 1959 when it was first elucidated in fire ants, Solenopsis invicta.
In essence, workers travel through a ‘vapor tunnel’ created by the diffusion of

the evaporating pheromone blend (primarily Z,E-β-farnesene) that flows from the
Dufour’s gland as the trail-laying ant draws the tip of its extruded sting lightly over
the ground surface. Recruitment in fire ants also provides a good example of insect
mass communication—the sharing of information that can be transmitted only by
a group of individuals. Far from being random, the number of S. invicta workers
leaving the nest to move along a particular foraging trail is in direct response to the
summed amount of trail substance laid down by those workers already returning
from the food source. Both quantity and quality are mass communicated. At first,
increasing numbers of ants arrive; then, as the food becomes crowded, the arriving
ants become increasingly unable to contact the food and begin to turn back without
laying trails. Thus, numbers equilibrate at a level that is a linear function of the area
of the food source. As this area declines, so does the number of trail layers and hence
the number of newly approaching workers. Quality of the food source can also be
mass communicated through an ‘electorate’ response in which individuals ‘choose’
whether to lay trail or not after inspecting the food find. Through such aspects of
mass communication, trail pheromones can provide a control that is more complex
than one might predict from individual responses alone.
The tent caterpillar Malacosoma americana is a familiar early spring sight in
eastern North America, with its glistening silken webs in crotches of apple and
cherry trees. These gregarious caterpillars forage in groups, laying down silk threads
impregnated with trail odors as they travel. Foraging episodes are cyclical, occur-
ring about every six hours, when hungry groups leave the tent in search of fresh
foliage. During the outbound initial search the chemical marks are intermittent,
termed exploratory trails. After locating a productive patch of new leaves and feed-
ing to repletion the group heads back to the web; this time individuals drag the
tip of their abdomens along the silk lining the branches, laying down recruitment
trails (Fig. 6.13). Unsuccessful foragers return to the nest periodically and if they
encounter a recruitment trail they will immediately follow it out to the food resource.
Fig. 6.13 Studying recruitment. (left) Apparatus for experimental analysis of recruitment to food
in the eastern tent caterpillar, Malacosoma americanum. Caterpillars roost in a silken tent to which
they continually add silk secreted by their labial glands. (right) At intervals, groups travel out along
a branch in search of food, laying silk along the route; if productive food patches are found, the
caterpillars feed and then return to the tent, this time heavily impregnating the silk trail with a
pheromone (principally 5β-cholestan-3-one) from a sternal gland at the tips of their abdomens.
Other hungry caterpillars selectively follow the trails back to the food source
The tent caterpillar system has feedback loops similar to those in the mass
communication system of the fire ant, in that recruitment is elective. Individual
caterpillars mark in response to the quality of the food site (which in addition to
nutrition, may also reflect density of caterpillars), and other individuals use this
information to evaluate food patch quality and preferentially move out only to
highest quality patches. The overall effect promotes a high degree of coordinated
synchrony among the entire group, increasing the efficiency not only of foraging,
but also of thermoregulation (important for food processing), and defense. Group
members are initially all siblings, but groups often merge when more than one egg
case was laid on the same tree. Apparently bigger is better—experimental sibling
groups of 100 caterpillars grew to 150% the body mass of their siblings that were in
groups of only 30 caterpillars.
6.3.3 Alarm and Alert
In 1609, in what may be the earliest explicit description of an insect pheromone,

Charles Butler commented about an aspect of honey bee behavior already common
knowledge among beekeepers:
When you are stung, or any in the company, yea, though a Bee have strike but your clothes,
specially in hot weather, you were best be packing as fast as you can: for the other Bees
smelling the ranke favour of the poison cast out with the sting will come about you as
thicke as haile: so that fitly and lively did he express the multitude and fierceness of his
enimies that said They came about me like Bees.
Alarm and defensive behavior often go hand in hand, and in some cases a single
substance simultaneously releases both behaviors. For example, disturbed nymphs
of the pyrrhocorid bug Dysdercus intermedius secrete a fluid from the third dorsal
scent gland; the compound is used simultaneously for defense and for releasing
alarm behavior. In other cases, different components of a pheromone blend release
different aspects of the response. For example, one of the multiple components of
the sting-released honey bee pheromone has been identified as isopentyl acetate,
which smells like bananas; alone on a cotton ball it continues to attract and anger
other worker bees but does not release attack behavior.
Alarm and alert pheromones are alike in that both are produced under con-
ditions of immediate or potential threat, but the responses they elicit grade into
several functional categories; defense, dispersal, agitation, aggregation, and recruit-
ment are among the more common. For example, subterranean ants of the genus
Acanthomyops have populous colonies crowded into relatively small caverns. When
stimulated by alarm pheromone, they rapidly converge toward the source of dis-
turbance and meet the danger head-on. Related ants in the genus Lasius normally
nest in smaller colonial groups in more exposed situations under rocks or logs; their
response to disturbance is to scatter as if in a panic and run hurriedly about. Both
species employ undecane as the principal component of their alarm pheromones,
but Lasius is much more sensitive at lower concentrations, in effect, possessing an
‘early warning system’ for rapid evacuation when danger threatens.
A similar concentration effect involves citral, a mandibular product of cer-

tain Neotropical stingless bees. In low concentrations droplets of citral are highly
attractive to workers of Trigona subterranea and serve as a trail pheromone used
in recruitment. High citral concentrations in the nest vicinity, on the other hand,
release alarm and attack behavior. Some unrelated stingless bees that have lost the
ability to collect their own pollen also use citral to disorient other bees whose nests
they raid for pollen. Lestrimelitta limao scouts that manage to penetrate nests of
susceptible Trigona species release citral inside, causing a complete breakdown in
the social organization of Trigona as well as being a strong attractant to other L.
limao workers.
Since insect aggregations, whether temporary or permanent, constitute a jackpot
for effective predators, selection for efficient protective behaviors has undoubtedly
been especially strong among aggregating species. The existence of alarm-alert
pheromones has been demonstrated for many. Gregariously feeding green peach
aphids discharge a volatile sesquiterpene from their abdominal cornicles when dis-
turbed (Fig. 6.14); this acts as a powerful dispersal agent for nearby aphids that
are likely to be clone-mates, possibly a behavior favored by kin selection (see
Chapter 10). Both nymphs and adults of the bed bug, Cimex lectularius, discharge a
Fig. 6.14 Spreading the alarm. As shown in this pheromone bioassay for the green peach aphid,
Myzus persicae, when a filter-paper triangle daubed with E-β-farnesene pheromone secreted from
the aphid’s abdominal cornicles is introduced next to a cluster of feeding aphids, it takes less than
five minutes before most aphids have left the area and some have dropped off. This alarm dispersal
behavior is characteristic of a variety of aphids, but depends on context. Ant-attended aphids tend
to remain feeding, relying on their ant guards for protection. Ants are attracted to E-β-farnesene,
as are some predators, such as the multicolored Asian ladybug. In their glandular leaf hair, some
solanaceous plants produce E-β-farnesene that may help to protect them from aphid attack
multi-component pheromone from thoracic glands in response to irritation; alerted

individuals exhibit a response similar to that of aphids.
However, the most widespread occurrence of chemical alert systems is among
the social insects, particularly in those species forming large colonies. Evidence of
alarm-alert pheromones exists for all of the more highly social Hymenoptera and
termites, where it is key to mustering the collective force of the colony for defense
of the fort. A concentrated resource bonanza such as the stored honey and helpless
brood in the honey bee hive establishes the need for effective recruitment against a
wide range of threats, including vertebrates and even conspecific robbers from other
nests. Japanese honey bees (Apis cerana japonica) are subject to raids by giant
hornets (Vespa mandarina japonica) that attempt to recruit additional raiders by
marking the victim’s nest with pheromones from sternal glands. In response, guard
bees release alarm pheromone that recruits a counterattack force massed just inside
the nest entrance. If a hornet tries to enter the nest, it is quickly engulfed by a buzzing
ball (see Plate 18). Hundreds of bees add more alarm pheromone and also mark the
hornet with additional secretions that transform it into a chemical beacon for still
more recruits. Surprisingly, instead of stinging the hornet to death, Japanese honey
bees kill by roasting it. Generating swarm heat, they raise the core temperature in
◦
the ball to about 47 C, just below the bees’ own upper thermal tolerance.
6.3.4 Host-Marking
After laying an egg inside or upon a host, bean weevils, various parasitoid wasps,
and a host of other insects leave a pheromone mark (Fig. 6.15); other female
Fig. 6.15 Marking a host. A female Trissolcus basalis (Hymenoptera: Scelionidae) draws her
ovipositor tip across the cap of the host egg, marking the spot where she has just laid her own egg.
Such chemical marking repels other conspecific females and minimizes the likelihood of multiple
oviposition on a single host
conspecifics usually avoid laying their own eggs in or on these marked hosts.
Ovipositing female Mediterranean fruit flies, Ceratitis capitata, readily discriminate
between infested and uninfested host fruits; in a graded response, they adjust their
propensity to oviposit accordingly, marking hosts after egg deposition and avoiding
or reducing egg clutches in already marked fruits. Apple maggot flies (Rhagoletis
pomenella) behave similarly.
The advantages of such a system seem obvious. Few hosts or host plants can
support an unlimited number of parasites or predators. Too many attacks, and a
host usually will die, often taking its attackers with it. Thus, there would be strong
selective pressure for attackers to develop some way to assess host suitability and
in particular, to detect the presence of preexisting eggs or larvae, whether they
be one’s own, those of a conspecific female, or another species entirely. Marking
pheromones confer advantages both to the signaler and to the receiver. For both,
they reduce competition; an ability to distinguish between occupied and available
hosts and selectively choose the latter enhances the survival of one’s offspring. In a
limited resource the earliest offspring would have a head start over any latecomers,
and likely would triumph in any competition.
Visual inspections can and undoubtedly do sometimes provide cues, but for
insects chemistry appears to be overwhelmingly favored. Host-marking pheromones
have independently evolved many times, and are now known to occur in more than
200 species of parasitoid wasps and more than 30 species of herbivorous insects
across eight families of Diptera, Coleoptera, and Lepidoptera. All lay their eggs in
or on hosts of restricted size that can support the successful development of only
a limited number of offspring. Where marking pheromones have been investigated
in detail, they generally have been shown to be non-volatile chemicals detected by
contact chemoreceptors. A study of this behavior appears in Case Study 6.3.
Case Study 6.3: Host-Searching by an Ichneumonid Wasp,

Pleolophus
Why is this solitary ichneumonid wasp, noted for low, slow egg production,
still consistently one of the most common parasitoids on its sawfly hosts?
Many sawflies in the hymenopteran families Diprionidae and
Tenthredinidae feed in the trees as larvae, but drop to the forest floor
below to pupate. Here they may be attacked by any of a number of predators
and parasitoids. In the case of the sawfly Neodiprion swainei, a serious
defoliator of jack pine trees in North America, one of these parasitoids is
the ichneumonid wasp Pleolophus basizonus. Compared to other parasitoids,
this ichneumonid is known to have a rather slow rate of egg production and
oviposition; a typical female lays only two to four eggs per female per day.
Surprisingly, when Peter Price collected sawfly cocoons from different areas
in Quebec, Canada, he found that P. basizonus was consistently the most
abundant parasitoid species to be reared from them. How were these ‘poor
layers’ able to compete so successfully against 18 other species of parasitoids
also competing for the same host? The question was of more than purely
academic interest, since this wasp had been purposely introduced several
years previously for biological control.
Maybe Pleolophus was simply more efficient than her competitors in some
other aspect than oviposition rate, Price thought. For example, the wasp might
search more efficiently or regulate her oviposition more carefully. He placed
three mated female Pleolophus with varying numbers of sawfly cocoons
into each of several caged ‘arenas’ that simulated the sandy, lichen-covered
forest floor in careful detail. Six days later, he split open the cocoons to
count the number of eggs laid within. At all cocoon densities he tested, the
frequency with which he found only one Pleolophus egg per Neodiprion
cocoon was much higher than expected from a random attack. In fact, at
the highest cocoon densities, the oviposition pattern was extremely regular,
with almost no wastage of eggs through multiple ovipositions in the same
host. Apparently, searching females could discriminate between parasitized
and unparasitized hosts. There were no visual signs of parasitism, so the
discrimination presumably had some chemical basis.
A number of other investigations with other parasitoid species had uncov-
ered cases in which conspecific females would interfere with each other
during oviposition, particularly if they were confined in a small area at high
parasitoid:host ratios. Price reversed his experiment and varied the numbers
of wasps but kept the number of cocoons constant in each cage. Then, for two
days he watched the cages at hourly intervals and recorded the wasps’ posi-
tions. As crowding was increased, he observed more females climbing on the
cage sides in what looked like an escape reaction. It appeared as though rather
than interfering with one another, Pleolophus were doing their best to avoid
one another.
Was this behavior also chemically mediated? Placing a barrier across the
middle of an arena, he released a single female in one side and allowed her
to search for 6 hours while watching continuously and recording her search-
ing pattern. She was then recaptured and released again in the center of the
arena with the barricade removed, and the amount of time subsequently spent
in each half was recorded. As a control, Price similarly observed the behavior
of a female released in a fresh arena. Analyzing the results of several such
trials, Price found that 7 of 19 females tested exhibited a clear recognition and
avoidance of areas they had previously searched (Fig. 6.16). Several variables
remained uncontrolled, but avoidance was frequent enough to suggest that
females were depositing some ‘trail odor’ on the substrate as they searched
and that it served later to repel them, promoting dispersal. In fact, later stud-
ies showed that even different species of parasitoids attacking the cocoons
are able to recognize and avoid areas previously searched by Pleolophus.
Price’s studies thus were some of the first to report long-distance repellency
of possible competitors by a solitary endoparasitoid.
Fig. 6.16 Host-searching. Movements of a female ichneumonid wasp, Pleolophus basizonus,

following her release in the center of a glass-covered experimental arena. The dotted line marks
the position of a removable barrier. (left) The upper half of the arena had been searched previously
for 15 minutes by the same female; her avoidance of it during the subsequent 15 minutes is clearly
shown. (right) The female’s movements for 30 minutes in a fresh (control) arena. The dashed line
depicts instances when the female crawled on the underside of the arena cover
It is important to remember that insects also can detect chemical and/or physical
changes that eggs or larvae induce and use these as signals; simple feeding dam-
age would be an example. Thus, three types of studies are used as evidence that
a marking pheromone is actually involved: description of a putative host-marking
behavior; biological assays of behavioral responses to extracts of putative marking
pheromones; and identification of sources of marking pheromones and mechanisms
for their detection.
Modeling studies suggest that a major selection factor for the evolution of
pheromonal marking has been to avoid self-superparasitism, that is, a female laying
eggs again on a host where she already has offspring, and potentially reducing the
survival chances for both clutches as a result. The models suggest that when hosts
are sufficient, other conspecific females might gain from taking notice of her marks
and using them; when few unparasitized hosts are available, they might do better
to ignore the marks. There is, in fact, abundant experimental evidence that some
female parasitoids can adjust their rate of superparasitism according to the propor-
tion of parasitized hosts that they encounter, as well as avoiding host areas that they
have themselves previously searched.
Moving from host-marking behaviors to true territoriality requires only a short
conceptual jump. Territoriality can be defined broadly as any space-associated intol-
erance of others of one’s own species, or more narrowly as defensive behaviors
based on real estate holdings. To be a true territory marking pheromone, the mark
by itself must induce aversive behavior in conspecific intruders.
In both vertebrates and invertebrates, territories (defended areas) occur in a vari-
ety of contexts that most commonly involve feeding, breeding, or raising young.
(We’ll be returning to territoriality in Chapter 9, in the context of courtship and

conflict where it plays a major role.) Both major animal groups rely heavily on
chemistry for territorial marking, but they generally employ it in different ways.
Among vertebrates, a common behavior is to mark the perimeter of the defended
area. This has not often been shown for invertebrates, probably simply because
their territories are small enough in scale not to require it. The exceptions that
have been noted generally have involved social insects with correspondingly larger
territories.
No behavior has all benefits and no risks. Among the risks of employing
marking pheromones is potential for attracting the attention of eavesdroppers (see
Section 6.4.3). Several examples are known where marking pheromones are used
as kairomones by parasitoids to more efficiently exploit a host-marking female’s
progeny. Other insects of one’s own species may also eavesdrop. Weaver ants
(Oecophylla; see Plate 19) are territorial ants that form extremely large colonies;
in Kenya, a single O. longinoda colony was reported with 500, 000 workers that
controlled the canopies of 17 trees. Their complex pheromonal communication
(Table 6.1) includes marking their territories with fecal pheromones in a manner
analogous to many vertebrates. In response to freshly exposed habitat, weaver ant
workers increase their rate of deposition of anal spots, which persist for many
days. When workers of an alien colony are given access to the marked surface
in the complete absence of the original ants, the first individuals to arrive dis-
play intense aggressive posturing, opening their jaws and raising their abdomens
while also showing intense interest in inspecting and antennating the spots. Soon
afterward, the intruders begin to deposit recruitment odor trails as they return to
their nest.
6.3.5 Recognition
Male and female Nicrophorus carrion beetles cooperating to rear a brood of young
on a small vertebrate carcass can accurately distinguish partners from conspe-
cific intruders after touching their antennae; the pair’s individual chemical profiles
become strikingly similar after several days together, but partner recognition appears
to be based upon the reproductive status that develops after prolonged contact with
the carcass, rather than upon a specific personal identification odor.
‘Body chemistry’ is more than simply a catchy phrase. Chemical profiles can
be as distinctive as fingerprints, and in the insect world they facilitate the recogni-
tion and identification of familiar and unfamiliar, friend and foe. These chemicals,
known as surface pheromones or contact pheromones, are based on cuticular
hydrocarbons, but in many cases, they may be composed of a combination of
species-specific or sex-specific odors and environmentally derived ones (as, for
example, from food sources).
An insect’s body is covered by a cuticle with an outermost layer composed of a
complex mixture of lipids. Some of these lipids are waxy substances that function
as moisture barriers; many insects can change the amount of wax on their cuti-
cle in response to seasonal changes, and species in hot, dry environments typically
Table 6.1 Recruitment in Oecophylla longinoda, one of the most sophisticated communication
systems known for any ant
Chemical signal(s) Tactile signal(s) Movement pattern(s) Apparent function
Odor trail from rectal Antennation; head Signpost marking; Recruitment of major
gland; regurgitation waving, mandible looping trails laid workers to immobile
of liquid crop opening with food around food source; food source,
contents offering main trail directly to especially sugary
nest materials
Odor trail from rectal Antennation; Signposts and looping Recruitment of major
gland occasional body trails only after workers to new
jerking physical contact terrain
with terrain; main
trail directly to nest;
increase in
frequency of anal
spotting
Odor trail from rectal Antennation; tactile Main trail directly to Emigration of colony
gland invitation for nest; no signposts; to new nest site
transport; physical predictable carrying
transport of sequence of
nestmates nestmate categories
Short looping odor None Trails short, looping, Short-range
trails from sternal and limited to recruitment to
gland; gland surface vicinity of contact enemies
exposed as abdomen with enemy
is lifted
Odor trail from rectal Antennation; at higher Main trail directly to Long-range
gland intensities, body nest; no signposts recruitment to
jerking enemies; used with
short-range
recruitment,
especially intense
during territorial
wars with
conspecifics
have greater amounts of wax than closely related species that live in more moderate
surroundings. However, for some time it has been suspected that the hydrocar-
bons in this cuticular lipid mixture play another important role—that of recognition
cues when insects encounter each other. Over the last few decades, technological
advances allowed the studies necessary to confirm this suspicion for many insect
examples (Table 6.2), and appreciation for insects’ use of cuticular hydrocarbons
for chemical communication has grown tremendously.
For the vast majority of insects, recognition of one’s own species is necessary
only for copulation. Contact pheromones are certainly important even in these
cases. Certain cerambycid beetles attempt to mate with conspecific females only
after touching them with their antennae. When a male cricket contacts another
male with his antenna, he becomes aggressive, but when the same individual anten-
nates a female, he initiates mating attempts; if filter paper impregnated with female
Table 6.2 Major recognition functions of insect cuticular hydrocarbons (CH)
Implied question Selected examples and mode of action
Are you a member of my Beetles: Conophthorus, 8 species recognize each other through
species? CH profiles.
Are you the same gender as Flies: Drosophila, sexes recognize each other by relative
I am? abundance in same CH set; Glossina, distinctive CH set in
each sex.
Are you a member of my Ants: Leptothorax, queen-produced CH;
colony?
To which caste do you Termites: Reticulitermes castes differ in relative abundance of
belong? CH components.
What is your dominance Ants: Dinoponera, dominance recognized by relative quantity
status? of one CH out of 80+; Harpegnathos, dominance recognized
by variations in CH correlated with oogenesis in both queens
and workers.
Do you want me to do some Ants: Pogonomyrmex, tasks determined by CH proportions,
task? with appropriate environmental input.
Are you closely related kin? Bees: Apis, CH profile plus exposure to hydrocarbons in comb
wax.
Can you recognize that I am Beetles: Zyras, Diaritiger, CH mimics that of host ant Lasius.
alien? Parasitoid wasp: Lysiphlebus, CH mimics that of host aphids.
Social wasps: Polistes sulcifer, a social parasite, changes its
CH to that of closely related host after nest takeover.
body-wash extracts is touched to a male’s antenna, it also elicites mating behavior.

However, it is among insects that have adopted a gregarious or societal existence
that unfailing recognition of conspecific individuals over a wide range of circum-
stances assumes utmost importance. Inasmuch as mutual antennal contact seems
to be the most universal greeting, olfactory and tactile cues appear to be primarily
responsible for this recognition.
At first glance, nest-mate recognition among social insects such as ants appears to
be a casual affair, usually no more than a pause and perhaps an exchange of antennal
strokes. However, should an alien insect be encountered within the nest, some of
the scope of this communication is swiftly and unmistakably revealed. Even if it
is the same species and only from a different colony, the intruder normally evokes
hostility; if it is of a different species as well, the resident insect’s inspection is
swift, the attack violent. Usually the alien is instantly killed or driven from the nest.
Such nest-mate recognition in most cases is based principally if not exclusively on
‘colony odor’. Colony odor is an example of a class of pheromones characteristic
of social insects, the so-called colony pheromones. These signals, which have both
genetic and environmental components, affect all colony members, causing these
members to respond in context-dependent ways.
A common feature of nearly all social insect recognition odors is that indi-
viduals learn them at a particular life stage, usually within a short time of their
emergence to the adult stage. Such an imprinting-like system has two advan-
tages. First, it allows colonies to have a unique colony odor that incorporates both
6.4 The Information Content of Pheromones 249
genetic and environmental components. Second, it serves to mask any underly-

ing genetic heterogeneity that can arise from a queen mating multiple times. The
learned odor bestows on each individual a reference template against which it can
compare the odor of others it interacts with in the future. Either they match or
they don’t, enabling fairly straightforward discrimination between nest-mate and
intruder.
By definition, the insects in social insect societies live as members of castes,
subsets of the colony or group that are morphologically distinct and behaviorally
specialized. Thus, for a complex insect society to function, the recognition capacity
of individuals must extend beyond colony odor to include distinguishing, and acting
appropriately toward, the caste, life stage, and dominance level of nest-mates. For
example, certain paper wasp individuals must be given preference in trophallactic
food exchanges, and worker ants must segregate eggs, larvae, and pupae into sepa-
rate piles. How a reliance on pheromones can mediate such complex discriminations
brings us to the subject of the next section.
6.4 The Information Content of Pheromones
Consider a single insect sitting on a leaf, releasing a pheromone into perfectly still
air. Obeying the law of gas diffusion, the pheromone expands at a calculable rate
in all directions surrounding the insect until it forms a sphere at whose outer limits
its concentration is zero. Nested within this sphere is a second, more behaviorally
interesting one, the active space—the zone in which the concentration is at or above
the threshold level needed to evoke a biological response from other insects. The
active space constitutes the pheromone signal itself.
Of course, perfectly still air is an oversimplification. If a breeze is blowing, the
pheromone expands to form a cone downwind of the insect. Within the cone, the
active space becomes not a sphere but a downwind semiellipsoid, rather like the
upper half of a giant blimp lying on the ground with the emitting insect sitting at
the tip of the airship’s nose (Fig. 6.17). The stronger the breeze, the shorter the
blimp (or ‘odor plume’) becomes. This simple model does not account for surface
drag nor for the fact that most attractant vapors, being heavier than air, will tend
to fall before being fully dispersed, thus effectively flattening the top of the active
space. In fact, most real odor plumes are turbulent and unpredictably meandering;
a swirling smoke cloud from a chimney provides an apt visual analogy. However,
the active space concept can be used in combination with linear measurements and
gas diffusion laws to generate some predictions and generalizations about the use of
pheromones in functionally different communication systems.
6.4.1 Physiological Adjustments: The Q/K Ratio
Suppose that one knows both the amount of pheromone released and the behavioral
threshold for perception for some insect. To quantify the amount of pheromone
Fig. 6.17 Modeling pheromone movement. A simplified depiction of the active space of the
gypsy moth sex attractant, as deduced from linear measurements and general gas diffusion models.
Height and width are exaggerated in the drawing. As wind speed increases, there is a contraction of
the space within which the pheromone from a single, continuously emitting female is sufficiently
dense to attract males
released, let us consider it to be the number of molecules the insect emits per unit
of time (or as a single puff), and call it Q. For purposes of brevity, let us call the
potentially responding insect’s behavioral sensitivity threshold K and measure it in
molecules per unit of volume. The interval between release of the pheromone and
disappearance of the active space is called the signal duration or fade-out time;
using the blimp-like model of Fig. 6.17, it is the time required for the longitudinal
axis of the active space to diminish to below the receiver’s perceptual threshold.
Behaviorally, variations in the active space of a signal would be expected when
pheromones serve different purposes. How might this be achieved?
Both intuitively and by studying the model, one can recognize two ways: vary the
rate at which the pheromone is sent or vary the receiver’s threshold of perception.
Basically, the characteristics of any pheromone system can be expressed as a Q/K
ratio. The highest Q/K ratio corresponds to a pheromone system with a great signal
distance and slow signal fade-out time. Conversely, low Q/K ratios characterize sys-
tems having rapid fade-out time and a relatively small active space. A theoretically
infinite variety of systems exist as intermediates along the continuum; they result
from changing the values of Q and K independently of each other.
Consider the respective information transfer strategies of mate-seeking moths
and alarm-sending ants. A virgin female moth may need to broadcast her desirability
over a span of miles for several hours before she attracts a male. Obviously, a slow
fade time for her pheromone would be advantageous; it would be desirable also to
increase the area in which a potential mate could encounter and respond to the scent,
i.e. to have a large active space. This necessitates a high Q/K ratio, accomplished by
increasing Q, decreasing K or both. Because there will be metabolic limits on Q (the
amount of pheromone the female moth could manufacture and continuously pump
into the environment) it is more efficient to change the threshold for perception. To
achieve a large Q/K, the value for K will have to be extremely low; in behavioral
terms, the male receiver must have a very low threshold of perception, an example
of which is the silkworm moth (see Fig. 6.4).
When an ant worker generates an alarm, it is advantageous that other workers
be able to locate it sharply in time and space. For this to happen, a signal must
have a relatively short fade-out time; correspondingly, it requires a much lower Q/K
value, which can be accomplished either by lowering the emission rate or raising the
threshold of concentration, or both. The citronella ant, Acanthomyops claviger, uses
an alarm pheromone with a Q/K between 103 and 105 cc/second; signals take about
two minutes to reach an effective radius of about 10 cm and eight minutes to fade
out. If the citronellal alarm pheromone had as high a Q/K as some moth sex lures,
a one-ant alarm might theoretically keep the colony in a perpetual unproductive
state of chaos and panic. On the other hand, were the citronella alarm pheromone
to have an even lower Q/K value, the signals would be relatively useless because
they would not travel beyond the distance within which other ants could perceive
the danger directly themselves.
6.4.2 Pheromones as Language: Syntax and Lexicon
Because they have tended to be more thoroughly studied and thus simpler to under-
stand, most examples so far have involved single chemicals as pheromones. Often
even such ‘simple’ systems are quite complex, communicating different information
at different concentrations or in varying contexts. For example, we have mentioned
(see Case Study 6.2) that males of the Douglas fir beetle stridulate when they
detect low levels of the pheromone MCH. However, under a high concentration
of MCH the chirps of the male beetles change from those of the female attract-
ing calls to a longer, interrupted chirp with quite different acoustic properties; this
latter call is characteristic of both courtship with females and rivalry contests with
conspecific males.
Intuitively, one would expect to find many insects using such multifunctional
single pheromones; being small animals, it would be particularly advantageous,
allowing economy in their receptor systems without sacrificing behavioral diversity.
This phenomenon, termed pheromonal parsimony, is particularly characteristic of
the class of pheromones releasing alarm in social insects. Among ants, some alarm
substances are produced by the pygidial gland, others by mandibular glands. For
example, both harvester ants (Pogonomyrmex badius) and leaf-cutter ants (Atta
texana) use the same alarm pheromone (4-methyl-3-heptanone) secreted from the
mandibular glands, and in both species, differing concentrations of this odor elicit
very different behavioral responses. Upon perceiving relatively low concentrations
ants are attracted toward the source; as they move they cross into a much higher
concentration of the pheromone. The respective active spaces can be visualized as
nested concentric hemispheres in what is actually an intensity gradient to which
the receiver can orient; within the inner of these, ant behavior changes from simple
attraction to frenzied aggression.
An alternative to the multifunctional single pheromone is to evolve different
pheromones that serve the different functions but produce them in different exocrine
glands. Recruitment pheromones in termites are apparently all produced in the
termite’s sternal gland, but despite considerable convergence in the trail-laying
behaviors of ants and termites, among the ants trail pheromones are produced from
at least eight different glands, and in some cases more than one gland is involved in
recruitment communication.
The insect pheromone language is clearly not composed of an infinite number
of chemicals different within each species and unique to each message. In order
to achieve and maintain chemical uniqueness and promote privacy during com-
munication with such single chemical systems, insects would need to synthesize
increasingly more complex molecules; this strategy obviously would have its upper
limits, dependent upon such factors as the biosynthetic capabilities (energy cost) of
the organism, physical characteristics (such as volatility) of the molecules used, and
the neurological complexity of the receiver. Thus, among distantly related species,
there has been considerable evolutionary convergence in pheromone chemistry, even
while the uses to which the pheromones are put remain quite distinct. For example,
both stingless bees (Trigona spp.) and Acanthomyops ants use citral as a pheromone,
for recruitment trails in Trigona, and for alarm in Acanthomyops. Asian elephants
even share the same sex pheromone with 140 species of certain small moths!
Despite the intuitive logic that it would seem adaptively advantageous for
insect chemical communication to be species-specific, in tests of single chemi-
cal compounds under laboratory conditions, many pheromones have appeared to
lack specificity even at the generic level. How can this apparent discrepancy be
reconciled?
First, what appears to be a lack of specificity may not be so after all. The odor
environment of an insect is enormously complex; an ordinary terrestrial community
may contain hundreds of thousands of animal species and scores to hundreds of
plant species each producing its own characteristic odors. Superimposed on this are
seasonal and diurnal cycles of production and multiple other factors operating under
Fig. 6.18 Interacting systems. Males of four North American species of giant silk moths
(Saturniidae) each have a differing time of peak flight activity. Seasonal separation also exists;
most cecropia males emerge in May; polyphemus moths, in late May to early June; cynthia and
promethea moths, mostly in June. While the respective female sex attractant pheromones are
known to be very similar, caged females of one species almost never attract males of another
species because of these temporal and seasonal differences
field conditions. For example, males of different species of giant silk moths differ in
the periodicity of their flight activity, and thus vary in their responsiveness to female
sex pheromones according to the time of day or night (Fig. 6.18).
Female silk moths also call only during certain time periods of the daily cycle,
and the periodicity of both male flight and female release of sex pheromones can be
influenced by ambient temperature during development. When pupae of one giant
silk moth species, Antheraea pernyi, were kept at 25◦ C, the adult females initiated
calling behavior about 6 hours after lights-off. However, when pupae were kept
at 12◦ C, the resultant females advanced the onset of their calling to 2 hours after
lights-off. Corresponding shifts in flight activity also occur in males under these
temperature regimes. Hormones (especially juvenile hormone, see Chapter 2) medi-
ate both male responses and female calling times by acting on the different parts of
the nervous system involved in processing olfactory information. Circadian rhythms
cued by photoperiod and perhaps by temperature or other environmental factors set
the hormone secretion cycles.
Second, tests of a single chemical seldom represent the natural situation. Insect
exocrine systems are extraordinarily diverse (see Fig. 6.1), with a multitude of
glands and wide variations in the ultimate uses of a given gland and its products.
Furthermore, the products themselves are often complex. For example, at least 50
volatile compounds are present in the Dufour’s gland of a carpenter ant, Camponotus
ligniperda.
In human language development, the use of a different word for every message
in every language has been superseded by the greater efficiency of coding systems
based upon combinations of words. While a single word can send a message, adding
others produces an almost infinite variety of meanings; consider the different behav-
iors elicited by ‘Help!’ ‘Help me,’ ‘Help me love,’ and ‘Help me kill.’ An analogous
strategy has been evolved by insects and allows them to send messages more
efficiently: pheromone blends or multi-component pheromones. Multi-component
pheromone blends are probably the rule in arthropods rather than the exception.
Such systems have arisen in two ways: by the use of specific blends of chemical
produced within a given gland and, perhaps less commonly, by the use of various
combinations of glands releasing pheromones simultaneously or in succession.
Honey bees have one of the most complex pheromonal communication systems
so far identified; they possess 15 known glands that produce an array of com-
pounds. A classic example of a pheromone blend is the so-called ‘queen substance’
or queen mandibular pheromone (QMP) secreted in the mandibular glands of the
queen honey bee. It affects social behaviour, maintenance of the hive, swarming (see
Plate 20), mating behaviour, and inhibition of ovary development in worker bees.
A principal QMP component is the fatty acid E-9-oxo-2-decenoic acid. On
its own, it is both a sex attractant and an aphrodisiac. Combined with E-9-
hydroxy-2-decenoic acid it acts as a primer pheromone and inhibits worker’s ovary
development, queen rearing, and rate of juvenile hormone biosynthesis; the latter
thereby regulates the onset of age-related tasks such as foraging.
In addition to these, QMP includes three more chemical components that act
synergistically to elicit retinue attraction. The queen’s retinue pheromone is now
known to have nine chemical components (not all synergistic components orig-
inate from the same gland) and is the most complex pheromone system known
in any organism. The complete blend is translocated to all her body surfaces
by self-grooming and elicits the distinctive retinue response among worker bees,
causing them to lick, antennate, and groom their queen (see Fig. 6.12). Virgin
queens produce significantly less (or none) of many of the nine components
and are thus much less attractive to workers than mated, actively egg-laying
queens.
It is not unusual to find a chemical blend produced within a single exocrine gland.
For example, in ten Myrmica ant species, a variety of natural products are used for
the same communicative purpose—alarm. Among eight of the species, the same
two chemicals (3-octanone and 3-octanol) are present in the mandibular glands,
but in highly varying ratios. In one, the two chemicals are combined with a third
major constituent. Two other species have none of the previously mentioned com-
pounds but shared a single compound characteristic of them alone. As an interesting
aside, it has been shown that many of these same ant alarm chemicals have signifi-
cant allomone effects, inhibiting the germination of grass seeds collected as dietary
staples by various myrmicine ants.
In general, the pheromone blends of closely related species have diverged and
diversified mainly by varying the ratios of chemical mixtures produced in the
same exocrine gland. In moths, for example, the multi-component pheromones
typically consist of one abundant component and several minor ones; differ-
ent related species tend to use the same components, probably because of
shared biosynthetic pathways, but in unique combinations or ratios. Synergistic
effects, in which a blend elicits stronger responses than any component alone,
are common.
One of the most complex chemical signaling systems occurs among weaver or
green tree ants, Oecophylla (see Plate 19); five different recruitment systems have
been identified, using pheromones from two different glands combined with vari-
ous tactile signals to mediate a range of tasks from foraging to emigration to new
nest sites to responses to enemy threats (see Table 6.2). This is a striking exam-
ple of signal economy, with specificity arising from the different combinations of
tactile and chemical elements. Recruitment to food and to new terrain are both
initially guided by pheromones from the rectal gland; subsequently, characteris-
tic head waving and jaw opening specify food whereas body jerking denotes new
terrain.
Allomones—the other major class of chemical releasers of behavior—often also
consist of chemical ‘cocktails’ and many of the above considerations apply equally
to them. In addition, as defensive secretions, multi-component allomones sometimes
may simply overload predator receptor systems, thereby effectively hiding the insect
chemically from its enemies.
6.4.3 Exploitation and Code-Breaking

The aptly named ‘shampoo ant’ Formicoxenus obtains the colony-specific odors
of its host Myrmica via active mutual grooming soon after eclosion. The larvae of
certain syrphid flies emit chemicals that cause their ant hosts to treat them as another
‘ant larva.’ Many parasites and predators use the aggregation pheromones of bark
beetles as kairomones to find their prey. In a similar manner, parasitic tachnid flies
find their spined soldier bug (Podisus maculoventris) hosts by zeroing in on sex
pheromones released by males.
Because chemicals are involved in so many aspects of insect behavior, it should
not be surprising to find that some species turn the tables and exploit chemical com-
munication systems for their own benefit (Table 6.3). We have already noted some
examples, such as the beetle guests of ants discussed in Chapter 4.
Illicit receivers that intercept the signals of their hosts are widespread, and
exploitation can take many forms. One is code-breaking (a behavior sometimes
called eavesdropping, although that term has auditory implications). An illicit
receiver of a different species (or sometimes, another insect of the same species)
may decipher a chemical message meant for another receiver, and modify its
own behavior as a result. At a more assertive level, the receiver may actually
Table 6.3 Classification of semiochemicals by benefit (+) and cost (–) to signaler and receiver
Semiochemical Effect on signaler Effect on receiver Communication
Pheromone + + Mutualistic
Allomone + – Deceit, propaganda
Kairomone – + Eavesdropping
steal the message. For example, the cuticular hydrocarbon ‘uniforms’ of host ants
are acquired by some of their obligate social parasites. Such camouflage exploits
the manner in which the ants themselves acquire their colony-specific odors, via
grooming and trophallaxis.
A second form of exploitation of the chemical system occurs when an insect
actively produces chemicals that mimic the chemical messages of others. Illicit
chemical signalers of this type abound in the insect world. The ‘adoption’ and
‘appeasement’ gland secretions of some species of rove beetles (see Fig. 4.11)
elicit host behaviors allowing the beetles to integrate themselves into ant soci-
eties. Among social species, a common method of gaining chemical congru-
ency with the host is to actively biosynthesize the host recognition odor. Still
another type of exploitation involves disrupting normal communication. Some
slave-making ants spread panic throughout a colony by mimicking the alarm
pheromones of their hosts; in the confusion, they take over the victims’ pupae
(see Chapter 10).
At least three factors appear to facilitate evolution of chemical code-breaking
among insects. One is the design of olfactory systems, with their broad sensitivity
to odors of all sorts. Another is the ability to evolve extreme tuning to any odor that
provides a selective advantage. A third is the relative ease with which many insects
can learn olfactory cues, an ability that has been widely reported for insects ranging
from crickets to parasitic wasps. Thus, it appears that the risk of exploitation is one
cost of reliance on chemical communication systems. Why does natural selection
not eliminate it? Perhaps if the deceiver is rare relative to the genuine signal sender
or receiver, then the cost of the deception can be tolerated, because the legitimate
use of the signal will be more selectively advantageous.
6.4.4 The Chemical Channel and Other Signal Modes
One of the most striking characteristics of chemical communication systems is cer-

tainly their ubiquity, not just in insects, but in most groups of organisms including
plants and bacteria. Some scientists have suggested that chemical communication
be considered one of the fundamental attributes of life itself. How does it compare
with other modes of communication?
From the simplest to the most complex, any communication system must fulfill
two major criteria—it must be ecologically appropriate and it must fall within the
sensory and motor ranges of the species concerned. An efficient system will likely
possess several of the following features as well: (1) qualitative and/or quantitative
specificity, (2) rapid rate of information dissemination, (3) efficiency over consid-
erable distance, (4) directionality of transmission, (5) wide range of information
content, and (6) persistence or ability to start and stop quickly.
As we have seen, chemical communication systems possess most of these
attributes and are certainly as fully versatile and sensitive as acoustic and visual
communication to be considered in the following two chapters. In addition, they
possess certain unique advantages. For one, they are the only major communicatory
mode to have signals capable of lingering in the environment. Second, for chemi-
cal signaling, time and space take on a special meaning, for the sender and receiver
do not have to be simultaneously coordinated as demanded by visual or acoustic
communication. As a third point, like sound, chemical signals are able to go around
environmental barriers although, unlike acoustic signals, chemical messages may be
difficult to modify on short notice.
In concluding this section it is appropriate to consider the odorant chemicals
themselves. Edward O. Wilson and William H. Bossert have convincingly argued
(Table 6.4) that airborne odorants should have a carbon number between 5 and 20
and a molecular weight between 80 and 300 due to volatility constraints. In addi-
tion, they predicted that chemicals used for purposes requiring a high degree of
specificity (such as sex attractants) should have greater molecular size than those
odor signals used to elicit behaviors (such as alarm) for which species uniqueness
is less essential. In the main, these predictive generalizations have been upheld by
accumulating empirical data although size restriction does not apply to waterborne
and contact odorants.
Table 6.4 Chemical criteria proposed for airborne pheromones
Molecular Olfactory Energy expense to

diversity efficiency produce Volatility Other considerations
Carbon number<5, molecular weight <80

Very limited Low Low Very high Possible difficulty in
glandular storage
Carbon number 5-20, molecular weight 80-300
Increasing Increasing steeply Intermediate Intermediate Differences in
exponentially; diffusion
great number of coefficient in this
unique range do not cause
compounds much change in
possible properties of
active space
Carbon number >20, molecular weight >300
Astronomical Further increases Great expense Low May lead to
probably confer required to difficulty in
little or no synthesize and maintaining
further transport large adequate active
advantage molecules space
6.5 Chemical Communication and Insect Control

Since 1959 when the term pheromone was coined, pheromone study has flourished,
the beneficiary of two major technological advances. The first was the introduc-
tion of gas chromatography, later coupled with mass spectrometry to permit the
identification of miniscule amounts of secretory products. The second was the devel-
opment of more sophisticated neurophysiological (especially electroantennogram)
techniques that have led to better understanding of insect chemoreceptor systems.
Concurrently, the development of predictive physical models about pheromone
behavior has allowed analysis and understanding of odor transmission in nature. In
the past 50 years, research on insect pheromones has grown steadily, driven by hopes
that pheromones might be put to work to control insect pests. The idea of species-
specific manipulation of insect communication systems without adverse effects on
other organisms has considerable appeal. The use of pheromones for pest control
offers several obvious advantages over conventional insecticides. Pheromones are
nontoxic and non-persistent; being species specific, they elicit behavioral responses
at extremely low doses.
Practical applications using insect pheromones are based on two main
approaches: (1) air permeation leading to communication and mating disruption,
and (2) attraction to point-source lures for population monitoring and for control
by mass trapping. Mating disruption has found widespread application primarily
against moth pests. Mass trapping has proved useful against various moths, beetles,
and flies; obviously, this method is most efficient when the pheromone lures attract
females. If attractants and repellents can be combined, it can result in an effective
‘push-pull’ pest management system (Fig. 6.19).
Fig. 6.19 Behavioral manipulation. Insect pest control approaches often incorporate semio-
chemicals that affect behavioral responses. Integrating chemical stimuli that act to make a potential
resource unattractive to the pests (e.g., repellants, alarm pheromones) while at the same time lur-
ing them to an attractant source (e.g., sex or aggregation pheromones) constitutes an appealing
‘push-pull’ pest management system
6.5 Chemical Communication and Insect Control 259
The main behavioral mechanisms underlying mating disruption are sensory

fatigue or habituation, competition between natural and synthetic sources of
pheromone, and camouflage of a calling female’s odor plume. A host of practi-
cal issues need to be addressed for each pest situation. Cost-competitiveness is
always an issue, and variable pest population densities complicate making blanket
recommendations. Pheromone dispenser technology is an important related critical
area of research; economical and efficient materials must be developed that can
release sufficient pheromone amounts over an extended period and that can be eas-
ily deployed over large area. Climatic variables such as temperature, humidity, and
wind must also be factored into the overall equation for implementing and achiev-
ing pest management, and every situation is different. Collaboration between the
various stakeholders (researchers, growers, chemical industries, and the public) is
essential.
Nevertheless, there has been exciting progress in many cases, and notable success
in some. For example, sustainable control of the codling moth (Cydia pomenella),
a major pest of apples that has become resistant to many insecticides, has been
achieved through pheromone-mediated mating disruption over large areas world-
wide. This commercially viable technique is now commonly used in control of
several dozen pest moth species.
Relatively inexpensive pheromone-baited traps are now also widely used to mon-
itor the occurrence and spread of pest insects, especially invasive species. The spread
of the gypsy moth (Lymantria dispar) is monitored with sticky traps that snare
attracted males. Boll weevils (Anthonomus grandis grandis), once the scourge of
the cotton crop in the southern United States, have been eliminated almost entirely
in that region by an intensive cooperative effort of growers and government agen-
cies. A key factor in the success of this program has been the area-wide use of
pheromone-baited traps used to detect local spot re-infestations that can be treated
immediately to prevent population establishment.
By predicting the outbreak of a large infestation, pheromone lures have been use-
ful in timing, calibrating, and eventually reducing the application of insecticides,
thereby facilitating integrated controls that include the use of natural enemies. For
example, consider a program in use against three of the world’s most destructive
fruit pests—the Mediterranean fruit fly, the melon fly, and the oriental fruit fly. All
three are present in Hawaii but not currently on the mainland of the United States;
accidental imports have, however, repeatedly occurred. Deployed about ports of
entry, triple-baited traps containing synthetic attractants combined with insecticide
have proven very effective in detecting new infestations. It is estimated that this
early warning system has saved millions of dollars in potential eradication costs.
Thus, today chemical communication ranks as a vigorous, growing area of study
within insect behavior, driven by both intellectual curiosity and practical concerns
such as the desire to better control pest species while simultaneously reducing the
use of pesticides.
Chapter 7
Visual Communication
7.1 Introduction
Small black paper balls pulled through the air on a string with sufficient speed
will attract male flies, which fly after and grab them. When dragonfly nymphs are
presented with various still objects against a background of moving stripes, they
will snap at them. Backswimmers, Notonecta, placed in a tank with white sides are
unable to avoid being swept downstream in a current. What is going on?
In all of these cases, the insects involved are using visual orientation (see Chapter
3) in the context of behavior. However, the above are not examples of insect
visual communication, because to communicate there must be at least one intended
receiver. In the context of orientation, vision is extremely important to most insects,
as it is to most animals in general. However, for communication, insects tend to rely
more heavily upon their other sensory systems.
Working primarily with vertebrates, the early ethologists found visual commu-
nication to be an unusually appealing subject for behavioral study for at least two
major reasons—the tremendous amount of information that visual signals can carry,
and the fact that visual systems play a dominant role in the behavior of humans.
Therefore, it is not surprising that in the development of ethology over the last cen-
tury, processes of visual communication have played a starring role. Konrad Lorenz,
for example, derived many of his theories from the study of fish and bird displays—
structures and behavior patterns that function as visual signals between conspecific
animals. Niko Tinbergen’s landmark 1948 paper on social releasers also took most
of its examples from visual signals. Ethologists have found that at least two impor-
tant signaling phenomena are more readily studied in visual signals than in auditory
or chemical signals. One is the origin of signaling structures; the other is the evolu-
tionary transformation of signals from their original purpose to assume new roles, a
process known as ritualization.
In addition to their central role in the development of ethological theories, visual
communication systems have formed a fertile ground for development of two of the
behaviorist’s most powerful tools: the use of models and the use of disguise. While
neurophysiologists might manipulate relatively few stimulus variables with good
result, the types of behavior of interest to ethologists often could be elicited only

262 7 Visual Communication
by more complex stimulus patterns. The development of ways in which to approach

the study of these was one of the outstanding achievements of pioneer ethologists.
The first method, using biologically meaningful simple models, gained its early
impetus in Germany from the use of artificial color patterns to study visual attrac-
tion in honey bees. However, while this worked well in some cases, at other times
it was just too difficult for an experimenter to effectively imitate all the necessary
stimulus configurations, particularly when they might involve more than one sen-
sory mode. This limitation could be overcome, however, by the use of disguise,
altering live animals in such a manner as to develop a model that looks and acts
like the live animal (which it is!) minus certain key characters. For example, the
butterfly Nymphalis io responds to predation attempts by rapidly lowering its wings,
exposing four eye spots (see Chapter 5) that mimic the vertebrate eye rather closely.
Wondering whether these eye spots help the Io butterfly escape predation, A. D.
Blest obscured the spots on some individuals. When he compared predation by
yellow buntings upon these models with predation on equal numbers of unaltered
individuals, the unaltered, eye-spotted individuals escaped 76% more often than the
disguised butterflies did.
7.2 Bioluminescence
Larvae and larviform adult females of Phrixothrix beetles are sometimes called
‘railroad worms’ because, like twin lanterns, luminescent organs of their body glow
green while those on their head glow red. The predatory larvae of some small dipter-
ans called fungus gnats are unique among flies in producing their own light. Groups
of fireflies sometimes blink in unison.
For centuries scientific investigators and curious lay observers alike have been
fascinated by the production of light by living organisms, or bioluminescence.
The phenomenon is more widespread than most people realize, and apparently
has evolved independently many times (some say up to 30 times or more).
Bioluminescence occurs widely in vertebrates, invertebrates, bacteria and fungi.
Ninety percent of deep-sea marine life is estimated to produce bioluminescence in
one form or another. Even symbiotic organisms carried within larger organisms are
also known to bioluminesce. Among the Insecta, self-luminescent species occur at
least nine families in five orders: Collembola, Hemiptera, Diptera, Dictyoptera, and
Coleoptera. The intensity of luminescence varies greatly from one insect species
to another. In some species, it is so low as to be visible to human beings only by
the completely dark-adapted eye. (If humans had more sensitive eyes, they would
probably consider more insects luminous.)
Bioluminescence has been described in more beetles than any other group.
Fireflies, lightning bugs, or blinkies—adult lampyrid and elaterid beetles—are the
most well known probably because there are over 2,000 species and many are widely
distributed. Some fireflies are exceedingly bright for their size. The common eastern
United States firefly, Photinus pyralis, has a flash that varies from 1/400th to 1/50th
7.2 Bioluminescence 263
of a candle. As in other bioluminescent organisms, the light contains no ultraviolet

or infrared rays; yellow, green, or pale red in color, its wavelength ranges from 510
to 670 nm (usually 550–580 nm).
7.2.1 The Physiology of Insect Light Production

All fireflies glow as larvae. In Europe and Asia, the latter are often called glow-
worms. In the Americas, ‘glow worm’ also refers to larvae in the related beetle
family, Phengodidae, and in the fly genus, Arachnocampa (see Plate 21). Among the
adults, some (mainly diurnal species) are not luminous, but in most fireflies, either
both sexes or females only emit light from the underside of the lower abdomen;
the well-innervated luminescent organs are specialized regions of fat body beneath
localized areas of transparent epidermis. The yellow-green glow comes from a layer
of light-producing cells (photocytes) with a massive tracheal supply; these rest upon
a layer of cells packed with urate crystals that reflect the emitted light. The lumines-
cent reaction is very efficient, with a yield of over 90% and almost no energy lost
as heat; in Pyrophorus beetles, it has been judged to be less than 1/80,000th of the
heat produced by a candle flame of equivalent brightness.
How is the light actually produced? In 1885 the French physiologist Raphael
Dubois experimented until he found the answer. In a dramatic demonstration, he
removed the light organ from Pyrophorus, a luminescent click beetle, ground it up in
water, and left it until the light went out. Then Dubois removed a second light organ
from another beetle and ground it into boiling water; after a short time, its light was
also extinguished. When he combined the two extracts the light reappeared! In this
manner, Dubois showed that two substances were required for the light, and that one
of these was inactivated by heat. These were named luciferin and luciferase, both
after Lucifer, the bearer of light.
We know today that luciferin and luciferase are not specific molecules. They are
generic terms for a pigmented substrate and its associated enzyme, respectively. In
the presence of oxygen, the luciferase, and the cellular energy source ATP (adeno-
sine triphosphate), luceriferin oxidizes to oxyluceriferin, an unstable compound that
is in such an excited state it yields a photon of visible light. Interestingly, because
ATP occurs among virtually all organisms, this bioluminescent reaction has been
used in distant space probes searching for extraterrestrial life. A probe scoops up
planetary soil into in a chamber that contains luciferin and luciferase; if even simple
life were present, its ATP would cause a glow that could be picked up by sensors
and transmitted back to earth.
Luciferin is more or less identical in all insect species that produce light, but
luciferases occur in such a diverse array that they have proven useful in molecu-
lar phylogeny studies. The most thoroughly studied, however, is that of Photinus
pyralis, the initial source for the genetically engineered luciferase now being pro-
duced in the laboratory for a number of purposes. In recent years, genes coding
for luceriferase production have been inserted into many different organisms, from
potatoes to fish to silkworms. Luciferase is also used in forensics and in human and
animal medicine.
When insect larvae glow, the light is on more or less continuously, but most adult
fireflies produce brief flashes of light that are vital for their courtship. How does
a firefly control its flash pattern? Pliny the Elder’s famous Natural History offered
the opinion that fireflies turn their lights off and on by opening and closing their
wings. Because this encyclopedic work covered almost the entire field of ancient
knowledge and was based on the best authorities of that time, this statement was
repeated again and again all the way into and through the Middle Ages.
Insect physiologists have now studied the system in some detail (Fig. 7.1), and
have shown that firefly flashes are controlled by gating the supply of oxygen to
the light-emitting cells. A nerve impulse that does not travel all the way to the
insect’s ‘lantern’ as one might expect, but instead via tracheal end cells to tracheolar
cells that are usually filled with fluid. The impulse results in changes in the cells’
osmotic potential, so that fluid levels drop and oxygen can reach the photocytes
for the oxygen-requiring light reaction. During the dark part of the flashing cycle,
the tracheoles refill with fluid, oxygen can no longer reach the photocytes, and the
flash turns off. Interestingly, flashing can be turned on and off experimentally. If
spontaneously flashing fireflies are placed in a chamber where the atmosphere can
be rapidly changed from oxygen to air, they being to glow continuously instead;
changing the gas stream back to air turns the continuous glow back to spontaneous
flashing.
7.2.2 Bioluminescence as a Communication Method
Why do railroad worms have red and green lanterns? Do fungus gnats in the family
Keroplatidae use their light to lure potential prey? When fireflies sometimes blink
in unison, what do they accomplish?
Bioluminescence serves many purposes in the animal world in general, and
among insect species in particular, and its functions are often different for vari-
ous life stages. For immature fireflies, for example, bioluminescence appears to be a
warning signal to predators, because many firefly larvae contain chemicals that are
distasteful or toxic.
For most adult fireflies, the production of light is used to locate other individ-
uals of the same species for reproduction. Courtship typically takes the following
sequence: Males initiate flashing during flight at species characteristic times (often
around sunset) in a well-defined habitat area; their flight paths are also species char-
acteristic, especially during moments of light emission. Females remain stationary
and, upon perceiving a male flash, answer with their own flashes, which follow that
of the male after a brief, again species-characteristic delay. Repeated flash–answer
sequences soon bring the sexes together, with copulation following.
In eastern North America, the commonly seen fireflies belong mainly to two bee-
tle genera; the larger ones are Photuris, and the slightly smaller ones are Photinus.
Fig. 7.1 Controlling the flash. (above) Adult fireflies produce discrete rapid bursts of light by
changing the length of a column of tracheolar fluids and thus controlling oxygen access to the
light-emitting cells (photocytes). These changes are actively brought about by nervous stimulation
of the tracheal end cells, which share similarities to other cells that function as sodium pumps.
P, photocytes; NE, nerve end; t, tracheole; TF, tracheolar fluid; T, trachea; TC, tracheolar cell;
TEC, tracheal end cell. (A) Increased fluid length blocks oxygen to the photocytes, producing the
dark phase of the flash cycle. (B) When fluid length decreases, the photocytes have the oxygen
necessary to produce light. (below) Only about 40–60 ms pass between the arrival of the nervous
action potential at the insect’s ‘lantern’ and the start of light emission, with maximal light occurring
within 100–150 ms after action potential arrival
Both genera include several cryptic species that are morphologically distinguish-
able only in minor details of body color. In the 1950s, Herbert Barber showed
that flash patterns clearly distinguished otherwise-confusing species of Photuris
fireflies. A decade later, building on this work, Jim Lloyd researched the flash pat-
terns of Photinus, using electronic devices to produce artificial flashes of known
duration and to accurately measure the female’s response delay (see Fig. 1.16).
By varying different signal parameters, Lloyd learned that female Photinus were
discriminating male light pulse length, the interval between pulses, and/or pulse
number. Furthermore, males were able to discriminate between differing female
answer delay times. In addition, Lloyd confirmed that the flash signals of sympatric
species differed significantly, but that those of species that normally did not occur
together were often very similar. In other words, refined isolating mechanisms had
evolved only where there was a possibility of reproductive mistakes being made.
Do Photuris speak a different flash language from Photinus? Female Photuris
had long been known to be carnivorous, but imagine his surprise when Lloyd
observed Photuris females attracting and devouring Photinus males by mimick-
ing the flash responses of Photinus females. This has become a classic example
of aggressive mimicry (see Chapter 4); conspecific males are not eaten.
As aggressive mimics, female Photuris are quite versatile and effective; they
can adjust their flashes to attract males of at least four species with distinctively
different flash patterns, and they generally succeed in capturing at least one male
for every ten attempts (see Plate 22). Later studies by others showed that a Photuris
female becomes a femme fatale only after having mated. Mating induces a suite of
behavioral changes, including locomotor activity, answering postures, predaceous
behavior, and response to flashes of males of different species.
The complex flash exchange systems found in Photinus and Photuris are not
typical of all fireflies, but instead probably represent a quite advanced level of evo-
lutionary development for the group. Some more primitive species known to use
pheromones have continuously glowing females, and in other species females turn
on their glow only in response to glowing males.
Many behavioral and ecological adaptations serve to enhance bioluminescent
communication efficiency. Females of most species climb up on perches during
hours of mating activity. Flashing males assume flight altitudes so that their light is
directed toward the ground ahead of them, and many species execute aerial maneu-
vers which enhance their chances of seeing or being seen by females (Fig. 7.2). In
addition, most fireflies exhibit habitat specificity and/or orientation to a mating site;
these further restrict the areas that males patrol, reducing ‘background noise’ and
the chances of interspecific interaction. Restricted periods of activity, seasonally or
diurnally, have similar advantages.
Perhaps the most extreme example of such ecological and behavioral restric-
tion is that of the ‘firefly trees’ found in parts of tropical Asia. At times the firefly
population in a grove of trees will flash on and off synchronously, hour after hour,
night after night, for weeks or even months. More than 30 similar reports have been
published describing these oriental firefly displays, occurring principally from man-
grove trees along brackish rivers. Such behavior is best known in tropical Southeast
Asia, but not restricted only to there. In the United States, one of the most famous
sightings of this phenomenon occurred near Elkmont, Tennessee in the Great Smoky
Mountains during the second week of June 2005.
Fig. 7.2 Flashing for mates.

Patterns of light flashes of
five species of Photinus
fireflies, illustrating
differences in habitat and
flash parameters that help to
maintain reproductive
isolation between the species
Many reasons have been proposed for these massive synchronized displays,
including ‘a sense of rhythm,’ ‘an organic law of rhythmic appreciation,’ ‘sym-
pathetic telepathy,’ and a whole host of other similar non-explanations with
anthropological overtones. One writer even went so far as to attribute it to the twitch-
ing of the observer’s eyelids, remarking that ‘the insects had nothing whatever to do
with it!’
Through extensive fieldwork with oriental fireflies in Thailand and Borneo and
use of photographic and photometric analyses, John and Elizabeth Bucks showed
that synchrony of great numbers of individuals is indeed nearly perfect. They
hypothesized that flash synchrony is controlled by an internal resettable pacemaker.
Contrary to earlier reports, both males and females occur in these trees, although
the females do not participate in synchronous flashing. In some instances, aggre-
gations in a given tree may include more than one species, resulting in a complex
combination of flashes that is still presumably effective for each species involved.
The Bucks considered synchronous flashing to be part of a complex of behav-
ior patterns (congregation, selection of certain trees, flashing, etc.) for enhancing
mating under otherwise difficult conditions. The firefly trees appear to serve as
quasi-permanent rendezvous, and synchrony appears to increase their efficiency as
beacons. Adult fireflies in these tropical areas live for only a few days and therefore
must find mates quickly if they are to reproduce before they die. A single prospective
mate is nearly invisible among the thick, tangled plants of these swampy areas, but
many males flashing synchronously are easily seen by other fireflies, both male and
female.
7.3 Light Reception

A preying mantis waits, motionless as a statue, until a small fly ventures within
strike distance. As observers, we watch with fascination as the insect’s eyes
silently track the fly, and we cannot help but wonder: Does it see the world as
we do?
Until late in the seventeenth century, people assumed that insect vision was essen-
tially similar to ours. Then the invention of the microscope gave curious humans a
close-up look at the eyes of a common fly and set off a scientific debate that took
about 300 years to resolve. When Antoni van Leeuwenhoek first saw at an insect’s
eye through his microscope, he was so surprised that he subsequently wrote a letter
to the Royal Society of London, saying:
Last summer I looked at an insect’s cornea through my microscope. The cornea was
mounted at some larger distance from the objective as it was usually done when observing
small objects. Then I moved the burning flame of a candle up and down at such a distance
from the cornea that the candle shed its light through it. What I observed by looking into
the microscope were the inverted images of the burning flame: not one image, but some
hundred images. As small as they were, I could see them all moving.
Do some or all insects really see the world as a mosaic of tiny images? Popular
works depict insect vision like that, and one can buy novelty viewers that allow
children to imagine that they ‘see as insects do’, but are they correct? What does an
insect actually see? And how? Taking into account both insect neurology and the
significant variation that occurs in this very large segment of the animal world, the
answers seem to condense down to one short statement—It depends on the insect—
but working out the details has taken over two centuries, and still continues today.
7.3.1 Receptors and Form Perception

Sometimes, arboreal Cephalotes ants fall or jump from trees; when they do, they
orient their trajectories toward light-colored objects, which in their environment are
usually tree trunks or lianas. Many caterpillars rest with the darker surface of their
body directed toward the brightest source of light, producing a counter-shading
effect (see Fig. 5.2), even though they appear not to have eyes. Moths will fly to
illuminated glass that allows ultraviolet light to pass through, even when it appears
opaque to us.
How an organism behaves depends greatly upon the way its world appears. When
an insect regards a complex scene, it obviously cannot and does not notice every
7.3 Light Reception 269
detail. (Nor do we.) What it sees, or rather, what it responds to, depends on what its
evolutionary history has preprogrammed it to ‘look’ for. An insect’s response is also
determined by properties intrinsic to the photoreceptors themselves and the manner
in which they relate to higher centers in the brain.
Insect eyes vary a great deal in form and location, but are of two basic types:
compound and simple. Simple eyes include the larval eyes, or stemmata, of
holometabolous insects and the dorsal ocelli present in most winged adult insects;
interestingly, the two are apparently not embryologically related. Larval stemmata
vary in size, number, and complexity. Fly maggots have only small pockets of photo-
sensitive cells in the cuticle. A lepidopteran caterpillar typically has six stemmata on
each side of its head, which look like simple white circles; with these, some caterpil-
lars can differentiate shapes and orient toward boundaries between black and white.
Other larvae, such as those of predaceous Cicindela tiger beetles, have stemmata
that apparently work as well as many compound eyes, making one wonder why the
insects did not retain eyes like this into adult life.
The winged adults of most orders and the nymphs of hemimetabolous species
usually have three true dorsal ocelli on the top of their heads. They appear to be
‘good’ eyes with a wide view and a lens; some also have a reflective layer and
some have a mobile iris. However, the retina is very close to the lens, so they are
apparently profoundly out of focus. Then what is their purpose? Currently the best
hypothesis is that they are horizon detectors that contribute to equilibrium when
the insect needs to make a fast correction during flight. Being unfocused allows the
receptors to respond to overall light distribution rather than every small leaf shadow.
The ocelli also are very sensitive to low light intensities. In certain bees and wasps
there appears to be a correlation between large ocelli and activity peaks at dawn
or dusk. (It is also known that insects can perceive light directly via the brain cells
as well as through their compound eyes and ocelli; some have other photosensitive
tissues as well.)
The best-studied and presumably most important visual organs in the majority
of insects are the compound eyes. With the exception of certain specially adapted
parasitic and cave-dwelling species, most adult insects have a pair of these promi-
nent organs, bulging to varying degrees from either side of their head so as to give a
wide field of vision in all directions. Composing each compound eye, and extending
inward like narrow columns, is a group of densely packed hexagonal units called
ommatidia. Most people think they know this from looking at insects, but in real-
ity this is modern knowledge gained from microscopic and photographic close-ups.
All we can view directly is a vague bumpiness, because the facets of an insect’s
compound eye are too small to be resolved with the naked eye. The number of
ommatidia in an insect eye varies greatly; while the common house fly has about
4,000, dragonflies may have as many as 28,000.
There is considerable diversity in the structural detail of ommatidia in differ-
ent insect groups. However, basically each ommatidium consists essentially of
two parts. One is an optical, light-gathering part (the lens or facet, and the crys-
talline cone). The other is the sensory part (retinal cells and their differentiated
margins called rhabdomeres or rhabdoms) that perceives the light radiation and
transforms it into electrical energy for nerve transmission. Each crystalline cone
also contains pigmented shield cells that help protect its actual light-sensitive
chemicals.
Yet, while all compound eyes look superficially quite similar, their insides vary
in some fundamental ways that relate to an insect’s behavior. Like all other organ-
isms, insects depend on light energy for their vision, but too much can be damaging.
Sunlight is particularly hard on pigment cells, so insect eyes basically come in
two different optical types (Fig. 7.3) that let different amounts of light strike the
rhabdoms.
Adult insects that are active during the day generally have apposition eyes that
limit the amount of light reaching the retinal rods. Each retinal rod receives light
only through its own lens. Light from other facets is intercepted by pigments that
act like a screen. The field of view of each rhabdom occurs next to (apposes) the field
of its neighbor, thus giving the eye its name. This is the kind of insect vision mim-
icked by children’s toys and popular literature. It produces an overall erect image
that is made up of a mosaic of adjacent fields of view. How coarse or fine a partic-
ular part of the image appears will depend on the number of facets per unit area.
Many nocturnal and crepuscular insects have a different system that gives a much
brighter image at night. A superposition eye converges (or superimposes) light rays
from many lenses onto each retinal rod. Because the effective aperture now is no
longer the width of a single facet but may be ten to thirty facets or more, the image
is brighter but more diffuse. By design, a superposition eye is much more sensitive
at low light levels than an apposition eye would be, but in addition, many super-
position eyes also have a tapetum, a mirror-like cell layer in the back that reflects
light outwards so that it passes twice along the retinal rod, doubling its effect on the
light-sensitive retinal pigments. Instead of a mosaic, all superposition eyes produce
a single deep-lying erect image in the vicinity of the retina. (This distinguishes them
not only from apposition eyes, but also from camera-type eyes in which the image is
inverted.)
When light is reduced, a day-flying insect usually just becomes inactive. But
if too much sunlight is damaging, what happens when a night-flying insect is in
daylight? Remarkably, whereas the shielding pigments in apposition eyes are sta-
tionary, those in superposition eyes can migrate. At low light intensity, the pigments
move into the distal part of the cells, allowing light from adjacent ommatidia to pass
through and strike the rhabdom below. Upon exposure to bright light, the pigments
disperse to form a protective curtain that prevents the spread of light from adjacent
facets, and the eye functions like an apposition eye.
This migration also explains why the eyes of many moths seem to glow at night
but not during the day. At night, the tapetum near the base of the moth’s eye reflects
light but during the day, the shielding pigments that have moved into place absorb
most of the light before it is reflected back.
Although we know that sight differs among animals, we do not know what
animals actually perceive. There is an important distinction between having light
illuminate the retina, and understanding what is being seen. Perception across the
Fig. 7.3 Seeing the light. (above) Robber flies (such as the one shown here feeding on a solitary
bee) typically have such acute vision that they are able to catch their prey on the wing. Like most
other diurnal insects, the Asilidae have apposition eyes that screen light from adjacent rhabdoms,
so that each lens forms its own relatively sharp image. (below) Nocturnal insects such as this firefly
typically have superposition eyes that gather light from many lenses to form a brighter but more
diffuse image. In addition, adjustments in both the colored screening pigments and the middle-
wavelength pigments in photoreceptors optimize the firefly’s reception of intraspecific signals
animal kingdom is the subject of much ongoing research. Consider the simple ques-
tion of whether an insect can distinguish between one shape and another, the ability
called form perception.
Much of our present knowledge of insect form vision comes from honey bee
behavior during training studies where choice of some shapes has been rewarded
with food. However, there are several limitations and difficulties implicit in these
studies. Perhaps the most striking has been that all such training studies have been
complicated by the bee’s spontaneous responsiveness to flicker (see Section 7.3.2,
below). Moving shapes of all sorts are more attractive than stationary ones, and
while bees could be trained to distinguish solid figures from broken ones, they could
not distinguish solid from solid or broken from broken. The adaptiveness of such
behavior is apparent if we consider a bee in its natural environment. Not only does
wind cause flower movement, but when bees fly low over the ground in search of
flowers they experience the passing of a radial cluster of flower petals as a burst of
flickering. The regularity and high frequency of such flicker patterns identify the
stimulus as readily as the total shape does to our eyes.
A second less immediately apparent limitation is that even these results are quite
possibly open to question because they may be confounded by stimulus filtering
appropriate to flower searching but not to other situations. Still a third limitation
is that honey bees may not be the appropriate ‘representative’ insects to generalize
from, any more than white rats are necessarily ‘representative’ mammals for all pur-
poses. Certainly some other Hymenoptera, such as the predatory wasps Philanthus
and Ammophila, use landmarks for topographic orientation (see Chapter 2) to an
extent which implies a better ability to discriminate forms than would be expected
from theories built upon honey bee observations.
7.3.2 Visual Acuity and Flicker Vision
Dance flies cruise just above the water surface, searching for drowning insect prey.
How do they keep their bearings? Males of the hoverfly Syritta discern and begin to
pursue a female while remaining beyond her visual detection range. What cues are
they discerning that trigger this pursuit?
The term visual acuity is used in various ways. It can describe how well two
close objects can be determined, or the smallest single object that an eye can detect,
or (as often used experimentally) the finest grating of equal light and dark stripes it
can resolve. A related term, resolution, is used in a loose way to mean the ability to
resolve fine detail or determine the parts that make up an image.
The better the acuity, the greater the distance at which prey can be seen, mates
can be located, or environmental structures can be used to navigate. The conditions
under which acuity is measured must be specified, of course, because the amount of
light available to the receptors affects the performance of any eye. As light becomes
increasingly dim, one eventually reaches a point where there simply are no longer
enough photons available to trigger the receptors to respond. In addition, contrast
becomes less and less.
Assuming the light is adequate, how acute is an insect’s vision in general? With
so many lenses, at first glance one might think that insects should have excellent
vision—yet it has long been clear that for stationary objects at a distance, the com-
pound eye is universally poor. Details of an object often fail to evoke responses
at all. With some insects such as the honey bee, the eye receives fuzzy images
even when the object is large and nearby. In 1894, a scientist named A. Mallock
described the compound eye vision like this: ‘The best of the eyes . . . would give
a picture about as good as if executed in rather coarse wool-work and viewed at
a distance of a foot.’
If insect eyes do not come even close to having the acuity and resolution of
human eyes, then why not? For one thing, form and motion perception are very
closely intertwined; just as in photography, image motion across the eye causes blur.
This is more than a theoretical consideration. Insects have been measured routinely
turning at speeds well into the hundreds of degrees per second, and in high-speed
maneuvers, rotating at speeds up to several thousand degrees per second.
Furthermore, just as we have noted for so many other phenomena, an insect faces
problems that arise from its relative size. Because each ommatidium has its own
lens, and an insect is small, the lenses are very small. Small lenses are diffraction
limited; to increase resolution by a factor of two requires doubling the diameter of
each ommatidium as well as doubling the number of ommatidia in a row. Another
way to state this would be to say that the eye would need to grow as the square of
the required acuity. For an insect compound eye to provide the same resolution as
ours, Mallock calculated, it would have to be huge—on the order of 38 ft (12 m) in
diameter.
As one might expect in a large group such as Insecta, many subtle changes in eye
structure have been found that increase visual acuity in specific ways that adapt their
owners to different environmental conditions. Nonetheless, the acuity challenges
inherent in the compound eye are great enough to have led one researcher, D.-E.
Nillson, to comment, ‘It is only a small exaggeration to say that evolution seems to
be fighting a desperate battle to improve a basically disastrous design.’
If the design is so poor, why has it persisted when single-lens alternatives were
and are apparently at hand? Researchers studying insect vision find this to be one of
the intriguing unsolved problems. Currently, the best guess seems to be that visual
acuity is less important than other advantages. One is that a compound eye excels
in processing speed. Because the images are processed in parallel, the design allows
for fast motion detection and image recognition.
For another, the compound eye excels in providing a panoramic view of the world
with a large field of vision. Even ants, which have a relatively small number of facets
on either side of their heads, can perceive almost the entire visual field above and
below the horizon. Their ‘blind spot,’ an area hidden below the thorax and abdomen,
has been estimated to be only about 10% of their total visual field.
A third major advantage is that some parts of the ommatidial mosaic can be
fine-tuned to be more visually acute than others. It has long been known that the
number of facets per unit area often varies in different directions in a given eye and
in different regions of the eye. Since the middle of the last century, major technolog-
ical refinements have allowed biologists to study insect visual acuity by noninvasive
methods rather than the older histological techniques. A major outcome has been the
realization that insect eyes vary in ways that reflect their life habits and ecology. For
example, one might reason that for many swiftly flying insects acuteness of vision in
the vertical axis would be more important than in the horizontal; in fact, the curva-
ture of their eyes lends very different dimensions to the ommatidial angles in these
directions. Three broad patterns have been identified in apposition eyes. One is an
overall pattern associated with forward flight (perhaps particularly through vegeta-
tion). The second is the presence of acute zones concerned with the capture of prey
or mates. The third is horizontal acute strips associated with flat environments such
as water surfaces.
Compound eyes also provide a unique way to determine distance. The eyes of
an insect are fixed; it cannot move them independently of each other. However, as
an insect directly approaches the object that it is viewing (or vice versa) the retinal
image gradually appears closer toward the inner part of the two compound eyes,
thus affording a method of judging distance. As the visual angle of the ommatidia
becomes progressively less over the inner part of the eye, the object comes into
sharper vision. Experimental work with predatory insects such as preying mantises
and dragonfly nymphs (Fig. 7.4) demonstrates this well.
Compound eyes are better adapted to perceive motion than static form. In fact,
reliance on an object’s real or apparent motion (flicker vision) is so widespread
that it may be regarded as one of the normal concomitants of vision with the
compound eye. Because ommatidia recover very rapidly from light impulse stim-
ulation, the insect eye has a remarkable capacity for seeing successively different
images at very short intervals and thus for scanning a moving object. As a result,
an insect may be able to resolve a finer pattern when it is flying than when it is
at rest. Many predatory species, for example, respond only to moving prey; and
given choices of stimuli, most insects show a preference for the shapes that cause
the most flicker.
Fig. 7.4 Assessing prey distance. At the front of its head, a dragonfly larva (Epicordulia) has a
two-jawed labium which is jointed in such a way that it cannot be moved sideways nor used except
in the fully extended position (right); thus there is only one point in space where a prey can be
caught. With its labium in the retracted position (left), the larva faces the prey directly and moves
toward it, viewing the prey with increasing definition. When the prey’s image falls on certain inner
ommatidial elements, it is seen the most clearly. Here it is exactly within proper striking distance,
which corresponds to the intersection of the optical axes of these ommatidia
Our own human attention is immediately attracted when sudden movement,

flicker, or novelty interrupts an otherwise static scene. The same is true for insects.
However, flicker is hardly noticeable to us when it exceeds a rate of about 20–30
changes per second. How do insects compare with us in this regard?
When a flickering light is presented to an animal, it is possible to determine the
rate at which its eye crosses a threshold from being able to distinguish the flashes of
light as discrete entities to see them as a single continuous light. This threshold value
is called the flicker fusion frequency. For the purposes of presenting moving images,
the human flicker fusion threshold is usually taken as 16 Hz. Flicker fusion is impor-
tant in all technologies for presenting moving images; if the frame rate falls below
the flicker fusion threshold for the given viewing conditions, flicker will be appar-
ent to the observer, and movements of objects will appear jerky. To overcome this
annoyance, frame rates are usually several times this threshold. (Current computer
CRT displays usually operate at a vertical scan rate around 100 Hz.) Research with
human subjects shows that the flicker fusion frequency can vary widely with such
factors as the relative brightness of the flicker and its surrounding environment, the
location on the retina where the image falls, and the fatigue level of the observer.
Common experience also shows that under some conditions, humans can discern
much higher flickers, such as the 100–120 Hz flicker produced by fluorescent lights
with magnetic ballasts.
Though much less well studied than that of humans, insects’ flicker fusion fre-
quency also varies widely. However, within limits that are undoubtedly set by retinal
mechanisms, there is a strong correlation between insect behavior and rate of flicker
fusion. ‘Slow eyes’ with a flicker fusion frequency of as low as 5–10 flashes per sec-
ond are characteristic of relatively slow-moving or nocturnal insects. Rapidly flying
diurnal insects such as bees and flies, by contrast, may resolve flicker frequencies
as high as to be 10 times as discriminating as the human eye. Male house flies,
for example, respond to flickering stimuli during courtship and will court models
with increasing responsiveness at flickers of up to 270 stimulus changes per sec-
ond before beginning to decline; their flicker fusion frequency has been estimated
separately through electrophysiological studies at 265 per second.
7.3.3 Polarized Light Perception

Hungarian mayflies were discovered laying egg masses on asphalt road surfaces
near the stream from which they emerged. What perceptual misunderstanding led to
such a completely unsuitable choice? Foraging in the evening, Scarabeus zambe-
sianaus dung beetles that find fresh dung quickly make it into a ball and roll
it away in a straight line; 40–50 minutes after the sun sets, they abruptly stop.
Why?
For some insects, an additional advantage of compound eyes is their ability (per-
haps aided by specialized ommatidia at the margins) to perceive light polarization
(see Fig. 3.12). As sunlight passes through the earth’s atmosphere, some of its waves
begin to vibrate in a specific direction. As perceived by an instrument or animal on

the earth, both the degree of this polarization and the plane of vibration vary with the
position of the sun above the horizon, and with the observer’s elevation above the
horizon. It is a complicated system, but insects as varied as house flies, European
red wood ants, and desert locusts have been shown to orient by it. Some of the
clearest examples of polarized light perception occur in many social Hymenoptera
that must orient to find food and return to their nests. Because the plane of polar-
ization changes as the sun moves across the sky, such homing requires these bees,
wasps, and ants to have not only a polarized light compass but also a way to measure
time lapse.
The first indication and then the first proof that animals might have the abil-
ity to detect this polarization was obtained through Karl von Frisch’s studies of
honey bee communication. The accuracy of the ‘dance’ (see Chapter 8) of a return-
ing forager, which shares information on direction and distance of the food source
with other workers, depends in part upon the polarized light pattern perceived from
blue sky. Because its dances are goal directed, the honey bee affords a unique
experimental opportunity to test orientation to polarized light. When the sky is com-
pletely cloud covered, bee dances are disoriented. If, however, a small patch of blue
sky remains (even when the sun is not visible), bees dancing on a horizontal sur-
face will orient correctly. If the plane of light vibration is altered artificially, as
can be done by interposing polaroid sheets, the orientation of the dances changes
correspondingly. For example, when the polaroid sheet is rotated clockwise by
30◦ , the bees will immediately shift the direction of their dances by about the
same amount.
Polarized light orientation and/or time compensation undoubtedly occur in a
number of contexts. When light bouncing off the asphalt that fooled mayflies in
the introductory example above was assessed with instruments designed to measure
polarized light, the road was found to reflect plane-polarized light the same way the
sunlit water surface in a stream did. The dung beetles in the other example forage
at the time of day when the light of the entire sky is polarized in one direction, and
researchers with polarization filters could switch the direction of the dung beetle’s
trips. Although other factors might also be involved in the beetles’ activity cessa-
tion, at the time at which they stop, the degree of polarization at the zenith of the
sky abruptly decreases from 45 to 5%.
In many cases, however, showing that an insect can detect polarized light or mark
time is not the same as showing it uses this information. Schistocerca gregaria,
the desert locust, clearly responds to polarized light, and this response has been
well characterized, but its use during migration has not been established. Monarch
butterflies use a time-compensated sun compass during long migratory flights to
Mexico (see Fig. 3.11). However, in both field and laboratory studies, the plane
of polarization did not appear to make any difference in their orientation. Many
Neotropical nymphalid butterflies reflect polarized light from their wings, but still
missing are behavioral studies to prove that these reflections are used to identify
conspecifics in the low light intensity of the forest foliage.
7.3.4 Color Vision

Mosquitoes and some flies have been shown to be attracted or repelled by specific
surface colors, and by specific colored sources of light; interestingly, the surface
colors they prefer do not necessarily correlate with the light source colors that
attract them.
Color vision and color perception across the animal kingdom are subjects of
much ongoing research; beyond a mere handful of species, science has a very lim-
ited understanding of the many ways that animals see. At its simplest, color vision is
the ability to discriminate between two wavelengths of light. In a bit fuller detail, it
is the capacity of an organism or machine to distinguish objects based on the wave-
lengths (or frequencies) of the light they reflect or emit. The nervous system derives
color by comparing the responses to light from photoreceptors in the eye that are
sensitive to different portions of the visible spectrum.
Three things are needed to see color: a light source, a detector (the eye) and a
sample to view. Color is a rather strange phenomenon because it is not a property
of the object—for example, a blue berry does not emit blue light. Instead, color is
a quality constructed by the brain; the blueberry is blue to us only because human
eyes can distinguish those frequencies of visible light that the berry reflects rather
than absorbs. The theoretical significance of developing color vision is thought to
be that it allows better discrimination of surfaces.
It appears that most insects have only two types of visual pigments—one absorbs
green and yellow light and the other absorbs blue and ultraviolet light. The color
vision of these species is limited, and somewhat resembles that of colorblind
humans. Dichromatic insects are generally considered to have difficulty distinguish-
ing between color blends. Other insects, notably honey bees, bumblebees, and many
diurnal butterflies, have true color vision. They have three visual pigments, with
absorption maxima in the vicinity of 360 nm (ultraviolet), 440 nm (blue–violet),
and 588 nm (yellow). Within the range of their spectral sensitivity, these trichro-
matic insects can perceive a complete spectrum of colors and can also discriminate
between single colors and mixtures of colors.
Humans also have trichromatic vision, so in this sense the color vision of these
insects resembles our own. However, at least in the species that have been studied,
their visible spectrum is shifted toward shorter wavelengths compared with ours (see
Plate 23). For humans, only the wavelengths between about 400 mμ (ultraviolet)
and 750 mμ (red) are visible; we cannot perceive the near-ultraviolet portion of
the sun’s rays that reaches the earth. In general, most insects respond to a range of
wavelengths extending from the near-ultraviolet (300–400 mμ) up to a maximum
around 600–650 mμ (orange).
Spectral sensitivity alone does not imply discrimination, but behavioral evidence
supports the interpretation that many insects can and do distinguish between dif-
ferent wavelengths and hence possess true color vision. Near-ultraviolet light is the
most effective region of the spectrum in directing phototaxes; the widespread use
of ultraviolet lamps in insect traps is a practical application of this. The adaptive
significance of ultraviolet phototaxis may be that such light signals ‘open space.’
Much of nature, especially green foliage, absorbs ultraviolet wavelengths; the open
sky, left as the only extensive source of ultraviolet rays, may signify room for free
flight and maneuvering.
In his pioneering work on insect color vision in 1914, Karl von Frisch showed
that honey bees were able to differentiate accurately between several major cate-
gories of color (see Plate 24): yellow, blue-green, blue (including violet), ultraviolet,
and ‘bee purple,’ a mixture of the spectral extremes, orange and ultraviolet. Von
Frisch’s techniques employed simple Pavlovian conditioning: marked bees were
allowed to feed at a sugar source while simultaneously being exposed to a particular
color stimulus, then tested to see whether they would be attracted to the color in
the absence of the food. Finally, if this proved successful, they were permitted to
choose between the original color and a different, closely similar one to see whether
the bees could discriminate between the two. In recent years these studies have been
supplemented with electrophysiological techniques.
Fig. 7.5 Sending ‘invisible’ messages. Like semaphore flags, some courting butterflies flash
UV-based wing patterns that are invisible to vertebrate predators. (left) Three species of pierid but-
terflies whose wings exhibit regions of high ultraviolet reflectance. The male is uppermost in each
photograph and often exhibits a pronounced dimorphism with respect to the ultraviolet reflectance
trait (right column, photographed in near-ultraviolet light, 300–400 mμ) which is not apparent to
our visual spectrum (left column, photographed in visible light, 400–700 mμ). (A) and (B) are
Colias eurytheme whose main visible color is orange; (C) and (D) are C. philodice, a predomi-
nantly yellow species; (E) and (F) are C. chrysotheme, another orange species. (right) Comparison
of the average reflectance spectra for a group of in-copula males (solid line) and free-flying males
(dashed line) of C. eurytheme from the same population shows no differences. Therefore such
signals are probably not used by females to discriminate differences in male quality
7.4 Functions of Visual Communication 279
Knowledge of the color sensitivity of insects has significant implications for

understanding and interpreting visual behavior. For example, to the eyes of a honey
bee many flowers glow with color we can see only with the aid of special equipment.
Many pale-yellow flowers, such as the evening primrose, Oenothera, or the cinque-
foils, Potentilla, reflect ultraviolet light over the greater part of their petals but have
patches at the petal bases that fail to do so; to honey bees, these latter nectar guides
are sharply differentiated visually (Fig. 7.5). In addition, some red flowers such as
poppies reflect ultraviolet light so strongly as to be conspicuous to the bees although
the rich red so striking to us is totally invisible to them. Color vision probably also
plays a part in the choice of backgrounds by cryptically colored insects, and such
choices may be based on a quite different Umwelt—a useful ethological term for the
perceptual ‘self-world’ of an animal.
Some butterflies and dragonflies have particularly well-developed color vision.
Courting butterflies may be especially sensitive to color patterns and iridescence
on the wings of their potential mates composed of colors that are totally invisible
to us (Fig. 7.5). The rhythmic flashing of these wing blotches during flight is of
communicative significance during courtship but is invisible to vertebrate preda-
tors. The combination of flight pattern and wing-beat frequency would seem to
have the potential to produce sign signals analogous to the flashing patterns of male
fireflies. Patches of closely appressed silvery hairs are commonly present on the
bodies of various insects and in at least one case, the fly Ochthera, these patches
reflect ultraviolet light and seem to serve as similar communicatory signals between
the sexes.
7.4 Functions of Visual Communication
Visual systems would seem to have several advantages over other communicatory
modes. The range of possible signal variations is theoretically almost limitless; one
has only to vary, independently and in combination, such basic signal aspects as
color, form or posture, movement, or timing. In addition, the rapid adaptation rates
of insect visual receptors could be exploited in development of a wide variety of
temporal patterns, developing a system that could be started or stopped immedi-
ately. Thus, an insect sensing a predator could freeze and need not communicate
its position by any lingering image such as those that might be left by a chemical
system. But in situations where it would be advantageous to do so, the insect could
make clear its exact position, so that the receiver could respond to it in terms of
precise location as well as general presence.
Functionally, any situation involving face-to-face interaction seems a potential
candidate for visual signals. For example, many visual signs are employed in inter-
specific contexts such as defensive behaviors (especially crypsis, threat displays,
and mimicry) and in pollination. In intraspecific contexts, the great majority of
currently known examples concern visual signaling associated with reproductive
activities (see Plate 46). This may, however, simply reflect the disproportionate
amount of scientific attention that has been directed toward courtship behavior as
compared to other behaviors in which visual cues may be functioning (Fig. 7.6).
Fig. 7.6 Signaling status. Portraits of nest-founding females of the paper wasp, Polistes dominu-
lus, illustrate some of the naturally occurring diversity in the shape and size of their conspicuous
black facial spots. Experimentally altering spots with paints indicates that these wasps can use
visual signals alone to assess quality differences among unfamiliar rivals, but there is some
controversy as to whether they routinely use these badges to recognize and assess one another
It is not unreasonable to expect that many additional examples of visual signaling

will be forthcoming as attention is directed toward other behaviors where frequent
close-range interactions between conspecific individuals are commonplace.
The fact remains, however, that insects do not appear to rely upon visual sig-
nals for intraspecific communication to as great a degree as they do upon chemical
and auditory ones. The same conspicuous patterns on an insect’s body that aid in
intraspecific communication can be a hindrance when it must hide from a predator.
Thus, permanent intraspecific visual signals tend to be exploited most fully by those
insects that are relatively immune to predators. In addition, except in unusual cases
such as fireflies, visual signals are useless at night, in dark places, where blocked by
the environment, and at long distances. The short visual range and limited capac-
ity for detail possessed by the insect eye also restrict the importance of fine detail
in insect releasers. Even within its useful range, a visual signal becomes simpler
and bolder, consequently carrying less information, as distance increases. A visual
signal normally cannot be increased by pumping more energy into it, as is possible
with sound or chemical signals.
7.4.1 Aggregation and Dispersion

Behaviors involved in the distribution of members of a species in space have a
crucial bearing on their exploitation of environmental resources and thus are of
great importance. At close range, such behaviors would seem ideal candidates
for visual messages. Aggregation by insects is relatively unstudied in compar-
ison with such behavior in fish, birds, and many mammals, all of which have
been demonstrated to utilize prominent visual signals in forming and maintain-
ing such groupings. In bringing together large numbers of widely scattered insects
from over large areas, it stands to reason, however, that auditory and chemical
cues would be relied upon more heavily than visual ones. The habits of most
insects are such that their environment simply does not permit long visual ranges.
In addition, the fixed-focus compound eye is not well adapted to long-distance
vision.
Dispersal through visual mediation is only slightly better known than insect
aggregation using these cues but theoretically might be encouraged by visual
signals of several levels of complexity. Insects might simply avoid other conspe-
cific individuals at sight, or they might direct signals to one another, eliciting
withdrawal. The signals could be simple ones generated by the insect alone, or com-
pound visual situations involving perception of a complex of environmental factors
as well.
Territoriality in dragonflies (see Plate 33, top and Fig. 7.7) provides a good exam-
ple of visually mediated complex behavior. Perhaps the most striking features of the
Odonata, aside from their two pairs of intricately netted wings, are their great pro-
truding eyes. Combined with a head that can be rotated readily upon a slender neck,
they allow a degree of motion perception (for some up to 40 yards away) that is
most unusual for fixed-focus eyes. In addition, the antennae, so prominent in most
Fig. 7.7 Displaying aggression. Two males of the dragonfly, Plathemis lydia, mutually display-
ing over an oviposition site, their abdomens raised to display the white upper surface. Success in
aggressive display is correlated with abdominal whiteness (which develops gradually with age)
and therefore with sexual maturity. Males with abdomens painted black are sometimes ignored by
other males
insects, are miniscule, and their removal appears to make no difference in navigation
or prey capture. As these facts might suggest, dragonflies are indeed among the few
insects in which the sense of sight is greatly dominant over the other senses. In fact,
Odonata behavior is so visually mediated that some have facetiously dubbed them
‘the bird watcher’s bugs.’
All dragonflies are more or less selective in their breeding sites. In many cases the
male, who generally arrives at the site before the female, confines his activities over
a certain pond or stream area that he will defend against intrusion by other males.
Within this area, he will court females. On a smaller scale, damselflies will often
establish similar territories closer to vegetation or to the water surface. Sometimes
several different species of different size classes will exhibit ‘stacked territories’
over the same pond.
As with most territorial animals, the dragonfly’s territorial behavior centers upon
certain ritualized aggressive displays, backed up by physical combat as a last resort.
In aggressive display, two males of the same species recognize each other as such
and then indulge in a formalized ritual that usually concludes with the departure of
one of the pair. One example can be seen in one of the most common pond inhab-
itants in North America, a rather large dragonfly, Plathemis lydia. The males have
abdomens that are bright silvery white above, and, as with the bright colors found
in many other male dragonflies, these play an important role in male interactions
(Fig. 7.7) The two males alternately dash at one another, flashing their silvery tail
patches in a pursuit display that continues until the new arrival finally restricts his
movements to the vicinity of another site.
Unlike the redundant stereotypy of most sexual displays, visual displays that
communicate aggression are often highly variable, ranging along a continuum from
attack to flight. Such graded displays make it possible to communicate slight
changes in motivation, an important advantage in facilitating the resolution of
territorial and other conflicts.
7.4.2 Alarm
In marked contrast with auditory and chemical communication systems, specialized
visual alarm systems have rarely evolved in insects. One explanation for this may
be that most predators place great reliance upon their visual sense while hunting. It
is most difficult for a potential prey to emit a visual alarm to its companions without
also making itself more conspicuous to the predator. Thus, the commonest visual
signals eliciting alertness, alarm, and flight tend to be provided not by specialized
systems but by the very actions of flight.
Adult butterflies of several families congregate, sometimes in large numbers,
around the margins of puddles of water or urine (see Plate 25) where they appar-
ently obtain needed sodium. To a predator such concentration of brightly colored
butterflies could represent a potential bonanza, but in fact it usually doesn’t. Upon
disturbance, masses of butterflies will suddenly fly up and around, surrounding
the predator with a whirling cloud of butterflies moving in unpredictable and
chaotic patterns, then gradually settle back again only as the source of disturbance
wanes. A predator finds it much more difficult to single out particular individu-
als among the swirling cloud than to pursue an isolated individual flying away
from the group. This holds true whether one or several species are involved. At
the same time, greater protection might be expected in a larger crowd. Thus,
there would be selective advantage in the convergent evolution of similar behav-
ior among many ‘puddling’ butterfly species. (Thus although several individuals
of the same species may be present, they are not necessarily closely related, and
such behavior is better interpreted in terms of survival benefit to individuals that all
respond similarly and simultaneously to threat of danger than as some form of alarm
communication.)
True alarm–alert systems are characteristic mainly of group-living organisms
such as social insects. Their evolution here forms part of the larger question of
the evolution of all types of altruistic behavior, that is, actions that result in self-
sacrifice of an individual benefiting others of its kind; this subject is covered in
Chapter 10. Even among social insects, though, visual signals are not the most
widely used channel for alarm communication, being upstaged by chemical and
sound systems. One possible exception may be found in paper wasps of the genus
Polistes. Close approach to a paper wasp nest will immediately alert some of the
Fig. 7.8 Alarming defense.

In response to the approach of
an intruder, resident Polistes
exclamans paper wasps
assume a characteristic
posture with raised wings and
front legs waving in the
direction of the source of
disturbance. Flight, attack,
and stinging will follow if the
intruder persists in disturbing
the nest, however some of the
participants are harmless
males
residents, which respond with a graded threat display in which front legs are raised
and wings are spread and vibrating (Fig. 7.8). Other workers detect this display
and respond by themselves showing an increased state of alertness and patrolling
activity, a mobilization to meet the potential threat. Although the possibility of
auditory and chemical cues being also involved in this system has not been ruled
out, to date no alarm pheromones are reported for any Polistes wasps. Moreover,
their relatively small colony size and their exposed comb nest environment make
it likely that visual and/or auditory alarm communication would be quite effective
here.
7.4.3 Sexual Signals
Perhaps nowhere else is the role of vision in insect communication as well studied
and amply documented as in sexual behavior, where visual signals often mediate a
chain of stimulus/response interactions between the partners (see Chapter 9). Such
signals are often highly redundant and stereotyped—there is strong evolutionary
pressure against making species-identification and sex-identification errors.
It is tempting to assume that Lepidoptera, with their wings of decorative colors
and patterns, are primarily visual communicators, making little use of odors and
touch in their sexual behavior. The validity of this assumption apparently depends
upon which species one is considering. The courtship of butterflies and moths ranges
along a continuum from those with distance responses in feeding and courtship that
are mainly evoked by airborne chemical stimuli and those responding mainly to
visual stimuli. The extremes are most easily recognized by differences in the behav-
ior of males. The ‘chemical type’ follows the scent upwind in a gradual zigzag flight
to find the female. The ‘visual type’ looks for the female and approaches quickly and
Fig. 7.9 Courting queens.

Stimulus-response reaction
chain in the courtship of the
queen butterfly, a species
closely related to the
monarch. Males may be
recognized by the glandular
pouch on either hind wing,
visible as a conspicuous black
dot. Resting males thrust their
unsplayed hairpencils into
them frequently, but the
pouches do not appear to be
essential to mating
directly. An interplay of signal systems is common, however. In the common orange

queen butterfly, Danaus gilippus berenice, of southern Florida, courtship consists of
an aerial phase in which the male pursues the female and of a ground phase that
begins once the female alights (Fig. 7.9). During the aerial phase, males are first
attracted visually through a combination of female movement, color, and shape;
this attraction is quite general, and males often mistakenly pursue improper objects
such as falling leaves. Once the male has overtaken the female, however, a chem-
ically mediated phase begins. Hovering over the female with everted hairpencils
(see Fig. 6.8), the male disseminates pheromones that inhibit female flight and wing
movement. Males deprived of their hairpencils are capable of courting females but
find it impossible to seduce them.
What precise components in a visual signal are important to a courting male
butterfly? A classic study in the 1950s attempted to answer this question for the
silver-washed fritillary, Argynnis paphia, a spectacular spotted orange butterfly
common in Europe (Case Study 7.1). The question fairly begged for the use of mod-
els, the technique that has played an important role in so many areas of behavioral
study because of the ways that it allows variables to be controlled.
Case Study 7.1: Visual Communication in the Silver-Washed

Fritillary, Argynnis paphia
This spectacular spotted orange butterfly, common in Europe, illustrates how

knowledge of the nature and limitations of the insect eye can be profitably
combined with the use of models to yield insights into an insect’s Umwelt.
As with many other butterflies, male and female silver-washed fritillaries
respond to blue and yellow when seeking nectar and to green when seeking a
place to rest. Sexually active male fritillaries, however, actively pursue many
different kinds of orange or yellow-brown moving objects. Only when within
about 10 cm of the object do they discriminate further, turning away unless
they encounter odor cues from a female.
In the 1950s, a German scientist, D. B. E. Magnus became interested in
unraveling the factors behind this male response. Upon learning that many
butterflies would court paper models dangled on a string from a wand, he
built more sophisticated versions (Fig. 7.10). The concept was simple but
effective—systematically varying different aspects of the model on each arm,
then tabulating male choices, would allow him to determine the relative
importance of aspects of the visual signal.
Fig. 7.10 Modeling courtship. Two of the several motor-driven carousel versions that Magnus
used to present different stimulus patterns to wild male silver-washed fritillary butterflies, Argynnis
paphia. The carousel arms carried either (right) a flapping butterfly model or (left) revolving
cylinders with alternating colors
To test his initial apparatus, Magnus presented male fritillaries with mod-
els of the same general form and color as females and with flapping wings.
Males readily followed the moving dummies and flew off only when they
were close enough to discover that the dummy did not smell like a female frit-
illary. Satisfied by this, Magnus methodically began to vary different visual
aspects of the dummies on the two arms of the carousel. He soon confirmed
that shape made no difference; a circle or triangle was just as effective as a
‘butterfly.’ A wide range of colors and patterns would elicit the first phase
of male courtship. Evidently the object only had to be moving and colored
somewhat like a female fritillary.
Size of the model mattered greatly, however (Fig. 7.11). Smaller models
were followed less than those as large as a real female. Surprisingly, models
that were twice the size of a natural female were even more effective; models
four times natural size were better yet.
What about wing beat? When he varied the wing-flapping rate, Magnus
discovered that the particular type of motion was less important than the
rate of flickering perceived by the male butterfly. On the ends of the arms
of his apparatus, he replaced the model butterflies with a much simpler type
of dummy—a mere rotating spool with alternating bands of color. At first, he
used segments of female wings on the spool but found that to a male fritillary,
a pattern of pure orange and black was even more attractive.
Fig. 7.11 Supersizing. Results of four different experiments comparing approaches of the male
fritillary butterfly, Argynnis paphia, to nonmoving colored paper dummies of different sizes. The
larger than natural size models are clearly more stimulating in each trial. Sizes were presented in
different combinations in each experiment; for example, the dashed line trial shows percentages of
113 males that responded to each of four sizes: 1/8, 1/2, normal, and 2×
Modifying his apparatus once again, Magnus developed a way to vary the
rotation of the spools on each carousel arm independently. When he charted
males’ choices between spools moving at different rates, he learned that
increasing the rate of flicker above a female’s normal wing-flapping rate also
made the model more attractive. In fact, the improved attraction continued to
increase even at flicker rates far greater than those that a flying female frit-
illary would be physically able to attain. Only when the flicker rate climbed
above about 140 stimulus changes per second—the upper limit of resolving
power of the butterfly eye—did the attractiveness begin to decline.
Thus, Magnus concluded, for a male fritillary the ‘ideal’ female would be
up to four times his size, pure orange, and flickering her colors as rapidly as
he was able to detect them.
Not only Magnus’ work but a great many studies with models have shown that
when it comes to attracting the opposite sex, ‘size matters’. Of course, simple phys-
ical impossibilities and opposing selective pressures make it unlikely that a female
like that in Fig. 7.11 would evolve. However, supernormal stimuli (see Chapter 2)
have been demonstrated many times in behavioral studies, and their experimental
use can be very helpful in interpreting the underlying physiological mechanisms
involved in stimulus filtering.
Over the years since Magnus completed his work, others also have confirmed
that visual long-range orientation of male silver-washed fritillary butterflies only
serves to bring the potential mates together. At close range, female pheromones
are necessary to initiate male courtship behavior, an apparently unusual situation
in butterflies. Receptive Argynnis females release a pheromone and when males
perceive it, they no longer respond to color. Furthermore, although receptive females
react to approaching males by directing their abdominal glands toward them in what
looks like a visually oriented response, this behavior might be strongly influenced
by perception of male forewing pheromones.
7.4.4 Multimodal Signaling
Traditionally, studies of the impact of insect sensory systems on various behaviors

have assessed the influence of a single trait in a single system. This is a very practical
choice that limits the number of variables that must be considered. However, this
approach has two problems. One is that it does not reflect the rich sensory input that
an insect is receiving during that behavior. The other is that it does not address the
evolution of multiple traits under complex selection regimes.
One of the themes of this book is that whenever a behavior is important to sur-
vival, it tends to rely on more than one type of signal. Consider the matter of insect
courtship, a subject we’ll be returning to in Chapter 9. Relying on many sources of
information was once thought to be an unstable evolutionary strategy for females
engaged in mate choice. However, it now appears that mate choice may generate
and maintain multiple secondary sexual traits in many animals, and that it is actually
quite adaptive.
If female butterflies, for example, choose between males—and considerable evi-
dence suggests that they do—then evolutionary theory suggests that they may be
receiving some benefit from this choice. Benefits generally come in two types:
material (or direct) benefits, and genetic (or indirect) benefits; examples of female
choice on each basis have been found. In the case of many butterflies, the benefit is
material, and comes in the form of a spermatophore; along with their sperm, males
deliver accessory gland secretions that contain protein and nutrients a female is able
to absorb and use for oogenesis and body maintenance. This is a significant gift,
because butterflies are on a ‘sugar high’; they are incapable of ingesting proteins as
adults and must completely rely on reserves they packed away as larvae. Thus, for
a female, receiving a spermatophore results in a higher oviposition rate and lifetime
fecundity and longevity. For a male, it is a significant investment; recently mated
males produce only about 40% of the quantity of material produced by males that
have not recently mated. Thus, there should be considerable pressure on both sexes
to ‘get it right’ by using every discriminatory tool at their disposal, not only during
courtship but even before it begins.
Many studies of visual sexual communication in butterflies have agreed in finding
movement to be universally necessary to elicit male butterfly courtship; Colias eury-
theme is no exception. Like many other butterflies, males tend to be most responsive
to the general color of their females’ wings, but complex color markings and pattern
details seem to have little effect, at least in this context. What are important, though,
are bright iridescent ultraviolet (UV) patterns that are only found in males and can
be seen over great distances.
Vision is not the only sensory modality, however. Pheromones are usually dis-
cussed in the context of night-flying moths, and wing displays in the context of
day-flying butterflies, and while these are true, butterflies also produce cuticular
hydrocarbon pheromones during courtship. Because UV signals can be seen over
a distance and cuticular hydrocarbons require proximity, it has seemed quite rea-
sonable to think that the ultraviolet signals of butterflies are used primarily to avoid
interspecific mating encounters and contact pheromones are used for intraspecific
female choice.
Newer evidence from several studies with Colias eurytheme suggests, however,
that bright structural coloration in the form of high UV reflectance plays a major role
in female mate choice. Under field conditions using free-flying males with naturally
occurring individual differences in UV brightness, female C. eurytheme preferen-
tially chose to mate with relatively younger males based on the greater brilliance
of their structural UV coloration. (As another way of looking at the situation, the
brightness of UV reflectance also was the strongest and most informative predictor
of male courtship success.)
At the same time, many studies have shown that C. eurytheme males produce
lipid-based cuticular hydrocarbons that they release when they brush their wings
against the female’s antennae. Individual males vary in both the quantities and
relative proportions of the three chemicals that comprise the pheromone, and the
characteristics of pheromones from field-caught mating males differ from those of

non-mating males. Moreover, there is considerable behavioral evidence that these
pheromones increase female receptivity and ganglia activity.
How does this tie together? As a male butterfly grows older, his chances of
having mated increase, and thus there is a greater chance that he will provide a
spermatophore with fewer proteins; he will also mate more slowly, increasing a
female’s vulnerability and exposure to predators. However, his wings will also have
become more worn, decreasing the brightness they reflect. Thus, in the parlance
of the evolutionary biologist, ultraviolet brightness is an honest indicator of mate-
rial benefit and females can, and apparently do, take advantage of this signal. Are
the pheromones also honest indicators? Perhaps not; one component that has been
investigated does not diminish with age. Is the pheromone a back-up system to the
UV signal, or does it provide additional information about other aspects of a male?
The question remains to be answered.
Possibly vision and chemical cues are still not the only systems employed. In a
study of the Asian corn borer moth, Ostrinia furnacalis, a Japanese research team
has discovered that courtship in this species includes extremely quiet ultrasonic
songs produced when the male rubs specialized scales on his wing against those
on his thorax, with the wing membrane underlying the scales possibly acting as a
sound resonator. The song increases his mating success by suppressing the female’s
escape behavior. Such vibration-based communication, which seems to point to a
whole undiscovered world of private messages, is the subject of the next chapter.
Chapter 8
Mechanocommunication
8.1 Introduction
Male grasshoppers and crickets stridulate, rubbing body structures against one
another to produce their familiar clicking calls, to which female conspecifics
respond. Death’s-head hawk moths squeak a warning to interlopers by forcing air
out of the pharynx through a short muscular proboscis. Male stink bugs challenge
rivals and court females by drumming their bodies upon a host plant; females listen,
then respond with their own species-specific calls.
Among all terrestrial animals, only vertebrates and insects have a rich sense of
hearing—the ability to detect those time-varying changes in air pressure that we
familiarly experience as ‘sound’ and to respond behaviorally in a biologically rel-
evant context. If one restricts this definition further by requiring hearing must also
involve ears—specialized organs sensitive to fluctuating airborne sound signals—
hearing has evolved in at least seven insect orders, including all the major ones
but Hymenoptera. However, expanding the definition to include detection of sound
waves through water and solids, then the list of insects with hearing expands
enormously.
Like the Russian nested dolls called matryoshkas, mechanocommunication keeps
revealing new layers of complexity and subtleness as humans develop the technol-
ogy to discern it. Since antiquity, people have known that many insects produce
relatively loud sounds, but it has only been during the past 200 years or so that sci-
entists have realized that many insects can hear. Detailed studies of insect ears and
behaviors associated with hearing only began in the early 1800s.
The concept that insects produce and sense vibrations began, again, with rather
loud and obvious examples, such as the tapping sound of deathwatch beetles in
wooden beams. However, paralleling the trajectory of so many other areas of sci-
entific endeavor, the development of new instruments for detecting and recording
acoustic signals (particularly those outside the realm of direct human perception)
has led to rapid developments in the field of insect bioacoustics. It is now clear that
insects also communicate through very soft vibrations, a private language whispered
at close range.
There is still much to learn, but already it is clear that insect acoustic communi-
cation is immensely more diverse than previously realized. It includes airborne and

292 8 Mechanocommunication
Fig. 8.1 Making sounds. Vibrational signaling occurs widely in the insects. The pie charts reveal
that whether one tallies the occurrence of such signaling by number of species or by number
of insect families, substrate vibration is by far the commonest modality. Members of the order
Hemiptera display the greatest diversity in the use of different acoustic modalities
underwater sound, substrate vibrations, and water surface ripples. Current estimates
are that over 195,000 described taxa use vibrational communication alone or in com-
bination with other forms of mechanical signaling (Fig. 8.1); because it is based only
on current published reports and named species, this number is probably low.
8.2 Producing and Sending Signals
The song of a single African cicada, Brevisana brevis, has been measured at 106
decibels at a distance of 20 inches (50 cm); considered the loudest insect sound on
earth, the call is equivalent to hearing a gas lawn mower running three feet away.
8.2 Producing and Sending Signals 293
Striking the tip of their abdomens against a tree branch, nomadic sawfly larvae
make soft tapping noises that summon lone individuals back into the aggregation.
The male waterboatman, Palmacorixa nana, strokes his peg-laden forelegs across
a female’s serrated mouthparts during mating, and she responds to the sound by
remaining motionless during copulation.
Clearly, in examples such as these the insects are making sound on purpose, as a
way to communicate. However, the first insect sounds undoubtedly were accidental.
The noises made by a flying insect, the scratching of burrowing larvae, or the vibra-
tions that occur when an insect crawls across a leaf all make sounds. Because insects
have a durable elastic cuticle, chance sound production occurs during locomotion,
feeding, or cleaning movements.
For communication to take place, however, there must be a sender and a receiver,
and the way in which a sound is produced must relate in some way to the manner in
which it is perceived. Many insects today produce sounds in the process of defensive
behaviors, such as the explosive anal discharges of bombardier beetles (see Fig. 5.9)
and the buzz of stinging bees and wasps. No one really knows whether these insects
make such noises in order to add to the effectiveness of their other defenses, but
such behaviors suggest a way in which some of the earliest sound communication
among insects may have arisen.
8.2.1 Sound Creation
Moving their forewings back and forth, katydids sing their familiar choruses; one
rainforest species sings in the extreme ultrasonic range, making the highest calling
note of any arthropod. Booklice, lacewings, stoneflies and many other insects have
taken a different route, tapping the substrate. Virgin queen honey bees produce a
‘piping’ sound by thoracic sclerite vibration; other queens still in their larval cells
respond with their own piping in short pulses at a lower frequency.
In some insect or other, almost every body part has become modified to produce
sound, and the majority of insect species produce communicative sound at some
stage of their life cycle, using an enormous variety of mechanisms. Various ways
to categorize these have been proposed. One is by the method by which they are
produced (Table 8.1).
Clicking and buzzing insects seem to be everywhere, so it is probably no surprise
to learn that friction-based methods of sound production predominate, particularly
among the Orthoptera, Hemiptera, and Coleoptera (Fig. 8.2). Rubbing of one body
part against another, or stridulation, is so common in fact that some authors have
broadened their definition of this useful term to include any sound produced by an
insect. One body part, the file, is usually a series of pegs, teeth or ridges; the oppos-
ing part, the scraper or plectrum, is generally a single edge or ridge. Associated
with many frictional sound production devices are various types of resonating sys-
tems that impart distinctive features to the resultant sounds. Frequency and pattern
may vary tremendously. Crickets produce some of the purest sounds known for
any insects and, in general, the simplest patterns; each movement of their wings
produces one pulse of sound. In other insects with more complex songs, such as
Table 8.1 Classifying insect air-borne sounds by the way they are made
Production mechanism Selected examples Use in communication
A. As a by-product of another activity

Feeding Wood-boring beetle larvae Spacing individuals
Emergence Megarhyssa wasps Sexual attraction
Flying Schistocerca locusts; Heliothis Maintaining aggregation;
moths, many mosquitoes sexual attraction
B. Impact against substrate (drumming)

Head Termites; deathwatch beetles Alarm; sexual attraction
Abdomen Booklice Female mating call
Legs Ametrus camel crickets Sexual duets
C. Friction (stridulation)
Many different body Widespread, especially in Sexual and aggressive
parts are used; Orthoptera, Hemiptera, Coleoptera behaviors
mechanisms are
essentially similar
for all.
D. Vibration of a membrane
Muscle-driven tymbals Cicadas, pentatomid bugs, arctiid Sexual and aggressive
moths behaviors, aggregation,
predator confusion
Thoracic sclerites Virgin queen honey bees Identification
E. Pulsed air stream

Epipharynx and Acherontia hawk moth Disturbance
pharynx
Spiracles Gromphadorhina cockroaches Disturbance, courtship
tettigoniid grasshoppers, each wing movement can produce many sound pulses.
Because of the mass and viscosity of water, the sound-producing structures of
aquatic insects that stridulate tend to be heavier and more massive than those of
terrestrial insects.
A second sound production method—using muscles to drive a vibrating mem-
brane, or tymbal—is most common among Homoptera and also occurs in some
Hemiptera (Pentatomidae) and Lepidoptera (Arctiidae). When the tymbal muscle
contracts, the tymbal produces a single sound pulse; when it relaxes, the tymbal pro-
duces another. The mechanism is analogous to a rounded tin can lid pressed inward
with one’s finger and then released. The insect’s tymbal muscle contracts with such
rapidity (170–480 contractions per second have been recorded) that to human ears
the sound output appears continuous. In a variation on this method, other insects
vibrate still other body parts.
In a third common method, many insects make noises by striking some body
part against the substrate. The death-watch beetle, Xestobium rufovillosum, attracts
mates by bending its head down and banging it against the floor of its wooden bur-
row seven or eight times a second, creating a tapping or ticking sound that can
be heard in old building rafters during quiet summer nights. Associated with quiet,
8.2 Producing and Sending Signals 295
Fig. 8.2 Chirping. Males of

the snowy tree cricket,
Nemobius fultoni, produce
their very audible chirps by
scissoring the wings rapidly
from side-to-side which rubs
the stridulatory file, located
on the underside of each wing
on a thickened vein (between
the joined arrows), against a
scraper on the medial edge of
the other wing. Sound is
produced only during the
closing stroke. The hearing
organ or tympanum is visible
on the front face of each fore
tibia (single arrow)
sleepless nights such as the vigil (watch) kept next to the dying or dead, they became
an omen of impending death.
In addition to these three ways of producing relatively loud sounds, one must add
the many methods by which insects produce a world of softer, closer-range sounds.
This brings us to the matters of distance and substrate.
8.2.2 Distance and Substrate
The love song of tephritid and drosophilid flies depends upon short pulses of air that
are only detectable at ranges of about 10 mm or less, driven by piston-like motions
of their small wings. During courtship, males of the parasitoid wasp Nasonia vit-
ripennis produce acoustic pulses that coincide with wing vibrations, but curiously,
the wings are not the source of the pulses because altering their surface or removing
them does not alter the signal. Crickets such as Phaeophilacris that dwell in cavities
no longer use airborne sound for calling and courtship; their wings make sound-
less flicks that only fan the air, but conspecifics can detect these motions from ten
centimeters away.
Another way to look at sound production is to consider the physics of vibration
transmission. This results in three broad categories: near-field sounds, substrate-
borne sounds, and hearing (or acoustic far-field sounds, which includes both
airborne and waterborne sounds).
Near-field sound involves displacement waves in air or water that occur close to
a sound (generally, a few body lengths from the sound source). At this short dis-
tance and small scale, sound production pushes air particles back and forth, making
low frequency waves that do not travel very far. The particles that are displaced are
highly directional, and displacement occurs in the same direction as the radiating
sound wave. Mechanoreceptive hairs are most often used to detect these particle
movements. Near-field communication is probably more common than scientific
literature would indicate, because technically, near-field communication is a diffi-
cult phenomenon to measure. Many good examples already have been identified.
The antennae of many insects function as near-field sound detectors. Specialized
dorsal hairs on some caterpillars can detect the near-field sounds produced by
the beating wings of a flying wasp; in response, the caterpillars freeze or drop
from a leaf, thus evading the potential parasite or predator. Similarly, cockroaches
can sense approaching predators through particle-sensitive setae on their cerci
(see Fig. 2.3).
Substrate-borne sound is vibration by its most restrictive definition. These sig-
nals have also been called seismic communication. Most insects can probably sense
vibration, and thus in this wider sense, one could say that most insects can hear.
(Section 8.3.1 returns to this subject.) However, airborne and waterborne sounds are
the sort we are most accustomed to thinking of as hearing. They involve ears that
are sensitive to traveling waves of changing pressure in air and water, known as the
acoustic far field.
Combining communication by air and water vibration is somewhat artificial,
because insects that produce sound under water must contend with the physical
problems of sound transmission in a relatively dense, viscous medium with sharp
boundaries. Aquatic sound fields also differ from those in air because the extent of
the near field is approximately four times greater in water, extending to a greater
distance than most aquatic insects communicate over.
8.3 Receiving Signals
In Kansas, great numbers of female cicadas are attracted to a tractor with its motor
running. In France, another cicada species responds to whistling. Female corixid
bugs of certain species come to the ultrasonic sound of a frequency generator.
Parasitic wasps that lay eggs within wood-boring larvae can locate their prey with
great accuracy even through an inch or more of bark.
What is happening? A simple answer easily comes to mind. These invertebrates
are hearing, and responding to, sounds. Probably the sounds are very similar to
those that would normally serve some communicative function for them. But if these
examples seem to clearly involve hearing, what should one make of the following
ones? When a noise is sounded near some types of caterpillars, they react by rear-
ing up either the anterior third of the body or sometimes the tail end; decapitated
caterpillars, or even isolated body pieces, show the same reaction. Whirligig bee-
tles normally swim about on the water surface film in freely moving swarms; when
their antennae are altered, individuals collide. Migratory locusts have hairs on the
fronts of their heads; stimulation of these always results in flying. How many of
these examples should be called hearing? Any? Some? All?
8.3 Receiving Signals 297
Table 8.2 A simplified functional classification of insect mechanoreception
Sub-sense Information content Receptors Remarks
Gravity and Spatial relations Displacement of various Cooperative with vision

pressure body parts
Movement Stabilization against Halteres; mass of ear, Surroundings as well as
disturbing forces, head and hair plates insect’s own
especially rotation in the neck movements
Current Orientation to currents Johnston’s organ; Often linked with
sensory hairs on head activation of sense of
and antennae smell
Touch Temporal and spatial Single hair sensilla Most primitive
distribution expression of
mechanoreception
Vibration Alert; social and sexual Sensory hairs on tarsi; Closely related to
messages chordotonal sensilla hearing, frequently
of legs; subgenual using same receptors;
organs very slow vibrations
may be considered
tactile
Hearing Phonotaxis, Auditory hairs and Hearing and sound
phonokinesis; sexual sensory spines; production often
behavior: attraction, Johnston’s organ; linked
alerting, localization tympanal organs
Help in answering these questions can be obtained by briefly considering the

whole of mechanoreception, the perception of any mechanical distortion of the
body. This broad category includes any reaction to pressure. Whether the body is
mechanically distorted by touching an object, being pulled by gravity, or being hit
by vibrations traveling through the air, water, or substrate, the task for receptors is
essentially the same—reaction to pressures of different sorts. Table 8.2 presents a
simplified functional classification of mechanoreception. The range and sensitivity
of all this equipment strikingly illustrates what an immense amount of information
an insect actually has about the outside world; without a doubt, mechanical stimuli
are involved in more behavioral activities than any other type of external stimulus.
At their most basic level, the majority of the sensory receptors of insects are quite
similar in form, namely, a small structure with a single sense cell and a single nerve
fiber. In fact, many of the tiny bristles or hairs that occur to one degree or another on
most insects are sense organs. These ‘little sense organs’ or sensilla (see Fig. 6.2),
provide the armor-encased insect with sensitive points of contact through its quite
insensitive cuticle. On the front of an insect’s head, sensilla perceive air movement.
Between head and thorax, they perceive gravity and position. On the antennae and
other appendages, many sensilla are touch receptors. On the tail filaments of some
insects, they clearly serve as hearing organs. Furthermore, in places sensilla may
join together to form a loose field or dense pad, functioning together as sense organs
of higher order. Inside the body, homologous sensilla may also join to form still
other types of mechanoreceptors.
In theory, these receptors fall into two general groups, those yielding information
about an arthropod’s position in space (the subject of Chapter 3) and those involved
in communication with other living organisms, the subject of this chapter. In prac-
tice, however, the divisions are often less clear-cut. One reason is that the same set of
information can be used for many purposes; for example, touch receptors function
in such diverse activities as avoiding obstacles, fighting, and copulating. Another is
that similar-appearing sensilla in different locations clearly serve different purposes.
All mechanoreceptors fall into two functional classes: pressure sensitive and
velocity sensitive. Pressure-sensitive hairs show a repetitive neural discharge during
a static deformation. In simpler language, they continue to fire all during the period
of time in which they are bent. Most common on those body areas where position
is important, they are the proprioceptors. Because they adapt slowly, proprioceptors
are not suited to registering sudden stimulus change. However, they give an accurate
measure of stimulus intensity, transmitting information about the state of muscles
even when contracted for long periods of time. Proprioceptors help the insect to
maintain its position, both the relation of various body parts to each other and the
relation of its whole body with respect to gravity (see Fig. 3.6). Other receptors,
particularly visual and tactile, assist in this capacity.
Velocity-sensitive sensilla, on the other hand, fire only while the stimulus is
changing, such as when a sensory hair is deflected, moves back to its original
position, and then is deflected again. Such waves of deflection are produced most
commonly by oscillations—vibrations that cause ongoing alternate compression
and expansion of the adjacent medium. Not surprisingly, tactile and auditory senses
depend upon this class of receptors. Velocity sensitive mechanoreceptors usually
have a very rapid adaptation rate and thus provide less accurate information about
differences in stimulus intensity. They are well suited, however, to record temporal
patterns of stimulation.
Because sound is vibrations spreading through a medium away from a source of
mechanical disturbance, hearing is a mechanical sense much like touch. However,
hearing and touch differ in the types of pressure alternations involved. Sound stim-
ulation has a phasic nature, whereas touch stimulation has a non-fluctuating or
irregularly fluctuating nature. More simply, the two senses differ on the basis of
distance. Touch is mechanoreception that involves contact; hearing is mechanore-
ception that involves a distance between sender and receiver. The differences thus
are akin to those between taste and smell in chemical communication.
8.3.1 Vibration
Water striders and backswimmers read the water’s surface waves for cues to food
and mate location (see Plate 26); they orient to the source of the pulsation through
differences in arrival times of the waves at receptors located on different legs.
Female Ephippiger bush crickets orient both to the audible songs of conspecific
males and to the tremors they produce; if the songs halt but vibration contin-
ues, females are still attracted, but only if the male is close by. Some solitary
plant-dwelling stink bugs find and recognize their mates by quivering their
abdomens to send signals through their host plants.
If the term ‘hearing’ is restricted to the detection of airborne sounds by special-
ized receptors, all other sound reception becomes relegated to a ‘vibration sense’ but
in the insect world sound detection and vibration detection often overlap. As a good
example, consider the chordotonal sensilla that are widely distributed in the insect
body, occurring in mouths, legs, wing bases, halteres, antennae, abdomens, and tra-
cheal systems. These sensilla are unique to the insects, but they have been found in
every insect in which they have been sought. Subcuticular, with no external evidence
of their presence, chordotonal sensilla were originally thought to be exclusively
audioreceptors. However, it now appears that they are derived from proprioceptors,
and at least some have a mixed function.
Upon the antennae, chordotonal sensilla are grouped into the Johnston’s organ
that detects movements of the shaft of the antenna. In culicine mosquitoes where
it was first discovered, and in midges as well, the Johnston’s organ is enormously
developed and has a clearly auditory function. On the other hand, in most other
insects where it has been studied, it appears to act primarily as a tactile organ that
relays several sorts of mechanical information. In aphids it is used in the control
of flight. Cutting off the antennal flagellum beyond the Johnston’s organ causes an
aphid’s flight to become erratic, but when an artificial antenna is reattached, normal
flight resumes.
Generally, insects also have chordotonal organs at various locations along their
legs, located in bundles in such a way as to sense vibrations of the surface upon
which they are standing. One set of particular importance are the ‘below the knee’
or subgenual organs. These are extremely sensitive at their optimal frequencies
because the amplitude of displacement required to stimulate the organ is very small.
Since they can localize a source of vibration that is not in contact with the insect,
subgenual organs could be called organs of sound perception as well as touch per-
ception. They may be particularly important among the wide variety of borers that
stridulate, transmitting vibrations through their wooden tunnels.
The fact that insects can detect vibration does not necessarily mean that they
communicate among themselves with it, of course. Most studies of vibrotaxis,
movement in relation to a substrate-borne source of vibrations, have involved par-
asites and predators honing in on and exploiting vibratory signals given off by
potential hosts, often unintentionally in the course of life activities. Thus, for exam-
ple, larvae of the leafminer Phyllonorycter produce wriggling vibrations that differ
in temporal pattern from the vibrations they make as they simply move through an
apple tree leaf; their pupae wriggle as well. One of their parasitoids, the eulophid
wasp Sympiesis, hones in on these signals to locate both life stages of their host.
Examples in which vibration is used for communication within a species have
been reported less commonly, but seem particularly to include insects that exhibit
some degree of sociality and regularly share some common substrate. Often, as
in the honey bee dance language detailed at the end of this chapter, the vibration
component of the communication is part of a larger picture that includes other
communication modes. However, a particularly well-studied example of primary
vibrotactic communication involves treehoppers, a group of membracid bugs that
feed on the sap of trees and shrubs (see Plate 27). Some inventive research with
Umbonia crassicornis (Case Study 8.1) provides a useful model for studying natural
mechancommunication behavior in the laboratory.
Case Study 8.1: Vibrotaxis in the Thorn Bug,

Umbonia crassicornis
To find potential mates, capture prey, or avoid predators, it’s not enough to just
notice vibration; an insect needs to be able to locate the vibration’s source.
Can it be done? If so, how?
Shaped somewhat like little thorns, treehoppers spend most of their lives
relatively motionless on host plant stems (a behavior that reinforces this
resemblance), leaving only to disperse or find mates. Fascinating in their vari-
ation and complexity, treehoppers have become the subject of quite a bit of
study. They exhibit the entire range of social behavior that ranges from asocial
individuals, to aggregations of nymphs and adults, to highly developed mater-
nal care with parent-offspring communication. The nymphs and new (teneral)
adults of some gregarious and subsocial taxa are colored aposematically (see
Plate 27); ant-treehopper mutualisms also are widespread.
Working with the thorn bug, Umbonia crassicornis, Reginald (Rex)
Cocroft found that when he simulated a predator’s approach, aggregated
nymphal offspring of this little treehopper produce substrate-borne vibration
signals that elicit their mother’s anti-predator behavior. Vibration was, he
showed, the primary communication mode; blocking vibration transmission
between signaling nymphs and their mother abolished the mother’s response.
The nymphs also communicated to one another, spreading the alarm
throughout the aggregation. It was clearly a complex system rather than a
simple response to random patterns of vibration because nymphs in otherwise
undisturbed aggregations signaled only in response to signals coordinated into
synchronized group displays. When signals from their siblings were played
back to undisturbed nymphs, they reacted as though they had received an
alarm. However, it was unclear whether they actually knew the direction from
which the threat originated.
A full-grown Umbonia is only a centimeter long. Could animals as small
as this send and receive vibration signals that encode location? To some scien-
tists it seemed unlikely on theoretical grounds. They pointed out that vibration
interactions take place at close range. Insect legs are close together anyway,
and the high conduction velocity of vibration waves will further minimize
arrival time differences between receptors. Cases of localization that had been
reported were in relatively larger arthropods such as orb-weaving spiders,
scorpions, and locusts, and in some cases, it also seemed that direction cues
were being gained from amplitude differences along a gradient. Whether the
same results would apply directly to smaller scales seemed questionable.
Experimental studies seemed to back up these theoretical concerns. When

researchers placed large desert scorpions on substrates in which conduction
velocities were five or more times higher than in their normal sand substrate,
the scorpions’ responses were no longer directional, apparently because of
insufficient time delays between their receptors at this larger scale. Wouldn’t
similarly short delays be experienced in normal substrates by insects whose
receptor distances were five or more times smaller?
Perhaps, Cocroft reasoned, insects detect vibrations by a mechanism that
does not depend on large amplitude and time differences among receptors.
Looking to other acoustic sound localizations in insects, he was struck by
the case of another small insect, a parasitoid fly whose mechanically coupled
ears react to the interaction of two modes of vibration. Might the treehop-
per’s whole body be acting in an analogous way? Considering Umbonia from
a mechanical viewpoint, it was like a mass (its body) resting on a set of
springs (its legs). Whenever two or more modes of vibration might rhyth-
mically swing the body about, its mechanical response would differ in ways
that would depend on the direction of wave propagation.
Earlier studies had shown that that bending waves are transmitted along
a plant stem that is mechanically vibrated in a plane perpendicular to the
stem’s long axis. By extension, it seemed logical that the signals of insects
that vibrate the stem also would be transmitted in the form of bending
waves. However, before he could examine the influence of vibration direc-
tion, Cocroft need to show that thorn bugs actually were communicating with
bending waves.
A straightforward way to test this would be to measure the propagation
velocity of signals the insects produce, and learn whether the measured values
were proportional to the square root of frequency, as expected for bending
waves. Cocroft collected adult U. crassicornis and placed them on potted host
plants in his laboratory. Then he attached two accelerometers with wax onto
the underside of a horizontal stem and placed female treehoppers individually
on its upper surface. By amplifying and digitizing the signals, he confirmed
that this was indeed the case.
When he tried to examine the influence of direction on the biomechnical
response to substrate vibration, however, Cocroft ran into problems. Natural
stems vary in filtering properties and identical sources would be difficult to
obtain; he would also need to present each stimulus both in front of and
behind the insect. Alternatively, between each pair of stimulus presentations
he would need to remove the treehopper, turn it 180◦ , and replace it in the
same location and posture on the stem. Both approaches seemed destined to
introduce too many unknowns. Clearly what he needed was an artificial plant
stem where propagation direction could be switched electronically between
measurements. With colleagues, Cocroft designed a simulator consisting of a
short length of dowel and two piezoelectric actuators (see Fig. 8.3). The actu-
ators were driven with two time signals that created the relationship between
slope and displacement that are characteristic of steady-state bending wave
motion. This would allow him to precisely control the frequency, intensity,
and direction of the vibration stimulus without moving the free-standing,
stationary bug.
Fig. 8.3 Picking up

vibrations. A female
Umbonia rests upon Cocroft’s
artificial stem. To mimic the
properties of bending waves
and allow their propagation
direction to be switched
electronically between
measurements, one end of a
pair of piezoelectric actuators
was attached with epoxy at
each end of a short dowel and
the other was mounted with
epoxy on a fixed metal plate.
As the actuators were driven
with two random noise
signals, a laser opposite the
dowel’s center measured
motion in the plane of the
long axis of the actuators and
the dorsal-ventral axis of the
treehopper
With laser-Doppler vibrometry, the researchers measured points on the

substrate and on the treehopper’s thorax and middle leg. Transfer functions
showed that vibrations resonated in the insect’s body at lower frequencies
but attenuated at higher frequencies. The measurements differed substan-
tially depending on whether the stimulus came from in front of or behind
the insect, and in addition to two modes of vibration along the length of its
body, in some cases the treehopper also experienced a side-to-side rocking.
Although behavioral directionality and the neural basis of vibration perception
remain to be investigated, Cocroft’s research demonstrated a means by which
straightforward comparison of signal amplitude, probably by the subgenual
and chordotonal organs, could provide a simple way for a small insect to
determine the direction of a vibrational signal.
To what degree can results like those in Case Study 8.1 be generalized?
Treehoppers have a very expanded pronotum, but it does not appear necessary for
generating a basic pattern of resonance and attenuation. The transfer function for
the legs and for the thorax was similar, and when Cocroft removed most of the
pronotum from three treehoppers, mechanical directionality remained unchanged.
Therefore, it seems as though other insects with different morphologies would show
similar patterns in mechanical response to vibration.
The bigger problem with generalizing lies in the many ways that environmen-
tal constraints on vibrotaxis can occur. For a treehopper on a plant stem, orienting
to a vibration source is essentially a one-dimensional problem. However, for other
insects such as a beetle searching for a mate on a leaf, localization is a two-
dimensional problem; results of studies on two-dimensional surfaces have given
mixed results, leading some scientists to question vibrotaxis as an explanation.
Alternatively, they say, insects may triangulate, sampling and comparing vibrations
at multiple points before orienting. Another possibility is that they may follow a
simple searching rule, making large turning angles whenever the amplitude of the
vibratory signal drops below a certain threshold. Further research will be needed
before all this can be straightened out but there is no reason, of course, to make such
explanations mutually exclusive; all insects need not follow a single rule.
8.3.2 Hearing
Some hawk moths have hearing organs in their mouthparts; while feeding on flowers
at dusk, they can listen for bat echolocation calls by inflating their palps. The South
African bladder grasshopper, Bullacris membracioides, has six pairs of ears running
along its abdomen.
A diverse group of specialized auditory organs has arisen in insects from organs
serving originally for proprioception. These organs are the ones most similar to
what most people might call insect ears. They show a great deal of morphologi-
cal diversity, and from one taxon to another, they can be found in many different
body locations (Fig. 8.4). Sometimes they are conspicuous; many other times, they
are not. Vertebrate ears are always on the head, behind the eyes and above the jaw.
How can insect ears appear in so many places? The answer seems to lie in evolu-
tionary and developmental differences. Vertebrate ears arise from the gill arches,
neural crest, and optic capsule, which confines them to the head. Insect ears, of
the other hand, have arisen through a few simple modifications to existing proprio-
ceptors and their surrounding cuticular tracheal structures. Because the chordotonal
proprioceptors are widely distributed, ears can be as well.
Despite this great diversity, each ear typically consists of the same three units:
a tympanal membrane, a tracheal air chamber, and a chordotonal sensory organ.
Fig. 8.4 Finding insect ears. A generalized insect showing 15 body locations where tympanal
organs have been identified as evolving independently among members of seven different orders.
Key: Auc = Hemiptera: Auchenorrhyncha, C = Coleoptera, D = Diptera, Het = Hemiptera:
Heteroptera, L = Lepidoptera, M = Mantodea, N = Neuroptera, O = Orthoptera. Locations of
hearing organs of Lepidoptera are most diverse, but all tympanal organs probably evolved from
proprioceptors
With the development of ears, the chordotonal sensilla have increased in number
and become attached to a thin, taut membrane that is set into vibration by waves
of airborne sound. This membrane, or tympanum, functions as a sort of eardrum. In
many nocturnal Lepidoptera that are ultra-sensitive to sounds, the typanum is so thin
that it is transparent and occurs in a protected location. By contrast, many Orthoptera
have a much thicker tympanum conspicuously positioned on the cuticle surface (see
Fig. 8.2). Tympanic organs are usually paired and they often interconnect through a
series of tracheal chambers.
In still another variation on insect hearing, hair sensilla that are primarily used
for tactile purposes often react to high-intensity airborne sounds. As a result, many
insects that lack tympanic organs are still quite sensitive to sound waves in air. Most
if not all caterpillars, for example, can detect sound and will react by becoming
immobile or thrashing as a protective response. It does however appear that those
sensilla types most highly developed for airborne sound reception differ from tym-
panic organs in several ways. Compared to tympanic organs, these hair sensilla
respond over much lower frequency ranges and fatigue fairly rapidly. They exhibit
only limited frequency discrimination, responding synchronously with stimulus
frequency over certain ranges. Finally, they tend to habituate or equilibrate, that
is, as a sound continues through time, the amplitude of the massed spikes in the
responding nerve declines.
8.3.3 Communication by Touch
The very large and complex nests of many social insects present an enigma. No
single colony member can oversee more than a small piece of the construction work
or envision the nest in its entirety. Sometimes, a nest may require a number of worker
lifetimes to complete, and each new part must be brought into balance with the old.
How can the workers communicate so effectively over such a long time period?
Who has the nest blueprint?
In the 1950s, termite researcher Pierre Grassé suggested a key to this puzzle.
Perhaps the insect nest-builders were responding to previously accomplished work
on the nest itself rather than to direct communication from their nest-mates. Then,
even if one could constantly renew the work force, the nest would be completed
‘according to plan’ because the nest structure already finished would determine
what further work should be done. In the years since, many researchers have chal-
lenged the completeness of Grassé’s explanation, but not its basic insight. The
most durable signals that communicate information between insects of the same
species are those incorporated in structures built by the insects such as nests. In his
groundbreaking book Sociobiology, Edward O. Wilson termed such communication
sematectonic, from the Greek words for sign and builder, and defined it as the evo-
cation of any form of behavior or physiological change by the evidences of work
performed by other animals, including the special case of the guidance of additional
work.
Sematectonic communication is not limited to social species. For example, in
the nests of many solitary bees and wasps the cells are arranged end to end in a
linear series. At emergence, adults always chew outward in the direction that they
are facing, which is almost invariably toward the nest entrance. In general, the nest
tunnel is not broad enough to permit the adult to turn around inside should it emerge
from its cocoon facing the wrong way. How does the mature larva correctly orient
itself with an accuracy that far exceeds that expected by chance alone?
As the wasp nest is constructed, the mother builds partitions between successive
cells. As a result of the techniques of construction, the inner and outer faces of
each partition differ in texture and concavity. The mother wasp works from the
outside, and the partition tends to become concave on the outer surface. It is also
smoothed on the outside, while the inner surface has an irregular, bumpy texture on
its convex face. By testing independently the effects of the four possible cues on
larval orientation, Kenneth Cooper was able to clearly show that concavity was the
primary cue used by the spinning larvae (Fig. 8.5); they always oriented their heads
away from the concave surface. Moreover, Cooper found that completely unrelated
Fig. 8.5 Getting cues from mother. The orientation of a solitary-wasp larva when spinning its
cocoon depends upon the way its mother built the linearly arranged cells. By facing the rough
convex side of the partition, when its development is complete the offspring will leave the nest in
the right direction. Artificially reversing partitions was the key to understanding this behavior; the
swollen end of each cocoon contains the head
wasps, including some parasitic species, orient correctly by the same cues. In twig-
nesting solitary bees, where the cell partitions tend to be rather amorphous walls of
resin or chewed leaves, the significant orientation cues seem to lie in the placement
of the pollen mass. Spinning larvae always orient cocoons facing away from the
pollen mass.
Theoretically, touch could form the basis for a complex and important com-
munication system, transmitting a great variety of messages by varying frequency,
pressure, and time of contact. However, tactile communication appears to be rela-
tively unimportant in comparison with the other communicatory modes. Most tactile
systems have one overriding limitation: the sender must be in contact with the
receiver. As a result tactile methods are usually restricted to close-range situations
such as courtship, mating, and occasionally alarm.
Except for the special case of sematectonic cues, touch communication as a
whole among insects is quite poorly known. The sense of touch is one of the most
difficult modalities to investigate in a communicatory context because it is hard for
an observer to interpose himself or his instruments in such a way as to register the
signals in the form that they are actually received without disrupting the system.
And because of this difficulty, even in cases where tactile signaling is involved,
other communicatory modes tend to receive the first attention.
8.4 The Acoustic Channel 307
8.4 The Acoustic Channel

Among insects hearing had seemed to be the exception, not the rule. This discon-
nect was probably only because hearing had been defined too narrowly. For insect
sounds the human ear has little to recommend it as an analyzing instrument. Its
frequency response range is too small, and its time constant is too long. However, as
noted above, to compare, contrast, or even properly describe insect songs, the use of
various instruments is essential, and with this continually improving technology the
extent and scope of insect mechanocommunication has turned out to be surprisingly
broad. Continued use of such technology promises to breach the gap between the
known occurrences of insect sound and the relatively little known significance of
them. It also promises fuller recognition and understanding of insect communica-
tion that is occurring in near-field and vibratory modes, which is undoubtedly going
on all around us, but unaided human ears have almost completely ignored.
8.4.1 Parameters of Insect Song

People hum melodies and whistle bird songs. But have you heard someone sing
along with an insect? Probably not. Why? One reason stands out—insects cannot
carry a tune. Most also appear to be tone-deaf; different tones are indistinguishable
to them.
Melody involves variations in pitch (the level of sounds in the scale, as defined by
their frequency), and insects sing in a monotone. The dominant frequency of most
air-borne insect songs stays at a more or less common pitch. Most insects produce
sounds that are in the upper part of the auditory range detected by their sound recep-
tion organs, where frequency discrimination appears to be rudimentary or absent.
The chirp of the domestic cricket is comparatively unusual, occupying a narrow
band of frequencies having a nearly pure tone that we recognize as a definite pitch.
Most insect sounds are noise-like, covering a broad band of many non-harmonically
related frequencies. These sibilant sounds are commonly below 18 kHz, but many
range into the ultrasonic.
If melody is a weak point for singing insects, rhythm is their strength. The
basis of insect acoustic communication lies in sound intensity and timing. Many
(and maybe most) insect songs occur as a nested pattern of time-related group-
ings of individual sounds at various volumes. The scale on which these elements
occur is normally hidden to us because the time constant of unaided human ears
is too long to discern and analyze them. For example, what we hear as a sim-
ple buzz is actually a set of individual behavioral song components (phonatomes)
and their first- and second-order groupings. Were the phonatome rate to be slowed
down enough, we could hear the individual phonatomes as countable ‘ticks’ or
‘smacks’ or, if they occurred slightly too rapidly to count, as a ‘rattle’. Because
temperature has a substantial effect on phonatome rate, the same song component
that is identified as a rattle at one temperature may be called a tick sequence at
a lower temperature. Sometimes this relationship to temperature is quite precisely

known. For example, snowy tree crickets (see Fig. 8.2) are well known to natu-
ralists as ‘thermometer crickets’ because by counting the number of chirps heard
over 15 seconds and adding 40, one can fairly accurately calculate the Fahrenheit
air temperature.
One phonatome includes all the sound produced during one cycle (one com-
plete stroke of the scraper over the file and return) during stridulation. Because it
is a behaviorally defined term, the phonatome label is difficult to apply precisely in
cases where the behavioral basis of sound production is not known. In contrast, a
pulse is the simplest element of amplitude that is convenient to recognize—a wave
train isolated or nearly isolated in time by a substantial amplitude modulation (fre-
quently, an interval of silence). With some simple sounds a pulse may be equivalent
to a phonatome, as in a single cricket chirp or one squeak of a restrained velvet
ant. In other cases, such as the meadow katydid, Orchelimum, the simplest sound
element is a pulse lasting a fraction of a millisecond, and the travel of the scraper
along the file generates a whole train of pulses that collectively are equivalent to a
phonatome.
Analogous to the way that a prism can separate the colors in light, special elec-
trical equipment can separate the components in a sound. The resultant images are
of two types, spectrograms and oscillograms (Fig. 8.6). A sound spectrogram, or
sonagram, records the frequency spectrum of a sound as a function of time; this
is commonly used to study bird songs because it permits one to distinguish pitch
differences. On the other hand, an oscillogram shows the volume of a sound; the
greater the deflection, or amplitude, of the tracing above and below the abscissa
baseline, the louder it is. This is important because insect songs often show rhyth-
mically alternating loud/soft patterns through time. Equally important, oscillograms
can resolve phonatome sequences that occur too rapidly for unaided human ears to
separate.
8.4.2 Song Synchronies
The alternating songs of crickets, the long raspy choruses of katydids by nights,
and the intense shrill buzzes and rattles of cicadas by day—these are some of the
most familiar and conspicuous insect sounds that humans recognize. Their precise
choreography in space and time often gives rise to striking alternation or synchrony
between neighbors, and one cannot help but wonder how and why they evolved and
are maintained.
Of all the acoustic responses an arthropod may give, the most widely observed
and easily demonstrated is the phonoresponse—upon hearing a noise, the insect
replies by making one. Acoustic stimulation may produce two different types of
phonoresponse, depending upon the species. In some, the responding insect alter-
nates its emission with that of the stimulus. Such alternation is frequently displayed
between two males, either at the emission of the calling song or of the rivals’ song;
it is also observed in species where a male and female emit an agreement song.
Fig. 8.6 Measuring songs. Sonagrams and oscillograms enable a viewer to appreciate song
characteristics and differences at a glance. Sonagrams show the carrier frequency on the verti-
cal axis (ordinate) and time on the horizontal axis (abscissa). Oscillograms also represent time on
the abscissa, but the ordinate expresses fluctuations in the pattern of sound intensity (loudness),
without specifying the frequency; the degree of blackening correlates with sound intensity. (a)
Sonagram of an easily recognized human ‘wolf whistle’. (b) Song of the long-horned grasshopper,
Phlugis, displayed on an oscillogram (top) and a sonagram (bottom)
Phonoresponses involving two or more individuals of the same sex are the most
common type.
Among other insects, particularly species living closely packed together, sound
emission by one individual may set off a collective and synchronous song of all the
population. Like their counterparts among bioluminescent species, these chorusing
displays of acoustically signaling insects are recognized as some of the great spec-
tacles of the living world. Chorusing most commonly involves males, and it may
result in large numbers of insects coming together in dense aggregations. In some
cases, individuals may synchronize or alternate calling phrases for indefinitely long
periods of time, as do Cicadidae and some Orthoptera. The forces that drive such
synchrony are undoubtedly many. The role of sexual selection has been a matter of
particular interest. Many play-back sound studies with various species have shown
that when females are presented with two male calls that are out of phase, they pref-
erentially orient toward the earlier call, a type of precedence effect. The importance
of this effect on mate choice in actual choruses and natural populations is still poorly
known, but there is some evidence that males strive to adjust their singing so as not
to lag behind others.
Studies of chorusing insects point to the powerful role of the social context in
influencing acoustic signaling, an area of research receiving increasing attention in
the context of communication networks. Chorusing of vertebrates has been most
thoroughly studied in various frogs and birds; chorusing of invertebrates, in periodi-
cal cicadas (Magicicada). In the latter, a three-species complex shows extraordinary
chorusing behavior. In each species, a calling or aggregating song produced individ-
ually but sung in chorus activates and assembles both males and females. Not only
do the songs of each species differ in their acoustic parameters, but each species has
a different time of peak chorusing activity as well (Fig. 8.7). These two factors result
in a slow but clearly defined grouping of each species in its habitat. In their first few
days of adult activity the three species may be intermixed, but after a week or two
the grouping has become so intense that a tree may contain hundreds of cicadas of
one species and only one or two of the other two species. Such sorting undoubtedly
helps female cicadas find a mate of their own species.
Fig. 8.7 Singing one’s species. Approximate diurnal times of maximal chorusing activity in the
three sympatric species of Magicicada found in the southeastern United States
Do female cicadas also choose between the males on some basis of some quality
difference signaled by the songs? Periodical cicadas have seemed to be strong can-
didates for participating in lek mating systems (see Chapter 9). However, a study
of patterns of mating success in actively chorusing males could find no consistent
differences between mated and unmated males on the basis of size and song pitch,
and the patterns of mating success in actively chorusing males were indistinguish-
able from random mating. Whether other insect choruses are leks is a matter of
ongoing debate.
A chorus is an extremely complex acoustic scene. What kind of information
do males and females acquire amidst this cacophony? How do they respond to
these signals? The reception of unintended signals—in a word, eavesdropping—has
become a particularly interesting area of theoretical study. Key conceptual questions
concern how listeners might use overheard information to their own advantage, and
how signalers might strategically direct their signal production to minimize possible
eavesdropping in certain contexts.
8.4.3 Active Acoustics

Based on early studies, hearing was thought to be a basically passive process. In
fact, Georg von Bekesy won a Nobel Prize in 1961 for showing that mammalian
hearing is place-coded; specific frequencies of sound excite hair cells correspond-
ing to their location on the basilar membrane in the cochlea. While this basic idea
still stands, more recent discoveries are causing a major paradigm shift in scien-
tists’ perception of auditory processing. At the level of cellular mechanisms, it is
becoming increasingly clear that acoustic reception is not a basically passive pro-
cess it was once believed to be. Instead, hearing involves active processes that take
place as hair cells acts like molecular micro-machines to enhance auditory sensitiv-
ity and tuning. These active processes occur only in living organisms, and for this
reason they have only come to light as newer technology has allowed such advances
as laboratory culture of vertebrate organs, tissues and cells in vitro and noninva-
sive cellular-level studies of hearing in intact, living insects under relatively normal
behavioral conditions.
In retrospect, active hearing in vertebrates seems perfectly obvious because of
a phenomenon we have all experienced: selective attention. Consider the situa-
tion in a crowded room; many conversations may be going on simultaneously,
but at the mention of your own name, instantly your attention zooms in on that
conversation. Probably nearly ever animal has some type of similar response that
momentarily heightens perceptual acuity when confronted by an important stimulus
such as a social signal or an environmental noise that might indicate a predator’s
presence. However, whereas in mammals, selective attention seems to be mostly
brain-centered, in the few insects where it has been studied, selective attention is
ear-centered.
Given the extraordinary auditory world of mosquitoes (see the case study later
in this chapter), it is probably unsurprising that the best-studied example of active
acoustics in insects involves a non-bloodsucking mosquito, Toxorhynchites bre-

vipalpis, that occurs in Tanzania. Like other mosquitoes that have been studied,
adult males of T. brevipalpis detect the presence of females on the wing by ear.
When they hear a nearby female’s flight tones, they pursue her. It sounds like a sim-
ple system—cruise, catch, mate—but studies have shown that there’s significantly
more to it than this. For one thing, the wings of a female mosquito are very small
compared with the wavelength of the sound they make. For the male mosquito, this
means that if a female is flying more than a few tens of centimeters away, he should
not be able to hear her, yet he does.
To investigate this puzzle, Joseph C. Jackson and Daniel Robert turned to
Doppler laser vibrometry, a powerful technique that allowed them to study
intact mosquitoes in a completely noninvasive way. Painstakingly, they measured
mosquitoes’ antennal responses to the sounds of real mosquitoes, tethered and
untethered, and to simulated mosquito sounds played through a loudspeaker. When
the results were analyzed, they showed that male responses were nonlinear; rather
than a straight-line response to female wing-beat sounds, the male antennae showed
three distinctly different levels of mechanical excitability, depending on the female’s
location. Significantly, after a female flew on by, the male’s antennal cells con-
tinue to be at heightened sensitivity for a while. (The scientists likened it to the
lingering scent of perfume, which briefly persists even after its owner has passed
by.) This mechanism would sustain and amplify the normally weak and transient
female signals.
The study revealed a remarkable twist to the mosquito auditory system—a mech-
anism that changes a male’s sensitivity to a biologically relevant sound. Because all
the responses occurred within the range at which female mosquitoes are well within
earshot of a male, it appears that this is a mechanism not so much for improved
detection as for focusing the male mosquito’s attention by making a female appear
closer than she is. Similar nonlinearities have been known in mammalian ears
(where they also enhance acuity) for some time. However, this is one of the first indi-
cations that the adaptive imperatives of functional hearing have resulted in similar
biomechanical processes in invertebrates.
Sensory tuning and selective attention will probably turn out to be quite
widespread. In 2006, James Windmill showed that the yellow underwing moth
changes its sensitivity to a bat’s calls when the moth is being chased; even after
the calls stop because the moth has escaped, its ear remained tuned in for several
minutes in case there is another attack.
It is also worth noting that one American field cricket species has been shown
to have an abruptly lowered auditory threshold at just the frequency (pitch) of its
song, about 4.5 kilocycles/second. If other crickets also have their auditory organs
so precisely tuned to their own frequencies, the songs of different genera might
be rendered almost inaudible. Then rhythm patterns of each could evolve without
reference to one another, and songs could be rhythmically quite similar without
confusion. Some weight is given to this theory by the observation that ground crick-
ets (Nemobiinae), which often live with field crickets, have songs usually pitched
quite differently from them (about 8–10 kilocycles/second). In a region such as
North America, where 40 or 50 insect species often may be stridulating simul-

taneously in the same locality, such a development would obviously be a great
advantage.
8.4.4 Sound as a Communication Method
Acoustic messages are an unusually complete and efficacious way to impart infor-
mation. They diffuse easily, resist disturbances, and exhibit the potential for creating
a vocabulary. They convey information instantaneously but they leave no lingering
traces to alert a predator or parasite. In many cases, an insect would not even need
to suspend other activities in order to transmit or receive sound signals.
Sound can be distinguished at very low levels; it also may be raised above envi-
ronmental noise by pumping energy into the system. Sound waves can travel at any
time or day or night, and not only through the air but through solid surfaces, thick
vegetation or muddy water. Sound also can be transmitted over long distances. If
one considers the distance at which signals are no longer physiologically percepti-
ble, chemical messages have perhaps the widest range and visual messages probably
the shortest; acoustic messages fall somewhere in between.
If audible sound is such a good way to communicate, why don’t more insects
use this method? One set of reasons has its roots in insects’ relatively small size.
First, for mechanical reasons small size generally results in the production of rel-
atively high-frequency signals. These attenuate and degrade more quickly than
low-frequency signals, and this limits communication range. To get around this
problem, some small insects increase their broadcast range by behavioral tactics
such as using plants as acoustic baffles, calling from elevated positions, or call-
ing from within burrows that act like horns; others have structural modifications
that allow them to produce sounds at lower frequencies than others of comparable
size.
Second, sound production may be energetically more costly and inefficient for
a small invertebrate organism. The density and elasticity of insect cuticle differs
significantly from that of air, so it takes a great deal of effort to convert the mus-
cular energy controlling a vibrating structure into airborne sound. Thus, it probably
should not come as a surprise to find that many insects have taken the route of pro-
ducing vibrations in the substrate instead. The energetic costs of tapping on a surface
are much lower than the costs of calling through the air. Tubular structures such as
leaves and stems allow information to be transmitted as bending waves, which can
travel efficiently for some distance.
Third, the narrow separation of insect ears makes it difficult to nearly impossible
for most insects to localize the source of an airborne sound by using differences in
sound wave arrival times at their ears. Some insects have gotten around this prob-
lem by adopting a different mechanism, pressure-difference receivers. However, this
has other consequences for acoustic communication, such as limiting the frequency
ranges that can be effectively used, as well as interacting with mechanisms used in
sound pattern recognition.
Another set of reasons arise from the ease with which potential predators and
parasites can home in on sounds. For example, Mediterranean house geckos ori-
ent to locations where male decorated crickets, Gryllodes supplicans, are calling.
Even though the male crickets call from burrows where they are protected from the
geckos, this behavior enables the geckos to intercept and consume female crickets
that also respond phonotactically to the cricket calls. Eavesdropping can occur with
any sensory system, of course, but acoustic messages are particularly apt to provide
clues to location.
Unlike bird songs that often feature a learned component, the song of arthro-
pods appears to be totally hereditary. What plasticity is present is preformed; it
cannot be modified through learning. Thus, song can be considered to be a species-
characteristic feature very much like morphological features. While external factors,
especially temperature, may considerably modify the song, these modifications are
only simple responses to changes of environment much like changes in pigmen-
tation or ornamentation. This means that with even a simple mutation, song may
become an important reproductive isolating factor. When new arrangements arise
(a rare but not impossible circumstance), they are capable of producing instant iso-
lation of these individuals from other members of their species. Sometimes, this
isolation may lead to speciation; more often it does not. Two sympatric grasshopper
species, Chorthippus biguttulus and C. brunneus, produce hybrids in the laboratory,
but few hybrids are found in nature. Because hybrid males produce hybrid songs
that are intermediate between those of either parental species, the females of both
species shun them. Thus although hybrid males are reproductively fertile, they are
behaviorally sterile.
8.5 Functions of Insect Communicative Sounds
To develop a complete functional classification for insect sound would be a gargan-

tuan task, but a relatively simple system covers most acoustic behavior in four basic
(but sometimes overlapping) categories. All of these fall naturally into two groups,
depending upon whether the sound shows any repetition of some basic element, that
is, whether the song is patterned.
Unpatterned sounds, while they might serve as simple behavioral releasers,
cannot transmit complex information. Signals between species tend to be char-
acterized by their high intensity and lack of pattern; they are the type of sounds
often released when an insect is touched, and most alarm and warning sounds are
of this type. Patterned sound, on the other hand, is found in all cases of com-
plex intraspecific behavior mediated by sound, and pattern is their most constant
feature.
8.5.1 Protest, Alarm, and Aggression
When threatened, both sexes of the velvet ant Dasymutilla occidentalis (Fig. 8.8)
stridulate with enough volume to be readily detected by the human ear a meter
8.5 Functions of Insect Communicative Sounds 315
Fig. 8.8 Making a ruckus. Oscillogram of the stridulation produced by an unrestrained adult
female velvet ant, Dasymutilla occidentalis, when disturbed. The unpatterned sound is produced
by an in-and-out movement of the abdomen segments; this rubs a series of transverse striations (the
file) on the dorsal anterior margin of some abdominal segments across raised ridges (scrapers) on
the posterior dorsal margins of the preceding segments. The top recording represents one direction
of abdominal movement; the lower trace represents the opposite direction
away. After an encounter with the stridulating carabid beetle Cychrus, the common
shrew will not attack this insect again.
Insects of almost every order are known to react with some noise when picked
up, pinched, probed, restrained, or otherwise disturbed. In fact, more descriptive
papers have been published on this general kind of sound communication than on
any other aspect of arthropod acoustical behavior. For most arthropods, the only
known acoustic communication is such a protest sound, a high-intensity unpatterned
sound of a broad frequency spectrum. These signals are elicited primarily upon tac-
tile stimulation. Their function is still unknown, and whether they are even primarily
airborne acoustic signals is open to question. It has been suggested that the effec-
tive stimuli are tactile or vibratory and the air-transmitted portions of the signal are
incidental.
The emission of protest sounds is so widespread in the Arthropoda that it is most
tempting to consider these as predator-escape mechanisms. However, after actual
observations one finds it at least as easy to persuade oneself that such signals do not
upset predators at all; the literature is certainly full of documentary notes on insects
being eaten while emitting such sounds. Of course, failure in a certain proportion
of cases does not necessarily mean that a protest sound does not have a defensive
function. Some startled predators may still release the prey. Even if it only worked
to the prey’s advantage some of the time, such behavior would still have positive
selective value.
Another suggestion has been that these protest sounds should be regarded merely
as displacement activity, an outlet for the intense nervous excitement involved in
capture or cornering. A variant of this theory states that these sounds have no mean-
ing at all. Why, it is asked, does the possible meaning of this acoustic display even
have to be considered when no one thinks of giving equal consideration to the role
and efficacy of the cries of a mouse captured by a cat? Of course, on this subject
one is truly unencumbered by the facts. Until some perceptive experimental work
is undertaken, the situation will remain so. It is likely that various hypotheses may
Fig. 8.9 Performing the play list. Sound spectrograms of a domestic cricket, Acheta domesticus.
The loudest and most commonly heard song is the calling song sung by males to attract sexually
receptive females to them. Once the sexes are together, a new song—the courtship song—facilitates
mating. A quite different song, the rivalry or aggression song, is sung when two males encounter
one another and serves to determine the relative dominance of an individual. In other species,
the male may sing a post-copulatory song that keeps the sexes together for a period after sperm
transfer
be supported in different situations. Furthermore, in still other situations, the very

novelty of an unfamiliar acoustic signal may be enough to elicit predator avoidance.
Actual acoustic alarm signals differ from protest sounds in some important ways.
Their sounds are almost always patterned, though they may range from simple to
(often) complex. Most operate within a species rather than between species. Alarm
signals also are characteristic of social insects and show a wide range of intensi-
ties. Releasing stimuli are often elaborate. On theoretical grounds an effective alarm
should have a very wide frequency range and be of high intensity, and indeed these
characteristics do typify most sounds in this category. Such sounds are both easily
heard and readily pinpointed as to source.
The termite Zootermopsis angusticollis provides a good example of an acoustic
alarm. When suddenly exposed to bright light or an air puff, soldiers and workers
tap their heads on the roof and floor of their tunnels. They often also produce sub-
strate vibrations when they meet other termites, and they will respond to an alarm
vibration by producing one. The vibrations are received through subgenual organs
on the tibia; these appear to be tuned somewhat to the frequency and pulse rate of
their particular disturbance response.
In human biology, it is common to speak of a ‘fight or flight’ response to per-

ceived danger. The two reactions are behaviorally linked in insects, as well. Thus,
a second type of intraspecific disturbance sound is the patterned aggressive signal
or rivalry song that is used to settle competitive interactions between individuals
an example of which is shown in Fig. 8.9. Another example among bark beetles
was mentioned previously (see Chapter 5). However, the role of aggressive acoustic
signals has been most thoroughly studied in crickets (Case Study 8.2).
Case Study 8.2: Aggressive Singing in Two Crickets, Acheta

and Gryllus
He who fights and runs away, lives to fight another day.
Over the years at least since the Tang Dynasty (618–906 AD), cricket
fights have been a popular sport in China. If two adult male crickets are caged
together with just one burrow to hide in, they will fight over it until a winner
is established. Escalating battles involve fencing with their antennae, flaring
their mandibles at one another, and finally, grappling with each other (see Fig.
8.10). As the two males approach each other or engage in battle, they also
produce an aggressive rivalry song quite unlike their sounds in other situations
(see Fig. 8.9). When a winner is decided, actual physical combat declines. The
subordinate cricket becomes unusually quiet, and repeatedly avoids encoun-
ters with the dominant one. Meanwhile, the winner continues to stridulate
aggressively, as though boasting of his success.
What determines the outcome of these fights? How is this outcome main-
tained? In the early 1970s, two Princeton University biologists, L. H. Phillips
and M. Konishi, decided to attack the question experimentally in each of two
common cricket species, the house cricket Acheta domesticus and the field
cricket Gryllus pennsylvanicus.
First, Phillips and Konishi marked adult males of each species with various
combinations of silver enamel dots. Two previously unacquainted conspecific
crickets were introduced at a time into a very small cage. Under these con-
ditions, they encountered each other frequently and fought or demonstrated
clear dominance. The procedure was repeated many times until a statisti-
cally sufficient number of encounters was recorded. Then, the losers in these
matches were anesthetized and the tympana in their forelegs carefully torn.
By electrophysiologically monitoring the responses of the tympanic nerve to
recorded chirps, the investigators satisfied themselves that the crickets were
deaf, their tympanal organs unable to respond to sounds of the frequency
involved in aggressive chirps. Monitoring of the cerci and the subgenual organ
in the foreleg showed these unable to register the stridulatory sounds as well.
While the subgenual organs in the middle pair of legs responded to the chirps
when they were played at supernormal intensity, control experiments ruled
out any behavior effects of such responses under normal situations.
Having thus carefully established the subordinates’ deafness, the

researchers re-matched them with the previous winners. In all cases, the deaf-
ened losers immediately became extremely aggressive! They initiated combat
and engaged in sustained battles, and often came to clearly dominate their
former superior rivals. In a second series of experiments, the researchers
chose sets of three or four crickets that had already established a dominance
hierarchy in an empty aquarium. The lowest-ranking individual in these sets
was then deafened and rematched in tournaments with all the higher-ranking
individuals. In every rematch, the deaf cricket significantly increased the
proportion of winning encounters.
Fig. 8.10 Fighting for dominance. The stereotyped sequence of escalating performances char-
acteristic of aggressive encounters between male crickets, moving from mutual avoidance (level
0) through to all-out fights with biting and grappling (6). At any point in the sequence, the fight
halts if one opponent retreats. This establishes the winner, who typically bounces around singing a
rival song
Figure 8.10 outlines the interactions that occur when two male crickets encounter
one another. How can they be explained? One might assume that the acoustic signals
of the dominant cricket’s song inhibit the aggressive tendencies of the subordinate
cricket and that inability to hear these signals frees a deafened cricket from this inhi-
bition. This would account for the increased aggression that Phillips and Konishi’s
deafened individuals displayed. However, in the rematches, the deaf individual did
not always come out on top. Therefore, it seems likely that the probability of a
cricket winning an encounter is additionally influenced by the existence of other
factors as well. In fact, tactile antennal stimuli are also important. Research has
shown that a dominant male can be ‘defeated’ by repeatedly exposing him to artifi-
cially produced aggressive sounds while lashing him with bristles simulating cricket
antennae. Other studies with staged fights indicate that crickets use such antennal
fencing to assess the fighting readiness of an opponent, and high willingness to fight
may help crickets win fights even when they actually may have inferior fighting
ability because they are younger, less experienced, or weigh less than their oppo-
nent. This is in line with game theory prediction that at least some behavior patterns
displayed during aggressive encounters are used to assess asymmetries in variables
that indicate fighting ability and resource value.
Territorial behaviors often involve aggressive/alarm sounds and vibrations. As
is the case with many vertebrates, ritualized signaling seems to allow contestants
to resolve contests without physical harm. For example, caterpillars of the hook-
tip moth Drepana arcuata defend their silk nest sites from conspecific intruders by
drumming and scraping their mandibles and specially modified anal ‘oars’ against
the leaf surface in a ritualized display that sometimes goes on for several hours. As
acoustic communication has come under closer examination, it is becoming clear
that not only the hook-tip moth but a great many other Lepidoptera larvae (and
probably many other insects as well) use acoustic signals to settle territorial disputes.
8.5.2 Aposematic Sounds and Acoustic Mimicry
Small insectivorous birds called tits will not approach stridulating Nicrophorus car-
rion beetles. Unless they are one of the few who are aware that male Hymenoptera
cannot sting, most people will not go near a buzzing male carpenter bee. Syrphid
flies hum like the bees and wasps that they mimic visually. Bats will avoid palatable
mealworms shot into the air if the sounds of unpalatable arctiid moths are played
concurrently.
It has long been noticed informally that many insects make warning sounds that
indicate some sort of special capacity for defense such as noxious secretions, an
offensive taste, or possession of chelicerae or a sting. A predator might, after one
or several experiences, associate protest sounds with earlier painful feelings. Once
thus conditioned, the predator would treat the sound as a warning to keep its dis-
tance. The ultrasonic clicks of certain moths (see Chapter 2) appear to serve as such
a device.
A major theme in evolutionary biology is the use of visual signals that warn of
special capacity for defense and the mimicry of these signals by other organisms,
both protected and unprotected (see Chapter 5). By comparison with the extensive
research attention these visually based phenomena have received, acoustic apose-
matism and mimicry remain almost uninvestigated but both may be more common
than scientists have recognized.
A particularly well-designed study of acoustic aposematism included not one

but three normally stridulatory taxa: mutillid wasps (Dasymutilla), water scavenger
beetles (Tropisternus), and round sand beetles (Omophron). At night in the field
under natural conditions, W. Mitchell Masters presented predatory wolf spiders with
individuals of these three species that had been either surgically silenced or sham-
operated. In all cases, the spiders gave up their attacks more quickly when attacking
phonic insects. More of the silenced insects were killed, and more quickly. Then
he constructed an artificial insect—a vibrating probe that mimicked the cuticle of
a stridulating insect. When this was presented to the spiders, they also attacked it
longer when it was silent than when it was vibrating.
Two maxims run throughout insect behavior. One is that whenever a behav-
ioral outcome is really important, there will be back-up systems. The other is that
whenever an organism invests in an important behavior, there will be others that
attempt to exploit that investment. An excellent example of both maxims in prac-
tice involves night-flying moths and their major predators, the bats (see Chapter 2).
As back-ups go, the tiger moth family Arctiidae has a particularly well-developed
array of defenses and displays. Many retain distasteful or poisonous chemicals
acquired from their host plants, and some species also have the ability to make
their own defenses. Most advertise their multimodal defenses with bright coloration,
unusual postures, odors—and in many of the adults, ultrasonic vibrations produced
by vibrating the membranes of a tymbal organ on the metathorax. These signals
interfere with the bat’s ability to precisely locate flying moths, and they also help
these nocturnal predators learn to avoid them. The tiger moth Bertholdia trigona
goes even further, actively jamming bat radar with loud and steady broadband sound
above the range of human hearing. Experiments show that when the moth’s noise-
making organs are disabled, bats easily catch and consume them; however, bats are
almost never able to nab intact moths.
With such an effective array, it probably is unsurprising to find that other insects
seek to take advantage of it. Acoustic aposematism is a defensive strategy that is
clearly open to mimicry. Seeking definitive experimental evidence for it, William
Conner and colleagues trained naïve laboratory-raised bats of two species to hunt
tethered moths. Then each night they presented the bats successively with various
combinations of noxious and palatable moths in varying order (see Plate 28). The
bats fed normally on palatable non-mimics but within five nights, they had learned to
avoid the noxious model. When a sound-producing Müllerian mimic was introduced
starting on night six, bats avoided it for a full five days more. On day 11, to deter-
mine whether the bats were generalizing on the basis of the clicks, the researchers
removed the moths’ tymbals and presented the silenced moths. The percentage of
tiger moths captured returned to control levels, but all of the moths were subse-
quently dropped rather than being eaten. Apparently other signals such as vibration
or chemical cues were still active. To investigate possible Batesian mimicry, Barber
and Conner conducted similar trials, this time introducing a sound-producing but
totally palatable mimic on day six. Again, the bats avoided the mimic initially.
However, a few bats of one species subsequently discovered its palatability, show-
ing again what a powerful selective force bats are capable of exerting on mimetic
resemblance. Interestingly, some tiger moths also fly during the day where they
are exposed to visually hunting bird predators. In the spring when birds are very
active and bats less so, tiger moths tend not to produce their ultrasonic clicks, and
those species active during the day are visually conspicuous, where their multimodal
defenses also serve them well.
Acoustic mimicry can occur in contexts other than predator-prey interactions,
of course. Courtship and mating provide another important set of behaviors open
to exploitation. For example, males and females of the Australian bush cricket
Caedicia normally sing in duets. However, when they cannot find a duet partner,
some males perform an extensive repertoire of acoustic mimicry tricks instead. They
make short-click calls that mimic the female’s reply within a duet. They disruptively
over-sing other males’ song elements that are critical for the female’s conspecific
recognition. Sometimes they even make a call that mimics the entire duet. As they
perform all this deception, they also search for the calls of any unmated female that
might be tricked into a response.
8.5.3 Sexual Signals

Males and females of the Neotropical stink bug Euschistus communicate by alter-
nating vibratory songs transmitted through plants. When male Aphodius beetles
encounter a female in dung, they begin vibrating the substrate in a complex song
that consists of a series of different patterns displayed in a specific order over a con-
siderable time; females rarely stridulate, but seem to use the information within the
song in the context of mate choice.
Acoustic sexual signals come in two types: long-distance acoustic signals (pair
formation and calling sounds) and close-range signals used after the two sexes have
come together (courtship and copulation sounds). Close-range acoustic courtship
signals have received less attention than the louder and more obvious long-range
calling. By definition, close-range courtship signals would include any sounds or
intentional patterned vibrations that occur after pair formation and before copu-
lation. Responses to acoustic courtship signals have not been widely confirmed
experimentally; often, they are described only in terms of a female ‘allowing’ the
male to copulate. However, most cricket species have both a calling song that attracts
females from a distance and a courtship song that is given only when a female has
come in contact with a male (see Fig. 8.9).
The acoustic behavior of more than half of the 1000 or so species of Orthoptera
and Homoptera in North America and Europe has been studied to some extent, and
pair-forming or calling signals have been experimentally demonstrated more fre-
quently than any other kind of intraspecific acoustic signal. In no case has identical
or confusingly similar sound signaling been discovered between sympatric species
singing at the same time and place. However, closely related but geographically
and/or temporally isolated species often have identical or very similar sounds.
In general, the evolution of these two aspects of sexual signaling appears to have
proceeded quite independent of each other. It seems likely that the long-distance
signals are used mainly for species recognition, whereas the courtship song may
contain information about an individual male’s quality. If so, a calling song might
be expected to be more stereotyped and less variable than a courtship song. Such
a difference has been demonstrated in the field cricket Teleogryllus oceanicus; the
higher variability in courtship songs did not depend on the response of the female
being courted, because similar results were obtained when males courted dead
females.
In long-range acoustic pair formation, one usually thinks first of a direct loco-
motory response of a silent female to a single sedentary male’s calling song, such
as occurs in most crickets (Fig. 8.11). Such oriented movements in response to
acoustic signals are called phonotaxes. Many variations in pair formation exist,
however. Some involve elaborate visual/acoustic displays by both sexes. Others
Fig. 8.11 Finding the singer. Positive phonotaxis used in experimental studies on the inheritance
of song in hybrid Teleogryllus crickets. Walking along a featherweight Y-maze, tethered recep-
tive hybrid adult females listened to different hybrid male calling songs played simultaneously by
equally distant speakers, and revealed their preference at fork points in the maze
depend upon alternating and distinctive songs between males and females. Habitat
ecology is undoubtedly a major determinant of the type of system that can be
employed. For example, diurnal desert grasshoppers that live in open habitats can
use visual–acoustic systems with ease, but tettigoniid grasshoppers that live in visu-
ally restricted habitats and have nocturnal and/or cryptic habits must rely more
heavily upon strictly acoustic elements.
During long-distance pair formation, one also usually thinks of each species
stridulating its own specific sex-related messages. If they do not, one immediately
assumes some other clear-cut isolating mechanism, but sometimes the case is less
well defined than this. One of the best examples of acoustically based reproduc-
tive isolation involves two sympatric grasshoppers, Chorthippus brunneus and C.
biguttulus. Morphologically they differ only by a very few small, nonoverlapping
characters. Although they have different songs, the two species are not separated
ecologically. Studies by A.C. Perdeck showed that the two species would cross-mate
and produce viable offspring in the laboratory, but very little hybridization occurred
in the field. How could this be? Careful field study revealed that the major difference
between the species lay in the degree of stimulating and orienting influence that their
songs possessed. In the presence of its own specific song, a male moved faster and
made more copulatory attempts; a female was also more ready to copulate in the
presence of the song of her own species. Hybrid females would answer the songs
of both hybrid males and parental males, but pure females of either parental species
would not respond to the song of hybrid males, which was intermediate between the
parental songs.
Long-range acoustic courtship signals are best known in cicadas, in some male
Orthoptera (Acrididae, Gryllidae, Tettigoniidae), and in some Diptera, but there is
suggestive evidence for their occurrence in males of many other arthropods. In the
classic case, a male insect stridulates sex-related messages and the female receives
them through her tympanal organs. However, sound production may also arise as a
by-product of other activities such as flight, and in some species the sounds used
in communication fall within the sensitivity range of hair sensilla and antennal
receptors. These sounds could, and do, also attract the opposite sex. Predictably,
the significant properties of these sorts of courtship sounds are rather different than
for the songs of Orthoptera.
Mosquitoes are a case in point. Many mosquitoes produce an audible hum during
flight, often so distinctive that some people can identify mosquito species solely by
their sound. Astute observers have long known that mosquitoes react to sounds as
well. Swarms of various species have been seen to alter their behavior following
acoustic stimuli as varied as pistol shots, locomotive whistles, musical instru-
ments, and the singing human voice. Furthermore, the antennae of male and female
mosquitoes of many species are structurally quite different. The female antennal
flagellum is about 2 mm long; the male flagellum, only about 1.6 mm. The flagel-
lum of the male antenna is plumose, with many more hairs of varying lengths, and
intriguingly, these hairs are absent in the males of some mosquito species for which
acoustic communication has not been demonstrated. As though to underscore the
hairs’ importance, males that have them can change their positions relative to the
antennal shaft; most of the time the hairs lie flat, but they are raised erect during the
period of swarming and mating.
Despite such circumstantial evidence for mosquito hearing, however, some care-
ful studies were needed to prove its existence and unlock the details. The first
researcher to elucidate the basis of mosquito sexual communication in detail was
Louis M. Roth. Case Study 8.3 presents a few aspects of Roth’s complex, wide-
ranging investigation; the interested student may wish to peruse his original 1948
study for further detail.
Case Study 8.3: Sexual Attraction in the Mosquito,

Aedes aegypti
Do mosquitoes listen to one another? If so, how? And what do they actually
hear? Before some perceptive research, the acoustic basis of mosquito sexual
communication was a matter of dispute.
In 1855, Christopher Johnston, a medical doctor in Baltimore, Maryland,
discovered a sensory organ in the second segment of the mosquito antenna,
and suggested it might play a role in mosquito hearing. Although the organ
was subsequently named after him, until about sixty years ago, scientists
were unwilling to give his observations a pivotal role in a behavior as vital
to survival as sexual attraction. Instead, one researcher claimed that males
were drawn to females by odor. Another postulated that the female’s motion
alone was sufficient for male recognition and attraction. A third suggest
that knoblike projections (halteres) behind the wings of the female produced
attractive ‘birdlike’ sounds both at rest and in flight.
As an entomologist in the U.S. Army during World War II, Louis M. Roth
had his own hunches. Observing yellow fever mosquitoes, he had noticed
that when a female mosquito was resting, she sometimes was surrounded by
males, with some so near that they touched her, but he never observed them
mating. It therefore appeared unlikely that odor or constantly produced sounds
were involved. It seemed more likely, he thought, that male mosquitoes were
attracted by the buzzing sounds produced by a female’s wings during flight
and that males received this sound through their enlarged antennal bases,
which have conspicuous Johnston’s organs. But how best to prove this?
Beginning with careful behavioral observations, Roth established that once
male Aedes aegypti became sexually responsive 15–24 hours after emergence,
they remained in a constant mating state through the rest of their lives. When
he forced caged females to fly continuously, males attempted mating with
them repeatedly; after a time, the females tended to resist further mating
attempts but as long as they continued to fly, males continued to be attracted.
However, when he killed female mosquitoes and presented them, the same
males that had readily mated with flying females were indifferent to these
freshly dead bodies.
Next, Roth exposed mosquitoes of both sexes to some mechanically pro-

duced sounds. He found no evidence that female A. aegypti were attracted
to any, though they sometimes gave shock reactions to certain frequencies.
Males, however, showed their characteristic phonotaxis and mating response
to any sound in the range between 300 and 800 vibrations per second. This
was true whether he produced the sound by a recording of another mosquito,
by an audio oscillator, or by a tuning fork.
His hunch further strengthened by this, Roth fastened fine wire loops
about the necks of female mosquitoes. Then he suspended them in a cage
full of males. As long as a female hung with motionless wings, the males
flying near her remained indifferent. Even when Roth swung the female to
and fro upon her tether, they did not respond. As soon as she began to
vibrate her wings, however, the males flying nearby immediately seized her
and began to mate. To put the theory of attractive birdlike sounds from the
halteres firmly to rest, Roth repeated the experiment again after surgically
removing the halteres from some females and removing the wings from oth-
ers. Males responded to the haltere-less females that still had wings and
vibrated them. However, males ignored wingless females that still had their
halteres.
What sets the range of sounds within which the male is attracted?
Measuring the wing-beat rate of mosquitoes of different ages and sexes, Roth
found the key in the sexual maturation of the mosquito. Mature females and
freshly emerged males beat their wings at about the same rate and are pursued
by older males, but young, freshly emerged females have a different wing-beat
rate and are not pursued. Evolutionarily, it is a sensible system. By setting
response limits such that only mature females are pursued, the likelihood is
increased that the female has ingested a blood meal and that a mating will
produce viable progeny. Young males avoid sonic misidentification and pur-
suit by older males by simply remaining quiescent until they complete their
sexual maturation. By the time a male begins to fly, the sound he makes in
flight is sufficiently high in pitch to be beyond the range that will stimulate
older males.
Is the female mosquito’s buzz simply a byproduct of normal wing flap-
ping during flight? When Roth cut off greater and greater portions of their
wings, the sound progressively decreased in volume while the pitch rose pro-
gressively. Yet males were attracted to and mated with females that had only
part of their wings, even stumps, vibrating. This suggested that some basal
organ might be the principal sound producer; subsequent morphological stud-
ies confirmed the presence of a stridulating organ at the base of the female’s
wings.
Spurred both by intellectual curiosity and by mosquitoes’ obvious medical and

economic importance, generations of researchers have continued to build on the
work of Johnston, Roth, and others. Their studies have revealed a very complex
system, and the mechanical sensitivity of the male mosquito’s antenna has been
shown to exceed that of all other arthropod movement receivers studied so far.
Research on physiological aspects of mosquito acoustic communication was
hampered initially by technical limitations and the antennae’s morphological com-
plexity and small size, but in the last decade or two, new technology has allowed
new analyses. Studies show that because the antennal hairs are stiffly coupled to the
shaft of the antennal flagellum, at the appropriate frequencies, the flagellum moves
like a stiff rod rocking about in its socket. This stimulates the Johnston’s organ at the
flagellum’s base. In both male and female mosquitoes, the antennae are resonantly
tuned mechanical systems that move as simple forced damped harmonic oscillators
when they are stimulated. The tuning of female antennae does not match the flight
sounds of males or of females, but male antennae respond to a range of attractive
frequencies that brackets the fundamental frequency of female flight sounds for their
species. This bracketing occurs because hairs of different lengths have different best
frequencies, with the result that males respond to a wider frequency band than would
be possible if all the hairs were the same length. This may be an advantage in that it
assures that females remain attractive even if they produce slightly different sounds
because of such uncontrollable factors as size, wing damage, or extent of distension
of their abdomen with food.
Is the resonance of the hairs themselves essential for mosquito hearing? Probably
not. These same studies have shown that the hairs resonate at frequencies much
higher than the best frequencies not only of female flight sounds, but of all arthropod
sensory hairs known to serve as near-field sound receivers. For this reason, some
researchers contend that the main function of the hairs may just be to increase the
antennal surface. At this scale, they may also increase the drag force of air passing
between the hairs on the antenna, much like a boat paddle dragged through water.
Earlier we mentioned acoustic tuning by males of Toxorhynchites brevipalpis.
Despite the fact that the Johnston’s organ is only slightly less sensitive in female
mosquitoes than in males, most studies have investigated ways in which males are
attracted to female flight tones. Gabriella Gibson and Ian Russell have done some
of the first work to show that the auditory attraction phase of sexual recognition
in mosquitoes involves interactive behaviors, just as in sexual behaviors of other
insects. When they flew opposite-sex pairs of tethered T. brevipalpis, each mosquito
altered its wing-beat frequency in response to the flight tone of the other, so that
within seconds their flight-tone frequencies were closely matched or completely
synchronized (Fig. 8.12). The behavioral thresholds in both sexes are about ten
times greater than neural thresholds, differences that could be explained by neural
processing (see Chapter 2).
Before leaving the subject of mosquitoes, it should be noted that once again, a
vital activity such as courtship relies upon more than one sensory mode. Swarming
Aedes aegypti males have been shown to produce a volatile pheromone that
stimulates the flying activity of females at a distance. Females also produce a
volatile attractant that appears to act as an aggregation pheromone. The resul-
tant pheromone-mediated swarm has been described as a 3-dimensional lek (see
Fig. 8.12 Flying in tune. Responses of tethered flying Toxorhynchites brevipalpis mosquitoes
show that males and females interactively synchronize their wing-beat frequencies to one another.
(above) The flight tone of a flying male is shown as an oscillogram, with the frequency spectrum
of the sound pressure levels (SPL) for the area outlined in the rectangle directly below. (middle and
below)The lower two graphs depict spectrograms of a flying male and a flying female, respectively,
over a 10-second recording; each shows fundamental components of flight tones before, during,
and after a pure stimulus tone (solid line) was played. The male’s spectrogram illustrates that
his flight tone first exceeds and then converges with the stimulus tone. The female’s spectrogram
shows that her flight tone first drifts upward, then decreases without reaching the stimulus tone,
and finally returns to the original frequency. Numbers on vertical axis indicate flight tone at onset
of record, peak frequencies during the record (dotted lines), and final frequency. The dashed line
indicates the flight tone in the second before the stimulus
Chapter 10), though there is still some question as to whether female choice is
involved in these swarms.
8.5.4 Social Sounds
Pogonomyrmex harvester ants trapped in a small landslide stridulate loudly, attract-

ing other workers who dig them out. Sibling groups of the sap-feeding treehopper
Calloconophora forage cooperatively; scouts that locate a new feeding site vibrate
the leaf, and newly recruited individuals signal in concert with those that have
already arrived. Hornets produce and are sensitive to a variety of sounds inside their
nests. Honey bees initiate swarming in part in response to a contagious buzzing
sound started by a few workers running excitedly through the hive.
Group-living insects range from those existing in simple aggregations to those
that cannot survive outside of their complex societies (see Chapter 10). To be social
necessitates having a way to integrate the cooperative behaviors of group members.
Social communication is rarely restricted exclusively to one particular mode. Most
often such integrative communication is primarily chemical, but sound also serves
as an effective avenue of communication in social insects, often as a back-up and
sometimes as a primary channel.
The interlocking combination of odor, sonic, and tactile signals that character-
izes the communication system of the Douglas fir bark beetle (see Chapter 6) is a
case in point. However, one of the classic and best-studied examples of the interac-
tion of communicatory modes is provided by that classic social insect, the common
honey bee. The behavior involved is recruitment to food, one of the most important
adaptations of nearly all social insects. Case Study 8.4 outlines the story.
Case Study 8.4: Communicatory Interaction in the Honey

Bee, Apis mellifera
Honey bee dance language remains one of the most striking, best-studied
examples of mixed channel communication and of a highly complex transfer
of symbolic information.
Beginning in 1919 with a few paint-marked bees and a flower enriched
with a drop of sugar syrup, Austrian biologist Karl von Frisch began a series
of thorough observations of honey bees that has culminated in what is proba-
bly the best known example of complex insect communication in the entire
world. In recognition of his pioneering work, von Frisch shared a Nobel
Prize in 1973, the only entomologist to date so honored; behaviorists Niko
Tinbergen and Konrad Lorenz received it with him. He continued to research
honey bee communication for many decades and in the 1960s, published an
outstanding summary of the work he and his students had conducted. The
story continues to excite new research, and has become part of the shared
human experience as the premiere example of symbolic communication aside
from human language.
What was it that excited people so? Von Frisch had learned that a forager
bee that found a good food source would bring back a full load of pollen and
nectar to the hive. It would then fly out again to the food, thus establishing the
most direct outward route. Back in the hive upon its second return, the suc-
cessful forager communicated its vital information to others by performing
specific, repeated, stereotyped behaviors, which von Frisch termed ‘dances’,
upon the vertical face of the honey-comb inside the hive (see Figs. 8.13
Fig. 8.13 Sharing good news. The dance language of a successfully foraging honey bee. (a) A
forager locating a source of food close to the hive (within a distance that varies among strains)
moves in circles of alternating direction, performing a round dance. Followers of the round dance
obtain information that food is near. As the distance to the food increases, the round dance perfor-
mance gradually changes to an open figure-8, then to the waggle dance (b) performed by foragers
discovering food sources far from the hive. As the bee passes through the straight run portion of
the dance, she waggles her abdomen rapidly from side to side. At the end of each straight run, she
circles back to the beginning, to the left one time and to the right the next, so that the dance takes
a figure-eight shape
and 8.14). In response, other foragers accurately flew directly to the food.
The discovery captured the common imagination and excited the scientific
community. Although observers had repeatedly noticed that a bee in a colony
sometimes would perform repeated circular movements, no one had firmly
established the connection between these movements and recruitment, and
few had guessed the complexity of the communication.
One aspect of the forager’s message was distance, which correlates with the
type of dance the forager displays. Relatively short distances evoked a round
dance; longer distances, a waggle dance. Within the waggle dance, a number
of features also correlate with food source distance. Von Frisch’s standard
measure was tempo, which he measured as number of circuits (a waggling
run plus the loop that returns the bee to the beginning of the next waggling
run) per 15 seconds. Correlations also occur with the speed or length of the
wagging run, the duration of the semicircular run, the duration of a single
tail-wagging phase, and the number of tail wags during the straight run. How
did the bee know how far it had flown? For a foraging bee, distance could be
assessed in various ways. To von Frisch, the most important aspect seemed to
be the energy that the bee had expended in covering it, because if a bee had
to fly into a headwind or uphill or if it was forced to walk, it would report a
distance farther than what was actually involved.
Fig. 8.14 Translating directions. When a forager waggle-dances to recruit others, her dance
velocity and number of figure-eight shapes per unit of time indicate distance to the food source.
Direction is more complex; whereas she learned the flight directions horizontally relative to the
sun azimuth, her dance takes place inside the dark hive on a vertical comb. To translate this to her
followers, the forager uses the direction upward as a reference for sun direction, and orients the
waggle portion of her dance at the same relative angle. Thus, a cross-run angle of 60◦ to the left
of the upward vertical axis as she dances means that the food was on a path 60◦ to the left of the
sun-hive line, going toward the sun. Interestingly, on sunny days a forager will sometimes dance at
the hive entrance; in this horizontal position, her tail-wagging run points directly toward the food
source
The other aspect was direction. The angle of the tail-wagging run in the wag-
gle dance with respect to the vertical is equal to the angle formed between the
sun’s position and the food source. Communication of this aspect is theoreti-
cally complex, because the dance is performed inside the hive out of sight of
the sky and (because it happens on a vertical comb) it must be translated to
occur with reference to gravity (see Fig. 3.6) rather than directly to the sun
itself. Furthermore, using the sun as a directional reference is complicated
by two problems—it moves, and sometimes it hides. Thus, both a celestial
compass and perception of polarized light are involved.
Some returning foragers do not dance at all; others may dance a hundred
circuits. As this would suggest, honey bee recruitment seems to be regulated
in a way that allocates outgoing bees to the resources that most benefit the
colony. The number of waggling runs a forager performs seems to indicate
the overall value of the resource, and bees dancing to good patches seem to
circle and dance more rapidly. Bees perform more dances to more sources of
nectar that are more highly concentrated or closer. When a colony is heat-
stressed, they perform more often and more vigorously to water sources than
to nectar sources.
A returning worker also changes her information in response to the activi-
ties of nest-mates with whom she interacts. When a successful forager returns
home, other workers greet her and, if she is carrying nectar, they induce her
to regurgitate her load to them. If many workers greet her enthusiastically,
she begins to dance a round or waggle dance. If, instead, she perceives a
long delay in unloading nectar or a shortage of receiver bees, she will per-
form a lesser-known variant called a tremble dance. Although von Frisch
described this dance in the 1920s, it has received less attention, and its func-
tion was unclear until 1993 when Wolfgang Kirschner discovered that the
tremble dance stopped nearby workers from flying to gather more nectar and
switched them to unloading nectar.
How do followers sense the dance? Given that the dance language is com-
munication, it is surprising how much remains undecided about this aspect
of the system. One reason is simply the complexity involved; the dances pro-
vide such a rich variety of potential information that it remains difficult to
know which information is actually being used and how it is being passed on.
Another reason is that the dance features help followers find and stay with the
dancer may differ from the features that carry the location signals.
Von Frisch emphasized the role of odors, specifically odors from flowers
and other environmental chemicals that might cling to the bees’ hairy bodies
and be detected by dance followers. He suggested that the dances got bees
into the flowers’ general vicinity, then odors from the forager guided them
in. In support of this idea, he noted that a powerful odor can even lead bees
to ignore dance information and instead find the food that the odor signals.
Foragers have in fact been found also to release an attractive pheromone as
they return to a familiar feeding place.
Might hearing also be involved? One of the first to tackle that question
experimentally with the aid of scientific instrumentation was Harald Esch,
who had spent some time studying bees with von Frisch. In the 1960s, Esch
built an artificial motor-driven bee capable of doing a waggle dance. When
it was introduced into the hive, bees clustered around it just as they would a
genuine forager, but no bees left the hive to search for food. Why? Perhaps
the robot didn’t sound right, Esch reasoned. So he placed a microphone in
the hive to record actual dancing bee sounds. To his surprise, not only did
the dancing bees produce strong sound signals with vibrations of their wings,
but the period of sound production coincided exactly with the duration of
the straight run of the dance. Since von Frisch had already found a correlation
between the latter and the distance to the food source, a logical conclusion was
that the bees might equally well be using sound as the cue to distance rather
than using the visual or tactile cues of the waggle itself. About the same time,
Adrian Wenner independently discovered that sound was produced during the
bee’s dance. Produced mainly at 250 cycles/second, the sound occurs in bursts
and the number of bursts correlates directly both with the distance and with
the sugar concentration of the food. The pulse rate of these sounds is directly
related to the rate of waggling, but they are approximately 2.5 times as fast.
Because the followers can be seen to have antennal contact with the dancer,
other researchers noted that it is also likely that tactile cues are important.
The speed of waggling could be sensed this way as a measure of distance.
Contact could also be necessary in another way. By following the dancer and
closely imitating the path of her dance, as many potential recruits do, they
might obtain information on direction as their proprioceptors monitored this
imposed movement. Von Frisch also showed that foragers could orient their
dances either to celestial clues or to gravity through proprioreceptive bristle
fields between the major body segments and the segments of the legs.
In general, as one moves closer and the insects involved become smaller,
hearing grades into vibration and near-field signals, and finally to touch. Many
studies have now provided evidence that dancing bees produce sound with
their wings and followers detect these with their antennae. They also produce
vibrations in the substrate that could be detected via the subgenual organs.
At very close inter-bee distances, near field airflows are generated by the
dancer’s vibrating wings and by its waggling abdomen. Observations show
that most follower bees place their heads directly into this zone of intense
airflows. During the waggle run, there is substantial physical contact between
dancers and followers. However, at all these levels, questions remain unsettled
as to whether the information is precise enough and/or is used beyond some
restricted circumstances.
What about vision? Though humans find it natural to describe what the
dances look like, in the total darkness of the hive bees could not observe the
dance and visually interpret its tempo. However, von Frisch recognized that
vision might have an important role at an earlier stage: foraging bees might
measure distance by monitoring optic flow, the streaming of visual texture
across the visual field as an animal moves through the environment. He noted
that bees flying to a feeder over calm water (which provides a weak optic flow)
signaled a shorter distance in their dances than bees that had flown over more
visually textured surfaces such as a wind-disturbed lake. Von Frisch was never
quite convinced that optic flow was as important as his energy hypothesis
as a way to encode foraging distance. However, through an extensive series
of studies, Esch and others have shown that optic flow is probably not only
the most important, but perhaps the only, odometer for foraging honey bees.
Furthermore, they have shown that many of the experimental results that von
Frisch interpreted as supporting the energy hypothesis are also consistent with
the optic-flow hypothesis.
With so many senses interacting, one finally must ask what systems are
not involved in the honey bee dance. Even though bees can orient to magnetic
fields in other contexts, so far there is no evidence that they use magnetic
orientation in their dance. The evidence for orientation via landmarks is also
equivocal; honey bees can use landmarks if they were first given the opportu-
nity to see them in conjunction with celestial clues, but the role of this ability
in nature remains to be shown.
When Karl von Frisch showed that bees could use waggles and circles to tell
other bees where to find forage, he moved honey bees and their close relatives to
front and center as a model of animal communication and social behavior. Ever
since, biologists have been trying to figure out the details of how bees communicate
this information and attract recruits in the darkness of the hive. Probably the most
important lesson learned in that search for answers has been that important com-
munication nearly always involves redundancy—more than one sensory modality
acting simultaneously and generally reinforcing one another’s messages.
Failure to consider such possibilities when analyzing communication systems
can be a pitfall, as was illustrated by the so-called dance language controversy
that arose in the 1960s when some researchers (particularly Adrian Wenner, Patrick
Wells, and Dennis Johnson) directly challenged von Frisch’s theories. suggested
that even though the dances might contain spatial information, it was irrele-
vant to honey bee recruitment. No interplay of sensory systems was needed;
odors alone were sufficient to explain the phenomenon. In fact, the entire con-
cept of bee communication was unnecessary. Wenner, for example, asserted that
bees are recruited solely by a conditioned response to food or odor. He felt that
bees lacking experience with a particular food source have great difficulty find-
ing it. Only a relatively few seem to be successful, and these take much longer
to arrive than would be expected if they were flying a straight line between the
hive and the source. Inexperienced foragers, Wenner suggested, locate the source
simply by dropping downwind of the hive and then searching for the right com-
bination of odors from food, locality, and other bees. Johnson and Wenner’s
experiments, done with strong scents and short distances, seemed to confirm their
expectations.
Some of von Frisch’s experimental results couldn’t be readily reinterpreted in
terms of an odor-only hypothesis, however. For example, when a hive is turned on
its side so that bees cannot use gravity as a reference, they do disoriented dances.
Under these conditions, recruits are less well oriented, even though odor cues have
not been affected.
Normally, odor information and dance vector information are highly correlated.
To settle the controversy, scientists knew they must tease these factors apart. In
the 1970s, James Gould did this quite successfully in a set of experiments that
unlinked the location and odor of the foragers’ food source from the directional
information in their subsequent dances. He knew that if one shines a bright light
from the side, dancing bees will use the position of this light as though it were
the sun, and recruits will interpret it that way. However, if a bee’s ocelli are coated
with opaque paint, the bee becomes less sensitive to light, and this shift in ref-
erence does not occur. Gould set up a situation in which dancing recruiters with
painted ocelli (and a reference of up, toward the sun) were followed by recruits with
unpainted ocelli who were reading those dances relative to whatever angle the arti-
ficial light produced. In this way, he was able to show that recruits could interpret
a direction from the dance that was independent of the direction to the food source.
When given the opportunity to hunt for the forager’s nectar source, these recruits
searched principally in the direction predicted by the modified dance information.
By the odor-only hypothesis, they should have still gone toward the true direction
of the feeder.
Many other imaginative investigations also took place as the scientific com-
munity subjected these two varying interpretations to critical examination and
experimental test, including development of a series of mechanical models and
robots capable of mimicking various aspects of honeybee dances (Fig. 8.15). One
of these was a computer controlled robot developed by Axel Michelsen, Martin
Lindauer, and Wolfgang Kirchner that mimicked the behavior of a dancing bee;
although it recruited rather imprecisely, changes in its dance caused significant shifts
Fig. 8.15 Mimicking a forager. An example of a complex model developed to simulate honey
bee foraging communication. (left) As an X-Y recorder (A) drives the model’s figure-eight and
wagging motions, a stepper motor (B) turns the model. Another motor (C) connected to a syringe
(S) pumps scented sugar water through a soft tube to the front end of a beeswax-covered model,
and an electromagnetic driver (D) from a bone-conduction hearing aid moves a wing (E) made of
a tiny piece of razorblade. (right) Potential recruits surround the mechanical model as it performs
a wagging run
in the search distribution of recruits in just the way that von Frisch’s dance language
theory predicted.
The odor search hypothesis has not yet been abandoned by its most loyal adher-
ents, and unfortunately some philosophers and science sociologists have greeted the
‘dance language controversy’ as a welcome example of what they see as severe lim-
itations of the scientific method. However, it is now clear that honey bee recruits can
use either the dance language and odor information or odors alone, and they most
probably can and do use other sensory inputs as well. The relative importance of
these various inputs undoubtedly varies from one situation to another. Furthermore,
as Gould has pointed out, the controversy might not have arisen but for the fact that
while von Frisch and Wenner were both studying situations involving exploitation
of an abundant food source, they focused on different stages of this process: von
Frisch on the early stages of discovery, Wenner on later stages.
The general consensus now seems to be that sounds, comb vibrations, tactile
cues from touching antennae, and odors all play roles in this elaborate transfer of
information, and that additional signals will probably continue to be discovered.
A study by Corinna Thom and associates in 2007 has added another component,
semiochemicals (see Chapter 6) in the form of four cuticular hydrocarbons emitted
by wagglers. Interestingly, the most energetic waggle-dancers produced the most
copious quantities of these compounds.
How did such a complex communication system evolve? Seeking an answer,
researchers such as Esch, Lindauer, and Kerr were among the first to focus on related
stingless bees (Meliponini) that occur in the tropics. All 11 species of these social
bees that they studied show some means of alerting their colonies to food, but the
efficiency and degree of development of these communication systems vary greatly.
In the simplest case, Trigona iridipennis workers that have found a food source sim-
ply fly back to the comb and excitedly run about, thus alerting other workers to the
existence of food but not to its location. In the most complex case, a Trigona pos-
tica worker constructs a trail system of marks left at short intervals by rubbing her
mandibles on convenient objects such as leaves to deposit mandibular gland secre-
tions. Back at the comb, the bee then runs noisily about, giving out food samples.
When a number of other bees have been acoustically and chemically alerted, the for-
ager leads these bees along the scent trail. Such a system is of special advantage to
these forest-dwelling species, because when living among trees vertical movements
may be as important as horizontal ones.
James Nieh has shown that another stingless bee, Melipona panamica, sends and
receives acoustic signals that allow inexperienced workers to communicate foraging
information. Foragers trained to a feeder at the top of the jungle canopy make short
pulses of sound while sharing food inside the nest, whereas those trained to a feeder
at the forest floor make much longer sound pulses. Immediately after unloading, the
forager dances briefly; sounds produced during this phase correlate with distance
of the food, with longer pulses given for greater distances. Nieh did various studies
in which he trained bees to a feeder, let these experienced foragers interact with
inexperienced bees, then released only the inexperienced bees to find a new feeder
placed at the training location (Fig. 8.16). In this way, he showed that both dis-
tance and height of the food source are communicated inside the hive, but direction
appears to be communicated outside, perhaps by following the foragers outside the
nest for a brief distance.
Honey bees represent a very tiny fraction of the approximately 20,000 known
species of bees. Among the present members of the genus Apis, the more basal
species make single, exposed combs, while the more recently evolved species nest
in cavities and have multiple combs (a development that has greatly facilitated their
domestication). Martin Lindauer’s pioneering study of three Apis species that differ
in their nesting habits seemed to indicate a progression in complexity that appears
to correspond to the phylogenetic development of the dance during Apis evolution
(Fig. 8.17). On a horizontal surface, both A. florea and A. mellifera orient similarly
(Fig. 8.17a). However, on a slope, differences are apparent. Dwarf honey bee
dancers invariably orient to the sun’s azimuth by using the sky horizon as a refer-
ence; placed on a slope, they compensate by head rotation (Fig. 8.17b). In contrast,
true honey bees on a slope rotate both their body and head (Fig. 8.17c), so that the
slope’s plane serves as an apparent horizon reference for orienting to celestial cues.
To Lindauer this suggested that dancing first evolved in an ancestral bee that built
a single comb in the open and oriented to celestial cues but could not yet substitute
gravity for the sun, much as A. florea still does today. Perhaps the earliest dances
were simply excited, disorganized movements, but other bees were aroused by the
excitement and thus were more apt to leave and search for food. In a later evolu-
tionary stage, still on an open nest comb, dancers began to translate their solar flight
angle into a dance angle relative to gravity and added acoustic signals. Over time,
bees became able to use celestial cues when they were available and gravity cues
when they were not, setting the stage for bees to begin moving their nests into the
protection afforded by cavities. As intriguing as Lindauer’s theory may be, however,
there have been a number of objections to it, based both on taxonomic uncertainties
as to ancestry of the various bees and on behavioral studies that have revealed more
plasticity than Lindauer observed. Researchers’ hopes for discovering ‘missing
links’ in the shape of simpler forms of honey bee dance language so far have not
been fulfilled, but however the story plays out, it promises to be an interesting one.
Before leaving the subject, it is interesting to note that elements of the bee dance
also have been uncovered in a diversity of insects outside of the Hymenoptera. For

Fig. 8.16 Communicating food location. The tropical stingless bee Melipona panamica uses a
wide range of communication techniques including sound. Some aspects of the information are
shared inside the nest; other aspects, outside. In the three types of trials shown here, bees were
trained to a feeder, then allowed to interact with inexperienced bees inside the nest. When the
inexperienced bees were released to find food, (a) more found the training feeder high in the forest
canopy than found the control feeder at the base of the trees; (b) many more went to the training
feeder than to the control feeder that was closer to the nest; and (c) equal numbers went to the
control feeder and training feeder when they were at equal distances but in opposite directions
from each other
Fig. 8.16 (continued)

Fig. 8.17 Translating the sun. (above) The exposed nest of the dwarf honey bee, Apis florea,
(left) with and (right) without the curtain of bees. Dances are performed on the upper horizontal
surface of the comb and are oriented directly toward the source of the food. (below) Various bee
species orient differently on horizontal and sloped surfaces (see text)
example, blowflies will perform a crude sort of dance after eating, with a vigor and
persistence related to the distance the fly has flown after feeding. Some saturniid
moths rhythmically sway upon settling down after any movement; the number of
oscillations is closely correlated with the distance they have flown before settling. Of
course, there is no evidence that these activities have any communicative function;
they may be simple by-products of the physiological state of the insect concerned.
But they do share the important characteristic of persistence, that is, they continue to
be performed for a considerable time after the cessation of the activity that produced
them. Such persistence is of paramount importance in communication, especially in
recruitment of foraging bees. If such by-products were to come to signal the sender’s
state to another animal, then they would be ideal materials for the construction of
a communication language. However, recruitment to food is considered primarily
an adaptation to social life. Why should this be so? Sociobiologists point out that
being able to recruit nest-mates to food sources allows insect colonies to capitalize
on one of the major advantages of living in groups—the ability to harvest food
that would not be readily available to an individual foraging alone. At the same
time, recruitment overcomes an inherent disadvantage of group life. Were it not for
the compensation that recruitment provides, solitary individuals would have better
access to food if they did not have to compete with and share with other group
members.
Chapter 9
Reproductive Behavior
9.1 Introduction
‘Lovebugs Menace Florida’s Tourists’ a recent newspaper headline declared. Not a

bug at all but a fly, the lovebug (Fig. 9.1) does not bite or transmit diseases, nor is
it a pest at picnic outings. How could such a poetically named insect constitute a
‘menace’?
Aspects of insect sexual behavior already have appeared in several contexts
in this book, from long-distance olfactory signals emitted by night-flying moths
to visual attraction by fireflies and dragonflies. The use of sequential cues from
a number of sensory modalities—a characteristic of the majority of courtship
systems—was illustrated by the interaction of pheromones and acoustic signals in
Fig. 9.1 Going to extremes. Lovebugs, Plecia nearctica (Diptera: Bibionidae), in typical copu-
latory position (female on the right). Attracted in large numbers both to car engine exhaust fume
components and to engine warmth, mating pairs can be a serious problem for motorists in the
southern United States. In the laboratory, lovebug pairs copulate for an average of 56 hours

342 9 Reproductive Behavior
Douglas fir bark beetles and the tactile, visual, and olfactory interplay involved in
butterfly courtship.
These examples left several questions unanswered, however. Why are the
courtship rituals of some species so intricate, even bizarre, whereas many others
mate with apparently only minimal preliminaries? What factors influence the choice
of a mate? Why are some species territorial? Why do individuals of some species
mate but once, whereas others promiscuously?
Insect reproduction has been one of the most intensively studied aspects of insect
biology in the past century. One reason is quite practical, based on the intersection
between insect populations and our own interests in agriculture and human health.
Learning about the reproduction of problem insects offers the promise of disrupting
that reproduction and reducing the number of those insects in the next generation.
The other reason is more theoretical, based on the diversity in insect reproduction,
which provides a rich source of material for discovering underlying rules of biology.
As intensively studied model systems in both field and laboratory, insects are ideal
for addressing questions like these.
9.2 Courtship and Mating
After an extensive aerial courtship, a male butterfly places a packet of sperm into
the female’s genital opening. Bed bug males inject sperm subcutaneously into the
female’s blood-filled body cavity. A male damselfly transfers sperm away from his
abdominal tip, freeing its claspers to grasp a female by the prothorax; the pair may
fly about for hours in this tandem position (Fig. 9.2, Plate 33).
Sexual reproduction, nearly all biologists agree, has evolved because of the much
greater speed with which new genotypes are assembled under this system. Such
diversity, they argue, is highly adaptive. Sexually reproducing populations are more
likely than asexual ones to adjust to changed environmental conditions through the
creation of new genetic combinations. Thus, sexual reproduction has increasingly
become the mode. In sexually reproducing animals, reproduction includes all those
events surrounding insemination (the transfer of a male’s sperm to a female) and
fertilization (the fusion of a sperm and egg to create a diploid zygote). We are
accustomed to the two events being closely linked, but in insects they often are
not, because insemination can be either internal or external and can either be direct
or through an intermediary step that involves a transfer container, or spermatophore
(see Plate 29). In addition, fertilization can occur as much as years later because
long-lived sperm can be stored in a special female pouch called a spermatheca.
Thus, rather than spanning all of reproduction, the narrower term ‘mating behavior’
is used to describe the more immediate actions that surround insemination itself,
from pair formation through courtship and copulation to the final breakup of the
mating pair.
The reproductive behavior of the primitively wingless (apterygote) insects has
received relatively little attention. In those that have been studied, an almost com-
plete dissociation of individual males and females occurs, each sex reacting to a
9.2 Courtship and Mating 343
Fig. 9.2 Indirectly transferring sperm. Prior to copulation, the male of dragonflies and dam-
selflies transfers sperm into a pouch-like structure on the underside of the second or third
abdominal segment. During copulation, the female bends the tip of her abdomen forward to con-
tact the male’s accessory genitalia. The male (above) uses claspers at his abdominal tip to grasp
the female’s prothorax, thus completing the “wheel position” characteristic of this group (right).
Pairs of damselflies or dragonflies often fly about for hours in tandem, the male holding the female
by his anal claspers (left); in some species the female oviposits during tandem flight
spermatophore rather than to each other. Such indirect sperm transfer seems to char-
acterize Archaeognatha, Thysanura, Collembola, Protura, and Diplura. In contrast,
in most winged (pterygote) species, sperm transfer is direct, through copulation. For
these insects, mating almost always involves chain-like interactions (see Fig. 7.9)
that begin with searching for mates, move on to pair formation and courtship, and
progress to copulation. For many, post-copulatory male behaviors culminate the
reaction chain.
Given the human interest in courtship and mating, it is perhaps no surprise to
find an enormous amount of published information on this aspect of insect behavior,
nor is it unusual that different approaches to the subject have arisen. The field has
become far too large to review comprehensively, but a brief overview can suggest
some of the directions that are being taken.
Early research on insect mating behaviors was almost solely descriptive. As
would be expected in a taxon as large as Insecta, the biological diversity in mating
behavior is enormous. Just to chronicle variation in the manner in which copulation
is accomplished has begun to produce an insect Kama Sutra. That task is still far
from finished and it is still a useful approach, not only in its own right but also in
providing the raw materials with which other analyses can proceed. Despite tremen-
dous diversity in the ways they may accomplish their goal, however, most insects
must solve a strikingly similar set of problems. They must find a partner, recog-
nize it as such, comport themselves appropriately so that insemination will occur
and be successful, and—for males—counteract potential sperm competition. Thus,
mating behavior studies increasingly are focusing upon broader and more theoret-
ical questions of causation and organization, fitness consequences, evolution, and
sexual selection in the context of costs and benefits to the individuals involved.
9.2.1 The Physiology of Mating Behavior

Females of the polyphemus moth, Antheraea polyphemus, start their distinctive
calling behavior only after exposure to a volatile chemical, trans-2-hexenal, that
is present in the leaves of their host plant, and then only at a specific time during
the photoperiodic cycle. Allatectomized females of the Cuban burrowing cockroach,
Byrsotria fumigata, fail to produce their sex pheromones and do not attract males;
when active corpora allata are implanted or synthetic juvenile hormone analogs are
injected into these females, their pheromone production ability is restored. Aging
male Nasonia vitripennis wasps court more and more slowly.
For something as vital as attracting a mate, it is not surprising to find that insects
use every sensory modality at their command, but chemical sex attractants certainly
predominate (see Chapter 6). For indicating the location of a receptive potential
sexual partner, the specificity and long-range effectiveness of pheromones are unex-
celled. These may function as attractants, excitants, arrestants, identifiers of species
and/or sex, or releasers or have several of such effects serially or concurrently.
Either sex may produce them, and sometimes both sexes have their own, each with
a distinctive role.
Neuroendocrine control of insect reproductive behavior tends to occur in those
insects that have a long adult life, a life history that includes repeated reproductive
periods, and a life span that includes time periods when mating is inappropriate or
even impossible. Cockroaches are an example. For most insects that feed as adults
and require neuroendocrine factors to mature oocytes, some type of neuroendocrine
sexual regulation is the rule. In insects that do not feed as adults, such as many
Lepidoptera, females from which the juvenile hormone-producing apparatus has
been removed can still lay fertile eggs; in at least some species, however, neuroen-
docrine factors are still necessary for proper environmentally cued release of sex
pheromones. For other insects that are very short-lived as adults—mayflies being
an ultimate case—mating and oviposition must happen quickly before they die.
With a premium on efficiency, mating is more efficiently built into the developmen-
tal process, appearing automatically with adulthood and divorced from extensive
endocrine controls. Most insects fall between these extremes, of course. A common
female reproductive pattern includes a single adult period of feeding and oocyte mat-
uration, followed by mating and then by an oviposition bout (or bouts interrupted
by additional feeding and often additional matings as well).
Not surprisingly, a great deal of research has focused on the triggers that switch
insect reproductive behavior on and off, because such knowledge promises both
theoretical and enormous practical benefits. Just as for much of the rest of the animal
world, the triggers are usually different for males and females.
Male sexual behavior usually matures along with the metamorphosis of the ner-
vous system; following ecdysis, expression of full copulatory behavior awaits only
suitable stimuli. In most such cases, the corpora allata are not involved, although
sometimes juvenile hormone lowers the male’s behavioral threshold to the female.
The endocrine control of male sexual behavior has been best studied in the grasshop-
pers and locusts, where some species require the corpora allata for the maturation
of male sexual behavior. For example, allatectomy of newly ecdysed males of the
desert locust, Schistocerca gregaria, and the red locust, Nomadacris septemfasciata,
completely prevents the onset of male sexual behavior; if, however, several pairs of
active corpora allata are transplanted into an allatectomized male, sexual behavior
reappears.
Particularly among longer-lived insects that have a restricted breeding season,
male sexual behavior often is modulated through an interaction between the cor-
pora allata, the brain’s neurosecretory system, and environmental parameters such
as photoperiod. For example, under long-day conditions, males of some locust and
grasshopper species show little or no sexual behavior, apparently due to cessation of
neurosecretory activity in the brain. Sexual behavior resumes, however, when these
uninterested males receive active corpora allata implanted from other locusts reared
under short day conditions.
Female insects usually pass through three distinct behavioral states: young vir-
ginity, unreceptive to courtship; mature virginity, responsive to male advances;
and mated, typically unresponsive toward males but showing intense oviposition
behavior. The switch from one state to another is often abrupt and often hormon-
ally induced. Hormones regulate pheromone production and release so that mating
will occur at the appropriate stage in the reproductive cycle of the emitter and/or
in the proper environmental context. Though chemical systems have received the
most attention, neuroendocrine control of sexual signaling should apply equally to
nonchemical communicative modes.
Cockroaches provided some of the first evidence that hormones mediate the
onset of insect sexual behavior. About fifty years ago, Franz Engelmann reported
that when he allatectomized newly emerged females of Rhyparobia (=Leucophaea)
maderae, 70% of them never displayed sexual receptivity. When he reimplanted cor-
pora allata, normal receptivity ensued. However, when other researchers attempted
to duplicate Engelmann’s results, to their surprise their allatectomized females
became receptive and mated normally. To resolve the conflicting results, the
researchers repeated their experiments together, this time both using Engelmann’s
cockroaches. Nearly all the control roaches mated, usually within a few minutes
after the introduction of males. Allatectomy, on the other hand, once again led to
a high percentage of non-mated females, and those that did mate required over an
hour before males could induce them to copulate. The different experimental results
were due to differences in the two laboratory populations, underscoring both the
importance of heredity in behavioral studies and the dangers of generalizing from a
single strain of laboratory animal.
In many species of insects, females will no longer accept courtship advances after
successful mating. Female grasshoppers, for example, show secondary defensive
behavior rather than receptivity toward males once a spermatophore is placed in the
genital tract. For some species, such refractory behavior is temporary; in others it
lasts for life. A great variety of mechanical or chemical stimuli may be responsible.
In some cockroaches that respond to mechanical stimuli, inserting a glass bead into
the bursa of a virgin female will stimulate refractoriness. Other insects respond to
chemical cues such as juvenile hormone withdrawal or the release of chemicals from
the maturing ovaries. In many cases, secondary reproductive structures such as male
accessory glands or female bursae release agents that trigger the switch.
An interaction between mechanisms is probably common, as has been shown in
some Diptera; mating first results in a rapid neurally mediated refractory state, then
is followed by a more slowly developing but more persistent hormonally induced
response. As the bursa copulatrix of the yellow fever mosquito, Aedes aegypti, fills
with seminal fluid, an immediate (probably neural) switch-off of receptivity occurs.
However it is the action of a chemical (matrone) from the male accessory glands
passing through the bursa walls into the hemolymph that keeps females behaviorally
unreceptive to further mating for the rest of their lives. In fact, injecting matrone into
the haemocoel is an equally effective way to produce refractoriness.
The implicit assumption that ‘mated’ equals ‘inseminated’ should be carefully
evaluated in each case, because the two are different events. For example, although
new adult female Aedes aegypti mosquitoes appear to mate readily as soon as
they begin to fly, they actually are unable to accept the transfer of sperm for at
least 24 hours after they eclose. Certain tiny parasitic wasp females will mate only
once; thereafter they rebuff the overtures of the males that nevertheless persistently
attempt to court them; an occasional virgin female, having been intensely stimu-
lated by male courtship, reacts as though inseminated and thereafter behaves like a
mated female even though her actual physiological status is still virgin. Such exam-
ples underscore the importance of using insemination rather than copulation as the
criterion for a successful mating.
9.2.2 Reproduction Modes
The ways in which insects reproduce vary in three important aspects: whether eggs
are fertilized, whether eggs are provisioned, and where embryonic development
takes place. The most common method of insect reproduction is by yolked eggs, fer-
tilized internally and laid outside the body. However, as one might expect in a taxon
as large as Insecta, other means are not uncommon, and some species approach the
task in quite unusual ways.
Take the matter of where embryonic development occurs. In Chapter 1 we con-
sidered oviparity, ovoviviparity, and viviparity among cockroaches, and it seemed
quite straightforward. Oviparous insects, by far the numerical majority, lay eggs.
Ovoviviparous insects do the same but retract their eggs back into a brood pouch
to develop, extruding them into the world again once they become nymphs. Going
a step further, viviparous insects ‘feed’ their developing young within the brood
pouch with nutrients and water.
The overall picture is not quite this simple, however, because discrete terms can-
not adequately describe a continuum, and insect reproductive modes clearly form
a continuum. For example, several chrysomelid beetle species have embryos that
develop in the ovary so quickly that they hatch a few minutes after deposition.
Should they be considered oviparous or ovoviviparous? They do not differ essen-
tially from close relatives in which the embryos hatch within the genitalia just a
few minutes earlier. Furthermore, in addition to the reproductive modes mentioned
above, insects also show a wide variety of more unusual types of development.
In a smattering of insects (and quite regularly among endoparasitic ones) each
egg develops into a number of larvae rather than just one. This process, called
polyembryony, theoretically would have the effect of increasing the reproductive
potential of the insect involved, but because polyembryonic forms tend to lay fewer
eggs than related ‘normal’ species the actual net effect is not always much greater.
However, the metabolic costs are lower.
In other cases insect eggs develop without being fertilized, a situation known as
parthenogenesis. Because it has been recorded as an occasional event in nearly every
insect order, sporadic parthenogenesis is probably widespread when a female fails
to find a mate. Clearly the phenomenon should have short-term advantages. When
only female offspring are produced, a mother’s immediate reproductive potential is
much greater than if half the population is male, and a parthenogenetically reproduc-
ing female can spend all her time and energy in feeding and reproduction, without
having to locate a mate and/or court.
Then why is sexual reproduction the rule, rather than the exception? The classic
argument has been that over the long run asexual reproduction has an impor-
tant drawback—it lacks the genetic recombination that normally occurs at mating.
Sexual reproduction is favored because it allows selectively advantageous inter-
genomic recombination that promotes adaptation to and persistence in novel and
rapidly changing environments; it also purges weakly deleterious mutations from
the genome.
Debate arises, however, with certain predictions that arise from this cost/benefit
theory. One is that the transition from sexual to asexual reproduction should occur
frequently because of the short-term success of asexual lineages. The other is that
asexual lineages should suffer rapid extinction as the disadvantages play out. From a
phylogenetic perspective, this means that obligate asexual lineages should be poly-
phyletic and young. However, ancient asexual lineages clearly exist, and in a number
of species, parthenogenesis has become the normal method of reproduction.
This is where cyclically parthenogenetic taxa fit in. A life history of alternating
parthenogenetic and bisexual generations looks like an evolutionary compromise,
and a successful one at that, for over 15,000 species in three animal phyla have
been discovered to utilize it. Cyclical parthenogenesis suggests that there must be a
short-term advantage to maintaining a sexual generation. If not, such taxa would be
expected to rapidly move to obligate asexual reproduction.
In most of the cyclically parthenogenetic arthropods that are known, multiple
asexual generations are punctuated by occasional sexual generations, with male pro-
duction often triggered by a change in environmental conditions. Among Andricus
Fig. 9.3 Alternating diversity. Some cynipid wasp species produce galls on two alternate host
plants at different stages of the wasps’ life cycle; among other cynipids, asexual and sexual gener-
ations produce different galls on the same host plant. In some species of Andricus, a large genus of
oak gallwasps, the adults of the sexual generation are morphologically quite similar, and the galls
they produce are small and structurally simple. However, asexual generation adults are morpho-
logically distinct and produce large, complex, structurally diverse galls such as those shown here
for (A) A. glutinosus, (B) A. coriarius, and (C) A. pictus. Scale bars are 1 cm
oak and sycamore gall wasps (Fig. 9.3), however, sexual and asexual generations are
produced in strict alternation. However, the adults and galls of the sexual and sexual
generations are so different morphologically that historically they have often been
placed in different genera. Furthermore, whereas the galls of the asexual generation
are large, complex, and long-lived, those of the sexual generation are small, cryp-
tic, and short-lived. Thus, it is not surprising that the sexual generations have been
under-sampled and misidentified, and many fewer species are known only from a
sexual rather than an asexual generation.
New light was shed on the debate in 2008, when an international team of
researchers led by Graham Stone used molecular techniques based on analyses of
allele frequency data to test for cryptic sexual generations in eight apparently asex-
ual European species in the gall wasp genus Andricus, and found that cryptic sexual
generations were widespread. Their results led support to the side of the debate that
suggests either that selection acts against the loss of sex or that constraints make loss
of sex difficult. The research team also found four Andricus species that appeared to
reproduce only sexually; because no known oak gall wasps have such a life cycle,
these discoveries imply the existence of unknown asexual generations in a genus
that scientists had thought were quite well known.
Aphids provide another example of relatively well-known taxa that continue to
yield surprises. Some aphids lay eggs. Others give birth to their young (see Plate
30). Still others are some of the few cases among insects in which immature insects
are able to reproduce. Apparently as a result of a hormone imbalance, development
of the offspring usually begins while the parent is still a larva itself. Most insects that
show such paedogenesis are also parthenogenetic and viviparous. When all of these
phenomena converge into a single life history, the situation can become incredibly
complex.
In tropical regions, most aphids are continuously parthenogenetic, but in tem-
perate zones they have developed a complex alternation of generations. Throughout
the summer, generations succeed one another so rapidly that while embryos are
developing in a mother’s egg tubes, they already have embryos developing in
them. Thus, the female aphid is not only an expectant mother but an expectant
grandmother! In autumn, males appear and normal sexual and oviposition behav-
ior are resumed. Eggs overwinter and, upon hatching the following spring, resume
rapid-fire parthenogenetic reproduction once again.
Overlaid upon this system there are usually also two other seasonal cycles: the
occurrence of several aphid morphs (physiologically and morphologically distinct
forms of the same species) and an alternation of host plants. For example, in Europe
members of the first spring generation of bean aphids, Aphis fabae (see Plate 10),
emerge from overwintering eggs on spindle trees, a species of Euonymus. This gen-
eration is entirely female and parthenogenetically produces either a winged female
generation or several wingless generations of females which themselves give rise
to winged forms. These migrate to bean plants, there to produce another series of
wingless morphs. Ultimately, these produce both winged and wingless sexual forms.
The winged aphids return to spindle trees and produce females; the wingless ones
go on to produce winged males that then join those females to mate and produce
winter eggs.
9.2.3 Complexity and Plasticity
Two closely related sympatric species of Melittobia parasitoid wasps differ strik-
ingly in the occurrence and sequence of different elements of their courtship. In
M. australica, males pump the female’s antennae slowly maintaining continuous
contact and only periodically lift their middle legs; complete courtship requires an
average of ten minutes. Males of morphologically similar M. megachilis pump the
female’s antennae discontinuously in rapid alternation with an up-and-down move-
ment of the hind pair of legs; courtship duration for this species is slightly less
than two minutes. The two ecological homologues may parasitize the same host
in a single back yard, but the different courtship patterns will not permit them to
cross-mate.
As this example illustrates, insect courtship displays have at least three common
functions: to promote the meeting of solitary individuals, to facilitate species and
sex identification, and to stimulate and maneuver females into copulation. Among
predatory insects such as robber flies and scorpionflies, an important additional
function is appeasement, the inhibition of the normal predatory instincts of the
participants, especially the female.
The traditional view of insect courtship and mating has been that these behaviors
are relatively constant within a species and that, by differing from one species to
ACTIONS CONTEXT
Emergence from hosts
O
O
Search for mating partner
Encounter Rivals compete for access

to females
Display
Stimulus quality
Becomes receptive Remains receptive Female choice
Gives up Signals receptivity How much courtship?
Switch to copulatory behaviour Switching-off procedures
Genital contact Copulatory mate guarding
Inseminates Stores sperm Post-copulatory courtship
Off Mate guarding Sperm depletion
Sperm competition
Off
Search for hosts
Host-quality assessment
Sperm recruitment (fertilisations)
Oviposition Offspring sex-ratio
Emergence from hosts Eventual mating sucess in

local mating pool
etc.
Fig. 9.4 Seeing the big picture. A schematic representation of the course of events in a typical
mating behavior sequence. Shaded and unshaded parts of the boxes denote the contributions of the
female and male, respectively
another, they serve as isolating mechanisms that effectively maintain and define
discrete species. An attractive analogy (Fig. 9.4) states that insect courtship operates
like a ratchet regulated by innate releasing mechanisms or physiological filters; at
each stage, some signal on the part of one participant elicits a response in the other.
This in turn elicits a new action by the initiator, and so on. Thus, the pair clicks along
toward copulation while the ratchet mechanism precludes inappropriate mating or
steps being performed out of sequence. Reproductive isolation is generally assumed
to be a principal function of such a reaction chain.
Does courtship actually perform this function? To answer such a question, one
must turn to analytical analyses of individuals reared in isolation in order to exclude
learning or imprinting. As an example, consider the investigation in Case Study 9.1,
which detailed the courtship reaction chain in crane flies.
Case Study 9.1: Courtship in the Crane Fly, Tipula oleracea
Crane flies generally look alike and tend to occupy the same sorts of habitats.
What courtship cues do conspecific males and females rely upon?
Crane flies, which look rather like delicate, fragile, oversized mosquitoes,
include a great number of species that are generally similar in appearance and
tend to occupy the same sorts of habitats. While the males mate repeatedly,
females typically mate but once. Under such conditions, a complex series of
reciprocal signals would be predicted; in what specific ways might this be
accomplished?
In the early 1960s, a Canadian researcher, H.J. Stich, raised crane flies
in the laboratory and kept them isolated from one another as adults. When
the adults were five days old, he began mating trials. The crane flies mated
readily, in a rapid series of actions and reactions that sometimes took as lit-
tle as 15 seconds and usually lasted only 90 seconds. Though the female
often began the courtship by touching a male’s long threadlike legs, it
was apparent that the male took the most active role in courtship as a
whole. True to the ratchet analogy, when Stich interrupted the courtship
pattern by removing the female, reintroducing her never allowed the pat-
tern to resume. Each time, the males began courtship anew from the
beginning.
The first detectable step of the crane fly courtship occurs when two individ-
uals happen to contact each other’s legs. A male’s response was to grab the leg
with his own. Stich began amputating legs before introducing live females to
males; six, five, four, three, two, or one leg—it made no difference to sexually
active males. Only when a female had no legs at all did males fail to respond.
Next, Stich picked up amputated legs in his tweezers and offered them.
Males immediately grabbed them. Was any discrimination occurring at this
step? Apparently not in relation to sex, because dead males or their legs
evoked the same grabbing reaction as females did. In fact, any leg-like tac-
tile stimulus seemed sufficient—males even readily grabbed paraffin-stiffened
threads dangled near them. But when stiffened threads of different thicknesses
were presented, those of the ‘wrong’ diameter were immediately released.
Stich repeated the experiment again, this time using legs of different diame-
ters from other species of Tipula. Male T. oleracea grabbed at them all but
immediately let them go. Here was Stich’s clue to an initial discrimination
that probably helped to ensure that the potential mate belongs to the proper
species.
Stich observed that receptive females respond to the male’s grasp by lift-
ing one or more of their legs upward, and that afterward, the male assumed a
mounting position above the female’s body. What exactly was the female sig-
naling with this behavior? To answer this question, Stich constructed a number
of models. He found that males refused to mount the model made from a dead
female with artificially stiffened legs, but they would mount a variety of mov-
able models, even a simple 1 mm long wire if it were fashioned into a pair
of movable legs. So, if to a male moving legs signify a female, what signifies
another male? Stich repeatedly caged pairs of males together and watched
their behavior. Males grabbed each other whenever their legs touched, but in
over 100 trials neither of them ever responded by upward leg movement nor
did they ever exhibit a mounting reaction.
After male leg-grabbing and female leg-lifting, a male crane fly tries to
pin down the female’s raised legs. Sometimes, this is accomplished in less
than 3 seconds; at other times, the female may resist, and the ensuing struggle
over leg position can extend over a period of 3 minutes. On the hunch that
this great variability might be associated with sexual receptivity, Stich pre-
sented males with previously mated females; these readily elicited the males’
grabbing and mounting reactions, but all their attempts at pinning down the
females’ legs failed. Sexually unreceptive females kept at least one leg contin-
ually raised, eventually causing the male to leave. Thus, previously mated or
otherwise unreceptive females effectively terminate courtship attempts at this
point.
After successfully pinning a female so that she remains motionless,
a male begins to locate her head, by contacting her with his mouth-
parts while he moves progressively forward over her body. When he
reaches her head, he touches his mouthparts to the back of it and stops
his search. To determine the stimuli for this particular element in the
courtship sequence (which Stich anthropomorphically called ‘kissing’),
the researcher exposed males to variously altered females (see Fig. 9.5).
Male reactions to these models clearly indicated that both the female’s
body and her head were necessary as a stimulus at this point in the
courtship.

Fig. 9.5 Searching reactions. Male crane fly behavior toward experimentally altered females. (a)
Decapitated female: searching extends beyond the body until male contacts substrate. (b) Female
head glued onto paper strip: male uses strip as guide to reach head, but percent completion declines
as strip length increases. (c) Female head attached to end of abdomen, model presented so that male
and female heads point in same direction: male orients toward head, attempts copulation through
neck cavity
Fig. 9.5 (continuted)

The final step in the courtship sequence occurs when the male slowly steps
backward, sliding his abdomen back over the female’s, and bends his abdomen
in such a way that the male genitalia come into contact with the female’s. Does
the female head or the ‘kissing reaction’ release this behavior? Or does the
abdomen function as the sign stimulus? To answer, Stich removed the heads
from females and attached them in various ways. Male responses to reversed-
head females showed clearly that the head, not the abdomen, was the sign
stimulus.
Thus, Stich was able to show that crane fly courtship facilitates dis-
crimination with respect to sex, species, and physiological condition of the
participants. At each stage of their courtship, different behaviors and differ-
ent body parts are employed as specific signaling devices in a touch-oriented
chain ideally suited for such weak-flying, drab denizens of the woods, and
courtship proceeds like a rigid sequence of filters excluding all but conspecific
sexually active males and receptive unfertilized females.
As this study illustrates, the use of experimental models can be particularly help-
ful in teasing apart the roles of individual behaviors in the complex interactive chains
that characterize many cases of insect courtship.
9.2.4 Pollination and Male Reproductive Behavior
Approximately 65% of all flowering plant species are known to be insect pollinated.
As a group, insect-pollinated plants generally have relatively showy and conspic-
uous flowers (which may however appear quite different to an insect’s eye than to
our own; see Plate 24). They produce pollen over a period of time and characteris-
tically produce nectar as well. In contrast, wind-pollinated plants have rather small
and inconspicuous flowers, generally lack nectar, and produce copious amounts of
pollen in concentrated flushes. Corn is a familiar example; most grasses and their
relatives are pollinated by wind, a not surprising fact since grassland habitats tend
to be relatively windy and open. An exception to this rule, certain bamboo-like
grasses that grow deep inside tropical forests where wind circulation is nil have
secondarily evolved insect pollination systems, an example that serves to reempha-
size the importance of ecology and especially population and community structure
in understanding insect–plant coevolution.
Pollination is often closely interwoven with other aspects of insect behavior, and
coevolved adaptations are common. One particularly striking case, the fig wasp, was
examined in Chapter 4. A few other examples will illustrate some of the variety that
occurs in the interrelationships between pollination and reproductive behavior.
Some of the strangest cases of insect pollination involve orchids. One of the
two largest families of flowering plants, the Orchidaceae includes some 880 genera;
there are said to be over four times as many orchid species as mammal species
on this earth. Although most are found in the tropics of Asia, South America and
Central America, orchids are cosmopolitan, and occur in almost every habitat but
deserts and glaciers.
While some orchids are self-pollinated and a few reproduce vegetatively, most
require cross-pollination. However, orchids have developed highly specialized pol-
lination systems and the chances of being pollinated are often scarce. In addition,
orchid seeds are generally almost microscopic and very numerous, in some species
over a million per capsule. After ripening they blow off like dust particles or spores.
They lack endosperm and must enter symbiotic relationship with various mycor-
rhizal fungi that provide them the necessary nutrients to germinate, so that all orchid
species are reliant upon fungi to complete their lifecycle.
In response to such long odds, orchid flowers usually remain receptive for very
long periods and most orchids deliver pollen in a waxy mass called a pollinium
(plural, pollinia) Each pollinium is connected to a stalk-like filament that ends in an
adhesive pad (viscidium) that sticks to the body of pollinators; taken together, the
entire apparatus is called a pollinarium (plural, pollinaria). In functional terms, this
means that the whole pollen content of a flower is removed during a single pollinator
visit and each time that pollination succeeds, a great many ovules can be fertilized,
making pollination a high-stakes game.
In orchids that produce pollinia, pollination happens as some variant of the fol-
lowing. When the pollinator enters into the flower, it touches a viscidium, which
promptly sticks to its body, generally on the head or abdomen. While leaving the
flower, it pulls the pollinium out of the anther, and the springy stalk bends, mov-
ing the pollinium forward and downward into a (generally orchid species-specific)
position such that when the pollinator enters another flower of the same species the
pollinium will stick to the stigma of the second flower, pollinating it (Fig. 9.6). Bees,
wasps, flies, ants, and moths are all exploited for orchid pollination in this manner.
Orchids go to great lengths to attract pollinators, guide them into the proper
positions, and assure that pollinia attach to an insect body location where they
will successfully be removed during the insect’s next visit to their particular orchid
species. The strangely contorted flower shapes that humans find so interesting are
actually adaptations that guide insect visitors in ways that promote pollination.
Sometimes, such promotion can grade into coercion. The bag-shaped bottom petal
of one group of orchids traps visiting insects; the only exit leads to the anthers
that deposit pollen on the visitor. Catasetum, a genus discussed briefly by Darwin,
actually launches its viscid pollinia with explosive force when an insect touches it.
For nearly 700 specialized orchids in the Neotropics, the pollinarium system
is carried even further. These flowers lack nectar, the usual reward for bees that
visit flowers, but produce such strong species-specific fragrances that they are often
called ‘perfume orchids.’ A bee attracted by the odor approaches a blossom from
downwind and upon landing, alternately brushes the surface of the basal flower
petal with his front legs, then hovers in front of the flower while scrubbing his legs
together as if to transfer some substance to his hind tibiae. Individual bees may stay
with a given blossom for up to 90 minutes, with repeated bouts of brushing and
hovering–transferring. Toward the end of prolonged visits the bee becomes much
less wary and its overall behavior suggests nothing less than increasing intoxication!
Fig. 9.6 Sticking to carriers. Outlines of Neotropical male euglossine bees showing the pollinia
of 11 different species of orchid flowers, each deposited in a precise location on the bee’s body.
Upon entering a flower, the bee either squeezes down a narrow pathway and brushes against the
sticky viscidium or triggers a mechanism that causes the pollinia to be ejected (sometimes with
rather great force) onto its body. On a visit to another conspecific flower, the same sequence occurs,
except that as the pollinator leaves, the attached pollinia are forced into the sticky stigmatic cavity
on the flower and are removed before the insect receives its new pollinia
Charles Darwin provided an early description of this strange behavior; at the

time, he believed the bees were females. However, this coevolved pollination
scheme is stranger than he guessed. The pollinators, it turns out, are males in a bee
tribe called the Euglossini; females never visit these flowers. These small, brilliantly
metallic ‘orchid bees’ differ in many aspects of their behavior, but all euglossine
males share two peculiar morphological traits: dense hair ‘brushes’ on their front
feet and greatly swollen but hollow hind tibiae. They pick up floral chemicals from
the orchid with special brushes on the forelegs, transferred from there by rubbing
the brushes against combs on the middle legs, and finally these combs are pressed
into grooves on the dorsal edge of the hind legs, squeezing the chemicals past the
waxy hairs which block the opening of the groove, and into a sponge-like cavity
inside the hind tibia.
The behavior of volatile collection is essentially unique in the animal kingdom,
but almost 200 species of orchid bees obviously find it important. What’s in it for
the bees? It is clear that later, the males will release their fragrant chemicals at sites
in the forest understory where matings are known to take place. However, the exact
reason behind their fragrance collection continues to be elusive.
In southern Florida, USA, where perfume orchids are absent, an introduced
orchid bee has become naturalized; an analysis of the contents of male tibial storage
organs found 55 fragrance compounds, half of which were identical to those pro-
duced the bee’s natural orchid mutualists in Mesoamerica. The fact that aromatic
leaves such as basil can serve as important ‘surrogate sources’ of volatiles for the
bees suggests that the orchid bee/perfume orchid mutualism may be facultative for
the bees, even though it is obligatory for the orchids.
As in other orchids, orchid bees collecting volatiles pick up pollinia specifically
placed to ensure that cross-pollination only occurs between orchids of the same
species, but for the orchid an added dimension of specificity has been obtained
because each orchid species has a characteristic fragrance spectrum, and different
orchid bee males are attracted to different chemicals. This can be particularly advan-
tageous in the tropics, where plants generally do not grow in groups, and individual
plants of the same species may be miles apart. Orchid bees are believed to forage
on specific plants along set routes, a behavior known as trap-lining.
Over 60 distinct attractive chemicals have been isolated; one of the most com-
mon is cineole (Fig. 9.7). When tested alone under field conditions, cineole attracted
35 of 57 different species of male euglossine bees native to the area. But when one
or more of the other compounds was combined with the cineole, the number of
attracted species declined dramatically. Taking advantage of the specificity of this
mutualism, scientists commonly use single synthetic compounds as bait to attract
and collect euglossine males for study. The attractant compounds include many
familiar flavorings and odors considered appealing to humans (such as methyl sali-
cylate, eugenol, benzyl acetate, and methyl cinnamate), and others (such as skatole)
that are not. The euglossine male Eufriesea purpurata is highly unusual in actively
collecting the insecticide DDT in huge amounts from houses in Brazil, without
suffering any harm from it.
Whereas perfume orchids simply offer chemical attractants, other orchids appear
to offer the promise of sex—they have evolved to resemble female versions of cer-
tain insects. This sexual deception strategy is practiced by various Australian orchids
and dainty Mediterranean orchids in the genus Ophrys. Known since the days of
Pliny the Elder (23–79 A.D.), Ophrys flowers are commonly called bee orchids
Fig. 9.7 Smelling attractive. A male euglossine bee attracted to cineole daubed on the label of
this commercial chemical
because even to human eyes the flowers of some species resemble the furry bodies
of bees and other insects. Like the orchids above, these do not secrete nectar, nor
is their pollen available to most insects. Female aculeate Hymenoptera are never
seen visiting these flowers, but males of certain species of wasps, bees, and beetles
approach the flowers, their mating instinct stimulated by chemicals exuded from
the orchid’s basal petals. At close range, the flower petals mimic female insects,
attracting and duping the right pollinator. As they shift about, attempting to copu-
late (Fig. 9.8), the ardent males loosen the pollinia, which then stick to their bodies
in positions specific to the orchid species involved.
Fig. 9.8 Tricking pollinators. Flowers of the orchid, Ophrys, trigger copulatory behavior by
males in a number of unrelated genera of aculeate wasps (a, b) and bees (c, d). After a chemically
mediated attraction, vital tactile stimulation is provided by the form and construction of the orchid
labellum and by characteristics of the hairs upon it, including their direction, length and grouping.
Each of the approximately 30 Ophrys species depends upon different Hymenoptera species for
pollination accomplished by entering the flower frontwards or backwards, depending on the insect
species: (a) Campsoscolia ciliata on O. speculum, (b) Argogorytes mystaceus on O. insectifera,
(c) Colletes cunicularius on O. “arachnitiformis-sphecodes” of sphecodes type, (d) Andrena sp.
on O. fusca
9.3 Courtship and Conflict 359
Given the number of orchids in the world, more examples of pseudocopula-

tion will probably be found. In Australia, sexually stimulated male ichneumonid or
thynnid wasps assault quite unrelated orchids. In some cases the orchid’s mimicry
seems near-perfect. For example, the Australian hammer orchids (Drakaea spp.)
have taken advantage of a mating ritual of a thynnid wasp species that involves a
wingless female wasp waiting on top of a branch or plant for a male to spot her
(see Fig. 6.7, right). The hammer orchid’s flower mimics the female wasp looking
upward for a male flying by, complete with a fake shiny head and furry body. The
orchid even releases an enticing female wasp pheromone mimic. Place a real female
wasp next to the orchid mimic, however, and the male wasp will spot the real deal.
Probably because males get nothing in return for their pollination services, natural
selection has favored flowers that bloom in the period when male wasps are flying
but females are not yet circulating.
There is no reason to think that wasps and bees should be the only insects
duped in this way. In 2005, Bradysia fungus gnat males were discovered to be the
sole visitors and pollinators of at least four species of Lepanthes orchids. These
orchids belong to one of the largest angiosperm genera, with over 800 species.
Their tiny, colorful, structurally complex flowers offer no nectar rewards; given
that most species in the genus have the same unusual flower structure, it seems
likely that pseudocopulatory pollination will be found to be prevalent in this genus.
The discovery that the fungus gnats ejaculate during their flower visit suggests
that the term pseudocopulation might better be replaced by a term such as sexual
deception.
Pollination is most often discussed as a feeding mutualism because generally
we think of flowers as offering food such as nectar to the insects that polli-
nate them. Cases of deception in feeding evolution have produced many flower
species that pretend to have food that insects want, emitting scents of coconut
or even rotting meat (see Plate 9). Some evolutionary biologists have suggested
that posing as a sexual suitor may not only facilitate pollination of already widely
separated orchids, but may actually allow the geographic spread of plants over a
wider area. After all, insect generally will travel further to find a mate than to
find a meal. Testing such hypotheses will require further studies of both the mat-
ing behavior of the insect pollinators and the growth and reproduction of deceptive
orchids.
9.3 Courtship and Conflict

Among some desert Orthoptera, silent satellite males become courtship parasites
on calling males by intercepting females attracted to the acoustic signaling of the
latter. In the parasitoid wasp Megarhyssa, males of three species gather and jos-
tle one another even before a female emerges; when she does appear, conspecific
males fight for access to her (Fig. 9.9). About 10 minutes before an adult female
crab-hole mosquito leaves her pupa, males begin to congregate about the emergence
Fig. 9.9 Listening closely.

Males of two ichneumonid
wasp species, Megarhyssa
macrurus and M. greenei,
assemble over an emerging
female. The stimuli that draw
them to the site are apparently
auditory but not
species-specific; once the
female chews out through the
bark, conspecific males
compete for the opportunity
to mate, while males of the
other species disperse
site. When the pupal cuticle begins to split, a frantic and violent scuffle breaks out
between the males as each attempts to establish possession. The victor monopolizes
the emerging female by standing over her and holding her in his legs; mating is
established by the time she is free from the cuticle.
When the product of an interaction has the potential to yield different optima for
different individuals or classes of individuals, conflict is the usual result, because
each party in the interaction will be selected to manipulate the transaction in
ways that move the result closer to its own optimum. Generally, these manip-
ulations take one of three forms: persuasion, coercion, or force. In persuasion,
one individual offers a fitness incentive that would increase the benefits provided
to the other party. In coercion, one party (generally the one furthest from its
optimum) attempts to gain control through actions that exert a cost to the other
party. Force, of course, involves taking control away from the other individual
by physical, behavioral or physiological means. The processes are not mutually
exclusive, of course, and individuals often switch from one approach to another
through time.
Sex would seem to be the ultimate social act. Yet, as Edward O. Wilson has
pointed out in the context of social insects, in many ways sex can quite prop-
erly be considered an antisocial force in evolution. The ‘ideal’ society would lack
conflict and would have a high degree of selfless behaviors and coordination, a
situation most likely to evolve where all members of the group are genetically iden-
tical. Within the invertebrates, the highest forms of sociality are found in groups
that create new colony members by budding—the sponges, tunicates, etc. Sexual
reproduction dilutes genetic relationships, causing a one-half reduction in genes

shared among parents and offspring. Thus, a conflict of interest inevitably arises
as each mate, parent, or offspring strives to increase its personal genetic fitness at
the expense of others. In this view, courtship and sexual bonding are means by which
organisms override the antagonism and tension that result from genetic differences
induced by sexual reproduction. The complex mixtures of conflict and cooperation
that result make reproductive behaviors both endlessly fascinating and challenging
to study.
Darwin’s theory of sexual selection remained dormant for almost a century after
he proposed it. However, once resurrected, interest in the subject grew with amazing
speed. For the past decade, the most active topic in behavioral ecology has been the
variety of ways in which males and females pursue, persuade, manipulate, control,
and aid one another so as to reach a common goal of producing offspring while still
assuring parity for their own individual interests and resources.
Until the mid-1960s, researchers generally viewed mating as a sexually cooper-
ative endeavor. Then, in an influential 1966 book, George C. Williams persuasively
presented another viewpoint—an evolutionary battle of the sexes. His critical
analysis suggested that because each gender essentially has its own best interests
at heart, ‘genic selection will foster a skilled salesmanship among the males and an
equally well developed sales resistance and discrimination among females.’
One of the first predictive theories of mating behavior and sexual conflict to
be based on individual selectionist thinking was the polygyny threshold model
proposed in 1969 by Gordon Orians. Based on studies of birds and mammals, its
simple premise was that females should select mates according to the quality of the
male’s territory and the potential for paternal care. The model (and the refinements
and elaborations that arose from it) soon spawned a host of new studies on sexual
selection, as researchers found that an emphasis on male behavior and the influence
of ecological factors allowed them to generate a host of testable hypotheses.
Developments in molecular genetics have further spurred this line of inquiry. DNA
fingerprinting techniques are now routinely employed to assign paternity, allowing
researchers to assess the connections between mating behavior and its genetic
consequences.
9.3.1 Dimorphism, Sexual Selection, and Mate Choice
Male and female insects sometimes differ so dramatically that they look like two
different species; in some aculeate wasps, in fact, matching has been possible only
when males and females have been discovered in copula. Other males are armored
(Fig. 9.10). Males of Hawaiian Drosophila, euglossine bees, Asian fireflies, and cer-
tain dragonflies repeatedly gather at specific sites where they display and compete
for the attention of females.
Why do males and females differ in their appearance? Why are males so often
the more elaborate sex? Why, for that matter, do individuals differ in appearance?
And why do members of the opposite sex care? Dimorphism may be driven by
Fig. 9.10 Wearing weaponry. An 1869 illustration by Émile Blanchard, a prominent early
French zoologist and entomologist, shows sexual dimorphism and the life cycle of a stag beetle
Lucanus cervus, a well known species across much of Europe. Male stag beetles use their stout
bodies and elaborate jaws to wrestle with each other for favored mating sites
food preferences, environmental factors, parasite load, or a host of other factors

with no direct relationship to reproductive behavior. In some cases, each sex sim-
ply may belong to a different mimicry complex. Still, sexual dimorphism relates
directly to courtship and mating in so many cases that such questions have inter-
ested biologists, psychologists, and philosophers from ancient times and are far from
settled today.
The questions certainly were central in the mind of Charles Darwin as he tried
to explain the evolution of what he and others called ‘secondary sexual characters’
(those not involving the reproductive organs), such as antlers on male deer. Natural
selection alone did not seem to explain what he saw, so Darwin posed an alternative
explanation—sexual selection. Competition for mates among the members of one
sex, he reasoned, was responsible for the evolution of those traits peculiar to that
sex, including all the various anatomical, physiological, and behavioral mechanisms
involved in mate selection. This would include both traits useful in combat between
members of the same sex and traits that increase the attractiveness of individuals to
members of the opposite sex.
Sexual selection actually has a number of subcomponents (Fig. 9.11), but two
major types of competition were apparent to Darwin. One depended upon choices
between the sexes (‘the power to charm,’ in Darwin’s words); this has been termed
epigamic (or intersexual) selection. The other involved interactions between males
(or less commonly between females); Darwin’s ‘power to conquer other males in
battle’ is now called intrasexual selection.
Epigamic sexual selection has been likened to a contest between salesmanship
and sales resistance. As expressed by Edward O. Wilson:
The sex that courts, ordinarily the male, plans to invest less reproductive effort in the
offspring. What it offers to the female is chiefly evidence that it is fully normal and physio-
logically fit. But this warranty consists of only a brief performance, so that strong selective
pressures exist for less fit individuals to present a false image. The courted sex, usually the
female, will therefore find it strongly advantageous to distinguish the really fit from the
pretended fit. Consequently, there will be a strong tendency for the courted sex to develop
coyness. That is, its responses will be hesitant and cautious in a way that evokes still more
displays and makes correct discrimination easier.
Fig. 9.11 Choosing a male. Types of mate choice open to females. (a) No choice, female accepts
any available conspecific mate. (b) Female accepts most valuable mate. (c) Female accepts any
victor of a competition among available mates. (d) Female chooses on basis of comparison against
some absolute standard. (e) Female selects an extreme phenotype, simultaneously rejecting other
equally available but less flamboyant males. (f) Female chooses among territories or position
in a lek
At first glance, Darwin’s sexual selection theory might seem a simple answer to
a simple set of questions, but its implications are far-reaching. The subtle ways in
which mate choices can influence male and female fitness are only beginning to be
more fully appreciated.
Explaining how female mating preferences have evolved has been relatively
straightforward in those situations in which a female’s mate choice clearly influ-
ences her immediate reproductive success (fecundity). For example, in mating
aggregations based on real estate, males generally (but not always) defend ter-
ritories, wait for arriving females, and perform courtship prior to mating. Many
produce long-range signals to attract females. In these resource-based mating sys-
tems, direct selection has always been considered to have an important influence on
mate preferences.
In other cases, the logic of female mate choice has been less apparent. Non-
resourced based leks are a prime example. A lek is a gathering of males for the
purposes of competitive mating display. In lek or lek-like mating systems, males’
only contribution to females is their gametes; they are not involved in parenting, and
if they have territories, they do not regulate a female’s access to any resources that
the territories may contain. Most matings occur in a mating arena, where females
freely select a mate. Females have a wide choice of males with which to mate, and
clearly prefer some to others. (Trivia-lovers note: A lek when females come together
to compete for males is called a mung; this role reversal is rare, but occurs in the
yellow-spotted millipede.)
Classical lek species were first identified in birds and mammals, and most lek
research still focuses upon vertebrates. At first, it seemed that species could be
strictly classified as to whether they had leks or not, but as more and more diver-
sity has been catalogued, it has become obvious that lek-like behaviors should be
considered as continuous variables that can differ independently with ecological
conditions, taxonomic affiliation, and other factors. This consideration is particu-
larly worth keeping in mind when confronted by the immense diversity in a group
as large as the Insecta. In addition, some researchers consider leks to include mating
swarms, whereas others see leks to include only substrate-based systems. Certainly,
swarms occur in many non-mating contexts, and even mating swarms differ from
substrate-based systems.
The so-called paradox of the lek has posed particular problems for evolutionists.
Because females seem to gain no material benefits or parental care from males, it
appears as though direct selection would have little relevance here. Furthermore,
one would expect that persistent female choice for particular male traits such as
elaborate ornamentation should erode genetic variation in them and thereby remove
the benefits of choice, Paradoxically, both variation and choice persist.
A number of competing theories have been proposed to settle the paradox. Some
emphasize male roles; others, female roles. The notion that females prefer clus-
tered males because clustering facilitates mate choice is not consistently supported
by field observations. In general, field and laboratory studies have failed to conclu-
sively support one theory over another. The paradox might be somewhat alleviated
by the occurrence of mutations introducing potential differences, as well as the pos-
sibility that traits of interest have more or less favorable recessive alleles. Moreover,
if leks are composed of closely related males and larger leks attract more females,
the paradox might be avoided because lower-ranking males could pass on their
genes indirectly by helping attract mates for their brothers. Heated debate arose over
two explanatory hypotheses: coevolution (and its famous extreme, runaway sexual
selection) and good genes.
The first of these hypotheses rests on the understanding that lack of direct selec-
tion on a character does not exclude that character from evolving; indirect selection
can occur on traits if they are correlated with traits under direct selection. In 1930,
R. A. Fisher presented a brief discussion of runaway sexual selection, which was
subsequently elaborated by others. The model postulated a sort of ‘self-reinforcing
choice’ in which an initial assortative mating generates a statistical association
between a trait and a preference that are both found in both sexes but only expressed
in the appropriate one. When this genetic correlation is high relative to the heri-
tability of the male ornamentation, as male traits evolve under the opposing forces
of natural and sexual selection a correlated evolution of the female preference
results. As a result, the trait and the preferences are exaggerated indefinitely as
they continue to evolve. If the genetic correlation is somewhat lower relative to
the heritability of the male ornamentation, the trait and preference will increase
only until viability selection against further trait elaboration balances sexual
selection.
Runaway sexual selection has been difficult to show empirically. Thus, it is
often relegated to consideration as a null hypothesis when other explanations can-
not be demonstrated. Evolutionary biologists note that the runaway process may be
so rapid as to be unlikely to be observed and it can be difficult or unfeasible to
conduct the experiments necessary to validate runaway. However, cases of genetic
correlation between traits and preferences are beginning to accumulate, as is evi-
dence showing that evolution of a trait can result in an evolutionary response in the
preference.
A particularly clear example has been shown in studies of stalk-eyed flies (see
Plate 31). When considered together, Wilkinson and Reillo’s results, outlined in
Case Study 9.2, provide very good support for a coevolutionary model of female
choice sexual selection. Female mate preference changed as a consequence of arti-
ficial selection on male relative eye span. Because females were caged separately
from males and selection was not exerted on females, the changes in female prefer-
ence appeared to result from a genetic correlation between male eye span and female
preference. However, Wilkinson and Reillo themselves admitted, their results did
not exclude another competing model for the evolution of male ornamentation—the
‘good genes’ hypothesis.
Case Study 9.2: Female Choices in the Stalk-Eyed Fly,

Cyrtodiopsis
Can female mate preference change as a consequence of selection on male
traits? Little flies with huge eye spans gather in mixed-sex groups each
evening, suggesting a way to test controversial mate choice theories.
Theories concerning the evolution of exaggerated male traits by female
mate choice have been highly controversial. One theory maintains that male
ornaments coevolve with female preferences. Another views male ornaments
as handicaps that signal increased viability. Still another holds that preexisting
sensory biases for conspecific recognition lead to trait exaggeration. Gerald S.
Wilkinson and Paul R. Reillo decided to test these models with a Malaysian
population of a little fly that has impressive male ornamentation–stalked eyes
separated by spans that often exceed a male’s body length (see Plate 31).
Several aspects of the flies’ life history suggested they might prove espe-
cially suitable for such study. Male eye spans are extremely variable, and
such variation has been shown to be heritable. Previous field observations on
Malaysian diopsids by other researchers suggested that eye span might influ-
ence mating success. During the day, the flies foraged alone, grazing on rotting
vegetative matter. However, each evening, they moved to streams and formed
aggregations on root hairs underneath overhanging stream banks. Each group
consists of a single male and several females, with males attempting to dis-
place one another from the root hairs as the aggregations formed. Males assess
the body size of competitors by their head projection, and males of similar size
engage in extended physical contests. Being the sole male in a roosting group
was important, because mating occurred at dusk or dawn in the aggregations.
A few years earlier, other researchers had published studies on a related
species showing that the male with the longest eye span usually won alterca-
tions between males, and that females preferred to alight on strings containing
model males with the longest eye span. This seemed to be a classic lek
situation—how, if at all, might a choice based on such a seemingly arbitrary
choice as eye span influence female survival or fecundity?
Wilkinson and Reillo approached the question quite systematically. In
January and October 1989, they counted male and female flies on root hairs
along a Malaysian stream bank. To quantify sexual selection, they collected
roosting flies at night, enclosing all flies on a root hair with a nylon stock-
ing. Flies were frozen in liquid nitrogen in the field and kept frozen until they
could be processed.
Analysis of their field data confirmed that females clustered, whereas
males dispersed beyond random expectation. But were the females sim-
ply returning to the same microhabitats night after night, irrespective
of the males that might be there? If so, the same sites should consis-
tently contain more females. Statistical analysis said no; female aggre-
gations did not appear to be determined by microhabitat choice alone.
In the field-collected aggregations the number of females per site varied
but the average number of females per male clearly increased with male
eye span.
With this background, the researchers decided to observe flies more
closely. They placed the insects in plastic mouse cages, fed them with ground
corn twice a week, and hung strings in their cages as substitute root hairs. The
flies lived well in captivity; groups of adults of both sexes could be sampled at
regular intervals up to 191 days old. Males could also be marked on the thorax
with typewriter correction fluid to identify individuals.
Would the number of females per male in a caged aggregation estimate
mating success? The researchers placed two males with different eye spans
and body lengths into each of 24 mouse cages containing an average of five
females. Then they counted and timed all the copulations for thirty minutes
after dawn on two successive days. Each morning within an aggregation, most
females mated repeatedly, but the larger male obtained 57% of the copula-
tions. In the field, they speculated, the skew in his favor would be even greater
because in cages females could not evade males and large males could not
exclude small males.
But were females really reacting to eye span itself, or to greater body size?
Wilkinson and Reillo scored frozen flies for eye span, body length, age, and if
female, fecundity. They measured eye span and body length with a video dig-
itizing system, and dissected each female to determine the number of mature
oocytes within her abdomen. Following procedures established by others, they
estimated fly ages from the amount of pteridine pigments in their eyes. Later,
this information would allow them to examine the body size/eye span relation-
ship and also female survival, fecundity, and other potential costs of female
choices.
What would happen if they could directly compare female reactions to
long and short eye-spanned males of the same body length? Measuring
the ratio of eye span to body length in their population, the scien-
tists selected ten males with the longest relative eye span and another
ten with the shortest relative eye span. They mated each with 25 ran-
domly chosen females. For thirteen generations, the scientists selected
for long and short male eye span distances in this way. At the end of
this time, they had two populations of males whose eye spans averaged
a difference of 1.57–1.47 mm, and whose body lengths averaged from
0.05 to 0.01 mm.
To quantify mate choice in the presence and absence of male interactions,
Wilkinson and Reillo decided to run four experiments. To make sure that
past association with males would not influence mate preferences, the females
from each selected line were caged together after eclosion without males until
tested two to three months later.
For the first set, groups of five females were placed into a cage with two
hanging strings and two selected line males. After two days of acclimation,
the number of females with each male before daylight was counted. In 24 of
31 cages, more unselected females roosted with wide-eyed males, suggesting
that male competition, female choice, or both occurred.
For the next three groups of experiments, male competition was excluded.
Clear partitions set up across the cages allowed the flies to see one another
but the partitions were perforated with holes that were just large enough
to allow females to move between the cage sides without allowing males
through. (Food was available on both sides of the partition.) For experiment
two, they used unselected females; as expected, the females preferred wide-
eyed males. For experiment three, they used wide-eyed females; these also
preferred wide-eyed male, but not to a statistically greater degree than in
experiment two. This was surprising; if genetic correlation exists between
male eye span and female preference, then selection for increased eye span
should produce more choosey females. Reviewing possible explanations, the
scientists decided it was quite possible that there may have been a difference
that went undetected because suitable test males with extra-long eye spans
were not provided. Supporting this idea were mate choice experiments with a
related species in which females preferred model males with longer eye span
than exist in natural populations. For experiment four, the researchers chose
females with short relative eye spans; these preferred not long- but close-eyed
males. In this way, they demonstrated a genetic correlation between female
preference and a sexually selected male trait.
Might it be more costly for females to select large eye span males than to
mate at random? If so, a negative correlation should exist between the relative
eye span of the largest male in an aggregation and either female age or fecun-
dity. In their laboratory analyses, Wilkinson and Reillo found instead that eye
span of the largest male in a field aggregation correlated positively with female
age, as estimated by amount of eye pigment. Thus, it appears that mate choice
did not appear to impair female survival, but actually improved female sur-
vival. However, because other explanations for this correlation are possible,
this conclusion must be approached with caution. One might also postulate
that a longer life could be expected to yield more offspring, but the data fell
short of demonstrating this advantage; the correlation between eye-span ratio
of the largest male and female egg count was not statistically significant.
Fisher’s theory begins at the point of rather arbitrary female preferences. As an

alternative, Amotz Zahavi proposed the idea that females prefer to mate with males
that have demonstrated their superior genetic quality for survivorship. He suggested
that females evaluate a male’s survival ability by assessing the magnitude of the
handicap with which he is able to survive. Because many extreme sexually dimor-
phic traits increase male mortality, these handicaps to survival evolve as honest
signals that allow females to assess male genetic quality. Thus, they serve as sig-
nals of ‘good genes.’ Several variations of Zahavi’s original handicap principle (or
good genes hypothesis) have been proposed. In one, because males without handi-
caps cannot be judged, females ignore them. In another, the handicap is condition
dependent; a male’s investment in the handicap varies with his condition such that
he is optimizing the tradeoff between mate attraction and survivorship to maximize
his reproductive success. The good genes hypothesis stirred up quite a bit of initial
controversy, but has come to gain much wider support since it has been shown that
female preferences can indeed evolve if they are genetically correlated to the ‘good
gene,’ which is taken to mean some heritable component for viability.
Evidence for the good genes hypothesis is lacking, but a number of studies pur-
port links between female choices, offspring viability, and paternal genetic effects
in a range of other animal species. Still it is important to recognize that costly dis-
plays by themselves cannot automatically be taken as evidence for female choice
for good genes. Because mate choice involves communication, it always involves
some sort of costs. The purpose of any signal is to increase the conspicuousness of
the signaler against background noise; it can easily become an unintended adver-
tisement to predators and parasites. There are physiological costs as well, either
in the growth of a structure used as a signal or in the behavior of displaying the
signal.
Increasingly, sexual selection studies are moving from antagonism between two
competing theories to consideration of a much broader range of choices. In some
cases, there is evidence that pre-existing sensory biases for conspecific recognition
lead to trait exaggeration. Direct selection on preferences and pleiotropic effects
can influence preference evolution. Furthermore, mate choice undoubtedly involves
female attention to a whole suite of sexually dimorphic traits that are under sex-
ual selection. Historical approaches have introduced a new dimension, as have the
modern phylogenetic methods.
Direct selection, runaway selection, selection for good genes, and such are often
portrayed as though they were mutually exclusive hypotheses, but in fact they may
represent something more akin to the ancient tale of the blind men’s encounter with
an elephant, with each describing a different aspect of the same reality. Thus, former
contentiousness between the holders of mutually exclusive hypotheses seems to be
yielding to a new emphasis on the relative strengths and influences of these different
forces when they act simultaneously.
9.3.2 Intrasexual Competition

Males of some damselflies scoop rivals’ sperm from the genital tract of females
they have mounted. Desert Centris bees gather around spots where females are
emerging from the ground and jostle with one another; larger males tend to be the
winners, with smaller males reduced to finding females that are missed in the first
go-around. Male Heliconius butterflies guard females that are still within their pupal
case, driving off rival males (Plate 32).
Intrasexual selection—which is almost always between males rather than
between females—is based on exclusion of courting rivals. We usually think of this
in terms of keeping other males from mating, often by intense fighting. Intrasexual
competition is not limited to behaviors before insemination, however. Animals have
developed a great number of ingenious post-copulatory devices as well. By far the
greatest diversity of these occur in the insects.
Unlike vertebrates, female insects can store viable sperm from a single insemina-
tion for use during the remainder of their reproductive lives (several years in the case
of some social species). At the junction where the female’s pouch-like spermatheca
opens to the genital tract, a sphinctor muscle parcels out sperm as eggs are laid.
Scientists theorize that in cases where sperm are deposited in packets and maintain
their cohesion during storage, they presumably would be used in the reverse order
from which they were obtained. This ‘last in, first out’ principle, called sperm prece-
dence, has become a widely accepted tenet of sexual selection theory, and there is
evidence for it in many insect species. Its importance lies in the fact that it sets the
stage for continued conflict because if the sperm of the last male to mate are the first
to be used in egg fertilization, a late-arriving male would still profit by attempting
to mate with an inseminated female. A second strategy is sperm removal. Odonata
provide some of the best examples. The male of a pair of copulating Calopteryx
damselflies (see Plate 33) spends most of his copulation time physically extracting
sperm of other males from the female’s spermatheca using his relatively elaborate
penis containing recurved spines, hooks, and a scraper to accomplish this task.
Obviously, it is not to a male’s advantage to have his own sperm subsequently
replaced. How can he ensure that his sperm will be the most likely to fertilize the
female’s eggs? One way is simply to monopolize the female so that other males do
not have access to her. Such a situation occurs in Scatophaga stercoraria dung flies.
Continuing to maintain a territory has a similar effect, as does merely staying in
physical contact with one’s partner without actual genital contact; the tandem flight
of many Odonata is a familiar example (see Fig. 9.2). So does prolonged copulation,
which occurs widely among insects such as the lovebug (see Fig. 9.1), walking
sticks, the queen butterfly, and the burnet moth (Fig. 9.12). Still another widespread
device is the mating plug, formed by coagulation of male accessory gland secretions
in the female’s genital tract; such plugs prevent sperm leakage and/or physically
prevent subsequent mating with other males.
Fig. 9.12 Prolonging copulation. The day flying burnet moth, Zygaena filipendulae, is
widespread in Europe, but the majority of zygaenids are tropical. The moths have red spots on
black, a warning to predators that they are distasteful; burnet moths contain hydrogen cyanide
throughout all stages of their life cycle
After members of the courting sex have aggressively excluded a portion of the
competition, classical intrasexual selection theory states that the more passive sex
simply chooses a potential mate from among the elite group of winners or the single
winner. In the process of acquiring a vigorous partner, he or she often also acquires
something else of great value—a share in his or her set of resources. How members
of the more actively courting sex partition these resources among themselves brings
us to the twin topics of territoriality and dominance.
9.3.3 Territoriality and Dominance
Male carpenter bees vigorously chase a variety of small objects tossed near them.
Male dragonflies drive off other conspecific males. Males of the ectoparasitoid
wasp Nasonia defend host fly puparia from which females are starting to emerge.
Mated females of the ectoparasitoid Melittobia megachilis aggressively fight other
females that they encounter upon their host, ultimately partitioning the host into
microterritories that are actively defended by amputating the legs and antennae
of rivals.
The phenomenon of territoriality—broadly, any space-associated intolerance
of others and, more narrowly, an intolerance based on real estate holdings—has
long been well known in vertebrates, often in association with aggressive behavior.
Among insects it has received less emphasis, but it is a surprisingly common and
widespread occurrence.
Territoriality is almost always associated with competition for mates or food. Its
function is to partition priority of access to resources that are limited and tied to a
particular area. Territoriality is not always associated with overt aggression; some
types of space exclusion can be settled solely through display. Nor does it even
always require the owner’s presence; some territorial insects such as female apple
maggot flies depend upon persistent chemical marks that have a repellent effect on
conspecifics.
Because territorial activities incur an expense of energy, selection for territorial
behaviors tends to operate only under a limited set of conditions. One condition
is that the desired but limited resource should be relatively localized in a readily
defensible situation. Another is that animals should stand to obtain more of the
resource by defending the area against competing individuals than by searching for
new resources to exploit. For example, males of the cicada killer wasp, Sphecious
speciosus, establish themselves on perches that overlook highly clumped nesting
areas containing emerging females; investment in defense of these perches may be
rewarded by extra copulation opportunities for the territory owner. Dragonflies and
damselflies establish similar perches along the shores of ponds and streams (see
Plate 33).
Territorial fights among vertebrate conspecifics are rarely fatal; when underlings
assume various submissive postures, the combatants slow or halt their battle. In
insects, in contrast, murder and even cannibalism may be routine phenomena. One
of the more striking examples of aggression occurs among males in Melittobia, a
Fig. 9.13 Fighting spirit.

The loser never let go in this
battle between males of
Melittobia digitata; the
winner, despite his ball and
chain and the loss of the distal
portion of his right antenna,
subsequently completed
successful courtships with
several females
genus of tiny parasitic wasps. Predisposing their behavior in this direction is an

unusual breeding structure—males are incapable of dispersal and thus their only
mating opportunities are with sisters emerging from the same host. Upward of
150 progeny from a single mother develop gregariously and emerge from a single
host, with an overwhelming preponderance (95%) being female. Inside the host’s
skin or pupal case, the few blind, flightless brothers aggressively compete among
themselves for the opportunity to copulate with their sisters, who will mate only
once. In mild forms, Melittobia aggression consists of intensive grappling bouts in
which contestants may lose an appendage; at its extreme, the first males to emerge
selectively decapitate their unemerged brothers (Fig. 9.13).
Particularly in species with long-lived adults, aggression may also be resolved
through the establishment of dominance, a ranking of individuals on the basis of
real or apparent authority, strength, or influence. The dominance concept implies an
ability to recognize individuals and remember relationships. Most examples occur
among social species, but male crickets confined to a limited space will establish
a rank order that is stable for some time (see Chapter 8). Age, size, possession of
a territory, opportunity to copulate, and the results of previous fights determine the
rank order, and the acoustic signal (rival song) of a dominant individual inhibits
the aggressive tendency of a subordinate one. If the rival male does not retreat, he
may respond with his own rival song. Until the hierarchy is established, aggression
includes not only singing but jumping at one another, biting, and lashing with
antennae.
9.3.4 Nuptial Gifts
A male of a non-predatory bug, Stilbocoris natalensis, collects a seed, injects it

with saliva that apparently partially predigests it, then offers it to a female; a male
without a seed may court, but he cannot persuade the female to copulate. In the
katydid species Isophya krausii, a male’s spermatophores include sufficient directly
absorbed nutritive material to meet a mate’s caloric needs for a full two days. Males
of the beetle Neopyrochroa flabellataduce ingest toxic cantharadin while feeding;
cantharadin transfer during mating has been shown to enable a female to use the
chemical to protect her eggs from predation.
As examples like this make apparent, during courtship the ‘salesmanship’ of
many male insects includes significant gift-giving, and their potential mates seem
quite open to such bribery. Nuptial gifts—male material donations transferred to a
conspecific female during mating—are delivered by two basic routes. One route
includes oral gifts that are donated by a male and ingested by the female dur-
ing mating (Fig. 9.14; see also Fig. 6.9 and Plate 29). The other type comprises
seminal gifts that are ejaculated into the female reproductive tract and absorbed
from there. Either way, the gifts may include nutrients and/or defensive chemi-
cals. These gifts are obtained from sources as diverse as leftover prey, carrion,
purposively collected foodstuffs, and secretions from male salivary or reproductive
glands.
The published literature holds many examples of nuptial gifts. When the imma-
tures of many insects feed, they accumulate various toxins, particularly alkaloids,
in their tissues that subsequently carry on to adulthood; in other cases, males gar-
ner the toxins as adults. In various species it has been shown that males that offer
toxin-bearing secretions either strengthen the female’s toxic protection to her eggs,
replace toxins she may have depleted during previous oviposition, or protect the
female herself.
The moth family Arctiidae seems particularly prone to host-derived chemical
defenses. The bella moth, Utetheisa ornatrix, is the most thoroughly studied in this
regard (Fig. 9.15), thanks to perceptive studies by Thomas and Maria Eisner and
Fig. 9.14 Courting with an oral gift. Nuptial feeding in a scorpionfly, Bittacus apicalis. Through
this habit, widespread in the Bittacidae, females gain increased reproductive success as measured
by increased oviposition
Fig. 9.15 Passing along alkaloids. The rattlebox moth, Utetheisa ornatrix, (left) feeds on the
toxic legume Crotalaria spectabilis (right) as a larva. Unaffected by the plant’s pyrrolizidine alka-
loids, the larva stores these toxic chemicals and carries them along into adulthood. Thus, females
are able to pass the alkaloids to their eggs; they also obtain an additional dose of the alkaloids from
males as a seminal gift during each copulation
their associates. This day-flying moth gains its alkaloid from its specialized larval
diet of rattlebox, Crotalaria, a legume that contains toxic pyrrolidizine alkaloids
(PA). The chemicals are stored systemically, retained into adulthood, and passed on
to eggs, thus providing all life stages of this moth with protection against spiders,
beetles and wasps. Moths of both sexes actively transmit PA to the next generation.
Females do it directly, in the process of laying hundreds of eggs. As a result, their
PA levels can drop over time. However, they can obtain more PA by mating, because
males gift them with a sizeable sperm package. Sometimes amounting to over 10%
of the male’s body mass, it contains both PA and nutrients.
Another example is the arctiid moth Cosmosoma myrodora (Case Study 9.3).
Males in this species produce fine cuticular filaments in abdominal pouches, and
discharge them in bursts during courtship (see Plate 34). If males feed upon plants
that produce PA, the filaments are laden with these alkaloids; a female laden with
these toxic filaments is protected against predators such as Nephila clavipes, a com-
mon spider, and she passes some of this protection on to her eggs. Males deprived
of alkaloids produce filaments that confer no such protection.
Case Study 9.3: Nuptial Gifting in the Arctiid Moth,

Cosmosoma
During courtship, a male of the scarlet-bodied wasp moth envelops its future
mate in a cloud of gauzy filaments that cling around her body. What is the
value of such a strange behavior?
Small Cosmosoma moths (see Plate 34) are not particularly abundant in
Florida, but their aposematic coloration and strange habits caught the atten-
tion of William E. Conner, Ruth Boada and their associates. For her master’s
thesis, Boada studied the courtship of these moths. Like many other moths, a
male C. myrodoa is attracted to a pheromone emitted by the female. However,
once he nears the female, he does not alight. Instead he flutters nearby, then
suddenly discharges bursts of fine filaments that swirl in the air around the
female like an explosively delivered net. These filaments cling all over her
body as the pair copulate, a lengthy process can take up to nine hours in this
species.
Unlike most adult Lepidoptera that are only known to feed on nectar,
Cosmosoma moths are attracted to juices from plants such as dogfennel
(Eupatorium capillifolium), a common roadside weed throughout the south-
ern and eastern United States; as its name suggests, the leaves and flowers of
this plant smell unpleasant when crushed. It also belongs to a plant group with
members known to produce strong pyrrolizidine alkaloids (PA). Watching the
moths closely, one might see them probing the plant’s surface with their pro-
boscis, feeding or attempting to feed from excrescent material. Because the
larvae of these moths feed upon climbing hempvine (Mikania scandens), a
relatively common and fast-growing southern weed that appears to have no
alkaloidal defenses, the researchers wondered if the males might be obtaining
PA as they probed about. If this were so, did the filaments they released during
courtship also contain pyrrolizidine alkaloids? Might the filaments offer any
protection against predators?
For answers, Conner would need many individuals of these relatively
rare moths. Happily, he was aware of a collecting technique that many
old-time naturalists had used to collect rare butterflies and moths. In
nature, a number of Lepidoptera visit flowers, stems, and leaves to obtain
alkaloid-rich juices, but roots of the plants contain the highest concen-
trations, and thus can be used as a lure. Collecting roots of E. capil-
lifolium, Conner hung them in moistened clusters about 1.5 m above
the ground. That night, Cosmosoma moths began arriving; all but one
was a male. Using mercury vapor light traps, he collected female moths
as well.
In the laboratory, the researchers raised the offspring of these field-
collected individuals on their natural larval food plant, climbing hempvine.
While they grew, he ran chemical analyses of both M. scandens and E.
capillifolium, and confirmed that the larvae’s natural food was PA-free, but
the Eupatorium roots were loaded with alkaloids, mostly intermedine and
lycopsamine but also four other PA, two of which apparently have not been
described from elsewhere in nature.
When adult males eclosed in his laboratory cultures, Conner presented one
batch with a crystalline offering of PA mixed with sugar, upon which the
males fed readily for an hour each day over the next five days. Then he killed
some, extracted flocculent out of their abdominal pouches with forceps, and
ran chemical analyses of their bodies and floss. Both contained PA, but it was
about 20 times more concentrated in the flocculent than in the moth’s body.
What protection, if any, did males gain from being PA-laden? As a test
predator, Conner decided to use Nephila clavipes, a common spider that coex-
isted with Cosmosoma, particularly along the edges of lakes and streams. The
researcher flipped a half dozen PA-fed males individually into the spider’s
web. Then he did the same with males that had no access to alkaloids. As
each became entangled, he watched the spider to see whether it ate the moth
or cut it free from the web. Every one of the PA-fed males were cut free, live
and uninjured. In contrast, five of seven PA-free males were eaten.
Would flocculent alone protect females that received it, or were PA neces-
sary as well? Conner placed conservative amounts of flocculent (about half
that present in a male’s pouches) in plastic containers with young virgin
females. As the confined females fluttered about, they stirred up flocculent
and coated themselves with it. Some females received flocculent from a male
that fed on the alkaloid/sugar mix; others received flocculent from a male
that had no access to alkaloids. Then Conner flipped each one individually
into the web of a N. clavipes spider. As they became entangled, he watched
the spider once again. All eight females coated with PA-free flocculent were
eaten. In contrast, the spider rejected and cut free over half the females whose
flocculent came from PA-fed males.
Wondering whether females might also gain PA during the lengthy copu-
lation process, Conner confined various combinations of males and females
overnight in cages; copulation typically began at dusk and lasted most of
the night. When he analyzed their bodies afterward, he found that all of the
females that mated with PA-fed males now contained alkaloids as well, but in
amounts that differed depending on whether they had laid eggs. This seemed
to indicate that the females not only received PA, but passed some of it onto
their eggs. Chemical analysis confirmed that the eggs did, in fact, contain PA.
Was enough PA passed to a female and her eggs in this way to be protec-
tive? Strikingly so. While five of seven virgin females and five of six females
mated with alkaloid-free males were eaten, all seven virgin females mated to
alkaloid-fed males were cut from the web by the spiders, without sustaining
injury. Furthermore, the amount of alkaloid in these protected females’ eggs
(on the order of 1 μg per egg) was shown in a study with another arctiid moth
(Utetheisa ornatrix) to be effective in egg defense against ants, coccinellid
beetles, and chrysopid larvae.
Oral gifts are particularly evident in different predatory taxa, where they appear
to have evolved independently. Thus, they were initially viewed wholly as a mech-
anism for appeasement, that is, inhibition of the normal predatory instincts of the
participants, especially the female. However, other explanations for this behavior
have also been put forth. One is that males with prey are more conspicuous to
females. Females probably also enjoy a selective advantage by consuming prey at a
point where they need extra protein and calories to convert into eggs; male fitness is
also increased by providing such prey because the success of his genetic contribu-
tion to the next generation is intimately tied to the survival and reproductive success
of each of his mates. In addition, the very act of feeding upon a gift tends to keep a
female occupied, making it likely that a male would benefit from longer copulation
that results in transferring more sperm.
Sexual cannibalism would seem like the ultimate oral gift, but this has been a
matter of some debate. Mantises generally begin feeding upon a prey insect by biting
off its head, and females sometimes do the same with males (Fig. 9.16), at least
under laboratory conditions. If mating has begun, the male’s copulation motions
may become even more vigorous.
Early researchers noted that ganglia in the abdomen, not the head, control
the male’s movements, and proposed that sexual cannibalism was a strategy that
females had developed to enhance fertilization while simultaneously obtaining
sustenance. Others have proposed that submissive males should be evolutionarily
Fig. 9.16 Living dangerously. With good reason, mantis courtship has sometimes been described
as a ‘sneak attack’. In this pair of copulating mantises, photographed in the field in Costa Rica, the
female has eaten the head, prothorax, and prothoracic legs of the male while their genitalia remain
coupled. Scientists are still debating whether males passively submit to such sexual cannibalism
because a well-fed female would optimize the number and condition of their jointly-produced
offspring, or whether males actively try to stay alive to mate again
favored because they achieve a selective advantage in their offspring when a female
copulates longer and is better fed; this theory is supported by the observation that
both copulation duration and chances of fertilization are as much as doubled when a
male is cannibalized. However, still other research with well-fed Tenodera sinensis
mantises indicates that the behavior of these extremely visually sensitive insects is
very open to human disturbance; when the mantises were given the opportunity to
copulate in a vacant room, with only a running video recorder to witness the event,
instead of sexual cannibalism they displayed elaborate courtship behaviors that
changed a female’s interest from feeding to mating. Other studies similarly suggest
that males mate repeatedly and do not merely submit to cannibalism; rather, they
appear able to assess a female’s hunger level and adjust their behavior accordingly.
Seminal gifts often appear to provide similar caloric or nutrient benefits to oral
gifts but simply delivered by a different system. However, evolutionary biologists
have been intrigued by an additional possibility. If mating is as much conflict as
cooperation, gifts that are directly produced by the male might contain manipula-
tive substances that enhance his potential at the expense of hers. For example, a
seminal gift might contain a hormone that induced female refractoriness (resistance
to further mating), thus reducing the chances that a later rival would overtop his
own sperm with theirs; in some cases, this has been demonstrated. However, it is
important to remember that there is no compelling reason to believe that all nuptial
gifts must function in the same way, nor is it necessary that the costs and benefits
to each sex play out in the same way for every species. Clearly, in many species the
females gain significant benefits from their gifts, and prefer males that bear larger
gifts. Likewise, in cases where gift-bearing males directly increase the number and
fitness of offspring as a result, they are making a contribution similar to those in
species that practice parental care (see Chapter 10).
To date, the concept of nuptial gifts as manipulation has been supported more by
theory than by empirical evidence. However, it has given rise to a number of predic-
tions that are potentially testable in various gift-giving species. The best known of
all gift-givers undoubtedly are the dance flies. This group of small Diptera received
its name from the flies’ habit of circling about in complex patterns as they search for
prey on the bark of trees. Dance flies present an enormous diversity in nuptial gifts
ranging from clearly useful items such as bits of prey to more puzzling tokens such
as elaborate silk balloons to inedible twig fragments. Why should females require
what appears to be a worthless gift as a condition of mating?
In many textbooks, dance flies are used as a model system to support an evolu-
tionary progression toward ritualization in courtship; a proposed sequence outlines
a decline in male investment in the mating effort from prey to a silk-wrapped dried
insect fragment to an empty silk balloon. Sexual conflict theory suggests a second,
simpler model in which token-giving arose in species that once gave nutritional
gifts, and males exploited female preferences by substituting more easily obtainable
but worthless items. Both models rely on male cheating and female inability to dis-
tinguish between worthwhile and worthless gifts, suggesting an evolutionary race
analogous to those that occur with other systems such as insect-plant coevolution
and predator-prey relationships.
Fig. 9.17 Cheating pays. Dance flies with oral gifts. (A) Rhamphomyia longicauda, male above,
mating with female while she feeds upon the gift of a small tipulid crane fly (B) Females of
Rhamphomyia sulcata usually accept a nutritious gift, but researchers substituted a cotton ball,
males presenting the token were rewarded with a copulation equivalent in duration to what they
would have received bearing a small genuine gift
An interesting experimental study questioned how easily cheating could invade a

species in which males normally provide valuable nuptial gifts. Males of the dance
fly Rhamphomyia normally provide an actual prey item, and dance fly pairs mate
on vegetation while the female eats the gift (Fig. 9.17a). Natasha LeBas and Leon
Hockman removed the gifts from mating pairs, replacing them with either a large
or small genuine gift or a large or small worthless cotton ball; then they timed the
copulation that followed. In this species, larger males usually carry larger gifts, so
by using this design, the researchers uncoupled the influence of the nuptial gift from
the influence of the male that produced, caught, or collected it. Their data showed
that females treated cotton ball tokens as equivalent to small prey gifts (Fig. 9.17b).
The researchers suggest that sexual selection in this species is acting to decrease
gift size because males are more maneuverable and flight-efficient in the lek with a
small gift; thus large and highly visible gifts, even if worthless, allow males to pro-
vide a large visual target to females while minimizing carrying costs. Presumably,
even though larger gifts increased duration of copulation, these shorter copulations,
whether natural or token-induced, are still long enough to result in adequate sperm
transfer. Selection would be expected to act strongly against males that fed females
for no reproductive return.
Three observations suggested that males were exploiting female sensory biases
with their cheating behavior. First, LeBas and Hockman observed that males given
cotton balls sometimes flew off with them after copulation, then returned to use
them again to pair up with new females. Second, the female reactions were similar
to responses to a supernormal stimulus. Cotton fibers are whiter than, but otherwise
quite similar to, the fibers of seed tufts that some dance fly males carry, and not too
different from the silk balloons and swollen foreleg projections that mimic a carried
gift in other species. Third, during copulation the females tried to feed upon the
cotton ball as though it were a genuine prey. (Similar behavior has been observed
in other species where males regularly present inedible tokens, suggesting that in
those species as well, males previously gave nutritious gifts.) However, copulation
duration was not significantly different for large versus small tokens, suggesting that
females take a set time to determine a gift’s edibility.
9.4 Mating Systems and Parental Investment
During her nuptial flight, a honey bee queen mates with several males in succession,
each of which dies after copulation. Sand fly males defend small arenas where they
emit pheromones; females choose freely, and a single male sometimes obtains all
the matings. A female of the hippoboscid fly Lynchia nurtures her single egg inside
her body until it has passed entirely through the larval stage and is about to pupate;
if a male encounters a female in this condition, he will dig his hook-like tarsal claws
into her back and ride about like this for hours before finally copulating with her
after she deposits her prepuparium.
In sociobiology and behavioral ecology, a mating system is any of the ways in
which animal societies are structured in relation to sexual behavior. These mat-
ing relationships also may or may not be associated with social relationships, in
which the sexual partners stay together to become parenting partners. Thus, there
has been some debate as to whether mating systems should be classified in a way
that reflects pair bonds and the rearing of offspring. The most straightforward clas-
sification simply considers the number of copulatory partners. The mating systems
most commonly recognized in animals are monogamy, polygamy, and promiscuity.
In monogamy, a male and female mate only with each other either during a particu-
lar breeding cycle or throughout their lives. Often called ‘pair bonding,’ monogamy
is rare among insects. Where it occurs, it is generally an evolutionarily derived
condition. Fidelity tends to evolve only under certain ecological circumstances, usu-
ally when the advantages of cooperation in rearing offspring outweigh the personal
advantages to both partners of seeking extra mates.
Polygamy is the state in which an individual has two or more mates, none of
which mates with other individuals. It has two forms. One is polygyny, where a
single male mates with several females; this is the most common system in insects.
The reverse case, polyandry, which occurs when a single female mates with several
males, as happens in honey bees, is relatively rare in insects, as in animals generally.
Thus, the terms polygamy and polygyny are sometimes used synonymously.
Promiscuity also takes two forms. In simple promiscuity, both males and females
mate repeatedly, but no pair bonds are formed. In polygynandry, two or more
males have an exclusive relationship with two or more females; this has been
best described among bird species in which males jointly help rear and/or protect
nestlings. The numbers of males and females need not be equal in a polygynandrous
system; in vertebrate species studied so far, the number of males is usually less than
9.4 Mating Systems and Parental Investment 381
the number of females, suggesting that polygynandry could be considered a variant

of polygyny.
It is interesting to note that virtually all the terms used to describe animal mating
systems were taken over from social anthropology, and were originally devised to
describe systems of human marriage. Any analogies should not be pressed too far; in
human societies, marriages typically have to be recognized by the entire social group
in some way, and there is no equivalent process in animal societies. Furthermore,
although mating systems can provide a convenient way to characterize a species,
variation can occur at the individual level. For example, DNA fingerprinting studies
in birds and mammals have shown that even in pair-bonding monogamy, mating
outside the pair (extra-pair copulations) occurs with fair frequency, and a significant
minority of offspring result from them. In addition, some species show different
mating systems in different circumstances, as for example in different parts of their
geographical range, or under different conditions of food availability, and, mixtures
of the categories may sometimes occur.
As in so many other areas of behavioral study, research on mating systems has
steadily been shifting from description toward prediction. After the question ‘what
sort of mating system does this insect have?’ has been answered, the second question
is ‘why?’ On a broader scale, how and why have particular mating systems evolved?
What factors might explain the variety that is observed? Do they have predictive
power?
One set of potential answers centers upon parental investment—contributions
of time and/or energy that increase the probability of some offspring surviving to
reproduce but occur at the cost of the parent’s ability to generate additional off-
spring. Different strategies of parental investment exist. For example, one species
may invest highly in a few offspring, while another may produce many but invest
very little in any one of its young.
Within a given species, the sexes generally differ in their degree of parental
investment (Fig. 9.18), beginning with the fact that males and females make unequal
physiological investments in the production of gametes. The clutch of eggs a female
lays may more than equal her total body weight, whereas even over an entire season,
sperm production usually represents a mere fraction of this. Moreover, the commit-
ment of time and energy required in such aspects as egg maturation, oviposition, or
birth are borne solely by the female, but in most insects a male’s only reproductive
role is to provide a set of genes.
Because of this general imbalance, a reproductive mistake has much greater con-
sequences for a female. ‘Coyness’ during courtship is one strategic result. Another
is that in almost every species, the female chooses a mate, not the reverse. Females
are under selective pressure to pick a male with a superior set of genetic instruc-
tions; and to the extent that such discrimination is possible, the outcome tends to
be that a select group of males with superior traits mates with a large proportion
of the female population. It should be to a female’s advantage to mate with a
male that signals genetic quality, no matter how many times he may have already
mated. Therefore, one should expect insect species, like most other animals, to be
fundamentally polygamous (or more properly, polygynous).
Fig. 9.18 Differing viewpoints. The central principle of sexual selection reformulated in terms
of parental investment. In the common situation illustrated here, the optimum number of offspring
(Of for the female, Om for the male) differs for mothers and fathers. Because the female must
expend a greater effort to create offspring, her greatest net production comes at a lower number
than in the case of the male. This sets the stage for a conflict between the sexes
9.5 Oviposition Behavior

A female locust or grasshopper begins tapping the soil with the tip of her abdomen
and probing with her ovipositor. Rejecting surfaces where the ground is hard, she
comes to a soft and sandy place where she begins to dig. Raising her body on her first
two pairs of legs, she arches the tip of her abdomen downward and scrapes particles
of soil sideways and upward with valves at its tip. As a hole forms and deepens,
her abdomen lengthens considerably. (If she is a migratory locust, Schistocerca, she
may dig to a depth of 14 cm.) To maintain necessary body pressure, she expands her
airsacs and swallows air into her crop and midgut caecae. Once the hole is dug, if
the soil is saline or dry, she rejects it. If the soil is moist, however, she begins to lay
groups of eggs (Fig. 9.19). To maintain hemolymph pressure, she continues to pump
air into her tracheal system by vigorous ventilatory movements of her head. About
20 minutes later, her airsacs occupy the whole cavity of the first five abdominal
segments, and the volume of her tracheal system has more than doubled from the
start of oviposition. Finally, when all her eggs have been laid, she forms a frothy
plug in the mouth of the hole, then withdraws her abdomen completely and scrapes
dirt over the top of the hole with her hind tibiae. All together, the process has taken
her about 2 hours.
Insects deposit their eggs in an almost infinite number of ways. Sometimes both
parents participate, but more commonly gravid females have sole responsibility for
the fate of their eggs. As the above example illustrates, the oviposition process can
9.5 Oviposition Behavior 383
Fig. 9.19 Ovipositing underground. The migratory locust, Schistocerca, (see Plate 4) digs a
shaft by opening and closing her dorsal and ventral ovipositor valves like a clam-shell posthole
digger, then lays a pod of eggs. After deposition, the eggs absorb water via a special pore in one
end and swell to twice their initial size. Successful breeding depends on rainfall patterns, and final
choice of site is positively influenced by the presence of other ovipositing locusts
be longer and more difficult than is generally assumed. Laying eggs is a signif-
icant expenditure of body energy reserves, as well. Because oviposition mistakes
may severely lower reproductive success and hence individual fitness, the stimulus–
response sequences tend to be quite complex, minimizing chances for error. In
general, the behavioral chain of events leading to oviposition closely parallels that
used in food location (see Chapter 4).
9.5.1 Selecting a Site or Host
Adult female mosquitoes are attracted to water, influenced by the presence of veg-
etation and amount of light reflected from the surface, but do not automatically
oviposit once they have reached it; when they land upon the water’s surface, some
first judge salt content and others assess the pH of the water through their tarsal
sensilla. Buprestid beetles of the genus Melanophila are attracted to burnt trees over
distances of several kilometers through an oriented response to infrared radiation
emanating from forest fires (see Fig. 3.7).
Site selection generally has two or more phases that vary in their sensory involve-
ment. First, there is a general reaction to the environment; this is followed by much
more specific final selection stages. For example, site selection by the tobacco horn-
worm moth includes both approach and landing. Approach is largely mediated by
Fig. 9.20 Tasting by foot-touch. Butterflies in the genus Pieris sense mustard oils in cruciferous
plants with their tarsi. Thus, in the laboratory they can be induced to lay their eggs on other surfaces
as long as their feet contact the proper foliar cue
visual cues and is not particularly discriminating. The decision to land, on the other
hand, involves subtle olfactory responses to smells emanating from host plants.
After landing, contact chemical stimulation elicits egg deposition. Similarly, female
cabbage butterflies are attracted to blue or yellow when searching for nectar, but not
when ready to lay eggs; an oviposition-ready female lands upon a plant and drums
upon it with her forelegs, sensing through her tarsal receptors whether the plant
contains mustard oil before she hangs upside down and oviposits her eggs on the
underside of its leaves (Fig. 9.20).
Among species that do not practice any form of parental care following egg
deposition—a situation that describes the vast majority of insects—proper egg
placement is particularly crucial. Considerable time and energy may be spent on
fine discriminations during site selection. Some stimuli that influence the behavior
of egg-laden female insects are negative, course. For example, certain plant-derived
chemical factors inhibit boll weevil oviposition (in one cotton species, by 40%). In
other cases, the stimuli for oviposition may be only indirectly related to actual larval
feeding preferences, a situation that may have unusual consequences (Fig. 9.21).
Many species have evolved a fine-tuned ability to assess host suitability fac-
tors such as the egg or larval load already upon a host. Female bean weevils,
Callosobruchus maculatus, do not deposit their eggs in mung beans at random but
compare the present bean with those previously encountered; a bean that is larger or
bears fewer eggs than the last one they came across is more likely to receive an egg.
The resultant nonrandom pattern of oviposition increases larval survival by 70%
Fig. 9.21 Differing cues. Although the larvae have a marked feeding preference for dandelion,
adult females of the noctuid moth Autographa precationis clearly prefer soybeans over dandelions
for oviposition, apparently because the shape of the dandelion leaves is a less effective oviposition
stimulus
over what would occur if eggs were laid randomly. Likewise, some species of hover
flies (Syrphidae) whose larvae feed on aphids are sensitive to aphid density and lay
their eggs in numbers proportional to the abundance of the aphid population. In a
similar vein, females of the parasitic wasp Tiphia popilliavora control the sex of
eggs they lay according to the size of their beetle host, Popillia; in third instar larval
hosts, the parasites lay fertilized female eggs, but in smaller second instar hosts they
lay smaller unfertilized male eggs.
At the same time, there is strong selective pressure for discrimination of poten-
tial danger from parasites and/or predators. Among cannibalistic species such as
preying mantises, even one’s own young may be a threat to siblings. For exam-
ple, the caterpillars of female heliconiine butterflies in the Neotropics will eat eggs
and one another. Ovipositing females often spend considerable time inspecting the
host plant prior to oviposition; it has been suggested that this is a visual search
for other Heliconius eggs or larvae. Probably the search is for other egg preda-
tors and parasites as well; one study estimated that over 90% of these butterflies’
eggs were killed by parasites. Such pressures may have led to uniformity in egg
dispersion becoming a widespread trait among many parasitic and phytophagous

insects.
Some parasitic wasps have eggs that keep on dividing, resulting in as may as
1,500–2,500 young from a single egg. Some termite queens can lay 6,000–7,000
eggs a day and the queen can live more than 15 years; one colony of the South
American Nasutitermes surinamensis had three million termites. Upon a small host,
a Melittobia parasitoid wasp may lay 150 eggs during her one-month life span; on a
larger one, up to 800.
At first glance, the reproductive capacity of a given species would seem to be
easily determined. In truth, however, records of egg laying or total egg output rarely
give a true picture of the reproductive potential of any species. A variety of fac-
tors, both internal and external, influence total egg production. Nutrition is probably
the single most important factor. Many insect species, particularly those having a
short adult life span, can lay eggs without having ingested any proteinaceous food.
The short-lived mayflies (Ephemeroptera) are a prime example of this phenomenon,
which is termed autogeny. Some autogenous insects never eat at all as adults; oth-
ers ingest only water or carbohydrates such as nectar. Most autogenous species,
however, lay relatively few eggs under such conditions; given the opportunity to
ingest proteinaceous foods, they will produce additional egg batches (which brings
us back to the potential value of nuptial gifts, see Section 9.3.4). In at least some
cases there appears to be a correlation between autogeny and larval nutrition, those
individuals with adequate nutritional reserves as larvae being more likely to be auto-
genous. In many other cases autogeny also appears as a consequence of adverse
climatic conditions. For example, autogeny in mosquitoes is common in temperate
and cold regions, where hosts and adult food sources are sometimes scarce. High
arctic mosquitoes are voracious biters, but there are few mammals available upon
which to feed. Confined to feeding on nectar, they can still lay a few eggs to assure
propagation of the species until an appropriate host appears (also see Fig. 4.1).
Environmental factors such as light, temperature, and humidity wield an influ-
ence not only by their direct effects but also indirectly through effects upon
feeding and mating activities. Day length, for example, has such a dual effect.
Reproductive diapause (see Chapter 3) is common, as is hibernation, which may be
thought of as a type of reproductive winter diapause. Both appear to be induced by
photoperiod.
Finally, any attempt to ascertain an insect’s reproductive capacity must take
note of the fact that mature eggs may simply not be deposited. For example, some
unmated females may mature a full complement of eggs but retain some or all of
them until after mating. In other cases, where the proper substrate or host is not
available, even fertile eggs may be retained and later reabsorbed.
In summary, it is doubtful whether data from laboratory colonies can ever real-
istically be equated with egg-laying in natural environments. Insects reared in the
laboratory generally live under better conditions, especially with respect to temper-
ature and the availability of food, but natural ecological relationships are difficult to
sustain under artificial conditions.
9.5.2 Reproductive Rates and Energy Allocation

Certain earwig and dipluran females brood their egg clutches prior to hatching (see
Plate 43). Females of the weevil Byctiscus populi construct rolls from young poplar
leaves, within which eggs are deposited and larvae develop (Fig. 9.22). Walking
sticks continuously drop single eggs apparently at random as they move through the
forest foliage. Parasitic wasps of the family Trigonalidae leave thousands of eggs
on foliage likely to be consumed by the caterpillars that serve as intermediate hosts.
What adaptive bases might there be for such diverse patterns?
To answer this question, we must consider an insect’s reproductive effort, a term
that includes not only the caloric content of eggs and sperm but the whole variety
of phenomena involved in the production of reproductively successful offspring:
the energies expended in seeking mates, searching for appropriate oviposition sites,
building nests, guarding eggs or young, feeding young, etc., and the risks result-
ing from the performance of these behaviors. For any given species, the question
becomes this: How much effort and/or exposure to danger will be optimal for max-
imizing the number of viable offspring, and how should this be apportioned among
those offspring that are produced? Phylogeny offers little help in answering this,
because birth rates often vary enormously from one species to another, even between
Fig. 9.22 Rolling one’s own. Construction of a leaf roll from a single aspen leaf by a female
weevil, Byctiscus populi. After notching the leaf petiole (arrow), the beetle walks around on part
of the leaf blade (stippled area), perforating the epidermis with its tarsal claws and sometimes its
jaws. Then Byctiscus positions itself parallel to the leaf edge, sinks in its claws, and draws its legs
up toward each other, using the snout to help guide the developing roll. Creeping inside the roll,
the weevil chews a longitudinal slit through one layer before reversing position to deposit a single
egg. Only then does she complete the roll and glue the last edge in place with anal secretions
closely related species. Instead, one must turn to ecological factors and consider
theories with an applicability that ranges beyond insects.
In an ideal environment, one with no predators and no intraspecific competition
for resources, a population would increase at its maximum rate, r, a value simply
obtained by subtracting the population’s death rate due to old age from its birth
rate. The whole reproductive thrust of every species is directed toward maintaining
r as close to maximum as possible. In the non-ideal real world, two very differ-
ent strategies for doing this are possible, depending on the amount of competition
and rigor of the environment. One may increase birth rate or decrease death rate.
Uncrowded and nearly ideal environments tend to favor increases in birth rate, rais-
ing r in the process. For this reason, selection that acts to raise the maximum rate of
population increase has been called ‘r-selection’ (more properly, ‘r max selection’).
Species that are r-selected are opportunists, able to quickly discover new habitats.
However, they are rarely able to persist successfully for long in such habitats, since
they tend to be poor competitors. They often have what may be termed ‘big bang’
reproduction, good colonizing ability, and short adult life spans. They can be termed
‘r-strategists’.
Other species live in approximate equilibrium with each other; their densities do
not fluctuate much. Their population levels persist at just about carrying capacity, K,
the number of individuals that the environment can support. There are no uncrowded
habitats to find and exploit; however, being able to survive where one is, particularly
for a long period, takes on new importance. There is little advantage to producing
large numbers of young, but behaviors that increase the survival of one’s young
become important. These species can be termed ‘K-strategists’.
It must be appreciated that r and K are theoretical constructs that represent the
endpoints of a continuum, and that a given organism may display some traits typical
of one strategy and other traits more representative of the other. Related to all of this
is the matter of energy allocation. Any organism has only a limited lifetime with a
limited amount of time and energy to partition among the various activities of its
life. Each animal must divide its time and energy among three major requirements,
which are (in their usual order of descending importance) food, defense against
predators, and reproduction. For each species this total time and energy will be
divided in a different manner depending upon environmental and evolutionary con-
straints. However, just as in any other budget, to the extent that one priority is easily
satisfied, more expenditure can be devoted to activities of the other priorities.
Chapter 10
Parental Behaviors and Social Life
10.1 Introduction
Webspinners (Embioptera) of all life stages live gregariously in a series of chambers

constructed from silk spun from glands in the forelegs of adults. Male and female
Geotrupes dung beetles dig a burrow in the soil together and store dung for their
larva, which is tended for most of its life by the female. Certain earwig females
(Dermaptera) lay their eggs in burrows and guard them until they hatch.
Since almost all insects are selective to one degree or another in their choice
of oviposition sites, one might say that the most fundamental rudiments of parental
care are nearly universal, namely, the laying of eggs in the correct place at the correct
time for proper development of young. Likewise, one might view internal incubation
and the birth of live young as elementary forms of parental care. However, after a
parent insect has fulfilled its responsibility for placing the egg (or larva in some
instances) in an appropriate environmental situation, most insects simply leave it.
Many insects go far beyond these basics, however, and tailor the environment to
the needs of their young through various behaviors collectively called parental care.
In most cases, a group of young are born and reared as a batch, or brood. Thus,
when behavioral involvement of the adults in feeding and protection of a batch of
young extends past the time of oviposition or birth, it has been called brood care.
‘Sociality’ is a more complex concept, and various scientists use the term itself
in different ways. Thus we will defer discussion of a more precise definition until
later (Section 10.3), here noting only that among the most highly developed social
insects, the brood has become the core around which all activity is centered, and
adults rear the brood all the way from egg to adulthood.
Parental care and social life in insects have been subjects of fascination and
speculation for millennia. Philosophers as diverse as Aristotle and Aesop found
inspiration and instruction for humankind reflected in these remarkable little soci-
eties that seemed so like ours (see Fig. 1.11). Natural histories of the most obvious
and extreme examples, particularly honey bees and ants, seemed only to underscore
the moral lessons to be learned. During the scientific and technological revolution
that coincided with the later part of the Middle Ages and through the Renaissance,
scientific ideas in physics, astronomy, and biology evolved rapidly. New organiza-
tions and institutions were established for the study of the natural world. Knowledge

390 10 Parental Behaviors and Social Life
of proximate causes of insect behaviors and information about the mechanisms

underlying certain behaviors advanced with surprising speed. However, ultimate
causes seldom entered people’s minds; insects were what they were, and always
would be.
Then, just under two hundred years ago, Charles Darwin placed biology, and with
it, insect parental care and sociality, in an entirely new context: that of evolutionary
history, adaptations, sexual selection and a host of other ultimate-causation ques-
tions. In this expanded context, insect parental care and social behavior has become
a major area of inquiry, not only in entomology but also in the biological sciences
as a whole.
10.2 Social Organization
Bumblebees are attracted to those flowers where other bumblebees are already for-
aging. Given a choice, male swallowtail butterflies prefer to land at artificial puddles
with a butterfly decoy. Female Mediterranean fruit flies reared in groups visit and
oviposit in more host fruits than solitary raised females do.
Most insects come together at least temporarily for mating; many also come
together for other reasons, forming temporary to relatively permanent groups. Some
of these are based on relatedness such as parent-offspring groups; others are not.
Some groups undoubtedly result simply from a common response to some partic-
ular environmental factor, such as presumably occurs when aggregations of bush
flies cluster on any available space (Fig. 10.1). At other times, mutual attraction is
clearly involved; some of the communicative devices involved have already been
mentioned in Chapters 6, 7, and 8 .
Might these be called ‘social responses’? The answer depends upon whom you
ask. Biologists use the term ‘social’ in a wide variety of ways. Some might speak
of the social relations between sexes during courtship and mating; others speak of
social interactions among animals in a herd, birds in a flock, or fish in a school.
Ecologists sometimes refer to all of the organisms in a habitat or community as a
society. However, a single thread is common to all these varied uses—the indication
of adaptively significant (often cooperative, but sometimes exploitative) interactions
between two or more individuals. Such interactions, mediated by reciprocal commu-
nication, produce effects that are qualitatively different from the mere summation
of the independent activities of the individuals.
Rather than getting mired in a bog of definitions, let us recognize that we are
dealing with a continuum of relationships. Attempts to classify the various kinds of
animal associations and interactions have been based on their form, basis for associ-
ation, degree of interactions of members, and the nature and duration of association.
One useful simple system is that represented in Table 10.1.
10.2.1 Aggregations and Simple Groups
By flashlight, one can sometimes find clusters of bees or wasps assembled together
to pass the night (Fig. 10.2). Up to 200 or 300 individuals may be involved, usually,
10.2 Social Organization 391
Fig. 10.1 Feeling crawly?. An aggregation of Australian bush flies, Musca vestustissima, clus-
tered on a man’s back. The harmless flies thrive in the vast semi-desert Outback, constituting a
continual annoyance by their sheer numbers and tenacity
but not always, of one species. Often they will return night after night to the same
location, but they do not cohabit a nest or rear young together, and the basis of their
individual attraction to the sleeping roost is unknown.
When an assemblage is composed of conspecific individuals including more than
just a mated pair or family, all gathering temporarily in the same place but not inter-
nally organized or engaged in cooperative behavior, it is termed an aggregation.
Many kinds of these ‘uncoordinated’ groups are common among various insects
(see Plate 5).
The causal factors behind aggregations are often unknown. Some result only
from limitations of suitable habitat. In other cases, aggregations grade into slightly
Table 10.1 A simple classification of insect associations, excluding those involving sexual
activitiesa
Interaction
among Nature of Selected
Association Basis members Duration association examples
Aggregations Facultative Uncoordinated Temporary Open; Butterfly

‘mutual (often members of communal
attraction,’ seasonal) other roosts; bee
mediated by populations sleeping
intrinsic or uncritically clusters;
environmen- accepted beetle
tal hibernation
factors clusters
Simple groups As above Coordinated Temporary, Open Locust and
movement facultative butterfly
migrations
Primitive Siblings Reciprocal May persist Open; Passalid
societies groups; communica- over a life members beetles; tent
parental care tion; may stage; often ± caterpillars;
cooperate in centers on related embiid
nesting nest webspinners
Advanced Obligatory Highly More or less Closed; Ants, termites,
societies interdepen- integrated, permanent members all other
dence of all cooperative; closely so-called
develop- division of related; all ‘eusocial’
mental labor others insects
stages excluded
a Overlaps are common.
more sophisticated associations in which members exhibit at least some coordi-

nation in their movements. Such simple groups, again comprised of conspecific
individuals remaining together for a period of time, are made up of group members
that interact with one another to a distinctly greater degree than they do with other
conspecific individuals. Swarms of midges, parasitic wasps, chorusing cicadas, and
migratory swarms of locusts (see Plate 4) or butterflies (see Plate 3) all provide good
examples of simple coordinated groups.
Every behavior has costs and benefits, and is expected to have evolved only if the
latter exceed the former. How might individuals benefit from associating with con-
specifics? In mating swarms, for example, individuals might benefit from increased
reproductive opportunities. Opposing this is the potential cost of increased inter-
ference and competition for mates; there also may be an increased risk of being
deceived or manipulated, as occurs with femme fatale fireflies (see Chapter 7).
Similar cost/benefit analyses apply to feeding assemblages. Food detection
and/or utilization may be more efficient in a group. When food supplies are patchy
and transient but rich, a group may locate and be able to defend such food supplies
more efficiently than isolated individuals, particularly if some sort of communica-
tion between the individuals has developed. (Fully social insects excel at recruitment
Fig. 10.2 Sleeping aggregations. (left) A cluster of males of the solitary bee, Chalicodoma, from
Australia ‘sleeping’ on a clothespin. These males returned every evening to the same site for about
three weeks. (right) A group of male Actinote surima surima butterflies roost at night in Brazil.
Similar sleeping aggregations of various butterflies and solitary bees and wasps occur commonly
worldwide
of large numbers of workers to exploit food resources.) Costs include heightened

competition for the food and the necessity to share.
Enhanced defense can be another benefit of groups. First, any assemblage of
individuals is likely to improve the chances of detecting potential predators, the
so-called group vigilance effect (see Plate 17). Additionally, gregariousness may
enhance predator confusion, reducing a predator’s hunting success, as well spread-
ing the risk that any particular individual in the group becomes a predator’s meal.
Massive periodical cicada aggregations are a striking example of a situation where
potential predators are simply overwhelmed.
Mollification of the environment through such effects as reduced convective heat
loss or evaporative water loss benefits individuals in a cluster. A possible disadvan-
tage is that individuals in such groups may be more vulnerable to pathogens and
disease outbreaks.
More subtle influences sometimes are also at work. Blattella germanica cock-
roach females apparently feed more and reproduce more quickly in the presence
of conspecifics than if kept alone. When confined in groups, Pogonomyrmex ants
show more intense digging behavior than isolated individuals given the same con-
ditions. Human beings show analogous behavior. For example, it takes a great deal
of resolution to avoid joining in when faced with a yawning or laughing person.
Social facilitation may be defined as an increase in the pace or frequency of a
given behavior due to the presence or activities of another conspecific individual. An
example of its workings may be seen in the Australian spitfire sawfly, Perga affinis,
Fig. 10.3 Burrowing battalion. Australian sawfly larvae of the genus Perga crawling on the
forest floor prior to cocooning. In order to successfully construct cocoons and pupate they must
first penetrate the crusty soil. Although individuals are poorly adapted for such burrowing, in larger
aggregations at least one larva usually succeeds in breaking through and the others are then able
to follow. If disturbed, the writhing mass of larvae rear up and regurgitate sequestered droplets of
liquid rich in the essential oils of their eucalyptus host plants, a behavior that gave rise to their
common name, spitfires
where an individual’s very survival depends upon it. Females lay their eggs in pods
within the tissue of eucalyptus leaves; after hatching the young must first gnaw an
exit hole before they can begin feeding. This is a difficult task, but if one succeeds,
the others follow. There is a strong correlation between pod size and larval mortality.
In one study, pods with fewer than ten eggs suffered 66% mortality, but among those
containing more than 30 eggs, mortality was only 43%. Social facilitation is also
involved when mature Perga larvae leave the tree to pupate (Fig. 10.3).
10.2.2 Parent-Offspring Interactions
Certain female sawflies guard their eggs and young (Fig. 10.4). Young Gargaphia
solani bugs orient to their mother and follow her from place to place. Scarab beetle
pairs cut out a chunk of freshly deposited manure, roll it off, and place it in an
elaborate underground burial vault where eggs are eventually deposited. The parents
of Brazilian Phloeophana longirostris pentatomids apparently provide nourishment
for their nymphs in addition to protecting them.
More than a dozen orders of insects include at least some species that care for
their young. Parent-offspring groups are the most common form of social assem-
blage in the insects and most often parental care equates to maternal care. Next
most common is biparental care, followed by exclusive paternal care, and finally
Fig. 10.4 Guarding young. Maternal care in the Brazilian sawfly, Themos olfersii. Throughout
the approximately 20 days required for incubation, the conspicuous orange and black female strad-
dles her eggs. If disturbed, she displays one of a series of at least ten different types of defensive
or threatening reactions. Later, even when the fully sclerotized gregarious larvae migrate together
to the basal portion of their host leaf and begin feeding, their mother continues to guard them,
sometimes accompanied by other females that happen to be nearby. Apparently because of their
distastefulness and warning coloration, the adults are only rarely subject to predation, and eggs and
young also benefit from this protection
by a taxonomically small group of truly miscellaneous assemblages that were once

lumped together under the term ‘communal insects.’
The roots of maternal care and biparental care are fairly easy to envision. Female
insects in any population probably vary in the alacrity with which they abandon
their eggs. For females in ecological situations where a tendency to stay around
might repeatedly result in greater offspring survival, such persistence would be evo-
lutionarily favored. However, in other cases, leaving immediately might be a better
female strategy because it would allow more opportunity for the additional feed-
ing necessary to mature further clutches of eggs. Biparental care probably had its
roots in an extended association of the sexes after mating; for a male, remaining
with one’s mate until eggs are laid has the result of increasing a male’s certainty
regarding parentage (see Chapter 9).
As one looks at cases where parental involvement stretches ever further into the
life of the offspring, various incipient social behaviors begin to appear. Among
scarabaeid beetles, for example, one finds a spectrum of behavior from species
in which the adults simply amass manure provisions upon which eggs are laid to
species such as Copris where the mother stays with her young until they reach
adulthood. In conceptualizing the development of sociality, it is tempting—but
misleading—to see such variation as proof of a linear, unidirectional evolutionary
march toward some sort of ‘perfection’ represented by ‘true’ sociality. Models of

insect sociality that arose in the early twentieth century relied heavily on such ideas;
one of the best known was that of William Morton Wheeler, in which social evolu-
tion progressed through seven steps of increasing parental care to an apex presided
over by his own favorite social insects, the ants. The caution here is that one must
remember to recognize these correlations as useful organizational devices, rather
than reflections of an evolutionary progression.
Parental investment theory predicts that because parental care adds to the inequal-
ity of male and female investment, under most conditions it will merely reinforce
a polygynous mating system (see Chapter 9). In highly polygynous systems and in
promiscuous systems, paternal care of young will be rare or there may be no parental
care at all.
As parental investment theory predicts, when parental care exists, females
almost always are the sole providers. Unassisted male parenting has been esti-
mated to occur in fewer than 150 of the nearly one million insect species
that have been described. However, it has been documented in diverse species,
including the Hymenoptera, Coleoptera, Hemiptera, and Thysanoptera, and in a
few species the male’s contribution to the fitness of his mate and offspring is
extensive.
Almost all cases of unilateral paternal care occur in one bug family, the
Belostomatidae, or giant water bugs. Water bugs in the genus Abedus are a striking
example. These moderate-sized to large aquatic hemipterans are predators, some
of the few insects that prey directly on small vertebrates such as tadpoles and fish
fry. Two aspects of their reproductive behavior are unusual. One is that females
invariably attach their eggs to the backs of males, where the eggs remain until
they hatch (Fig. 10.5). The other is that the females go their own way shortly after
mating and oviposition rather than sharing in parental care as some other insect
pairs do.
Reflecting a prevailing chauvinistic attitude as to the ‘proper’ sex roles, early
observers generally believed that the egg-bearing water bugs were females carrying
their own eggs. Some authors even described the oviposition process, attributing the
feat to a long protrusile ovipositor that could be extended over the female’s body.
In 1899, F. W. Slater (a woman) set the record straight in a delightful account not
totally emancipated from chauvinistic overtones:
That the male chafes under the burden is unmistakable; in fact, my suspicions as to the sex
of the egg-carrier were first aroused by watching one in an aquarium which was trying to
free itself from its load of eggs, an exhibition of a lack of maternal interest not expected in
a female carrying her own eggs.
A few years later another entomologist, while mentioning the ‘indignity’ of such
male servitude, noted the ‘peculiar fact’ that copulation took place in connection
with oviposition in the species he observed. However it took almost 75 years before
Robert Smith’s pioneering study (Case Study 10.1) began to make the situation
understandable, and even today studies of male parental care in these bugs continue
to yield new insights.
Fig. 10.5 Reversing roles. The copulation-oviposition cycle of the giant water bug, Abedus her-
berti. After preliminary sparring between the sexes, receptive males perform vigorous pumping
(a). Females respond by climbing on the male’s back (b) as if to oviposit. With one hind leg, the
male manipulates the female off into a copulatory position (c). About one minute after intromis-
sion, the male abruptly ‘scrubs’ one hind leg on the female’s hemelytra. Immediately the female
repositions herself (d) and begins to lay eggs, starting at the apex of the male’s hemelytra. After
about five minutes, although the female has laid fewer than four eggs, the male’s temporary qui-
escence ends. Forcing the female out of position with his hind leg, the male begins a new bout
of vigorous pumping. This cycle continues until the female’s total egg clutch is deposited on the
male’s back (e)
Case Study 10.1: Reversed Sex Roles in the Giant Water

Bug, Abedus
Female water bugs abandon their eggs to males that proceed to care for them.
What has caused such an unusual behavior?
Among the giant water bugs, Abedus herberti, females oviposit on the
backs of males, and the latter carry them about until the first instar nymphs
emerge and swim away. An egg-encumbered male of the common giant water
bug, Abedus herberti, can be placed in an aquarium with a bit of water and a
spot to perch, and an unusually high number—about 97%—of the eggs will
usually hatch. In the mid-1970s, Robert Smith began with some simple ques-
tions. Does this high egg survival rate depend upon eggs being carried about
upon the back of a live male water bug? If so, why?
Removing egg pads from the backs of six males, Smith placed each egg
pad in a separate fingerbowl in the laboratory; three were covered with water
and three were allowed to dry. None of the eggs survived; those kept in water
developed a fungus after about four days, and those allowed to dry became
desiccated. Repeating the experiment but first washing the eggs with distilled
water, he got the same result. Was something about the male exoskeleton
necessary to their development? Smith killed egg-bearing males and left
them with eggs intact in their containers, but these eggs also failed to hatch.
Attachment to a healthy male appeared essential for eggs to develop normally
and hatch. Why?
Brooding males, Smith noticed, tended to prefer shallow water and seemed
to continually be adjusting their position to keep their egg pad at or above
the water’s surface. Perhaps regular exposure to atmospheric air was a requi-
site for normal development. To test this, he placed an egg pad in a finger
bowl and added water to a depth of less than 3 mm. Evaporative losses
tended to expose the unattached ends of the eggs in the center of the pad,
so he added water each day. After two weeks, he was excited to find that
some of the centermost eggs began hatching. And although only 14 of the
72 eggs in the pad finally hatched, the success was a notable improvement
over previous results. Nevertheless, it still appeared that the live brooding
male was providing services beyond simply exposing the eggs regularly to
atmospheric air.
Observing brooding males of a related species of belostomatid bug in the
field, Smith noticed that submerged egg-bearing males frequently pulled their
hind legs vigorously and repeatedly over their egg pad from front to rear,
whereas unencumbered males only rarely stroked their backs, and then appar-
ently only as part of normal grooming behavior. Perhaps such stroking also
was necessary. Watching A. herberti males, he saw what seemed to be anal-
ogous behavior of a slightly different form. Egg-carrying males spent nearly
one third of their submerged time in ‘brood pumping’; resting on submerged
vegetation or the stream bottom, they rocked their bodies repeatedly forward
and backward, pivoting on their middle legs while remaining in a fixed posi-
tion. He also noted that before they began to rock or pump, males of both
species patted or touched the eggs with their hind legs.
Was patting an essential precursor to pumping or stroking? Smith quickly
peeled off egg pads after a pumping bout had been initiated but while the bug
was still under water. The males all continued to pump. Later, when they felt
about for the eggs and found them gone, they no longer would pump. Thus,
Smith concluded that male water bugs used patting and stroking to inform
themselves of the status of their egg pad.
Why should males take on such a task? It seemed evident that a brood-
ing male is exposed to extra risks. The additional weight of developing eggs
would impair his swimming and perhaps reduce his ability to escape preda-
tors. Smith measured swimming speeds of egg-carrying and unencumbered
males and found that the latter could swim considerably faster. The behav-
iors involved in the father’s parental care undoubtedly also increase risk of
predation. Because water bugs normally rest motionless, protected by their
cryptic coloration, pumping or stroking behaviors probably make them more
vulnerable by betraying their location.
Perhaps males were merely being tricked into doing this work. If so, it
seemed likely that they could be tricked into carrying any female’s eggs with
equal likelihood. Were male A. herberti carrying eggs that they had not fertil-
ized themselves? Through detailed observations of Abedus mating behavior,
Smith showed that courtship and oviposition were under overwhelming male
domination. The male waterbug commanded the female to copulate, guided
her into the copulatory position, signaled his desire to uncouple, and deter-
mined the length of time she was given to oviposit. He allowed no oviposition
to take place at all until after the female had mated with him at least once
and usually several times. Then, as the female began laying her eggs, he inter-
rupted her frequently to insist upon additional copulations. As predicted by
sperm precedence, by insisting on copulation prior to oviposition, the male
water bug ensures that the eggs he receives will have been those that he has
fertilized. By limiting the period of oviposition he also minimizes the chance
that he will mistakenly accept eggs fertilized by another male. Does this mean
that belostomatids are monogamous? Apparently not. Smith was able to show
that males are able to simultaneously brood and fertilize multiple clutches
of eggs from different females; bug’s strict control over the situation would
assure that these offspring were his own, as well.
Thus, Smith established that male giant water bugs have developed a whole
set of courtship and mating behaviors that maximize individual male fitness by
ensuring paternity and maximize female fitness by placing brooding behavior
under male care and thus freeing themselves to capture the prey necessary to
make additional clutches of eggs.
Smith proposed that male brooding in water bugs arose by ancillary selection,
a form of natural selection that produces characters supported by primary traits. In
this case, the primary trait was selection for large body size that occurred more than
150 million years ago. Larger bugs were able to take advantage of vertebrate prey,
but larger size also meant larger eggs. Because of the way that insect eggs develop,
those beyond a certain size cannot develop unattended under water, and thus the
stage was set for parental care in these species.
Why should it not be the female water bugs, rather than males, that brood the
young? An enhanced fecundity hypothesis that was originally proposed for birds
may provide an answer. Basically, this model suggests that if one of the costs that a
female incurs by brooding is a reduced ability to obtain nutrients needed to produce
eggs, selection will favor male traits that reduce this cost and improve the number
of offspring the pair can produce. Because males require only enough nutrition to
maintain their body, rather than to maintain it and produce energetically costly eggs,
they can invest in parenting behaviors at a lower cost than females can.
Research suggests that this theory applies very well to giant water bugs. In sev-
eral feeding studies in which adults have been given generous prey opportunities
immediately after mating, female water bugs have consumed many more prey than
males have eaten. (In one case with goldfish, females killed five to nine times as
many fish as males did over a seven to nine day period.) Males seldom feed while
brooding, and also appear to be inhibited from feeding on prey in the size class of
first-instar nymphs, an adaptation that may minimize the risk of males consuming
their own hatchlings. Even stronger evidence was provided by a study that compared
the effects of food limitation on fitness of males and females. One group was fed
every day; the other received one prey every ten days. The results were dramatic.
Both well-fed and poorly fed males sired statistically similar numbers of offspring.
However, well-fed females laid 3.5 times as many viable eggs as poorly fed females
did. Thus, as a result of freeing females to pursue prey while males assume the task
of parental care, both sexes benefit through an increase in offspring numbers.
Because Abedus males insist on a copulatory bout after every few eggs, it can
take hours before a female finishes depositing her full complement of eggs. Is the
threat from other water bug males really so great as to require going to such lengths
for paternity assurance? Smith mated a female to a male that was homozygous for
a dominant genetic marker. Then he paired that female to another male that he had
‘ductectomized’ so that it was unable to fertilize eggs. When this sterile male’s
brood hatched, all bore the genetic label from the female’s previous mate (Fig. 10.6),
demonstrating both sperm precedence and a very real threat of cuckoldry in this
species.
10.2.3 Solitary and Communal Nesters
All stages of the house cricket, Acheta domestica, spend the day concealed in nooks
and crannies. As soon as they hatch, the larvae of bagworm moths (Psychidae) begin
Fig. 10.6 Brooding about baby. This male giant water bug has been made an experimental cuck-
old; the stripe on the first hatched nymph’s back identifies its father as a genetically marked male
that copulated with his mate before she was placed with him
making a mobile case of silk and environmental materials; extending the front as
they grow, and excreting waste out the back, they move about, consuming vegeta-
tion. Adult males leave to mate, but females remain in the bag; birds that eat them
pass the hard-shelled eggs unharmed, spreading the population. Across much of
Australia, homeowners are familiar with the massive nests constructed in unusual
locations by the continent’s largest mud dauber wasps, Abispa (see Plate 35).
Many larvae take advantage of refuges, from naturally occurring crevices to cav-
ities formed as a side effect of their own feeding. Others construct their own shelters
as feeding contrivances, such as the funnels made by the ant lions and worm lions
(see Fig. 4.6) and the cases constructed by caddisflies. Only the broadest definitions
consider these to be nests, however. A nest is more often defined to be a structure
that an adult animal builds to shelter its young, using available natural materials that
it modifies in some way. When adults build nests for their offspring, the behavior is
considered to be a hallmark of a high level of parental care. A colony, on the other
hand, is the social unit that occupies the nest, consisting of brood as well as adults.
(This distinction will become quite useful in discussing fully social insects later in
this chapter.)
Leaving aside the complex nesting behaviors outlined in the ‘social register’ later
in this chapter, let us consider here the solitary nesters, in which one adult or at
most a pair of adults build the nest and in some species, defend it. These are not
numerically common behaviors, but they have appeared in a taxonomically diverse
array of insects, often occurring in those groups closely associated with the soil. For
example, nesting is found in certain earwigs, burrowing crickets (Fig. 10.7), some
cockroaches, various solitary bees and wasps, and a variety of beetles, including the
dung beetles.
As an example, let us consider the latter. The term ‘dung beetle’ has commonly
been applied to three subfamilies of the large beetle family Scarabaeidae. Only
Fig. 10.7 Burrowing crickets. A female of Anurogryllus arboreus in her brood chamber. To the
right is the exit burrow; to the left, a filled defecation midden. A single newly hatched nymph may
be seen touching antennae with the mother, while an egg pile lies beneath her abdomen partly
covered by soil. The chamber floor has bits of grass she has gathered, upon which nymphs feed. In
this species, special miniature trophic eggs laid by the mother also serve as baby food, an unusual
trait; these nonviable eggs are very attractive to the offspring, which crowd around them and fight
for their possession
in the Scarabaeinae, however, are the vast majority of the roughly 4,500 species
coprophagous, most feeding on the excrement of large mammals. No known adult
Scarabaeinae have mouthparts capable of chewing or cutting solid food, hence when
dung begins to desiccate it becomes unusable to them. This, plus the clumped
discontinuous distribution of dung pats of many grazing animals, creates a fierce
competition. Food relocation to constructed nests is a key behavioral feature of this
subfamily and is thought to have evolved to reduce predation, parasitism, and com-
petition, especially from coprophagous flies. Nest construction in this subfamily is
nearly always the cooperative endeavor of a male–female pair. The association is a
close one, and once a pair bond is formed, it endures, at least until after the dung
has been lodged in its final resting place.
Gonzalo Halffter and his colleagues, the team responsible for the most com-
prehensive studies available on the natural history of dung beetles, recognize two
major behavioral groups, burrowers (Fig. 10.8) and rollers. (A third small group are
the dwellers.)
Burrowers, which include the well-known genera Copris and Synapsis, compact
a piece of dung, cut it into spheres and place them in a nest they have excavated
below or adjacent to the dung pat. The female beetle attends the offspring until
they pupate, cleaning mold growth from the surface of the spheres and applying
antimicrobial chemicals derived from her sternal glands. Males help dig the nest,
Fig. 10.8 Burrowing dung beetles. In burrowers such as Onthophagus gazella, couples excavate
burrows below a dung source, then make a series of brood balls that the beetle’s larvae consume as
they complete their development
provision it, and compact the dung. During this period of participation, they also
actively defend the nest using (in many species) the prominent horns on their head
or prothorax.
Rollers (also called tumblebugs) (see Plate 36) are the second major group; they
include most of the tribe Scarabaeini. The sacred scarab of ancient Egypt belonged
to this group. The ability to roll a ball of dung overland (sometimes up to 40 m from
the source) sharply distinguishes this group and results in a quite homogeneous
group behaviorally. Generally the male pushes and the female pulls or rides on top,
walking in place like a hamster on a treadmill. Because morphological adaptations to
dung rolling (especially longer and more slender legs) often mean a partial or com-
plete loss of digging ability, there are no elaborate nests or impressive excavations
in this tribe, and some do not bury the brood ball at all.
The Australian continent lacks native large mammalian herbivores. Conse-
quently, native dung beetles have evolved as specialists on the small bits of
excrement characteristic of marsupials such as wallabies and kangaroos. With the
advent of sprawling cattle ranches in northern Australia, cattle dung became an
unanticipated blight on the countryside as large amounts accumulated in the pas-

tures. To restore the ‘balance of nature,’ Australian entomologists have imported
several species of dung beetles from Africa. There is, however, one ‘odd-ball’ native
Australian dung beetle that perhaps more properly should be called a ‘compost’
beetle. This beetle (Cephalodesmius armiger) is quite common in parts of northern
Queensland where bisexual pairs dig elaborate burrows that they stock with dead
leaves and other plant debris gathered from the forest floor. This plant matter is then
prepared into small food balls to which the beetles add their own excrement and
possibly a particular type of fungus, and eggs are laid individually in each ball. As
a larva grows, biparental care continues; the male brings additional leaves that the
female prepares and progressively adds to the food ball as the larva consumes it
from within. When the larva pupates the female covers the exterior of the brood ball
with a mix of excrement and remains in the nest until she dies. This species displays
the most complex social behavior in its subfamily.
Extensive biparental care also is the hallmark of burying or carrion beetles of
the family Silphidae, another group that has convergently evolved behaviors simi-
lar to dung beetles. They face a problem similar to that of the dung beetles, in that
the cadavers that serve as their food are valuable, spatially unpredictable resources,
for which competition is keen. To secure a small carcass prize such as a mouse
against competing flesh flies and other carrion-feeding insects, Nicrophorus beetle
pairs transport the carrion away from the discovery site, then bury it. Working below
ground, they proceed to extensively process the carcass, removing its fur and adding
embalming antibiotic secretions. They then cooperatively raise a brood of young,
feeding them partially predigested meat, for which the larvae even appear to beg in
a manner analogous to hungry fledglings in a nest. The parents also defend their off-
spring from predators such as staphylinid beetles and conspecific nest usurpers. One
curious aspect of carrion beetle behavior is the propensity of females to lay more
eggs than the carcass resource can support, only to later cull out some of the brood.
In their nest-building activities, wasps, ants, and bees exhibit some of the most
complex behaviors known in invertebrates. We have discussed some of these before:
the ‘fixity of instinct’ shown in the nesting activities of Ammophila pubescens, the
interaction between predator and prey in the cricket-hunting Liris nigra, and the
remarkable ability of the bee wolf, Philanthus triangulum, to learn the locality of
her nest and the stereotyped sequence of behaviors involved in prey capture. All of
these predatory wasps are in the family Crabronidae (formerly called Sphecidae), a
large and diverse group.
Consider another example. In early summer, sand dunes throughout the world
are alive with large yellow and black wasps flying briskly about low over the dune
surface. Later, these flying wasps seem to dwindle in number, but closer inspection
will usually reveal wasps quietly at work, each digging, filling with prey, then clos-
ing nest burrows in the sand. Each female does everything alone, without assistance
from the others. A patient observer may be rewarded by the sight of a wasp sud-
denly appearing with a large, brown fly under her body held by her middle legs.
The wasp alights, makes a few digging movements synchronously with her front
legs, then quickly disappears through this previously invisible door. At the burrow’s
end, several inches below the surface, one finds a chamber with a legless, helpless
yellow-white grub feeding on the flies brought in by its mother. After eating about
two dozen, the grub will spin a cigar-shaped cocoon inside which it will eventually
metamorphose into an adult.
Like these sand wasps, the vast majority of wasps live solitary lives. Their nests
belong to three broad types: those dug in a substrate such as soil, rotten wood, or
plant pith; those constructed in preexisting cavities such as hollow twigs; and those
constructed wholly of foreign materials such as plant pulp, mud (see Plate 35), or
resin. Many of the details of solitary wasp nesting behavior appear to have been
molded by biotic factors, notably various natural enemies and nest associates. For
example, among ground-nesting crabronid wasps several species construct acces-
sory burrows, blind-end false tunnels close beside the true nest burrow. In every
case the accessory burrows are left open and the true nest entrance closed off.
There is evidence that such accessory burrows divert the attention of parasites; bom-
byliid flies may lay their eggs in them (Fig. 10.9), and velvet ants often spend time
investigating and digging in the bottom of such holes.
Species nesting in the soil have another problem relative to pests, namely, soil
removal. A mound of dirt at the nest entry can be a dead giveaway of the presence of
that nest. So while some species simply let the displaced soil accumulate outside the
nest entrance, others carry out the soil a bit at a time in their jaws, flying off a short
distance before dropping it to the ground. Still others push or rake the dirt back-
ward out of the nest and then by a variety of behaviors remove or at least disperse
Fig. 10.9 Deceiving enemies. A bombyliid bee fly hovers over an accessory burrow of a solitary
wasp. With repeated flicks of her abdomen she flings several eggs into the hole, which is actually
a fake nest entrance made by the wasp
Fig. 10.10 Hiding the evidence. Patterns created by mound-leveling behavior of three Australian
Bembix sand wasps during nest construction. Throughout both digging and leveling, synchronous
motions of the long-spined front tarsi rake the sand, producing a species-characteristic pattern
the conspicuous mound. These leveling behaviors, which are a species characteris-
tic, typically consist of repeated passages over the mound while scraping with the
front legs and turning slightly from side to side (Fig. 10.10). The enigmatic resul-
tant patterns, highly diagnostic of the species that made them, are often elaborate
but ephemeral due to the homogenizing action of wind and weather. In addition to
dispersing the mound, it has also been postulated that leveling may serve to disperse
or cover any wasp scent that might be detected by foraging ants or other predators,
parasites, or scavengers.
Solitary nesting species sometimes clump their nests into aggregations with con-
specifics, in this way gaining some measure of support in defense, but many solitary
bee and wasp species have evolved another strategy. Each female provides the food
for her own offspring, but several females occupy a communal nest together. Such
nest sharing takes various forms that differ in duration of association and degree
of cooperation between females. In some species, females co-occupy nests soon
after emergence and simultaneously provision different cells for several days, but
10.3 The Insect Social Register 407
finally all but one of the females is forced out. In other species, females stay together
apparently amicably for prolonged periods.
Several advantages for this communal behavior have been proposed, including
increased efficiency at nest building, reduced parasitism, and improved nest defense.
However, because communal nesting is uncommon, there must be substantial, if as
yet largely unidentified, costs to this system as well. One almost certainly is the
ever-present temptation for kleptoparasitism (see Section 4.2.5) and prey theft.
10.3 The Insect Social Register

Asked to name an example of a social insect society most of us would quickly
come up with at least ants or honey bees, if not also paper wasps or yellowjackets.
Awe-inspiring due to their complex division of labor, fierce defense of their nest,
and sheer numbers, they fascinate scientists and nonscientists alike (Fig. 10.11). Yet
these highly social species are in many respects only the tip of the iceberg. Diverse
social arrangements have evolved in many other insect groups from bugs to beetles,
and even caterpillars and cockroaches; insights gained from such diversity can help
us to better understand and appreciate the highly social forms.
Fig. 10.11 Going places. The Rev. H. C. McCook, an early American myrmecologist, wrote this
illustration’s original caption, ‘Winged female ants at play on the plaza.’ Actually, these are virgin
females preparing to leave on their mating flight
The name of this section is drawn from an old custom apparently specific to the
United States; the Social Register is a directory of names and addresses of prominent
American families who form the social elite in a number of major cities. These tra-
ditionally were the ‘members of polite society’ (i.e., those with ‘old money’); recent
editions now also include those with ‘newer’ wealth, the political or corporate elite.
Like that human social register, the insect social register includes the well-
established examples, with a nod to newcomers. The level of social organization
recognized as a society includes most instances of parental care discussed in the
first part of this chapter. It also embraces some groups of developing individuals
of the same generation (see Plate 37). Certain tent caterpillars, for example, spin
a home web at which they gather gregariously to sleep and a feeding net which
also serves as a protective community shelter; successful foragers deposit chemical
trails that elicit preferential following by unfed individuals (see Fig. 6.13). In most
insect species parents die before progeny mature, or else offspring disperse and do
not remain associated with the parent. In some, however, at least some young per-
sist with their parent(s) for one or more generations as an extended family living in
some type of domicile (Fig. 10.12).
As in so many areas of biology, many of the first serious scientific efforts to
understand social behaviors involved attempting to classify them. However, trying
to cram the ever-expanding number of examples of insect social groups into existing
Fig. 10.12 Living in a log. Horned passalid beetles, Odontotaenius disjunctus, sharing a common
tunnel system in a decaying oak log during the fall. During this nonreproductive season, both adult
sexes are present but young are absent. In the following breeding season, pairs cooperate to rear a
brood of larvae, sometimes with the help of other adults. Both adults and larvae produce a diversity
of audible sounds in one of the richest auditory repertoires known for any insect, but their functions
remain poorly understood
classification frameworks that date back to the early 1900s has led to such confusion
as to call into question the value of such an activity. Even the widely cited tradi-
tional lexicon from the 1960s (pre-, sub-, quasi-, semi-, para-social) is cumbersome,
somewhat teleological, and has difficulty accommodating the range and complex-
ity of cooperative interactions of insects. Beginning in the mid-1990s a number of
alternative classifications were proposed that ranged from conceptual expansion, to
narrowing or redefining, to total abandonment.
Today there is a growing consensus in favor of a return to a broader, more
inclusive interpretation of social behavior, with a shift in focus to the case-by-case
examination of natural histories and a focus upon particular ecological constraints
that promote group behavior. Thus, we retain only the term applied to the most
advanced societies: eusocial.
While it seems that all levels of intraspecific interaction from completely solitary
behavior to highly complex coordinated societies are found in insects, compar-
atively few species are eusocial. They number a few thousand species and are
included in five orders, Hemiptera (bamboo and gall-forming aphids), Thysanoptera
(gall thrips), Coleoptera (a bark beetle), Hymenoptera (ants, bees, and wasps), and
Isoptera (termites). Molecular phylogenetic study suggests that the termites should
properly be considered as a family of cockroaches (Blattodea), an idea that has a
long and often contentious history (see Chapter 1). To minimize confusion we will
mostly avoid using either ordinal name and simply refer to this group as the termites.
Eusocial insect species are traditionally considered to include only those whose
societies meet three criteria: they live as groups of adults of different generations,
with cooperative activity, and with different individuals obligatorily performing dif-
ferent roles essential for the success of the group. The last of these three is the
essential defining criterion that sets the bar relatively high for admission to this
exclusive club. As originally applied, it refers to the presence of a reproductive divi-
sion of labor, in which there are sterile workers and one or more fertile queens.
A broader but still succinct definition of eusociality has been offered by Edward
O. Wilson, ‘care across generations of the offspring of a reproductive caste by a
non-reproductive or less reproductive worker caste’.
Castes are defined as a subset of the colony or group that is both morphologi-
cally distinct and behaviorally specialized. The physiology of caste determination
is beyond the scope of this book, but appears to involve changes in regulation of
highly conserved molecular pathways interacting with life history attributes in the
different social lineages.
Division of labor is a prominent feature across all the eusocial insects, and worker
and queen castes are the most common job descriptions. Among eusocial bees,
wasps and ants, workers are female adults and usually sterile, although some may
occasionally produce eggs. Among honey bees, highly developed task-related spe-
cialization has evolved; despite both being genetically female, the worker caste and
the queen caste differ in at least 53 quantitatively distinct morphological features in
addition to their obvious physiological and behavioral differences.
Caste polymorphism reaches its most spectacular development in termites, the
other large eusocial group. Workers are of either sex and include immature stages as
well as adults. Queens of some species develop huge abdomens so distended with
developing eggs as to resemble a small sausage (Fig. 10.13), and soldiers specialized
for defense may exhibit either enlarged heads with powerful jaws or conical nozzle-
like heads that eject a sticky substance (see Fig. 5.10).
In addition to caste polymorphism, as colony complexity increases and the range
of duties surrounding the care of the young becomes correspondingly greater, in
some cases an age-related division of labor also develops. For example, in the
ant Myrmica scabrinodis, worker individuals that have emerged during the present
season function as nurses; those that became adults during the previous season
are builders, and even older individuals act as foragers. Such temporal polyethism
reaches its greatest development among the honey bees. On emergence, a young bee
works as a cleaner for about three days. After this, coincidental with labial gland
development, she becomes a nurse, producing secretions with which she feeds lar-
vae. About the tenth day of her life her abdominal glands begin to produce wax
and her labial glands atrophy; she becomes a builder. At about the sixteenth day she
begins to receive nectar and pollen loads from foragers and stores them in the comb.
On about the twentieth day, she fills the post of guard. Finally, she becomes a for-
ager, working at this for the rest of her life. Age polyethism, though well developed,
is not inflexible. If a hive is divided so that one half contains only young bees and
the other only older foraging bees, after a few days of adjustment, workers in each
half carry out all normal hive tasks (see Case Study 1.1).
10.3.1 The Ants

Of all eusocial groups, ants are the most widespread and numerically abundant
and contain more known genera and species (approximately 12,000) than all other
eusocial groups combined. In variety of ecological and social adaptations they are
unparalleled.
Like the termites, early ants nested in soil and leafy compost, putting themselves
in position to exploit an extremely rich microhabitat and to build and maintain
long-term nests in a protected location. However, because they began as predators
upon other arthropods (see Plate 38), primitive ants were not bound like termites
by a cellulose diet dependent on intestinal symbionts. The development of a wing-
less worker caste, an innovation already present as far back as the mid-Cretaceous
period, increased ants’ ease of access into soil and plant crevices. Acids secreted
from the metapleural glands, an anatomical development possessed by all ants but
by no other Hymenoptera (see Fig. 6.1), inhibit the growth of microorganisms in
their moist nest chambers.
Most ants live in soil still, either excavating chambers or constructing elaborate
mound structures. Many others, however, have adopted an arboreal existence with
tough carton-covered nests or homes in hollow twigs or specialized plant parts such
as hollow thorns (see Plate 8). Others like the leaf-cutter ants (see Chapter 4) cul-
tivate gardens of fungi. Weaver ants (see Plate 19) are unique in their use of larval
silk to fasten leaves together to form large baglike nests. Members of the largest ant
Fig. 10.13 Social insect castes. (above) Two castes of the African termite Macrotermes. The
soldier’s huge head serves in colony defense, but soldiers must depend upon the smaller workers
to feed them. (below) The extreme physogastry of this termite queen dominates the picture; her
sausage-like abdomen greatly overshadows her tiny head and thorax. Her abdominal tergites have
become pulled widely apart by the stretching of the intersegmental membranes and appear as
crescent-shaped islands on a sea of whitish membrane. Attending the queen are numerous workers
Fig. 10.14 Blocking entry. A colony fragment of the myrmecine ant Zacryptocerus varians in a
hollowed-out stem of red mangrove. Three female castes are shown: the queen rests on the floor
of the nest to the left, while on the right a large major worker blocks the nest entrance with its
saucer-shaped head. Behind the queen another major worker receives regurgitated liquid from a
minor worker
genus—the carpenter ants (Camponotus), with about 1,500 species—excavate nests

in wood. In a number of wood nesters, larger workers possess distinctive disc-shaped
heads used to form living plugs at nest entrances (Fig. 10.14).
Some ants make no permanent nest. The term ‘army ant’ popularly includes any
of the 200 or more carnivorous ant species that live as nomads and send out sorties
of great numbers of workers. In a stricter sense, ‘army ant’ applies just to members
of the ant subfamily Dorylinae, awesome predators that pioneer myrmecologist
William Morton Wheeler called the ‘Huns and Tartars of the insect world.’ Nearly
blind, these predominantly tropical ants feed almost entirely upon live arthropods
captured during mass expeditions or raids. Moving like great armies across the land,
overcoming every obstacle in their path, even bridging small streams with living sus-
pensions of workers, they seem to have military precision. Columns of ants maintain
infantry formation while, like a real army, being accompanied by both scouts and
fighting soldiers.
The first investigator to study army ant behavior objectively in detail was
Theodore Schneirla, who spent over 35 years studying Eciton in both laboratory
and field. Case Study 10.2 presents some of his work, in which he demonstrated
that the interrelationships among colony members provide the driving force for the
cyclical behavior of army ants (Fig. 10.15). Over time, it has become increasingly
clear that pheromonal mediation takes place in the army ants and that the ultimate
adaptive significance of the alternation of quiescent and nomadic phases is to pro-
tect not only the helpless young but the queen as well through a time when both
are most vulnerable. While little mention of the queen’s role in the two phases has
been made, early in his studies Schneirla observed that queens captured during the
nomadic phase had considerably less swollen abdomens than those captured during
the statary phase. The queen lays her clutch of thousands of eggs all during just
a few days in the mid-statary phase. As the colony enters the nomadic phase, her
ovaries cease activity.
Case Study 10.2: Cyclical Foraging in Army Ants, Eciton
Any first-time visitor to lowland rainforests of the New World tropics will
quickly note the ubiquity of ants patrolling nearly every surface. Periodically,
conspicuous hordes of army ants ominously swarm over every piece of vege-
tation driving hapless arthropods before their onslaught. How is this system
maintained?
Each day within the tropical forest, Eciton workers stream out at dawn to
begin their raids anew, their branching columns quickly overrunning areas up
to 100 m from their nest. At the front line, biting, stinging ants attack insects
and other arthropods, tearing their prey apart and carrying the softer pieces
back to the nest so that the forest floor soon has a series of two-lane highways
traveled by steady streams of advancing raiders and returning victors. So effi-
ciently does the massive raiding progress that one colony may haul in more
than 100,000 other arthropods in a single day. Then, as night begins to fall,
the whole colony begins to emigrate along one of the day’s principal raiding
trails. Moving in solemn procession, sometimes through most of the night,
the colony finally settles in a new bivouac, or temporary camp, often under
a low-hanging branch or vine. After about two weeks, however, the colony’s
behavior abruptly changes. It appears to quiet down and enter a statary phase.
Few workers go out on raids; when they do, the forays are much smaller. No
longer do nightly migrations occur. Remaining at the same site for about three
weeks, the colony acts as though it had gone partially dormant. Then, just as
abruptly, a new nomadic phase of intense foraging activity begins.
What factors underly such spectacular phasic behavior? Walter Schneirla
decided to investigate. At the time, most scientists accepted a straightforward
and seemingly logical explanation: depletion of food supply. The army ants,
they argued, simply stay in one place until the food supply is exhausted and
then move on to new hunting grounds. However, an alternative explanation
also had some appeal. Perhaps the cycles were cued to some environmental
phenomenon, such as phases of the moon, or perhaps changes in temperature,
humidity, or air pressure.
Detailed observational data on colonies in the field were sorely needed.
Schneirla began following a single colony, then another and another, through
one or more complete cycles. Painstakingly, he logged a dozen armies through
whole cycles and more than a 100 more through partial cycles, taking down
data in the field and repeatedly sampling the internal colony composition at
various stages of the behavioral cycle. One thing quickly became clear. Within
a single environment, several colonies were often present; of this number,
some were generally in the nomadic phase, others in the statary phase. This
seemed to rule out major environmental factors as the determinant of the
Fig. 10.15 Raiding ant cycles. The alternation of statary and nomadic phases in the colony cycle
of the army ant, Eciton burchelli. Nomadic phases are triggered by worker–callow interactions
and maintained by worker–larval interactions. As cocooning of mature larvae begins, intensity of
mutual stimulation between adults and brood declines temporarily and the colony lapses back into
the statary condition until the pupal brood emerges
nomadic cycle. The second hypothesis was more difficult to substantiate or

reject, but persistent observation eventually provided a clue. Schneirla
was able to confirm that a nesting site vacated by one colony would some-
times be moved into by a second colony the following night; in some cases
the newcomers would remain for the whole three-week statary period. These
observations seemed clear evidence that the food supply around the bivouac
had not been depleted.
This, of course, left Schneirla without an established hypothesis for the
colony cycle. As he began to analyze brood samples from different stages in
the behavioral cycle, however, a new picture began to take shape, implicating
the breeding cycle within a particular colony as the determinant of its cyclical
behavior. The nomadic phase, it appeared, always coincided with the period
when a larval brood was developing in the colony. The statary phase, on the
other hand, always seemed to begin at the point when mature larvae started to
spin their cocoons. In fact, the statary phase coincided with a point in which
only pupae or newly laid eggs were present, neither of which required daily
feeding by the workers.
A sentimental explanation was this: while helpless young exist, the work-
ers protect them with a fixed bivouac, but once eggs hatch and pupae eclose,
the workers must suddenly feed this massive number of new mouths, so
they become nomadic. Although this might explain the adaptive significance
of such behavioral cycles, it does little to elucidate proximate mechanisms.
Impressed by the role of trophallactic interactions among all social insects,
Schneirla turned to that phenomenon for his answer. Although by the end of
his first trip to the tropics he could only phrase it loosely, Schneirla was able to
theorize that the regular fluctuations in Eciton colony activity were regulated
by some sort of stimulative interactions between the colony’s large developing
broods and the adult workers. What might these ‘stimulative interactions’ be?
In the field and in the laboratory, Schneirla watched as workers cared for
pupae of different ages. During maturation in their cocoons, pupae appeared to
become progressively more attractive to the workers; these handled them more
frequently, even partially tearing the cocoons open. Upon emergence, the new
adults (callows) became the objects of much licking, stroking, and handling
by the older adults; in turn, the callows themselves became quite hyperactive.
This reciprocal stimulation resulted in a rising crescendo of excitement that
soon seized the whole community and led invariably to colony emigration
initiating the nomadic phase.
Although he had no evidence of pheromone involvement at this time,
Schneirla had no trouble calling this a trophallactic interaction because he
interpreted trophallaxis broadly to include not only the actual exchange of
food but the mutual exchange of all ‘equivalent’ stimuli, including touch.
Whatever the mode, some sort of bond among colony members was obviously
being forged. Colonies from which he removed all callow brood appeared
lethargic in contrast and invariably failed to initiate nomadic behavior at the
expected time.
But while trophallaxis between workers and callows was appealing enough
in theory, Schneirla’s observations at this point did not rule out a second pos-
sibility, namely, that the necessity to feed developing larvae and new adults
was causative. In fact, during later stages of the nomadic phase when the lar-
vae had grown and developed voracious appetites, they seemed to become the
source of the colony ‘drive.’ To test the role of larval brood in stimulating the
workers’ activity, Schneirla split a colony into two parts of comparable size.
In one group he left the larvae intact; in the other he removed them. The work-
ers in contact with larvae continued to show considerable activity, but those in
the broodless portion were much less active. In another experiment, Schneirla
removed the entire larval brood from a colony that was in the nomadic phase.
True to expectations, the colony stopped emigrating, and the intensity of its
daily raids diminished.
Army ants display a remarkable syndrome of traits—nomadism, obligate collec-

tive foraging, and permanently wingless queens whose abdomens alternately swell
to enormous proportions and then shrink. What controls the queen’s reproductive
cycle? Schneirla postulated that the queen was stimulated to feed in excess because
of the intense worker activity that occurred when mature larvae began to reduce
feeding demands prior to spinning cocoons, thus freeing an abundance of food.
The brood cycle may well serve as an endogenous pattern generator that triggers
a cyclical cascade of physiological changes in the queen.
10.3.2 The Eusocial Wasps
Eusociality in wasps is almost wholly restricted to some members of a single

family, the Vespidae. The paper wasps (Polistinae) and the yellow jackets and hor-
nets (Vespinae) are entirely eusocial; eusociality is facultative in the hover wasps
(Stenogastrinae) of the Old World Tropics. All of the most commonly encoun-
tered eusocial vespids construct nests of paper made from plant fibers chewed off
weathered wood and/or dead plants, but hover wasp nests are often constructed
of mud.
The nests of eusocial wasps are begun by mated females called foundresses. Life
on a social wasp nest is a matter of uneasy and shifting alliances, so no consideration
of social wasp behavior is complete without a discussion of dominance, a rank-
ing of individuals on the basis of real or apparent authority, strength, or influence.
The dominance concept implies an ability to recognize individuals and remember
relationships, and results in a ranking called a dominance hierarchy. Particularly evi-
dent in species with long-lived adults, dominancy hierarchies can be found not only
among social wasps, but among dragonflies, crickets (see Fig. 8.10), and passalid
beetles, among others.
In the life of a social wasp nest, a colony cycle is the period of development that
lasts from the end of one reproductive episode to the end of the next. In temperate
species nests are normally annual; one foundress or a group of them starts a nest,
which grows and expands until late summer or autumn, when a reproductive phase
culminates in mated females overwintering to start the cycle anew the next spring.
Polistes paper wasps provide a good example. Sometimes, a single queen must find
a new nest location on her own, and bear the risk of nest building and foraging for
herself and her first batch of young. In other cases, groups of females join together
in the spring to begin their nests, but each fertile potential queen then competes in
laying eggs, and soon relatively rigid hierarchies are established. Dominant Polistes
maintain their reproductive superiority by three means: laying the greatest num-
ber of eggs, physically removing and eating the eggs of any subordinates that may
have succeeded in ovipositing, and demanding and receiving the greatest share of
food. Interactions between dominant and subordinate individuals are often matters
of posture; the dominant individual rises on her legs above the subordinate, which
crouches and lowers its antennae (Fig. 10.16). The dominant female also performs
a conspicuous side-to-side vibration of the abdomen (tail wagging) more frequently
than other individuals do. Polistes castes are difficult to distinguish morphologically
but they are behaviorally clear. During the first half of the summer, only workers
are produced. In late summer and early fall, sexual forms (males and reproductively
competent females called gynes) start to appear; concurrently, the colony slowly
declines, finally abandoning the nest entirely.
Fig. 10.16 Bullying nestmates. Dominance and subordinance behavior in the paper wasp Polistes
fuscatus. The female on the left shows the relatively elevated posture characteristic of a dominant
individual. The subordinate (right) has been seized at the hind leg and crouches with antennae
lowered
For tropical social wasps, on the other hand, seasonal constraints may be less rig-
orous, and in many species the components of the colony cycle seem to have become
dissociated. Several cycles of brood may be produced before a reproductive brood
is produced, a nest may be used for variable lengths of time of up to several years
before it is abandoned, or a colony may routinely occupy two nests successively
before producing any reproductives. Many of these tropical species start their colony
cycle with swarm-founding. A swarm can be thought of as a reproductive unit that is
essentially a colony without a nest or brood. In honey bees, it consists of a hive’s old
queen and about half of the workers; the remainder stay behind with a new queen
and the brood that are already in the hive. In swarm-founding wasps, it consists of a
group of reproductives that disperse, leaving a fraction of the population back at the
nest to repeat the colony cycle.
There are many advantages to swarm-founding, including reduced risk for the
queens. Rather than scouting for a new nest site, they remain in the security of
the natal nest while workers take on this job, and are exposed only while flying to
the site. In addition, they never have a period where they must take on the risks
of foraging. Finally, having a large worker force from the start provides the strong
defense that is needed against predators, in particular ants, a formidable enemy force
in tropical regions (Fig. 10.17).
Nests covered by envelopes are universal to the second group of eusocial
wasps, the vespines, a predominantly temperate group that includes the hornets and
yellowjackets. Their annual nest cycles (Fig. 10.18) are basically similar to that of
Polistes, except that the queen and worker castes are morphologically distinct. Most
species of yellowjackets construct subterranean nests; hornet nests are typically
aerial or built within hollow trees. Most wasps feed upon a variety of arthropods
Fig. 10.17 Defending the fort. (left) Polybia occidentalis workers rear back and buzz their wings
in short bursts, directing intermittent blasts of air at two fire ant (Solenopsis) foragers that are
attempting to climb on the nest. Ants that face this windstorm usually retreat, If an ant does
pass the defenders, a wasp will grasp the ant in her mandibles, carry it away from the nest, and
drop it. (right) ‘Ant guards’ made by the hover wasp Parischnogaster on a rootlet above the nest
attachment. The substance is secreted by the Dufour’s gland and applied to the substrate
Fig. 10.18 Seasonally changing. The annual colony cycle of a yellowjacket wasp (Vespula)
includes solitary and colonial phases. In early spring fertile queens awaken from hibernation and
alone initiate nests by constructing a few paper cells hung from the top of a preexisting soil cavity
such as an abandoned rodent burrow (upper left). After the first workers emerge the colony grows
rapidly through the summer, with the nest ultimately reaching the size of a basketball. By late sum-
mer, cells of the bottom combs (here two) are distinctly larger than those in the uppermost combs.
These bottom cells give rise to males and virgin queens, while the cells in the upper combs produce
only workers. From this point the nest declines. Almost always, only the newly produced queens
overwinter to repeat the cycle the following year
caught alive and carried back to the nest piecemeal in their jaws, but some yel-
lowjacket species become nuisance pests, scavenging around picnic and recreation
areas, especially in late summer.
The Stenogastrinae, the third social vespid group, are delicate slender wasps
with hovering flight that are denizens of rainforests of the Indo-Pacific region. The
approximately 60 species are poorly known, and their colonies never become very
populous.
Outside the Vespidae, almost no other wasps are eusocial. However, a group of
four genera in the family Crabronidae are apparently unique in producing silk from
abdominal glands that have not been found elsewhere in this wasp family. The silk
is used in nest construction, and more than one adult is commonly found in these
nests, suggesting the possibility of social behavior. The most completely studied
species, a tiny Central and South American wasp called Microstigmus comes, preys
exclusively upon Collembola. Inside its pendent bag-like nests (Fig. 10.19) as many
as 11 females cooperate in nest defense and cell provisioning; genetic evidence
indicates that colonies include a mother and her daughters, with one female having
well developed ovaries.
10.3.3 The Bees

Few insects are more familiar than the ubiquitous honey bee. Elsewhere, we have
discussed its unique dance language, its ability to sense the earth’s magnetic
field and other sensory capabilities, and the multiple roles of the queen substance
pheromone in colony life. At this point it is appropriate to put the honey bee into per-
spective as but one of several hundred species of eusocial bees, surrounded by many
times that number of relatives showing all manner of social behaviors, because the
bees, more than any other insect group, display the full spectrum of social evolution.
Bees have arisen from a different part of the Hymenoptera phylogenetic tree than
ants and vespid wasps, but are so like crabronid wasps that they can be regarded as
being simply a sort of wasp that has specialized on collecting pollen instead of insect
prey as larval food. For over 50 million years bees have evolved in close contact with
the angiosperm plants upon whose flowers they depend. Social behaviors of various
kinds crop up repeatedly, but eusociality appears independently in two of the nine
families of bees recognized today—the Halictidae and the Apidae.
The Halictidae are a behaviorally diverse family of over 5,000 species of mostly
small insects that are often called ‘sweat bees’ because of their often-annoying habit
of lapping up perspiration from people’s skin during hot weather. In many areas of
the world, they are more important elements of the bee fauna than most people
recognize; for example, in one study, halictid bees accounted for up to half of all
the individual bees collected on flowers at one site. Most halictid bees nest in soil; a
few nest in rotting wood. Many halictid bees nest in aggregations that can be dense
(150–200 nests per square meter have been reported) and sometimes long-lasting
(35 years in one report). Because aggregations are easier to find than individual
nests, they are better studied. They also are easier for predators to locate; in a study
of Halictus rubicundus, cell mortality was twice as high for aggregated nests as for
isolated nests.
Social behavior is extremely variable in sweat bees. Solitary, variously social,
and parasitic behaviors occur not only among lineages but sometimes among and
within populations. Three independent clades in the family show eusocial behavior.
Fig. 10.19 Crafting with silk. Thumb-sized nest constructed by the eusocial crabronid wasp,
Microstigmus comes. The nest is formed of plant fibers bound together by silk thread produced
by unique glands found in the tip of the adult female’s abdomen (arrow). The plant fibers are first
loosened and scraped from an area on the underside of a palm leaf (a, b), then gathered together
to form an undifferentiated mass which is bound with the silk (c–e) and then slowly lowered on
a silken thread pedicel (f). The pedicel coil and entrance hole at the base of the pedicel are then
added and the upper portion of the interior hollowed and prepared with the first cell (g–i)
Within certain sweat bee colonies, even if there are only two individuals, one is a
female that has mated, has developed ovaries, lays eggs, eats others’ eggs, dominates
others, rarely forages for pollen and nectar, and is the oldest or largest; the other
has not mated, has underdeveloped ovaries, lays few eggs and has these eaten, is
dominated, frequently forages or guards the nest, and is younger and/or smaller
than others.
Whether halictid bees were once fully social and are evolving toward solitary
behavior or vice versa continues to be a matter of academic debate. As might be
expected with such a large group, the majority of sweat bees are still unknown
behaviorally, and information is limited for many others. Furthermore, most of the
species that have been studied are from the temperate zone, where environmental
correlates such as flowering synchrony may influence social biology.
The family Apidae includes many eusocial members. The true honey bees in the
genus Apis are the best known, but another group shares many behavioral similari-
ties with them. The Meliponini, or stingless bees are among the most conspicuous
and numerous bees in these tropical regions. Like honey bees, these bees live in
perennial colonies that produce new colonies by swarming; individual queens or
workers cannot survive alone for long. Furthermore, both have queen and worker
castes that differ, behaviorally, physiologically, and morphologically. They both
develop large populations, up to 180,000 workers in some stingless bee species.
Colony integration and communication are necessarily complex, and both of these
bee groups have evolved elaborate means of communicating the location of food
sources (see Chapter 8). Both honey bees and stingless bees also build nests with
wax secreted from abdominal glands. Larvae of both are reared in individual cells
arranged in combs within the nest, while separate cells are used to store honey
and pollen (Fig. 10.20). However, honey bees and stingless bees belong to differ-
ent clades, and do show other differences. One major difference is that meliponine
larvae are mass provisioned; those of Apis are fed progressively. Another is that
honey bees defend their nest by stinging; meliponines defend theirs by aggressively
biting intruders.
Another social group familiar to almost everyone is the genus Bombus, which
includes the robust hairy bees commonly called bumblebees. Bumblebees are char-
acteristic of cool north temperate climates around the world; with few exceptions,
they have not succeeded in invading the tropics, except at high elevations where
they may become very abundant locally. A behaviorally cohesive group, all of the
more than 250 Bombus species have relatively small annual colonies (Fig. 10.21)
founded by a single overwintered queen. Abandoned rodent burrows are preferred
nest sites; inside, a queen fashions her nest from wax secreted by abdominal glands.
Rather than directly feeding her brood as do other eusocial groups, the bumblebee
queen characteristically stocks pollen and honey separately from the brood cells in
special wax storage pots. She then lays several eggs together in a single distensi-
ble wax brood cell; thus, the bumblebee larvae are reared in groups but since they
are contained in a capped cell with their entire food supply, there is no trophallaxis
contact between parent and offspring.
Finally, some unusual eusocial behaviors are found among the so-called allodap-
ine bees. Unlike the nests of all other known bees, those of allodapines have no
cells. Larvae are kept together on the nest floor like ant larvae, and like the young
of ants they are moved about the nest by the mother and arranged in groups accord-
ing to age (Fig. 10.22). Allodapine larvae are fed progressively, being given more
frequent pollen meals as they grow, a behavior that promotes mother–offspring con-
tact. In most species, the mother dies before her young emerge, but in some species
Fig. 10.20 ‘Beeing’ stingless. One of the species of stingless honey bees of the genus Trigona
from Costa Rica builds large aerial nests (right) of secreted wax mixed with large amounts of resin
to form a tough material called cerumen. Inside the nest, pollen and honey pots (left, below and
above respectively) fill the brood chamber. While Trigona cannot sting, they actively defend their
nests by vigorously biting intruders
the mother’s life overlaps the adulthood of her offspring, who often remain in the
parental nest and help rear a second generation.
10.3.4 The Termites

A cosmopolitan but predominantly tropical group, termites include more than 2,600
species. All of them are eusocial. Termitidae is by far the largest family, and contains
the most advanced species. Because termites mostly feed on dead plant material,
they are of considerable ecological significance, particularly in subtropical and
tropic regions; a significant number also are economically significant as pests that
can cause serious structural damage to buildings, crops or plantation forests. Often
the first external evidence of infestation may be when they swarm (Plate 45).
Along with the closely related Cryptocercus cockroaches (see Plate 39), termites
are the only wood-eating insects that depend on symbiotic intestinal protozoans
and bacteria to break down energy-rich but nitrogen-poor cellulose into usable sug-
ars (see Chapter 4). It appears that termite societies began as feeding communities
and only later evolved social brood care, in a sequence that is the reverse of social
evolution in the Hymenoptera.
Termite nests are essential to termite life, because termite workers have a soft
cuticle and are easily desiccated if exposed directly to the outside environment. A
Fig. 10.21 Living in a mouse nest. A colony of the European bumblebee, Bombus lapidarius.
The larger individual at the lower center is the queen, resting on a cluster of pupae-containing
cocoons. At the upper left are several communal larval cells. Open cells contain pollen or honey
Fig. 10.22 Developing progressively. A nest of an allodapine bee, Exoneura, showing eggs and
developing brood grouped together in an opened pithy plant stem. Larvae are fed progressively on
small pollen balls brought in by cooperating female bees
Fig. 10.23 Ventilating the fort. (left) Environmental control in a termite nest is evident in this
cutaway view of the interior of the nest of Macrotermes bellicosus, an African fungus-growing
termite, that shows the temperature and percentage carbon dioxide concentration at different posi-
tions. Metabolic heat from the huge biomass of colony members is concentrated in the central
core of the nest, rises by convection to the large upper hollow cavity, then diffuses toward the
sides where it flow into a network of narrow channels close to the surface; here the air is cooled,
and gaseous exchange occurs by diffusion through the thin dry walls. Refreshed air then sinks to
the lower passages of the nest and eventually recycles. (right) Cathedral mound of Nasutitermes
triodiae from northern Australia
typical colony contains eggs, nymphs, workers, soldiers, and reproductive individ-
uals of both genders, plus one to several egg-laying queens. Nests are punctuated
by a maze of tunnel-like galleries that effectively ameliorate air temperatures
(Fig. 10.23) and control carbon dioxide/oxygen balance, as well as allowing the
termites to move through the nest; foraging is typically done along radiating cov-
ered tunnels built out from the nest. Constructed of soil, excrement, and saliva, nests
vary in location from completely subterranean, to inside wood, to entirely arboreal.
The most conspicuous nests are begun underground and as they grow begin to pro-
trude progressively, eventually assuming impressive species-characteristic shapes.
Although termites are sometimes called ‘white ants’ they are not at all related to
ants. In a very literal sense, termites are ‘social cockroaches’ that have attained their
eusociality from a base extremely remote in evolution from the Hymenoptera, yet
surprisingly the differences between these eusocial groups are not as great as one
Table 10.2 Major distinctions between termites and eusocial Hymenoptera
Termites Eusocial Hymenoptera
Similarities
1. Caste number and kind are analogous, especially between termites and ants.
2. Trophallaxis occurs, and is an important mechanism in social regulation
3. Chemical trails are used in recruitment; comparable trail laying and following behaviors
4. Inhibitory caste pheromones exist, and act in similar ways
5. Frequent grooming between individuals transmits pheromones
6. Nest odor and territoriality are of general occurrence
7. Comparably complex nest structures have precise temperature and humidity regulation
8. Cannibalism is widespread, but not universal
Differences
1. Castes in ‘lower’ termites are 1. Castes are determined chiefly by nutrition; pheromones
determined chiefly by sometimes play a role
pheromones; sex and other
factors are involved in some
‘higher’ termites
2. Worker castes consist of both 2. Worker castes consist of females only
sexes
3. Larvae and nymphs contribute to 3. Larvae and pupae are helpless and almost never
colony labor, at least in later contribute to colony labor
instars
4. No dominance hierarchies occur 4. Dominance hierarchies are commonplace but not
among individuals in the same universal
colonies
5. Social parasitism between 5. Social parasitism between species is common and
species is almost wholly absent widespread
6. Anal trophallaxis occurs in all 6. Anal trophallaxis is rare, but trophic eggs are exchanged
‘lower’ termites, but trophic in many bees and ants
eggs are unknown
7. Primary reproductive male stays 7. One or more males mate with the queen during the
with queen after nuptial flight, nuptial flight and die soon afterward; female stores
helps construct first nest. No gametes in spermatheca for up to several years
mating during nuptial flight;
mating occurs periodically as
colony develops.
8. Diploid sex determination 8. Haplodiploid sex determination
might expect (Table 10.2). The most striking differences in the two groups con-
cern the nature and care of their young. The Hymenoptera are holometabolous,
and because the larvae of eusocial Hymenoptera are helpless grubs that must be
nursed continuously, their care requires a great deal of adult labor. Termites are
hemimetabolous, so the young stages of termites, on the other hand, are active crea-
tures quite similar to the adults and quite able to fend for themselves, particularly
among the so-called ‘lower’ termites. Many termite workers are nymphs that may
eventually develop into winged adults. Others, termed pseudergates, are workers
that for chemical reasons will probably never grow up, although they retain the
potential for developing into any of the castes. Thus, up to a point it may be said
that termites rely upon ‘child labor’ for the maintenance of their societies. Another
difference is that the male sex in termites participates in colony labor, which is never
true for male hymenopterans.
10.3.5 Lesser Known Candidates

Though ants, wasps, bees, and termites are the best known members of the insect
social register, new members continue to seek admission as studies turn up eusocial
behaviors in diverse groups. A small number of thrips and aphids—two unrelated
groups of small insects that suck plant juices—plead a specially convincing case for
inclusion.
Thrips (the name is both singular and plural) are small hemimetabolous insects in
the order Thysanoptera; many are serious pests upon hundreds of different crops. In
about 300 species in the suborder Tubulifera, a single female (or sometimes a male
and female together) alter plant growth to form galls, in which they live and feed;
another dozen species do not actually form galls, but have other ways of making
closed spaces in which to feed on plant tissues. Females in some of these species
develop extreme physogastry and lay hundreds of eggs. In many species the life
span of gall foundresses overlaps with that of their adult offspring. Interestingly, all
bisexual thrips species are apparently haplodiploid, like Hymenoptera.
The dwellings that thrips produce on Australian acacias are often contested by
other thrips species that do not build domiciles of their own. Some of these species
simply take over abandoned galls, but at least one species fights and sometimes kills
the domicile owner. As though in response to such threats, some gall-forming thrips
have what amounts to a soldier caste; some adult males and females have enlarged,
heavily sclerotized forelegs (see Plate 40) that are lacking in other group members.
These soldiers are often (but not always) sterile, and have smaller wings than the
other thrips.
Aphid societies consist of clones established parthenogenetically from a single
foundress, and the soldier caste appears early in development in the larval rather
than the adult stage. Unlike the haplodiploid thrips, aphids are colonial diploids.
Given their clonal condition one might expect most aphids to be eusocial, but in
fact eusociality is mostly restricted to some members of just two closely related
families, which also happen to be the two groups capable of inducing galls on their
host plants. Thus whereas not all gall-making aphids are social, all (known to be)
social aphids are gall-makers. Some 40–50 aphid species have a life history that
roughly parallels the gall-making thrips, including development of a soldier caste.
One’s mental picture of aphids as docile ‘plant lice’ or tiny cow-equivalents
tended by ants (see Plate 10) needs some readjustment in the face of these aphid
soldiers—fierce defenders that patrol the colony, grappling with intruders and stab-
bing with their short thick beaks. Recall that even non-soldier aphids are not
defenseless; many secrete chemicals from their cornicles that incite clone mates
to quickly scatter (see Fig. 6.14). In some social forms, the cornicle secretions
recruit soldiers to confront a threat. Galls clearly are resources that are generally
worth defending; in some cases, tending ants are also part of the ecological mix;
one study showed that excluding ants dramatically increased the aphids’ soldier to
non-soldier ratio.
One more unusual social insect that deserves mention is the Australian beetle
Platypus (=Austroplatypus) incompertus. Unlike most bark beetles that attack dead
or dying hosts, this ambrosia beetle attacks the heartwood of living Eucalyptus trees,
forming galleries that sometimes persist for several years. Each gallery houses a
cooperative family group composed of a single fertilized female, accompanied by
several unfertilized females that maintain and patrol the galleries and process the
mycelia of ambrosia fungi, ensuring a rich food supply.
10.4 Implications and Correlates of Social Life

Evolutionary biologists look for finely tuned adaptive behavior in the animals they
study, and are happy when they find it. Insects have a long evolutionary history,
and most of their behavior patterns apparently do, as well. However, sometimes a
seemingly imperfect system can be even more interesting and in puzzling about it
scientists can be led to insights that might never have arisen otherwise.
Such is the case with golden egg bugs (see Plate 41), a group that has captured
the attention of a number of biologists, including Arja Kaitala and his students and
the same Robert Smith who studied Abedus water bugs (see Case Study 10.1). Most
Hemiptera oviposit on foliage or other environmental objects, and very few tend
their eggs in any way. Golden egg bugs and giant water bugs are only distantly
related, but their convergence in a behavior as unusual as egg-carrying fairly begged
for comparative study (Case Study 10.3).
Case Study 10.3: Egg-Carrying in the Golden Egg Bug,

Phyllomorpha
If both sexes of an insect carry eggs, and the eggs are almost never their own,
can it be parental care?
Phyllomorpha laciniata are plant-sucking insects with a bizarre appearance
(see Plate 41) and equally strange behaviors. In most Mediterranean countries,
a careful observer who managed to spot these little coreid bugs (which are
actually camouflaged by their elaborate leafy spines) would find both sexes
to be wandering around on vegetation, many with variable numbers of small,
golden eggs fastened to their backs.
Do the bugs deliberately attach their eggs to other bugs? Is such behav-
ior obligatory? Is either sex more likely to carry eggs? What is the carrier’s
genetic relationship to its eggs? Do carriers even know the eggs are there? Do
carriers show any other behaviors indicative of care toward their eggs?
10.4 Implications and Correlates of Social Life 429
Field observations and collections by members of Arja Kaitala’s research

team established the basics. The bugs they caught each held about five eggs,
but numbers varied from one to 30; most of the eggs were glued on the
bugs’ backs, but sometimes the researchers found them on undersides, legs,
heads, and even antennae. Males carried only slightly more than two-thirds of
the eggs; females carried the rest. Laboratory observations and experiments
revealed a portrait of female golden egg bugs eager to lay eggs on any conspe-
cific, regardless of its sex. In most situations, female potential recipients could
resist by simply avoiding other females. Males were in a more difficult bind.
They persistently courted females, but were often unsuccessful in procuring
copulation; instead, they ended up with the female’s egg attached somewhere
on their body. When they did successfully mate, the pair remained coupled for
long periods, often from several hours to an entire day. Other mated females
sought out such pairs, and took advantage of the situation by laying eggs on
both participants.
Knowing that close relatives of the golden egg bug lay their eggs exclu-
sively on their host plant, Smith and Kaitala compared the reproductive
condition of field-captured and laboratory-held female P. laciniata. Those
taken directly from the field had several eggs in their oviducts; those that spent
time in the laboratory enclosure had few to none. Apparently the crowded con-
ditions of the enclosure allowed females to quickly unload their eggs. In the
natural setting, they could have simply dumped their eggs on vegetation but
instead they were retaining mature eggs until they could locate conspecific
carriers.
Why should carrying be so important? Another member of their team,
Xavier Espadaler, provided the answer. Seven ant species with a shared
taste for golden eggs were frequent visitors to the host plant. When the
researchers removed eggs from bugs and placed them on host plant flowers,
ants discovered and ate the eggs within minutes or hours.
Having identified a major benefit, the researchers began looking for the
costs of egg-carrying, and soon found them, but not quite where they expected.
Reasoning that the added weight of the eggs might slow bugs down, they
conducted bug-race experiments; bugs bearing eggs were statistically just as
likely to win as unencumbered bugs were. Instead, the major cost seems to be
that bugs with golden eggs are more conspicuous to predators, and probably to
parasitoids as well. Gangs of one particularly aggressive ant species attacked
egg-laden bugs, and often killed them.
Might egg-carrying in P. laciniata be a case of parental care, perhaps anal-
ogous to that in giant water bugs? For several reasons, Smith and Kaitala
thought it was not. By definition, parental care requires parents and their
offspring, a bond that was strangely lacking in golden egg bugs. Females,
of course, cannot be the mothers of the eggs they carry; they do not lay
eggs on themselves. Furthermore, there is little reason for golden egg bug
females to show preferences in terms of the sex of the carrier or the degree
of relatedness between egg and carrier simply because any conspecific car-
rier probably increases the chances that eggs will survive by about the same
amount. However, unlike the situation in water bugs, golden egg bug males
did not appear to be carrying their own offspring either. Samples sent to
Stockholm University for molecular genetic studies showed that egg-laden
males had fathered only a very small percentage of the eggs they carried.
Furthermore, hundreds of hours of observation did not reveal any type of
parental nurturing analogous to the behaviors of male water bugs.
Thus, Smith and Kaitala concluded, although it has obvious benefits, egg-
carrying in Phyllomorpha laciniata appears to be a purely passive behavior,
rather than parental care, at least at this point in time. Perhaps, however, its
imperfection is a clue that we are being given a rare glimpse at something we
almost never see—an evolutionary work in progress.
Three aspects of Phyllomorpha laciniata biology—flexibility of oviposition

behavior, the fact that males carry most of the eggs, and the demonstration that
males carry both genetic offspring and unrelated eggs—have made this system the
subject of continuing controversy. However, García-González and colleagues make
a strong case that sperm mixing is the key to explaining golden egg bug behavior.
In most insects, ‘sperm precedence’ is presumed to occur, so that the most recent
male to mate with a female fertilizes most of her eggs, but in P. laciniata, their
studies show that the last male to copulate with a female sires only about 43% of
her eggs, and this percentage does not change significantly over a period of several
days. Because females lay single eggs at intervals for four to five months rather than
producing clutches, they could be expected to benefit by promoting a sperm com-
petition mechanism that would maximize the number of males accepting eggs, and
the context in which they are likely to accept them. Sperm mixing ensures that each
male that has copulated with a female has some chance of fathering her offspring,
but he also has no clue as to when his own offspring will be produced. Despite this
lack of confidence in paternity, male egg-carrying is favored because males do not
suffer mating costs when they carry eggs, and the survival of eggs (including per-
haps their own) is so much higher when eggs are carried rather than being left on
plants.
10.4.1 The Ecology of Parental Care
Over the past seventy years, it has become increasingly apparent that the evolution of
parental care—an important behavior both in and of itself, and as a key component
of more comprehensive social life—correlates with many aspects of an organ-
ism’s ecological milieu. A number of general aspects of the physical and biological
environment have been identified as being particularly important, not only among
insects but also among a wide variety of vertebrates that practice parental care. It has
been suggested that two extremes are likely to push animals toward parental care:
a very stable environment and an unusually harsh one; let us follow the reasoning
behind this idea.
The first of these, a stable, predictable environment, favors the evolution of
parental behaviors because the K-selection (see Chapter 9) that tends to prevail in
these conditions has certain demographic consequences: the animal tends to live
longer, to grow larger, and to reproduce at intervals. As an example, consider rotting
logs on a forest floor; these would seem to fit the stable, relatively constant environ-
ment criterion. Two lineages of cockroaches have evolved extended bi-parental care
of slow-growing, highly dependent offspring in such an environment; the most thor-
oughly studied exemplars are members of the genera Panesthia and Cryptocercus
(see Plate 39).
Extended bi-parental care has also evolved independently in Odontotaenius
(=Popilius) disjunctus, a large, glossy black beetle in the family Passalidae (see Fig.
10.12). The young of most cockroaches and beetles can forage independently from
an early age; they are analogous to many types of precocial birds that begin to peck
and feed themselves soon after hatching. The young of these three wood-feeding
groups of insects are more like the altricial young of robins and wrens—reared in
small slow developing clutches, highly dependent on their parents to provide food
and protection, and unable to forage for themselves. Both of the wood-feeding cock-
roaches require the help of bacterial symbionts to digest their cellulose diet; the
parents produce bacteria-rich soups to inoculate the digestive systems of nymphs
that line up like pigs at the food trough to imbibe the hindgut fluids. In contrast,
unlike most other wood-consuming species, bess beetles apparently lack digestive
symbionts. Instead, an important part of their diet is fecal frass, which acts as a
substrate for bacterial and fungal development and is reingested. Adult bess bee-
tles live two or more years; between rearing their yearly generations of young,
they are somewhat gregarious, many adults being found in the same tunnel sys-
tem. However, during the reproductive season, each pair of beetles maintains its
own tunnel system in which 20–60 eggs are laid. The developing young feed on
material prepared by their parents, and when ready to pupate they cooperate with
the adults in construction of pupal chambers.
A second circumstance that may propel an animal toward parental care is pen-
etration of a physically stressful environment where offspring would require some
sort of protection, at least during the most vulnerable period of their development.
In extremely harsh environments, many insect species guard their offspring. For
example, the staphylinid beetle, Bledius spectabilis, occupies intertidal mud; the
female of this beetle constructs a bottle-shaped burrow in which eggs are laid. She
remains with her eggs and young, twice each day plugging the burrow entrance to
prevent the tide from entering, then reopening it to prevent the young from suffo-
cating; this care continues until the larvae complete their first instar and disperse to
establish their own individual burrows. Eggs in burrows from which the mother was
removed all perished, either to mold or to predation by other beetles.
Many ecological factors are probably involved in every insect’s life, and there is
often considerable challenge in finding what factors actually lie behind an observed
case of parental care. For example, adults of the tropical stink bug Antiteuchus
tripterus possess potent chemical defenses and are aposematically colored. They
generally live relatively long and are able to produce young two or three times dur-
ing their life. However, their offspring start life quite vulnerable to predation and
parasitism because their stink glands do not become functional until they reach their
second instar. William Eberhardt’s studies of maternal care in this species (Case
Study 10.4) showed how parental behavior and ecology are well correlated, as pre-
dicted, with risks of predation and parasitism acting as important selective factors.
Case Study 10.4: Parental Care in the Stink Bug,

Antiteuchus tripterus
Stink bugs are aptly named. Their distinctive odor, derived from abdominal
stink glands, is familiar to most everyone. Less well known is the extent to
which some mother stink bugs care for their young.
Antiteuchus tripterus is a tropical pentatomid bug common on a variety
of landscape trees and shrubs in urban areas of Cali, Columbia. It lays its
barrel-shaped, unusually thin-shelled eggs, in compact masses almost always
numbering 28, on the undersurface of leaves. For 15–16 days following
oviposition, females remain with their clutch without feeding while the eggs
develop and hatch and the nymphs grow to reach the second instar. Such
parental behavior is not uncommon in the Pentatomidae, and representatives
of several genera have been reported to have similar behavior. Why? William
(Bill) Eberhard reasoned that the parent bug was defending her eggs against
predators. His first experiment was simple. He removed brooding females
from some batches of eggs while leaving others nearby untouched to serve
as controls. Not one bug survived from 48 unguarded egg masses, most of
which vanished, whereas 50% of the protected eggs produced bugs! Later
observations revealed that several species of insects, especially foraging ants,
fed on undefended eggs. Thus, the females’ defense of their eggs against such
‘generalized’ predators appeared to be highly effective, and necessary.
Nevertheless, a 50% mortality of the guarded eggs seemed surprisingly
high, considering the time investment of the mothers. Eberhard presented arti-
ficial insect-sized models to brooding females. These elicited a repertoire of
defensive behaviors, including waving antennae in the direction of the stim-
ulus, tilting the body to form a shield between the threat and the nymphs,
general body ‘shuddering,’ scraping of the front legs along the periphery of
the egg mass, and kicking backward with the middle and hind legs. Although
like all other stink bugs Antiteuchus possessed glands which could spray a
highly repellent chemical, this defense was never employed against the insect-
sized stimulus. Eberhard assumed that such chemical defense was reserved
for larger potential predators, for when the body was tilted for spraying, the
bugs also vividly displayed aposematic orange bands along the sides of their
abdomens.
Attempting to rear the eggs, Eberhard discovered that a number of them
yielded, instead of bugs, tiny scelionid parasitoid wasps of two different
species. Why were they so successful despite constant maternal vigilance?
Stereotypy in maternal orientation provided a clue. Antiteuchus females con-
sistently aligned their bodies parallel with the axis of the leaf upon which their
eggs were laid, generally facing toward the leaf tip. Observing them care-
fully when their overt behavior (kicking, scraping, etc.) indicated they were
aware of the presence of a wasp attacking their eggs, Eberhard was unable
to document a single instance when a mother bug turned and assumed a new
orientation over its eggs. The consequence of this constant maternal orienta-
tion behavior (Fig. 10.24) was that, while females’ defenses did effectively
deter many wasp oviposition attempts in eggs situated on the front and sides
of the egg mass, eggs to her rear were not effectively guarded. Analysis of the
distribution of parasitized eggs within the mass confirmed this.
Fig. 10.24 Defending against attack. A female Antiteuchus tripterus in characteristic guarding
position over her eggs. As the bug lowers her antennae and scrapes with her front foot along the
egg mass in an attempt to drive off Phanuropsis, a parasitic wasp which attacks from the front, a
Trissolcus wasp oviposits in an egg at the poorly guarded rear edge
Would wasps attack more if eggs were not guarded at all? To exclude the
foraging ants and other walking predators that had decimated unguarded eggs
in his previous experiment, Eberhard ringed the trunk and branches of the host
tree with a sticky barrier. One of the scelionid wasps, Trissolcus, attacked the
eggs at the same rate in the mother’s absence. However, removing the mother
caused significantly reduced parasitism levels for the other, Phanuropsis. In

fact, parasitism was lessened by over half, from 64% on controls to 29% on
unguarded eggs!
These unexpected results led Eberhard to conclude that perhaps the pres-
ence of egg-guarding females served more or less as a visual ‘beacon’ to
this parasite, serving to identify location of potential hosts. Several obser-
vations seemed to support this idea. For one, Phanuropsis often approached
or rested near quiescent individual bugs of both sexes. When a guarding
female was removed several centimeters to one side of her eggs, female
Phanuropsis waiting nearby typically ignored the undefended eggs and moved
instead to realign themselves with the bug’s new position. That this response
was visually mediated was supported by the additional observation that
the wasps’ searching activity was limited to daylight hours, peaking about
midday.
Thus, even though the parental behavior of Antiteuchus seems well suited
to combat activities of parasites and predators in general, she still suffers a net
loss of eggs to at least one parasite species that has turned her defensive habit
to its own advantage for finding hosts.
10.4.2 Paradoxes of Insect Sociality
Because all living members of both the termites and the ants are totally eusocial,
they can tell us little about the origins of social behavior. Attention has therefore
centered about the bees and wasps, since both of these major groups exhibit a range
of stages that suggest a ‘progression’ to eusociality. The other orders with eusocial
species also display a range of solitary to cooperative behaviors, making them a
rich resource for comparative study as well. All of these suggest that social groups
can form either by cooperation among adults of the same generation or via parent-
offspring (family group) associations (Fig. 10.25). The latter is thought to be the
most common pathway, and formed the cornerstone of William Morton Wheeler’s
treatises on social insects in the 1920s.
Regardless of which evolutionary route has been taken, however, the resultant
society seems to present an evolutionary paradox in the existence of altruism, or
self-sacrificial behaviors. One example is the awesome willingness of many bees
to throw themselves into suicidal battle (Fig. 10.26). Charles Darwin wrestled with
another aspect of the paradox: if natural selection favors the individual able to pro-
duce the greatest number of viable offspring that live to reproduce, then how can
one explain worker sterility? Social insects must represent a special case, he rea-
soned, an example where natural selection was operating on the level of the colony,
or family group, rather than on the single organism.
Group selection is defined as the differential survival and reproduction of
entire cooperative groups. Perhaps the foremost later proponent of group selec-
tion was a Scottish scientist, Vera C. Wynne-Edwards. In an influential 1962 book,
Fig. 10.25 Evolving eusociality. Postulated pathways to complex social behavior in

Hymenoptera. Examples of each stage can be found in extant species of bees and wasps. However,
one should not consider that that all species are inexorably progressing toward eusociality.
Rather each type of group association is presumed optimally adaptive given the particular set of
environmental conditions
Fig. 10.26 Injecting venom. The ultimate self-sacrifice is to be sterile and die defending one’s
home. Honey bees have evolved a barbed sting; upon delivery, it remains in the victim and
disembowels the bee, resulting in the bee’s death
he interpreted a vast array of animal social behaviors as adaptations to prevent

overexploitation of shared resources, a sort of density-dependent population con-
trol. He asserted that such adaptations would evolve over time, because groups
of self-sacrificing restrained individuals would out-compete groups that contained

selfish individuals.
Wynne-Edwards’ version of group selection reflected ideas that were generally
accepted by biologists at the time. For example, social insect colonies were regarded
as superorganisms, a term coined by Wheeler in 1911, in reference to the fact that
individuals acted in ways that benefited the good of the group or colony, even if
and when their behavior was individually disadvantageous. However, ideas began
to change after a seminal 1966 book by George C. Williams persuasively renewed
emphasis on individual selection, and the latter became the dominant paradigm into
the twenty-first century. When scientists observed behavior apparently beneficial
to groups but disadvantageous to the individual donor it was almost manda-
tory that they explain it in terms of individual selection, most often by invoking
pedigree.
Although some of the basic aspects of kin selection theory had been advanced by
others, it was William D. Hamilton who formalized the concept in the mid-1960s,
John Maynard Smith who gave it a name, and Edward O. Wilson who promoted
it in broader and more generally understandable form. In essence, kin selection
refers to changes in gene frequency across generations that are driven at least in
part by behavioral interactions between related individuals; in theory, the enhanced
fitness of relatives sometimes can more than compensate for fitness losses that an
individual incurs.
The combined sum of an individual’s direct fitness (via personal reproduction)
and indirect fitness (genes contributed through help to non-descendent kin to pro-
duce relatives) is termed inclusive fitness. Inclusive fitness is not the simple sum of
an individual’s own offspring plus offspring produced by relatives, however; rather,
each must be adjusted by a coefficient of relationship (r). In essence, this is the prob-
ability that a random gene of the recipient of an altruistic act has an identical copy
in the donor.
Kin selection can operate at any level of relatedness; although the large propor-
tion of unlike genes carried by individuals outside an immediate family group slows
the progress of selection, it does not alter its direction. Under certain situations
altruism is advantageous even toward quite distant relatives.
In the Hymenoptera, many different degrees of sociality occur, and eusocial
behaviors have evolved independently several times. Why has eusociality evolved
so often in this one insect order? And why are hymenopteran societies organized
as matriarchies in which the workers are always genetically female? Hamilton’s
special insight concerned the importance of haplodiploidy—the mode of sex deter-
mination in which haploid individuals derived from unfertilized eggs are male
and diploid individuals arising from fertilized eggs are female. Haplodiploidy, he
showed, results in relatedness asymmetries whereby sisters in ants, bees, and wasps
are more closely related to each other than to their brothers or mothers (Table 10.3).
Since the father is haploid, the half of their genes which female offspring receive
from him are all identical. Within the gene set they receive from their diploid mother,
on average one half of the genes are identical. Thus, full sisters share through com-
mon descent an average of three fourths of their genes. As a result, a daughter is
Table 10.3 Maximal average coefficients of relationship (r) among close kin due to haplodiploidy
in Hymenoptera
Mother Father Sister Brother Son Daughter
Female 0.5 0.5 0.75a 0.25 0.5 0.5

Male 1 0 0.5 0.5 0 1
aA female’s relatedness to sisters (and/or nieces) will be lower if their
mothers have mated more than once, which is often true in eusocial
Hymenoptera
more closely related to her full sisters than she is to her own mother, a biologically
unusual situation. Just as unusual is the male hymenopteran. He has a grandfather
but no father. Because he receives all his genes—in one of two alleles—from his
diploid mother, he shares an average of one half of his genes with his brothers
but possesses only an average of one fourth of his genes in common with his
sisters.
In theory, an individual’s own altruistic sacrifice in fitness could be counterbal-
anced by an increase in the fitness of some group of relatives. Hamilton proposed
that this increase must be by a factor greater than the reciprocal of the coefficient of
relationship to that group (i.e., greater than 1/r) if an altruistic trait is to spread. In
the majority of animals, when an altruistic female sacrifices her life or reproductive
success for a sister or daughter, there is one half chance that the latter shares the gene
for that altruistic trait. For it to be statistically probable that the altruistic gene will
be fixed, the reproductive success of the sister or daughter must be at least doubled
as a result of the sacrifice. However, among full sisters where there is a three quar-
ter chance that the gene is shared, altruistic behavior is statistically probable when
the recipient’s gain in fitness is equal to only four thirds, or 1.33, times the donor’s
loss. Hamilton’s Rule is usually stated as br-c>0, or more simply as br>c, where b
= benefit to the recipient, c = cost to the actor altruist, and r is the coefficient of
relationship between the two (which ranges from 0 to 1).
Thus, because of haplodiploidy, when a hymenopteran mother lives to reproduce
beyond the adulthood of her first female offspring, these offspring may increase
their inclusive fitness more by care of their younger sisters than by an equal amount
of care given to their own offspring. Since sisters are more closely related than are
brothers, altruism will be more favored among daughters than among sons. Haploid
males have more to gain by fathering daughters, with whom they share an average
of one half of their genes, than by assisting in the production of more sisters, with
whom they share an average of only one fourth of their genes.
Thysanoptera are also haplodiploid and since 1990 a number of eusocial thrips
have been discovered (see Section 10.3.5). They differ in that soldiers of both
sexes actively defend the colony and can also reproduce. Colonies of some
soldier-producing species have been found to have extremely high levels of inbreed-
ing; this reduces the relatedness asymmetry between brothers and sisters. Also,
unlike the hymenopteran sting that can only be possessed by the female gender,
spines and enlarged forelegs found in thrips soldiers are not sexually dimorphic
traits.
All other things being equal, it would seem that Hymenoptera should all tend to
become social! Yet the truth is that although apparently all of the thousands of soli-
tary and parasitic Hymenoptera species have this same haplodiploid reproductive
system, most show no trace of sociality. Why?
As new technology has made relatedness one of the easier things to quantify
among social group members, the striking conclusion from numerous studies is
that there is no overwhelming link between sociality and degree of relatedness.
Thus, while genetic relationships and hence genetic mechanisms may be involved
in furthering the initial evolution of cooperative behavior, other factors must also be
considered to be important.
What might such factors be? Searching for the earliest stages of social evolu-
tion has traditionally focused on groups such as halictid bees or the aculeate wasps,
because in these sociality has risen both relatively recently and repeatedly, whereas
in the termites or ants all known species are eusocial and are derived from stocks that
evolved eusociality long ago. However, when we step back and look at the distribu-
tion of eusociality, we find that it has arisen independently at least 13 times, mostly
among the arthropods (including once in a crustacean) but also at least once in a
vertebrate, the naked mole rat. At the same time it is rather striking that eusociality
is relatively rare, considering the huge diversity of arthropods. Pressures inhibit-
ing the attainment of eusociality must be extraordinary. On the flip side, one must
envision some equally extraordinary set of conditions favoring flexibility that could
foster cooperative breeding to overcome environmental contingencies and individ-
ual selfishness. Such a scenario seems to require selection to operate at multiple
levels—genes, individuals and groups—in context-dependent ways.
A revolution in thinking about the origin and evolution of eusociality has
occurred over the past decade. Facilitated by the tremendous progress in whole-
genome sequencing and comparative genomics through various large-scale tech-
nologies (proteomics, epigenomics, etc.), the revolution has been driven in part from
new knowledge about the interaction of genes and the environment. It is increasingly
possible to probe relations between genes and social behavior. Several examples of
social influences on gene expression in insects are now documented. One instance
(see Case Study 1.1) involves shifts in the expression of large populations of genes in
honey bee brains that relate to the point at which hive bees switch to foraging outside
the nest. In particular, differences in the foraging (for) gene expression are related to
social activity, rather than to genetic differences between individuals. Thus, emer-
gent traits arising from interactions among group members constitute a new frontier
for ultimately understanding the social evolution puzzle.
The significance of quantifying relatedness as it relates to the evolution of euso-
ciality is also being reevaluated. Emphasis has shifted to the other parameters of
Hamilton’s Rule: costs and benefits of group life. Besides haplodiploidy, several
other preadaptations are now thought likely to be equally or even more important
for the evolution of eusocial behavior (Table 10.4). Previously existing behavior
patterns, physiological processes, and morphological structures that are already
Table 10.4 Some preadaptations that are considered to be important for development of insect
sociality
Preadaptation Leading to:
1. Reproductive ground plan Differentially expressed genes and resultant

(condition-dependent larval phenotype plasticity upon which natural
developmental pathways and shared selection can act in different environments;
ancestral genes) evolution of castes
2. Haplodiploid sex determination Development of altruistic behaviors depending on
coefficient of relationship
3. Chewing mouthparts Manipulation of nest materials, brood, and other
objects
4. Construction or possession of a nest in Potential family gathering place and behavioral
which young are reared and to which equipment for gathering there; central place
female returns repeatedly foraging
5. Plasticity in stereotyped nesting Progressive provisioning; direct contact between
patterns: a. altered nest and cell larvae and adults opens opportunities for
construction sequence reciprocal interactions
b. placement of numerous cells at one Clumping of brood and increased chances of
site neighbors being relatives
6. Increased female longevity Overlapping lifespan of mother with young
7. Mutual tolerance of conspecifics Communal nesting; possibility of cooperation
8. Control over sex ratio Strong female bias; male production postponed
until late in nest cycle
9. Effective channels for intraspecific Way to integrate and regulate multiple interactions
communication (especially chemical) among many individuals
functional in some other context may serve as bridges to new adaptations. For exam-
ple, the reproductive ground plan hypothesis proposes that the gene network that
once regulated the foraging and reproductive phases that coexisted in solitary ances-
tors was modified in the process of evolution to form the basis for the caste divisions
that are evident in today’s social insects. Thus, queens express genetic traits that
are characteristic of the reproductive phase of solitary insects, whereas workers
express traits characteristic of the foraging phase. A modification of this hypoth-
esis suggests that in some wasps at least, the co-opted gene regulatory networks
were instead from ancestors with two summer generations; workers evolved from
networks switched on in early-emerging first generation individuals, whereas queens
evolved from networks switched on in late-emerging, second-generation individuals
that diapause.
Certainly, many studies have illustrated that relatedness alone is not necessarily
the best predictor of colony success. For example, comparisons of Florida harvester
ant colonies (Pogonomyrmex occidentalis) revealed that colonies with relatively
low relatedness among the workers had strikingly higher colony growth and repro-
duction rates than those with high worker relatedness. Structured kin groups like
harvester ants illustrate that genetic variation between colonies may be greater and
of more evolutionary significance than within-colony variation. It is upon this inter-
colony variation responding differentially to environmental or genetic factors (e.g.,
disease resistance) that group selection processes can act, resulting in some groups
being better adapted than others (Fig. 10.27). Indeed, the principal selective regime
Fig. 10.27 Selecting groups. The group selection model for the evolution of eusociality. Some
individuals in an initial population possess eusocial-prescribing alleles (black thorax) that induce
cooperation, but whose phenotypic expression is flexible (curved and knobbed antennae). Common
possession of eusociality alleles trumps relatedness in the early stages; subsequent selection based
on environmental factors sorts subsequent groups without requiring a kin network. Higher related-
ness in downstream populations can arise by further group selection with or without kin selection;
thus relatedness is more a consequence than a cause of eusociality
leading to and maintaining sociality is differential reproductive success among

colonies of a population.
In many halictid bees, individuals join in groups of females that at best may
be sisters but most often are more distantly related. Suppose they are sisters; if a
joiner becomes a worker, then she is enhancing the production of nieces to whom
her coefficient of relationship is only 0.375. Reasoning from kin selection theory,
the better strategy would seem to be to live alone and produce offspring with a
relationship of 0.5. However, the balance may be tipped by enhanced nest defense
and protection from parasites; early-season cooperation by adult females effectively
produces an ‘instant’ colony with a head start in protection. Such behaviors simul-
taneously increase the fitness and productivity of all individuals involved without
the requirement of any close genetic relationship of the partners, although some-
times such does occur. In other words, the selective effect of colony-level benefit
outweighs the costs of individual fitness loss.
10.4.3 Interspecific Social Interactions
Scaptotrigona stingless bees make a living by plundering the nests of other species,
stealing the honey and provisions for their own nest. Queens of Vespula squamosa
intrude into established nests of Vespula maculifrons and either seize the queen’s
position or lay their own queen-producing eggs alongside those of the host, to be
raised by the host’s worker force.
In Chapter 4, we considered a number of symbiotic interspecific alliances in the
context of feeding strategy. Here, we focus on some exploitive interrelationships
between fully social species. Nest building and the care and feeding of an individ-
ual’s young are usually thought to be very conservative, physiologically deep-seated
behavior patterns. However, even these may be selected against if increased fitness
results.
For example, various solitary Hymenoptera species (see Fig. 4.8) have lost the
behaviors of nest building and provisioning and rely entirely upon the nests and
labors of others. Social parasitism occurs when the offspring of one species are
raised by another species. Among the eusocial Hymenoptera, it has evolved inde-
pendently many times and in many ways. Ants show a greater diversity of forms of
social parasitism than any other group of animals, with over 160 species of social
parasites, no two of which are exactly alike in their adaptations for this mode of life.
The term ‘social parasite’ is unfortunately somewhat misleading, because the
parasite does not need to be social nor does the label refer to every parasite of a
social species. Rather, a social parasite is a species that uses its host as a work force
(something social insect species are particularly suited for) rather than as a direct
source of food. However, because many social parasites are closely related to their
hosts, they often are social themselves. In its most extreme forms, social parasitism
is a condition of permanent and complete dependence, in which the parasite’s entire
life cycle is carried out within the host’s nest. Commonly, when the parasite is a
social species, its own workers are nonexistent or conspicuously degenerate. Such
extreme social parasites, or inquilines, appear to have arisen convergently via at
least three evolutionary routes.
The name Teleutomyrmex means ‘final ant,’ a name most appropriate for what
is probably the ultimate social symbiont. In the Swiss and French Alps, one may
find this ant in small isolated populations that include no workers. Although quite
an assemblage of ant species surround it, Teleutomyrmex is parasitic upon only
one—its closest phylogenetic neighbor, Tetramorium. It has never been found out-
side the nests of its hosts. Teleutomyrmex queens spend much of their time riding
upon the backs of the queens of their host colony. Very delicate, the symbionts sel-
dom move independently; they apparently feed only upon regurgitates passed from
workers to the host queen. Placed in an artificial nest they cannot survive, even
if host workers are present. Their brains, mandibles, nervous system, and skin, in
fact nearly all of their morphology, shows extensive degeneration, except in their
reproductive system—each of the tiny physogastric parasites lays an average of one
egg every 30 seconds. Having up to six or eight ectoparasitic ants riding upon her
back may slow down a host queen; infested colonies tend to be somewhat smaller
than non-infested ones. Significantly, infested colonies produce no sexual forms
of the host. Teleutomyrmex adults, especially older females, apparently produce
a very potent attractant; host workers lick them continuously. As it is circulated
throughout the colony this substance may impose ‘reproductive castration’ upon its
hosts. Such a castration phenomenon has been demonstrated in a number of similar
parasite–host colony relationships, though its physiological mechanism has yet to be
determined.
Two more examples illustrate the diversity of social inquilinism. One involves
slavery. Red Amazon ants (Polyergus rufescens) with their saberlike mandibles are
fierce fighters but totally inept housekeepers. In their home nest, their only activities
are grooming themselves and begging for food. Living as a pure colony, they would
surely perish, because they neither excavate nests nor care for their own young.
Fig. 10.28 Raiding the neighbors. Amazon ants of the species Polyergus rufescens (light) con-
ducting a raid upon a colony of the slave species Formica fusca (dark) nesting in dry soil beneath
a stone. Killing resistors by piercing them with their saber-like mandibles, the Amazons rush off
with captured brood. Such behavior occurs quite often among cold temperate ant species, where
slave labor is found in at least 35 species from six independently evolved groups
What keeps them alive is a unique slave-making habit. Periodically, the Amazon
ants swarm out of their nest, marching swiftly to the nests of Formica fusca to
launch a raid (Fig. 10.28). With feverish haste, they pour over the colony in a body;
any defenders who resist the attack are punctured and killed. Then, like pillaging
soldiers, the Amazons proceed to carry off fusca pupal cocoons.
Back in the Amazon nest, the pupae soon begin to hatch; genetically programmed
to perform various housekeeping tasks, they begin work. Some bring food into
the nest, while others tend the eggs, larvae, and pupae of their captors. Still oth-
ers actually feed the adult Amazons who have made them slaves, responding to
their begging by regurgitating liquid droplets. Throughout all this activity, the fusca
slaves make no distinction between their genetic siblings and the Amazons, fully
accepting their captors as sisters. Eventually, their numbers dwindle, because as
members of the worker caste they cannot reproduce. In response, the slave-making
Amazons set out to pillage alien colonies once again.
The nest-raiding techniques of slave makers are among the most sophisticated
behavior patterns found in the insect world. Some physically overpower their vic-
tims. Others depend on various chemical ruses. Workers of two species of Formica
spray acetate chemicals at resisting nest defenders; these chemicals act as ‘pro-
paganda substances’ that imitate the alarm pheromones of the slave species so
powerfully they throw the resistors into utter helpless panic. In addition to this
disruptive effect, the acetates serve as an attractant to workers of the slave maker,
quickly assembling them where fighting has broken out.
A second example shifts our focus to queen behavior. Sexual forms of most social
Hymenoptera leave their home colony for their mating flights. Some mated queens
fail to find their own colony again following such flights, or after hibernation fail
to successfully initiate their own nest the following season. Should they locate a
colony of another species and somehow eliminate the host queen, however, they
gain an established nest complete with a work force. For a brief period, when the
usurper queen has begun to reproduce but not all the original queen’s brood have
reached old age and died, the colony consists of workers of both species. Eventually,
however, the nest will come to contain only the usurper and her offspring. This
situation, termed temporary parasitism may evolve from a facultative affair to an
obligatory one where usurper queens depend entirely upon this method to found new
colonies. Temporary parasites have increased their chances of successfully entering
the host nest in a variety of ways, from stealth to aggression to deceptions involving
chemistry, morphology, or behaviors such as ‘playing dead.’
The apparent ease by which symbionts have repeatedly intruded themselves
into their host’s colony is somewhat surprising, particularly in view of the excep-
tional ability of social insects to defend themselves from larger predators. The
clue to understanding this paradox lies in the organization and integration of insect
societies. As stressed in Chapter 6, insect societies are integrated primarily by non-
personal colony odors, an easier code to break than the visually based individual
recognition or learned roles and dominance-based interactions characteristic of ver-
tebrate societies. Furthermore, organization into castes results in role specialization
in which the individuals of one caste largely lack a broad awareness of the roles
of other colony members. These combine to lend an impersonal nature to social
insect organizations, which has apparently made it relatively effortless for social
symbionts to insert themselves into the colony regime.
Credits
Chapter One
Fig. 1.1 Diagram by the authors.

Fig. 1.2 From Gullan, P. J. and P. S. Cranston. 2004. The Insects. An Outline of
Entomology, 3rd edition. Wiley-Blackwell, Oxford. By permission of
Wiley-Blackwell.
Fig. 1.3 Photograph courtesy of S. Alexious and The Archaeological Museum of
Heraclion, Crete. See also LaFleur, R. A., R. W. Matthews, and D. B. McCorkle, Jr.
1979. A reexamination of the Mallia insect pendant. American Journal of
Archeology 83:208–212 & PI. 29.
Fig. 1.4 Modified from Bodenheimer, F.S. 1928. Materialien zur Geschichteder
Entomologie bis Linné, Vol. I, Junk, Berlin.
Fig. 1.5 Engravings reproduced from M. S. Merian’s caterpillar books, bound
together in Erucarum Ortus, Alimentum et Paradoxa Metamorphosis, a
compilation published in Amsterdam 1718. See also Todd, K. 2007. Chrysalis.
Maria Sibylla Merian and the Secrets of Metamorphosis. Harcourt, Inc.
Fig. 1.6 (left, right) Portraits from Kelly, H. A. 1906. Walter Reed and Yellow
Fever. McClure, Phillips and Co., NY. (center) Mosquito photograph by James
Gathany, downloaded 8 May 2009 from http://commons.wikimedia.org/
wiki/File:Aedes_aegypti_bloodfeeding_CDC_Gathany.jpg/
Fig. 1.7 Photographs courtesy of Marla Spivak and Gary Reuter, University of
Minnesota. See also www.extension.umn.edu/honeybees/
Fig. 1.8 Diagram by the authors after Rothenbuhler, W. B. 1967. Genetic and
evolutionary considerations of social behavior of honey bees and some related
insects. pp. 61–l06 In J. Hirsch (Ed.), Behavior-Genetic Analysis, McGraw-Hill,
NY. See also Lapidge, K., R. Oldroyd, and M. Spivak. 2002. Seven suggestive
quantitative trait loci influence hygienic behavior of honey bees.
Naturwissenschaften 89:565–568.

DOI 10.1007/978-90-481-2389-6_BM2, C Springer Science+Business Media B.V. 2010
446 Credits

Fig. 1.11 Drawing by Arthur Rackham, reproduced from Aesop’s Fables,
translated by V. S. Jones. Published 1912, William Heinemann, London.
Fig. 1.12 (left) Roth, L. M. 1970. Evolution and taxonomic significance of
reproduction in Blattaria. Annual Review of Entomology 15:75–96. Reprinted, with
permission, from the Annual Review of Entomology, volume 15 © 1970 by Annual
Reviews www.annualreviews.org. (right) Drawing by the authors. See also Roth, L.
M. 2003. Systematics and phylogeny of cockroaches. Oriental Insects 37:1–186;
and Klass, K.-D. and R. Meier. 2006. A phylogenetic analysis of Dictyoptera
(Insecta) based on morphological characters. Entomologische Abhandlungen
63:3–50.
Fig. 1.13 Photograph by the authors.
Fig. 1.14 Photograph by Andy Phillips.
Fig. 1.15 Diagram by the authors, based on updated information and Evans, H. E.
and R. W. Matthews. 1973. Systematics and nesting behavior of Australian Bembix
sand wasps (Hymenoptera: Sphecidae). Memoirs of the American Entomological
Institute No. 20, 386 pp.
Fig. 1.16 Lewis, S. M. and C. K. Cratsley. 2008. Flash signal evolution, mate
choice, and predation in fireflies. Annual Review of Entomology 53:293–321.
Reprinted, with permission, from the Annual Review of Entomology, volume 53 ©
2008 by Annual Reviews www.annualreviews.org.
Fig. 1.17 Holzenthal R. W., R. J. Blahnik, A. L. Prather, et al. 2007. Order
Trichoptera Kirby 1813 (Insecta), Caddisflies. pp. 639–698 In Z.-Q. Zhang and
W. A. Shear (Eds.) 2007. Linnaeus tercentenary: Progress in invertebrate
taxonomy. Zootaxa 1668:1–766.
Fig. 1.18 Toma, D., G. Bloch, D. Moore, et al. 2000. Changes in period mRNA
levels in the brain and division of labor in honey bee colonies. Proceedings of the
National Academy of Sciences USA. 97:6914–6919. Copyright (2000), National
Academy of Sciences, USA.
Table 1.1 By the authors, based on Tinbergen, N. 1963. On the aims and methods
of ethology. Zeitschrift fur Tierpsychologie 20:410–463.
Chapter Two
Fig. 2.1 Diagram by the authors, based on Snodgrass, R. E. 1935. Principles of

Insect Morphology. McGraw-Hill, NY.
Credits 447
Fig. 2.2 From Wilson, D. M. 1968. The flight-control system of the locust.
Scientific American 218:83–90 (May). Copyright © (1968) by Scientific American,
Inc. All rights reserved. See also Marder, E., D. Bucher, D. Schulz, et al. 2003.
Invertebrate central pattern generator moves along. Current Biology
15(17):R685–R699.
Fig. 2.3 (above) Redrawn after Roeder, K. D. 1967. Nerve Cells and Insect
Behavior, revised edition. Harvard University Press, Cambridge, MA. (below)
From Camhi, J. M. 1980. The escape system of the cockroach. Scientific American
243:151–172. Copyright © (1980) by Scientific American, Inc. All rights reserved.
See also Levi, R. and J. M. Camhi. 2000. Wind direction coding in the cockroach
escape response: winner does not take all. Journal of Neurosciences
20(10):3814–3821.
Fig. 2.4 Modified from Dethier, V. G. 1971. A surfeit of stimuli: a paucity of
receptors. American Scientist 59:706–715.
Fig. 2.5 Roeder, K. D. 1970. Episodes in insect brains. American Scientist
58:378–389. Reprinted by permission, American Scientist, journal of Sigma Xi,
The Scientific Research Society of North America.
Fig. 2.6 From Alcock, J. 1975 Animal Behavior. An Evolutionary Approach. 1st
ed. Sinauer Associates, Sunderland, MA.
Fig. 2.7 (above) Redrawn after Rains, G. C., J. K. Tomberlin, and D. Salasiri.
2008. Using insect sniffing devices for detection. Trends in Biotechnology 26(6):
288–294. (below) From Olson, D. M., G. C. Rains, T. Meiners, et al. 2003.
Parasitic wasps learn and report diverse chemicals with unique conditionable
behaviors. Chemical Senses 28: 545–549. See also Salazar, B. A. and D. W.
Whitman. 2001. Defensive tactics of caterpillars against predators and parasitoids.
Chapter 8 In T. N. Ananthakrishnan, Insects and Plant Defence Dynamics. Science
Publishers, Enfield, NH, USA.
Fig. 2.8 Reproduced with permission from Lent, D. D. and H-W. Kwon. 2004.
Antennal movements reveal associative learning in the American cockroach
Periplaneta americana. Journal of Experimental Biology 207: 369–375.
Fig. 2.9 Tinbergen, N. 1951. The Study of Instinct. Clarendon Press of the Oxford
University Press, London. By permission of Oxford University Press. See also
Tinbergen, N. 1972. The Animal in its World. Explorations of an Ethologist,
1932–1972. Vol. 1. Field Studies (especially pp. 103–145). Harvard University
Press, Cambridge, MA.
Fig. 2.10 Drawing by Charles Clare from Janzen, D. H. 1974. The deflowering of
Central America. Natural History 83:48–53.
Fig. 2.11 (above, center) Modified from Gandolfi, M., L. Mattiacci and S. Dorn.
2003. Preimaginal learning determines adult response to chemical stimuli in a
parasitic wasp. Proceedings of the Royal Society of London B 270:2623–2629.
448 Credits
Reproduced with permission of the Royal Society of London. (below) From

Gullan, P. J. and P. S. Cranston. 2004. The Insects. An Outline of Entomology, 3rd
edition. Wiley-Blackwell, Oxford. By permission of Wiley-Blackwell.
Fig. 2.12 Modified from Gandolfi, M., L. Mattiacci and S. Dorn. 2003.
Preimaginal learning determines adult response to chemical stimuli in a parasitic
wasp. Proceedings of the Royal Society of London B 270:2623–2629. Reproduced
with permission of the Royal Society of London.
Fig. 2.13 Drawing by the authors, based on Kerfoot, W. B. 1967. The lunar
periodicity of Sphecodogastra texana, a nocturnal bee. Animal Behaviour
15:478–485.
Fig. 2.14 From Truman, J. W. and L. M. Riddiford. 1970. Neuroendocrine control
of ecdysis in silkmoths. Science 167:1624–1626. Reprinted with permission from
AAAS.
Chapter Three
Fig. 3.1 Drawings by Paul H. Matthews.
Fig. 3.2 Photograph courtesy of Poramate Manoonpong, Bernstein Center for
Computational Neuroscience, Goettingen, Germany.
Fig. 3.3 Drawing by the authors. See also Dickinson, M. 2005. Insect flight.
Current Biology 16(9): R309–314.
Fig. 3.4 Drawing by Joan W. Krispyn. See also Fraenkel, G. S. and D. L. Gunn.
1940. The Orientation of Animals. Kineses, Taxes and Compass Reactions. Dover,
NY.
Fig. 3.5 Redrawn and modified from Schöne, H. 1951. Die Lichtorientierung der
Larven von Acilius sulcatus L. und Dytiscus marginalis L. Zeitschrift Vergleichen
Physiologie 33:63–98, with kind permission of Springer Science+Business
Media.
Fig. 3.6 From Alcock, J. 1975. Animal Behavior. 1st ed. Sinauer Press,
Sunderland, MA. See also Lindauer, M.1971. Communication Among the Bees.
Harvard University Press, Cambridge, MA.
Fig. 3.7. Photographs courtesy of Prof. Dr. Helmut Schmitz, Universität Bonn,
Institut für Zoologie. See also Schmitz, H, H. Bleckmann, and M. Murtz. 1997.
Infrared detection in a beetle. Nature (Lond.) 386:773–774.
Fig. 3.9 Redrawn after Dingle, H. 1972. Migration strategies of insects. Science
175:1327–1334. See also Dingle, H. 1996. Migration. The Biology of Life on the
Move. Oxford University Press, New York.
Credits 449
Fig. 3.10 (left) From Askew, R. R. 1971. Parasitic Insects, Heinemann

Educational Books Ltd. and American Elsevier. (right) Redrawn from Evans, H. E.
1969. Phoretic copulation in Hymenoptera. Entomological News 80:113–124.
Reproduced with permission from the American Entomological Society.
Fig. 3.11 From Solensky, M. J. 2004. (Above) Chapter 10. Overview of monarch
migration and (below) Chapter 15. Overview of monarch overwintering biology. In
Oberhauser, K. S. and M. J. Solensky (Eds.) 2004. The Monarch Butterfly. Biology
and Conservation. Cornell University Press, Ithaca, NY. Above by permission of
Cornell University Press; below by permission of Wiley-Blackwell.
Fig. 3.12 Reproduced with permission from Brower, L.P. 1996. Monarch butterfly
orientation: missing pieces of a magnificent puzzle. Journal of Experimental
Biology 199:93–103.
Fig. 3.13 From Wehner, R. 1989. Neurobiology of polarization vision. Trends in
Neurosciences 12:353–359.
Chapter Four
Fig. 4.3 Stoffolano, J. G. Jr. 1974. Control of feeding and drinking in diapausing
insects. In L. Barton Browne (Ed.), Experimental Analysis of Insect Behaviour,
Springer Verlag, NY, with kind permission of Springer Science+Business Media.
Fig. 4.4 Weires, R. W. and H. G. Chiang, 1973. Integrated control prospects of
major cabbage insect pests in Minnesota-based on the faunistic, host varietal, and
trophic relationships. University of Minnesota Agricultural Experiment Station
Technical Bulletin 291, 42 pp.
Fig. 4.5 Diagram by the authors. See also Charnov, E.L. 1976. Optimal foraging:
the marginal value theorem. Theoretical Population Biology 9:129–136.
Fig. 4.6 From Alcock, J. 1975. Animal Behavior. An Evolutionary Approach.
1st ed. Sinauer Associates, Sunderland, MA.
Fig. 4.8 Drawing by Joan W. Krispyn.
Fig. 4.9 Courtesy of Ulrich G. Mueller. From Mueller, U. G. and N. Gerardo.
2002. Fungus-farming insects: multiple origins and diverse evolutionary histories.
Proceedings of the National Academy of Science 99(24):15247–15249.
Fig. 4.10 Courtesy of Bert Hölldobler. From Hölldobler, B.1971. Communication
between ants and their guests. Scientific American 224:86–93 (March). Copyright
© (1971) by Scientific American, Inc. All rights reserved.
450 Credits
Fig. 4.11 Drawing by Turid Hölldobler-Forsyth. Hölldobler, B. 1970.

Orientierungsmechanismen des Ameisengastes Atemeles (Coleoptera:
Staphylinidae) bei der Wirtssuche. In W. Herre (Ed.) Verhandlungen der
Zoologischen Gesellschaft (Zoologischer Anzeiger Supplement)
33:580–585.
Fig. 4.12 Geiselhardt, S., K. Peschke and P. Nagel. 2007. A review of
myrmecophily in ant nest beetles (Coleoptera: Carabidae: Paussinae): linking early
observations with recent findings. Naturwissenschaften 94: 871–894, with kind
permission of Springer Science+Business Media.
Fig. 4.13 Photograph by the authors. See also Ruehlmann, T. E., R. W. Matthews
and J. R. Matthews. 1988. Roles for structural and temporal shelter-changing by
fern-feeding Lepidopteran larvae. Oecologia 75:228–232.
Fig. 4.14 (left) Image downloaded 24 April 2009 from
http://www.cefe.cnrs.fr/coev/albums/ficus_carica.htm. (right) Drawings adapted
from Wigglesworth, V. B. 1964. The Life of Insects. The New American
Library, NY.
Fig. 4.15 Courtesy of Braulio Dias. See also Dias, B. F. 1975. Comportamento
pre-social de sinfitas do Brazil Central. 1. Themos olfersi (Klug) (Hym., Argidae).
Studia Entomologia 18:401–432.
Fig. 4.16 Modified from Wilson, E. O. and T. Eisner 1957. Quantitative studies of
liquid food transmission in ants. Insectes Sociaux 4:157–166, with kind permission
of Springer Science+Business Media.
Chapter Five
Opening quote from Owen, D. 1980. Camouflage and Mimicry, p. 15. University
of Chicago Press.
Fig. 5.1 Photographs by Robert E. Silberglied.
Fig. 5.2 From Cott, H. B. 1940. Adaptive Coloration in Animals. Methuen, NY,
with kind permission of Springer Science+Business Media.
Fig. 5.3 Kettlewell, H. B. D. 1973. The Evolution of Melanism. The Study of a
Recurring Necessity, with Special Reference to Industrial Melanism in the
Lepidoptera. Clarendon, NY. See also Hopper, J. 2003. Of Moths and Men: An
Evolutionary Tale: The Untold Story of Science and the Peppered Moth. W. W.
Norton & Co., NY.
Fig. 5.4 Courtesy of Terrence D. Fitzgerald. See also Fitzgerald, T. D. 2008. Toxic
hairs enable some caterpillars to venture forth in conspicuous processions. Natural
History 117(7):28–33.
Credits 451
Fig. 5.5 (left) Photograph by Douglas W. Whitman. (right) From Greene E., L. J.
Orsak, D. Whitman. 1987. A tephritid fly mimics the territorial displays of its
jumping spider predators. Science 236:310–312. Reprinted with permission from
AAAS.
Fig. 5.6 Photographs by the authors.
Fig. 5.7 Photograph by Robert E. Silberglied.
Fig. 5.8 (left) Photograph by the authors. (right) Photograph by Douglas W.
Whitman.
Fig. 5.9 (left) Photograph by Abraham Hefetz. (right). Photograph courtesy of
Thomas Eisner. See also Aneshansley, D., T. Eisner, J. M. Widom et al. 1969.
Biochemistry at 100o C. The explosive discharge of bombardier beetles
(Brachinus). Science 165:61–63.
Fig. 5.11 (above) Photograph by the authors. (center and below) Photographs by
Robert E. Silberglied.
Fig. 5.13 From Stradling, D. J. 1976. The nature of the mimetic patterns of the
brassolid genera, Caligo and Eryphanus. Ecological Entomology 1:135–138. By
permission of Wiley-Blackwell.
Fig. 5.14 From Cott, H. B.1940. Adaptive Coloration in Animals. Methuen, NY,
with kind permission of Springer Science+Business Media.
Fig. 5.15 Drawing courtesy of Daniel Otte. See also Otte, D. 1977. Acoustical
communication in Orthoptera. In T. A. Sebeok (Ed.), How Animals Communicate.
Indiana University Press, Bloomington, IN.
Chapter Six
Fig. 6.1 Drawing by Paul H. Matthews, based on Billen, J. and E. D. Morgan.

1998. Pheromone communication in social insects: sources and secretions.
pp. 3–33 In R. K. Vander Meer, M.D. Breed, M.L. Winston et al. (Eds.),
Pheromone Communication In Social Insects: Ants, Wasps, Bees, And Termites.
Westview Press, Boulder, CO.
Fig. 6.2 Photograph by Albert Dietz.
Fig. 6.3 Hangartner, W. 1967. Spezifität und Inaktivierung des Spurpheromons
von Lasius fuliginosus Latr. Und Orientierung der Arbeiterinnen in Duftfeld.
Zeitschrift für Vergleichen Physiologie 57:103–136, with kind permission of
Springer Science+Business Media.
452 Credits
Fig. 6.4 Photograph by Kevin Wanner, Montana State University. See also
Steinbrecht, R. A. 1999. Olfactory receptors. pp. 155–176 In E. Eguchi,
Y. Tominaga & H. Ogawa, Atlas of Arthropod Sensory Receptors. Dynamic
Morphology in Relation to Function. Springer-Verlag.
Fig. 6.5 Adapted from Birch, M. C. 1984. Aggregation in bark beetles.
pp. 331–354 In W. J. Bell and R. T. Cardé (Eds.) Chemical Ecology of Insects,
Chapman and Hall, London, with kind permission of Springer Science+Business
Media.
Fig. 6.7 (left) Photograph by Eleanor Smithwick and U. Eugene Brady. (right)
Photograph by Rob Peakall. See also Schiestl F. P., R. Peakal, J. G. Mant, et al.
2003. The chemistry of sexual deception in an orchid-wasp pollination system.
Science 302(5644):437–438.
Fig. 6.8 Photograph by Robert L. Silberglied.
Fig. 6.9 Photograph by Patricia J. Moore. See also Moore, A. J. and P. J. Moore.
1999. Balancing sexual selection through opposing mate choice and male
competition. Proceedings of the Royal Society of London Series B – Biological
Sciences. 266:711–716.
Fig. 6.10 Drawing by Lee C. Ryker. See also Ryker, L.C. 1984. Acoustic and
chemical signals in the life cycle of a beetle. Scientific American 250:113–124.
Fig. 6.11 From Ishii, S. 1970. Aggregation of the German cockroach Blattella
germanica L. pp. 93–109 In D. L. Wood, R. M. Silverstein, and M. Nakajima
(Eds.). Control of Insect Behavior by Natural Products, Academic Press, NY.
Fig. 6.12 From Beggs, K. T., K. A. Glendining, N. M. Marechal, et al. 2007.
Queen pheromone modulates brain dopamine function in worker honey bees.
Proceedings of the National Academy of Sciences of the US. 104(7):2460–2464.
Copyright (2007) National Academy of Sciences, USA.
Fig. 6.13 (left) Drawing by Paul H. Matthews. (right) Photograph by Terrence D.
Fitzgerald. See also Fitzgerald, T. D. 1995. The Tent Caterpillars. Cornell
University Press, Ithaca, NY.
Fig. 6.14 Drawings by Joan W. Krispyn. See also Beale, M. H., M. A. Birkett,
T. J. A. Bruce, et al. 2006. Aphid alarm pheromone produced by transgenic plants
affects aphid and parasitoid behavior. Proceedings of the National Academy of
Sciences of the USA 103(27):10509–10513.
Fig. 6.15 Courtesy of CSIRO Division of Entomology, Canberra, Australia.
Fig. 6.16 From Price, P.W. 1972. Behavior of the parasitoid Pleolophus basizonus
(Hymenoptera: Ichneumonidae) in response to changes in host and parasitoid
density. Canadian Entomologist 104:129–140. Courtesy of the Entomological
Society of Canada.
Credits 453
Fig. 6.17 From Wilson, E. O. and W. H. Bossert, 1963. Chemical communication

among animals. Recent Progress in Hormone Research 19:673–716.
Fig. 6.18 From Wilson, E. O. and W. H. Bossert, 1963. Chemical communication
among animals. Recent Progress in Hormone Research 19:673–716.
Fig. 6.19. Drawing by Paul H. Matthews. See also Cook, S. M., Z. R. Khan, and
J. A. Pickett. 2007. The use of push-pull strategies in integrated pest management.
Annual Review of Entomology 52:375–400.
Table 6.1 Based on Hölldobler, B. and E. O. Wilson. 1978. The multiple
recruitment systems of the African weaver ant Oecophylla longinoda (Latreille)
(Hymenoptera: Formicidae). Behavioral Ecology and Sociobiology 3:19–60.
Table 6.2 Based on Howard, R. W. and G. J. Blomquist. 2005. Ecological,
behavioral, and biochemical aspects of insect hydrocarbons. Annual Review of
Entomology 50:371–395.
Table 6.3 Compiled from various sources.
Table 6.4 Based on Wilson, E. O. and W. H. Bossert, 1963. Chemical
communication among animals. Recent Progress in Hormone Research
19:673–716.
Chapter Seven
Fig. 7.1 Reproduced with permission from Timmins, G. S., F. J. Robb, C. M.

Wilmot, et al. 2001. Firefly flashing is controlled by gating oxygen to
light-emitting cells. Journal of Experimental Biology 204(16):2795–2801.
Fig. 7.2 Drawing by D. Otte. See also Lloyd, J. E. 1966. Studies on the flash
communication system of Photinus fireflies. Miscellaneous Publications of the
Museum of Zoology, University of Michigan, No. 130, 95pp.
Fig. 7.3 (above) Photograph by Douglas W. Whitman. (below) Photograph by
Susan Ellis, Bugwood.org. See also Cronin, T.W., M. Jarvilehto, M. Weckstrom, et.
al. 2000. Tuning of photoreceptor spectral sensitivity in fireflies (Coleoptera:
Lampyridae). Journal of Comparative Physiology A 186:1–12.
Fig. 7.4 Drawings by Joan W. Krispyn.
Fig. 7.5 (left) Photograph by Robert E. Silberglied. See Silberglied, R. E. and
O. R. Taylor. 1973. Ultraviolet differences between the sulfur butterflies, Colias
eurytheme and C. philodice, and a possible isolating mechanism. Nature
241:406–408. (right) Kemp, D. J. 2006. Ultraviolet ornamentation and male
mating success in a high-density assemblage of the butterfly Colias eurytheme.
Journal of Insect Behavior 19:669–684, with kind permission of Springer
Science+Business Media.
454 Credits
Fig. 7.6 Photographs copyright by Elizabeth A. Tibbetts. See also Tibbetts,

E. A. and R. Lindsay. 2008. Visual signals of status and rival assessment in Polistes
dominulus paper wasps. Biology Letters 4:237–239. For an alternative view see
Cervo, R., L. Dapporto, L. Beani, et al. 2008. On status badges and quality signals
in the paper wasp Polistes dominulus: body size, facial colour patterns and
hierarchical rank. Proceedings of the Royal Society B 275:1189–1196.
Fig. 7.9 From Brower, L. P., J. V.-Z. Brower, and F. P. Cranston. 1965. Courtship
behavior of the queen butterfly, Danaus gilippus berenice (Cramer). Zoologica
50:1–39. Reprinted with permission of the Wildlife Conservation Society.
Fig. 7.10 Magnus, D.1958. Experimentalle Untersuchunger zur Bionomie und
Ethologie des Kaisermantels Argynnis paphia L. (Lep. Nymphalidae). Zeitschrift
fur Tierpsychologie 15:397–426, with kind permission of Springer
Fig. 7.11 Magnus, D. 1958. Experimentalle Untersuchunger zur Bionomie und
Ethologie des Kaisermantels Argynnis paphia L. (Lep. Nymphalidae). Zeitschrift
fur Tierpsychologie 15:397–426, with kind permission of Springer
Chapter Eight
Fig. 8.1 Courtesy of Rex Cocroft. From Cocroft, R. B. and R. L. Rodríguez. 2005.
The behavioral ecology of insect vibrational communication. BioScience
55(4):323–334. Copyright, American Institute of Biological Sciences.
Fig. 8.2 Photograph by Douglas W. Whitman.
Fig. 8.3 Courtesy of Rex Cocroft. From Cocroft, R. B., T. D. Tieu, R. R. Hoy,
et al. 2000. Directionality in the mechanical response to substrate vibration in a
trehopper (Hemiptera:Membracidae: Umbonia crassicornis). Journal of
Comparative Physiology A 186:695–705.
Fig. 8.4 Drawing by Paul H. Matthews, based on Yak, J. and R. Hoy. 2003.
Hearing. pp. 498–505 In V. H. Resh and R. T. Cardé (Eds.), Encyclopedia of
Insects. Academic Press, New York.
Fig. 8.5 Drawing by the authors. See also Cooper, K. W. 1957. Biology of
eumenine wasps. V. Digital communication in wasps. Journal of Experimental
Zoology 134:469–514.
Fig. 8.6 Drawings by the authors.
Fig. 8.7 From Alexander, R. D. and T. E. Moore. 1962. The evolutionary
relationships of 17-year and 13-year cicadas, and three new species (Homoptera,
Cicadidae, Magicacada). Miscellaneous Publications of the Museum of Zoology,
Credits 455
University of Michigan No. 121, 59 pp. See also Cooley, J. R., G. Kritsky, M. J.
Edwards, et al. 2009. The distribution of periodical cicada brood X in 2004.
American Entomologist 55(2):106–113.
Fig. 8.8 Photograph by Justin O. Schmidt. See also Schmidt, J. O. and M. S. Blum.
1977. Adaptations and responses of Dasymutilla occidentalis (Hymeoptera:
Mutillidae) to predators. Entomologia Experimenta et Applicata 21:99–111.
Fig. 8.9 Courtesy of Richard D. Alexander.
Fig. 8.10 Stevenson P. A., Dyakonova, V., Rillich J., et al. 2005. Octopamine and
experience-dependent modulation of aggression in crickets. The Journal of
Neuroscience 25(6):1431–1441.
Fig. 8.11 Modified from Hoy, R., J. Hahn, and R. C. Paul. 1977. Hybrid cricket
auditory behavior: evidence for genetic coupling in animal communication.
Science 195:82–83. Reprinted with permission from AAAS.
Fig. 8.12 Modified from Gibson, G. and I. Russell. 2006. Flying in tune: sexual
recognition in mosquitoes. Current Biology 16(11):1311–1316.
Figs. 8.13 and 8.14 Drawings copyright Emily A. Matthews, based on Frisch,
K. von 1967. The Dance Language and Orientation of Bees. Harvard University
Fig. 8.15 From Michelsen, A., Andersen, B. B., Storm J., et al. 1992. How
honeybees perceive communication dances, studied by means of a mechanical
model. Behavioral Ecology and Sociobiology 30(3–4):143–150. With kind
permission of Springer Science+Business Media.
Fig. 8.16 From Nieh, J. C. 1999. Stingless bee communication. American Scientist
87(5):428–435. Reprinted by permission of American Scientist, journal of Sigma
Xi, The scientific Research Society of North America.
Fig. 8.17 (above) Photographs by Thomas D. Seeley. (below) From Dyer, F. C.
2002. The biology of the dance language. Annual Review of Entomology
47:917–949. Reprinted, with permission, from the Annual Review of Entomology,
volume 47 © 2002 by Annual Reviews www.annualreviews.org.
Table 8.1 Based on DuMortier, B. 1963. Morphology of sound emission apparatus
in Arthropoda. pp. 277–345 In R. G. Busnel (Ed.), Acoustic Behaviour of Animals,
and on Haskell, P. T. 1974. Sound production. pp. 353–410 In M. Rockstein (ed.),
The Physiology of Insecta, Volume II, 2nd ed., Academic Press, NY. See also,
Bailey, W. J. 1991. Acoustic Behaviour of Insects. An evolutionary perspective.
Chapman and Hall, London.
Table 8.2 Based on Schwartzkopff, J. 1974. Mechanoreception. pp. 273–352 In
M. Rockstein (Ed.), The Physiology of Insecta, Volume II, 2nd ed., Academic
Press, NY. See also Ewing, A. W. 1989. Arthropod Bioacoustics. Neurobiology and
Behaviour. Comstock Press of Cornell University Press, Ithaca, NY.
456 Credits
Chapter Nine
Fig. 9.1 Drawing by Joan W. Krispyn. See also Thornhill, R. 1976. Reproductive
behavior of the lovebug, Plecia nearctica (Diptera: Bibionidae). Annals of the
Entomological Society of America 69:843–847.
Fig. 9.3 From Stone G. N., R. J. Atkinson, A. Rokas„ et al. 2008. Evidence for
widespread cryptic sexual generations in apparently purely asexual Andricus
gallwasps. Molecular Ecology 17:652–665.
Fig. 9.4 From Hardy, I. C. W., P. J. Ode, and M. T. Siva-Jothy. 2005. Mating
behavior. pp. 219–260 In M. A. Jarvis (Ed.), Insects as Natural Enemies: A
Practical Perspective. Springer, Dordrecht Netherlands, with kind permission of
Springer Science+Business Media.
Fig. 9.5 Drawing by Joan W. Krispyn. See also Stich, H. F. 1963. An experimental
analysis of the courtship pattern of Tipula oleracea (Diptera). Canadian Journal of
Zoology 41:99–109.
Fig. 9.6 From Dressler, R. L. 1968. Pollination by euglossine bees. Evolution
22:202–210.
Fig. 9.8 From Kullenberg B. and G. Bergström, 1976. Hymenoptera aculeata
males as pollinators of Ophrys orchids. Zoologica Scripta 5:13–23. By permission
of Wiley-Blackwell.
Fig. 9.9 Photograph by the authors. See also Crankshaw, O. S. and
R. W. Matthews. 1981. Sexual behavior among parasitic Megarhyssa wasps
(Hymenoptera: Ichneumonidae). Behavioral Ecology and Sociobiology 9:1–7.
Fig. 9.10 Drawing by Emile Blanchard, In Figuier, L. 1869. The Insect World.
Chapman and Hall, NY.
Fig. 9.11 From Otte, D. 1974. Effects and functions in the evolution of signaling
systems. Annual Review of Ecology and Systematics 5:385–417. Reprinted, with
permission, from the Annual Review of Ecology, Evolution, and Systematics,
volume 5 ©1974 by Annual Reviews www.annualreviews.org.
Fig. 9.12 Photograph by Garry Wall.
Fig. 9.13 Photograph by the authors. See also Matthews, R. W., J. M. González,
J. R. Matthews, et al. 2009. Biology of the parasitoid Melittobia (Hymenoptera:
Eulophidae). Annual Review of Entomology 54:251–266.
Fig. 9.14 Photograph courtesy of Randy Thornhill. See Thornhill, R. 1976. Sexual
selection and nuptial feeding behavior in Bittacus apicalis (Insecta: Mecoptera).
American Naturalist 110:529–548. See also Thronhill, R. and J. Alcock. 1983. The
Evolution of Insect Mating Systems. Harvard University Press., Cambridge, MA.
Credits 457
Fig. 9.15 (left) Photograph by Douglas W. Whitman. (right) Photograph by Judy

Baxter, Hahira, GA. See also Eisner, T., M. Eisner, and M. Siegler. 2005. Secret
Weapons: Defenses of Insects, Spiders, Scorpions, and Other Many-Legged
Creatures. Belknap Press of Harvard University Press, Cambridge, MA.
Fig. 9.16 Photograph by Robert E. Silberglied. See also F. R. Prete, H. Wells,
P. H. Wells, and L. E. Hurd (Eds.). 1999. The Praying Mantids. Johns Hopkins
University Press, Baltimore, MD.
Fig. 9.17 From LeBas, N. R. and L. R. Hockham 2005. An invasion of cheats: The
evolution of worthless nuptial gifts. Current Biology 15(1):64–67.
Fig. 9.18 Drawing by the authors. Based on Trivers, R. L. 1972. Parental
investment and sexual selection. pp. 1871–1971 In B. Campbell (Ed.), Sexual
Selection and the Descent of Man, Aldine, Chicago, IL.
Fig. 9.20 Photograph by Paul H. Williams, University of Wisconsin.
Fig. 9.21 From Kogan, M. 1975. Plant resistance in pest management. pp. 103–146
In R. L. Metcalf and W. H. Luckmann (Eds.), Introduction to Insect Pest
Management. Wiley, NY. By permission of Wiley-Blackwell.
Fig. 9.22 Drawing by Joan W. Krispyn. See also Daanje, A. 1975. Some special
features of the leaf-rolling technique of Byctiscus populi L. (Coleoptera:
Rhynchitini). Behaviour 53:285–316.
Quote p. 363 from Wilson, E. O. 1975. Sociobiology. The New Synthesis. Harvard
University Press, p. 320.
Chapter Ten

Fig. 10.2 (left) Photograph by Bonnie S. Heim. See also Evans, H. E. and
J. E. Gillaspy. 1964. Observations on the ethology of digger wasps of the genus
Steniolia (Hymenoptera: Sphedcidae: Bembicini). American Midland Naturalist
72:257–280. (right) From Paluch, M., M. M. Casagrande, O. H. H. Mielke. 2005.
Comportamente de agregaçåo noturna dos machos de Actinote surima surima
(Schaus) (Lepidoptera; Heliconinae; Acraeini). Revista Brasileira de Zoologia
22(2):410–418.
Fig. 10.4 Courtesy of Braulio Dias. See also Dias, B. F. 1975. Comportamento
presocial de Sinfitas do Brazil Central. I. Themos olfersii (Klug) (Hymenoptera:
Argidae). Studia Entomologia 18:401–432.
458 Credits
Fig. 10.5 Modified from Smith, R. L.1976. Male brooding behavior of the water
bug Abedus herberti (Hemiptera: Belostomatidae). Annals of the Entomological
Society of America 69:740–747.
Fig. 10.6 Photograph by Robert L. Smith. See also Smith, R. L. 1997. The
evolution of paternal care in the giant water bugs (Heteroptera: Belostomatidae).
pp.116–149 In J. C. Choe and B. J. Crespi (Eds.), The Evolution of Social Behavior
in Insects and Arachnids. Cambridge University
Press, UK.
Fig. 10.7 From West, M. J. and R. D. Alexander. 1963. Sub-social behavior in a
burrowing cricket, Anurogryllus muticus (De Geer). Orthoptera. Gryllidae. Ohio
Journal of Science 63:19–24.
Fig. 10.8 From Gullan, P. J. and P. S. Cranston. 2004. The Insects. An Outline of
Entomology, 3rd ed. Wiley-Blackwell, Oxford. By permission of Wiley-Blackwell
See also Halffter, G. 1997. Subsocial behavior in Scarabaeine beetles. pp. 237–259
In J. C. Choe and B. J. Crespi (Eds.), The Evolution of Social Behavior in Insects
and Arachnids. Cambridge University Press, UK.
Fig. 10.10 Drawings by Sarah Landry from Evans, H. E. and R. W. Matthews
1973. Systematics and nesting behavior of Australian Bembix sand wasps
(Hymenoptera: Sphecidae). Memoirs of the American Entomological Institute
20:1–386. See also Evans, H. E. and K. M. O’Neill. 2007. The Sand Wasps.
Natural History and Behavior. Harvard University Press, Cambridge, MA.
Fig. 10.11 From McCook, H. C. 1909. Ant Communities and How They Are
Governed, Harper & Brothers, NY.
Fig. 10.13 (above) Photograph by Robert E. Silberglied. (below) Photograph by
Carl W. Rettenmeyer.
Fig. 10.14 Drawing by Turid Hölldobler-Forsyth, In Wilson, E. O. 1976. A social
ethogram of the Neotropical arboreal ant Zacryptocerus varians (Fr. Smith).
Animal Behaviour 24:354–363.
Fig. 10.15 From Tophoff, H. 1972. The social behavior of army ants. Scientific
American 227:70–79 (November). Copyright © (1972) by Scientific American,
Inc. All rights reserved. See also Gotwald, W. H. 1996. Army Ants: The Biology of
Social Predation. Comstock Press, Ithaca, NY.
Fig. 10.16 From West-Eberhard, M. J. 1969. The social biology of polistine wasps.
Miscellaneous Publications of the Museum of Zoology, University of Michigan
no. 140, 101 pp.
Fig. 10.17 (left) Drawing by Amy Bartlett Wright from Jeanne, R. L. 1991. The
swarm founding Polistinae. pp. 191–231. (right) Photograph by C. K. Starr from
Credits 459
Turillazzi, S. The Stenogastrinae. pp. 74–98. Both chapters in Ross, K. G. and

R. W. Matthews (Eds.). The Social Biology of Wasps. Cornell University Press,
Ithaca, NY. Reproduced by permission of Cornell University Press.
Fig. 10.18 Drawings by Joan W. Krispyn.
Fig. 10.19 Drawing by Amy Bartlett Wright from Matthews, R. W. 1991.
Evolution of social behavior in sphecid wasps. pp. 570–602 In Ross, K. G. and
R. W. Matthews (Eds.). The Social Biology of Wasps. Cornell University Press,
Ithaca, NY. Reproduced by permission of Cornell University Press.
Fig. 10.20 Photographs by Robert L. Jeanne., University of Wisconsin.
Fig. 10.21 Photograph by Elaine Evans, courtesy of the University of Minnesota
Extension Service.
Fig. 10.22 Photograph courtesy of Michael Schwarz. See also Schwarz, M. P.,
N. J. Bull, and K. Hogendoorn. Evolution of sociality in the allodapine bees: a
review of sex allocation, ecology and evolution. Insectes Sociaux 45: 349–368.
Fig. 10.23 From (left) M. Lüscher, M. 1961. Air-conditioned termite nests.
Scientific American 205:138–145. Copyright © (1961) by Scientific American, Inc.
All rights reserved. (right) Photograph by the authors.
Fig. 10.24 From Eberhard, W. G. 1975. The ecology and behavior of a subsocial
pentatomid bug and two scelionid wasps: strategy and counterstrategy in a host and
its parasites. Smithsonian Contributions to Zoology No. 205. 39 pp. Reprinted by
permission of the Smithsonian Institution.
Fig. 10.25 Diagram by the authors. Based on Lin, N. and C. D. Michener, 1972.
Evolution of sociality in insects. Quarterly Review of Biology 47: 131–159.
Fig. 10.26 Photograph by Justin O. Schmidt.
Fig. 10.27 Drawing by Paul H. Matthews. Based on Wilson, E. O. and
B. Hölldobler. 2005. Eusociality: origin and consequences. Proceedings of the
National Academy of Sciences of the USA 102(38):13367–13371. See also
West, S. A., A. S. Griffin, and A. Gardner. 2007. Social semantics: altruism,
cooperation, mutualism, strong reciprocity and group selection. Journal of
Evolutionary Biology 20:415–432.
Fig. 10.28 Drawing by Sarah Landry from Wilson, E. O. 1975. Slavery in ants.
Scientific American 232:32–36 (June). Copyright © (1975) by Scientific American,
Inc. All rights reserved.
Table 10.1 Compiled from various sources, but primarily based on Wilson, E. O.
1975. Sociobiology. Harvard University Press, Cambridge, MA and on Lindauer,
M. 1974. Social behavior and mutual communication. pp. 149–228 In
M. Rockstein (Ed.) The Physiology of the Insecta, 2nd ed., Vol. 3, Academic
Press, NY.
460 Credits
Table 10.2 Based on Wilson, E. O. 1971. The Insect Societies. Harvard University
Plates
Plate 1 (above) Photograph by the authors. (below) Image downloaded 31 January

2009 from the Virtual Atlas of the Honeybee Brain,
http://www.neurobiologie.fu-berlin.de/beebrain/
Plate 2 Photograph by Darren Wong and David Merritt.
Plate 3 Photograph copyright by Dave Bonta.
Plate 4 (above) Photograph by Greg Sword. (below, left) Image downloaded 3
February 2009 from http://www.nri.org/images/migrantpests2.jpg. (below, right)
Photograph by Diane Earl.
Plate 5 (above) Photograph by the authors. (below) Photographs by Douglas
W. Whitman.
Plate 6 Photograph by Sean McCann, copyright 2006.
Plate 7 Photograph by the authors.
Plate 8 Photographs © Dan L. Perlman/EcoLibrary.org. See also Janzen, D. H.
1967. Interaction of the bull’s horn acacia (Acacia cornigera L.) with an ant
inhabitant (Pseudomyrmex ferruginea F. Smith) in Eastern Mexico. University of
Kansas Science Bulletin 67:315–558.
Plate 10 Photograph by Charlie Charlton.
Plate 11 Photograph by Dave Bonta.
Plate 12 Photograph by Dean Gugler.
Plate 13 From Majerus, M. E. N., F. F. A. Brunton and J. Stalker. 2000. A bird’s
eye view of the peppered moth. Journal of Evolutionary Biology 13:155–159. By
permission of Wiley-Blackwell. See also Majerus, M. E. N. 1998. Melanism:
Evolution in Action. Oxford University Press.
Plate 14 Photographs by Carl W. Rettenmeyer.
Plate 15 Photographs by Carl W. Rettenmeyer.
Plate 16 Downloaded 24 August 2009 from http://topicstock.pantip.com/wahkor/
topicstock/2006/02/X4102342/X4102342-0.jpg. See also Birch, M. C., G. M.
Poppy and T. C. Baker. 1990. Scents and eversible scent structures of male moths.
Annual Review of Entomology 35: 25–54.
Credits 461
Plate 18 Photographs copyright of Masato Ono, Tamagawa University, Tokyo. See

also Ono, M., T. Igarashi, E. Ohno, et al. 1995. Unusual thermal defence by a
honeybee against mass attack by hornets. Nature 377:334–336.
Plate 19 Photograph by Douglas W. Whitman.
Plate 21 (above) Photograph by Anthony O’Toole and David Merritt. (below)
Photograph by David Merritt.
Plate 22 Photograph by James E. Lloyd. See also Lloyd, J.E. 1975. Aggressive
mimicry in Photuris fireflies: signal repertoires by femmes fatales. Science
197:452–453.
Plate 23 Courtesy of John R. Meyer. See also Briscoe, A. D. and L. Chittka. 2001.
Evolution of color vision in insects. Annual Review of Entomology 46:471–510.
Plate 24 Photograph by Klaus Schmitt. See http://www.phase.com/kds315/uv_
photos
Plate 25 Photograph by Robert Duncan, courtesy of Douglas W. Whitman.
Plate 26 Photographs by D. L. Hu. Reprinted by permission of Macmillan
Publishers, Ltd. from Hu, D. L., B. Chan, and J. W. L. Bush. 2003. The
hydrodynamics of water strider locomotion. Nature 424:663–666. See also Wilcox,
R. S. 1979. Sex discrimination in Gerris remigis: role of a surface wave signal.
Science 180:1325–1327.
Plate 27 Photograph by Rex Cocroft. See also Cocroft, R. B. 1999.
Offspring-parent communication in a subsocial treehopper (Hemiptera:
Membracidae: Umbonia crassicornis). Behaviour 136(1):1–21.
Plate 28 Moth photographs from Hristov, N. I. and W. E. Conner. 2005. Sound
strategy: acoustic aposematism in the bat–tiger moth arms race. Naturwisschaften
92(4):164–169, with kind permission of Springer Science+Business Media. Bat
series from Barber, J. R. and W. E. Conner. 2007. Acoustic mimicry in a
predator-prey interaction. Proceedings of the National Academy of Sciences of the
USA 104(22):9331–9334. Copyright (2007) National Academy of Sciences, USA.
Graph by the authors.
Plate 29 Photograph by Douglas W. Whitman. See also Brown, A. W. 1999. Mate
choice in tree crickets and their kin. Annual Review of Entomology 44:371–396.
Plate 30 Image downloaded 14 February 2009 from
http://commons.wikimedia.org/wiki/File:Aphid-giving-birth.jpg
Plate 31 Photograph by Gerald S. Wilkinson, University of Maryland. See also
Wilkinson, G. S. and P. R. Reillo. 1994. Female choice response to artificial
selection on an exaggerated male trait in a stalk-eyed fly. Proceedings of the Royal
Society of London 255:1–6.
462 Credits
Plate 32 Photograph by the authors. See also Gilbert, L. E. 1982. Coevolution of a

butterfly and a vine. Scientific American 110–118.
Plate 33 Photographs by Marshall M. Kerr. See also Waage, J. K. 1986. Evidence
for widespread sperm displacement ability among Zygoptera (Odonata) and the
means for predicting its presence. Biological Journal of the Linnean Society
28:285–300.
Plate 34 Photographs by W. E. Conner and N. Hristov. Conner, W. E., R. Boada,
F. C. Schroeder, et al. 2000. Chemical defense: Bestowal of a nuptial alkaloidal
garment by a male moth on its mate. Proceedings of the National Academy of
Sciences of the USA 97(26):14406–14411. Copyright (2000) National Academy of
Sciences, USA.
Plate 36 Photograph by Susan VanMeter, Hampshire County, WV.
Plate 37 Photograph by Terrence D. Fitzgerald. See also Costa, J. T. 2006. The
Other Insect Societies. The Belknap Press of Harvard University Press, Cambridge,
MA.
Plate 39 Photograph by Christine A. Nalepa. See also Nalepa, C. A. and
W. J. Bell. 1997. Postovulation parental investment and parental care in
cockroaches. pp. 26–51 In J. C. Choe and B. J. Crespi (Eds.), The Evolution of
Social Behavior in Insects and Arachnids. Cambridge University Press, UK.
Plate 40 (left) From Crespi, B. J., D. C. Morris and L. A. Mound. 2004. Evolution
of ecological and behavioural diversity: Australian acacia thrips as model
organisms. Australian Biological Resources Study, Canberra and Australian
National Insect Collection, Canberra. 328 pp. (right) photograph by Laurence A.
Mound.
Plate 41 Photograph by Robert L. Smith. See also García-González, F.,
E. R. S. Roldán, F. Ponz and M. Gomendio. 2007. The adaptive significance of
male egg carrying in the golden egg bug. Ecological Entomology 32:578–581.
Plate 43 Photograph by Alan Melville. See also Kölliker, M. 2007. Benefits and
costs of earwig (Forficula auricularia) family life. Behavioral Ecology and
Sociobiology 61(9):1489–2497.
Plates
Plate 1 Proximal analysis of honey bee foraging. (above) A foraging worker depends on its highly
developed spatial navigation abilities to search for pollen and nectar outside its hive or nest, return
home, and communicate this information to others. (below) Clusters of neurons called mushroom
bodies (shown in red) located at the top front of its brain are involved in spatial learning; the yellow
globes are the optic lobes of the bee brain
463
464 Plates
Plate 2 Amplifying power for a remarkable jump. As the flea crouches before takeoff, a resilin
pad (insert) at the base of the leg is squeezed and two cuticular catches are cocked. When these let
go, all of the energy imparted via the leg muscles is released from the pad in about a millisecond,
thrusting the flea’s hind trochanters against the substrate
Plates 465
Plate 3 Two monarch butterflies, Danaus plexippus, on their host plant, the milkweed Asclepias
tuberosa
466 Plates
Plate 4 Swarming behavior. (top) Two morphs of Schistocerca gregaria; the gregarious form is
on the left, the solitary form is on the right; (bottom left) locusts swarming in Africa; (bottom right)
a group of adults and nymphs feeding on cabbage in a laboratory colony
Plates 467
Plate 5 Aggregating for hibernation. (above) Paper wasps, Polistes carolina, move to sheltered
sites in the fall, then eventually disperse to establish new nests in the spring. (below) Some
coccinelliid beetle species gather in large numbers before hibernating, as this aggregation from
Arizona illustrates
468 Plates
Plate 6 Stylopsized paper wasp. Three strepsipteran parasites protrude from between the abdom-
inal segments of this Polistes exclamans worker. The parasites modify the behavior of the wasp to
their benefit
Plates 469
Plate 7 Cuckoo wasps such as this one, Stilbum cyanura, attack their host, in this case a mud
dauber wasp, by first chewing through the nest wall to reach the helpless offspring. They then
lay a egg on the host using their telescoping terminal abdominal segments. If disturbed, the heav-
ily armored cuckoo wasp can retract its abdomen into a tight ball; its thick cuticle is relatively
impenetrable
470 Plates
Plate 8 Living in a thorn. (above) Nest of Pseudomyrmex ants in a swollen thorn of the bull’s
horn Acacia, which the ant hollows out to house its brood. The entrance hole near the tip is clearly
visible, as are extrafloral nectaries (three swellings on the adjacent leaf petiole). (below) Portion of
an Acacia leaf, showing the protein-rich yellow Beltian bodies on the leaflet tips being collected
by a Pseudomyrmex worker ant
Plates 471
Plate 9 The carrion plant, Stapelia, produces a strong odor like decaying carrion; this is highly
attractive to calliphorid and sarcophagid flies that are duped into depositing their eggs at the
flower’s base. Two flies are visible at the flower base here
Plate 10 An ant, probably Formica, tends a batch of black bean aphids

472 Plates
Plate 11 Vivid aposematic colors characterizes these nymphs of Oncopeltus fasciatus, the
milkweed bug; this warning is backed by chemical defenses gained from the milkweed plant
Plates 473
Plate 12 Two adults of the willow leaf beetle, Plagiodera versicolora; their larvae practice both
group feeding and cannibalism
474 Plates
Plate 13 The two forms of peppered moths, Biston betularia, on foliose lichen, (above) as they
would look in normal ‘visible’ light, and (below) under UV illumination
Plates 475
Plate 14 Three examples of crypsis among insects of the tropical rain forests. Top: nymph of an
unidentified preying mantis. Middle: a well-camouflaged walking stick. Below: a katydid resting
on a moss- and lichen-covered branch demonstrates both crypsis and disruptive coloration
476 Plates
Plate 15 Four examples of apparently Batesian mimics resembling many tropical ‘tarantula hawk’
wasp species. (upper left) a carnivorous katydid from Panama; (upper right and lower left) two
different coreid bugs from Ecuador; (lower right) an assassin bug from Panama
Plates 477
Plate 16 The expanded coremata of a displaying male of an arctiid lekking moth, Creatonotos
478 Plates
Plate 17 A cluster of
Neodiprion sawfly larvae on a
pine branch exhibit their
characteristic defensive
posture, with the anterior
parts of their bodies tilted
backward and droplets of
regurgitated fluids exposed
from their mouths. Vigorous
jerking movements enhance
the effectiveness of this group
display. The secretion
contains primarily
plant-derived substances
Plates 479
Plate 18 The Japanese hornet Vespa mandarinia japonica preys on worker honeybees that it cap-
tures at the hive entrance; instead of attempting to sting or succumbing as most domestic honeybee
species would, workers of the native Apis cerana honeybees grasp the intruder, and dozens more
quickly surround and engulf it into a living, buzzing ball of warm bees, producing a temperature
the bees can withstand but the hornet cannot
480 Plates
Plate 19 Green tree ants

build nests by stitching leaves
together with larva-produced
silk. Dominant canopy ants
throughout the Old World
tropics, Oecophylla form
colonies many thousands
strong and can be ferocious
when their nests are disturbed
Plates 481
Plate 20 Attracting a crowd. The secret to creating a bee beard lies in knowing that the honey
bee queen’s pheromones makes her the center of attention. To make the beard, the queen must be
located in a bee swarm and moved into a small screen cage with an attached string that can be
quickly hung like a necklace. Within minutes the workers sense that their queen is missing and
take flight in a huge cloud; when they discover her, the workers quickly settle as close as possible
to their queen. (She is under the author’s chin, hidden by the swarm.)
482 Plates
Plate 21 Dinner by glowworm light. (above) The predatory Australian glowworm Arachnocampa
flava inhabits caves, where the larvae spin silk hammocks adorned with glistening droplets. (below)
A time-lapse photo of a group of glowworms; the source of one glowworm’s luminescence is
spotlighted in the upper photo
Plates 483
Plate 22 Playing femme fatale, a female Photuris firefly has seized a male of another firefly
species in a fatal embrace after attracting him by mimicking the mating signal of females of that
species. Photuris are such significant predators on other fireflies in the Americas that they are
thought to be the driving force that has caused several firefly species to become diurnal
484 Plates
Plate 23 Perception of color by humans (top), bichromatic insects (center), and (bottom)
trichromatic insects such as honey bees
Plates 485
Plate 24 A bumblebee, here just leaving a Mexican zinnia, sees a very different color palette
than the human who planted the flower. (above, left), human view; (above, right), the same flower
under ultraviolet (UV) light. (below) Simulated bee vision shows the way a flower with a visiting
bumblebee would look to humans if our light sensitivity were like that of the bees. In this photo-
graph of a bee on a yellow flower, colors have been remapped so that UV reflectance is shown as
violet/blue and the whole image only contains UV-blue-green
486 Plates
Plate 25 Butterflies drinking at a mud puddle. When disturbed, they will swirl up together to form
a confusing mass of colorful forms
Plate 26 Using a thin layer of thymol blue on the water surface reveals the way in which a water
strider (Gerridae) propels itself across the surface of a pond or river by hemispherical vortices shed
by its driving legs
Plates 487
Plate 27 An adult female thorn bug treehopper (Umbonia crassicornis) guards her nymphs. They
will use substrate vibrations to signal her as a group if a predator such as a coccinellid beetle
should approach, and she will respond by blocking the invader, fanning her wings aggressively,
and sometimes buzzing
488 Plates
Plate 28 ‘Speaking’ to bats. (A) All possible combinations of palatability and sound production
traits occur naturally in four different species of sympatric North American tiger moths: C+ and
C– refer to presence or absence of defensive chemistry and S+ and S– refer to ability or inability
to produce ultrasonic clicks. (B) Four stages in prey capture. (graph) Predictions arising from three
alternative hypotheses for the function of moth sound (jamming, startle, or warning) are depicted;
any would result in a trained bat aborting its usual prey capturing sequence. Results were consistent
only with a hypothesis of acoustic mimicry (‘observed’ line on graph); apparently the sounds are a
warning to bats that the moths are unpalatable, and the bats quickly learn to ignore clicks that are
not associated with distastefulness
Plates 489
Plate 29 Mating pair of snowy tree crickets, Oecanthus fultoni. For as much as an hour, the female
(above) will remain in position, chewing on a thick glutinous liquid from the male’s metanotal
gland; later she will also consume the white spermatophore that is visible here. Mating in this
subfamily is entirely female-controlled
Plate 30 An aphid giving birth to live young, one of the ways in which aphids reproduce
490 Plates
Plate 31 Stalk-eyed flies, Cyrtodiopsis dalmanni, gather to roost. A male (above) will fight off
other males to be the sole male in such aggregations, in which mating occurs at dusk or dawn.
Males with wider eye spans usually win these altercations, and females prefer to group with males
with the longest eye spans
Plates 491
Plate 32 Heliconius hewitsoni butterfly male, guarding a pupa (attached to a Passiflora vine in a
Costa Rican rain forest) from which his future mate will soon emerge
492 Plates
Plate 33 Ebony jewelwing damselflies, Calopteryx maculata. (above), a male on his territorial
perch. (below), a mating pair in the classic “wheel” position; the female has white wing spots
Plates 493
Plate 34 Chemical defense in the arctiid moth, Cosmosoma myrodora. (A) An aposomatic male
rests on its larval food plant, a source of pyrrolizidine alkaloids (PA). (B) A courting male has just
ejected a flocculent cloud (arrow) that will festoon the female in PA-rich fibers
494 Plates
Plate 35 Mud nest of a large Australian potter wasp, Abispa ephippium. Strictly solitary, one
female builds this fortress and progressively provisions each of up to seven cells. The nest entrance
funnel, thought to play a role in parasite deterrence, is dismantled and constructed anew with each
added cell
Plates 495
Plate 36 A tumblebug (Scarabaeinae) rolls its ball of dung carved from manure. The flesh fly
(Sarcophagidae) riding on the ball also breeds in dung and is a competitor for this rich food
resource. After rolling the ball for some distance, the pair of beetles will cooperatively excavate a
burrow and bury it, thereby making it inaccessible to flies. Underground in their burrow, the pair
may spend long periods preparing the ball to receive their egg
496 Plates
Plate 37 Phelypera distigma weevil larvae in their ‘circle the wagons’ (cycloalexic) defensive
formation between feeding bouts
Plate 38 Mastotermes darwiniensis worker being attacked by three green tree ants, Oecophylla
sp. Worldwide, ants undoubtedly pose the single greatest threat to termites
Plates 497
Plate 39 Parental care in a wood-feeding cockroach, Cryptocercus punctulatus. The mother pro-
tects a clutch of offspring that infect themselves with needed cellulose-digesting symbionts by
feeding on the mother’s feces as well as on fluids from the mother’s hind gut
Plate 40 (left) Australian gall-making thrips, Kladothrips morrisi, display striking polymorphism,
shown by a stout foundress female and her more heavily sclerotized soldier daughter (right) A
foundress female (black) and her offspring inhabit a domicile (here opened) formed from gluing
together two phyllodes of an Acacia
498 Plates
Plate 41 Golden egg bugs, Phyllomorpha laciniata, carry one another’s eggs, keeping them safe
from ant predation. The odd, leafy spines help the bug blend in with dried parts of its host plant
Plate 42 A Manduca hawk moth paying a night visit to a moonflower, Ipomea alba. The flow-
ers have coevolved for moth pollination; they do not open until evening, and they close the next
morning
Plates 499
Plate 43 Maternal care in the earwig, Forficula auricularia. Here, a female retrieves displaced
eggs and returns them to her nest. In addition to guarding the eggs, she will bring food to her
nymphs while they are very young
500 Plates
Plate 44 After spending most of their life feeding on roots underground, cicada nymphs emerge
from the soil in great numbers, split their exoskeleton, and become short-lived adults whose sole
purpose is reproduction. Males produce noisy songs, using their tymbals, and receptive females
respond with timed wing flicks that attract males for mating
Plates 501
Plate 45 Termites swarming from a railroad tie
Plate 46 To a human observer, all bumblebees and carpenter bees appear quite similar, but to a
female Xylocopa virginica, this yellow face mask indicates a conspecific carpenter bee male
Index
A Allee effects, 232, 236

Abedus, 396–400, 428 Allelochemicals, definition, 226
Abies, 127, 169 Allomone, 194, 204, 226, 227, 254–256
Abispa, 401, 494 Altruism, 197, 236, 434, 436, 437
Ablation, 49, 50, 89 Ambush, 50, 146–148
Acanthomyops, 240, 251, 252 Amitermes, 108
Acheta, 316, 317–319, 400 Ammophila, 58, 81, 82, 170, 272, 404
Acid, formic, use in defense, 187, 207 Ampulex, 145, 146
Acilius, 104 Analysis, proximate and ultimate, 40–41
Acripeza, 209 Anaphylaxis, 205
Actinote, 393 Anax, 68
Action potentials, 47, 53, 55, 59, 60, 219, 265 Andrena, 151, 358
Active space, 249, 250–252, 257 Andricus, 348–349
Adjustment, spatial, 93–129 Androconia, 229
Adrenaline, 86 Anisomorpha, 206
Aedes, 12, 132, 324–327, 346 Anolis, 209
Aggregations, 127, 151–168, 197, 214–215, Antennae
406, 420, 472 behaviors involving, 7, 70, 71, 78, 146,
adult, 111, 116, 127, 151–152, 267, 159–162, 183, 219, 228, 236, 246,
281–283, 294, 310, 364–368, 247–248, 254, 317–319, 349, 402,
390–394, 466, 467, 490 417, 433
classification of, 392 as sensory receptors, 49, 60, 107, 109,
feeding, 177–180, 478 134, 182, 221–223, 229, 230, 258,
of immatures, 293, 300, 394, 472, 478, 496 296–297, 298, 312, 323–326, 333,
pheromones in, 224, 231–241, 255, 258, 335
326 Antheraea, 90, 253, 344
See also Clustering Anthonomus, 259
Aggression, 14, 153, 201, 252, 282, 283, Anthropomorphism, 26
314–317, 318, 371, 372, 443 Antiteuchus, 432–434
See also Mimicry, aggressive Ant lions, 147, 401
Agriculture, insect, 154–157 Ants, 2, 13, 26, 29, 31, 36, 68, 69, 72, 81,
Agrotis, 45 105, 109, 113, 125, 128, 129, 139,
Alarm, 27, 172, 207, 208, 214, 226, 240–242, 147, 149, 151, 156–158, 161–163,
251, 252, 254, 256, 257, 258, 170, 179, 181, 183, 184, 197, 201,
283–284, 294, 300, 306, 314–317, 205–208, 213, 214, 220–222, 227,
319, 442 231, 248, 249, 251, 256, 273, 276,
in aphids, 171, 241 300, 389, 392, 296, 404, 406, 407,
in honey bees, 240, 242 409, 418, 420, 422, 427–429, 432,
Allatectomy, 84, 345 434, 436, 438, 441, 442, 498
503
504 Index
acacia, 171–173, 470 Attractants, sexual, 27, 83, 221, 225, 226, 229,
Amazon, 442 250, 253, 254, 257, 344
aphid-tending, 155, 171, 241, 349, 428, Austroplatypus, 428
471 Autographa, 385
Argentine, 19 Autotomy, 213
army, 116, 158, 237, 238, 412, 413–416 Axons, giant, 53
attine, see Ants, leaf-cutter
carpenter, 207, 254, 412 B
citronella, 251 Backswimmers, 261, 298
fire, 131, 204, 237, 238–240, 418 Ballooning, 125
formicine, 186, 208 Bats, 62–68, 99, 194, 303, 319, 320, 321, 461,
green tree, see Oecophylla 488
harvester, 27, 252, 327, 439 Bee beard, 481
leaf-cutter, 107, 124, 139, 154–155, 183, Bee fly, 405
252, 410 Bees, 3, 7, 26, 42, 59, 72, 73, 74, 102, 105,
myrmecine, 412 106, 116, 129, 143, 144, 151, 153,
shampoo, 255 163, 174, 181, 183, 202, 205, 221,
thief, 153, 154 231, 248, 269, 272, 275, 355, 359,
trail-following, 222, 239 391, 404, 409, 420, 424, 427, 434,
velvet, 308, 314, 315, 405 436, 485
weaver, 29, 108, 246, 255, 410 allodapine, 422, 424
wood, 159–160 bumble, see Bumblebees
Anuraphis, 155 carpenter, 319, 371, 501
Anurogryllus, 402 coevolution with orchids, 355–357
Aphids, 59, 121, 125, 140, 141, 144, 148, 155, dances of, 45, 59, 72, 105, 106, 238, 276,
167, 185, 194, 197, 227, 231, 241, 299, 323–338, 420
248, 299, 348, 389 euglossine, 72, 356, 357, 361
alternation of generations, 349 gravity perception, 104, 106
bean, 117, 471 halictid, 420, 422, 438, 440
birth, 348, 489 honey, 8, 20, 21, 42–44, 45, 59, 60, 69, 72,
gall-forming, 409, 425, 427–428 81, 87, 94, 102, 105, 106, 108, 109,
mutualisms, 171 114, 128, 133, 139, 151, 175, 182,
social, 428 221, 231, 238, 240, 242, 254, 262,
Aphis, 349 272, 273, 276–279, 293, 294, 299,
Aphodius, 321 328–339, 380, 389, 407, 409, 410,
Aphrodisiacs, 195, 254 418, 420, 422, 423, 435, 438, 463,
Apiomerus, 170 479, 481, 484
Apis, 42–44, 248, 336, 422 hygienic behavior, 20–22
See also Bees, honey mimicry of, 197, 199, 319
Aposematism, 152, 188, 198–204, 232, orchid, see Bees, euglossine
319–321, 433, 472 retinue behavior, 237, 238, 254
Appeasement, 160, 161, 256, 349, 377 sleeping, 390–393
Apterygote, 342 sociality in, 163, 299, 409, 420–423,
Arachnocampa, 263, 482 434–440
Argogorytes, 358 solitary, 149, 151, 271, 305, 306, 393, 401,
Argynnis, 58, 285, 286–288 406
Argyrotaenia, 218 stingless, 185, 241, 252, 335, 336, 422,
Arms race, see Coevolution 423, 440
Asobara, 145 sweat, 87, 420, 421, 422
Assembly, 231–240 Beetles, 3, 4, 6, 10, 19, 75, 81, 96, 97, 104,
Associations, classification of, 392 105, 110, 111, 113, 116, 121, 122,
Atemeles, 158, 159–161 125, 132, 150, 155, 158, 162, 163,
Atta, 107, 139, 156, 183, 252 165, 170, 186, 197, 201, 206, 208,
Index 505
231, 248, 258, 303, 320, 321, 347, Bioluminescence, 262–268

358, 373, 374, 385, 392, 407 See also Light production
ambrosia, 155, 156, 428 Biston, 191–194, 474
bark, 117, 122, 155, 167, 168, 224, Bittacus, 373
232–236, 251, 255, 317, 328, 342, Bivouac, of army ants, 413, 415
409, 428 Blastophaga, 175–177
blister, 195, 204 Blatella, 236, 237, 393
bombardier, 207, 208, 293 Bledius, 431
buprestid, 47, 107, 383 Bleeding, reflex, 205
carabid, 162, 207, 315 Blood, as insect diet, 86, 114, 132, 148, 325
carrion, 246, 319, 404 Blowflies, 58, 59, 81, 137, 138, 338, 495
cerambycid, 93, 109, 165, 189, 197, 200, Bombus, 115, 422, 424
247 See also Bumblebees
click, 94, 97 Bombykol, 226
coccinellid, 167, 185, 376, 467, 487 Bombyx, 89, 223, 225–227
deathwatch, 291, 294 See also Moths, silkworm
diving, 97, 104 Booklice, 293, 294
dung, 58, 105, 152, 153, 275, 276, 389, Brachinus, see Beetles, bombardier
401–404, 495 Bradysia, 359
jewel, 107 Brain, insect, 40–51, 59, 67, 68, 75, 83, 85–86,
ladybug (lady, ladybird), 116, 144, 203, 89, 90–91, 106, 113, 137, 146, 219,
232 221, 269, 277, 345, 438, 441, 463
See also Beetles, coccinellid Brevisana, 292
lampyrid, see Fireflies Bristletails, 4, 158, 164, 213
leaf, 170, 179–181, 206, 473 Brood, 21, 30, 87, 114, 115, 116, 148, 150,
long-horned, see Beetles, cerambycid 151, 153, 155, 156, 157, 159, 160,
lycid, 194, 200, 232 161, 163, 183, 186, 234, 242, 246,
passalid, 392, 408, 417, 431 346, 387, 389, 398, 399, 400, 402,
potato, 111, 134, 137, 165 403, 404, 408, 414, 415, 416, 418,
rove, see Beetles, staphylinid 422, 423, 432, 439, 442, 443, 470
scarab, 29, 394, 395, 401–402 Bugs (Hemiptera), 84, 101, 105, 107, 111, 117,
scolytid, see Beetles, bark 148, 157, 170, 189, 195, 205, 207,
stag, 362 214, 240, 255, 262, 292, 293, 294,
staphylinid, 98, 158–161, 404, 431 296, 300, 304, 372, 394, 396, 407,
tenebrionid, 208 409, 428
tiger, 94, 147, 204, 269 assassin (reduviid), 86, 92, 147, 170, 189,
whirligig, 96, 109, 296 206, 213, 476
wood-boring, 93, 109, 294, 296 See also Rhodnius
See also Weevils bed, 131, 148, 157, 241, 342
Bee wolf, 59, 73, 133, 404 belostomatid (giant water), 396–401, 428
Behavior bomb-sniffing, 69
altruistic, see Altruism fire, 168–169
biological basis, 2, 3–5, 16, 245, 249, 291, fulgorid, 209, 210
343 golden egg, 321, 428–430, 498
group, advantages of, 56, 64, 83, 105 lightning, see Fireflies
history of study, 1–44, 45, 140, 201 milkweed, 24, 118–121, 472
hormonal coordination, 83–89 pentatomid (stink), 157, 185, 204, 291,
‘retinue’, 237, 238, 254 294, 432–434
social, see Sociality Bullacris, 303
Beltian bodies, 171, 470 Bumblebees
Bembix, 32, 34, 215, 406 foraging, 133, 143
Bertholdia, 320 heat regulation, 114–116
Bioassay, 78, 80, 224, 225, 241 sociality in, 390–397
506 Index
Burrow, accessory, 405 sound detection, 296, 304

Butterflies, 3, 4, 7, 58, 60, 72, 101, 104, 107, tent, 177, 239, 240, 392, 408
110, 11, 114, 115, 116, 160, 167, as wasp prey, 32, 45, 58, 82
174, 191, 197, 199, 200, 203, 212, Cecropia (plant), 171
213, 227, 262, 276, 277, 278, 279, Celerio, 110
283, 289, 293, 375, 384, 392, 393, Central pattern generator (CPG), 51, 52, 53,
486 56, 96
courtship, 284–285, 290, 342 Centris, 369
fritillary, 286–288 Cephalodesmius, 404
grayling, 101, 103, 229 Cephalotes, 139, 156, 268
heliconiine, 164, 170, 385 Ceratitis, 243
See also Heliconius Cerci, 30, 47, 53, 54, 55, 56, 93, 296, 317
monarch, 109, 119, 120, 122, 127–128, Ceropales, 153, 154
167, 187, 195, 197, 200, 229, 276, Cerula, 186
465 Chaeborus, 228
‘owl’, 209, 211 Chalicodoma, 393
queen, 284–285, 370 Chemoreception, 60–62, 137, 219, 221, 222,
swallowtail, 198, 199, 202, 204, 390 231, 243, 258
Byctiscus, 387 Chorthippus, 323
Byrsotria, 344 Chorusing, see Songs, insect
Chrysomela, 206
C Chrysopa, 83, 148
Caddisflies, 38, 39, 56, 97, 98, 125, 148, 204, Chrysoperla, 45
213, 401 Cicadas, 86, 88, 94, 110, 114, 116, 292, 294,
Caedicia, 321 296, 308, 310, 311, 323, 392, 500
Caligo, 209, 211, 212 Cicindela, 94, 269
Calling, 88, 228, 229, 253, 259, 293, 295, 308, Cimex, 241
310, 313, 314, 316, 321, 322, 344, Clade, definition of, 28
359 Cladistics, 33–36, 38
Calloconophora, 328 Clock, 86–88, 89, 90, 94, 99, 112, 128, 135,
Callosobruchus, 384–385 139, 193, 276
Calopteryx, 370, 492 Clunio, 88
Camouflage, see Crypsis Clustering, 114, 152, 163, 179–181, 197, 231,
Camponotus, 254, 412 232, 237, 241, 364, 368, 478, 479
Campsoscolia, 358 See also Aggregations
Cannibalism, 159, 179–181, 183, 371, 377, Clypeadon, 31
378, 385, 426, 473 Coadaptation, see Coevolution, mutualism,
Cantharadin, 195, 373 symbiosis
Caprification, 174–177 Cockroaches, 45, 83, 88, 94, 145, 157, 208,
Care, parental, 177, 300, 380–382, 384, 230, 294, 344, 393, 425, 431, 497
389–407, 428–434, 487, 499 aggregation in, 236–237
Cassida, 206 classification of, 29–31, 409
Castes, 29, 248, 249, 409–412, 417, 418, 422, escape behavior, 52–56
426, 427, 439, 442, 443 feeding responses, 71–72
Cataglyphis, 45 reproduction in, 345–349
Catasetum, 355 Code-breaking, see Eavesdropping
Caterpillar, 61, 77, 85, 96, 186, 225, 269 Coefficient of relationship (r), 436, 437, 439,
defenses, 186–187, 190, 191, 195, 196, 440
202–204, 211, 319 Coevolution, 141, 164–171, 173, 204, 354,
feeding by, 135, 164, 195, 196 364, 365, 378
as hosts, 70, 79–80, 148, 177, 387 Colias, 278, 289–290
orientation, 101, 103, 105 Colletes, 358
Index 507
Colony Crotalaria, 374

cycle, 413–416, 417, 418, 419 Crypsis, 186, 187–191, 194, 197,
odor, 248, 249, 443 279, 475
Coloration Cryptocercus, 423, 431, 497
aposematic, 203, 209, 232, 375, 472 Cuticular hydrocarbons, 231, 246–248, 256,
disruptive, 188, 189, 192, 475 289, 335
flash, 209, 212 Cychrus, 315
Color vision, 277–279 Cyclic-reflex hypothesis, 51
Commensalism, 161, 163 Cydia, 76, 77, 259
Communal activities, 72, 151, 177–184, 215, Cyrtodiopsis, 365–368, 490
392, 395, 400–407, 424, 439
Communication D
acoustical, 257, 291, 307, 313, 315, 319, Damselflies, 4, 33, 282, 343, 369–371, 492
323, 326 Danaus, 119, 127, 285, 465
chemical, 60, 78, 166, 181, 214, 217–259, Dance flies, 228, 272, 378–379
283, 298, 345 Dances, honey bee, 45, 59, 72, 105–106, 238,
mass, 239, 240 276, 299, 328, 329–338, 420
Dance language controversy, 333–335
mechanical, 291–340
Dasymutilla, 314–315, 320
sematectonic, 305
Defense, 185–216
tactile, 306
See also Predation
visual, 257, 261–290
Deinacrida, 213
Competitive exclusion principle, 153
Dendroctonus, 232–234
Conceptual pitfalls, 25–28
Description, pitfalls of, 25–27
Conditioning
Deutocerebrum, 48, 49, 219
classical, 69, 70, 71, 72
Diamesa, 113
instrumental, 69, 72 Diapause, 83, 86, 111, 112, 113, 117, 119–121,
preimaginal, 75 126, 152, 386, 439
Congruency hypothesis, 136 Diaretiella, 140
Conophthorus, 248 Diaritiger, 248
Copidosoma, 148 Diceroprocta, 114
Copris, 395, 402 Dicrocheles, 67
Coprophagy, 402 Dimorphism, sexual, 361–369
Coptotermes, 183 See also Polymorphism
Copulation, 58, 59, 125–126, 247, 293, 343, Dimorphothynnus, 126
346, 352, 370–380, 381, 397–399, Dinoponera, 248
429 Diplura, 343, 387
See also Mating Disguise, 148, 209, 261, 262
Coremata, 219, 229, 477 See also Crypsis
Corixa, 94, 293 Disparlure, 221
Cornicles, 227, 241, 427, 471 Dispersal, 93, 116, 117, 120, 123, 124–129,
Cosmosoma, 374–376 214, 240, 241, 244, 281, 372
Cost/benefit analysis, 393 Displacement activities, 59, 315
See also Optimality theory Displays, 186, 188, 201–203, 228–229, 246,
Cotesia, 75 261, 266–267, 279, 282–284, 289,
Countershading, 188, 190, 191 300, 310, 320, 322, 349, 364, 371,
Courtship, 37, 38, 50, 58, 177, 228, 229, 231, 395, 433, 477
275, 279, 284–290, 295, 316, 326, See also Startle
341–379, 399 Dissoteira, 202
Creatonotos, 477 Division of labor, 392, 407, 409, 410
Crickets, 4, 248, 294, 298, 312, 314 Dominance, 183, 248, 249, 316–318, 371–372,
ground, 74, 295, 312 417, 426, 443
snowy tree, 86, 295, 308, 489 Dormancy, 110–113
Critical period, 75, 85 See also Diapause
508 Index
Dragonflies, 4, 68, 98, 100, 104, 115, 116, 118, Eyes, insect
141, 261, 269, 274, 279, 281, 283, apposition, 270, 271, 274
341, 343, 361, 371, 417 compound, 49, 88, 91, 269, 270, 273, 274,
Drakea, 229, 359 275, 281
Drepana, 319 superposition, 270, 271
Drosophila, 12, 19, 60, 68, 69, 75, 87, 88, 94, Eye spot, 209, 210, 212, 262
97, 109, 113, 121, 132, 145, 231,
248, 361 F
Dysdercus, 101, 117, 240 Fables, insect, 6, 8, 26, 389
Facilitation, social, 393, 394
E Feedback, 51, 52, 56, 57, 59, 84, 138, 240
Ears, insect, 303–305 Feeding, 70–71, 77, 79, 111, 117–119, 122,
Earwigs, 387, 389, 401, 499 123, 124, 126, 131–184, 188, 196,
Eavesdropping, 236, 255, 256, 311, 314 203, 228, 232, 239, 241, 271, 303,
Ecdysone, 85, 168, 169 327, 338, 359, 373, 378, 384–386,
387, 392, 394, 395, 402, 404–405,
Ecdysteroid, 85, 86, 92
408, 415–416, 422, 466, 473, 497
Echolocation, 62, 65, 67, 303
Femmes fatales, 200, 266, 392, 483
Eciton, 116, 412, 413–415
Ficus, 175
Eggs, trophic, 184, 402, 426
Figs and fig wasps, 174, 177
Elaphrosyron, 154
Filter, sensory, 60, 219
Elasmucha, 214
Fireflies, 37, 147, 262–268, 271, 361, 392, 483
Electroantennogram, 222, 258
Fixed action patterns (FAPs), 57, 58, 83, 84
Eleodes, 186
Flash coloration, 209, 212
Endocrine system, 2, 46, 48, 83–85, 117, 226,
Fleas, 86, 94–97, 228, 464
344–345
Flicker vision, 272–275
See also Hormones
Flies
Endosymbiont, 157 bombyliid (bee flies), 405
Endothermy, 114–116 bush, 390, 391
Enhanced fecundity hypothesis, 400 crane, 209, 213, 351–354, 379
Ephippiger, 298 dance, 228, 272, 378–380
Epicordulia, 274 dung, 370
Eryphanis, 209 fruit (Drosophila), 12, 19, 69, 121, 132,
Escape behavior, 46, 52–56, 62–65, 84, 98, 231
101, 209, 212, 262, 290 fruit (Rhagoletis), 243
Ethogram, 13 glowworm, 263, 482
Ethology, 2, 13–15, 42, 53, 81, 261 house, 40, 99, 137, 153, 201, 269, 275, 276
Eucalyptus, 294, 428 Flight, 117, 119, 379
Eufriesea, 357 muscles, 51, 99, 100, 114–117, 212
Euplusia, 74 tandem, 343, 370, 492
Eupatorium, 375 Food web, 141, 142
Euschistus, 3 Foraging, 42–44, 72, 74, 87, 113, 124,
Eusociality, 409–428, 434–440 131–184, 239, 240, 254, 255, 275,
in aphids, 427–428 329–339, 390, 413, 416–418, 425,
in bees, 420–423 463
in thrips, 427 Foraging (for) gene, 44, 438
in wasps, 416–420 Foraging strategies, 140–163
Evolutionary convergence, 36, 187, 198, 252 Formica, 81, 158, 159, 160, 182, 207, 213,
Evolution by natural selection, 2, 9, 15–16, 24, 442, 471
27 Formicoxenus, 255
See also Selection, natural Form perception, 268–272
Exocrine glands, see Glands, insect Foulbrood, 20, 21
Exoneura, 424 Foundress, 151, 416, 417, 427, 497
Index 509
Free run, 88 Hemileuca, 196

Fungiculture, 155–156 Herd, selfish, 214, 236
Heritability, 23, 24, 365
G Herpetogramma, 169
Gall insects, 133, 143, 348, 409, 427–428, 497 Heterotheca, 170
Game theory, 27, 141, 319 Heterothermy, 110, 114
Ganglia, 47–51, 54–56, 72, 81, 145–146, 219, Hibernacula, 111
290, 377 Hipparchia, 229
Gargaphia, 394 Hippodamia, 116
Genetic drift, 17–19, 27, 200 Homeostasis, metabolic, 138
Geotaxis, 105 Home range, 72, 109, 116
Geotrupes, 389 Homing, 109, 127, 229, 276
Gestalt, 57, 188 Homologies, 36, 37, 99
Glands, insect Hopkins’ host-selection principle, 75
Dufour’s, 239, 254, 418 Hormones, 13, 42, 46, 83–89, 90, 121, 168,
endocrine, see Endocrine system 225, 228, 253, 345
exocrine, 219, 220, 252, 254, 255 Hornets, see Wasps, hornet
mandibular, 208, 220, 236, 252, 254, 335 Host-marking, 242–246
prothoracic, 83, 85, 86, 92, 187 Host-searching, 77, 243–244
pygidial, 252 Hyalophora, 90, 218
tergal, 230 Hygienic behavior, 20, 21, 22
Glossina, 114, 248 Hyla, 209, 211
Glowworms, 263, 482 Hyssopus, 76, 77–81
Gomphocerippus, 84
Good genes hypothesis, 368–369
I
Grasshoppers, 18, 26, 48, 72, 84, 94, 97, 110,
Ichneumon, 148, 243–245
121, 122, 123–124, 136, 206, 209,
Imprinting, 248–249
212, 294, 303, 309, 314, 323, 345,
Inclusive fitness, 150, 436, 437
382
Infochemicals, 226
See also Locusts
See also Allomone; Kairomone;
Gravity perception, 104–106
Pheromones; Semiochemical
Gregarization, 123
Innate releasing mechanism, 58, 60, 350
Grooming, 1, 50, 59, 146, 158, 159, 183, 186,
Inquiline, 157, 441
254, 255, 256, 398, 426, 441
Insecta
Group selection, see Selection, group
Groups, simple, 177–181, 390–394 abundance, 199
Group vigilance effect, 393 classification, 3
Gryllus, 317–319 Insemination, 342, 343, 346, 369
Gynes, 152, 417 Insight, 82
Gyrinus, 109 Instinct, 10, 13, 45, 57, 82, 196, 404
Intelligence, 81–83
H Interneurons, giant, 53, 54, 56
Habituation, 69, 221, 259 Interspecific social interactions, 440–443
Hairpencils, 229, 230, 285 Investment, parental, 380–382, 396
Halictus, 420 See also Care, parental
Hamilton’s Rule, 437, 438 Ips, 224, 234
Handicap Principle, 368–369 IRM, see Innate releasing mechanism
Hardy-Weinberg equilibrium, 19 Isolation, reproductive, 136, 231, 267, 314,
Harpegnathos, 248 323, 350
Heat production, 114–116 Isophya, 372
Heliconius, 58, 72, 164, 170, 369, 385, 491
Helicoverpa, 124 J
Heliothis, 122, 124, 294 Johnston’s organ, 299
Hemeroplanes, 197 Jumping, mechanisms, 97, 164, 372
510 Index
Juvenile hormone (JH), 85, 86, 92, 152, 168, M

169, 253, 254, 344, 345, 346 Macrotermes, 156, 411, 425
Maculinea, 148
K Magicicada, 310
Kairomone, 226, 227, 236, 246, 255, 256 Magnetic field orientation, 108–109
Katydids, 114, 189, 209, 213, 308, 372, 475, Malacosoma, 239–240
476 Malvolio (mvl) gene, 44
Kinesis, 101, 102, 297 Manduca, 61, 190, 498
Kleptoparasitism, 152–154, 407 Manipulation, adaptive, 150, 152
Klinotaxis, 102 Mantises, 45, 50, 67, 69, 132, 137, 141, 146,
K -selection, 431 147, 268, 274, 377, 378, 385, 475
Marginal value theorem, 144, 145
L Mating
Lacewings, 45, 67, 83, 148, 213 disruption of, 258, 259
Lanternaria, 210 evolution of types, 321–322, 361–365
Lasius, 105, 155, 222, 240, 248 function and complexity, 45–92, 227–249
Learning, 13, 27, 49, 60, 68–81, 93, 102, 107, physiological control, 84, 114, 133, 168,
139, 172, 199–204, 248, 256, 330, 344–346, 350, 354, 381
404 systems, 380–382
Legs, modifications, 96, 97 See also Copulation; Courtship;
Lekking, 152, 477 Insemination
Leks, 231, 311, 364–365 Mayflies, 4, 88, 98, 132, 275, 276, 344, 386
Lepanthes, 359 Mechanocommunication, 291–339
Leptinotarsa, 165 Megarhyssa, 359–360
Leptothorax, 248 Melanism, in peppered moths, 191–194
Lestrimelitta, 241 Melanophila, 107, 383
Lice, feather, 126, 157 Melipona, 335
Lice, plant, 231, 427 Melittobia, 19, 149, 349, 371, 372, 386
See also Aphids
Memory, 49, 68–81, 198
Light, orientation to, 104–105
Microplitus, 70
Light, polarized, see Polarization
Microstigmus, 420, 421
Light production, physiology of, 263–264
Midges, 62, 88, 99, 113, 115, 125, 143, 228,
Light reception, physiology of, 109, 268–279
299, 392
Limenitis, 197
Migration, 116–129, 276, 392, 413–415
Linepithea, 19
Mikania, 375
Lions, ant and worm, 147, 401
Liris, 81, 404 Mimetic polymorphism, 188, 198
Locomotion, 50–51, 94–100, 103, 105, 117, Mimicry, 95, 162–164, 188, 191, 197–202,
322 209–212, 248, 256, 266, 319, 321,
Locusts, 51–52, 97, 103, 123–124, 135, 136, 358–359, 379, 483
191, 202, 222, 276, 296, 345, acoustic, 319–321, 488
382–383, 392 aggressive, 148, 200, 266
See also Grasshoppers automimicry, 200
Lomechusa, 158 Batesian, 188, 198–202, 320, 476
Lovebugs, 341, 370 Müllerian, 188, 198–200, 320
Lucanus, 362 transformational, 201–202
Luciferase, 263, 264 Wasmannian, 188, 201
Luciferin, 263 Mites, 3, 67, 125, 146, 163, 165
Lures, use of, 147, 258, 259, 375 Mobbing, 215
Lycorea, 230 Model, use in research, 58, 261, 262, 285, 334,
Lymantria, 259 354
Lynchia, 380 Monophagy, 133, 136
Lysiphlebus, 248 Morgan’s canon, 25
Index 511
Mosquitoes, 12, 57, 84, 99, 121, 125, 132, Nuptial gifts, 372–378
148, 157, 277, 294, 299, 311–312, Nutrition, 131, 134, 181, 240, 386, 400, 426
323–327, 346, 359, 383, 386 Nymphalis, 262
Moths, 30, 60, 88, 103, 104, 111, 115, 121,
122, 140, 146, 164, 167, 169, 185, O
186–187, 190, 196, 197, 204, 205, Occam’s razor, 25
209, 217, 218–219, 229, 290, 319, Ocelli, 269, 334
344, 370, 373, 374, 400 Ochthera, 279
arctiid, 219, 294, 319, 374–376, Odontotaenius, 408, 431
477, 493 Odors, 60–61, 70–71, 75–81, 133, 134, 150,
cecropia, 90, 111, 218 160–162, 167, 172, 182, 206, 211,
codling, 75–80, 259 217–260, 331–335, 426, 471
gypsy, 85, 111, 125, 131, 164, 221, 250, See also Pheromones
259 Oecophylla, 246–247, 255, 480, 496
hawk (hornworm, sphinx), 110, 116, 133, Oenothera, 87, 279
174, 291, 294, 303, 383, 498 Oligophagy, 133
noctuid, 62–68, 120, 124, 148, 385 Ommatidia, 104, 269–275
peppered, 191–194, 474 Omophron, 320
polyphemus, 60, 167, 253, 344 Oncopeltus, 118–119, 121, 472
silk, 89–90, 132, 253 Onthophagus, 403
silkworm, 9, 89, 221, 223, 225–226, 251 Oogenesis-flight syndrome, 120–121
tiger, 194, 320–321, 488 Ophrys, 357–358
Motivation, 58–59, 68, 83, 283 Optimality theory, 41, 140–145
Musca, 137, 391 Orchid pollination, 34, 58, 229, 354–359
Mushroom bodies, 42, 43, 49, 60, 463 Organization, social, 241, 390–407, 408
Myrmecophiles, 158–163 Orientation, 73, 100–109, 126–129, 276, 297,
Myrmica, 160, 161, 254, 255, 410 305–306, 333
Myzus, 241 Oscillogram, 308–309
Oscinella, 121
Osmeterium, 204, 211
N Ostrinia, 290
Nasonia, 295, 344 Oviposition behaviors, 30, 243–244, 346,
Nasutitermes, 208, 213, 386 382–387, 396–400
Nature–nurture controversy, 82
Nauphoeta, 230 P
Navigation, 42, 101, 105, 109, 124, 126, 128, Paedogenesis, 349
129, 282, 463 Palmacorixa, 293
Nectar guides, 279, 485 Panesthia, 431
Nemobius, 72, 295 Paper factor, 168
Neobarettia, 213 Papilio, 198, 202, 211
Neoconocephalus, 114 Parasites and parasitoids, 31–33, 67, 75–81,
Neodiprion, 232, 243, 244, 478 124, 125, 134, 140, 142, 145,
Neopyrochroa, 373 148–152, 243–245, 248, 255, 295,
Nephila, 374 299, 301, 306, 349, 359–361,
Nerve cord, 49, 54–56, 67 371–372, 385, 386, 387, 405–406,
Nervous system, 46–68, 83–90, 121, 139, 146, 433–434, 441–442, 468
219, 226, 227, 253, 277, 345, 441 Parasitism, social, 426, 441
Nests, 105, 157–163, 401, 421, 470 Parasitism, temporary, 443
Neural inhibition, 50 Parental behavior, see Care, parental
Neuropil, 47, 49 Parischnogaster, 418
Nicrophorus, 246, 404 Parthenogenesis, 347–349
Nomadacris, 135, 345 Parsimony, principle of, 25
Noradrenaline, 86 Passiflora, 164, 170, 491
Notonecta, 261 Patterns, fixed action (FAP), 57–58, 83–84
512 Index
Patterns, repeated motor, 50–56 Polymorphism, 123, 188, 198–199, 409–410,

Pemphigus, 125 497
Perga, 393, 394 Polyphagy, 132, 136
Period (per) gene, 112 Polyphenism, phase, 123, 466
Periplaneta, 30, 53–56, 69, 71, 145 Potentilla, 279
Phaeophilacris, 295 Predation, 53, 62–64, 75, 141, 146–148,
Phanuropsis, 433, 434 164, 185–216, 262, 274, 320, 392,
Phase, nomadic vs. statary, 412–416 396–399, 412, 479, 483, 488
Pheidole, 113 Prey capture, 81, 94, 133, 145–148, 274,
Pheromonal parsimony, 252 378–380
Pheromones, 217–260, 284, 285, 288, Prey theft, see Kleptoparasitism
289–290, 326, 331, 344–345, 359, Prey transport, 31–33, 73
375, 380, 415, 420, 426, 442, 481 Prociphilus, 171
See also Allomone Promiscuity, 380–382
Philanthus, 73, 133, 272, 404 Proprioceptors, 104, 298, 299, 304, 332
Phloeophana, 394 Proteomics, 146
Phlugis, 309 Prothoracicotropic hormone (PTTH), 85, 92
Phonatomes, 307–308 Protocerebrum, 48, 49, 50
Phoresy, 125–126 Pseudabispa, 32
Phormia, 137, 138 Pseudergates, 426
Photinus, 37, 38, 200, 262, 264, 266, 267 Pseudocopulation, 359
Phototaxis, 102, 105, 278 Pseudomyrmex, 171–173, 470
Photuris, 37, 147, 200, 264, 266, 483 Psychology, 2, 13, 14
Phrixothrix, 262 Ptilocerus, 147
Phyllomorpha, 428–430, 498 Puddling, 283, 486
Phyllonorycter, 299 Pulse, in insect song, 308
Phylogenetic systematics, see Cladistics Pyrophorus, 263
Phylogenetic tree, 35, 36, 38, 115, 420 Pyrrhocoris, 168
Phylogeny, 28–38, 263, 387 Pyrrolizidine alkaloids (PA), 374, 375, 493
Physogastry, 163, 411, 427
Phytophagy, see Herbivory Q
Pieris, 135, 384 Q/K ratio, 249–251
Pimpla, 140 Queen substance, 254, 420
Plagiodera, 179–181, 473 Quiescence, 111, 112, 397
Plathemis, 282–283 See also Diapause
Platymeris, 206
Platypus, 428 R
Plecia, 341 Radiation, adaptive, 32–34, 165
Pleolophus, 243–245 Ranatra, 94
Plodia, 229 Reaction chain, 58, 285, 343, 350, 351
Podisus, 255 Reactions, types of, 104–105
Pogonomyrmex, 248, 252, 327, 393, 439 Recognition, chemical, 228, 246–249
Polarization, 105, 128, 129, 275–276 Recruitment, 231–242, 246, 247, 252, 255,
Polistes, 7, 150–152, 248, 280, 283–284, 417, 328–339, 350, 392, 426
418, 467, 468 Reflex bleeding, 205
Pollination, 11, 173–177, 279, 354–359, 498 Reflexes, 50–56, 89, 96, 205
Pollinia, 355–358 Releaser, 57–60, 84, 134, 182, 188, 226, 227,
Polybia, 418 255, 261, 281, 314, 344
Polyergus, 442 Reproduction, modes of, 346–349
Polyethism, 410 Reproductive ground plan hypothesis, 439
Polygamy, 380 Resilin, 97, 464
See also Mating, systems Reticulitermes, 183, 248
Polygyny threshold model, 361 Rhabdoms, 269, 270, 271
Index 513
Rhagoletis, 243 Social Register, 401, 407–428

Rhagovelia, 94 Solenopsis, 204, 238–239, 418
Rhodnius, 84–85, 92, 94 Sonagram, 308–309
Rhopalus, 205 Songs, insect, 51, 307–311, 392
Rhyparobia, 345 Sound, 186, 202, 207, 233–234, 257, 281, 283,
Rhythms, circadian, 43, 87, 88, 89, 118, 253 291–340, 408, 488
See also Clock Spanish Fly, 195
Rhythms, gated, 89–91 Sperm competition, 343, 350, 430
Rhythms, reiterative, 86–88 Sperm mixing, 430
Ritualization, 261, 378 Sperm precedence, 370, 399, 400, 430
Robots, 56, 94, 95, 331, 334 Sperm transfer, 343
Rotational orientation hypothesis, 128 Sphecodogastra, 87
SPM, see Secondary plant metabolites
S Spodoptera, 124
Sarcophaga, 113 Spontaneous generation, 6, 7, 9
Sawflies, 96, 167, 177–178, 187, 232, 243, Springtails, 146, 164
244, 293, 393, 394, 395, 478 Startle, 54, 55, 56, 188, 204, 209–213, 215,
Scaptotrigona, 440 315, 488
Scatophaga, 370 Stenus, 98
Scarabeus, 275 Stictia, 152
Schistocerca, 123–124, 276, 294, 345, Stilbocoris, 373
382–383, 466 Stimulus, 59–83
Search image apostatic, 199, 204 conditioned, 13, 69–70, 278
Secondary plant metabolite (SPM), 166–169 filtering (tuning), 60–68, 219, 272, 288,
Seducin, 230 312
Selection, artificial, 22, 23, 24, 365 generalization, 200, 202
directional, 22–23, 191 orientation to, 101–109
disruptive, 22 sign, see Releaser
group, 181, 436, 440 supernormal, 58, 288, 379
kin, 436–440
token, 58, 134
natural, 2, 9, 15–29, 40, 58, 140, 214, 235,
Stoneflies, 293
256, 359, 362, 400, 434–435
Strepsipterans, 150–152, 468
r and K, 388
Streptomyces, 155
sexual, 199, 228–229, 310, 344, 361–371,
Stridulation, 293
379, 382, 390
Selective attention, 311–312 Stylopization, 150, 152
Semiochemicals, 226, 227, 256, 258, 335 Stylops, 151, 468
Senotainia, 33 Subgenual organ, 299
Sensilla, 47, 107, 219, 221, 297, 298, 299, 304, Summation, spatial and temporal, 53
323, 383 Superorganism, 436
Sensitization, 69, 80 Suprachiasmatic nucleus (SCN), 86
Sex attractant, see Pheromones Swarm-founding, 418
Signaling, multimodal, 288–290 Swarming, 214, 326–328, 392, 501
Signals, honest and dishonest, 198–199, 290, honey bee, 242, 254, 481
368 locust, 123–124, 266, 466
Silverfish, 4, 86, 213 mating, 364
Sinigrin, 140, 166, 167 Swimming, see Locomotion
Slavery, 441 Syconium, 175, 176
Sociality, 249, 389, 392, 395–396, 409, 439 Symbionts, 155–159, 163, 410, 431, 441, 443,
implications and correlates, 428–443 497
paradoxes, 434–440 Sympiesis, 299
pathways to, 434–435 Synapsis, 402
See also Eusociality Syritta, 272
514 Index
T Vespa, 242, 479

tau gene, 87 Vespula, 114, 183, 419, 440
Taxis, 101–103 Vibrations, communication by, 295–303, 313,
See also Geotaxis; Klinotaxis; Phototaxis; 321
Vibrotaxis See also Communication, tactile
Taxon, definition of, 28 Vibrotaxis, 299
Teleogryllus, 322 Vision
Teleology, 26 acuity, 67, 272–275
Teleutomyrmex, 441 receptors, 279
Termites, 30–31, 108–109, 155–157, 163, 181, ultraviolet, 277, 278, 289, 474, 485
183, 208, 213, 220, 248, 252, 294, See also Polarization
316, 386, 392, 409–411, 423–427,
496, 501 W
Termitophiles, 157, 163 Wasps, 32, 82, 134, 145, 358, 361, 371, 407,
Territoriality, 153, 245, 246, 371–372, 416–420, 438
426 coevolution with orchids, 358–359
Tetramorium, 441 cuckoo, 153, 154, 469
Thaumetopoea, 196 fig, 174–177, 354
Themos, 178, 395 gall-making, 348–349
Thermoregulation, 93, 110–116, 119, hornet, 242, 328, 416, 418
240 ichneumonid, 140, 243–244, 360
Thrips, 94, 409, 427–428, 438, 497 paper, 7, 150, 249, 280, 283, 407, 416, 417,
tim gene, 87 467, 468
Tinbergen’s Four Questions, 40 parasitoid, see Parasites and parasitoids
Tiphia, 365 sand, 32, 45, 56, 58, 215, 405, 406
Tipula, 351–354 scelionid, 124, 433
Token stimuli, 134 social, 416–420
Tommy Tucker syndrome, 173–177 solitary, 31, 149, 152–154, 204, 231, 306,
Touch, see Communication, tactile 405
Toxorhynchites, 312, 326, 327 spider, 153, 154
Triatoma, 12 thynnid, 229, 359
Trichogramma, 134 yellowjacket, 407, 419
Trigona, 241, 252, 320, 335, 423 Webspinners, 389, 392
Trissolcus, 242, 433 Weevils, 122, 170, 242, 259, 384, 387, 496
Tritocerebrum, 48, 49 Wings, see Flight
Triungulin, 150, 151, 152
Trophallaxis, 181–184, 237, 256, 416,
X
422, 426
Xenos, 150–152
Trophic eggs, 183–184, 402, 426
Xestobium, 294–295
Tropisternus, 320
Trypanosoma, 12
Trypodendron, 156 Y
Yellow fever, 11, 12, 324, 346
U Yellowjacket, see Wasps, yellowjacket
Ultrasound, 63–68
Ultraviolet, perception, 277 Z
See also Vision, ultraviolet Zacryptocerus, 412
Umbonia, 300–303, 487 Zeitgeber, 88
Umwelt, 279, 286 Zombification, 145, 146, 149
Zonosemata, 201
V Zooptermopsis, 183, 316
Venom, 145–146, 205, 435 Zorapterans, 146
Venturia, 75 Zygaena, 370
Vermileo, 147 Zyras, 248

Insect Behavior 2ed PDF

Uploaded by

Copyright:

Available Formats

Insect Behavior 2ed PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Insect Behavior 2ed PDF

Uploaded by

Copyright:

Available Formats

Insect Behavior

Robert W. Matthews · Janice R. Matthews

ISBN 978-90-481-2388-9 e-ISBN 978-90-481-2389-6

© Springer Science+Business Media B.V. 2010

Cover design: Boekhorst Design BV

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)

Like any writers of a general textbook, we recognize a deep obligation to many

Athens, Georgia Robert W. Matthews

1 The History and Scope of Insect Behavior . . . . . . . . . . . . . . 1

5.3.2 Systemic Defenses: ‘I’m Noxious!’ . . . . . . . . . . . 194

8.2.2 Distance and Substrate . . . . . . . . . . . . . . . . . 295

10.4.1 The Ecology of Parental Care . . . . . . . . . . . . . . 430

1.1 Foraging Onset in the Honey Bee, Apis mellifera . . . . . . . . . . . . . . . . . . 42

1.1.1 What Is Insect Behavior?

R.W. Matthews, J.R. Matthews, Insect Behavior, 2nd ed., 1

1.1.2 Insect Behavior’s Biological Context

1.1.3 Historical Foundations

1.1.4 The Watershed Years

establish standards of observational patience and accuracy that subsequent workers

1.1.5 The Rise of Ethology

philosophical foundation. Whereas the comparative psychologists focused on labo-

1.2 Conceptual Frameworks

1.2.1 Evolution by Natural Selection

Saint-Hilaire, a prominent colleague of Jean-Baptiste who expanded and defended

1.2.2 Genetics and Behavior

been traced to mutations of single genes. Intensive inbreeding of laboratory stocks

Parents F1 hybrids Backcrosses (n=29)

Females showing Non-hygienic generation F1 males X original hygienic females = ?

If 1/2 hygienic: 1 gene

If 1/4 hygienic: 2 genes

14 non-hygienic (6 of these would

Rothenbuhler’s model has been widely cited as an instance where a natural

With artificial selection of the type called a ‘truncation selection experiment’ a

Fig. 1.10 Determining heritability. See text for further explanation

What would happen if artificial selection were attempted on an inbred homozy-

1.2.3 The Comparative Approach

only data gathered by direct experimentation and observation could be used to

1.2.4 Conceptual Pitfalls

A third language-related problem is that, because most behavior is described in

1.3 Phylogeny’s Role

Phylogeny, the presumed history of ancestry of a group of taxa, provides a strong

1.3.1 Microevolution and Macroevolution

easily recognized phenomena: domestication, and observations of natural selec-

1.3.2 Phylogenetic Systematics and Cladistics

To deduce the evolutionary direction of apparent behavioral progressions, a useful

1.3.3 Behavior and Speciation

Can behavioral patterns be used like morphological or genetic characters in con-

in minor morphological details. In many places, two or more species of Photinus

1.4 Questions and Perspectives

Types of Types of explanation

Proximate—how Causational—mechanistic Developmental—explanations based on

1.4.1 Proximate and Ultimate Analyses

1.4.2 Types of Approach

In addition to the differing viewpoints of proximate and ultimate analyses, scientists

Case Study 1.1: Foraging Onset in the Honey Bee, Apis

Proximate causation can be considered to have four components—hormonal,

their mushroom bodies resembled those of the normal-aged foragers. Thus,

R.W. Matthews, J.R. Matthews, Insect Behavior, 2nd ed., 45