L0020 Saito 2010 Plant Mites and Sociality Diversity and Evolution
L0020 Saito 2010 Plant Mites and Sociality Diversity and Evolution
L0020 Saito 2010 Plant Mites and Sociality Diversity and Evolution
i
iii
Yutaka Saito
© Springer 2010
This work is subject to copyright. All rights are reserved, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting,
reproduction on microfilm or in any other way, and storage in data banks.
The use of general descriptive names, registered names, trademarks, etc. in this publication does not
imply, even in the absence of a specific statement, that such names are exempt from the relevant protective
laws and regulations and therefore free for general use.
Front cover:
Top: Male prey counterattacking larval predator
Lower left: Two-spotted spider mite, Tetranychus urticae
Back cover:
Upper left: Social cheyletid mite waiting for prey in a sit-and-wait stance
Upper right: Symbiosis between a spider mite and a psyllid?
Background, front and back covers: Sasa ground cover in Hokkaido forest
This book is on mite sociality. I have studied spider mite diversity for 35 years
mostly from the perspective of evolutionary ecology, and this led to the discovery
of several new phenomena on sociality in plant mites. My main interest is similar
to that of Jean-Henri Fabre, the famous French entomologist, although my studies
have been restricted to a few mite groups and yield no comparison to his broad,
encyclopedic knowledge. On the other hand, I used some modern methodologies
that could not have been adopted in his period. Of course I would not dare to com-
pare my contributions to those of such a great entomologist, but I believe that the
studies described in this book bear some resemblance to his work. At least I have
been fortunate to learn about the wonderful world of mites, and I hope the readers
of this book will experience this excitement as well.
When I first planned to write this book, several colleagues advised me not to
focus just on my own work but to address wider themes in mite biology. I have
surely tried to do so, but now in retrospect I am wondering whether I achieved this
goal, because my target mite species are limited to a narrow area of Asia and may
not be representative of all mites. However, I suspect that these “extraordinary”
species may well be distributed over a much larger part of the world: they may
simply not have been discovered and studied so far. Therefore, if someone wishes
to study them after reading this book, it will give me considerable satisfaction.
Mites are small animals that live in a world they cannot see. They are also hard
to see by the naked eye because of their very small body size. Perhaps this is why
they were named “mite” instead a more impressive name like “dino-” or “dyna-” mite.
However, as I will explain below, this animal group comprises a large number of
species showing great diversity. My studies have been restricted to only a few
groups, in particular the family Tetranychidae, which are generally known as spider
mites. If I succeed in showing how diverse such a small group of mites in nature
can be, it may leave the reader with an impression of how little is known and how
mysterious the world of mites really is. In this book, I will address several impor-
tant findings in spider mite evolution, species radiation, and behavioral ecology.
These findings may reflect the complex evolutionary history of these taxa and also
help to point out clearly what is known and what is not yet known to date. I believe
that such a book could be useful for young researchers with an interest in this
v
vi Preface
vii
viii Contents
8 Further Research for Mite Behavioral and Evolutionary Ecology ...... 153
Afterword......................................................................................................... 155
Contents ix
The Acari represent a subclass of the class Arachnida. They are divided into two
orders: Parasitiformes and Acariformes. Although there have been several discrep-
ancies in the classification of higher taxa, I consistently adopted the classification
of Krantz and Walter (2009) in this book, this being the most standard treatise on
systematics of higher taxa above family level. In this book, I mainly address the
mite species of two suborders: the Actinedida (Prostigmata), to which belong the
families Tetranychidae, Eriophyidae, Cheyletidae, and Tenuipalpoidini; and the
Gamasida (Mesostigmata), to which belong the families Phyotoseiidae, Laelapidae,
and Varroidae. I have omitted the authority names in the text when describing spe-
cies to improve the readability of the text, hoping that taxonomists will not criti-
cize me for this approach. Instead, I have attempted to spell out the full species
epithets in the tables whenever possible.
The features that discriminate Acari (mites and ticks) from other arthropods derive
from their special morphological characters. Among these, body size should receive
attention first, because most mites are smaller than 1 mm (Walter and Procter 1999),
Amblyseius
Euseius
Genera appear Neoseiulus Details in Box 2.1
in this book Phytoseius
Diccrocheles Sancassania
Phytoseiulus Agistemus Tarsonemus
Hypoasis
Typhlodromus
Hemicheyletia
Macrocheles Varroa Brevipalpus Ensliniella
Teranychidae Tarsonemidae
Laelapidae
Family Phytoseiidae Varroidae Eriophyidae
Stigmaeidae Acaridae
Macrochelidae Cheyletidae Tenuipalpidae
Winterschmidtiidae
Ixodida Acaridida
Suborder Gamasida (Mesostigmate) Oribatida
Holothyrida Actinedida
(Prostigmata)
Opilioacarida
Order Acariformes
Parasitiformes
Subclass Acari
Fig. 1.1 Brief classification of subclass Acari according to Krantz et al. (2009). Only main
groups involving species addressed in this book are shown
Fig. 1.2 Two-spotted spider mite, Tetranychus urticae Koch, on kidney bean leaf, the most
famous species worldwide as an agricultural pest. {Micrograph by Y. Saito under scanning elec-
tron microscope (SEM) with cryo-system attached}
Perhaps because they are wingless, some mite species have silken threads to
assist in movement and attachment (Saito 1977a; Walter and Proctor 1999). This
ability must diversify the lives of several mite taxa in their habitats, as described
hereafter. Spider mites, the main “cast” of this book, typically produce silken
threads and use these in various ways that represent an adaptation to survive on the
leaf surfaces of their host plants.
Mitchell’s (1973) study on the life history of Tetranychus urticae in connection
with the r/K selection theory, Helle and Overmeer’s (1973) work on genetics, and
the work by Potter et al. (1976) on sexual selection are considered to be classic
studies on spider mite evolution. After that, various studies on spider mites were
reviewed in the book Spider Mites. Their Biology, Natural Enemies, and Control
(Helle and Sabelis 1985), published about 24 years ago. In the same book, Crozier
(1985), a researcher famous for his work on social insects, stressed the availability
of spider mites as model organisms for studying evolutionary ecology and genetics.
I also had the honor of participating in this book, addressing the diversity of spider
mite life types in relation to the use of silk for constructing webs and the associated
defecation behavior (Saito 1985). At that time, diversity was studied quite sepa-
rately from these behavioral features, but I now believe that these two are really
closely associated with each other.
Chapter 2
Plant Mites
“Plant mites” are not a taxonomically defined group, but a collective term that is usu-
ally denoted as “plant-inhabiting mites.” This group consists of many mite species
belonging to different taxa, such as the phytophagous mite families Tetranychidae,
Eriophyoidea (superfamily) and Tenuipalpidae, and the predacious mite families
Phytoseiidae, Stigmaeidae, Cheyletidae, Cunaxidae, and Anystidae. Furthermore,
several mite taxa including Tydeidae, Tarsonemidae, and arboreal Oribatidae mites are
also included (Table 2.1), most of which are considered to be polyphagous. On a par-
ticular plant species, researchers tend to focus on a particular mite species and to
ignore the others occurring there that are from different taxa. However, as is discussed
in this book, almost all mite species must interact with each other on the plant, and
such interactions must be important to understand the real world where these micro-
arthropods are living, i.e., their ecology. This is one reason why we adopt such a col-
lective term, plant mites. Furthermore, some plant mite species other than phytophagous
species interact directly with host plant species; e.g., Walter (1996) reported that leaf
domatia made by host plants greatly affect the plant mite fauna, which may be
regarded as a strategy of host plants. Fournier et al. (2003) also reported that plant
physical structures such as domatia, depressions, and leaf roll formation by insects
must have important effects on competition and predator–prey interactions. As such,
the plant leaf is a kind of micro-sphere where various mites play evolutionary games,
such that we call them plant mites in the same sense as aquatic and soil mites. Among
plant mites, Tetranychidae and Phytoseiidae are the most general families in agricul-
ture and forestry, and many studies have been conducted on these groups of mites so
far (Jeppson et al. 1975; Helle and Sabelis 1985; Walter and Proctor 1999). How these
mite groups diversified on plants and what factors have caused their evolution on par-
ticular host plants are the main themes of this book. For this purpose, I have to address
many mites with their Latin scientific names and their abbreviations (Box 2.1).
Spider mites (Tetranychidae) are the main group of plant-inhabiting mites (a detailed
classification appears in Box 2.2). According to Bolland et al. (1998), who provided
a catalogue of spider mite species of the world, 1,189 species had been described
by the end of December 1996. In the more than 10 years since this publication,
there have been many newly described species, however.
In this family, there are two subfamilies, Bryobiinae and Tetranychinae. The latter
is a very famous group because it involves many pest species of agricultural crops,
orchards, and forest trees. Within the subfamily Tetranychinae, there are three
tribes, Tenuipalpoidini, Eurytetranychini, and Tetranychini, and the species belonging
to the latter two tribes form the main “cast” of this book. Although it is taxonomi-
cally not accurate, I refer simply to “spider mites” as a collective name for the last
two tribes. Thirty-two genera are known in spider mites: 11 in Eurytetranychini and
21 in Tetranychini (Bolland et al. 1998).
Ehara (1999) reviewed the Japanese species of Tetranychidae and showed that
there are three genera in the Eurytetranychini and eight (seven in accordance with
Bolland et al. 1998) genera in the Tetranychini. In these two tribes, four species in the
former family and 66 species in the latter family have been described (Ehara 1999).
Since 1999, one genus and several new species have been added to the Tetranychini
(e.g., two species of Stigmaeopsis were newly described by myself). The present
book is not a taxonomic review. Here, it suffices to know that there are more than
1,000 species of Tetranychidae in the world and also more than 70 species in Japan.
8 2 Plant Mites
It is known that there are special food habits among the genera of the Tetranychini
(Table 2.2). For example, the genus Stigmaeopsis only inhabits graminid plants,
such that it is regarded as monophagous (Saito et al. 2004). On the other hand, the
genus Tetranychus inhabits a wide variety of plants and is regarded as polyphagous
(Gutierrez and Helle 1985). The genera Panonychus and Oligonychus are considered
intermediate to Tetranychus and Stigmaeopsis. Eotetranychus and Schizotetranychus
also occur on a wide variety of tree species. However, such an overview at the genus
level is not always appropriate, if we focus on the individual species of each genus:
the species of Eotetranychus and Schizotetranychus are almost all host specific,
e.g., Eotetranychus tiliarium occurs only on Tillia spp., Eotetranychus shii only on
chinquapins, Schizotetranychus schizopus only on willows, S. cercidiphylli only on
katsura, S. lespedezae on Japanese wisteria, and so on. Thus, at the genus level, they
have a wide variety of hosts, but at the species level, almost all species inhabit only
a very narrow range of species of host plants. Therefore, there is little sense in
saying that a spider mite genus has a certain type of food habit, as pointed out by
2.2 Life History and Host Plants 9
Table 2.2 Brief summary of food habits of genera in Tetranychidae (restricted to the genera
distributed in Japan)
No. of No. of Evergreen Annual or Host
species in species Tree or or deciduous perennial specificity at
Genus the world in Japan grass tree grass species level
Aponychus 19 2 Grass – Perennial High
Eutetranychus 32 1 Tree Deciduous – High
Amphitetranychus 3 2 Tree Deciduous – High
Eotetranychus 175 20 Tree Both – High
Oligonychus 194 17 Both Both Perennial Relatively
high
Panonychusa 15 8a Both Both Perennial High
Schizotetranychus 114 6 Tree Both Perennial High
Stigmaeopsis 7 5 Grass – Perennial High
Tetranychus 135 10 Both Deciduous Both Lowb
Yezonychus 1 1 Grass – Perennial High
a
Sasanychus inhabiting perennial bamboo is considered to be synonym of Panonychus in accor-
dance with Bolland et al. (1998)
b
There are exceptions such as Tetranychus ezoensis Ehara.
Gutierrez and Helle (1985). Rather, such trends at the genus level mentioned above
should be considered historical trends and should be reanalyzed from a phyloge-
netic point of view.
Saito (1979a) and Saito and Ueno (1979) showed that there is much variation in life
history patterns among six spider mite species (see the typical three species in
Fig. 2.1) and discussed that such diversity might be explained by the spatiotemporal
stability of host plants. For example, a famous pest species, the two-spotted spider
mite [Tetranychus (Tt.) urticae], has nearly the highest rm (intrinsic rate of natural
increase) among spider mites and Stigmaeopsis (Schizotetranychus) longus (long
seta form of Schizotetranychus celarius, before Saito 1990a) shows very low rm.
Mitchell (1973) stated that the high rm of Tt. urticae is a characteristic of colonizing
species that are found on ephemeral plants (annual host plants) carrying a huge
amount of resources, and that factor rapidly increases its population. This type of
life history is advantageous if resources emerge randomly. A high potential for
population increase must evolve in a species to use such habitats effectively and to
produce sufficient dispersers to find new resources arising unexpectedly. On the
other hand, in stable habitats with sufficient resources, so many competitors and
natural enemies will coexist (or rather accumulate) and prey species have to spend
much energy in the struggle for survival. Therefore, they should develop a kind of
protection and/or defense mechanism instead of high reproductive rates as an evo-
lutionary trade-off. In general, the spatiotemporal stability of host plants must be
very low in the host plants of Tt. urticae in comparison with those of St. longus.
10 2 Plant Mites
Time in days
Fig. 2.1 Various life histories known in Tetranychinae explained from the spatiotemporal stability
of host plants under 25°C, 50–60% r.h. (after Saito and Ueno 1979). Horizontal bar shows mean
developmental period of each species.
The leaf disc method (sometimes called detached leaf culture), which has
been used as a convenient method for rearing plant mites, involves a fatal
problem for life history study. In this method, a detached leaf is placed on
water-soaked cotton or polyurethane sponge in a Petri dish filled with water.
Water surrounding the leaf effectively prevents mites on the area from
escape, and it maintains the leaf alive for a long period. Therefore, it has
been used as a simple applicable method to maintain mite colonies under
laboratory conditions.
However, the temperature decrease by water evaporation from the cotton
and Petri dish is sometimes not negligible. This effect is obviously serious
under low-humidity conditions, such as in air-conditioned insectariums. Saito
and Suzuki (1987) showed that the surface temperature (mites must be
affected by this) of a detached leaf under 47% r.h. is about 21.5°C when air
temperature is 24.8°C. Because the developmental time of Tetranychus (Tt.)
urticae under 21.5°C is roughly corresponds to 1.5 times of that under
24.8°C, we cannot know the accurate developmental time, if there is a
description of air temperature only without that of air humidity. Therefore,
we should pay great attention to the temperature (and/or air humidity) of the
microhabitat (i.e., leaf surface) when we conduct mite life history study by
this method.
2.3 Life Type Diversity in Spider Mites 11
Thus, Saito and Ueno (1979) concluded that such a difference in life history param-
eters reflects host plant stability, as expected by Mitchell (1973). It was still an
application of the well-known r/K selection or habitat template theories (McArthur
and Wilson 1967; Southwood 1977) prevailing at that time in ecology, but they
might have contributed to showing the importance of diversity in spider mite life his-
tory (Box 2.3). At present, we know that life histories vary depending upon phylog-
eny, antipredator adaptations (Sect. 3.6), and so on, but there can be no doubt that
host plant stability has also fundamentally affected spider mite life histories
(Sabelis 1985).
There is great variation in life type (or life style) between spider mite species
(Fig. 2.2; Saito 1979b, 1985, 1995b). Life types are defined as “strategies by which
mite species utilize silken threads on the leaves of their host plant.” Even though
the species studied are mostly restricted to those being distributed in Japan, I was
able to show that spider mites have diversified over various plants and exhibit vari-
ous kinds of life types on them. I have studied not only pest species but also various
non-pest spider mite species to enable comparative ecological studies (Box 2.4).
LW
Egg
Feces
Leaf under surface
CW
Web
WN
Fig. 2.2 Spider mite diversity in life type, little web type (LW), complicated web type (CW), and
woven nest type (WN). Typical subtypes are shown later
12 2 Plant Mites
Box 2.4
Spider mites had been thought to be somewhat homogeneous animals and
were simply referred to as “red mites” in agricultural fields, because of their
very tiny bodies with red color (not always so). When I started my biological
study on spider mites, most researchers were thinking in such erroneous terms.
One day Mr. Moriyama, who was my senior by 4 years, showed me several
species of spider mites that occur on Sasa, which are the main cast of this
book. This was when I first recognized the diversity of life style (life type) in
spider mites. Therefore, observing them changed my research life and pro-
vided the opportunity that these spider mites became my long-term partners.
Let us describe several of the typical life types observed so far. The simplest life
type, called LW (little web), was observed in Aponychus firmianae, which never
produces threads during its lifetime (Fig. 2.3). This species belongs to
Eurytetranychini, which is thought to be more ancient than Tetranychini. In the
same tribe, we know another species, Aponychus corpuzae, which produces web
covers on eggs (see Sect. 3.2); this is the simplest usage of silken threads in the
Tetranychinae. Panonychus citri (and Panonychus ulmi) always pull out threads
through their pedipalpi when walking (Sect. 2.3) and make “guy ropes” extending
from their eggs to the leaf surface. These species are also considered to have the
LW life type (Fig. 2.3). On the other hand, most Tetranychus species and some of
the species of Oligonychus, Schizotetranychus, and Eotetranychus have another life
type, the CW (nonsystematic and complicated web; Fig. 2.4) type. Saito (1977a)
showed that Tt. urticae always pulls out threads when walking (Sect. 2.3). Because
females of this species aggregate in leaf depressions (e.g., along leaf midribs), such
threads accumulate in particular areas and form web structures. This process is
considered to be a base of the CW life type. Furthermore, the mites tend to walk,
to deposit eggs and feces (Hazan et al. 1974; Gerson 1985), and to be quiescent
within the accumulated web. Thus, I hypothesized that these complicated webs
must serve to avoid predation and hazards from adverse climatic conditions (Saito
1985, 1995b). Relative to the foregoing, it is known that leaf depressions on which
CW webs are made are sometimes created by the spider mites themselves (Oku
et al. 2006). These suggest that CW type webs are not only a result of mite move-
ment with threads but also are produced for such functional reasons from aggre-
gated life, as mentioned later (Sect. 2.3).
On the other hand, there are many species that build systematic structures by
threads on the host leaf surface. I called this life type the woven nest (WN) type
(Figs. 2.5, 2.6). Eotetranychus suginamensis, E. shii, and Schizotetranychus brevisetosus
make web nests over depressions on leaves (on the underside), as shown in Fig. 2.6
2.3 Life Type Diversity in Spider Mites 13
Defense by parent(s)
Parental cohabi-
Adult cooperation
Web functions Interior of web tation with its
as a shelter for Web serves as nest as a offspring stages
Cooperation in
sanitizing nest
as a place for place for
next gener.1
next gener.1
quiescence
quiescence
egg laying
egg laying
defecating
defecating
egg cover
feeding 2
walking
walking
feeding
nymph
nymph
Species
adult
adult
larva
larva
egg
egg
Aponychus
firmianae
corpuzae
Stylophronychusvannus
Eutetranychus
orientalis LW
Eurytetranychoides
japonychus
Panonychus
citri
ulmi
mori
(Sasanychus)
akitanus
CW WN
Fig. 2.3 Various life patterns observed in the genera of Tetranychidae. Rectangles indicate what
stages of offspring usually share in the benefit of the trait or what traits were observed in the
species. g, guy ropes; 1, web serves for the next generation (gener.) (grandchildren); 2, feeding
within and under web; LW, little web type; CW, complicated web type; WN, woven nest type (after
Saito 1995b)
quiescence
quiescence
egg laying
egg laying
defecating
defecating
egg cover
feeding 2
Species
walking
walking
feeding
nymph
nymph
adult
adult
larva
larva
egg
egg
Oligonychus
castaneae g
shinkajii
uruma
Tetranychus
bambusicola
piercei
kanzawai
desertorum
fijiensis
urticae
Amphitetranychus
viennensis
quercivorus
CW WN
Fig. 2.4 Various life patterns observed in the genera of Tetranychidae. See Fig. 2.3 for
explanations (after Saito 1995b)
14 2 Plant Mites
Defense by parent(s)
Adult cooperation
Web functions tation with its
Web serves as nest as a offspring stages
Cooperation in
as a shelter for
sanitizing nest
as a place for place for
next gener.1
next gener.1
quiescence
quiescence
egg laying
egg laying
defecating
defecating
egg cover
feeding 2
walking
walking
feeding
nymph
nymph
Species
adult
adult
larva
larva
egg
egg
Eotetranychus
celtis
smithi
tiliarium3
tiliarium4
spectabilis
uncatus
querci
suginamensis
shii ?
CW WN Cooperative sociality
Fig. 2.5 Various social life patterns observed in Eotetranychus. 3, On Alnus hirsuta; 4, on Alnus
japonica. See Fig. 2.3 for other explanations (after Saito 1995b)
(WN-s in Fig. 2.7). Individuals living under the web nests always deposit their feces
on the outside nest roof (web), and their nest webs become coated with feces.
Yokoyama (1932) observed this behavior and supposed that E. suginamensis forti-
fies its nest by using feces. However, I rather supposed that this behavior has the
primary function of nest sanitation, because many species having nest-weaving
habits show more or less different kinds of feces manipulation behaviors (Fig. 2.7;
Saito 1995b). Among mites having the WN life type, the most sophisticated nesting
behavior is seen in Stigmaeopsis species. This genus is novel to the Tetranychidae
(Saito et al. 2004; the details are described in Sect. 3.5). All species belonging to
this genus have a common life subtype, WN-c, which is characterized by their solid
woven nests and special feces-depositing behavior. For example, in the largest nest
builder, Stigmaeopsis (St.) longus, all nest members deposit their feces at a particu-
lar place in their nests (WN-c in Fig. 2.7), suggesting that there is a kind of coopera-
tion in this species (Sect. 4.4). A single united nest is used for a long period, such
that a big pile of feces within the nest is visible to the naked eye. Most species hav-
ing the WN life type have a monophagous food habit and live on the leaves of
evergreen trees and perennial grasses (e.g., Sasa and Miscanthus). Anyhow, the
nest-weaving and fecal manipulation habits are two important features for defining
the WN life type and also serve as criteria to define life subtypes (Saito 1995b).
We can thus recognize three basic life types, LW, CW, and WN, through web-
building and feces manipulation patterns in Tetranychini. However, we also know
that there are several life types that are now conveniently categorized into these three
types Box. 2.5. In the CW type, Panonychus (Sasanychus) akitanus makes two to
three layers of web over depressions on the host leaf undersurface (Sasa senanensis).
Individual mites move on these layers and deposit eggs and feces there. Furthermore,
2.3 Life Type Diversity in Spider Mites 15
Fig. 2.6 Various social life patterns observed in three genera in Tetranychinae. 5, A different
means of egg protection and quiescent stage is known (see Sect. 3.3); 6, feces are deposited
outside of nest; 7, outside of particular places near nest entrance; 8, inside of particular place(s)
near nest entrances; 9, web mat is made under eggs. See Fig. 2.3 for other explanations (after
Saito 1995b)
all quiescent stages rest on these layers and never on the leaf surface. Whether we
have to classify this life pattern into CW or WN is another question, but this in
itself may not serve any biological insight other than showing their diversity in
life types. Schizotetranychus recki makes individual nests with leaf hairs
(Sect. 3.4). Eotetranychus asiaticus shows another curious habit: many mite indi-
viduals live with CW webs in a gall of psyllid larvae that cover their gall openings
(Fig. 2.8). Such a closed leaf hole must provide a very safe area from predator
invasion, such that it is supposed that this habit has a kind of antipredatory func-
tion as well as functioning to protect mites from adverse climatic conditions
(Saito, unpublished data).
As described above, such variation in leaf physical features is undoubtedly impor-
tant for spider mites to establish their own life type by webs. There are many studies
on competition between herbivorous mites, on interactions between herbivorous
mites and their predators (Huffaker et al. 1963; Bernstein 1984; McMurtry 1985; Hoy
1985; Saito et al. 1996; Osakabe et al. 2006), and on indirect effects of plant structure
(Fournier et al. 2003), which are more or less related to the web structures of these
16 2 Plant Mites
Eggs
Feces
WN-u WN-t
Web
Oligonychus gotohi Eotetranychus tiliarium
Schizotetranychus shizoupus
WN- s WN - c
Box 2.5
Recently I observed a curious life type on a Chinese bamboo plant that has
quite dense hairs. These hairs force female mites to walk through a narrow
space between the parallel-growing leaf hairs. The females of Schizotetrany
chus sp. have very slender bodies, perhaps because of the physical structure
of this habitat, and they move on the leaf surface in a column. Flat web nests
are built on the tips of leaf hairs, as shown below. It is a little bit difficult to
imagine the function of this web, but if the mites use the dense hairs sur-
rounding their nest as a stockade (then there is a lower risk of being attacked
from the lateral sides of nest) and protect themselves by the flat web from
predators attacking the upper surface of the nest, the function of this kind of
life type is understandable. However, I have still no information about such
predator species.
2.3 Life Type Diversity in Spider Mites 17
Fig. 2.8 Extraordinary life type known in Eotetranychus asiaticus Ehara living with CW life type
in an open gall of Trioza cinnamomi (Boselli). The larval T. cinnamomi that had closed the gall was
replaced (arrow). Is this symbiosis between spider mite and a psyllid (in Tanegashima, Japan)?
herbivorous mite species. Thus, we learned that spider mites have surely evolved
under the selection pressures of host plant physical features, host plant phenology,
and natural enemies (Chant 1959; Dosse 1961; Van de Vrie et al. 1972; Helle and
Sabelis 1985; Sabelis and Bakker 1992; McMurtry and Croft 1997).
From the foregoing discussion, we know that spider mites have very diverse lives
that vary among species and among host plant species. If so, we have next to
address basic problems about how and why spider mites evolved to produce silken
threads and to have their own life types.
Before describing the function of each life type, it is useful to determine what the
basic function of silken threads is. I previously introduced that Panonychus (Pn.)
citri and Tt. urticae always produce silken threads when walking (Saito 1979a). It is
thought that the name “spider mites” was originally given to these animals because
several well-known pest species produce silken threads just as do spiders (Lindquist
1985). There are several taxa in the Acari other than Tetranychidae in which some
species produce silken threads: Cheyletidae, Anystidae, Bdellidae, and so on (Walter
and Proctor 1999). However, most of them only use the threads to make cocoon-like
capsules when they are quiescent (Anystidae) or to make traps to capture prey
(Bdellidae). Hence, the Tetranychini and Cheyletidae (see Sect. 4.6) are the only
groups in the Acari that use silken threads so diversely throughout their lives.
18 2 Plant Mites
Spider mites produce two thread lines from a pair of organs (called terminal
eupathids) on the pedipalpi, as shown by the scanning electron microscope (Jeppson
et al. 1975; Box 2.6). They never eject these silken threads but rather use ambulatory
movement to pull silk out from the organs; i.e., they stick the tips of the silk threads
from the silk production organs onto the substrate (leaf surface) and then silken
threads are pulled out as they walk (Fig. 2.9; Mori and Saito 2006) or sometimes by
their falling down.
Box 2.6
Before Jeppson et al. (1975), there were two opposing hypotheses about where
spider mites produce silken threads; i.e., one has been mentioned in the text
(Grandjean 1948), and the other is that a single thread is produced from the
rostrum (Mills 1973). Because of the extreme thinness of silk threads, such a
discrepancy in observation was inevitable. Saito (1977a) also observed that a
Tetranychus urticae female produces a pair of threads simultaneously. From very
closely located pairs of pedipalpi and silk-producing organs (see below), two silk
threads readily stick to each other. Gerson (1979) estimated the thickness of
Tt. urticae thread to be the diameter of a single thread is 0.03 µm and that of a
wound thread is 0.06 µm. These measurement values are reasonable if there are a
single thread and stuck (double) threads.
Terminal euphathids
(Silk producing organs)
of St. miscanthi
Silken threads
Fig. 2.9 How a spider mite draws out its silken threads
2.3 Life Type Diversity in Spider Mites 19
Box 2.7
Silk threads are almost invisible even under dissecting microscope. How did
I idealize this method for the quantitative evaluation of webs? In the 1950s,
in Japan, children had no toys such as they do today, and we often played in
fields using homemade apparatus to capture insects. One such tool was made
by circles of wire. Spider’s webs were swept into these circles for use as
adhesive sweeping nets. This memory, as well as the method used to evaluate
walking velocity by Fransz (1974), provided the hint to use lattices with thin
nylon threads. Of course, the minimum thickness of the available nylon
thread was 50 µm, such that it was very laborious for spider mites, which are
only 400 µm in body length, to move over such relatively big obstacles.
(x100) (x100)
5 5
Panonychus citri on polyethylene film Panonychus citri on citrus leaf
Y=0.80X + 22.49 Y=0.92X - 1.22
4 4 r=0.98**
r=0.96**
3 3
No. of threads crossing quadrat (1x1 mm)
2 2
1 1
0 0
1 2 3 4 5 1 2 3 4 5
(x100) (x100)
(x100) (x100)
5 Tetranychus urticae on polyethylene film 5
Tetranychus urticae on red clover leaf
3 3
2 2
1 1
0 0
1 2 3 4 5 1 2 3 4 5
(x100) (x100)
No. of walking tracks crossing quadrat
Fig. 2.10 There is almost a 1:1 correspondence between threads and walking tracks in these two
spider mite species (after Saito 1977a)
Fig. 2.11 Tetranychus (Tt.) urticae female suspended from leaf surface by threads. Because the
threads are invisible, a white line was added on the original photograph to show the thread (from
VTR made by Nihon Nouyaku Co. Ltd)
2.3 Life Type Diversity in Spider Mites 21
Therefore, Saito (1977a) concluded that one of the key – and perhaps primary –
functions of silken threads is as “lifelines” that enable mites residing on the host
leaf undersurface to avoid the disadvantage of losing contact with the food plant.
Such a function of threads must be regarded as a convergence of both spider and
some larval insect threads. In relationship to the foregoing, Yano (2008) investi-
gated the role of trails in collective and solitary behaviors of Tt. urticae females. He
found that dispersing females simply follow the trails left by preceding females,
then they disperse collectively. Although he did not ascertain what constitutes a
trail, he discussed the possibility of a pheromone and/or threads. As Penman and
Cone (1972) and Oku et al. (2005) reported the importance of web plus female sex
pheromone, which simulated male searching behavior for females (quiescent
deutonymphs), I believe that Yano’s (2008) discovery is closely related with the
threads (and pheromone attached to them) produced by female walking.
On the other hand, spider mite threads are not always lifelines to fix them on to
the host plant leaves but can be apparatuses for dispersal from their residence (Helle
and Sabelis 1985). Fleschner et al. (1956) reported such a paradox in their paper
titled “Air drift of spider mites.” They reported that among spider mites, several
species disperse with silken threads (ballooning) from leaves when their host plants
are severely damaged. This behavior is thought to be resemble the popular dispersal
behavior of spiders, i.e., gossamer. Furuhashi and Nishino (1978) observed that the
females of Pn. citri disperse from tree to tree by drift (extending threads) in calm
winds. We also observed that many females of Oligonychus castaneae (ununguis)
suspend by threads from heavily infested chestnut leaves and gradually extend
threads. By swinging from leaf to leaf by the wind, they disperse and colonize new
leaves (Wanibuchi and Saito 1983). Thus, spider mite threads have at least two
opposite functions: “lifelines” when residing on leaves and tools for moving when
dispersing from plant to plant. Occasionally, we were fortunate enough to see that
both sides of the same coin apply under natural conditions.
In the manner mentioned above, silken threads are indispensable for many spider
mite species. Such threads sometimes form web structures on the leaf surface, i.e.,
the webs of CW and WN life types. The phenomenon that attracted our attention
was the huge webs (CW) of Tetranychus spp. (e.g., T. lintearius; plate 1B in Helle
and Sabelis 1985) and Tt. urticae (Fig. 2.12), which almost entirely cover the host
plant (Box 2.8). Why they should build such a web wrapping over almost withered
(and heavily damaged) plants is a difficult question, because if webs are primarily
produced to protect mites from predators and/or adverse climate conditions, they
should first build such webs just after host plant colonization (as observed in the
web nests of the WN life type; see later). Thus, I concluded that such huge webs
are primarily a result of mite infestation, in other words, “by-products.” After
revealing that the threads function as lifelines, it is understandable to consider such
web accumulation as the result of walking by a tremendous number of mites to
search for food within deteriorated host plants.
On the other hand, the CW type web might have subsidiary functions in several
spider mite species. It is known that there are two types of predator species, those
that can intrude into CW webs and those that cannot (Sabelis and Bakker 1992;
22 2 Plant Mites
Fig. 2.12 Tent-like webbing produced after a heavy infestation of Tt. urticae on red clover (from
VTR made by Nihon Nouyaku Co. Ltd)
Box 2.8
My first interest as a spider mite researcher was in this huge web. I was a
graduate student of Prof. H. Mori, who had studied the biological control of
the two-spotted spider mite by phytoseiid predators. The first theme that he provided
me was a search for alternative food for phytoseiids. To keep phytoseiids in
the laboratory, the rearing of spider mites as prey in a greenhouse was important
work, but I often failed at it when the prey population overshot. As a result of
this overshoot, all host plant leaves withered and tremendous mite webs covered
entire plants as well as pots. The reason why spider mites produce such huge
webs after they had exploited their host plants thus became a stepping stone
to my web research.
McMurtry and Croft 1997), such that the webs sometimes function to protect spider
mites from predation (Sect. 2.5). In several Tetranychus and Oligonychus species
having a CW-u life type, larvae and nymphs enter quiescence on the web (see
Fig. 2.4; note that the quiescent stage is as long as each active immature stage). Oku
et al. (2003) reported that Tetranychus kanzawai deutonymphs more frequently
enter quiescent stages (teleiochrysalis) on the web than on the leaf surface, when
they (of course before becoming quiescent) are exposed to a phytoseiid predator,
Neoseiulus womersleyi. Such a behavior suggests that quiescence on the web is
safer from predator attack than quiescence on the leaf surface.
2.3 Life Type Diversity in Spider Mites 23
expected to be functional for grasping the leaf substrate but must get hooked on
threads when walking on a Tt. urticae web. Actually, Pn. citri frequently becomes
entangled when it walks on Tt. urticae webs. Therefore, Pn. ulmi may have diffi-
culty living on leaves where Tt. urticae CW webs have accumulated, thus opening
the way for its displacement by the latter, as shown by Osakabe et al. (2006).
However, although there are several reports about the functions of CW and WN
life types, they are restricted to the species that are serious pests in agriculture. If we
turn our eyes to natural vegetation, there are many spider mite species that build
woven nests on leaf surfaces, i.e., species having WN life types. Shortly I address in
detail the functions of this life type for the species inhabiting Sasa senanensis (Chap. 3).
But before that, it is necessary to discuss spider mite diversity a little more.
2.4 Diversity of Physiological Adaptation in Spider Mites 25
There is another important trait in the spider mite life cycle, i.e., the diapause that
enables them to overwinter during the severe winters of cold and temperate zones.
Because spider mites include so many agricultural pest species in temperate zones,
many researchers have studied their diapause attributes.
the host phenology hypothesis (Gutierrez and Helle 1985). Among Oligonychus spe-
cies, most of which inhabit trees, the grass-inhabiting species Oligonychus pratensis
(Banks) and O. rubicandus undergo diapause at the adult (female) stage (see
Table 2.3; Jeppson et al. 1975). Furthermore, Tetranychus species, most of which are
grass inhabiting and very polyphagous, also overwinter as adult females (Table 2.3).
Therefore, Gutierrez and Helle (1985) suggested that an advantage of the female
diapause type mainly observed in grass inhabitants may be that after reactivation the
adult females have better chances of survival and a broader dispersal area, compared
with the larvae hatching from diapausing eggs. In contrast, in species living on trees,
overwintering eggs deposited on the bases of their own host branches do not need
any such dispersal ability because the larvae hatching from the eggs can easily reach
new spring leaves. Because most overwintering eggs (diapaused) have thick shells
that guarantee higher tolerance to cold weather and desiccation than summer eggs
(Lees 1953), it is advantageous for species living on trees to overwinter at the egg
stage (if larval hatching and leaf foliation is sufficiently synchronized). This host
phenology hypothesis in the diapause stage may also partly apply to most spider
mites species distributed in Japan (Table 2.3). However, Stigmaeopsis species over-
winter at the adult female stage on Sasa plants, most of which never defoliate, such
that a single leaf can live at least for 2 years (Saito, personal observation).
Furthermore, Eotetranychus species that hibernate at the adult female stage (Table 2.3)
cannot be explained, because most of these are tree inhabitants. Therefore, the host plant
phenology hypothesis is applicable to several genera, but not to these two genera.
Diapause
So far I have avoided using the term diapause because, with the exception of several
species (Lees 1953; Veerman 1985; Gotoh 1989; Ito and Saito 2006; Saito et al.
2005; Kawakami et al. 2008), we have little evidence of whether the overwintering
stages of spider mites are truly induced diapause. Diapause-induced females stop
oviposition, and some of them change their body color into carmine red or yellow.
However, diapause induction in females is not always accompanied with a body
color change (Saito et al. 2005; Box 2.9). Therefore, to determine diapause induc-
tion, we have to confirm the cessation of reproduction (reproductive diapause) for
females or nonhatching of eggs (egg diapause) without a cold period, such that the
data for species in Table 2.3 are not always sufficient. Precisely how short a day-
length is required to induce diapause (critical photoperiod) and how long a period
they have to be kept under cold conditions before diapause termination varies
between species and between populations distributed in different latitudes were
studied mainly for Tt. urticae (Veerman 1977; Koveos et al. 1999) and Pn. ulmi
(Lees 1953). Gotoh’s (1989) study on the diapause of several spider mite species
inhabiting deciduous trees revealed that the occurrence of diapause-induced
females that deposit diapause eggs is strongly influenced by the quality of food
(plant) on which the females developed. Entering diapause must be advantageous
28 2 Plant Mites
Box 2.9
It had been believed that diapausing female spider mites always changed their
body color thus, body color change is a useful criterion for diapause induc-
tion. In most studies examining female diapause in spider mites, changes in
female body coloration to bright orange (or sometimes yellow) are used as
the sole criterion of diapause induction (Veerman 1985; Mochizuki and
Takafuji 1996). In this study, we discovered an important fact: that Stigmaeopsis
females induced to diapause (as measured by the cessation of reproduction)
do not change their body color under experimental conditions (Saito et al.
2005). Therefore, these observations indicated that this criterion is not appli-
cable to the diapause of St. miscanthi and St. longus females (nor to that of
Yezonychus sapporensis; K. Ito, personal communication, 2002). On the other
hand, under field conditions, we observed bright orange females in midwin-
ter, suggesting that the body coloration of overwintering St. miscanthi females
does not always synchronize with diapause induction. Therefore, we must
consider that reproductive diapause is different from the acquisition of cold-
hardiness that is generally known to be related to body coloration (Leather
et al. 1993; Saito et al. 2005). Because the overwintering females of St. mis-
canthi and St. longus show a bright orange color in the field, identifying the
factors that change the body color of diapause females is an open problem for
future study.
for mites developing on low-quality leaves because no recovery of leaf quality can
be expected in deciduous trees within the year. The fact that host difference greatly
influenced female diapause incidence in Tetranychus kanzawai may also show that
the host phenology through food quality influences diapause performance (Ito and
Saito 2006). Furthermore, Goka and Takafuji (1991) reported that the diapause
induction rate in tree-inhabiting Tt. urticae populations shows clinal variation with
latitude in Japan, whereas herb-inhabiting Tt. urticae show a low diapause induc-
tion rate and no such trend. This finding means that differences in host quality and
in host species partly explain the variation in diapause traits of spider mites.
On the other hand, Goka and Takafuji (1990) studied the genetics of diapause
attributes of Tt. urticae. They suggested that the diapause of this species is deter-
mined by two factors: one is represented by multi-locus nuclear genes and the
other by cytoplasmic factors. Furthermore, Yano and Takafuji (2002) reported
that the diapause intensity of Tt. urticae is correlated with its dispersal ability:
they were able to artificially select mites that had either a high or a low dispersal
capacity. Then they observed the diapause incidence rate for these selected lin-
eages and showed that mites with a genetically high dispersal capacity have a
higher diapause incidence and a lower performance compared to mites with a low
dispersal capacity. Their results suggest that differential dispersal capacity is
2.5 Diversity of Plant Mites Other Than Spider Mites 29
In her famous book Silent Spring, Carson (1962) allegorically described a relationship
between pesticides, pest spider mites, and their predators: After spraying of chemical
pesticides, spider mites that have high tolerance to such chemicals can cease the
web production necessary to protect themselves against predators. Then, the pest
mites can increase their populations tremendously. We now know that Tetranychus
and Panonychus spider mites, to which she must have been referring in her book,
cannot cease silk production (Sect. 2.3), but her suggestion that the web has some
function to prevent predation is true. We know various species of predators occur
on plants. Predacious mites belonging to the families Phytoseiidae and Stigmaeidae
are common on plants and are considered to be the other main components of plant
mites (see Sect. 1.3). In Japan, 74 species of phytoseiid mites and six species of
stigmaeid mites are known (Ehara et al. 1994; Ehara and Shinkaji 1996), and
the number of species in the former is very similar to that of spider mite species,
suggesting their evolutionary relationship.
Phytoseiids
Chant (1959) first stressed the importance of phytoseiid mites as natural enemies
of spider mite species by his own experiments as well as field observations. His
report must be considered a classic among studies about predator–prey interactions.
After his work, several important studies on phytoseiids were made by Dosse
(1961), McMurtry and Johnson (1966), and Van de Vrie and Boersma (1970).
After their pioneer studies, many studies on the predatory efficiency of phytoseiids
have been conducted for various species (Helle and Sabelis 1985), although most
of them were concerned with agriculturally important species such as Phytoseiulus
(Pl.) persimilis (Fig. 2.14), Amblyseius orientalis, Neoseiulus (Amblyseius)
cucumeris, and Neoseiulus womersleyi (McMurtry and Croft 1997). One of these,
Pl. persimilis, is famous worldwide as a natural enemy that effectively suppresses
Tetranychus mites, which are serious pests on greenhouse vegetables, fruits, and
crops, and this species is now mass produced and supplied by several biological
control companies.
30 2 Plant Mites
500 µm
Fig. 2.14 Phytoseiulus persimilis Athias-Henriot, the most well known phytoseiid species in the
world (scanning electron microscopy, SEM)
Predator population
Prey population
2000 40
1000 20
0
40 80 120 160 200 240
Time in days
Fig. 2.15 Stable predator (Typhlodromus occidentalis, open circle)–prey (Eotetranychus sex-
maculatus, black circle) interaction under complex (heterogeneous) environment conditions
(redrawn from Huffaker 1958)
On the other hand, phytoseiid mites were also used as model animals to observe
predator–prey interactions from the aspect of animal ecology. Huffaker (1958)
indicated that there is a condition under which predator and prey can maintain
stable populations, as in a Lotka–Volterra model (Lotka 1925; Volterra 1926), with
some oscillation in a single system (Fig. 2.15). It is well known that most predator–
prey systems are unstable (Kuno 1987), because there was a serious limitation of
assumption (strict density dependence) in the Lotka–Volterra model. Therefore,
Huffaker’s experimental results showing that Eotetranychus sexmaculatus and
Typhlodromus occidentalis populations can persist for long periods within a single
system were extraordinary, and up to the present day these findings are frequently
2.5 Diversity of Plant Mites Other Than Spider Mites 31
referred to in many biology texts (e.g., Begon et al. 1999). The reasons why his
system could persist for so long are possibly the heterogeneity of habitat conditions
(in the experimental system) under which prey (but not predator) can easily move
between structured sub-patches within habitats. Anyway, this study on the interac-
tions between spider mites and phytoseiid mites has contributed so much to showing
“mites” as useful organisms for testing basic hypotheses in ecology.
Let me further discuss the diversity in predatory mites. McMurtry and Johnson
(1966) first reported that there is variation of response to web structures between
several phytoseiid mites: Amblyseius hibisci finds it difficult to penetrate Oligonychus
punicae webs, but Pl. persimilis and Typhlodromus (Tp.) occidentalis can easily do
so, such that the former species cannot be an effective predator of this prey species.
McMurtry and Croft (1997) reported a comprehensive review about phytoseiid life
styles in relationship to their roles in biological control. They categorized phytoseiids
into four types based on food habits and related biological and morphological traits:
type I, specialized predators of Tetranychus spp.; type II, selective predators of
tetranychid mites; type III, generalist predators; and type IV, specialized pollen
feeders/generalist predators (Box 2.10). Why such types are recognized in phytoseiids
may be closely related to the ability to handle spider mite threads (web and nest).
Type I phytoseiids must have a strong ability to penetrate CW type webs and
effectively feed on spider mites within webs. Type II phytoseiids may include two
subgroups: one feeds on LW type prey and the other feeds on WN type prey. Aponte
and McMurtry (1997) reported how Oligonychus perseae, a serious pest of avocado,
builds web nests on avocado leaves, thus showing a WN life type. Montserrat et al.
(2008) recently showed how the predatory mite Neoseiulus californicus rips and
Box 2.10
then penetrates the densely woven nests (WN type) of O. perseae. They observed
triangular openings after this predator penetrated the nests of O. perseae, such that
they supposed the phytoseiid can somehow rip nest threads. Very recently Shimoda
et al. (2008) observed how Pl. persimilis, Neoseiulus womersleyi, and N. californicus
can all cut the threads of a CW web using their chelicerae and palpi. On the other
hand, Typhlodromus (Tp.) bambusae, a specific predator of Stigmaeopsis species,
can easily intrude into woven prey nests using the nest entrances without ripping
threads. Furthermore, the result of Mori and Saito (2004) suggest no mite predators
are able to destroy the dense nest web of Stigmaeopsis species (Sect. 3.5). These
studies suggest that there is much variation both in the strength of spider mite
threads and in the ways phytoseiids manipulate them.
In relation to the foregoing, Sabelis and Bakker (1992) supposed that the long
dorsal setae of some phytoseiids must be an adaptation to manipulate prey webs,
because species that adapt well to webs commonly have long dorsal setae, i.e.,
the so-called “chaetotaxy hypothesis.” Pl. persimilis has long dorsal setae (see
Fig. 2.14). In Japan, N. womersleyi, an effective predator of Tetranychus species
(CW type), also has long dorsal setae, as does Tp. bambusae (Fig. 2.16), a spe-
cific predator of Stigmaeopsis species that build dense-woven nests (WN type;
Fig. 2.7). Therefore, their hypothesis may have a point. Sabelis and Bakker
(1992) suggested that such setae are used to extend moving space and to prevent
the adhesive threads from sticking to the dorsum. I think that the long dorsal
setae of Tp. bambusae have an additional function to recognize the prey nest
web, as proven in Stigmaeopsis species (Sect. 3.5), because this predator has a
strong tendency to live inside the prey’s nests. Therefore it must have a sensor
to recognize “nests” to realize such a habit. Although this is a little bit different
Fig. 2.16 Typhrodromus (Tp.) bambusae attacking Stigmaeopsis (St.) longus larva. Note that this
photograph is a simulated scene (by SEM)
2.5 Diversity of Plant Mites Other Than Spider Mites 33
Table 2.4 Larval feeding type and oviposition pattern in Phytoseiidae (from Chittenden and
Saito 2001)
Average Preferred prey life type (%)
no. of eggs/ Larval
Phytoseiid species day (±SD) LW-j LW-s CW-u WN-c feeding type
Typhlodromus bambusae Ehara 2.00 ± 0.22 0.00 4.17 0.00 95.83 NFL
Neoseiulus californicus (McGregor) 3.38 ± 0.25 0.00 8.64 90.13 1.23 FFL
Phytoseiulus persimilis Athias- 3.08 ± 0.94 1.33 12.00 78.67 8.00 NFL
Henriot
Amblyseius sp 1. 3.04 ± 0.50 4.11 0.00 38.36 57.53 NFL
Phytoseius tenuiformis Ehara 1.38 ± 0.22 0.00 0.00 93.94 6.06 NFL
Neoseiulus womersleyi (Schicha) 3.83 ± 0.59 0.00 0.00 71.43 28.57 NFL
Ambliseius orientalis Ehara 1.08 ± 0.19 11.54 73.08 0.00 15.39 OFL
Iphiseius degenerans (Berlese) 1.00 ± 0.25 25.0 29.17 20.83 25.00 FFL
Amblyseius sp. 2 1.25 ± 0.22 20.00 16.67 3.33 60.00 OFL
Amblyseius eharai Amitai & Swirski 1.03 ± 0.41 12.50 50.00 12.50 25.00 OFL
NFL nonfeeding larvae, FFL facultative feeding larvae, OFL obligatory feeding larvae, LW little
web type, CW complicated web, WN web nest. LW-j , subtype appears in Eurytetranychus
japonichus (Fig. 2.3); LW-s, that in Yezonychus sapporensis in Fig. 2.6; CW-u, that in Tetranychus
urticae in Fig. 2.4; WN-c, that in Stigmaeopsis celarius in Fig. 2.6.
various spider mite species having different life types (Sect. 2.3), such that all phy-
toseiid species targeted could feed, develop, and oviposit by feeding on some of
these spider mites. Then we used Sasa leaves as experimental arenas and four spi-
der mite species inhabiting Sasa stands as prey (species having four life subtypes
appear in Figs. 2.3–2.6). We prepared four sections by using water-soaked cotton
strips on a leaf area (4.5 cm × 4.5 cm). Then we first released several females of
each prey species on one of the four areas. Several days after the prey females laid
sufficient eggs (and some species made web structures), a female of each predator
species was introduced onto the area after eliminating cotton barriers and permitted
to select any of the four subareas (thus four prey species).
There were considerable differences in preference between these predator species
(see Table 2.4). Tp. bambusae only located to the area inhabited by St. longus, having
a WN life type; Pl. persimilis, N. californicus, N. womersleyi, and Phytoseius tenui-
formis located to the subarea with Pn. akitanus, having a CW life type. These
results indicate that there are strong preferences in these phytoseiid mites to
particular prey life types. On the other hand, there were three phytoseiid species
that located over all subareas, suggesting that they have no preference to prey life
types. Simultaneously we checked the larval feeding types of these ten predator
species, finding that five species have nonfeeding larvae, three have obligatory
feeding larvae, and two have facultative feeding larvae (Table 2.4). To know the egg
deposition patterns of these species, the distances between all eggs laid on the
experimental arena were measured, such that we could learn there is much variation
in egg-laying patterns between species. The species having obligatory feeding lar-
vae tend to lay eggs more sparsely than those having nonfeeding larvae. Species
having facultative feeding larvae deposited eggs at an intermediate distance, and
those having nonfeeding larvae deposited eggs in masses (Fig. 2.17).
36 2 Plant Mites
40
30
20
10
0
Am. sp. 1
Am. sp. 2
Pl. persimilis
I. degenerans
N. californicus
Ps. tenuiformis
Tp. bambusae
Am. orientalis
Am. eharai
N. wormersleyi
Fig. 2.17 Average distance between eggs of ten phytoseiid species. White, obligatory feeding
larvae; gray, facultative feeding larvae; black, nonfeeding larvae. Vertical bars are SD. (after
Chittenden and Saito 2001)
There are certainly many species of predators other than phytoseiids on plants. In
comparison with the number of studies done on phytoseiids, relatively few studies
have been conducted on other predators. Santos and Laing (1985) reviewed studies
on stigmaeid predators (Fig. 2.18) and stated that most species of this group have a
38 2 Plant Mites
low ability to control phytophagous plant mites because of their K-selected life
histories (low oviposition rates, high longevity, and low dispersal ability). Even
though I am not so familiar with stigmaeids, we have discovered that some species
play important roles in nature as natural enemies of particular spider mite species
as well as other phytophagous mites (Saito et al. 2008a; see Sect. 3.5).
According to Lin and Zhang (2002), Tarsonemidae is a large family of mites
present in many different habitats worldwide. Among them, some species occurring
on plants are important pests of agricultural crops (Lin and Zhang 2002). We found
one predator species, Tarsonemus praedatorius, which is the only reported tarson-
emid predator of spider mites in Japan (Lin et al. 2002). Females of this species
mostly prey on the eggs of St. longus in early summer. St. longus females cannot
recognize such small predators (even though St. longus has a high counterattack
ability against phytoseiid predators; cf. Sect. 3.5), and they have to face a situation
in which no offspring will develop in their nests unless they lay many eggs. I do not
yet know what cues they use, but St. longus females disperse from such nests to
newly grown leaves after heavy infestations by T. praedatorius (Saito, unpublished
data). Thus, this predator should perhaps be regarded as a parasite of St. longus
nests, but its importance as a factor enhancing the dispersion of St. longus from old
to new leaves of Sasa should not be underestimated.
Other predators on plants belong to Cheyletidae, Cunaxidae, and Anystidae. I
address Cheyletidae in Sect. 4.6 in relationship to social organization. Species of
Cunaxidae and Anystidae sometimes occur on plant leaves, but little is known
about their ecology other than they are predators. Their nimbleness and ability to
prey on phytoseiids and stigmaeids (A.R. Chittenden, personal communication)
lead one to suspect that they may belong to a higher trophic level.
Chapter 3
Diverse Microcosmos on Sasa
Sasa spp. are typical vegetation components of the forest understory (Fig. 3.1) in
Hokkaido, Japan. There are at least three species of Sasa, namely, Sasa senanensis,
Sasa kurilensis, and Sasa nipponica, and they are distributed separately in response
to snow depth in Japan (Suzuki 1978). On leaves of S. senanensis (I refer to this
species simply as Sasa), which is distributed over southern and central parts of
Hokkaido where deep snow cover accumulates in winter, at least nine species of
spider mites are known to occur (Saito 1990e; Table 3.1).
When I began to study the comparative ecology of spider mites about 35 years ago,
my target species from Sasa was solely St. longus (called the long seta form of
Schizotetranychus celarius at that time). Therefore, I had no idea of the diverse mite
fauna occurring on Sasa, although I had noticed that there are several different
spider mite species there. Hereafter I will describe the mite diversity on Sasa, but
this does not reflect the order of my study sequence. Many years later, we revealed
that the nine species of spider mites are all host specific to Sasa and most are some-
times sympatrically found in a single Sasa stand. Furthermore, perhaps because of
such a diverse and numerous phytophagous mite fauna, there are also more than
seven species of predaceous mites living on Sasa (see Table 3.1), and five of these
are also considered “bamboo-specific” species (Saito 1990e; Chittenden 2002;
Gotoh and Shida 2007).
I previously stated that there is diversity in the life history and life type of spider
mites (Sects. 2.2 and 2.3; Saito 1983), and I initially thought that such diversity
must simply reflect different host plants and environments. However, we have
noticed that all the basic life types (see Fig. 2.2) that are currently known in the
subfamily Tetranychinae appear on Sasa. Furthermore, we have also observed
variation in life history among spider mite species on Sasa (Fig. 3.2). As mentioned
before, life history variation in spider mites has been basically explained by host
plant spatiotemporal stability (Saito 1979a; Sabelis 1985). However, it is now
apparent that such a hypothesis cannot explain the variation in life history in
Sasa-inhabiting spider mites because they all live on the same host plant. How then
can we interpret the diversity in species, life history, and life type (Fig. 3.3)
observed on Sasa? We started our study with two working hypotheses. (1) The
spider mite diversity on Sasa was caused by predation pressure. (2) If so, then the
predacious mite diversity is a result of various antipredator strategies of the prey in
relationship to its webbing behavior, namely, life type and the host plant physical
features (for example, leaf hair density and depth of leaf depression).
3.1 Diversity in Sasa-Inhabiting Mites 41
5
4.5
4 CW
3.5 Schizotetranychus bambusae*
3
2.5
2
1.5
Panonychus akitanus
1
0.5
0
lxmx
2.5
2 Yezonychus sapporensis Aponychus corpuzae
LW
1.5
1
0.5
0
2.5
2 Stigmaeopsis longus WN
1.5 Stigmaeopsis takahashii
1
0.5 Schizotetranychus recki
0
0 10 20 30 40 50 60 70
Days after eggs laid
Fig. 3.2 Life history variation in seven species of spider mites on Sasa under 25 ± 2°C (Saito and
Ueno 1979; Saito and Takahashi 1982; Saito et al., unpublished data). Asterisk: species inhabiting
moso bamboo (Phyllostachys pubescens)
Eggs
Feces
Aponychus corpuzae
Yezonychus sp. 1
Panonychus bambusicola
Stigmaeopsis longus Stigmaeopsis saharai
Stigmaeopsis takahashii
Yezonychus sapporensis
Schizotetranychus recki
Panonychus (Sasanychus) akitanus
Quiescent Web
When we observed spider mites inhabiting Sasa, Aponychus (Ap.) corpuzae immediately
attracted our attention because of its special appearance. Ap. corpuzae has an
extremely flat body (Fig. 3.4). Whether this flat body, and the quiescent and feeding
postures, really function as antipredator traits is our question. Furthermore, does the
web cover (Sect. 2.3) the female puts over each egg serve any function? Chittenden
and Saito (2006) used quiescent (inactive stage) deutonymphs, which normally
assume a flat posture on the leaf surface, for this experiment, because it is impos-
sible to keep active mite stages in the optional position. Deutonymphs are quiescent
for at least 1.5 days under 25°C conditions, such that we could use them for 24-h
predation experiments. Half of them were tipped on their sides using a very fine
needle and the other half were left untouched in their natural positions (Fig. 3.5).
We also tested the function of egg webs by removing them by a fine needle. Of
course, by using a control experiment we had already confirmed there were few
effects from such artificial treatments. Then we tested the ability of four predator
species that usually co-occur with Ap. corpuzae to prey on Ap. corpuzae eggs and
deutonymphs.
It was apparent that Amblyseius (Am.) orientalis frequently ate Ap. corpuzae,
although it was still difficult for it to eat prey mites in the normal position (Fig. 3.5).
We also learned that the egg web cover has a strong function to prevent predation
(Fig. 3.6). Two species, Am. orientalis and Tp. bambusae, could increase their egg
consumption rates when the egg webs were removed. Furthermore, the deutonymphs
left in the normal position were eaten less frequently by all the predators (Fig. 3.5).
Therefore, we reached the following conclusion: The egg webs, the flat body, and
the quiescent posture of Ap. corpuzae help it avoid at least four predator species
(Chittenden and Saito 2006). Thus, we considered that such a body feature evolved
as an antipredator trait (Box 3.1).
The fact that most predator species occurring on Sasa have difficulty eating
Ap. corpuzae in its normal position (and with egg cover intact) surprised us. This
species has a very wide distribution on various species of bamboo (including Sasa);
i.e., its distribution is known over North and South-East Asia (it is known to occur
Fig. 3.4 Photographs of Aponychus corpuzae: adult female (left); egg (right)
3.2 Is the Flattened Body an Antipredator Strategy? 43
**
7
Nymph at normal position
(mean+SD)
6
Nymph attitudinally erected
deutonyphs eaten
No. of quiescent
(mean+SD)
5
** P<0.01
4
* P<0.05
3 ** - insignificant
-
2
1 -
0
Agistemus
Amblyseius
Typhlodromus
tenuiformis
Phytoseius
summersi
orientalis
bambusae
Predator species
Fig. 3.5 Function of flattened body of Aponychus corpuzae as an antipredatory strategy (after
Chittenden and Saito 2006)
**
16
** P < 0.01
12
No. of eggs eaten
** - insignificant
-
-
4
Agistemus
Typhlodromus
Amblyseius
tenuiformis
Phytoseius
summersi
orientalis
bambusae
Predator species
Fig. 3.6 Function of egg web as antipredatory strategy in Aponychus (Ap.) corpuzae (after
Chittenden and Saito 2006). White bars, with egg web cover; shaded bars, without web cover
44 3 Diverse Microcosmos on Sasa
Box 3.1
A flat body shape, like a flat fish, is generally thought to be such an adaptation,
and no one doubts this. However, to our surprise, there can be no evidence that
a flat body has a “true function” because it is difficult to make such an animal
body thicker or thinner. In this context, by using artificial manipulation of
Aponychus corpuzae quiescent stages, can we find it for the first time?
in Japan, China, Korea, the Philippines, Thailand, Malaysia, and Indonesia, all
regions with extensive native Sasa and bamboo vegetation). Such a widespread
distribution may be explained from its overall ability to escape predation.
Fig. 3.7 Photographs of Yezonychus sapporensis show egg and quiescent larva positioning on the
tips of leaf hairs
all positioned at the tips of leaf hairs, but never so in winter (Yanagida et al. 2001).
Thus, the individuals of this species alternate their behaviors between seasons,
although the ecological reason why they change the behaviors is still unclear.
The reason why Y. sapporensis females change their oviposition sites from the
tips of leaf hairs to the leaf surface in late autumn may be related to diapause
induction, because females are known to change oviposition sites from the leaf
surface to twigs after diapause inducement in several spider mite species that
overwinter in the egg stage (Lees 1953; Gotoh 1989). This change may be analo-
gous to the case of Y. sapporensis, although both adults and eggs overwinter in
this species (Sect. 2.4).
46 3 Diverse Microcosmos on Sasa
14
12 Manipulated
(20 eggs)
10
8 , Normal position
6
, Manipulated position.
4
2
Amblyseius
Amblyseius
Agistemus
Agistemus
Typhlodromus
womersleyi*
tenuiformis
iburiensis
summersi
Phytoseius
Neoseiulus
orientalis
bambusae
sp1.
Predator species
Fig. 3.8 Are oviposition and quiescence on tips of leaf hairs antipredatory behaviors? (Redrawing
from Yanagida et al. 2001)
Third, we address Schizotetranychus (Sz.) recki (Fig. 3.9), which makes compartment-
type nests by using two or three leaf hairs on Sasa leaves. All active individuals
make their own nests when they feed and enter quiescent stages. Furthermore,
females weave tent-like nests around each egg they produce (Fig. 3.9). Because of
such a life habit, it is difficult for females of this species to reproduce on hairless
Sasa leaves (Saito, unpublished data). Horita et al. (2004) tested whether the nests
really have a role in avoiding predation. They artificially removed the nests cover-
ing Sz. recki eggs and compared the probability of predation between eggs with and
without nest protection.
Figure 3.10 clearly shows that most predator species can eat eggs without nest
protection more easily than those within nests. The compartment-type nests effec-
tively protect eggs from five species of predators. The only exception observed was
Agistemus (Ag.) summersi, which tended to eat more eggs contained within nests
than eggs outside of nests. Therefore, the most important predator species interact-
ing with Sz. recki in nature is considered to be Ag. summersi (Horita et al. 2004).
Here nature once again shows us two sides of the same coin: Primarily, the anti-
predator apparatus that evolved in Sz. recki is now used by Ag. summersi in a con-
tradictory way, i.e., as a cue for prey existence.
So far, I have described several relationships between spider mites and their
natural enemies. These results have shown that spider mite diversity as well as
predacious mite diversity must be the result of complex evolutionary interactions
3.4 Do Compartment-Type Nests Have Antipredator Functions? 47
Female Egg
Woven nest.
Fig. 3.9 Schizotetranychus (Sz.) recki makes compartment-type nests. Female is shown walking
outside nest
prefers (mean+SD)
12
eggs in nests
10
8
6
4
2
Agistemus iburiensis
Amblyseius orientalis
Phytoseius temuiformis
Typhlodromus bambusae
Agistemus summersi
Neoseiulus womersleyi
Predator species
Fig. 3.10 Compartment-type nests of Sz. recki function to prevent predators. Choice experiment:
between these two treatments all are significant at P < 0.001 (after Horita et al. 2004)
between them in relationship to the physical features of Sasa leaves, i.e., flat sur-
faces for Ap. corpuzae and leaf hairs for Sz. recki and Y. sapporensis. Dense leaf
hairs are necessary for Sz. recki and Y. sapporensis, as mentioned before, because
they are essential materials for the life types of these species, i.e., antipredator
48 3 Diverse Microcosmos on Sasa
traits. In other words, these two species must be limited in their habitat choices by
the physical features of Sasa leaves (Yanagida et al. 2001; Horita et al. 2004).
Actually, there is much variation in leaf hair density among Sasa stands, and the
distributions of these two spider mites are strongly restricted to Sasa stands having
dense leaf hairs (Chittenden 2002; Gotoh and Shida 2007). If we focus only on
these findings, we cannot understand why some Sasa plants grow dense leaf hairs
that pest species utilize to protect themselves from predation. However, if we
change our focus to other pest species, the situation changes dramatically. We know
that the feeding behavior of an aphid species (Melanaphis bambusae), which can
severely damage Sasa, is greatly disturbed by dense Sasa leaf hairs (Saito, unpub-
lished data). As mentioned in the previous section, some trees produce leaf domatia
that natural enemies use as refuges (Walter 1996). Whether the variation in leaf hair
density between Sasa stands reflects an antiherbivore apparatus is not easy to deter-
mine. All I can say is that the leaf hairs provide important support for herbivorous
mites to construct antipredatory refuges.
What Is Stigmaeopsis?
Before describing the behavior and ecology of the genus Stigmaeopsis, I have to
provide a brief history of this genus, because it is still unfamiliar even to spider mite
researchers. Saito and Takahashi (1982) first noticed that there was variation in nest
size and morphology between populations of Schizotetranychus (Sz.) celarius.
Hereafter, I used, for the time being, “Sc” to express primarily a single species, Sz.
celarius, which was then separated into several species of the “celarius species
group,” until the redefinition of the genus “Stigmaeopsis” is completed (Box 3.2).
Saito and Takahashi first observed several dorsal setae for many Sc populations and
found a positive correlation between the lengths of two dorsal setae [P2 (sc1), D2
(d1); Fig. 3.11]. There were at least four distinct groups having different seta
lengths in this species distributing over Japan. These findings were the starting
point for long-term studies on the morphology, ecology, sociobiology, and specia-
tion of Stigmaeopsis (Schizotetranychus) species.
At first, we had thought that nest size variation was simply a reflection of complex
physical features on the Sasa leaf surface: Deep leaf depressions may enable the
mite to construct larger nests and shallow depressions may force the mite to build
smaller nests. To eliminate the influence of leaf physical features, we developed a
method to render leaf depressions constant by using a right-angled acrylic support
3.5 Woven Nests of Stigmaeopsis 49
Box 3.2
The genus Stigmaeopsis had once disappeared from mite taxonomy 54 years
ago (McGregor 1950) and was reinstated by Saito et al. (2004). Stigmaeopsis
celarius Banks was first described by Banks (1917) from an introduced bam-
boo plant (Bambusa metake) as a new genus in Tetranychidae. The Greek
name “stigma” refers to the small nest-like structure that is built by this spider
mite. Thirty-three years after the first description, this species was redescribed
by McGregor (1950) and moved to the genus Schizotetranychus. The genus
Stigmaeopsis, which had only ever contained one species, was then deleted.
After that, Ehara (1957) reported Sz. celarius occurring on Sasa in Japan.
µm
L & LS
LM
100
D2
M
50
60 80 100 120 µm
P2
Fig. 3.11 Correlation of two dorsal setae of Sc [Stigmaeopsis (St.) celarius] in Japan and Taiwan
(from Takahashi 1987). L long, LM long-medium, M medium, S short (seta forms)
and double-faced adhesive (Saito and Ueno 1979; Mori and Saito 2006; Fig. 3.12).
The (shiny) upper surface of a flat Sasa leaf was attached to the support with adhe-
sives, to create a right-angled leaf depression on the undersurface of the leaf. An Sc
female was released onto the right-angled depression of the leaf and permitted to
construct a nest (or nests).
Both the nest sizes and the length of the dorsal setae were measured after nest
construction, and the measurements were subjected to analysis. We confirmed that
the longer are the dorsal setae, the larger are the nests constructed in Sc (Fig. 3.13).
Hereafter the four groups discriminated by seta length are tentatively called L
(long), LM (long-medium), M (medium), and S (short) seta forms of Sc in descending
order of P2 seta length (Saito and Takahashi 1982) (see Fig. 3.11).
50 3 Diverse Microcosmos on Sasa
Adhesive barrier
Double-faced adhesive
Acrylic support
Fig. 3.13 Mean area (±SE) per nest of Stigmaeopsis. Results of the post hoc test are shown
(n = number of females analyzed) (from Mori and Saito 2004, with small changes) L, Stigmaeopsis
(St.) longus; LM, St. celarius; M, St. takahashii; S, St. saharai
Behavioral observations about mites within their nests showed that several
dorsal setae always touch the ceiling of the web nest, leading us to hypothesize
that some dorsal setae of Sc must function as sensors to recognize nest existence
(Box 3.3). Saito and Takahashi (1980) supposed that the dorsal setae extending
like the pantographs of electric trains must have such a function (Fig. 3.14).
3.5 Woven Nests of Stigmaeopsis 51
Box 3.3
As expected, animals living within nests or other closed spaces must recog-
nize whether they really live within such closed spaces. Most animals have
eyes, so that there is no need to develop special ways to recognize nests.
However, as mentioned in the text, most mites have no functional eyes, such
that they must develop other ways to recognize the existence of a nest if they
need to live within it.
Fig. 3.14 Schematic expression of the function of dorsal setae in recognizing the nest roof. This
figure is artificially manipulated with a scanning electron micrograph of St. longus and drawn lines
This hypothesis was finally proven by a genetic study between M and S forms.
Mori (2000) revealed that the length of dorsal seta, P2, is heritable in an incomplete
dominance manner. He discovered that some populations of M (St. takahashii)
and S (St. saharai) produced F1 hybrids when they were interbred (see later). By
use of the hybrid of these two species, he observed the relationship between
dorsal seta length and nest size. Then he showed that the seta length of hybrid
females became intermediate, and also the nest size built by the hybrid was inter-
mediate between that of the two parental species. This finding clearly showed that
the nest size is primarily governed by dorsal seta length, and that such a behav-
ioral trait is heritable.
Thus, we identified the mechanisms that covary with nest size and the length
dorsal setae in Sc. The next problem was whether the populations showing different
52 3 Diverse Microcosmos on Sasa
dorsal setae lengths (and different-sized nests) are considered variations within a
single species or different species. Because the differences in setal length are
discrete (Fig. 3.11), any reproductive isolation between these forms may determine
whether they are separate biological species in the sense of Mayr (1963). Thus, we
had to conduct crossing experiments between populations. Table 3.2 shows that
there is strong reproductive isolation between almost all combinations of populations
having different setal lengths (Saito and Takahashi 1982; Mori 2000). Furthermore,
a population collected from Miscanthus sinensis, newly designated as LS (dorsal
seta lengths are very similar to L) also showed strong reproductive isolation with L
(Sakagami et al. 2009). Although one-way reproductive convertibility does exist
between M and S, the reproductive isolation is still ecologically important (68–97.9%
deleterious; Mori 2000).
From the foregoing results, as well as from morphological and host plant differ-
ences, Saito (1990a) could describe two new species, Schizotetranychus longus and
Schizotetranychus miscanthi, from Sc (Sz. celarius). After that, we had an opportu-
nity to observe two Chinese species considered to be closely related to Sc. Based
on this, Saito et al. (2004) were able to describe two more new species and defined
the genus Stigmaeopsis (Fig. 3.15; Appendix 1). Hereafter, I describe these species
as those belonging to Stigmaeopsis.
Because all Stigmaeopsis species can be assumed to be very closely related sibling
species, as shown above and by Sakagami et al. (2009), we were able to restrict our
comparisons between sibling species to only those phylogenetic constraints that
could be minimized. I then tried to solve an evolutionary question as to why there
is such variation in nest size between these closely related species.
Nests made from solid and dense webbing undoubtedly function to protect the
individuals within them against something. One hypothesis about nest function is that
the nests function as protective refuges from predators, as in Sz. recki (Sect. 3.4).
3.5 Woven Nests of Stigmaeopsis 53
100 µm 100 µm
100 µm
100 µm
Fig. 3.15 Four species of Stigmaeopsis (St.) (females) on Sasa and bamboo (after Saito et al.
2004). Note the length of dorsal setae
Mori and Saito (2004) could prove this hypothesis by observing that Stigmaeopsis spp.
nests can more or less protect nest members from certain predator species. Seven to
ten females of each Stigmaeopsis species were introduced from stock cultures onto
detached leaf experimental arenas. After the females had constructed their nests and
deposited a sufficient number of eggs within them, the females were removed and two
treatments were prepared, i.e., the “web-removal” and “web-intact” treatments, thus
somewhat resembling the Ap. corpuzae and Sz. recki experiments (Sects. 3.2 and 3.4).
Adult-stage females of five predator species, which syntopically occur on Sasa (see
Table 3.1), were introduced individually onto each experimental arena. Two days
after predator introduction, we recorded the number of prey eggs eaten. The results
clearly showed that the nests have a strong protective effect on the eggs within them
(Fig. 3.16). Furthermore, nest size greatly influenced predatory intrusions. The largest
St. longus nests could only protect eggs from a single predator species, Amblyseius
sp., but the smaller nests of St. takahashii and St. saharai could protect against three
predator species, although all predator species could eat the eggs of all Stigmaeopsis
species if there were no web nests.
54 3 Diverse Microcosmos on Sasa
, Tp. bambusae;
, Ag. iburiensis;
, Ps. tenuiformis;
, Ag. summersi;
, Amblyseius sp.
No. of eggs eaten
St. longus St. celarius St. takahashii St. saharai St. longus St. celarius St. takahashii St. saharai
Stigmaeopsis species
Fig. 3.16 Prey consumption of five predator species over 48 h when there is a nest web (left) or
no nest (right) web (from Mori and Saito 2004). Numerals in right side of figure are numbers of
eggs eaten. The native host plant of St. celarius is considered to be some bamboo species (i.e.,
Phyllostachys spp. native to Japan), and it seldom occurs on Sasa. We used this species as an
example of a middle-sized nest builder, because it can normally reproduce on Sasa
The key point from these results is that the protective effect varies between different-
sized nests: Larger nests are vulnerable to the invasions of many predator species,
whereas smaller nests can prevent more predator species from invading. If so, from
the aspect of the efficiency of the nest as a refuge, the question of why there are
species making larger nests inevitably arises. What has happened during evolution-
ary history – namely, why did nest size diversify? – becomes the next question from
an evolutionary point of view.
One of the reasons why smaller nests can repel many species of predacious mites
is apparently because most predator species can eat prey eggs only when they can
intrude into the nest first. Amblyseius sp. is a large-bodied predator species that
cannot penetrate even the largest nests of St. longus. On the other hand, the middle-
sized Ag. summersi and Ps. tenuiformis can intrude into St. longus nests through the
nest entrances, but are unable to enter St. celarius, St. takahashii, and St. saharai
nests. This means that there are no predator species capable of ripping open
Stigmaeopsis nests on Sasa, as O. perseae does on avocado (Sect. 2.5; Montserrat
et al. 2008). Thus, we could determine that Agistemus iburiensis and Tp. bambusae
must be important common predator species to Stigmaeopsis spp.
3.5 Woven Nests of Stigmaeopsis 55
However, the data shown in Fig. 3.16 are egg death rates over 2 days, such that
they are not an accurate indication of the nest effect from the context of their entire
lives. If we intend to know the fitness difference between no-nest and nest-intact
conditions, it is necessary to convert the egg consumption data into egg survival
rates. Let us assume that the survival rate of eggs is zero when there is no nest.
When the number of eggs eaten is n (this means that all predators ate the maximum
number of eggs), and the number of eggs eaten by predators when the eggs were
covered by nests is n1, n − n1 can be considered the number of eggs surviving as a
result of the nest protection effects. Therefore, the survival rate of the eggs covered
by nests can be calculated as sv = (n − n1)/n. There is no information about the attack
rate of each predator in nature, so I assumed all predators are equally able to attack
each Stigmaeopsis spp. Under this assumption, sv could be calculated as shown in
Fig. 3.17. Although the difference in survival rate between different nest sizes is not
so large in this calculation, we may infer that fitness increases with a decrease of
nest size in a concave manner (Fig. 3.17). Furthermore, a small increase in survival
rate during a short time span (2 days) has a great effect on fitness over a long time
span (15 days are necessary for development), if prey mites are repeatedly exposed
to the same factors (Box 3.4; Mori 2000). Therefore, smaller nest builders have
much greater advantages than larger nest builders.
0.8
Fitness (survival rate)
0.7
Fig. 3.17 Fitness curve (survival rate) calculated from the prey consumption rates obtained in
Fig. 3.16. The procedure to convert the number of eggs eaten to the survival rate of eggs appears
in the text. Asterisk, St. celarius
56 3 Diverse Microcosmos on Sasa
Box 3.4 Difference in short-term survival rate greatly affects that of long-
term survival
Difference in survival rate measured for a short span causes a big difference
in fitness for a long period, because the rate is basically a power function of
the number of repeats of such spans: Let us denote s1 and s2 as survival rates
of two species measured during 1 day when they are exposed to the same
mortality factors, and t as days when these two species are exposed to the
same factors. The survival rate of each species is calculated as s1t or s2t. How
is such a difference in survival rate crucial? For example, if t = 5, s1 = 0.5 and
s2 = 0.4, and the total fecundity = 100; species 1 (s1) get 3.13 offspring
(0.55 × 100) but species 2 (s2) only get 1.02 offspring (0.45 × 100). In other
words, 0.1 difference in the 1-day survival rate results in a threefold difference
in the fitness (number of offspring) after 5 days.
Is protection against predator invasion the sole function of smaller nests? The
answer may be no. We noticed that there is another kind of variation in Stigmaeopsis
species nesting patterns. It is closely related to nest size variation, but not always
the same. The large nest builders, St. longus and St. miscanthi, make their nests in
a continuous manner, while the smaller nest builders, St. takahashii and St. saharai,
scatter their nests over the leaf surface. As the former two species can scatter their
nests and the latter two can build their nests in a concentrated fashon, these differ-
ences in nesting patterns are not caused by nest size differences alone. As such, we
realized that there are unknown factors forcing species to conduct different nesting
patterns. One hypothesis is as follows: for smaller nest builders, nests scattered
randomly over the leaf surface (as is actually the case; Saito et al. 2008a) have a
hindering effect on predators, even after the nests have suffered predation (note that
no predator species can destroy web nests), and this must be related to the searching
efficiency of the predator.
To understand the nest-scattering effects, Saito et al. (2008a) conducted several,
quite complex experiments to artificially manipulate nesting patterns. If we were to
use five Stigmaeopsis species, difficulties caused by host plant differences and nest
size differences would inevitably arise. Therefore, we decided to use a single species,
St. takahashii, which builds intermediate-sized nests. Many St. takahashii females
were introduced into leaf arenas and allowed to build nests and to oviposit.
Thereafter, the females were removed, and all the eggs in nests other than a single
nest containing many eggs (adjusted to 18 and hereafter referred to as “valid nest”)
were crushed to create void nests. Finally, several void nests were arbitrarily
removed with a fine brush and distilled water to establish the appropriate number
3.5 Woven Nests of Stigmaeopsis 57
of void nests in the experimental arena. The crushing of several surplus eggs was
unavoidable in this procedure. Thus we had to ascertain that the crushed eggs have
no effect on the following experiments. One day after the artificial manipulation of
nesting and egg-depositing patterns, we introduced Ag. iburiensis females (a specific
predator, as mentioned before) individually onto each arena and observed their
predation for 1 day. If this predator could discriminate nests without eggs (void
nests) from those with eggs (valid nests), then there must be no difference in the
number of eggs eaten, even if there are various numbers of empty (void) nests.
The results showed that the number of prey eggs eaten decreases significantly
with the increase of void nests. This finding showed that it becomes increasingly
difficult for this predator species to feed on prey eggs if there are many void nests
surrounding each valid nest. Furthermore, an increase of variance among replicates
with time was observed when there were many void nests (Fig. 3.18). If predatory
success depends more heavily upon chance than on systematic searching, then vari-
ance must increase with the number of void nests and with time (conversely, such
variance will decrease with time in systematic searching): If there are only a few
void nests, predators can easily find the valid nest(s), such that there is little vari-
ance among the replicates. On the other hand, if there are many void nests, several
fortunate predators may find valid nests quickly, but most will not find any for a
certain period. Therefore, in the latter case, the number of eggs consumed will vary
strongly between replicates. Moreover, if the predators search for valid nests in a
nonsystematic manner (namely, there is no improvement in searching efficiency
with time), then variance among the replicates must increase in the latter part of the
experiments with many void nests, because the fortunate predators that found a
Fig. 3.18 Effects of the number of void nests on prey consumption rate of Agistemus (A.) iburiensis
for starved (left) and fully fed (right) predator females. Because the predation processes at the
three categories of the number of void nests were unique, median (symbols) and range (bars)
values are shown instead of means and SD. There are significant differences in variance between
different characters by F test (after transforming the count data by the root) at each observation
time (P < 0.01 with the sequential Bonferroni correction for post hoc multiple comparisons) (from
Saito et al. 2008a)
58 3 Diverse Microcosmos on Sasa
valid nest earlier will have eaten many eggs by that time, whereas predators that had
still not found any valid nests will still not have eaten. Therefore, we learned from
Fig. 3.18 that nests that have suffered predation, i.e., have become void nests, exert
a decoy effect to trick predators.
However, the foregoing experiment could only show the effect of empty nests in
St. takahashii. Whether the nesting patterns observed in the five Stigmaeopsis
species really have the effect of decreasing predation pressure was still unknown.
To accurately address this problem, we conducted a more realistic experiment using
St. takahashii to ascertain whether the egg-depositing pattern per se affects preda-
tory efficiency. We arranged the number of nests and the number of eggs within
them as shown in Fig. 3.19: i.e., from the number of nests, x = 1 with 18 eggs to the
number of nests, x = 18, each with one egg. The former is analogous to St. longus
building only one nest with many eggs and the latter to St. saharai building many
nests each with small numbers of eggs. Then we introduced an Ag. iburiensis
female onto each arena. Two days after predator introduction, we observed how
many eggs had been eaten. As expected, there was a big difference in predator
success between nesting patterns. The predator could consume many eggs when
they were concentrated in a single nest, but it could only eat a small number of eggs
if they were subdivided into many nests (Fig. 3.20).
The reason why there is such a trend is understandable from the simple probability
exercise shown in Box 3.5. If eggs were subdivided into randomly scattered small
nests, the predator repeatedly gets trapped inside nests the contents of which have
already been consumed (i.e., void nests), and in doing so its chances of finding
further valid nests decreases. A simulation result using a practical model of repeated
Number of nests 18 6 2
9 3 1
Fig. 3.19 Experimental design to determine the effect of ovipositing and nesting patterns in
Stigmaeopsis.
3.5 Woven Nests of Stigmaeopsis 59
Fig. 3.20 Number of eggs consumed by the predator Agistemus (Ag.) iburiensis over 48 h as a
function of the number of nests (Saito et al. 2008a). A total of 18 eggs were artificially distributed
evenly over all nests (Fig.3.19) for each of the six trials. Numerals in the figure show the number
of replicates. Bars indicate standard deviations of the observed mean. There is a significant cor-
relation between the number of nests and number of prey eggs eaten (by Kendall’s rank correla-
tion: z = −2,24, t = −0.87, P < 0.015)
blind sampling with actual time budgets observed under VTR (Appendix 2) clearly
showed the time to eat 18 eggs increases with the number of nests (Fig. 3.21), indi-
cating the effect of nest-scattering behavior.
Therefore, we discovered an overlooked effect of the nesting pattern of small
nest builders (Saito et al. 2008a): scattered small nests containing small numbers of
eggs have the effect of extending predator searching behavior and thereby protect
the offspring contained inside. However, we met a difficulty as to whether we can
reconcile the nesting patterns of Stigmaeopsis species as a strategy (or function) in
the sense of Williams (1966) (Box 3.6). If a female makes many empty nests and
scatters them before commencing oviposition, this can be considered a defensive
trick against predators. However, small nest builders of Stigmaeopsis species do
not adopt such a “strategy.” Our discovery is that all females make only valid nests
(i.e., nests with eggs), and as a result of predation these change into void nests.
60 3 Diverse Microcosmos on Sasa
Box 3.5
Repeated build sampling trials: At the beginning there are only white balls
in the bag. At each sampling trial, when a player samples a white ball, he
exchanges the ball with a black one and returns the black ball into the bag.
If the player samples a black ball then he simply returns it into the bag.
Subsequently, the probability of drawing a white ball decreases rapidly with
the number of sampling trials. If we substitute the white balls with valid
nests and the black ones with nests void of eggs, this process is an exact
replica of the tricking effect modeled in Appendix 1 and shown in the text.
Therefore, someone may consider that it is not a strategy but only a by-product of
ordinary nesting behavior; this means that we should strictly consider this effect as
an indirect effect of nesting pattern. However, I now believe that such nest-scatter-
ing behavior per se could evolve as a tricking strategy by natural selection because
individuals scattering their nests in such a manner undoubtedly get higher offspring
fitness. Whether such an effect is authorized as one of a new suite of predator avoid-
ance adaptations such as mimicry, crypsis, schooling, effective escape (Taylor
1977; Schmidt 1990), and the laying of egg masses (selfish-herd in Hamilton 1971;
delusion effect in Calvert et al. 1979; and Foster and Treherne 1981; Vulinec 1990)
should be still tentative.
In any event, the nesting behavior of constructing small nests has another advan-
tage: they are more easily scattered than large nests. Therefore, we know that there
are double advantages for small nest builders. We again infer that fitness increases
with a decrease of nest size in a concave manner (by converting the small number
of nests to large nests, and vice versa in Fig. 3.20). If so, then the crucial question
now becomes why there are large nest builders at all in Stigmaeopsis.
3.5 Woven Nests of Stigmaeopsis 61
80
60
50
40
30
20
10
1 2 3 6 9 18
Number of nests (x)
Fig. 3.21 Simulation of hindering effect of nesting pattern on time required to consume 18 prey
eggs in Ag. iburiensis. x is number of nests per arena. Number of prey eggs in a leaf arena is 18
(constant), such that number of eggs per nest varies from 1 to 18 depending upon x. Searching and
eating time only were considered (Table A-1 in Appendix 2)
Box 3.6
The answer to this question may come from the discovery of sociality in the large
nest builders St. longus and St. miscanthi (Saito 1986a, 1990c). St. longus is known
to show strong counterattack behavior against predators, and attack efficiency
increases with the number of defenders (Fig. 3.22; Saito 1986b). Because a united
62 3 Diverse Microcosmos on Sasa
%
80
Defense by 2 males
20
No male
0
hr
0 8 16 24 32 40 48
Fig. 3.22 Male density and counterattack success (from Saito 1986b) shown by 2-day survival of
predator larvae. There can be no further increase in predator mortality because after 48 h all larvae
have developed to protonymphs, which show strong predatory activity
Fig. 3.23 Typical WN-c life type of Stigmaeopsis (St.) longus inhabiting Sasa bamboo leaf (SEM
micrograph). This photograph shows only approximately one-fourth of the united nest
St. longus nest can sometimes exceed 20 cm in length, large numbers of adults must
be required to defend their nests effectively against predator invasion (note that this
species has a 0.5-mm body length and no functional eyes). Therefore, it is easy to
understand that there is a great advantage for St. longus to live gregariously within
its large united nests (Fig. 3.23).
In relation to this, Mori and Saito (2004) compared the counterattack efficiency
of each Stigmaeopsis species against the larvae of a specific predator (Tp. bambusae).
Figure 3.24 indicates that St. longus is the strongest defender among the Japanese
3.5 Woven Nests of Stigmaeopsis 63
Stigmaeopsis spp. and that the defensive effect decreases with nest size. It was also
known that two large nest builders, St. miscanthi (HG form; see later) inhabiting
Chinese silvergrass (Miscanthus sinensis) and St. nanjingensis inhabiting moso
bamboo (Phyllostachys pubescens) in China, also show very strong counterattack
behavior against phytoseiid predators. In the former, a single male had a 0.67 prob-
ability of excluding a predator larva for 48 h (Box 3.7) and 0.75 probability for 24 h
in the latter (Zhang 2002). Therefore, there is no doubt that all the large nest build-
ers have strong counterattack abilities against invading predators. As described
later, this behavior is one of the important criteria for defining these species as
cooperatively social (Sect. 4.4). The middle-sized nest builder St. celarius (note that
this species is not native to Sasa) also shows counterattack behavior against predators,
although its success rate is lower than that of St. longus (Fig. 3.24; Mori and Saito
2004). On the other hand, the two builders of small nests, St. takahashii and St.
saharai, showed no such effective antipredator behavior. Both males and females
tended to move out of their nests when predator larvae had invaded. Thus we deter-
mined that there is apparent variation in social behavior among Stigmaeopsis spe-
cies (Mori and Saito 2004).
Box 3.7
0.8
Nest survival rate
0.6
0.4
0.2
0
2 females
+ 1 male
4 females
1 female
1 male
1 female
longus
1
.6
celarius*
predator exclusion)
.4
.2
saharai takahashii
0
.4 .5 .6 .7 .8 .9 1 1.1 1.2 1.3
Nest size (cm2 )
Fig. 3.24 Counterattack efficiency (= nest survival rate) varies among Stigmaeopsis species
having different nest sizes. Nest survival rate = killing rate + driving out rate of predator when two
males of each species defend a nest during 1 day (after Mori and Saito 2004). Remember that the
native host plant of St. celarius is considered to be some bamboo species and it seldom occurs on
Sasa. We used this species as an example of a middle-sized nest builder because it can normally
reproduce on Sasa
Why is there such variation in the counterattack behavior? One answer may be
that the large nest builders preserve their inclusive fitness through counterattack
(see Chap. 6) because such large nests can contain many offspring. Conversely,
small nest builders should escape from predators, because they keep only a few
offspring in their nests and will have ample opportunity to reproduce again somewhere
else. Another possible explanation is that the former must do so, but the latter has
no need to, because they make such highly protective nests. In other words, we
expect that there are two different strategies: One is to make small, scattered nests
that can effectively prevent predator invasion and hinder searching behavior. The
other is to make large, continuous nests where many individuals can cohabit and
perform cooperative defense of their offspring against intruding predators.
Hereafter I will attempt to explain the division of these two strategies in relation to
their speciation process. I have already addressed how nest size affects predatory
intrusion and showed that the prey that build their nests smaller and in a more scat-
tered manner obtain greater fitness. The increased trend of fitness (survival rate)
with the decrease of nest size and that with the level of nest scattering (linking with
the nest size decrease) both showed a concave manner (see Figs. 3.17, 3.20).
3.5 Woven Nests of Stigmaeopsis 65
small large
Nest size
Fig. 3.25 Expected effect of nest size on survival rate of Stigmaeopsis species by counterat-
tack (when the length of parents attendant and number of defenders in nests is considered)
(Mori 2000)
high
longus
saharai
Fitness
defense
protection
Fig. 3.26 Hypothetical relationship between offspring fitness and nest size variation (from
Mori 2000)
As such, extremely small and extremely large nests are thought to be the two
fittest strategies that inevitably evolved soon after Stigmaeopsis adopted nesting
behavior as an antipredator strategy. Such traits, namely large nests with coopera-
tive sociality and small, scattered nests functioning to prevent predatory intrusion,
might lead these mites in different directions. Therefore, Mori and Saito (2004,
2005) proposed the possibility of predator-mediated speciation. Figure 3.27 shows
the outline of this process. If there was an ancient species having middle-sized nests
and a mutation for making smaller nests occurred, the builder of smaller nests
would benefit from lowered predation pressure. On the other hand, if another
mutation to make larger nests occurred that incorporated defensive behavior, such
3.5 Woven Nests of Stigmaeopsis 67
a mutation also benefits from lowered predation pressure. Therefore, two extreme
strategies, namely, making both the largest and the smallest nests, might be
selected if all other things are equal. This scenario suggests that some Stigmaeopsis
species may have evolved through “predator-mediated speciation” (Mori 2000;
Mori and Saito 2005). The discovery that nest size is determined by incomplete
dominance gene(s) that govern the length of dorsal setae (as sensors), as men-
tioned previously, lends indirect support to this hypothesis because such traits
are heritable.
A Further Question
As such, we can explain why two species having very different-sized nests occur
sympatrically on Sasa leaves. Actually, we could regularly observe two Stigmaeopsis
species, either St. longus (large nest builder) and St. takahashii (small nest builder)
or St. longus and St. saharai (small nest builder), in Sasa stands sympatrically, but
St. takahashii was seldom found sympatrically with St. saharai (Chittenden 2002).
The largest extreme is observed in St. longus, the females of which each build large
and continuous (united) nests over their lifetimes. A further question then arises:
What is the minimum nest size in Stigmaeopsis?
As described before, there are two small nest builders, St. takahashii and St.
saharai, occurring on Sasa leaves, even though they are distributed allopatrically
(Chittenden 2002). There was a significant difference in nest size between these
two species, i.e., the nest of St. takahashii is larger than the nest of St. saharai
(see Fig. 3.13). The reasons for this difference must be related to minimum nest
size. First, we can easily conclude that the minimum nest size must be larger than
mite body size, because if the nest is the same size as an individual body, then
it is not a “nest” but more like a body wrapping (like a suit). As previously
described, small nests prevent predators from invading, but this is only effective
when the nest has a protective refuge function. A nest that is too small has no such
effect and no protective function against predators, because predators can feed on
mites directly through the webbing. Such a phenomenon has occasionally been
observed in interactions between St. saharai and Ag. summersi: the latter is a
relatively large predator, occurring on Sasa, that sometimes preys upon the
former from outside the nest (see Fig. 3.16), although it seldom preys upon St.
takahashii in the same manner (Yoshida et al., unpublished data). This observation
suggests that the minimum nest size might be determined by the risk of predation
from outside the nest. The fact that St. saharai is never distributed with Ag.
summersi on Sasa leaves in the field (Chittenden 2002) suggests that this
proposition may have reality. Furthermore, small nest builders must frequently go
outside their nests to make new nests, but walking outside is risky for females
because many predacious mites occur on Sasa.
68 3 Diverse Microcosmos on Sasa
In the previous sections, I have addressed several examples thought to have resulted
from ecological interactions between predator and prey species. If so, is it plausible
to consider the notion that such interactions may be the sole cause of mite diversity?
Are there any other explanations for this phenomenon? As a reply, let us again
address the diversities in life history and behavior, as well as some other special
features of these mite species.
We have observed variation in life history and life type among spider mite spe-
cies on Sasa (see Table 3.3; see Figs. 3.2, 3.3). Furthermore, I have shown that there
are differences in the predator–prey relationships among Sasa-inhabiting species.
The variation in life history between these spider mites cannot be explained from
the host plant stability hypothesis (Sect. 2.2; Saito and Ueno 1979), because all the
mites live on the same host plant. On the other hand, there does appear to be a trend
that species having high rm show CW life types and those having lower rm show WN
life types (see Fig. 3.2), whereas species having LW life types show intermediate
rm values. Therefore, these trends may be closely related to the differences in anti-
predatory traits: As stated, mites building CW type webs have no special behavior
against predators, and such webbing is merely the result of mites walking and/or
aggregating (Sect. 2.3). If the web has some additional roles in mite living, such as
the location of eggs and feces deposition or the exclusion of some predator species,
then the production of silken threads is itself not costly; i.e., the web effects are
basically by-products of lifeline threads, and as such may not be subject to the cost
of “time.” Therefore, I expect that these mites are colonizers of predator-free space
on Sasa. They increase their populations rapidly on newly occupied Sasa leaves
where predators are scarce and then disperse after predatory mites begin to accu-
mulate. This function is considered to be a kind of r-strategy, if we can replace
“high probability of predator accumulation” with “easily exploitable and ephemeral
habitat” in the original r/K theory.
On the other hand, if mites could develop some apparatus and/or behavior(s)
to decrease predation pressure, they can share the same habitats (Sasa leaves) for
Table 3.3 Life type and overwintering peculiarities of spider mites on Sasa
Species Leaf surface infested Life type Overwintering stage
Aponychus corpuzae Underside LW Female
Panonychus bambusicola Upperside LW Egg
Panonychus akitanus Underside CW Female and egg
Schizotetranychus recki Underside WN Egg
Stigmaeopsis longus Underside WN Female
Stigmaeopsis saharai Underside WN Female
Stigmaeopsis takahashii Underside WN Female
Yezonychus sapporensis Underside LW Female and egg
Yezonychus sp. 1 Underside CW Unknown
3.7 Other Variation in Spider Mites on Sasa 69
longer without dispersing away. Mites showing WN life types have evolved such a
trait. Because constructing solid nests requires both time and resources, there must
be a trade-off (in the evolutionary sense) between the ability to reproduce rapidly
and web construction. Furthermore, we know that most species having WN life
types are relatively small in size, possibly because they live in nests. Such morpho-
logical traits may also decrease their fertility. Therefore, they evolved as K-strategists
by acquiring a good ability to resist high predation pressure.
In relationship to the foregoing, I reported that there is an interesting correspon-
dence between spider mite life history and predator life history (Saito 1990b): The
life history of Tp. bambusae, which feeds upon K-selected St. longus, closely
resembles that of its prey, i.e., low rm, but N. womersleyi, an important predator of
Tt. urticae (considered an r-selected species), has a very high rm compared to that
of its prey. Of course, these are only two prey species and their specific predators,
but such a phenomenon can provide a valuable insight into understanding mite life
history evolution (also see Sabelis and Janssen 1993).
Generally speaking, when reasoning about r/K selection on Sasa, another kind
of evolutionary trade-off must be considered: rapid increase with high dispersal
costs to escape predators may favor CW types. On the other hand, parental defense
and cooperation without dispersal must incur considerable costs to realize the WN
type. LW type mites having high mobility and spacing abilities do not need any web
structure but should incur the costs of dispersal as well as the costs of locating
mates and food. Therefore, an important point to consider is that life history diver-
sity is created not only by the effects of host spatiotemporal stability but also by
predator–prey interactions in the same habitat.
We also noticed that all three overwintering types appear in Sasa-inhabiting spider
mites (see Table 3.3). As with life history variation, we must search for an alter-
nate explanation for such diversity in overwintering because there is no way to
adapt the host plant phenology hypothesis (Sect. 2.4) to species living on the same
host plant. Furthermore, Y. sapporensis and Pn. akitanus have two overwintering
stages – egg and adult female – on Sasa. (Gotoh 1986; Table 3.3).
How can we explain such overwintering diversity between spider mites living on
Sasa? Addressing the conclusion first, I have no clear explanation of the reason(s)
for this diversity. However, I can suggest several possibilities that may be related to
this phenomenon. As pointed out by Harvey and Pagel (1991), there are great dif-
ficulties associated with studying evolutionary processes solely by between-species
comparisons without accurate phylogenetic information. Because the mites inhabiting
Sasa are categorized into various genera of Tetranychinae, such diversity in over-
wintering types appears to be simply a reflection of phylogenetic constraints. If so,
namely that for some reason an ancient species of a certain genus had a particular
overwintering stage, then such a trait is not always an ecological trait on Sasa.
70 3 Diverse Microcosmos on Sasa
Gutierrez and Helle (1985) have proposed several hypotheses on spider mite phy-
logeny on the basis of morphology and karyotype, and both Navajas et al. (1996)
and Sakagami (2002) have considered the issue from a molecular perspective.
Although neither is sufficiently comprehensive and both contain several inconsis-
tencies, we may say that the differences in overwintering stages evident in Sasa-
inhabiting mites are consistent with the major trends in their respective genera, such
that it may be regarded mostly as a phylogenetic constraint: Although Stigmaeopsis
is still not positioned in the Tetranychinae phylogeny, in accordance with Sakagami
(2002), it may be placed near the genus Tetranychus in which all species overwinter
as adult females (see Table 3.3; Veerman 1985; Niino and Sakagami, unpublished
data), such that we can also consider overwintering at the adult female stage to
reflect phylogeny. However, the overwintering in two stages observed for both Y.
sapporensis and Pn. akitanus is an extraordinary and rare phenomenon in
Tetranychinae (although Sz. brevisetosus also has two overwintering stages; Saito,
unpublished data). Yezonychus is thought to be very close to Schizotetranychus in
which all species overwinter in the egg stage; thus, we can suppose that the ancestor
of Y. sapporensis primarily overwintered in egg stage. Therefore, I consider it plau-
sible that overwintering adults recently evolved through ecological factors, such as
predation pressure on Sasa. Anyhow, all we can say at the present moment is that
there is no single concise explanation for the overwintering stage diversity seen in
Sasa-inhabiting spider mites, other than an ecological null hypothesis, i.e., phylo-
genetic constraints. From this reasoning, the observed species can all live on Sasa
leaves regardless of differences in overwintering type, because there may be no
strong selection pressure on such a trait in this habitat.
Chapter 4
Mite Sociality
Within the world of mites, we can find almost any kind of genetic system
(genome-transporting systems; Table 4.1), i.e., diplo-diploidy, haplo-diploidy, and
thelytoky (Walter 1996). Furthermore, an extraordinary system, “apparent” hap-
loid-thelytoky, occurs in Tenuipalpidae (Table 4.2; Weeks et al. 2001).
Understanding the diversity in genetic systems in mites (in all organisms for that
matter) is an extremely important issue that is as yet unresolved (Bull 1983).
Furthermore, there is much variation in how such genome inheritance systems are
attained within each genetic system. For example, haplo-diploidy involves at least
two different mechanisms, i.e., arrhenotoky and paternal genome loss (PGL), and
thelytoky involves both automixy and apomixy. Furthermore, as stated by Wrensch
et al. (1994), inverted meiosis, which is hypothesized in mites, is not negligible.
Such differences in the mechanisms tend not to influence the number of genomes
per zygote but rather greatly influence genetic variability. From the perspective of
evolutionary ecology, the latter is rather important, because selection basically
operates on such genetic variability.
Meanwhile, it is erroneous to say that genetic systems are equal to sex determi-
nation mechanisms. As Helle and Overmeer (1973) properly stated for haplo-diploid
organisms, the statement that haploid individuals are males and females are diploid
is true (but see Weeks et al. 2001), but stating that males are always haploid and
females always diploid is not always true. This is a problem of causal sequence
(cause-and-effect problem). In social bees having haplo-diploidy, it is well known
that diploid males are produced from inbred females (Crozier 1979; Bull 1983),
Table 4.2 Genetic systems in Tetranychidae and its related family, Tenuipalpidae
No. of species
Family or subfamily Thelytoky Arrhenotokya Others checked
Teytranychidae
Bryobiinae 35% 65% – 17
Tetranychinae 2% 98% Tychothelytokyb 116
Tenuipalpidae + + Haploid-thelytokyc –
a
Including species in which unfertilized oviposition and male existence were only observed
b
In Oligonychus ilicis, unfertilized females reproduce males, but sometimes they also produce
females (Helle and Pijnacker 1985)
c
See Table 4.1
a phenomenon explained by the following two models. One is that two alleles at a
locus determine sex, i.e., hemizygotes (haploid) and homozygotes (diploid) become
males and heterozygotes becomes females (the CSD model, Bull 1983; Van
Wilgenburg et al. 2006). The other is the multi-locus homo-hetero relationship, i.e.,
hemizygotes and homozygotes at multiple loci become males and heterozygotes at
least at one locus become females (Bull 1983). Therefore, it is usually appropriate
to say that hemizygotes (gametes) develop into males, but not that zygotes (diploid)
always develop into females.
Another recently discovered example is Brevipalpus phoenicis (Tenuipalpidae,
Fig. 1.1), which reproduces parthenogenetically. Weeks et al. (2001) observed the
details of this species by molecular and bacteriological methods, and revealed
that the “female” is actually a “male!” This paradoxical result has surprised many
researchers, including myself. Weeks et al. treated B. phoenicis females with an
antibiotic and got adult male offspring, which had been hitherto unknown in this
4.2 Sexual Behavior 73
Males show at least three kinds of “mate guarding.” One is a monogamous system,
where males guard or defend partner females from pair formation to development
of offspring, as is known in many birds and mammals. The second is called
74 4 Mite Sociality
Aponychus corpuzae
strong tendency to guard older quiescent females (just before ecdysis). However, they
introduced active deutonymphs onto the experimental arena to establish different-
aged quiescent females. Thus, their findings do not always mean the existence of
male strategy to guard older quiescent females to save guarding time; this is because
the guarding frequency must be also a function of the length of time quiescent
females are exposed to males. If we wish to solve this question in a clearer manner,
different-aged third quiescent stage females must be simultaneously provided to a
single male (the reason for using a single male is to minimize the effect of the male-
to-male competition observed by Potter et al. 1976); then the age of any female
guarded by the single male could be observed. Therefore, we do not know whether
spider mite males actually possess any such strategy as theoretically predicted by
Grafen and Ridley (1983). I believe that such a study would be very relevant.
We do have a detailed understanding of the variation in pre-copula behavior and
male-to-male competition behavior for securing mates between spider mite species;
e.g., Tt. urticae males use stylets and saliva (Fig. 4.2), Pn. citri males fight each
other intensely by using their forelegs in a boxer-like manner (also sometimes using
stylets and saliva; Fig. 4.3), and St. miscanthi males use their stylets to cannibalize
opponents (Chap. 6; Saito 1990c). Because of the difficulty in evaluating the intensity
of male-to-male combat, such differences in male behavior have mostly been
ignored in comparative behavioral studies so far (but see Sect. 6.1).
On the other hand, Radwan (1993, 1995) reported a very interesting phenomenon
in acarid mites. Male dimorphism is known in many acarid species. The “fighter”
males of Sancassania (Caloglyphus) berlesei are typically characterized by a thick-
ened and sharply terminating third pair of legs, whereas “scrambler” males have no
such morphology and are benign. The “fighter” males frequently kill rivals using
the thickened pair of legs. Because males can be triggered to become fighters by
decreasing the density of these mites, the variation in fighter expression in this mite
76 4 Mite Sociality
Fig. 4.2 Male-to-male combat in Tt. urticae. On the left is a photograph from a 16 mm movie and
on the right is a drawing based on the movie. Arrows show saliva secreted on stylets
Male
Quiescent female
Fig. 4.3 Combat-like boxing in Panonychus (Pn.) citri males. Leg movement was so swift that
their first legs were filmed as a two-fold thickness (filmed in 16 mm format by Saito)
4.2 Sexual Behavior 77
is caused by the density conditions under which they developed. He supposed that
the fitness benefits of the alternative tactics were that at low colony densities fighter
males are likely to have an advantage over scramblers (benign), because they are
able to kill all their rival males and dominate a group of females (Radwan 1993).
At high colony densities, on the other hand, fighter males would derive no advan-
tages because of the high cost (frequency) of fighting (if all fighters fought together
then males would have little time to mate). Using this species, Tomkins et al. (2004)
tested the existence of status-dependent evolutionary stable strategy (ESS) and
showed that this mite species is quite appropriate for testing such a general theory
in evolutionary ecology. These are undoubtedly nice examples for demonstrating
the suitability of mites as materials for testing general theories in evolutionary
ecology and sociobiology. However, I have a further question about why such a
strategy occurs exclusively in acarid mite groups because I also observed a kind of
mixed strategy in male-to-male antagonism in a spider mite that does not show any
male dimorphisms (Sect. 6.5) within a population. I think that identifying the factors
which cause such high male-to-male antagonism in animals and why some animals
evolved dimorphic males and others did not are important problems that will soon
be challenged by utilizing these mite groups.
Sexual behavior between males and females has received much attention in
mites, because there is so much variation in the mode of sperm transfer. According
to Walter and Proctor (1999), almost all known sperm transfer modes are listed in
the Acari. Therefore, it is difficult to describe them one by one. Here I only want
to introduce several examples and recommend any reader wishing more detailed
information to check Walter and Proctor’s (1999) fascinating review. All spider
mite males have an aedeagus (penis) and use it to directly transfer sperm. If the
sperm’s path inside the female spider mite involves traveling through the hemo-
coele, then first male sperm precedence (P1) is thought to be caused by the traveling
mode (this traveling process was also hypothesized in M. muscaedomesticae)
(Walter and Proctor 1999). Other fascinating studies on intersexual behavior appear
in Tilszer et al. (2006) and Prokop et al. (2007). Tilszer et al. (2006) used an acarid
mite species, Rizhoglyphus robini, to show how sexual conflict decreases female
fitness. They revealed that the decrease in harmfulness (to partner females) of
monogamous males was accompanied by a decline in reproductive competitive-
ness by thirty seven generations of selection lines either enforced monogamy or
polygamy. Prokop et al. (2007) discovered that the female offspring obtained by
mating with older males show significantly decreased fitness, probably through the
accumulation of deleterious mutations in the male germline. These are also excel-
lent examples showing the appropriateness of mites as materials for testing general
ecological theories, as reviewed by Radwan (2007).
One of the interesting phenomena related to sexual behavior in spider mites
underlies the large variation in the size (and/or bending direction) of the aedeagus
between species. From the point of view of the taxonomist, the spider mite aedea-
gus is relatively simple in shape and not so distinctive between related species, such
that it is only used as a secondary character for species description. However, if we
switch our perspective to the aedeagus size differences between species, then its
length varies from about 13 µm in Tt. urticae to about 67 µm in Eotetranychus tiliarium
78 4 Mite Sociality
(Ehara and Shinkaji 1975). If we compare the length of the aedeagus to male body
size, in the former it is only about 2% but in the latter it reaches about 23% of body
length. There is no behavioral information about why such variation exists, but I
believe that it is some reflection of a mating strategy.
It is well known that the sex ratio of mites varies considerably. Hamilton (1967)
first addressed the evolutionary reason why biased sex ratios evolved in various
organisms, even though a 1:1 ratio is primarily an evolutionary stable strategy as
shown by Fisher (1930). Fisher (1930) solved the reason why sex ratios attain even-
ness in both sexes from the theory of frequency-dependent selection: Under random
mating conditions, as the frequency of one sex increases, mothers who produce
more offspring of the rarer sex will get more advantage (many grandchildren). Thus
the sex ratio = 1/2 attains an evolutionary stable strategy (ESS). Hamilton (1967)
considered this principle as “a starting point of the sex ratio problem.” Then
he considered “nonrandom mating” by introducing spatially structured mating
populations (Box 4.1). By the use of a game theoretical approach, he showed that
the sex ratio will become biased toward the female sex if the following conditions
Fertilized females
n++
n++ n++
Habitat
Offspring Offspring
Offspring
females mate females mate
females mate
with males
within habitat
with males
within habitat
with males
within habitat
.....
Fertilized females
4.3 Sex Ratio 79
Box 4.2
As pointed out by Hardy and Mayhew (1998), there is a well-known negative
correlation between brood size and (male) sex ratio in Bethylidae. If a patch
is occupied by a single fertilized female only, there will be an extreme number
of males; i.e., a single male must be expected by the LMC model (note that
the equation predicts no males, but this is impossible for bisexual organisms).
In such an extreme case, the female fecundity must greatly influence the sex
ratio observed, because a single male is always produced regardless of fecun-
dity. Fecundity increase must increase female offspring and then decrease
male ratio. Therefore, we cannot neglect the effect of fecundity when we
attempt to test the LMC model.
As shown in Table 4.3, there is much variation in the sex ratios of spider mites and
almost all are more or less female biased. It was very difficult to determine the “pri-
mary sex” ratio of spider mites, although now it is possible by the use of molecular-
level methods (but few attempts have been made). Therefore, the data shown in
Table 4.3 are all progeny sex ratio at maturity (note that the offspring survival rate is
sufficiently high to be considered them an approximation of primary sex ratio as
well). To prove LMC in spider mites, two studies have been conducted that revealed
there are phenomena which are consistent with LMC predictions. Roeder (1992)
showed that Tt. urticae females produced more males when they were reared under
high-density conditions, although the numbers of eggs laid per female changed with
female density (Box 4.2). Sato and Saito (2007a) also reported that St. longus and
St. miscanthi females (both living in nests with cooperative sociality) changed their
sex ratios according to the nest foundress density, i.e., singly foundresses produce
significantly fewer males than two or more foundresses. They simultaneously showed
that there is no difference in female fecundity between different foundress densities,
such that they could conclude that the results showed the existence of LMC. However,
because the observed sex ratios were higher than those expected from the LMC
model for haplo-diploidy (Taylor and Bulmer 1980), the question of why there is
“surplus female-biased sex ratio” in these species arose.
The foregoing two studies were conducted under experimental conditions, such
that they could only show the fact that foundress density changes offspring sex
ratio. As stated before, the most important prerequisite of Hamilton’s LMC is the
presence of spatially structured habitats in which mating takes place (Hamilton
1979; Taylor and Bulmer 1980). In nature, as we have no accurate information
about mating population structures in space and time for mite species, we must
devote much more attention if we intend to apply the LMC model to these spider
mite species (Box 4.3). This is a problem common to the other reasonings of sex
ratio variation in mites mentioned later.
There are two other models that try to explain biased sex ratio in haplo-diploid
mites. Nagelkerke and Sabelis (1996) adopted the “haystack model” of Maynard
4.3 Sex Ratio 81
Box 4.3
In relationship to these, Roeder et al. (1996) experimentally showed that sex
ratio shifts in response to the relatedness between foundresses when the
number of foundresses is equal. They suspected that such immediate
responses in sex ratio suggested that Tetranychus urticae female has some
method of discrimination between kin and non-kin.
Time
Fertilized females
Beginning
1
of season
1 1
Haystack
population
.....
growth
Females mate
with males
End of season
Fertilized females
Smith (1964) for explaining the biased sex ratio of phytoseiid mites (haplo-diploidy).
The haystack model is also explained in mite social evolution in Sect. 4.7, such that
I only describe it here in essence. The basic conditions of the model are as follows
(Fig. 4.4): Haystacks refer to resources that accumulate or appear once in the begin-
ning of a season (start of animal reproduction). These resources are distributed
patchily and are altogether isolated. Animals that were inseminated within their
natal patches in the previous season migrate randomly and singly into these new
patches and reproduce there for several generations. There is no dispersion of indi-
viduals between patches (thus they are isolated together) after immigration, and all
the haystacks deteriorate (become exhausted) simultaneously at the end of the sea-
son. Just before the disappearance of the haystacks, all adult females mate within
their natal haystacks and disperse (males never disperse). Then the life cycle starts
again at the beginning. There is no difference in immigration success between
82 4 Mite Sociality
Box 4.4
The haystack model itself is an individual-based (1-locus and 2-allele) model.
Maynard Smith (1964) presented these conditions as improbable assumptions
by use of this model. In the 1960s, there was much debate about the existence
of group selection, i.e., organisms have evolved to maintain their group, e.g.,
flock, population, or species. The most famous and influential opinion was
provided by Wynne-Edwards (1962). However, as stressed by Williams
(1966), selection basically operates on individuals, not on groups, and
situations where group selection operates are uncommon. The haystack model
has hypothesized such conditions, but as Maynard Smith stressed himself, the
necessary conditions are unrealistic.
individuals regardless of the haystacks from which they originated. Under these
assumptions, animals should produce which sex ratios? The answer is obvious.
Females reproduce as many females as possible by keeping male offspring to a
minimum. In other words, females adopt an rm (intrinsic rate of natural increase)
maximizing strategy by increasing female offspring. However, this situation is very
curious, if we remind ourselves that individuals in a species usually exchange genes
through sexual reproduction (thus they belong to a species in the sense of Mayr
1963). Because there is no interbreeding between haystack populations, each hay-
stack population could be considered a different species or thelytokous organism in
which sex ratio strategy never evolves. Therefore, as stated by Maynard Smith
(1964), these conditions must be not realized in a species with sex (Box 4.4), such
that it is difficult to utilize this model in its original state to explain female-biased
sex ratio in mites.
Then, Nagelkerke (1993) relaxed several conditions of the original haystack
model as follows (some of them were also analyzed by Maynard Smith 1964) and
simulated the optimal sex ratio. The number of immigrant females per haystack was
changed from 1 to n (more than 1) individuals (non-kin), and the maximum number
of individuals attainable at the last generation (=dispersal generation) was restricted.
If sex ratio is fixed for an individual’s traits (inflexible), the increase of n gradually
decreases the female ratio in the population, but if sex ratio is flexible for females at
each generation, it immediately reaches evenness (1/2) other than n = 1 (Nagelkerke
1993). Furthermore, he attempted to change the mating structure within a haystack
and found out that subdividing haystacks in one-generation mating groups promotes
the female bias and the unbeatable (optimal) sex ratio can change radically with
generations within haystacks.
I think that a distinct difference between the LMC model and the haystack with
unlimited growth model (Nagelkerke 1993) exists only in the number of generations
in a patch; namely, the former model assumed that only one generation (from
foundress reproduction to offspring development and mating) passes in a patch and
the latter that several generations pass there. Therefore, the latter and more complex
4.3 Sex Ratio 83
Mori and Saito (2005) suggested that the nest size is correlated with the effect of nest
defense (counterattack). Twenty-two years ago, I reported counterattack behavior by
females and males in St. longus (known as the long seta form of Schizotetranychus
celarius at that time), i.e., large nest builders, against a specific predator species, Tp.
bambusae, as mentioned previously (Sect. 3.5, Saito 1986a). This finding is a kind of
social behavior that must be considered to arise from biparental care. However, I
received a criticism from Crozier (1985) about this reasoning:
4.4 Sociality in Stigmaeopsis 85
Box 4.5
There was a reason why I could get his comments in the book Spider Mites.
Their Biology and Control (eds. W. Helle and M.W. Sabelis, 1985) before the
publications of my original studies (Saito 1986a,b); when I visited Prof. Helle
to consult about my part in the manuscript after the 17th International Congress
of Entomology in Hamburg (1984), I told him of my recent discovery about
mite parental care, and he recommended that I send my manuscript immediately
to Prof. Crozier, who was also a contributor to the book. Thus I had an
opportunity to get his comments in this influential book.
The sociobiology of spider mites has received a significant boost from the recent discovery
(Saito 1986a) that adults of Schizotetranychus celarius (which lives in a multi-chambered
web nest) eject or kill invading predatory phytoseiid mites, thus protecting immature mites
in the nest (which are not always their own offspring). If further experiments show that
these adults are acting to protect the immatures rather than just themselves, then the case
would clearly be one of parental care
(Box 4.5). Actually, showing whether St. longus (Sz. celarius) females and males
attack predators to defend their offspring or simply as self-defense posed a difficult
problem. Although I briefly wrote the reasons why I believe the former in a later
publication (Saito 1986a), a more detailed explanation is as follows.
In behavioral studies, researchers often encounter such kinds of questions when
they focus on “altruistic” traits. For example, when a raptorial bird approaches a
prey nest containing several chicks, a prey parent bird often flies away from the
nest and begins a distraction display. This behavior has the effect of startling the
predator and diverting its attention away from the nest. Why do we immediately
regard such behavior as parental care (chick defense) behavior without any
hesitation? The reasoning must be as follows: Parent birds can escape from the
predator if they need to (in other words, the predator is not dangerous to the
parents, but dangerous only to their chick(s)). Even in such a situation, parents
perform a behavioral trait that must increase the fitness of their offspring. As such,
even though the researchers cannot understand the direct “motivation” of the
parent birds, they can understand that the function of such parental behavior is
“parental defense.” Of course, if we intend to say such a behavioral trait evolved
through natural selection (raptor predator), we have to evaluate the effect of this
behavior on the fitness of offspring.
Thus, if I intend to define the behavior of St. longus parents as biparental care,
I have to at least have shown evidence such as in the case of the parent birds just
mentioned, although I had evidence of offspring fitness gain by this behavior.
One of the important criteria for stating this may come from predator–prey
interactions as observed in Saito (1986a,b). I used Tp. bambusae larvae as
“predators” for the behavioral observations. Initially I had no idea about the concept
of parental care and simply wanted to examine the prey adult behavior against the
86 4 Mite Sociality
predator. By using the weakest predator stage, I simply intended to get clearer
evidence of counterattack. After receiving Crozier’s (1985) criticism, I felt very
fortunate he had mentioned the concept of parental care. As stated before (Sect. 2.5),
Tp. bambusae larvae are nonfeeding (Chittenden and Saito 2001), such that this
predator stage never attacks the prey. Therefore, there is no need for the prey adults
to attack the predator, at least until it develops into a feeding stage. However, the
prey adults (parents) very strongly attack the predator larvae and frequently kill
them (Saito 1986a; Fig. 4.5). Furthermore, I observed that whenever an adult
female predator invaded a nest where two prey females or two males resided, one
of the parent females soon abandoned its offspring and nest while the other was
preyed on by the predator, but both the parent males attacked and were most often
killed by the predator (Fig. 4.6). There was a chance for one adult male to escape
from their nest when the other male fell victim to the predator. So why doesn’t he
take it? These observations clearly showed that the males attack the predators not
to serve themselves, but to serve their nest and offspring, i.e., they really show
biparental defense! Soon after, Saito (1986a,b, 1990c), Zhang (2002), and Mori
and Saito (2004) discovered similar counterattack behavior in St. miscanthi,
St. nanjingensis, and St. celarius, respectively, as mentioned earlier (Table 4.4; see
Box 3.7; Sect. 3.5).
Another origin of social behavior in Stigmaeopsis comes from cooperative nest
building and usage. In Fig. 4.7, an example of the nesting pattern of St. longus on
Sasa in the field is shown. Several important features are observed. First, this nest
consisted of 15 nest cells that were progressively added by nest members, one by
4.4 Sociality in Stigmaeopsis 87
1.0
0.5
2 males
0
2 4 6 8 24 48
Hours after predator introduction
Fig. 4.6 Predation by Typhlodromus (Tp.) bambusae on St. longus females and males
one. If a new nest cell was added after an older cell was exhausted, nest individuals
ceased using the older cell. However, this figure shows that all nest cells of a
continuous (united) nest are more or less utilized by nest members. Such large,
united nests (about 15 cm in length) are thought to be a result of St. longus longevity
and philopatry, and the oldest nest cell must have been built at least 1 month
88 4 Mite Sociality
Total
60
40
20
0
EG LR PC PN DC DN TM TF AM AF
Fig. 4.7 Age structure and number of St. longus in all cells of united nests. EG egg, LR larva,
PC protochrysalis, PN protonymph, DC deutochrysalis, DN deutonymph, TM teleiochrysalis
male, TF teleiochrysalis female, AM adult male, AF adult female (from Saito 1986a)
previously (note that the generational turnover occurs about every 15 days under
25°C; Saito and Ueno 1979). This observation means that some nest members had
already commenced new nest cell construction before the deterioration of the
previous one. If they have no cooperative tendencies, nest members will not begin
building a new nest cell before the old one has deteriorated, because to build a new
nest cell they have to venture outside the nest where the danger of predation is
greatly increased (cf. Sect. 3.1). Furthermore, several females sometimes initiate
new St. longus and St. miscanthi nests during the season when new leaves develop
(Saito 1987; Saito, unpublished data). These facts suggest that there is some merit
in cooperative nest construction.
A fascinating element of the social organization of Stigmaeopsis species is its
fecal manipulation behavior. As shown in Fig. 3.23, we observed one or two brown-
ish grains on the leaf surface covered by the nest (most exist at the ends of the major
nest axis). Behavioral observations showed these to be accumulations of feces
deposited by nest members (Saito 1997; Sato et al. 2003). Because mites feed off
the host plant juices, their feces are watery. Touching these watery feces is very
dangerous for the mites, because they can easily become trapped by the feces and
die (Sect. 2.3). If the nest members defecate everywhere in a nest, they may face
great difficulties. Therefore, this fecal depositing behavior strongly suggests that
4.4 Sociality in Stigmaeopsis 89
the nest members perform some cooperative nest sanitation behavior. In other
words, they maintain public lavatories within their nests.
Sato et al. (2003) and Sato and Saito (2006) determined the mechanisms by
which Stigmaeopsis mites select the defecation site. Under the hypothesis that
mites use tactile stimuli and/or chemical cues to deposit their feces at particular
sites, we designed manipulation experiments to find out whether female mites
use these two cues to maintain their defecation sites: if females use only the
tactile stimulus from the nest structure to maintain the defecation sites, they
should always defecate in the same places regardless of the location of the fecal
pile. And if they depend upon olfactory stimuli from their feces, they should
deposit their feces at the site of a relocated fecal pile, or an odor extract
thereof.
First, we determined whether all nest members tended to defecate at specific
sites. The results clearly showed that nest members of St. longus and St. miscanthi
(St. nanjingensis in China also does so; Zhang and Saito, unpublished data) defe-
cate at only “one site” inside or outside the nest. Furthermore the waste manage-
ment behavior is maintained by two simple rules. First, mites defecate near the nest
entrances if no volatile chemical cues are available, and second, when chemical
cues are available from feces deposited previously, St. longus and St. miscanthi
defecate at this site (see Fig. 3.23).
According to these rules, a foundress that does not have access to chemical cues
fixes the defecation site by tactile stimuli just after constructing the framework of
her nest. In this nesting and defecation process, Sato et al. (2003) discovered an
interesting behavior in foundresses. Just after the introduction of a St. miscanthi
female onto a new Chinese silvergrass leaf, there is no nest web, such that the
female does not defecate for a long time. After the framework of the nest has been
completed, she defecates at an entrance of her nest. Thereafter, the foundress, as
well as any other individuals that may subsequently join the nest, deposit their feces
at the fixed site based on odor cues (Fig. 4.8). This observation suggests that the
foundress delays its defecation until the completion of the nest framework!
Furthermore, Sato and Saito (2006) discovered that the chemical cues originate
from chemicals produced by host plants.
In addition, if these mites can use tactile stimuli for such behavior, why have
they additionally developed chemical cues? The nest has a simple tunnel-like struc-
ture, with two similarly shaped entrances. If the entrance structure is the only tactile
cue for recognizing the defecation site, there will inevitably be two defecation sites.
Because the nests are often extended continuously at either end (united nests) in
St. longus and St. miscanthi, double defecation sites may disturb the nest extension.
This explanation is strongly supported by the fact that St. takahashii and St. saharai
(both small nest builders) build separate nests and defecate on both sides of the nest
entrances (Sato et al. 2008) by tactile stimuli only. No chemical cues are employed.
We believe the sophisticated mechanisms that enable Stigmaeopsis mites to deposit
their feces at particular places strongly suggest that their defecation behavior
evolved under strong natural selection.
90 4 Mite Sociality
0
1st 2nd 3rd 4th
Fig. 4.8 The average time (h) from the beginning of nest foundation to the first defecation (1st)
and the average intervals from one defecation to the next after the first defecation (2nd, 3rd, 4th)
with SE bar (from Sato et al. 2003) in St. miscanthi females
From the biparental defense as well as from the cooperative nest building and
utilization (nest sanitation) behavior, it can be concluded that St. longus and
St. miscanthi have a highly developed level of sociality (now considered
communal sociality; Mori and Saito 2005), the first such discovery in the Acari.
Furthermore, Mori and Saito (2005) indicated that as nests become smaller, the
counterattack behavior decreases (Fig. 3.24), meaning that a conflict exists
between nest size and social development. As expected, large nests can contain
many individuals and can persist longer as both food sources and living spaces
(Sect. 3.5). Because the defensive success increases with adult density per nest
(Fig. 3.22, Box 3.7), an additional strategy becomes apparent in St. longus and St.
miscanthi, i.e., making larger nests to enhance the efficiency of cooperative nest
defense by many members, even though these larger nests are more vulnerable to
predator invasion. As such we see cooperative nest building, cooperative nest
sanitation, and cooperative nest defense in the large nest building Stigmaeopsis
species. Although cooperative feeding by adults is an important criterion for
defining cooperative breeding (Wilson 1975), I have not yet observed such a
behavior. One reason why there is no cooperative feeding might be related to
phytophagy, because juveniles always live on a food source, i.e., host leaves, on
which there is no difficulty in feeding by themselves. In other words, there may
be no need for parents to feed their young (Box 4.6).
Meanwhile, I wish to discuss some by-product discoveries from our studies
on fecal manipulation mechanisms. St. miscanthi lives on Miscanthus grass and
St. longus on Sasa; both these species are large nest builders. Therefore, whether
4.4 Sociality in Stigmaeopsis 91
Box 4.6
The possibility that there are other kinds of parental behavior in these mite
species cannot be rejected, e.g., egg sanitation as protection from fungal disease,
as known in earwigs (Lamb 1976). Mori et al. (1999b) observed that the eggs of
Stigmaeopsis longus with a parent female showed a slightly higher survival rate
than those without a parent female under field conditions, although the difference
was not significant. Furthermore, Kikuchi (unpublished) discovered that a
fungus species that is dangerous to St. longus sometimes occurs in its nests. A
recent study suggested that parent female attendance is important to serve eggs
within nests under field conditions (Kanazawa, unpublished data).
Not effective
Effective
80 80
Frequency (%)
Frequency (%)
60 60
40 40
20 20
0 0
miscanthi miscanthi longus longus
to to to to
longus
its own feces miscanthi longus feces its own feces miscanthi
to to
feces
other individual’s other individual’s
feces Treatments feces
Fig. 4.9 Effectiveness of feces of different species as cues of waste management (from Sato et al. 2008)
of evidence discovered is that St. miscanthi immatures can develop on Sasa, but
St. longus immatures cannot develop at all on Miscanthus (Sakagami et al. 2009).
Taking these three observations together strongly suggests that St. miscanthi
evolved from a common ancestor of St. longus living on Sasa bamboo and that the
host plant shift from Sasa to Miscanthus might have played a great role in this
speciation event (Sakagami et al. 2009). Therefore, we are now thinking that communal
sociality evolved in St. longus (or its ancestor) once, then a social St. miscanthi
branched out through host plant shift.
So far, I have only addressed sociality in the genus Stigmaeopsis. Are there any
developments in sociality in other spider mite species? The answer may depend
upon how sociality is defined, i.e., in the narrow or broad sense. As stated briefly
when discussing life type variation, there is diversity in defecation behavior among
spider mite species. In the WN life type especially, we now know that most species
have more or less fixed defecation locations in their nests; e.g., in Schizotetranychus
brevisetosus, Eotetranychus suginamensis, and Eotetranychus shii, nest members
deposit their feces on the outside surface of the woven nest roof, Eotetranychus
querci and Schizotetranychus shizopus place their feces around the inside margin of
their nests, and Schizotetranychus saitoi, Schizotetranychus approximatus, and
Schizotetranychus laevidorsatus defecate just outside of their nests (Box 4.7). Such
life type features strongly suggest that this behavior is to prevent nesting spaces
from becoming dirty. In their nests, eggs and immature stages are cohabiting, such
that this behavior could be considered a kind of parental care. Furthermore, the
nest-building behavior itself serves to protect offspring from predators, as shown in
Sz. recki in Sect. 3.4. Therefore, we could say that species having WN life types
possess parental care behavior, which has long been considered a prime criterion of
subsociality (Wilson 1975; Saito 1995b). Waste management is not always
restricted to the WN life type, however. As stated previously, several species having
CW life types only deposit their feces on a CW web. This behavior must prevent
the leaf surface from becoming dirty, but I hesitate to define such mites as subsocial
based solely on this behavior.
In Figs. 2.3–2.6, I have shown the various social behaviors known in spider
mites. In these figures, I adopted three criteria to define the level of mite social
development as follows: the length of parent–offspring attendance, nest building by
cohabiting parents, and offspring and feces management behavior, as considered in
Stigmaeopsis. Readers of this book may differ in their views and some may perhaps
think that there is no merit in defining such species as subsocial or primitively
social. Even if so, I believe that close relationships (other than sexual ones) between
individuals must give us some background into social evolution in animals. The
first criterion is closely related to life history patterns in spider mites because par-
ent–offspring attendance is primarily determined by whether parents have sufficient
4.5 Sociality in Spider Mites 93
Box 4.7
Why don’t Thailand species, whose life types closely resemble Japanese
Stigmaeopsis species, have the habit of depositing their feces at particular
sites? My co-researcher, Y. Sato, had a fascinating suggestion, i.e., the flush
toilet hypothesis. In tropical regions, very heavy rainfall frequently washes
plant leaves. She thus supposed that there is no need for such species to
manipulate their feces, because rainfall always removes any feces that have
been deposited outside nests. Actually, these species deposit feces all around
the perimeter of the nest under experimental conditions, but we could not
observe any feces there in the fields. Of course, it is still just a story.
In relationship to the above, another curious finding is that Schizotetranychus
saitoi (Thailand species) lives on the upper surface of bamboo leaves, although
the other two species in Thailand and all species of Stigmaeopsis live on the
undersurface. As described, although very heavy rainfall in tropical areas
must make it more difficult for Sz. saitoi to live on upper surfaces of leaves,
why has it such a habit? The fact that the nest made by this species is very
solid seems to be related to this question, but I have no idea what factors have
caused Sz. saitoi to live there.
time to attend their offspring. In Sect. 2.2, I addressed life history variation among
several species (see Fig. 2.1). It is apparent from this figure that there is a great
overlapping of generations (at most three generations) in St. longus. Mothers ovi-
posit for a long time and the relatively short developmental time of immature stages
creates generational overlapping. Such life history peculiarities thus provide a
background for Stigmaeopsis sociality (a reminder that human beings also have
such peculiarities). If we agree to this, all spider mite species enjoy more or less the
same background, although there is variation in the extent of generational overlap-
ping among species.
Aggregating at a certain place for a long period is also a prerequisite of social
organization, because close interaction between individuals is only possible if they
have frequent contact (viscosity). This requirement is realized by nesting behavior
or chemical control (aggregation and dispersion pheromones are known in acarid
mites; Kuwahara et al. 1982; Kuwahara 1990). In spider mites, there is little evi-
dence of aggregation by chemical cues [although there are sex pheromones (Cone
1979) and attractive odor from the plants infested by conspecifics (Pallini et al.
1997)], but we do know that there are plant mite species that form large aggrega-
tions on the plant leaf surface without any nest-like structures (Box 4.8).
As such, in the narrow sense at least, subsociality is only recognized in spider
mite species having WN life types. Of course, if we extend the definition of subso-
ciality to include species showing short-term egg care, i.e., egg web covers, most
spider mite species could be categorized as primitively social or subsocial.
94 4 Mite Sociality
Box 4.8
We have observed a tremendous aggregation of Tydeus sp. on the flat
undersurface of durian leaves in Thailand (see below). We sometimes
observed similar aggregations of tydeiid and asciid mites on plant leaves.
Kuwahara et al. (1982) reported aggregation pheromones in acarid mites
living on stored products, but there is no information about the functions of
such aggregations on plant leaves. As stated in the text, aggregation is one
of the prerequisites of social evolution, such that these are interesting from
an evolutionary point of view.
Are there any mite species other than spider mites that have social behavior? It has
been reported that several mite species belonging to various families have sociality.
Some predacious cheyletid mites belonging to the order Actinedida (spider mites
belong to the same order) are supposed to show sociality because they live
gregariously in nests built with silken threads. Hemicheyletia (Hm.) morii shows
several special behaviors (Mori et al. 1999a). They use thin threads to construct
nest-like structures that resemble woven mats at the curled edges of the undersides
of leaves. The web-mats serve as footing to prevent individuals catching powerful
prey from being dragged away from their nesting site. Eggs are oviposited on the
web mats and then covered by anchoring threads (Fig. 4.10). A nest usually
contains many individuals of various stages (Fig. 4.11), but there are no males,
suggesting thelytokous reproduction (rearing under experimental conditions
showed this species is thelytokous; Mori et al. 1999a). All active mite stages stand
in a circle around the periphery of their woven mat, with their powerful pedipalpi
opened toward the outside (see Fig. 4.10). When a prey item ventures too close
and comes into contact with a cheyletid’s pedipalpi, the cheyletid seizes a prey
appendage (leg, pedipalpus, bristle, and so on) with its pedipalpi, then inserts its
4.6 Overview of Mite Sociality 95
Fig. 4.10 Micrographs of Hemicheyletia (Hm.) morii. Left, using an optical microscope; right,
SEM with cryo-system attached
6
Number of individuals
5
4
3
2
1
0
33
32 30
31 29
28
272625
24 22
23 21
20
1918
1716 Eggs
15
Nest numbers 1413 Larvae
1211
10 Proto-
98
76 nymphs
54 Deuto-
32
Adults
nymphs Mite stages
1
Average
Fig. 4.11 Stage structure of Hm. morii nests (from Mori et al. 1999a)
96 4 Mite Sociality
chelicerae into the appendage (perhaps injecting toxin) and commences feeding.
Immediately after a predator captures prey, other nest members swiftly respond to
the prey’s struggle by taking hold of other appendages and feeding on them.
Through such group predation behavior, this mite species can capture very large,
strong prey species relative to their body size, such as aphid larvae, thrips, and
phytoseiid adults. There is no aggression between nest members even during
feeding. This was the first discovery of a sit-and-wait group hunting species in
Acari (Mori et al. 1999a; but also see Cloutier and Johnson 1993).
Cooperation between predacious conspecifics is not a common phenomenon in
arthropods, although several examples have been reported in spiders and non-spider
arachnids (Aviles 1997; Uetz and Hieber 1997; Rayor and Taylor 2006). The reason
for such extraordinary hunting behavior, i.e., group predation in Hm. morii, can be
explained by both their thelytokous reproduction, which inevitably increases
the relatedness between nest members, and their small size, which must require
cooperation to extend their menu of victims. We sometimes observed that if an
Hm. morii individual captured a large prey item – an adult phytoseiid female, for
example – without any help from other individuals, it was soon dragged away from
its web mat, suggesting that the cooperation of several individuals is necessary to
capture large and strong prey items. Furthermore, through this behavior, parent
females can provide food to their offspring, because immature stages always get
feeding opportunities when their parents (and other group members including them-
selves) capture prey. A marvelous behavior observed in this mite is that one of the
nest members usually tries to throw the victim’s dead body from their nest after
predation, which may function as nest sanitation as well as cleaning the trapping
site (Mori et al. 1999a). From the aspect of social organization, such a behavior
may look like a kind of altruism for the other individuals, but I think that this is only
under the rule that the last individual feeding the victim must do so even if it is
alone. Anyhow, we should categorize this species as having cooperative sociality.
In relationship to this, Boczek (1959) reported that the Cheyletus eruditus females
protect their clutches until the larvae emerge, suggesting that there are other chyl-
etid mite species that show a social behavior.
Another example appears in Varroa jacobsoni (Donzé and Guerin 1994). This
species is a male-haploid mite parasitizing honeybee broods. The mites accumulate
their feces at a particular site on the wall of the honeybee’s brood cell and use it as
a rendezvous site for mating. Furthermore, the foundress only feeds at a single site
on the bee larva, probably increasing the survival of her offspring through the
avoidance of drowning as a result of the hemorrhaging of the host.
There is also an interesting example of finely organized subsociality in moth ear
mites. I believe that subsociality in the Acari must have originally been discovered
in this group by Treat (1975), although he understatedly said that he had insufficient
data. Several Dicrocheles phalaenodectes females, which parasitize the ear chambers
(tympanic organ) of noctuid moths, utilize different chambers of the ear for differ-
ent purposes: the first and second chambers are used for feeding and oviposition,
and the third is used for mating and the deposition of feces. Thus there is a regulatory
mechanism for public sanitation between room members. Furthermore, adult mites
4.6 Overview of Mite Sociality 97
defend their chambers against intruders. Because the chambers include many off-
spring, this defense behavior may represent a kind of parental care, even if they do
not give any special care directly to offspring. This species is assumed to be male-
haploid (Norton et al. 1993).
Some phytoseiid mites aggregate and sibling individuals interact without
cannibalism. In one interesting report of predatory behavior in phytoseiid species,
Cloutier and Johnson (1993) observed that young nymphs of Neoseiulus (Amblyseius)
cucumeris have difficulty preying on thrip larvae by themselves, such that they can
develop normally only when conspecific gravid females hunt the prey for them.
Although Cloutier and Johnson did not consider this phenomenon as social organi-
zation, I suggest that these mites perform a kind of parental care, i.e., feeding their
young. As previously mentioned (Sect. 2.5), there is another kind of social organi-
zation, i.e., avoiding cannibalism in phytoseiids and Hp. aculeifer (Usher and Davis
1983). Macrocheles superbus shares oligochaete worms (Blazak et al. 1990), but
Walter and Proctor (1999) said that cooperative hunting is unlikely, because the
initial attack is by a single mite and then either the thrashing of the prey attracts
others or they stumble across the dying prey and commence feeding.
Furthermore, I intend to address some fragmental knowledge about Tp. bambu-
sae and Ag. iburiensis, both of which have been introduced as important predators
of Stigmaeopsis. As I have described in Sect. 2.5, Tp. bambusae has nonfeeding
larvae, which was hypothesized to be an adaptation for avoiding sib-cannibalism
(Chittenden and Saito 2001), and females of this species seldom cannibalize their
own offspring (Saito, unpublished data). When I first discovered the counterat-
tacking behavior of St. longus (Saito 1986a,b), I simultaneously reported on the
difference in counterattacking behaviors between St. longus females and males (see
Sect. 4.4): When an adult female Tp. bambusae invades an St. longus nest where
two females are cohabiting, one of the females could usually escape from the nest
while the other female was being eaten (see Fig. 4.6). On the other hand, the preda-
tor female will always eat two male St. longus if they co-defend a nest (Fig. 4.6).
From the point of view of St. longus, these phenomena can be easily explained as
follows: St. longus females have another chance of reproduction elsewhere, but
males do not, such that males are strongly concerned with the survival of their
offspring even if they can rarely win against such a strong predator (Sect. 4.4).
However, if we change our point of view from the prey to the predator, we may
attain a different explanation: The predator female that intrudes a new prey nest first
sweeps for dangerous males (and females) that will counterattack against her
offspring (Saito 1986a). The observation that Tp. bambusae females seldom canni-
balize their own offspring under no-food conditions now inclines me to believe the
latter possibility.
Ag. iburiensis also shows a kind of social behavior. In Saito et al. (2008a), we
discovered that the scattered nest effect hinders this predator (Sect. 3.5). At the same
time, however, we observed that this predator female has a tendency to repeatedly
return to nests that it has previously occupied (although this trend was statistically
insignificant). Because there are usually several eggs laid by the predator female in
such nests, this tendency may be related with maternal care, i.e., subsociality.
98 4 Mite Sociality
Although this habit seems to enhance the effect of void nests as decoys, there are
undoubtedly reasons for the predator female to do this. It may not be so unusual as
to think that the predators of social prey species could also be social. Of course, this
is still only an idea.
Anyhow, with the exception of tetranychids, the level of social organization that
suggests a beginning stage of sociality is only known in a small number of species
scattered over several different taxa. An important point from such a rough review
of sociality in mites is that all examples other than tetranychids are predators or
animal parasites. I have no idea how to explain this. On the other hand, it has been
suggested that gregarious living within narrow spaces, such as in bee brood cells,
and the associated contact of individuals can result in close interactions between
sibs. Fecal manipulation behavior is inevitable to realize long-term gregarious
living in the same space; thus Donzé and Guerin (1994) suggested that this is a key
behavior for the evolution of sociality. In the same context, Jackson and Hart
(2009) recently discussed about the importance of the reducing mechanism (blind
gut) of fecal load in eusocial evolution in hymenoptera. I do not think that the
statement by Donzé and Guerin (1994) is always true, because such behavior is
merely a prerequisite for gregarious living and/or only a result (not a cause) of
social organization.
Last, I briefly consider the Rayor and Taylor (2006) review of sociality in non-
spider arachnids. They examined the patterns of parental care, duration of associa-
tion, and the presence of social traits found in the most social taxa of non-spider
arachnids. Species in most arachnid orders have transient parental care with defense
of eggs, a brief period of association with newly emerged young before indepen-
dent foraging and explosive dispersal from the natal nest. More prolonged sociality,
with long-term associations among mother–offspring–sibling, is rare and has only
been described in a few species in the Amblypygi, Scorpionida, Pseudoscorpionida,
and Acari. All such species have subsocial origins (i.e., there is no example of
parasocial origin; cf. Wilson 1975).
There is no doubt that several factors or backgrounds are important for the evolu-
tion of sociality in mites. Some of these have been already mentioned. Furthermore,
Tallamy and Wood (1986) stressed the peculiarities of resources, and Crespi (1994)
suggested that one of the requisite conditions for the evolution of eusociality (caste)
is possession of shelter-like structures providing food. Rather than repeat such
comprehensive reasoning here, I would prefer instead to concentrate on factors
peculiar or common to mites. Of course, the peculiarities mentioned previously,
such as life history, genetic systems, sex ratio, and sexual behaviors are all funda-
mentally related to mite sociality (Box 4.9).
Hamilton (1964) first formalized that cooperative behavior, especially altruistic
behavior, evolved under kin selection (inclusive fitness concept; Box 4.10 ):
4.7 Background of Social Evolution in Mites 99
Box 4.9
Phylogenic constraints, behavioral and ecological traits, and environmental
factors related to the evolution of the sociality of several Stigmaeopsis species.
Phylogenetic
and genetic Behavioral and
Environmental backgrounds ecological traits
Species
factors (constraints) and outputs Social elements
Stigmaeopsis saharai
Stable and Stigmaeopsis takahashii
Nesting habit Small nest
abundant Ability to
resources in produce threads
habitat Large nest
(evergreen and Philopatry
Long life
perennials)
High predation
Spatio-temporal
pressure
overlap of generations
Relatively
Stigmaeopsis celarius
large eggs
Stigmaeopsis nanjingensis
Iteroparity
Stigmaeopsis miscanthi
Cooperative nest
Low reproduc- repairs and
tion rate enlargements
Stigmaeopsis longus
(arrhenotoky) Female biased Cooperative brood
sex ratio defense by females
(maternal care)
Low inbreeding High relatedness Kin
depression between females selection
Inbreeding in a nest Brood defense by
Male can
a single male
overwinter
Outbreeding Low relatedness (paternal care)
(in warmer Egoism
region) between males
Brood defense by
Male cannot Inbreeding High relatedness Kin multiple males
overwinter between males selection (cooperative
(in cooler sharing)
region)
Box 4.10
The concept of inclusive fitness provided by Hamilton (1964) is a key to
understand his kin selection theory. He defined two fitness components: one
is personal (individual) fitness and the other is the fitness in other kin
individuals; he formalized that gene evolution must occur through these two
passageways. This is the famous kin selection theory: if your genes exist in
your full sib-brothers and sib-sisters at the probability of 0.5, then your life is
just equal to two brothers or sisters from the aspect of gene evolution.
300
200
n = 601
m = 2.94
100 SD = 5.36
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
No. of foundresses per nest
Fig. 4.12 Spring nest foundation pattern in St. longus (from Saito 1987)
Box 4.11
How the single foundation of an unfertilized female increases relatedness
between male nest members in Stigmaeopsis (St.) miscanthi. No male
immigration or male survival until the next generation was assumed.
Unfertilized female
A/ B
Mothe-son mating
A
First generation
A A A B
A A 1 1 1 0
A B A
A 1 1 1 0
. . A 1 1 1 0
. A B A A .
. B 0 0 0 1
.
r = 10/16=5/8
102 4 Mite Sociality
Late Apr. to
early May Overwintered
aggregation
No male
period
Spring nest Feb. to Apr.
foundation on Sasa
Sib-
Population increase mating Overwinter
on old Sasa leaves under snow
Sib- Partial dispersal
mating and new foundation
Sib-
Middle Jun. mating
to early Jul. New Sasa leaves
coming out Dispersion
Sib-
mating Sib-
mating
Dec. to Jan.
Dispersion
Population increase
on old and new Termination
Early Oct. to leaves of diapause
middle Nov.
Enter diapause
Overwinter
aggregation
, nest; * unfertilized
females. Some of these mated females then disperse to establish new nests, although
most females reside inside their parental nests and begin oviposition (Fig. 4.13).
Such a life cycle suggests a high relatedness between nest members. V. jacobsoni,
D. phalaenodectes, Hm. morii, and WN type tetranychids are regarded to be phylo-
patric and are also more or less under sib mating conditions.
Such circumstantial evidence suggests high relatedness between the individuals
of the mite species that live within nests or closed spaces, although no direct data
about relatedness are available yet. How does the benefit increase by such kin-
selected traits? What kinds of cost are there? Without such data, kin selection,
despite its obvious appeal, remains only a plausible explanation at best.
There may be another line of explanation about social evolution in mites; it
concerns the group selection model. Maynard Smith (1964) proposed the “haystack
model” to show the improbability of group selection, at a time when the Wynne-
Edwards (1962) group selection reasoning had been prevalent (see Box 4.4).
Although this model has been referred to for explaining biased sex ratio in mites
(see Sect. 4.3; Nagelkerke and Sabelis 1996), I once again address whether this
model is applicable to mite sociality.
Let us consider that a number of haystacks (habitats with food) emerge at the
beginning of an “animal season” in a field, where fertilized females can singly
4.7 Background of Social Evolution in Mites 103
invade and reproduce for several generations (Fig. 4.4). All the haystacks are
exploited for a certain period; thereafter, all individuals leave their haystacks simul-
taneously and gather in a particular place to mate (note that this situation is a little
bit different from Sect. 4.3). The mated females again individually invade “haystacks”
that have regenerated in the field. The necessary conditions of this model are
(1) haystacks must, for a time, be reproductively isolated, (2) haystacks must
be started by one foundress (randomly selected from all mated females), and (3) there
is no dispersion between haystacks (the food supply is discontinuous in space; each
patch of food supports its own haystack group). Let us denote that the probability
of assortative mating is R (mating between the individuals from the same haystack)
and that of random mating is 1 − R. The haystacks invaded by single “timid”
females mating with “timid” males can produce “1 + G” offspring, while those
invaded by single “aggressive” females mating with “aggressive” or “timid” males
only produce “1” offspring (“1” does not mean the real number of offspring here,
but a unit, thus a relative value). Under such circumstances and life cycles, there are
conditions under which genes coding for cooperative traits (timid–scramble) can
increase their frequency by suppressing egoistic traits (aggression–contest) depend-
ing upon G and R values. Maynard Smith (1964) showed this with a one-locus and
two-allele model. He showed that the condition necessary for the evolutionary
spread of “timid” behavior is R (1 + G) > 1. If there is no inter(out)breeding between
colonies, even at migration (R » 1), timid behavior will evolve, provided it is advan-
tageous to the group; i.e., genes producing more offspring in a haystack increase.
On the other hand, if all dispersers mate randomly, i.e., R = 0, there is no condition
where the “timid” trait will evolve. For an intermediate R value, there must be con-
ditions under which the “timid” trait will evolve depending upon the G value,
although Maynard Smith (1964) stated:
If the admittedly severe conditions listed here are satisfied, then it is possible that behavior
patterns should evolve leading individuals not to reproduce at times and in circumstances in
which other members of the species are reproducing successfully. Whether this is regarded
as an argument for or against the evolution of altruistic behavior by group selection will
depend on a judgment of how often the necessary conditions are likely to be satisfied.
As described so far, the more or less inbreeding nature of spider mites and other social
mite species, as well as their nesting habits, are thought to partly satisfy the conditions
hypothesized in the haystack model: St. longus females tend to establish nests singly
(Fig. 4.12), the nests persist long enough for the reproduction of several generations,
offspring females mate with their sib males in their maternal nests before dispersal (it
is analogous to high R), and their cooperation in nest defense may increase their
offspring survival (higher G). Therefore, I have to agree with the possibility that the
haystack model can also explain the social development of mites, although we have no
accurate data whether the three necessary conditions of Maynard Smith are satisfied.
Rather, I think, however, that the kin selection and the haystack models are not
alternative models, because habitat structure is not considered in the former and
relatedness is not considered in the latter. If we calculate the relatedness between
the members of the haystack in the latter model, we may immediately notice that
104 4 Mite Sociality
most of them are very close kin. In other words, the habitat structure hypothesized
in the haystack model inevitably creates a situation under which individuals having
beneficial traits for kin interact strongly. Thus, these two forms of reasoning may
be regarded as two sides of the same coin, I believe.
Chapter 5
Inbreeding Depression in Haplo-diploidy
As stated before, there is little doubt that kin selection is an important prime driver
of social evolution. Inbreeding sometimes causes high relatedness between indi-
viduals, as is the case with social spider mites (note that inbreeding is a result of
life patterns but is not a prerequisite for kin selection). However, in the animal
kingdom, inbreeding is believed to be deleterious because of inbreeding depression
(Charlesworth and Charlesworth 1987). There are two contradictory views on the
genetics of haplo-diploidy in relation to inbreeding effects. One prediction is that
there are few deleterious genes in haplo-diploidy because such genes had been
selected out through haploid (hemizygotic) males (Smith and Shaw 1980; Atmer
1991). The other prediction is that there is a sufficient number of deleterious genes
that only affect diploid females even in haplo-diploid organisms (Crozier 1985).
I once contemplated the possibility that the genetic peculiarities of haplo-diploidy
might be related to the fact that there are so many eusocial species in haplo-diploid
organisms (Saito 1994b). We then tried to find out whether there are such genetic
peculiarities in spider mites.
Crozier (1985) pointed out that spider mites are good model organisms for
understanding the genetics of haplo-diploid organisms, because hymenopteran
insects having haplo-diploidy and eusociality are difficult animals for genetic stud-
ies. Using a social spider mite, St. miscanthi, we observed how strong inbreeding,
namely mother–son mating, causes inbreeding depression. As a result, we could
observe significant inbreeding depression in female fertility with the increase of
Wright’s f-value (Wright 1922), but no depression in immature survival (Fig. 5.1;
Saito et al. 2000b). Furthermore, we detected that the genes causing this depression
effect are mildly deleterious recessive genes (Fig. 5.2b; Saito et al. 2000b). Through
the inbreeding experiments, we could establish two lineages by mother–son mating
for 11 generations; i.e., one lineage had no depression (normal lineage) and the
other lineage showed about one-half decrease in fertility by strong depression
(depressed lineage). Then we conducted a cross-experiment between them. As shown
in Fig. 5.2a, the F1 females produced by the cross between these two lineages ovi-
posited at a similar rate to the normal lineage. Then we back-crossed the F1 hybrids
with normal and depressed lineages. The F2 females obtained from the back-cross
showed an intermediate oviposition rate between normal and depressed lineages
a 10 b 10
(corrected values)
-10
-5
-20
-10
-30 -15
0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 0 .2 .4 .6 .8 1
f-value f-value
Fig. 5.1 Inbreeding effect on oviposition was very significant, but not on immature survival, in
Stigmaeopsis miscanthi (from Saito et al. 2000b). a Indirect inbreeding effect on the survival rate
of progeny females (produced by inbred pairs). b Parental inbreeding effect on the reproductivity
of female progeny (observed). Number of eggs laid for 10 days was corrected by environmental
variance observed in the outbreeding control. Thin lines in b show significantly depressed lin-
eages.
a b
9
Average no. of eggs laid for 10 days
Average no. of eggs laid for 10 days
9 b b 8
8 b 29 28 36 35 32
26
7
7
6 34
6 35
5
5 a
4
4 19
31
3
3
2
2
1
1
0
L2C5B2
0
L5C2B2
L5C2B5
L5C5C5
L2C5B5
L2C2C2
L2C2
L5C5
L2C5
L5C2
Combinations of cross
Back cross combinations
Fig. 5.2 Cross-experiments clearly showed that depression in female fecundity is caused by some
deleterious recessive genes. L2 is a lineage depressed by inbreeding; L5 is a lineage not depressed
by inbreeding in Fig. 5.1b. a Fecundity of fertilized females produced by parental crossing. L2C2
offspring of L2-female × L2-male, L5C5 those of L5-female × L5-male, L2C5 those of
L2-female × L5-male, L5C2 those of L5-female × L2-male. Between different alphabet characters,
differences are significant at P<0.01. b Results of back-cross experiments using the crossed lin-
eages. Between the combinations linked by horizontal bars, there are significant differences at
P < 0.03. Vertical bars are standard deviations. (from Saito et al. 2000b)
(Fig. 5.2b). Therefore, we could determine that the genes governing the depression
of oviposition are recessive.
The foregoing discovery (Saito et al. 2000b) is important, because it was the first
paper in which clear inbreeding depression by recessive genes was detected in
5 Inbreeding Depression in Haplo-diploidy 107
haplo-diploid organisms (see also Helle and Overmeer 1973). Thus we could prove
Crozier’s (1985) hypothesis that recessive deleterious genes exist mostly in the
female. However, we met a difficulty in our attempts to publish this discovery,
because we started the inbreeding experiment from parents that originated from two
distant (~300 m) populations to gain genetic variability in the parental lineage.
There is a difficulty in regarding this result as evidence of inbreeding depression,
because of the possibility of heterosis in the parent generation. It was impossible to
determine whether the depression was caused by the breakdown of heterosis or by
the expression of deleterious genes retained in a population, because we had started
our inbreeding experiment from a mixture of two adjacent populations. If we focus
on a short-time process only, then this is a valid criticism, but we should expect
such heterosis to readily occur among field populations in nature, because these
sub-populations are adjacent. Anyhow, our primary purpose was to ascertain
whether there are deleterious genes masked by dominance-recessive genetics in
haplo-diploid females, and we presently believe that our findings have sufficiently
proven this to be so.
However, the doubts still lingered. Then we (Mori et al. 2005) conducted a simi-
lar inbreeding experiment using parental lineages originating from a single popula-
tion occurring far from the populations used in Saito et al. (2000b). There was very
little depression on immature survival of inbred lineages in all populations. On the
other hand, in the two of eight inbred lineages, all of which originated from small
populations, female oviposition decreased significantly with the increase of
Wright’s f-value, showing that mildly deleterious genes are actually retained even
in natural populations of haplo-diploid organisms (Box 5.1).
From the foregoing two studies (Saito et al. 2000b; Mori et al. 2005), we could
demonstrate two important facts; i.e., deleterious genetic factors causing depression
of female fecundity actually exist in haplo-diploid mite populations, and such
depression is not always caused by the breakdown of artificially created heterosis,
but by genetic factors retained at the semi-isolated natural population level. If the
inbreeding depression observed in Saito et al. (2000b) was a chance result of an
artificial crossing, we would have a very slim possibility of observing it again in
Box 5.1
Henter (2003) recently reviewed the nature of inbreeding depression in
haplo-diploidy in comparison with diplo-diploidy and showed several impor-
tant phenomena. She detected severe inbreeding depression in longevity that
is not obviously sex limited in a haplo-diploid species. The discrepancy
between her results and ours may be caused by differences in approach and
materials. In our two studies (Saito et al. 2000b; Mori et al. 2005), inbreeding
effects on several traits such as male insemination ability and adult male
longevity were ignored, such that further studies on these traits are also necessary
to understand in detail the genetic structure of haplo-diploid mites.
108 5 Inbreeding Depression in Haplo-diploidy
field populations. However, we could show that the factors causing depression are
frequently retained in wild mite populations (Mori et al. 2005). This result means
that effects such as the high frequency of outbreeding (it may sometimes cause
heterosis), pleiotropic effects, and/or mutation–selection balance are operating in
natural populations (Crozier 1985; Saito et al. 2000b).
As such, we know that most inbreeding effects in haplo-diploidy are female sex
limited and that animals possessing this genetic system have a high tolerance to
inbreeding, i.e., low depression in immature survival. These peculiarities might
permit spider mite inbreeding, such that the high relatedness between interacting
individuals through mother–son and sib mating may create the conditions under
which kin selection frequently operates.
Meanwhile, another plausible factor for social evolution can be detected from
the nature of inbreeding depression in haplo-diploidy. As observed by Saito et al.
(2000b), reproductive inequality among the diploid females of haplo-diploid organisms
is frequently caused by inbreeding depression. This means that we must observe
“semi-sterile” females in haplo-diploid organisms more frequently than diplo-dip-
loid organisms, because most females produced by inbreeding in the latter are
selected out, especially as immature stages, by their low tolerance to inbreeding
depression. Therefore, Saito (1994b) hypothesized that such differences in toler-
ance to inbreeding must be related to the frequent occurrence of nonreproductive
castes in females of haplo-diploidy. For example, St. miscanthi females having
deleterious homozygous recessive genes suffer depressed reproduction, but most
can still construct their nest webs normally (Saito, unpublished data). If such web
nests serve as protective refuges and females perform nest defense for their rela-
tives (possibly including sisters having recessive and normal genes heteroge-
neously), we may simultaneously recognize “semi-sterile castes = depressed
females” and “reproductive castes = normal females” within a nest. By this reason-
ing, I have proposed a possible genetic constraint that may explain some of the
still-unresolved problems in eusocial evolution, i.e., a mechanism for the occur-
rence of reproductive castes, an explanation for why caste differentiation only
occurs in females in haplo-diploidy, and for the frequent evolution of eusociality in
haplo-diploidy. Whether such reasoning is flawed, as per Keller’s (1995) objection,
or plausible must be determined by future studies (Saito 1995c). It was recently
shown in ants that interlineage crossing, thus outbreeding, causes caste differentia-
tion (Glennis et al. 2002). This process is completely contradictory to my reason-
ing, because it results in reproductively suppressed castes (workers), which are
heterozygotes. However, the important point is that both my own reasoning and the
discovery in ants showed that eusocial evolution is sometimes caused by “genetic
factors.” In any event, I believe my reasoning may contribute to future hypotheses
about the origins of eusociality in other haplo-diploid organisms (Saito 1997).
Chapter 6
Kin Selection
So far, I have discussed that mite sociality may evolve under kin selection, although
there still remain several other possibilities. In this chapter, I address more detailed
and more realistic phenomena on kin selection in mites.
While observing the social behavior of Stigmaeopsis (St.) longus and St. miscanthi,
I noticed a considerable difference in male pugnacity between these two species.
I subsequently observed male-to-male interactions in St. longus and St. miscanthi
as follows. Web nests in which St. longus or St. miscanthi females deposited several
eggs before the following experiments were provided. After all females had been
removed from their nests, two strange St. longus or St. miscanthi males (from
different stock cultures of the same species) were released there. These nests were
then observed for 5 days. There was a considerable difference in male-to-male
behavior between these two species (Fig. 6.1). St. longus males cohabited in nests
for long periods and showed no aggression toward each other. On the other hand,
St. miscanthi males always engaged in mortal combat that inevitably resulted in the
death of one of the paired males during the 5-day period.
In the nests containing many individuals, multiple St. longus males show pre-
copulatory mate guarding of the same female (Sect. 4.2) without any aggression
whatsoever (Fig. 6.2), but St. miscanthi males fight to the death when they meet in
the nests. Furthermore, the victor often sucked the loser’s body contents, i.e.,
showed cannibalism (Fig. 6.2). Since these initial findings, I have carefully
observed many St. miscanthi nests in stock cultures and found numerous dried-up
dead bodies of males inside and outside the nests. I am convinced that such strong
killing behavior is ordinal in this species.
I was really surprised to observe such extraordinary male behavior in St. mis-
canthi because spider mites are herbivorous. Just what is it that St. miscanthi males
are competing for through mortal combat? I supposed that the aggressive behavior
of St. miscanthi males must be to acquire females in a nest, i.e., to form a harem.
S. miscanthi S. longus
100
80
60
100
50
0
0 1 2 3 4 5 0 1 2 3 4 5
Days after introduction of 2 males
Fig. 6.1 There is a major difference in male pugnacity between St. (S.) miscanthi and
St. (S.) longus when two males are introduced into a nest (after Saito 1990c; Saito and Mori 2005)
200 µm
Males
Quiescent female
Fig. 6.2 Multiple male precopulatory guarding in St. longus (right); St. miscanthi male killing
conspecific male (left)
6.1 Variation in Male Aggressiveness Between Species 111
86
70
62
62 70 80 86
Loser’s leg I length (µm)
Fig. 6.3 Difference in size of leg I between winners and losers in male-to-male combat of
St. miscanthi (from Saito 1990d)
Table 6.2 Death probability and time until death on one of paired males in kin-paired and
stranger-paired male-to-male combat of Stigmaeopsis (St.) miscanthi (from Saito 1994a)
Five days after pairing Time until one male’s
Experiments No. of tests Both alive (%) One killed (%) death (mean ± SD days)
Between brothers 53 9(17) 44(83) 2.8 ± 2.0
Between non-kin males 52 13(25) 39(75) 2.9 ± 2.9
Difference in death probability between two experiments was insignificant (Fisher’s exact prob-
ability test, P = 0.44) and that in mean length of the time until a death was also insignificant
(t test, P = 0.93)
females can produce only males among which there is at least 1/2 relatedness (of
course, if the mother was also produced by inbreeding, then the relatedness between
them is much higher). Thus, kin males were obtained from unfertilized females.
Non-kin males were taken from different stock cultures, which had been established
from at least 500 individuals collected in the field (Ashizuri population). The
relatedness between the non-kin males could not be strictly estimated, but it was
considered to be much lower than that between kin males. Then two kin males or
two non-kin males were paired and released into a woven nest containing several
eggs. During the 5-day observation period, there was no difference in male mortality
rates between kin pairs and non-kin pairs (Table 6.2). In both cases, males fought
strongly with and killed cohabiting males. Although St. miscanthi males do not seem
to exhibit any kin-discriminative behavior, whether they have kin recognition ability
is still unknown because we did not conduct any experiments to prove it.
6.2 Variation in Male Aggressiveness Within a Species 113
As such, I learned that St. miscanthi males show very high aggressiveness
against conspecific males regardless of kinship. At first I thought that this aggres-
sion was a kind of shift from “counterattack against predators” (Sect. 3.5) to attack
against “conspecific males.” In other words, the aggressive traits males exhibit
against conspecific males are the same that they exhibit against predators. However,
this kind of impression must be rejected because St. longus males also show very
strong counterattack against predators yet show no aggressive behavior against
conspecific males (see Fig. 6.1). Therefore, we have to separate the male aggres-
sion against predators from that against conspecific males. So, what are the possible
reasons for the variation in male aggressiveness seen between St. longus and
St. miscanthi? Saito (1990c,d) could only suggest that such a difference in male
aggression may be related to a difference in male relatedness at the population level
between these two species, because St. longus is mainly distributed in cooler
regions and St. miscanthi in warmer regions in Japan, which may cause lower dis-
persal (or mixed colonies) in the former and higher in the latter.
After completing the foregoing studies, I had an unlikely idea: If there is such
variation in male aggressiveness between closely related species, then could male
aggressiveness vary within species? Fortunately, St. miscanthi has a very wide
distribution range, from the northernmost point of Honshu Isl. (lat. 26°6¢) to Ryukyu
Isl. (lat. 41°30¢), because its host plant, Chinese silvergrass (Miscanthus sinensis),
occurs throughout Japan (and in many other Asian countries, as well). Thus I started
to consider whether there is variation in male pugnacity between populations of a
single species (Saito 1995a). St. miscanthi was collected from various regions around
Japan (Box 6.1), and many strains were established in our laboratory. As described,
it is very easy to evaluate paired male mortality in a nest, such that I regarded the
death rate of one of the paired males as “male relative aggressiveness” and evaluated
each population by the same-garden analysis (populations originated from various
regions are tested under the same experimental conditions). Very large variation in
male aggressiveness among populations was detected; e.g., one of the paired males
was killed immediately in almost all nests in the Ashizuri population, but no males
were killed over 5 days in the Mashiko population (Saito 1995a).
What kind of environmental factors were responsible for this variation was the
next theme. To search for factors, we need a working hypothesis. One plausible
factor was differences in the relatedness of interacting males among populations.
Hamilton (1979) observed interspecies variation in male aggressiveness between
species whose wingless males were expected to have different kinship from the
difference in female oviposition pattern into figs. He explained the variation by
the following three hypotheses: The reasons why there is low aggression between
close kin males in a fig must be (1) many of the rival males are brothers, and a male
doesn’t care so much whether he or his brother does the mating; (2) many of the
114 6 Kin Selection
Box 6.1
Localities where Stigmaeopsis (S.) miscanthi were collected in Japan. Circles
indicate where populations in Japan were observed by Saito (1995a). Hatched
circles show populations whose males were demonstrated to be HG; open
circles were LW. Squares indicate points where S. miscanthi was confirmed
to exist but where aggressiveness has not yet been observed. Filled circles
and squares (all must be LW) indicate heavy snowfall regions. (from Saito
and Sahara 1999) HG and LW, see Sect. 6.4.
females are his sisters, and he doesn’t wish to risk that some sisters remain
unmated; and (3) for those that are his sisters, a male may actually prefer to have
them mated by a unrelated male because of the opportunities this will give for use-
ful recombination in the next generation. These points are all from his kin selection
(inclusive fitness) theory. On the other hand, West et al. (2001) showed that there
is no correlation between male aggressiveness and relatedness between interacting
males per fig and the aggressiveness is rather correlated with number of females per
male. They provided the “resource (mates) competition hypothesis” to explain the
variation in male aggressiveness of fig wasps. We initially attempted to analyze the
variation in accordance with Hamilton’s (1979) kin selection theory.
6.2 Variation in Male Aggressiveness Within a Species 115
From the kin selection hypothesis, the variation in male aggressiveness must be
correlated with the relatedness of interacting males. If so, what factors can change
the relatedness of males between populations? As stated before, I knew that the
St. longus males distributed in cooler regions have no aggressiveness and those of
St. miscanthi in hotter regions show high aggression, such that one of the candidate
factors was temperature. I then attempted to analyze the correlation between the
male relative aggressiveness and temperature of each sampling location. However,
there are several kinds of temperature values, i.e., annual mean temperature, mean
maximum daily temperature, mean minimum daily temperature, and so on. Among
these, I selected mean temperature for summer (June to September) and the mean
lowest daily temperature for winter (December to February), because mite life his-
tory, especially the number of generations (it must reflect the intensity of mixing
by dispersal among populations), may mostly depend upon the former and male
overwintering probability may depend upon the latter. The reason why I focused on
the mean lowest daily temperature for winter should be explained further: St. mis-
canthi overwinters at the adult female stage (see Table 3.3) and the overwintering
females enter diapause (tolerant to winter hardiness) in autumn (Saito et al. 2002).
On the other hand, it has been pointed out that spider mite males have no such
adaptations (for Tt. urticae; Veerman 1985). Although there are few adult males in
spring (overwintered) populations of St. miscanthi in cooler regions (there are
no males in overwintered St. longus populations in Sapporo; Saito 1987), some
non-diapausing males sometimes overwinter in warmer regions in Japan (Saito
1995a). This finding must mean that the lowest temperature in warmer regions is
sufficiently high to permit non-diapausing males to survive during winter (how
many males can overwinter varies year to year depending upon winter severity).
Male existence in spring may greatly influence the kin structure of this mite spe-
cies, if unfertilized (unmated) females overwinter (Saito 1995a; Saito and Sahara
1999). As shown in Figs. 4.12 and 6.4, the foundation of spring nests is mostly
carried out by a small number of females, although sometimes many females will
build a nest. Therefore, if single unfertilized females found nests, they often mate
with their sons (mother–son mating, as described in Sect. 4.7). Because such strong
inbreeding has no short-term fatal effects for this mite species (Chap. 5), the relatedness
between offspring produced by mother–son mating must increase in such nests (cf.
Box 4.11). On the other hand, if males can overwinter, they must fertilize overwin-
tering females in the winter aggregations (St. miscanthi and St. longus females
usually form great masses in certain nests in winter; Saito, unpublished data). Thus,
the probability of male overwintering (or survival during winter) is thought to affect
the probability of inbreeding (mostly through mother–son mating; Saito 1987) in
spring nests (Fig. 6.4). Therefore, winter minimum temperature, which may deter-
mine overwintering male survival, would reflect the relatedness among nest mem-
bers in nests.
Figure 6.5 shows that the relative male aggressiveness is significantly correlated
with the minimum temperature of the regions where mite populations were col-
lected, although the mean temperature in summer did not show such a clear trend.
If the minimum temperature in winter indirectly affects the relatedness of interacting
35
30
25
15
10
0
1 2 3 4 5 6 7 8 <
Numbers of foundresses per nest
Fig. 6.4 Numbers of foundresses observed in spring (middle May) nests of St. miscanthi. Data
from six populations from north to south Kyushu Island. Only nests with adults, eggs, and larvae
were included, because only these could be regarded as newly founded nests
%
100
r = 0.922
Relative male aggressiveness
P < 0.001
50
-4 -2 0 2 4 6
Average of daily minimum temperature in winter (AMT, Dec.-Feb.)
Fig. 6.5 Correlation between relative male aggressiveness and winter coldness. Relative male
aggressiveness was estimated as the relative death rate of one of the paired males introduced
into a nest for 5 days (after Saito 1995a). Each experiment included more than 25 replicates and
controls
6.2 Variation in Male Aggressiveness Within a Species 117
males, this correlation must suggest that male aggressiveness evolved by kin
selection (Saito 1995a).
Next, let us verify whether the other hypothesis, “resource competition,” is
applicable to the present case. After publishing (Saito 1995a) the kin selection
hypothesis, I met with several counter-arguments. Among these, the most popular
opposition was as follows: a higher male mortality for whatever reason should bias
the sex ratio. If so, the value of females as resources to males would decrease, and
this might result in selection against male aggression. In other words, if females are
easy to get, there is no need to fight for males, i.e., “resource competition” (West
et al. 2001; see earlier). Then, I examined (Saito 2000; Saito et al. 2000a) whether
this hypothesis holds in the case of St. miscanthi.
For the convenience of discussion, I divided the life cycle of St. miscanthi into
two seasons: winter to early spring (the overwintering season) when they must
overcome adversity, and summer when mites actively reproduce and increase their
numbers. From winter to spring, male overwintering probability is low in cooler
regions and high in warmer ones (Saito 1995a); this inevitably causes a difference
in male abundance in winter to early spring. Under the kin selection hypothesis, this
difference causes a difference in the relatedness of nest mates in the spring and
summer seasons, whereas under the resource competition hypothesis the difference
in male abundance directly alters the aggression tactics of males. As stated previ-
ously, male overwintering probability is lower in cooler regions than in warmer
regions. With males abundant in warmer regions, we could expect higher aggres-
sion there in accordance with the resource competition hypothesis.
However, other difficulties associated with the resource competition hypothesis
become apparent if we focus on the detailed population structure during the over-
wintering season. Even in relatively high aggression populations (such as the
Tsuyazaki and Keya districts), overwintered males are much more scarce than
females: males made up only 1.3% and 1.7%. respectively, of all adults observed in
March (Saito 1995a). Therefore, I find it difficult to believe that such a small num-
ber of males, compared to females (resources), would display mortal aggression
toward each other when competing for mating opportunities in this season. In other
words, I believe that there are very few situations under which competition between
males occurs during the overwintering season (including early spring) in St. mis-
canthi. On the other hand, as a rule of population genetics, a small number of
immigrant males can greatly affect the genetic structure of a gene pool (Ims and
Yoccoz 1997). Thus, male overwintering probability (male immigration for space
is analogous to male survival over time in this situation) greatly affects the genetic
structure, i.e., the relatedness of the colonies of this species (see also Sect. 6.4).
In the summer season, we studied the sex ratio at maturity of six populations of
St. miscanthi (two of the “low-aggression” groups and four of the “high-aggression”
groups mentioned hereafter) under experimental conditions and found very similar
sex ratios among them; i.e., approximately 15% males (sex ratio of offspring
produced by isolated female; Sato et al. 2000). This finding means that there is no
difference in the basic sex ratio trait between populations. Moreover, Saito (2000)
determined that there is no significant correlation between (experimentally evalu-
118 6 Kin Selection
ated) male aggression and the number of adult females per nest for 32 field popula-
tions. Rather, male aggression was negatively correlated with the number of males
per nest in the fields. To avoid any confusion between the cause and consequence,
namely, that mortal aggression of males decreases the number of males or vice versa,
I investigated the relationship between a potential female sex ratio = (third quiescent
females)/(third quiescent males + third quiescent females) and male aggression
among populations (“third quiescent” is the last quiescent stage before maturation).
The results showed that there is no significant correlation between them (Kendall’s
tau = 0.09, P > 0.96), indicating that the abundance of available (potential) females
for potential males is not related to the variation in male aggression even in the sum-
mer season. Therefore, I could conclude that there is no strong evidence supporting
the “resource competition hypothesis” (Saito 2000; Saito and Mori 2005).
If so, then the reader may wonder why St. miscanthi males do not show kin
discriminative behavior (as shown in Sect. 6.1; Saito 1994a). If they had this ability,
then males would behave amicably toward their kin males and aggressively against
strangers. I have no clear idea of how to definitively answer this question, but I can
suggest that it is related to the phylopatric nature of this organism; i.e., they do not
encounter males having very different relatedness within their nests frequently
enough for such an ability to evolve (Saito 1994a), such that they do not use such
precise information.
There is also the question of whether the variation in aggression seen in
St. miscanthi is affected by the variation of predation pressure (e.g., S. Aoki, per-
sonal communication; Saito 1995a). If there were big differences in predation
pressure between local populations of St. miscanthi, then males should decrease
their male-to-male competition, because they should defend their offspring in a
cooperative way. In other words, if it has killed all the conspecific males within its
nest, the fitness of the victorious male can completely disappear because it becomes
unable to effectively defend its nest, such that the intensity of predation is related to
the variation. There are actually several co-occurring predator species in the habitats
of St. longus and St. miscanthi (Saito 1990b; Chittenden and Saito 2001; Mori and
Saito 2004), and although these predator species are hypothesized to be one of the
driving forces behind speciation in Stigmaeopsis (Sect. 3.5, Mori and Saito 2004),
we have no strong evidence that there is a large difference in predator fauna among
populations. However, if such an effect is important, there is no doubt that kin selec-
tion, i.e., relatedness, is also responsible for male cooperation, because high
relatedness always makes it easy for interacting males to cooperate (see Sect. 6.6).
As such, the variation in male aggressiveness of St. miscanthi was explained by kin
selection. However, we later noticed that there were two distinct trends in Fig. 6.5;
most plots of the more aggressive populations were positioned on the upper side of
the regression line and those of less aggressive populations on the lower side of the
regression line. If I had erroneously observed two groups having different male
aggressiveness, all the foregoing scenarios would be meaningless. Therefore, I was
forced to study many more populations from various districts.
With the steady accumulation of data, it became apparent that there are two
groups showing different levels of male aggression in Japan. As seen in Fig. 6.6, at
an approximately intermediate value of relative male aggressiveness, there are two
distinct groups, both of which show clinal variation (hereafter the low aggression
group is called the LW form and the high aggression group is the HG form).
Furthermore, with two parameters, i.e., relative aggressiveness and winter mini-
mum temperature, a cluster analysis (Ward method) revealed that there are two
forms separated by a long distance (Saito and Sahara 1999). In relation to this,
120 6 Kin Selection
Fig. 6.6 Relative male aggressiveness (y) and average daily minimum temperature in winter (x)
in populations from low-snowfall regions. Open circles are HG (high aggression) form; filled
circles (grey) are LW (low aggression) form (after Saito and Sahara 1999)
coasts of Honshu Isl. (facing the Pacific Ocean, Box 6.1). Since Kojima et al.
(1970) reported that the temperature under snow varies according to depth and
approaches 0°C at ground level, it was expected that the temperature affecting male
overwintering survival is not the ambient temperature (i.e., from meteorological
data), but the under-snow temperature in the regions of heavy snowfall. Therefore,
it was expected that the relationship between male aggressiveness and average
minimum winter (air) temperature varies between areas of light and heavy snow-
fall. Under this supposition, Saito and Sahara (1999) compared the relationship
between male aggression and winter minimum temperature in two categories, light
snowfall and heavy snowfall regions. As a result, male relative aggressiveness
really depended upon the coldness of light snowfall regions (lower cline in Fig. 6.6),
but there was no correlation between aggressiveness and temperature in heavy
snowfall regions (Fig. 6.7). Therefore, the correlation between male aggressiveness
and male overwintering probability must change according to the insulating effect
of snow cover. The present comparison between heavy and light snowfall regions
may demonstrate that male overwintering probability is not affected by the ambient
temperature conditions of localities but by the temperature of the plant leaf surface
where males are living. Simultaneously, this connection means that “male survival
in winter” is a key to explaining male aggressiveness. Although it remains indirect
evidence, it does lend support to the hypothesis proposed by Saito (1995a) that
male aggressiveness (=intensity of intrasexual selection) evolved from a change in
relatedness among interactors (=advantage of male cooperation).
Fig. 6.7 Relationship between relative male aggressiveness and average daily minimum tem-
perature in winter in heavy snowfall regions (after Saito and Sahara 1999)
122 6 Kin Selection
The following may be the final question concerning the variation in male aggressive-
ness, but it has not been sufficiently resolved. Saito et al. (2002) observed the diapause
attributes of the two forms (HG and LW) of St. miscanthi. We started the study under
the working hypothesis that female diapause attributes also affect the relatedness
between nest members in spring. For example, we have sometimes observed all stages
of St. miscanthi (HG) on Chinese silvergrass in low-altitude southern regions even
during the coldest season (Fig. 6.8 left). This observation suggests that some HG
populations do not enter diapause but continue to reproduce during winter. Under
such conditions, strong inbreeding (mother–son mating), which is thought to be the
main factor increasing relatedness, must not occur. On the other hand, there are
only females in winter in northern and/or high-altitude populations (Fig. 6.8 right),
suggesting they enter deep diapause. Although the reason is still unknown, if such
a difference in female diapause is a qualitative one, i.e., there are diapause and
non-diapause populations, we should be able to explain the existence of the two
groups showing different clinal trends. This idea came from several previous dis-
coveries about diapausing and non-diapausing populations of Tt. urticae and Pn. citri
Fig. 6.8 Stage structure of six populations of St. miscanthi in midwinter. Numerals in parentheses
are the number of individuals (left) and the number of nests (right) observed. HG, high aggression
form, LW, low aggression form (from Saito et al. 2002)
6.4 Two Groups Showing Different Levels of Male Aggressiveness 123
(Shinkaji 1979; Gotoh and Shinkaji 1981; Osakabe and Saito 1991). At the present time,
Pn. citri is divided into two species, Pn. citri (non-diapause) and Panonychus mori (dia-
pause) by morphological differences (Osakabe and Sakagami 1994; Ehara and Gotoh
1992). The status of Tt. urticae is still disputed (Dosse 1967; Takafuji et al. 1991).
We then conducted an experimental study on the diapause attributes of the LW and
HG forms (Saito et al. 2002). Two populations of each form were selected, and the
diapause induction and termination patterns of females were then observed. Generally,
spider mite diapause is induced under short-daylength and low-temperature conditions
(Veerman 1985). First, we reared eggs of each population to adulthood under these
conditions. After female maturity and mating, they were kept under different
daylength conditions and their oviposition was observed. Figure 6.9 shows that the
females of all populations developing under short-day and low-temperature condi-
tions significantly decreased their reproductivity after moving to long-daylength con-
ditions, indicating they enter diapause (Fig. 6.9b,d). Furthermore, there was a big
difference in diapause elimination patterns between the LW and HG populations
(Fig. 6.9), suggesting that there is a distinct difference in female diapause attributes.
However, data from only two populations of each form cannot provide sufficient
information to draw conclusions. We next compared the stage structure of three
populations of each form in the field (Fig. 6.8) and the winter temperature data of
Cumulative percentage of females beginning oviposition
a b
c d
Fig. 6.9 Oviposition (ovi.) patterns at 18°C of females developing (dev.) under 18°C and under
long-daylength and short-daylength conditions. a, development under 15L-9D / oviposition under
15L-9D; b, dev. 9L-15D / ovi. 9L-15D; c, dev. 15L-9D / ovi. 9L15D; d, dev. 9L-15D / ovi. 15L-9D
Ashizuri and Tobuko are St. miscanthi HG form and Kubokawa and Unzen are LW form (from
Saito et al. 2002)
124 6 Kin Selection
Fig. 6.10 Average winter (December–February) temperatures of six regions where St. miscanthi
populations were observed. Black circles, average daily mean temperatures from December to
February; white bars, 3-month (Dec.–Feb.) average daily minimum temperatures; black bars,
monthly minimum temperatures for 10 years (calculated from AMEDAS, data base provided by
the Japan Meteorogical Agency during 1985–1995) (from Saito et al. 2002)
their respective localities (Fig. 6.10). Many eggs and developmental stages were
found in the Ashizuri, Awa, and Tobuko populations (all HG), whereas the
Kubokawa, Odochi, and Unzen populations (all LW) consisted mostly of adult
females and a small number of eggs (Fig. 6.8). There were significant differences
in stage structure among all populations. On the other hand, the three kinds of tem-
perature data in the six localities where St. miscanthi were collected change more
or less in a gradual manner (Fig. 6.10), such that it is difficult to understand why
the stage structures of the HG form (Ashizuri, Tobuko, and Awa populations) are
very different from those of the LW form (Unzen, Kubokawa, and Odochi popula-
tions). In addition, the stage structure of the Awa population is very different from
that of Odochi, although the temperatures are not so different. This finding means
that winter climate does not fully explain the variation in St. miscanthi stage struc-
ture in winter and also suggests that there may be some qualitative difference in
diapause intensity and/or in physiological response (e.g., the threshold temperature
for oviposition) between them (we have continued this study for many LW and HG
populations, and this difference has become increasingly evident; Saito et al.,
unpublished data). Thus, the working hypothesis that the difference in female dia-
pause attributes causes disruptive change in the mean relatedness of nest members
became increasingly probable.
6.5 Sexual Selection 125
An overview of sexual selection in mites has already been presented in Sect. 4.2.
In this section, I digress a little to describe a case study on this problem in
St. miscanthi. I address whether the variation in male antagonistic behavior has
an influence on male weaponry through intrasexual selection. Sexual selection
theory predicts that male body size and/or weaponry must evolve, as known in
horned beetles, deer, and elephant seals, if males fight strongly to acquire mates
(Wilson 1975; Alcock 1979). As already mentioned, Saito (1990d) observed that
leg length is one important factor determining the winner of male-to-male combat
in a population of this species. As there is variation in male aggressiveness, it
is expected that there is some variation in leg length between populations of
St. miscanthi, especially between LW and HG forms. Saito (1995a) measured
male leg length to address this prediction. Leg lengths vary between populations
for several factors, such as host plant conditions (Saito 1990d), but the leg length
ratio (I/III) does not change (Nakano, unpublished data). The softness of spider
mite bodies makes it impossible to measure their body sizes. Thus, I adopted the
ratio of leg-I length / leg-III length (the relative length of leg-I), because behavioral
observations indicated that the male leg-III is only used for supporting the body
and not for aggressive behavior (although whether leg-III is a true reflection of
body size is unknown).
Figure 6.11 shows the correlation between male relative aggressiveness and
relative length of male leg-I. There is a significant difference in the relative
length of leg-I between HG and LW, with the former being more exaggerated.
However, I could not observe any clinal trend in leg length along with
aggressiveness within either form. As stated previously, there are significant
correlations between the aggressiveness and the winter minimum temperature
within both LW and HG forms, such that this result is not of constant concern to
me. This is one of the important questions left for future study. Anyhow, the
difference in the relative leg-I length between LW and HG has become a good
criterion to morphologically discriminate these forms in successive studies (e.g.,
Sato et al. 2008).
126 6 Kin Selection
0.16
0.15
Relative length of leg I (log)
0.14
0.13 HG
LW
0.12
0.11
0.10
0.09
0 10 20 30 40 50 60 70 80 90
Relative measure of male aggressiveness (arcsin÷)
Fig. 6.11 Relative length of male leg I was not correlated with the intensity of male aggression
(after Saito 1995a)
Earlier I have described cooperation and aggression in males of St. miscanthi and
St. longus. From the aspect of animal behavior, we have learned that there are two
extremes, namely, aggression and cooperation in these two closely related species.
Although aggression and cooperation, as fundamental themes in sociobiology, have
always attracted a great deal of attention and research (Wilson 1975; Alcock 1979;
Maynard Smith 1982a; Trivers 1985; Krebs and Davis 1987; Dugatkin 1997), they
have always been treated as “separate” phenomena. For example, Dugatkin (1997)
discussed animal cooperation through the game theory approach, but the strategies
he hypothesized are a dichotomy of “neutral” and “cooperation.” The famous
hawk–dove game provided by Maynard Smith (1982a) and Hines and Maynard
Smith (1979) also focused on two strategies, i.e., “aggression = hawk” and
“neutral = dove.” There are several reasons why these two behaviors have been
treated separately. One must come from the fact that the prime movers (selection
pressures) that evolved these two behaviors are considered to be different: aggression
will evolve through egoism (individual selection), and cooperation through kin
selection and/or group (or multilevel in the modern day) selection (Hamilton 1964;
Maynard Smith 1982a; Dugatkin 1997; Lehman and Keller 2006). Furthermore,
doubts remain over how an increase in inclusive fitness through cooperation
6.6 Theoretical Explanation of Male Aggression and Cooperation 127
Saito and Mori (2005) reviewed several hypotheses about variation in male
aggressiveness and discussed how the diverse nature of aggression and cooperation
can be understood if we focus on where aggression reaches a compromise with nonag-
gression and/or cooperation in response to the relatedness of the interactors. To
further clarify this discussion, I regard aggression and cooperation as a “continuous
trait” in group-living animals whenever possible. So, by considering relatedness,
just how aggression reaches a compromise is addressed here. In the beginning, the
outline of a famous and inclusive theory of aggression and/or cooperation must be
redescribed, although I have previously introduced its essence (Sect. 6.2).
Hamilton (1979) stated, with respect to the variation in aggression of wingless fig
wasp males, “a difference in mean relatedness between rivals accounts for the different
male behaviours.” Although he did not say that this is the sole factor accounting for the
difference in behaviors (see Sect. 6.2), the above statement is sometimes believed to
correspond to Hamilton’s rule of altruism (Michod and Hamilton 1980), i.e., rB − C > 0,
where r is Wright’s coefficient of relatedness to the recipient, B (³0 per the primary
definition) is the benefit associated with the trait for which the gene codes, and C (³0)
is the donor’s cost, which accrues from the decrease in mating opportunities. This rule
was then applied to explain the case of the male fig wasps (Trivers 1985; Frank 1985;
West et al. 2001). However, as mentioned before, Saito (2000) doubted the rationale that
high relatedness between males must decrease male pugnacity. If there is even a small
difference in mating success between males through aggression, sexual selection will
favor more aggressive males; i.e., aggression might evolve in males of every population
regardless of their relatedness when r < 1.
To ascertain the foregoing prediction, let us see whether “rB − C” tends to
increase above 0 as r increases in male-to-male competition. The cost (C ³ 0) accrues
from the withdrawal of an actor male (donor) from the competition for females and
the benefit accrues from the increased mating chances for the other males (recipi-
ents) resulting from the actor’s withdrawal. Assuming that females can be insemi-
nated at all, let us imagine that there is a single male and n females in an arena and
mating occurs between them. If x males join the arena, then the first (actor) male
should compete with them. Are there any conditions under which the first actor male
raises its inclusive fitness by sharing females? When the actor male decreases its
aggression level and loses mating chances for d (=C), then the other males x will get
a surplus benefit, d/x, each (assuming that the males joining the arena have primarily
no chance of getting partners). Because B = x(d/x) = d, the inequality rB − C > 0 is
d(r − 1) > 0. Because d > 0, r > 1 is always required. In other words, the cost d paid by
128 6 Kin Selection
the first actor male as decreased mating opportunities is “intact” and is evenly
divided among the other males as their benefits, so that there is “no way” of increas-
ing the inclusive fitness of the first actor male under r £ 1. From the beginning,
therefore, it is unlikely that Hamilton’s rule can ever be applied to the varying com-
petition in mating between males, if both the competition and mating opportunity
are not repeated (i.e., once in a lifetime; Maynard Smith 1982a).
Of course, the conditions under which fig wasp males compete are different
from the foregoing example because they inevitably join multimale competition
within a fig fruit. It is thus suggestive that the level of fighting between fig wasp
males shows no correlation with the estimated relatedness of interacting males but
is negatively correlated with mating opportunities (West et al. 2001). If we still
intend to apply Hamilton’s rule to male-to-male competition, we have to search for
plausible conditions where any decrease in an actor’s mating opportunity can
significantly increase its kin’s mating chances (i.e., B >> C). I know that the above
calculation is so simple that most people might consider it a self-evident conclu-
sion. However, it is undeniable that some researchers believe that Hamilton’s rule
(note that this rule is only a part of his kin selection theory) is applicable to such a
case. This problem is thought to be caused by an erroneous application of the
Hamilton’s rule: his rule is a comparison of fitness between solitary-living and
group-living (interacting) individuals and shows the conditions under which the
latter individual gets more inclusive fitness than the former one. In other words, it
is not the rule to explain behavioral traits (aggression, cooperation, and so on) of
individuals when they are always situated in game conditions (interaction is
inevitable).
Several other hypotheses have also been put forward to explain variation in male
aggression. One is the increase in lifetime mating success in nonfighting males
(Enquist and Leimar 1990) or the high risk of disability to the winner male
(Maynard Smith 1976; Tomkins et al. 2004; addressed later). Another is that if
males are confronted with another different selection pressure (e.g., the synergistic
effect of multiple males in Maynard Smith 1982b), then the selection pressure will
also change male aggressiveness.
For the moment, however, let us examine whether such a different selection
pressure can improve Hamilton’s rule for explaining male aggression (i.e., by com-
parison between individual fitness and inclusive fitness). Let us recall St. longus
and St. miscanthi. The males of both species are characterized by their counterat-
tack behavior against nest-intruding predators (Saito 1986a, 1990e; Sect. 4.4). If
there are multiple males in a nest, they may lose mating opportunities even if they
effectively defend their offspring and/or mates. Therefore, males must inevitably
adopt either of two traits for two different selections, namely, (1) cooperation in
offspring defense against predators and (2) aggression to obtain mates. As described
in Sect. 6.2, the variation in the aggressiveness of S. miscanthi males cannot be
explained by the resource (=mate) competition hypothesis proven in the fig wasp
case by West et al. (2001) but is thought to be closely related to the effects of the
high risk of disability to the winner male and of cooperative defense by males. We
first tried to formalize factor (1) as follows.
6.6 Theoretical Explanation of Male Aggression and Cooperation 129
If a single male mates with females and defends his (female) partners in a nest,
he will have f offspring over his lifetime. Thus, f is considered the net income
(fitness) of a solitary male after his reproductive and defensive behavior. Cooperation
of x males increases the survival of their offspring “1 + s” times (s ³ 0). The reason
why I express the fitness gain through cooperation as 1 + s is to maintain consistency
with the game model mentioned later. If the cooperators are his kin, he increases his
inclusive fitness through the cooperators’ offspring at the rate of r (relatedness). If x
males live altogether, mating opportunities decrease and thus the number of
offspring decreases because of the competition. Any decrease in mating opportunity
for a male can be converted into the number of offspring “produced” by him, such that
the cost is regarded as the decrease in the male’s offspring to apply the same term as
the benefit of cooperation. The fitness of a male belonging to an x-male group is
(1 + s ) f (1 + s ) f ( x − 1)
+r .
x x
Then we can search for the condition under which cooperation will evolve, i.e.,
“cooperating male’s inclusive fitness” > “solitary male’s fitness.” Thus, we can
obtain the following inequality:
(1 + s ) f + rf (1 + s )( x − 1)
> f.
x
Because f > 0, it can be reduced to
(1 + s)(1 – r) + x(r + rs – 1) > 0.
If only two individuals interact (x = 2), then the above inequality is simply expressed
as (1 + s)(1 + r) > 2. This inequality can be further simplified as r > (1 – s)/(1 + s)
because 1 + s > 0. The last formula suggests that if s ³ 1, cooperation will evolve
regardless of the r-value, meaning that there is a condition under which cooperation
occurs unless the interacting individuals are non-kin (e.g., Fowler and Gobbi 1988;
Uetz and Hieber 1997; Matsuura et al. 2002). Furthermore, r > (1 – s)/(1 + s) clearly
indicates that male relatedness is also important; namely, r is the primary determinant
of the evolution of cooperation, if 0 < s < 1. Therefore, the effect of cooperation (s) and
relatedness (r) must be essential to understand the cooperation, and here we once
again encounter Hamilton’s inclusive fitness concept (Hamilton 1964).
Next, let us see whether r > (1 – s)/(1 + s) is applicable to the spider mite case.
The parameter s » 1 (a twofold defensive success) evaluated experimentally in the
case of St. longus (see Fig. 3.22) is now meaningful in accordance with this
inequality. If s > 1, there is no option for the male mites other than cooperation
regardless of r, while Saito (1990c, 1997, 2000) stressed the importance of higher r.
Although the relatedness (r) between males may vary in these mite species, males
have low dispersal trends, such that there is a high probability of interaction
between close relatives. Therefore, both higher s and r can now explain why we
only observed cooperative males in this species.
130 6 Kin Selection
The foregoing explanation (Saito and Mori 2005) showed that aggression would
be replaced by cooperation in relationship to the relatedness of interacting individu-
als at r > (1 – s)/(1 + s). However, it does not show whether there are conditions
under which aggression (or cooperation) gradually change (as seen in Fig. 6.5) with
relatedness, the effect of cooperation, or other factors. Furthermore, the formula
shows only the threshold at which “solitary living” changes to cooperation. In other
words, it is still insufficient to resolve the question of whether the clinal change in
male aggression observed by Saito (1995a) and Saito and Sahara (1999) can be
logically supported.
C A
(1 + s ) f
C 0
2
(1 − k ) f
A f .
2
When a C strategist interacts with another C strategist, the net fitness is an addi-
tive benefit, and s (s ³ 0, hereafter called the helping effect) is shared equally
between the two C strategists who defend their families; i.e., the payoff is repre-
sented by [(1 + s)f ]/2. We have defined this situation as “cooperative sharing.” If
two males interact together in a patch (nest) with a constant number of females
(mating partners), as known in St. miscanthi, one of the behavioral traits of such
males is to equally share one-half of the partners (1/2). In cases where males pos-
sess the ability to defend and protect their patches (including their partners and
offspring) from predators, then s > 0, as shown in Sect. 3.5. This helping effect (s)
can be measured practically by comparing the success of defending offspring from
predators by two C strategists with that of a single C strategist (an example is shown in
Appendix 3).
If a C strategist interacts with an A strategist, it always loses its direct fitness and
receives nothing (payoff = 0; i.e., no offspring). Although this situation may seem
quite severe for the C strategist, the mortal male-to-male aggression in
St. miscanthi (Fig. 6.1; Saito 1995a) indicates its reality.
If an A strategist interacts with a C strategist, it receives a payoff, f, because it
always wins and can monopolize all females.
When an A strategist confronts another A strategist, they always fight. The
winning combatant will receive a payoff of [(1 – k)f]/2, where k (0 ≤ k ≤ 1) is
the winner’s risk of injury through aggression. Thus, we observe strong “aggres-
sive interactions” between A strategists. The variable k is simply the cost of
aggression that can be measured as the probability of a winner’s death or injury
in an aggressive confrontation between A strategists (an example is shown in
Appendix 3).
In the above basic payoff matrix, by introducing the relatedness of interactors, r, we
obtained the following “inclusive fitness matrix” (Hines and Maynard Smith 1979):
C A
(1 + s )(1 + r ) f
C rf
2
(1 − k )(1 + r ) f
A f
2
From the inclusive fitness matrix, we could identify a possible evolutionary stable
strategy (ESS), q* (the frequency of A strategy, Maynard Smith 1982a), as follows
(we can also calculate q* directly from the basic matrix; Saito and Takada 2009):
132 6 Kin Selection
The candidate ESS frequency, q*, is obtained when E [A, (mixed strategy, q)] = E [C,
(mixed strategy, q)] (Maynard Smith 1982a). We could then get q* = [1 – r – s – rs]/
[(1 + r)(k – s)]. After several additional calculations, we confirmed that this formula
showed a mixed ESS under 1 − s > 0 and k − s > 0 conditions (Grafen 1979).
Next, let us see how the ESS (frequency of A strategy, q* = [(1 – r) – s(1 + r)]/
[(1 + r)(k – s)]) varies with the three parameters, r, k, and s, by simulation. If a risk
of aggression, k = 0.35 (measured value in Appendix 3) exists, then three ESS types,
namely, pure A strategy, mixed strategy, and pure C strategy, can each become ESSs
depending upon the r-values at every s-value satisfying k − s > 0 (Fig. 6.12a). The
area of mixed strategy gradually changes with r and s values similar to the skirts of
mountain having an “aggression” peak. Thus, we have herein learned that male
Fig. 6.12 Calculated evolutionary stable strategy (ESS) (frequency q* of the A strategy) surface
with various values of r and s at k = 0.35 (a) and ESS surface with various values of r and s at
extremely high k = 0.998 (b). Furthermore, q* gradually changes with k value even when s = 0 (c)
6.6 Theoretical Explanation of Male Aggression and Cooperation 133
which competition and cooperation occur. Thus, the effective relatedness between
individuals should be determined from two kinds of relatedness, i.e., when cooperat-
ing and when competing. West et al. (2002) concluded that the reason why there is
no relationship between relatedness and male aggressiveness in fig wasp males is
that the effect of the relatedness when cooperating is exactly the same as the cost of
competing (i.e., the former is completely offset by the latter).
However, we here proposed that there is a condition where aggression compro-
mises cooperation at the same spatial scale through “classic” relatedness. Thus, the
effect of individual selection is not always completely offset by kin selection, if we
focus on the selection intensities, i.e., by introducing practical parameters, such as
k and s. Because there is no rational explanation that the magnitudes of kin and
individual selections are equal even at the same spatial level, thus where egoism
compromises with cooperation on r must be rather an important question in socio-
biology. In other words, there is no need to idealize “effective relatedness” (West
et al. 2002) other than classic r, if we can appropriately determine the magnitudes
of kin selection and individual selection. The case of the polyembryonic encyrtids
reported by Giron et al. (2004) in which soldiers’ aggressiveness between lineages
is correlated with r may be another example of the present scenario. In these cases,
there is no difference in spatial scale between competition and cooperation, but the
levels of aggression and cooperation should vary with the relatedness between
individuals.
In any case, I believe that the Saito–Takada model is applicable not only to male-
to-male interactions but to many kinds of aggression and cooperation in nonre-
peated games as well. The conditions required in this model are that there are two
different kinds of selection against group-living animals, such as intrasexual selec-
tion or resource competition for egoism, and kin selection(s) beneficial to coopera-
tion. These situations frequently occur in animals, such as male-territorial birds and
fishes, and the present model may explain why these males do not always fight
mortally (Van Hoff 1990). Thus, if we change our focus to multiple selection pres-
sures that operate on both individual lifetime fitness and inclusive fitness, then we
may find many examples explainable from the present model in nature.
In the previous section, we could show how r determines male behavior in St.
miscanthi and St. longus. Regrettably, we have still no practical data on
relatedness from molecular-based methods between interacting males because of
the homogeneous nature of St. miscanthi genes (we have found several
microsatellite loci, but they do not show sufficient variation). Therefore, it is
difficult to judge whether the model’s prediction is correctly realized in nature.
However, if we consider the opposite perspective, we can evaluate relatedness
from the Saito–Takada model by substituting the practical parameters that can
be obtained experimentally.
6.6 Theoretical Explanation of Male Aggression and Cooperation 135
The male aggressiveness of St. miscanthi was evaluated as the probability of the
death (by combat) of one of the two males paired in a woven nest (Sect. 6.1; see
Figs. 6.5, 6.6). So as to fit the present model, I translated this as the frequency of A
strategy, q, by the procedure shown in Appendix 3. The cost of aggression, k, was
estimated from the previous experiments in which two males killed each other
simultaneously (k = 0.35 for HG form; see Appendix 3). The net effect of coopera-
tion, s = 0.25, was also evaluated as the increase of counterattack success (»survival
rate of offspring; Fig. 3.24) against predators by two males (using the data obtained
in LW form; Appendix 3).
Then s and k were substituted in q* = (1 – r – s – rs)/[(1 + r)(k – s)], and the
change of the ESS curve (the frequency of A strategy) with r-values could be
simulated (Fig. 6.13; see white line in Fig. 6.12a). Next, setting q* calculated from
the experimental data (Appendix 3) at the vertical axis and the winter minimum
temperature (=regarded as relative relatedness between males) at the horizontal
axis, I obtained Fig. 6.14. By indexing the plot points of HG (open circles) in
Fig. 6.14, I expanded and contracted the horizontal axis of Fig. 6.13 to obtain a
curve that fits the almost all population plots on a computer. Finally, I obtained
Fig. 6.15, on which the cline of HG form is well fitted. The same procedure was
applied for the LW form, providing Fig. 6.16. From these figures, we could
estimate the relatedness of males as the values from 0.46 to 0.60 in HG form
(Fig. 6.15) and 0.57 to 0.62 in LW form (Fig. 6.16). The aggression frequency
cline may form at some point between these values.
s = 0.25
0.8 k= 0.35
0.6
q*
0.4
0.2
0
0.4 0.5 0.6 0.7
Relatedness (r)
Fig. 6.13 Simulated cline by using model of Saito and Takada (2009) and practical data in
Appendix 3
1
0.8
0.6
q*
0.4
0.2
0
8 6 4 2 0 -2 -4 -6 -8
Average winter minimum temperature (AMT ˚C)
Fig. 6.14 Observed cline in the field (q* are transformed from Fig. 6.6 data as shown in
Appendix 3). Note that the horizontal axis (AMT) is reversed in scale (against Fig. 6.6.,6.6) to
adjust AMT to the expected change of relatedness, i.e. high temperature at left and low tempera-
ture at right
0.8
s = 0.25
0.6 k = 0.35
q*
0.4
0.2
0
8 6 4 2 0 -2 -4 -6 -8 AMT
Fig. 6.15 Fitting of EES curve (Fig. 6.13) to the observed cline (open plots) in HG form. AMT is
an index of winter harshness (average daily minimum temperature in winter); r is the relatedness
between males, and q* is the frequency of aggression strategy. Arrows with numerals show the
estimated r values
6.6 Theoretical Explanation of Male Aggression and Cooperation 137
0.8
s = 0.25
0.6 k = 0.35
q*
0.4
0.2
0
8 6 4 2 0 -2 -4 -6 -8 AMT
Fig. 6.16 Fitting of EES curve (Fig. 6.13) to the observed cline (gray plots) in LW form. AMT
is an index of winter harshness (average daily minimum temperature in winter); r is the related-
ness between males, and q* is the frequency of aggression strategy. Arrows with numerals show
the estimated r values
Thought Exercise
Are the relatedness values ranging from r = 0.48 to r = 0.62 at which the mixed
strategy (cline) appears in Figs. 6.15 and 6.16 reasonable estimates? As stated
before, I have been unable to address this directly, but it is possible to roughly
evaluate the r-value by the use of indirect data from field observations. Even though
there is a danger of double-quoted hypothesis errors, I am not afraid to do so here.
Let us discuss this as a thought exercise only. One available data item is the
proportion of unfertilized females in overwintering females in the field.
Furthermore, the overwintering probability of males by spring is available.
Because St. miscanthi is a haplo-diploid organism, there is little relatedness
between father and son if they are under outbreeding conditions. On the other hand,
if the mites are under inbreeding conditions, the relatedness between males drasti-
cally increases (see Box 4.11). Therefore, the structure of the mating population
must influence relatedness between males. To estimate the relatedness from such
data, we need several presuppositions for simplification, as follows. Assumption 1:
There are no overwintering males. Assumption 2: There are no immigrant males
138 6 Kin Selection
(mating occurs only between nest members), such that unfertilized females produce
only males and they mate with their sons. Assumption 3: all nests are founded by
single females.
I first attempt to evaluate the relatedness between males from the fertilization
status of overwintering females only. Under the three assumptions, the average
relatedness between males in a nest can be estimated as follows: Between males
produced by an unfertilized female in the first generation (produced by overwin-
tered females), r = 0.5 on average (because they are all sibs), and between males
produced by a mother–son mating female in the next (second) generation, r = 0.625
(Box 4.11). In the other nests where single fertilized females reproduce, the average
relatedness between males is 0.5 regardless of the generation. Here let us denote the
unfertilized rate as u. The observed proportion of unfertilized females ranged from
a minimum of 0.03 in Wakamiya to a maximum of 0.36 in Unzen populations.
Then, in the first generation, the relatedness between males in a nest is r = 0.5
regardless of fertilization status, but in the second generation it will increase to
r = u × 0.625 + (1 − u) × 0.5 = 0.503 in Wakamiya and r = 0.545 in Unzen populations.
Of course, the above relatedness value must be an overestimation because of
assumptions 1 through 3. Relaxing assumption 3, the proportion of nests founded
by i foundresses is denoted as vi (i = 1, 2, 3…) in spring nests (just after overwinter-
ing). The relatedness must decrease depending upon i and vi, if the relatedness
between foundresses is low. Let assume that there is no relatedness (note that this
assumption is too strict) between foundresses of a nest (r = 0, an extreme case); then
we can calculate the change of r (under assumptions 1 and 2) between males in the
first generation as r = 0.5 ∑(vi/i) (Box 6.2). By using vi observed in several popula-
tions (see Fig. 6.4), we can estimate r = 0.29 in the first generation. In the second
generation, the r must increase by the effect of mother–son mating, such that r must
increase. However, it is too complex to estimate here. Then, r = 0.29 must be an
underestimation.
Let us try to calculate r from the effect of male overwintering probability (relax-
ing assumptions 1 and 2 under assumption 3). Overwintering males must affect the
relatedness between males in two ways, i.e., they decrease unfertilized females dur-
ing winter to early spring, and increase outbreeding in spring nests especially in the
second generation in spring. Because the former effect primarily changes the rate
of unfertilized females in spring, thus it is a synonymous case with the first attempt.
Here I considered the latter effect only, i.e., no unfertilized females overwinter.
How the relatedness between males is affected by overwintering males can be cal-
culated from the proportion of nests where overwintered males exist in the first
generation as follows.
If there is no relatedness between overwintered males and the newly produced
males in spring nests, and the number of newly born males is assumed to be two on
average (males are rare in this species; see Table 4.3), the relatedness of males in a
nest where an overwintered male immigrated decreases from 0.5 to 0.167. By
denoting the proportion of nests in which an overwintered male exists as m, we can
roughly calculate as r = 0.167m + 0.5(1 – m). In the most aggressive population
(HG) collected in Ashizuri, 0.81 males per nest were observed in winter (February).
On the other hand, only 0.02 males per nest were observed in the Odochi popula-
tion (LW) that shows lower male aggressiveness. I could assume that 81% of nests
involved at most one overwintered male in the former population (remember that
only single males can exist in such nests; Sect. 6.1) and 2% of nests in the latter
population. Therefore, I obtained r = 0.2 in Ashizuri and r = 0.493 in the Odochi
population. Because the assumption that there is no relatedness between overwin-
tered males and newly born males must be too severe in this species, these estima-
tions must be underestimations.
As such, the several attempts to estimate r between St. miscanthi males interact-
ing in their respective nests may range from 0.2 to 0.55 in spring nests. These
estimations are not independent, and one may influence the other. Furthermore,
there are expected to be other factors that may increase the mean relatedness
between interacting males, such as the primary relatedness (Wright’s coefficient of
inbreeding) between the foundresses of the first generation and that between over-
wintered males and newly born males, both of which were assumed here to be zero.
Thus it must be noted that the relatedness evaluated here may be a rather “underes-
timated” value as a “thought exercise.” Actually, Antolin (1999) showed that total
inbreeding coefficient (Fit) in a population of Trichogramma pretiosum in a field
was 0.246. If we could assume that the value is near to St. miscanthi and the related-
ness is additive (Lynch and Walsh 1998), the range of relatedness becomes
r = 0.45 − 0.80. It is noticeable that the range of the relatedness predicted here well
overlaps with the range predicted from the Saito and Takada (2009) model, i.e.,
r = 0.48 − 0.62 (see Figs. 6.15, 6.16).
Chapter 7
Is Basic Science the Father of Applied Science?
So far, I have described the behavioral and evolutionary ecology of spider mites,
mainly by drawing upon several social spider mite species. As scientists, we some-
times struggle to reconcile social (applied) requirements with our own interests. As
I am employed in an agricultural college, I have often struggled with such a dichot-
omy: How do your studies contribute to agriculture and/or society? Although it is
perhaps an old-fashioned example, I intend to describe my limited experiences with
this theme.
In the autumn of 1997, I received a letter from a researcher at the Fujian Plant
Protection Institute, China. In that letter, an unbelievable situation was described:
Outbreaks of pest mites on moso bamboo (Phyllostachys pubescens) had been
occurring since the late 1980s, and many bamboo forests in Fujian province had
suffered serious damage (Zhang and Zhang 2000). Moso bamboo (giant bamboo)
is a species native to China and is thoroughly utilized in industry, housing, and
handicrafts, and its shoots are a prized food item in many Asian countries.
Furthermore, moso bamboo forests play a vital role in preventing soil erosion in
mountainous areas. Spider mite outbreaks are not uncommon in agricultural fields
and forests, but the outbreak species are mostly restricted to Tetranychus,
Panonychus, and Oligonychus species – all of which are regarded as r-selected
species (Sect. 2.2) with great potential for population increase. However, the main
species of this particular outbreak was Stigmaeopsis (Schizotetranychus) nanjin-
gensis, Dr. Zhang said. This species is a very close relative of St. longus and St.
miscanthi, the life histories and social life style of which I have previously described
at length. Several photographs as well as the dried-up leaves covered with dead
mites enclosed in the letter suggested this was an extreme situation (Fig. 7.1). The
primary cause for my surprise was my belief that Stigmaeopsis spp. are K-selected
species that have evolved in stable habitats with many natural enemies, as described
in Sect. 3.6. I am convinced that this is why they evolved sociality.
Fig. 7.1 Damaged moso bamboo forest in Fujian Province in China (photo by Zhang YX)
to increase the yields of bamboo culms and shoots, which inevitably lead to a
decrease in both plant and arthropod (pests as well as their natural enemies) diver-
sity (Zhang et al. 2004). Zhang (2002) studied the mite fauna of polyculture moso
bamboo forests in comparison with monoculture ones, showing that there is a great
difference in the species richness of predatory mites between them. Furthermore,
she noticed that the time of the initial mite outbreaks corresponded almost exactly
to the time when the moso bamboo plantations in Fujian Province were shifting
from polyculture cultivation practices to monoculture ones (Zhang et al. 2004).
Through repeated surveys on the mite fauna in Fujian moso forests, we became
more and more inclined to believe that the lack of natural enemies, which would
have ordinarily suppressed St. nanjingensis populations before the change in culti-
vation practice, must be a key factor. Figure 7.2 shows the population fluctuations
of St. nanjingensis and its phytoseiid natural enemies (predators) in heavily damaged
moso bamboo forest (monocultures). There are no consistent trends in the dynamics
of either prey or predator, showing no effective prey regulation by the predators.
If there is a classic-type predator–prey population interaction, the prey population
initially increases, then the predator population increases by consuming the abun-
dant prey. The increased predation pressure then causes the prey population to
decrease, which in turn leads to a decrease in predator numbers decrease as a result
450
350
300
250
200
150
Tp. bambusae(predator)
100
50 ?
Sep. 26, 96’ Jan. 6, 97’ Sep. 6, 97’ Jan. 6, 98’ Sep. 29, 98’
Date
Fig. 7.2 Outbreaks on monoculture Moso bamboo forests in Fujian, China (number of mites per
ten leaves). Arrows show predator extinction (from Zhang 2002; Saito et al. 2008b)
144 7 Is Basic Science the Father of Applied Science?
of a lack of prey. Thus, we may observe a more or less delayed oscillation of the
predator population following the prey population dynamics (Lotka 1925; Volterra
1926; Huffaker 1958 in Fig. 2.15). We ascertained that the most abundant predator
species occurring on moso bamboo was Tp. bambusae. Needless to say, I was sur-
prised, because this predator is known to be a formidable regulator of Stigmaeopsis
spp. in Sasa forests in Japan (Sect. 3.5). Summing up this information, I got the idea
that polyculture plantations contain several key plants other than moso bamboo on
which other Stigmaeopsis species inhabit. These Stigmaeopsis species are a vital
alternative (secondary) prey for Tp. bambusae. An important point is that this
predator species only occurs with Stigmaeopsis species in Japan (Sect. 3.5),
because it has a strong preference for living under Stigmaeopsis woven nests and
has a bit of difficulty reproducing without such nests (Saito 1990b). Based on these
facts, I doubted whether Tp. bambusae could persist in monocultured moso bamboo
forests.
Approaching a Factor
If so, then what are these key plants? Fortunately we soon got the answer from a
survey of an old abandoned moso bamboo forest near the offices of the Fujian Plant
Protection Institute. One day when we were taking a rest with cups of Ulong tea in
Dr. Zhang’s office, her assistant, Mr. Ji, brought a bundle of Chinese silvergrass
(Miscanthus sinensis) collected in the nearby forest to ask me to identify the spider
mites infesting it. At that time, I had no information about whether Stigmaeopsis
miscanthi was distributed in China. Hence, I was very surprised to see so many
injury spots of this mite species on the leaves. This observation later became very
important to hypothesize on the origin of the two forms of St. miscanthi in Japan
(Sect. 6.4), but at that time it was a crucial hint in our search to discover the key
causative factor of the St. nanjingensis outbreaks in moso bamboo forests. Taking
the silvergrass, I observed the leaves one by one under a dissecting microscope, and
I found numerous Tp. bambusae in the nests of St. miscanthi! Tp. bambusae is
frequently found in St. miscanthi nests on Chinese silvergrass in Japan (Chittenden
2002). Furthermore, Chinese silvergrass must be native in China, as expected from
its scientific name. Right then, I felt that I could fully understand the cause of mite
outbreaks in moso bamboo forests.
After the foregoing discovery, we conducted two experiments to further clarify
our hypothesis: (1) Can St. miscanthi infest moso bamboo? (2) Can the Tp. bambu-
sae collected from St. miscanthi really feed on St. nanjingensis? We got our
answers very quickly: St. miscanthi never reproduces on moso bamboo leaves, and
Tp. bambusae obtained from M. sinensis is the same as that obtained from moso
bamboo forests and can effectively feed upon St. nanjingensis (Zhang unpublished
data). Furthermore, Chinese silvergrass was a very common undercover plant in
moso bamboo forests, but as it is now used as a material for papermaking, it is
periodically harvested and sold.
7.2 Simulation of One-Predator–Two-Prey Systems 145
250
No. of mites (predator x 5)
200
100
50 Tp. bambusae
0
Apr. 1, ‘01 Aug. 1, ‘01 Dec. 1, ‘01 Apr. 1, ‘01
Date
Fig. 7.3 Predator (Tp. bambusae) and prey (St. miscanthi) populations on Chinese silvergrass
(Miscanthus (M.) sinensis) growing in a mixed (polyculture moso bamboo) forest (number of
mites per 50 leaves) (from Saito et al. 2008b)
However, whether the existence of nursery plants for Tp. bambusae in moso forests
really helps regulate St. nanjingensis is another problem. From a theoretical per-
spective, predator–prey interactions have generally been thought to be unstable
because the conditions under which the Lotka–Volterra equation is hypothesized
are usually hard to find in practice (Kuno 1987). On the other hand, Tansky (1978)
and Matsuda et al. (1985) showed that predator–prey systems under which preda-
tors can switch from one prey species to another in a prey density-dependent man-
ner easily attain a stable state. In the latter studies, it was hypothesized that predator
individuals have the ability to switch between two prey species depending upon
respective prey density. Thus, the switching threshold of prey density is regarded as
an evolutionary trait of individual predators. We thought that this model situation
might be analogous to the systems observed in moso bamboo and Chinese silver-
grass in China. Although there is no way for Tp. bambusae to switch prey species
on a single host plant species, because only a single species occurs on it, they could
146 7 Is Basic Science the Father of Applied Science?
realize the switch simply by moving to the other nearby host plant to access another
prey species. We attempted to test this idea by using a computer simulation (Saito
et al. 2008b; Tsuji et al., unpublished data).
The purpose of the simulation described below is to show what happens in pest
mite populations when moso bamboo plantations are switched from polyculture to
monoculture practices and to define the causal keys to the outbreaks. To ascertain
the factors causing the trends observed in moso bamboo (abbreviated as PP) and
silvergrass (MS), we developed a simple system model that can simultaneously
simulate “one predator–two prey” interactions on two host plant species (Saito
et al. 2008b). Most data for St. nanjingensis were referred from Zhang and Zhang
(2000), those for St. miscanthi from Sato et al. (2000), and for the predator, Tp.
bambusae, from Saito (1990b) and Zhang (2002). In addition, several unknown
parameters for prey species were obtained by substituting data for closely related
species. Climatic conditions were deemed constant, and all life history parameters
were fixed as being under 25±1°C, favorable humidity (50–80% r.h.), and no rain
conditions. Furthermore, host plants were all considered to be infinitely abundant,
such that there were no food limitations on the two prey species. The ultimate
purpose of this simulation model was thus to ascertain both the previously undeter-
mined effect of successful predator migration between PP and MS and the effect of
predation success under field conditions on the stability of the overall system.
The program consisted of two distinct host plant subsystems, i.e., MS and PP
(Box 7.1). No compatibility in prey species exists between the host plants, with
St. nanjingensis occurring only on PP, and St. miscanthi occurring only on MS, as
mentioned before. The predator Tp. bambusae, on the other hand, can readily eat
both prey species by migrating between the two host plants with particular proba-
bilities of immigration success. We adopted two stages – one is nonreproductive
(stage 1; immature females from egg to preovipositional female) and the other is
reproductive (stage 2; mature females) – to make the situation as realistic but as
simple as possible. The model was defined as difference equations (a more detailed
explanation is given by Tsuji et al., manuscript submitted). Strictly defining sys-
tems stability in systems simulation is usually difficult, such that we tentatively
defined the “stable state” to be when the predator–prey system continued for at least
1,100 days (about 3 years), except for cases when both the predator and prey popu-
lations increased indefinitely (overshooting the equilibrium).
First, we simulated whether two subsystems (PP and MS) reach stable states
when isolated: initial prey and predator densities are fixed. The parameters of
predator immigration success coefficients from one subsystem to the other (from
MS to PP and from PP to MS) were 0 (completely isolated), and the predation suc-
cess coefficients vary artificially (trial-and-error method). Then we got basically
three output types with time. If the predation success coefficients were sufficiently
high, the prey population became extinct and shortly thereafter the predator popula-
tion did so as well. If these coefficients decreased, then the predator population
became extinct first, and shortly thereafter the prey population entered a phase of
overshooting the equilibrium (Box 7.2), or the two populations increased gradually.
Figure 7.4a,b shows the result of a simulation where the predator population first
Box 7.1 Brief flow of simulation model in Chinese silvergrass and moso bamboo
Temp.rainfall,
TB:Typhlodromus nutrition are
bambusae constatnt
(1=immature, 2=adult)
SM:Stigmaeopsis miscanthi
(1=immature, 2=adult) SN:Stigmaeopsis
STV:starvation najingensis
FD: food intake Death and (1=immature,
PR: preyed predation Death and
Death 2=adult)
predation
Death
PR
STV
PR
TB(1) STV
SM(1)
TB(1) SN(1)
7.2 Simulation of One-Predator–Two-Prey Systems
FD
Immigration FD
success Immigration
SM(2) success
TB(2)
TB(2) SN(2)
STV PR
STV PR
FD Death and
predation Death and
Death and predation
emigration to FD
Death and
moso system
emigration to
silvergrass
T. bambusae-S. miscanthi module on Chinese silvergrass (Miscanthussinensis) system
T. bambusae-S. nanjingensis module on moso babboo(Phyllostachys pubescens)
147
148 7 Is Basic Science the Father of Applied Science?
Box 7.2
It is questionable whether the prey population is suppressed completely and
then the predator population becomes extinct in nature. In agricultural closed
systems, most examples showed such dynamics (Kuno 1987; Saito et al.
1996). However, if such relationships commonly existed in nature, we would
not observe any predator–prey relationships, because they all have become
extinct. In nature, of course, habitat heterogeneity and some complex rela-
tionships must operate to conserve prey and then predator populations.
Rather I should point out the famous “life-dinner” principle provided by
Dawkins and Krebs (1979). This principle predicts that predators never
evolve their hunting ability to overexploit prey at all, because the latter face
the risk of losing life but the former have only the risk of losing a dinner.
Especially in moso bamboo forests covering several hectares, predators may
not able to discover and exploit all prey individuals when densities of both
predators and prey become so low. Therefore, we think that it is more realistic
to suppose that predators must become extinct because of starvation; then
usually some prey individuals can escape predation.
Bamboo Silvergrass
500 a b
400 No predator
disperse
300
200
100
0
0 200 400 600 800 1000 0 200 400 600 800 1000
Populations
1000 c d
Predators
800 disperse at
high rates
600
Prey
400
200 Predator
0
0 200 400 600 800 1000 0 200 400 600 800 1000
Days
Fig. 7.4 a, b There is no dispersion of predator mites between bamboo and silvergrass systems. c,
d Predators can move (disperse) at high rates between the two systems (after Saito et al. 2008b)
7.2 Simulation of One-Predator–Two-Prey Systems 149
became extinct and thereafter the prey population entered the overshooting phase.
Here we assumed that the predator could not find the prey at low densities because
of habitat heterogeneity in a large-scale forest (Box 7.2). These two coefficients
were used as constant values in the following simulation.
Next we simulated when the predation success coefficients were fixed at the
values in Fig. 7.4a,b and when the migration success coefficients varied arbi-
trarily. Whether or not both subsystems (=whole system) attained stability was
dependent upon the values of the migration success coefficients. There were three
outcomes: both subsystems became stable, one system became stable but the
other system did not (both predator and prey gradually increase infinitely), and
both systems became unstable. Several combinations of the two migration suc-
cess coefficients can make the systems stable at low levels of prey and predator
density (Fig. 7.4c,d).
From these outlined simulation trials, we could infer that a “two-plant–two host-
specific pest species – single natural enemy” system can sometimes attain stability
through predator migration rates between two plant species. Although this systems
simulation is too simplistic and its outcomes too abstract to mimic the actual
dynamics of mite populations in Chinese bamboo forests, it at least shows that a
“two plants / two pests / single natural enemy” complex (here, P. pubescens / M. sin-
ensis / St. nanjingensis / St. miscanthi / Tp. bambusae, respectively) can potentially
attain stability, suggesting that effective control of pest mite outbreaks might be
possible simply by recovering “biodiversity.” We now believe that these findings
can partly explain the phenomenon that has occurred in Fujian: A change in bam-
boo cultivation practice from polyculture to monoculture, which simultaneously
eliminated Chinese silvergrass vegetation, may have triggered the outbreaks of St.
nanjingensis.
As such, we have established that reintroducing Chinese silvergrass (M. sinen-
sis) is necessary to restore moso bamboo plantations in China. Although traditional
polyculture moso plantations involve high plant and arthropod diversity, it is note-
worthy that even the recovery of just a single plant species can greatly improve
system stability, i.e., sustainability in agricultural fields. Thus, we may show that
the conservation of diversified systems must contribute to the stability of biosys-
tems as well as provide biological control of pest species by using endemic natural
enemies. In other words, we may prove that biodiversity works as an effective regu-
lator of pests.
So far I have described an attempt to apply the results of basic research to an
applied field, i.e., pest management. If I had not had access to any data about
Stigmaeopsis biology or their predators before the Chinese survey, I might not
have detected any way to solve the mite outbreak in moso bamboo forests.
Studies on spider mites occurring on Sasa still have no practical value from the
point of view of agriculture and forestry in Japan (Box 7.3), such that I repeat-
edly encounter ironic questions from applied researchers, such as “What is the
purpose of your study?” and “Is a study on giant panda diet necessary?” I believe
that basic science has provided the key ideas for how to solve the pest problems
in China.
150 7 Is Basic Science the Father of Applied Science?
Box 7.3
The discovery that two prey species on different plants stabilize predator
populations may suggest another reason why Stigmaeopsis species on Sasa
never experience outbreaks in Japan. In Sect. 3.5 I described that there are at
least two species of Stigmaeopsis that have different nest sizes. Such a differ-
ence in nest size is possibly related to the switching behavior of Typhlodromus
bambusae and Agistemus iburiensis, as modeled by Tansky (1978). If there
are such switching traits in these predators, the predator–prey populations
must become stable, as predicted by the model.
I would like to address an epilogue to the Chinese survey in this chapter. Even
though we could provide a possible means of regulating pest mite outbreaks in
Chinese bamboo plantations by reintroducing biodiversity, the question of why
Japanese Sato-yama have been destroyed by the overvegetation of introduced moso
bamboo remains unanswered.
Moso bamboo, believed to have been introduced into Japan about 270 years ago
from China through the Ryukyu Islands (according to one account, it is 1,200 years
ago), is now destroying traditional Japanese forests, called “Sato-yama” (Fig. 7.5).
This destruction occurs because poorly maintained bamboo forests rapidly cover
over and eliminate other native plant species, which loss subsequently triggers
landslides. Cheap imported bamboo shoots from Fujian Province have discouraged
Japanese farmers from maintaining their bamboo forests. The biological pest man-
agement introduced here, if widely adopted in China, might have further adverse
affects on the sustainability of Japanese Sato-yama. Thus, a fundamental conflict
exists between Chinese and Japanese bamboo plantations. We have learned that the
establishment of sustainable agriculture in one country sometimes causes unex-
pected problems for sustainability in another.
To resolve this problem, we must identify the mechanisms that prevent moso
bamboo from covering over and eliminating other plant species in its native areas,
such as Fujian Province in China. During the moso bamboo study, I learned that
there are two bamboo rat species that feed on moso bamboo shoots. Furthermore,
in the old days giant pandas were prolific consumers of bamboo shoots, even in
Fujian Province (Fig. 7.6). How these species regulated moso bamboo in the past
is now difficult to ascertain, but it is very suggestive that bamboo rats prefer to use
Chinese silvergrass for their den sites (Fig. 7.6; Xu 1989). As the introduction of
exotic pest species (mites, bamboo rats) carries inherent danger to indigenous flora
and fauna, we cannot recommend such a course of action in Japan, however. Thus,
conservation of Japanese Sato-yama through the prevention of unchecked moso
7.3 Epilogue: Japanese Sato-yama 151
Fig. 7.5 Moso bamboo climbing up traditional Japanese “Sato-yama” (Karatsu City, Kyushu
district, Japan)
regulation regulation
damage
damage
¥
Predatory mite
Typhlodromus bambusae
Chinese
Moso bamboo
Overvegetation
Symbiont? silvergrass
Phyllostachys pubescens Miscanthussinensis
¥ damage damage
damage providing
Giant panda den sites
Ailuropoda melanoleuca
Bamboo rats
Rhizomys sinensis and/or Rhizomys pruinosus
Fig. 7.6 Possible relationships between moso bamboo and Chinese silvergrass and their herbi-
vores and predators in ancient China (from Saito et al. 2008b)
Evolutionary ecology and sociobiology are sciences for studying the selection pressures
operating in nature. As I have described, I believe that spider mites and phytoseiid
mites are ideal materials for such studies, because they are often directly exposed
to natural selection. Their short life spans and our ability to observe their entire life
cycle in small experimental areas are also great advantages. By these and other
merits, we have been able to conduct much of the research described so far.
However, there are also some difficulties involved in using mites for ecological
studies. Each trait of the species that we focus on, the nesting behavior of St. longus,
for example, has to be checked against every kind of selection, i.e., natural selection
operating on survival and reproduction, sexual selection on mating success, and
synergistic and/or kin selection on social living, all of which must be operating
simultaneously in natural habitats. The minute bodies of mites often make it impos-
sible to observe individual lifetime success in the field. Furthermore, environmental
conditions surrounding such minute organisms may be quite different from our own
sense of realism, as pointed out by Schmidt-Nielsen (1984) and Saito and Suzuki
(1987). Difficulties underlying the study of spider mite genetics must be also men-
tioned. Saito et al. (2000b) revealed that there are recessive genes that govern the
depression of female fecundity (Chap. 5). However, they could not determine how
many genes (alleles and loci) are responsible for this trait or how intense this effect
is in comparison with the wild type. To acquire such information, we have to
observe the fitness changes between generations under constant conditions without
environmental variance. Because spider mites are phytophagous, it has been (and
still is) difficult to keep host plant conditions constant. Therefore, we still have only
limited means of obtaining this important information. It is necessary to develop
some kind of artificial diet that can keep the food conditions of spider mites constant
to overcome such a hurdle.
As is well known in many insects, as well as some mites, chemical cues are
a very important means for recognizing the external world (e.g., Dicke et al.
1998; Sato and Saito 2006). This importance partly comes from the fact that
mites basically live in a world they cannot see. Are there any possibilities that
they might use sound as a cue for communication? Because of their tiny bodies,
such a study has never been attempted. Although it may be difficult to separate
sound from tactile stimulus, we must focus more attention on such possibilities
as well as on chemical cues.
I have described several unresolved questions in Chaps. 5 and 6. Among these,
the origin of the two forms of St. miscanthi is our current research focus. In relation
to this theme, Sato et al. (2008) recently reported that there is a contact zone of
these two forms and a possibility of interbreeding between them. If such interbreeding
frequently occurs in the field, “introgenesis” between the two forms may be
expected. Such “hybrid gene effects” may simply explain the variation in male
aggressiveness of St. miscanthi (M. Kawata and Y. Sato, personal communication)
and represent the antithesis of my kin selection hypothesis. However, I believe that
this cannot be an alternative to kin selection, because the former is a proximate and
the latter is an ultimate factor to explain such a phenomenon. A more important
question is why two forms showing “different” intensities of male aggressiveness
exist in Japan, even though we know there is a quantitative difference in diapause
attributes between the forms (Sect. 6.4). To solve this problem, we have to both
widen our perspective, to the Asian continent, for example, while at the same time
narrowing it down to embrace the molecular level.
Finally, I address our recent attempts to understand the structure and function
of mite communities. Sasa bamboo is known to harbor a complex mite community
(Chittenden 2002; Mori and Saito 2004). We have only focused on several spider
mite species on Sasa bamboo, but at least nine species of spider mites and perhaps
an equal number of predacious mite species live either syntopically or allotopically
in the same forests. I do not know whether the mite diversity observed on Sasa
bamboo is extraordinary or not, but the reasons why there are so many host-specific
species should be explained from the perspective of evolutionary community
ecology. The importance of the diversity of organisms has been stressed through-
out this book, but do we really know the reasons why there are such complex com-
munities? The relationships between spider mites and their natural enemies are
very easy to observe under seminatural experimental conditions as described in
this book, such that studies on the complex community existing on Sasa from the
aspect of evolutionary and behavioral biology may provide an answer to this very
important question.
Afterword
After I had completed the draft of this book, I sincerely thought that I had only
written about our own studies on spider mites and their natural enemies. This is a
concern of the advisers for this book. Anyway, how do you feel as a reader of this
book? Some have criticized that such narrow and fanatical research is not scientific,
and others have said that scientific research should be made with working hypoth-
eses to be proved before the start. On the other hand, mite researchers may say that
I have to write in a more detailed and comprehensive manner. These comments all
concern my style of ecological research methodology. Current ecology and evolu-
tional biology have strong trends that hypotheses or theories should come first and
then research be conducted to solve or prove them. This is how I have taught my
graduate students. However, I do not know when I first developed such habits, but
I have tended to adopt rather traditional natural history methodology, which is now
considered to be old-fashioned. This is the reason why I define my studies as Fabre
like, even though my contributions are hardly comparable to his great works.
However, it is apparent that no biological study can be conducted without observing
“phenomena” in nature. Therefore, I think that theoretical and or hypothesis verifi-
cation studies can only be conducted after we observe phenomena that need to be
resolved. In this context, ecology and evolutionary biology are inductive. In this
book I have introduced several fascinating researchers who have detected general
results by using mites, which can then be applied to “zoology.” However, if they
also pay more attention to mites as “actual animals” to be observed in relation to
their evolutionary histories, we are bound to learn more deep insights from their
results, I think. From the hypothesis verification style, my contributions are not so
general and must be restricted to a narrow area of zoology as well as acarology.
I now believe that the dichotomy of scientific methodology mentioned above is
unimportant. Theoretical studies should always be accompanied by natural history
studies and vice versa. Therefore, if we develop ecology, behavioral, and evolutionary
biology, we will always be paying attention to the traditional zoological sciences
– taxonomy, phylogeny, and physiology – as well as to theoretical biology. If what
I have presented in this book successfully resembles something like this type of
integrative biology, I will have achieved my goals.
155
Appendices
One of the major difficulties involved in reaching the reinstatement of the genus
Stigmaeopsis arose from the condition of the type specimen made by Banks (1917).
I asked the Smithsonian Institution (USA) to render type specimens of Stigmaeopsis
(Schizotetranychus) celarius and had the opportunity to observe them (by courtesy
of Dr. Ochoa R). Regrettably, the holotype and paratypes had disintegrated beyond
recognition, making morphological observations impossible. This is an unfortunate
and inevitable problem in mite taxonomy, because until Saito et al. (1993) developed
a new method for making permanent specimens using Canada balsam, no way to
prepare permanent specimens had existed (Jeppson et al. 1975). Furthermore, the
drawings made by Banks (1917) for species description were not always accurate
and made it difficult to identify characters. On the other hand, the descriptions by
Pritchard and Baker (1955) were very detailed and we could compare our specimens
with their drawings. We then decided to assume that Stigmaeopsis celarius described
by Banks was the same as Pritchard and Baker’s (1955) description of
Schizotetranychus celarius (herafter Sc). Under this assumption, we could compare
all populations of Sc distributed in Japan (Fig. 3.11) and China as follows: the LM
form is a synonym of Sc (redescribed by Pritchard and Baker 1955), L and MS are
different species from Sc as reported in Saito (1990a), and S and M are also different
from Sc, L, and MS in their morphologies. However, the question remained as to
whether the five Japanese species and two Chinese species should be categorized
into the genus Schizotetranychus or into the deleted genus Stigmaeopsis, because we
had also discovered an important difference in the chaetotaxy of the distal segment of
the palpus: In all seven species, the distal segment has two simple setae in both females
and males, but it has three simple setae in the other species of Schizotetranychus as
well as in most genera of Tetranychini (Lindquist 1985). Thus, we concluded that
such a difference must be an important criterion for separating these seven species
from the genus Schizotetranychus. Although there is no information about this dif-
ference in Banks’s (1917) first description, we chose to reinstate the genus
Stigmaeopsis and determined that these seven species should belong to it. As such,
we described M and S as two new species, Stigmaeopsis takahashii and Stigmaeopsis
157
158 Appendices
saharai, respectively, and moved the other five species, including the two Chinese
species (Schizotetranychus nanjingensis and Schizotetranychus tenuinidus), to the
genus Stigmaeopsis.
After reinstating Stigmaeopsis, we faced another problem. Takahashi (1987)
observed a species from Thailand that had a similar web nest habit (WN life type)
to Stigmaeopsis spp. and reported that the chaetotaxy of the distal segment of the
palpus was different from the Japanese Stigmaeopsis (Sc complex) species. We
subsequently rediscovered such a difference between Stigmaeopsis and the other
Tetranychini, as mentioned before. Recently, we (Sakagami et al. 2009) conducted
a survey of mite fauna in Thailand and could collect several species with WN life
types (Schizotetranychus saitoi, S. approximatus, and S. laevidorsatus). Using these
specimens and Stigmaeopsis spp., we conducted a molecular phylogenic study and
revealed that Stigmaeopsis spp. and the Thailand species inhabiting Bambusa spp.
and Gigantochloa spp., form two distinct but very close clades. Furthermore, we
found that dorsal chaetotaxy of the Thailand species (parallel-arranged dorsocentral
setae) more closely resembles the drawings of Banks (1917) than Stigmaeopsis spp.
(all species of this genus have dorsocentral setae arranged in a V-shape). This was
puzzling. We had no way to continue our taxonomic studies because of the lack of
an observable holotype. However, we now think that the Stigmaeopsis celarius
described by Banks (1917) is one of these three species from Thailand. There is no
way of deriving a correct answer from this confusion, but I am now considering
whether we must change the present Stigmaeopsis into a new genus, then return
Stigmaeopsis to the Thailand species some time in the future.
The point that I wish to convey from this story is the importance of a holotype.
The main difficulty of our taxonomic studies on Stigmaeopsis was caused by the
disintegration of specimens because of the shortcomings of Hoyer’s solution
(a kind of gum chloral fluid), the only mounting medium that was available at the
time for mite slide preparation. This water-soluble medium is very useful for short-term
preparations but has a short life span (10 years at most using sealing plastic).
To preserve such specimens periodic remounting is essential, but because of the
decreased numbers of taxonomists and curators in universities and museums
around the world, this is not always possible. I strongly disagree with this trend,
believing instead that a viable means of making permanent specimens is always
required before beginning taxonomic study. Now that we have the new Canada
balsam methods (Saito et al. 1993) to prepare all type specimens as permanent
specimens, spider mite taxonomy faces no such constraints.
In order to adapt the repeated blind sampling model for the nest-scattering trait of
Stigmaeopsis species in Saito et al. (2008a), we formulated a mathematical model.
Appendix 2. Model of Repeated Blind Sampling 159
Model
The many small Stigmaeopsis nests randomly scattered over the leaf surface are
considered to be a situation analogous to the tricking effect of prey dummy nests,
because under natural conditions such nests persist on leaves for a certain period
even after suffering predation. We developed a simple model to show how void
nests can effectively trick predators. From the VTR (video tape recorder) and field
observations, the conditions assumed for the tricking model are as follows:
1. A prey female makes x nests randomly over a certain area throughout her
lifetime.
2. A prey female reproduces N eggs throughout her lifetime.
3. A prey female evenly oviposits N/x eggs in each nest (N ³ x).
4. A predator female finds all prey nests with equal probability regardless of the
number of nests or the presence of prey eggs within them.
5. When a predator female finds a nest, it always intrudes into the nest regardless of
the presence of prey eggs and eats all available prey eggs.
Assuming that i nests have already been found and the prey eggs within them
have been consumed [i.e., there are i already preyed (void) nests], the number of
nests that have not been found is x − i. Thus, the expected number of searching
attempts required to find a “nest with prey eggs” is x/(x − i), but x ¹ i.
From the above, the number of trials (R) in which a predator has found k valid
nests (k £ x) is
k −1
R = x ∑ [1 / ( x − i )] (A.1)
i=0
The searching time required to find a nest (regardless of whether it is valid or void)
was denoted by s (including the handling time required to intrude a nest). The time
required for a predator to eat an egg is denoted by e. Introducing these parameters
into (A.1), we obtain the time (T) required for a predator to eat all eggs (N) as the
following equation (the resting time including oviposition time must occupy a cer-
tain proportion of the total time spent but is ignored here):
x −1
T = sx ∑ [1/ ( x − i )]+ eN (A.2)
i=0
From the point of view of the prey, x is a strategy by which prey fitness will increase
with time T, because as T lengthens, the fewer prey eggs will be eaten.
Table A.1 Behavioral parameters of Agistemus (Ag.) iburiensis at x (number of nests) = 6
determined by VTR observations (n = 5) for 24 h
Parameters Values estimated (h)
Searching and handling time required for a nest, s 1.27
Time for eating an egg, e 0.33
160 Appendices
By using the raw data without any adjustment (therefore, the data are not consistent
with Figs. 6.5 and 6.6), q was calculated as follows: One male death observed
experimentally during 5 days (most results were settled within 3 days) is regarded to
occur on the following three occasions: the combat between A strategy and A strategy,
that between A strategy and C strategy, and that between C strategy vs. A strategy.
The frequency of A strategy could be calculated. Let us denote the frequency of A
strategy as q and that of C strategy as p. The frequency of A vs. A confrontations is
q2, that of C vs. C is p2, that of A vs. C is pq, and that of C vs. A is pq. Considering
that all the confrontations including A strategy result in the death of one of the paired
males, the frequency of p2 can be calculated from the experimental data as
p2 = 1 − “the probability of the death of one of the two males in the previous studies.”
Thus the frequency of A strategy can be obtained as q = 1 − p.
The aggression cost between A strategists can also be calculated by using the
above calculation as follows: The probability of killing each other was 0.081,
observed in the Sumoto population of S. miscanthi, showing 0.76 probability of
being killed by the other male. However, this is not the required value of k, because
k is only the risk when males adopting A strategy fight each other. The frequency
of A vs. A fighting can be calculated from the frequency of C vs. C (p2) as above.
For example, in the Sumoto population, the frequency of p2 = 1 - 0.756; thus I
obtained p = 0.494. Because p + q = 1, then q = 0.51 and k = 0.081/q2 = 0.31. I have
also calculated k for several populations in the same manner and obtained values
from 0.1 to 0.35 (Saito, unpublished data). Because these estimations are derived
from the extreme results of male fighting, they do not include any injury risks that
disable the winner, such that it is reasonable to adopt the maximum value of
k = 0.35.
The net effect of cooperation was newly estimated as “defense success against a
young predator by two males” / “that by a single male” in a nest with one female.
St. miscanthi and its predator species, Neoseiulus (N.) womersleyi (this species is
one of the predators frequently observed in St. miscanthi LW form nests), were
collected from Mt. Unzen, Nagasaki Prefecture, Japan, on August 12, 2004 (Saito,
unpublished data). Two young males and two unfertilized females of St. miscanthi
were introduced onto a detached host leaf (Miscanthus sinensis), and mating
occurred. Four days after their introduction, one female, one male or two males,
and eggs (4-12) were left in a nest, and the excess males and females were removed.
A predator larva was then introduced into the St. miscanthi nest. Three days after
the predator introduction, success of offspring defense by St. miscanthi adults was
observed (n = 23~27). The reason why we used St. miscanthi LW form for this
experiment came from the difficulty of keeping two HG males in a nest simultane-
ously because of their strong male-to-male aggressiveness. As stated before,
“aggression between males” and “counterattack ability against predators” are inde-
Appendix 3. Calculation of Practical Parameters in Game Model 161
pendent (remember the St. longus case), such that we assumed that the data
obtained for the LW form could be commonly used for the HG form.
When the predator was killed, the defense was considered to be complete (crite-
rion 1). Furthermore, if the predator was excluded from the nest at every observa-
tion (3 times) and if half of the prey immatures survived for 5 days, the counterattack
was considered to be successful (criterion 2). By the use of these criteria, the
defense success of a single male is 0.130 (criterion 1) and 0.522 (criterion 2), and
that of two males is 0.148 (criterion 1) and 0.609 (criterion 2). The net effect of
cooperation, s, is calculated as 0.14 at criterion 1 and 0.17 at criterion 2. If we con-
sidered the long-term effect [e.g., assuming double predatory intrusion events, t = 2
(for 10 days); Box 3.4], s is calculated as 0.29 at criterion 1 and 0.36 at criterion 2.
Here I used s = 0.25 as the middle value of 0.14 to 0.36. These estimations are con-
sidered to be tentative because of the small number of replications, i.e., n = 23 (one-
male defense) to 27 (two-male defense).
References
Abe J, Kamimura Y, Ito H, Matsuda, H, Shimada M (2003b) Local mate competition with lethal
male combat: effects of competitive asymmetry and information availability on a sex ratio
game. J Evol Biol 16:607–613
Abe J, Kamimura Y, Kondo N, Shimada M (2003a) Extremely female-biased sex ratio and lethal male–
male combat in a parasitoid wasp, Melittobia australica (Eulophidae). Behav Ecol 14:34–39
Alcock J (1979) Animal behaviour: an evolutionary approach, 2nd edn. Sinauer, Sunderland, MA
Antolin MF (1999) A genetic perspective on mating systems and sex ratios of parasitoid wasps.
Res Popul Ecol 41:29–37
Aponte O, McMurtry JA (1997) Damage on ‘Hass’ avocado leaves, webbing and nest behaviour
of Oligonychus perseae (Acari: Tetranychidae). Exp Appl Acarol 21:265–272
Atmer W (1991) On the role of males. Anim Behav 41:195–205
Aviles L (1997) Causes and consequences of cooperation and permanent sociality in spiders. In:
Choe JC, Crespi BJ (eds) The evolution of social behaviour in insects and arachnids.
Cambridge University Press, London, pp 476–498
Banks N (1917) New mites, mostly economic (Arach., Acar.). Entomol News 28:193–199
Begon M, Harper JL, Townsend CR (1999) Ecology: Individuals, populations and communities,
3rd edn. Blackwell Publishing, Oxford
Bernstein C (1984) Prey and predator emigration responses in the acarine system Tetranychius
urticae – Phytoseiulus persimilis. Oecologia (Berl) 61:134–142
Blazak C, Ehrensberger, Schuster R (1990) Beiträge zur Kenntnis der Lebenweise der Litoralmilbe
Macrocheles superbus Hull, 1918 (Acarina: Gamasina). Osnabrücker naturewiss Mitt 16:51–62
Boczek J (1959) Biologia I ekologia sierposza rozkuszkowka (Cheyletus eruditus) (Schrank,
1781) (Acarina, Cheyletidae). Press Naukowe IVR 175–230
Bolland HR, Gutierrez J, Flechtmann CHW (1998) World catalogue of the spider mite family
(Acari: Tetranychidae). Koninklijke Brill NV, Leiden
Boudreaux HB, Dosse G (1963) The usefulness of new taxonomic characters in females of the
genus Tetranychus Dufour (Acari: Tetranychidae). Acarologia 5:13–33
Bull JJ (1983) Evolution of sex determining mechanisms. Benjamin/Cummings, Advanced Book
Program, Palo Alto, California
Calvert WH, Hedrick LE, Brower LP (1979) Mortality of the monarch butterfly (Danaus plexip-
pus L.): avian predation at five overwintering sites in Mexico. Science 204:847–861
Carson R (1962) Silent spring. Houghton Mifflin, Boston
Chant DA (1959) Phytoseiid mites (Acarina: Phytoseiidae). Part I. Bionomics of seven species in
southeastern England. Part II. A taxonomic review of the family Phytoseiidae, with descrip-
tions of thirty-eight new species. Can Entomol 91 (suppl) 12:1–166
Charlesworth D, Charlesworth B (1987) Inbreeding depression and its evolutionary consequences.
Annu Rev Ecol Syst 18:237–268
Chittenden AR (2002) Diversity in predator–prey interaction of plant-inhabiting mites (Acari:
Tetranychidae, Phytoseiidae and Stigmaeidae). PhD dissertation, Hokkaido University, Sapporo
163
164 References
Chittenden AR, Saito Y (2001) Why are there feeding and nonfeeding larvae in phytoseiid mites
(Acari, Phytoseiidae)? J Ethol 19:55–62
Chittenden AR, Saito Y (2006) Tactile crypsis against non-visual predators in the spider mite,
Aponychus corpuzae Rimando (Acari: Tetranychidae). J Insect Behav 19:419–428
Cloutier C, Johnson S (1993) Interaction between life stages in a phytoseiid predator: western
flower thrips prey killed by adults as food for protonymphs of Amblyseius cucumeris. Exp
Appl Acarol 17:441–449
Cone WW (1979) Pheromones of Tetranychidae. In: Rodriguez JG (ed) Recent advances in
acarology, vol 2. Academic Press, New York, pp 309–317
Crespi BJ (1994) Three conditions for the evolution of eusociality: are they sufficient? Insect Soc
41:395–400
Crozier RH (1979) Genetics of sociality. In: Hermann HR (ed) Social insects, vol 1. Academic
Press, New York, pp 223–286
Crozier RH (1985) Adaptive consequences of male-haploidy. In: Helle W, Sabelis MW (eds) Spider
mites. Their biology, natural enemies and control, vol 1A. Elsevier, Amsterdam, pp 201–222
Dawkins R, Krebs JR (1979) Arms races within and between species. Proc R Soc Lond B
205:480–512
Dicke M, Sabelis MW. (1988) How plants obtain predatory mites as bodyguards. Neth J Zool
38:148–165
Dicke M, Sabelis MW, Takabayashi J, Bruin J, Posthumus MA (1990) Plant strategies of manipu-
lating predator–prey interactions through allelochemicals: prospecters for application in pest
control. J Chem Ecol 16:3091–3118
Dicke M, Takabayashi J, Posthumus MA, Schutte C, Krips OE (1998) Plant-phytoseiid interac-
tions mediated by herbivore-induced plant volatiles: variation in production of cues and in
responses of predatory mites. Exp Appl Acarol 22:311–333
Donzé G, Guerin PM (1994) Behavioural attributes and parental care of Varroa mites parasitizing
honeybee brood. Behav Ecol Sociobiol 34:305–319
Dosse G (1961) Uber die Bedeutung der pollennahrung fur Typhlodronus (T.) pyri Scheuten
(= tiliae Oud.) (Acari, Phytoseiidae). Entomol Exp Appl 4:191–195
Dosse G (1967) Beitrage zum Diapause-Problem von Tetranychus urticae Koch und Tetranychus
cinnabarinus Boisduval-Komplex im Libanon (Acarina, Tetranychidae). Pflanzenschutz Ber
34:129–138
Dugatkin LA (1997) Cooperation among animals. An evolutionary perspective. Oxford University
Press, New York
Ehara S (1957) On three spider mites of Schizotetranychus from Japan. J Fac Sci Hokkaido Univ
Ser 6 Zool 13:15–23
Ehara S (1993) Colored guide to the plant mites of Japan. (Ehara S ed), Zunkoku Noson Kyouiku
Kyokai, Tokyo (in Japanese)
Ehara S (1999) Revision of the Spider Mite Family Tetranychidae of Japan (Acari, Prostigmata).
Species Divers 4(1):63–141
Ehara S, Gotoh T (1992) Descriptions of two Panonychus spider mites from Japan with a key to
species of the genus in the world (Acari: Tetranychidae). Appl Entomol Zool 27:107–115
Ehara S, Okada T, Kato H (1994) Contribution to knowledge of mite family Phytoseiidae in Japan
(Acari: Gamasina). J Fac Educ Tottori Univ 42:119–160
Ehara S, Shinkaji N (1975) An introduction to agricultural acarology. Zenkoku Noson Kyoiku
Kyokai, Tokyo (in Japanese)
Ehara S, Shinkaji N (1996) Principles of plant acarology. Zenkoku Nouson Kyoiku Kyokai, Tokyo
(in Japanese)
Enquist M, Leimar O (1990) The evolution of fatal fighting. Anim Behav 39:1–9
Faraji F, Janssen A, van Rijn PCJ, Sabelis MW (2000) Kin recognition by the predatory mite
Iphiseius degenerans: discrimination among own, conspecific and heterospecific eggs. Ecol
Entomol 25:147–155
Fisher RA (1930) The genetical theory of natural selection. Oxford University Press, Oxford
References 165
Fleschner CA, Badgley ME, Ricker DW, Hall JC (1956) Air drift of spider mites. J Econ Entomol
49:624–627
Foster WA, Treherne JE (1981) Evidence for the dilution effect in the selfish herd from fish preda-
tion on a marine insect. Nature (Lond) 293:466–467
Fournier V, Rosenheim JA, Brodeur J, Laney LO, Johnson MW (2003) Herbivorous mites as
ecological engineers: indirect effects on arthropods inhabiting papaya foliage. Oecologia
(Berl) 135:442–450
Fowler HG, Gobbi N (1988) Cooperative prey capture by an orbweb spider. Naturwissenschaften
75:208–209
Frank SA (1985) Hierarchical selection theory and sex ratios. II. On applying the theory, and a test
with fig wasps. Evolution 39:949–964
Fransz HG (1974) The functional response to prey density in acarine system. Center for
Agriculture Publication Documents, Wageningen
Furuhashi K, Nishino M (1978) Studies on the dispersal of citrus red mite, Panonychus citri McG.
Bull Shizuoka Pref Citrus Exp Stn 14:29–42 (in Japanese with English summary)
Fujimoto H, Takafuji A (1993) Oviposition of diapause-induced females of Panonychus mori
Yokoyama (Acari: Tetranychidae) Appl Entmol Zool 28:106–108
Gerson U (1979) Silk production in Tetranychus (Acari: Tetranychidae). In: Rodriauez JG (ed)
Recent advances in acarology, vol 1. Academic Press, New York, pp 177–188
Gerson U (1985) Webbing. In: Helle W, Sabelis MW (eds) Spider mites. Their biology, natural
enemies and control, vol 1A. Elsevier, Amsterdam, pp 223–232
Gerson U, Aronowitz A (1981) Spider mite webbing. Part IV: The effect of acaricides on spinning
by the carmine spider mite Tetranychus cinnabarinus (Boisduval). Pestic Sci 12:211–214
Giron D, Dunn DW, Nardy ICW, Strand MR (2004) Aggression by polyembryonic wasp soldiers
correlates with kinship but not resource competition. Nature (Lond) 430:676–679
Glennis EJ, Fewell JH, Gadau J, Johnson RA, Larrabee D (2002) Genetic determination of the
queen caste in an ant hybrid zone. Proc Natl Acad Sci U S A 99:8157–8160
Goka K (2001) Geographical variation. In: Aoki J (ed) Biology of ticks and mites. University of
Tokyo Press, Tokyo (in Japanese)
Goka K, Takafuji A (1990) Genetical studies on the diapause of the two-spotted spider mite,
Tetranychus urticae Koch (1). Appl Entomol Zool 25:119–125
Goka K, Takafuji A (1991) Genetical studies on the diapause of the two-spotted spider mite,
Tetranychus urticae Koch (2). Appl Entomol Zool 256:77–84
Gotoh T (1984) Ecological studies on diapause in Tetranychidae, with special reference to dia-
pause attributes in relation to host plant phenology. PhD dissertation, Hokkaido University (in
Japanese)
Gotoh T (1986) Local variation in Panonychus akitanus Ehara (Acarina: Tetranychidae). Exp
Appl Acarol 2:137–151
Gotoh T (1987a) Annual life cycle of the spider mite, Eotetranychus uncatus Garman (Acarina:
Tetranychidae). Appl Entomol Zool 22:52–58
Gotoh T (1987b) Intraleaf distribution of Panonychus ulmi (Koch) (Acarina: Tetranychidae) on
dwarf bamboo. Appl Entomol Zool 22:248–258
Gotoh T (1987c) Life-history parameters of two species of the genus Eotetranychus (Acarina:
Tetranychidae) on deciduous trees. Appl Entomol Zool 22:45–51
Gotoh T (1987d) Comparison of life-history parameters of two forms of Panonychus akitanus
Ehara (Acarina: Tetranychidae). Appl Entomol Zool 22:119–124
Gotoh T (1989) Seasonal life cycles and diapause in three species of Schizotetranychus (Acari:
Tetranychidae). Appl Entomol Zool 24(4):407–417
Gotoh T, Shida T (2007) Life cycles and interactions in spider mites (Acari: Tetranychidae) on
dwarf bamboo, Sasa sinensis (F. & S.) (Poaceae), in Japan. Int J Acarol 33:259–273
Gotoh T, Shinkaji T (1981) Critical photoperiod and geographical variation of diapause induction
in the two-spotted spider mite, Tetranychus urticae Koch (Acarina: Tetranychidae), in Japan.
Jpn J Appl Entomol Zool 25:113–118 (in Japanese with English summary)
166 References
Grafen A (1979) The hawk–dove game played between relatives. Anim Behav 27:905–907
Grafen A, Ridley M (1983) A model of mate guarding. J Theor Biol 102:549–567
Grandjean F (1948) Quelques caracteres des Tetranyques. Bull Mus Hist Nat Paris (2) 20:517–524
Gutierrez J, Helle W (1985) Evolutionary changes in Tetranychidae. In: Helle W, Sabelis MW
(eds) Spider mites. Their biology, natural enemies and control, vol 1A. Elsevier, Amsterdam,
pp 91–107
Hamilton WD (1964) The genetical evolution of social behaviour. I, II. J Theor Biol 7:1–52
Hamilton WD (1967) Extraordinary sex ratios. Science 156:477–488
Hamilton WD (1971) Geometry for the selfish herd. J Theor Biol 31:295–311
Hamilton WD (1972) Altruism and related phenomena, mainly in social insects. Annu Rev Ecol
Syst 3:193–232
Hamilton WD (1979) Wingless and fighting males in fig wasps and other insects. In: Blum MS,
Blum NA (eds) Sexual selection and reproductive competition in insects. Academic Press,
London, pp 167–220
Hardy CWI, Mayhew PJ (19a98) Sex ratio, sexual dimorphism and mating structure in bethylid
wasps. Behav Ecol Sociobiol 42:383–395
Harvey PH, Pagel M (1991) The comparative method in evolutionary biology. Oxford University
Press, Oxford
Hazan A, Gerson U, Tahori AS (1973) Life history and life tables of the carmine spider mite.
Acarologia 15:414–440
Hazan A, Gerson U, Tahori AS (1974) Spider mite webbing. I. The production of webbing under
various environmental conditions. Acarologia 16:68–84
Helle W (1967) Fertilization in the two-spotted spider mite (Tetranychus urticae: Acari). Entomol
Exp Appl 10:103–110
Helle W, Overmeer WPJ (1973) Variability in tetranychid mites. Annu Rev Entomol 18:97–120
Helle W, Pijnacker LP (1985) Parthenogenesis, chromosomes and sex ratio. In: Helle W, Sabelis
MW (eds) Spider mites. Their biology, natural enemies and control, vol 1A. Elsevier,
Amsterdam, pp 129–139
Helle W, Sabelis MW (eds) (1985) Spider mites. Their biology, natural enemies, and control, vols
1A, 1B. Elsevier, Amsterdam
Henter H (2003) Inbreeding depression and haplodiploidy: experimental measures in a parasitoid
and comparisons across diploid and haplodiploid insect taxa. Evolution 57:1793–1803.
Herre EA (1985) Sex ratio adjustment in fig wasps. Science 228:896–898
Hines W, Maynard Smith J (1979) Games between relatives. J Theor Biol 79:19–30
Horita M, Chittenden AR, Sato Y, Saito Y (2004) Function of web box as an anti-predator barrier
in the spider mite, Schizotetranychus recki. J Ethol 22:105–108
Hoy M (1985) Almonds (California). In: Helle W, Sabelis MW (eds) Spider mites. Their biology,
natural enemies and control, vol 1B. Elsevier, Amsterdam, pp 299–310
Hoy MA, Groot JJR, van der Baan HE (1985) Influence of aerial dispersal on persistence and
spread of pesticide-resistant Metaseiulus occidentalis. Entomol Exp Appl 37:117–131
Huffaker CB (1958) Experimental studies on predation: dispersion factors and predator–prey
oscillations. Hilgardia 27:343–383
Huffaker CB, Shea KP, Herman SG (1963) Experimental studies on predation: complex dispersion
and levels of food in an acarine predator–prey interaction. Hilgardia 34:305–329
Ims RA, Yoccoz NG (1997) Studying transfer processes in metapopulations: emigration, migra-
tion, and colonization. In: Hanski IA, Gilpin ME (eds) Metapopulation biology: ecology,
genetics, and evolution. Academic Press, San Diego, pp 247–266
Ito K, Saito Y (2006) Effects of host-plant species on diapause induction of the Kanzawa spider
mite (Tetranychus kanzawai Kishida). Entomol Exp Appl 121:177–184
Ito Y (1993) Behaviour and social evolution of wasps: the communal aggregation hypothesis.
Oxford University Press, Oxford
Jackson DE, Hart AG (2009) Does sanitation facilitate sociality? Anim Behav 77:e1–e5
(doi:10.1016/j.anbehav.2008.09.013)
Jeppson LR, Keifer HH, Baker EW (1975) Mites injurious to economic plants. University of
California Press, Berkeley.
References 167
Maynard Smith J (1982a) Evolution and the theory of games. Cambridge University Press, Cambridge
Maynard Smith J (1982b) The evolution of social behavior – a classification of models. In: King’s
College Sociobiology Group, Cambridge (ed) Current problems in sociobiology. Cambridge
University Press, London, pp 29–44
Mayr E (1963) Animal species and evolution. Harvard University Press, Cambridge
McArthur RH, Wilson EO (1967) The theory of island biogeography. Princeton University Press,
Princeton
McEnroe WD (1969) The photoresponse of the two-spotted spider mite, Tetranychus urticae
Koch, to skyshine. Acarologia 11:281–283
McEnroe WD, Dronka K (1969) Eyes of the two-spotted spider mite. II. Behavioural analysis of
the photoreceptors. Ann Entomol Soc Am 62:466–469
McGregor EA (1950) Mites of the family Tetranychidae. Am Midl Nat 44:257–420
McMurtry JA (1985) Avocado. In: Helle W, Sabelis MW (eds) Spider mites. Their biology, natural
enemies and control, vol 1B. Elsevier, Amsterdam, pp 327–332
McMurtry JA, Croft BA (1997) Life-styles of phytoseiid mites and their roles in biological con-
trol. Annu Rev Entomol 42:291–321
McMurtry JA, Huffacker CB, Van de Vrie M (1970) Ecology of tetranychid mites and their natu-
ral enemies: their biological characters and the impact of spray practices. Hilgardia 40:
331–390
McMurtry JA, Johnson HG (1966) An ecological study of the spider mite Oligonychus punicae
(Hirst) and its natural enemies. Hilgardia 37:363–402
Michod RE, Hamilton WD (1980) Coefficient of relatedness in sociobiology. Nature (Lond)
288:694–697
Mills LR (1973) Morphology of glands and ducts in the two-spotted spider mite, Tetranychus
urticae Koch, 1836. Acarologia 15:218–236
Mitchell R (1973) Growth and population dynamics of a spider mite (Tetranychus urticae K.,
Acarina: Tetranychidae). Ecology 54:1349–1355
Mochizuki M, Takafuji A (1996) Effect of photoperiod and temperature on the diapause ter-
mination of the Kanzawa spider mite, Tetranychus kanzawai Kishida (Acari: Tetranychidae).
J Acarol Soc Jpn 5:83–88
Montserrat M, Peña FDL, Hormaza JI, González-Fernández JJ (2008) How do Neoseiulus califor-
nicus (Acari: Phytoseiidae) females penetrate densely webbed spider mite nests? Exp Appl
Acarol 44:101–106
Moraes DE, McMurtry JA, Denmark HA, Campos CB (2004) A revised catalog of the mite family
Phytoseiidae. ZOOTAXA 434, Magnolia Press, Auckland
Mori H, Saito Y, Tho Y (1999a) Cooperative group predation in a sit-and-wait cheyletid mite. Exp
Appl Acarol 23:643–651
Mori K (2000) Factors which caused variation in social systems of spider mites, Schizotetranychus
celarius species group (Acari: Tetranychidae). PhD dissertation, Hokkaido University,
Sapporo
Mori K, Saito Y (2004) Nest size variation reflecting anti-predator strategies in social spider mites
of Stigmaeopsis (Acari: Tetranychidae). Behav Ecol Sociobiol 56:201–206
Mori K, Saito Y (2005) Variation in social behavior within a spider mite genus, Stigmaeopsis
(Acari: Tetranychidae). Behav Ecol 16:232–238
Mori K, Saito Y (2006) Communal relationships in a social spider mite, Stigmaeopsis longus (Acari:
Tetranychidae): an equal share of labor and reproduction between nest mates. Ethology
112:1–9
Mori K, Saito Y, Sakagami T (1999b) Effects of nest web and female attendance on survival of
young in a subsocial spider mite, Schizotetranychus longus (Acari; Tetranychidae). Exp Appl
Acarol 23:411–418
Mori K, Saito Y, Sakagami T, Sahara K (2005) Inbreeding depression of female fecundity by
genetic factors retained in natural populations of a male haploid social mite (Acari:
Tetranychidae) Exp Appl Acarol 36:15–23
Morimoto K, Furuichi H, Yano S, Osakabe M (2006) Web-mediated interspecific competition
among spider mites. J Econ Entomol 99:678–684
References 169
Müller JK, Braunisch V, Hwang W, Eggert A-K (2006) Alternative tactics and individual repro-
ductive success in natural associations of the burying beetle, Nicrophorus vespilloides. Behav
Ecol 18:196–203
Nagelkerke CJ (1993) Evolution of sex allocation strategies of pseudo-arrhenotokous predatory
mites (Acari: Phytoseiidae). PhD dissertation, University of Amsterdam, Amsterdam
Nagelkerke CJ, Sabelis MW (1996) Hierarchical levels of spatial structure and their consequences
for the evolution of sex allocation in mites and other arthropods. Am Nat 148:16–39
Navajas M, Gutierrez J, Lagnel J (1996) Mitochondrial cytochrome oxidase I in tetranychid mites:
a comparison between molecular phylogeny and changes of morphological and life history
traits. Bull Entomol Res 86:407–417
Nickel JL (1960) Temperature and humidity relationships of Tetranychus desertorum Banks with
special reference to distribution. Hilgardia 30:41–100
Nishiyama R, Wada Y, Mibu M, Yamaguchi Y, Shimogawara K, Sano H (2004) Role of a nonse-
lective de novo DNA methyltransferase in maternal inheritance of chloroplast genes in the
green alga, Chlamydomonas reinhardtii. Genetics 168:809–816
Norton RA, Kethley JB, Johnston DE, O’Connor BM (1993) Phylogenetic perspectives on genetic
systems and reproductive modes of mites. In: Wrensch DL, Ebbert MA (eds) Evolution and
diversity of sex ratio in insects and mites. Chapman & Hall, New York, pp 8–99
Okabe K, Makino S (2008) Parasitic mites as part-time bodyguards of a host wasp. Proc R Soc B
275:2293–2297
Oku K, Yano S, Osakabe M, Takafuji A (2005) Mating strategies of Kanzawa spider mite,
Tetranychus kanzawai Kishida (Acari: Tetranychidae), in relation to the mating status of
females. Ann Entomol Soc Am 98:625–628
Oku K, Yano S, Takafuji A (2003) Spider mite’s use of a refuge during the quiescent stage in the
presence of a predator. The Netherlands Entomological Society. Entomol Exp Appl 108:71–74
Oku K, Yano S, Takafuji A (2006) Host plant acceptance by the phytophagous mite Tetranychus
kanzawai Kishida is affected by the availability of a refuge on the leaf surface. Ecol Res
21:446–452
Osakabe M, Hongo K, Funayama K, Osumi S (2006) Amensalism via webs causes unidirectional
shifts of dominance in spider mite communities. Oecologia (Berl) 150:496–505
Osakabe M, Saito Y (1991) Morphological differences between the diapausing and non-diapaus-
ing forms of the citrus red mite, Panonychus citri (McGregor) (Acari: Tetranychidae). Bull
Fruit Tree Res Stn 21:95–115
Osakabe M, Saito Y, Sakagami Y (1993) Protein differences detected by two-dimensional electro-
phoresis in the species complex of Schizotetranychus celarius (Banks). Exp Appl Acarol
17:757–764
Osakabe M, Sakagami Y (1994) RFLP analysis of ribosomal DNA in sibling species of spider
mite, genus Panonychus (Acari: Tetranychidae). Insect Mol Biol 3:63–66
Overmeer WPJ (1985) Rearing and handling. In: Helle W, Sabelis MW (eds) Spider mites. Their
biology, natural enemies and control, vol 1B. Elsevier, Amsterdam, pp 161–170
O’Neill SL, Hoffmann AA, Werren JH (1997) Influential passengers: inherited microorganisms
and arthropod reproduction. Oxford University Press, New York
Pallini A, Janssen A, Sabelis MW (1997) Odour-mediated responses of phytophagous mites to
conspecific and heterospecific competitors. Oecologia (Berl) 110:179–185
Penman DR, Cone WW (1972) Behaviour of male two-spotted spider mites in response to quies-
cent female deutonymphs and to web. Ann Entomol Soc Am 65:1289–1293
Potter DA, Wrensch DL, Johnston DE (1976) Guarding, aggressive behaviour and mating success
in male two-spotted spider mites. Ann Entomol Soc Am 69:707–711
Pritchard AE, Baker EW (1955) A revision of the spider mite family Tetranychidae. Pac Coast
Entomol Soc Mem Ser 2:1–472
Prokop Z, Stuglik M, Radwan J (2007) Male age, mating probability, and progeny fitness in the
bulb mite. Behav Ecol 18:597–601
Queller DC (1985) Kinship, reciprocity and synergism in the evolution of social behaviour. Nature
(Lond) 318:366–367
Queller DC (1994) Genetic relatedness in viscous populations. Evol Ecol 8:70–73
170 References
Radwan J (1993) The adaptive significance of male polymorphism in the acarid mite Caloglyphus
berlesei. Behav Ecol Sociobiol 33:201–208
Radwan J (1995) Male morph determination in two species of acarid mites. Heredity 74:669–673
Radwan J (2007) Sexual selection and conflict in the bulb mite, Rhizoglyphus robini (Astigmata:
Acaridae). Exp Appl Acarol 42:151–158
Radwan J, Unrug J, Tomkins JL (2002) Status-dependence and morphological trade-offs in the expres-
sion of a sexually selected character in the mite, Sancassania berlesei. J Evol Biol 15:744–752
Rayor LS, Taylor LA (2006) Social behavior in amblypygids, and a reassessment of arachnid
social patterns. J Arachnol 34:399–421
Ridley M (1983) The explanation of organic diversity: The comparative method and adaptations
for mating. Clarendon Press, Oxford.
Ridley M (1985) The problem of evolution. Oxford University Press, Oxford
Roeder C (1992) Sex ratio response of the two-spotted spider mite (Tetranychus urticae Koch) to
changes in density under local mate competition. Can J Zool 70:1965–1967
Roeder C, Harmsen R, Mouldey S (1996) The effects of relatedness on progeny sex ratio in spider
mites. J Evol Biol 9:143–151
Sabelis MW (1985) Reproductive strategies. In: Helle W, Sabelis MW (eds) Spider mites. Their
biology, natural enemies and control, vol 1A. Elsevier, Amsterdam, pp 265–278
Sabelis MW, Bakker FM (1992) How predatory mites cope with the web of their tetranychid prey:
a functional view on dorsal chaetotaxy in the Phytoseiidae. Exp Appl Acarol 16:203–255
Sabelis MW, Dicke M (1985) Long-range dispersal and searching behaviour. In: Helle W, Sabelis
MW (eds) Spider mites. Their biology, natural enemies and control, vol 1B. Elsevier,
Amsterdam, pp 141–160
Sabelis MW, Janssen A (1993) Evolution of life-history patterns in the Phytoseiidae. In: Houck
MA (ed) Mites. Ecological and evolutionary analyses of life-history patterns. Chapman &
Hall, New York, pp 70–98
Sabelis MW, van de Baan HE (1983) Location of distant spider mite colonies by phytoseiid preda-
tors: Demonstration of specific kairomones emitted by Tetranychus urticae and Panonychus
ulmi. Entomol Exp Appl 33:303–314
Sabelis MW, Vermaat JE, Groeneveld A (1984) Arrestment responses of the predatory mite,
Phytoseiulus persimilis, to steep odour gradients of a kairomone. Physiol Entomol 9:437–446
Saito Y (1977a) Study on the spinning behavior of the spider mite (Acarina: Tetrnychidae).
I. Method for the quantitative evaluation of the mite webbing, and the relationship between
webbing and walking. Jpn J Appl Entomol Zool 21:27–34 (in Japanese with English summary)
Saito Y (1977b) Study on the spinning behavior of the spider mite (Acarina: Tetrnychidae). II. The
process of silk secretion and the webbing on the various stages of Tetranychus urticae Koch
under different environmental conditions. Jpn J Appl Entomol Zool 21:150–157 (in Japanese
with English summary)
Saito Y (1979a) Comparative studies on life histories of three species of spider mites (Acrina:
Tetranychidae). Appl Entomol Zool 14:83–94
Saito Y (1979b) Study on spinning behavior of spider mites. III. Response of mites to webbing
residues and their preferences for particular physical conditions of leaf surface (Acarina:
Tetranychidae). Jpn J Appl Entomol Zool 23:82–91 (in Japanese with English summary)
Saito Y (1983) The concept of “life types” in Tetranychinae. An attempt to classify the spinning
behaviour of Tetranychinae. Acarologia 24:377–391
Saito Y (1985) Life types of spider mites. In: Helle W, Sabelis MW (eds) Spider mites. Their biol-
ogy, natural enemies and control, vol 1A. Elsevier, Amsterdam, pp 253–264
Saito Y (1986a) Biparental defence in a spider mite (Acari: Tetranychidae) infesting Sasa bamboo.
Behav Ecol Sociobiol 18:377–386
Saito Y (1986b) Prey kills predator: counter-attack success of a spider mite against its specific
phytoseiid predator. Exp Appl Acarol 2:47–62
Saito Y (1987) Extraordinary effects of fertilization status on the reproduction of an arrhenotokous
and sub-social spider mite (Acari: Tetranychidae). Res Popul Ecol 29:57–71
Saito Y (1990a) Two new spider mite species of the Schizotetranychus celarius complex (Acari:
Tetranychidae). Appl Entomol Zool 25:389–396
References 171
Saito Y (1990b) Life history and food habit of Typhlodromus bambusae Ehara, a specific predator
of Schizotetranychus celarius (Banks) (Acari: Phytoseiidae, Tetranychidae). Exp Appl Acarol
10:45–51
Saito Y (1990c) “Harem” and “non-harem” type mating systems in two species of subsocial spider
mites (Acari, Tetranychidae). Res Popul Ecol 32:1–16
Saito Y (1990d) Factors determining harem ownership in a sub-social spider mite (Acari,
Tetranychidae). J Ethol 8:37–43
Saito Y (1990e). Natural enemies counterattacked – in acari predator–prey interactions. In:
Kiritani K, Shiga M (eds) Ecology of natural enemies. Tokai University Press, Tokyo,
pp 38–45 (in Japanese and Korean)
Saito Y (1994a) Do males of Schizotetranychus miscanthi (Acari, Tetranychidae) recognize kin in
male competition? J Ethol 12:15–18
Saito Y (1994b) Is sterility by deleterious recessives an origin of inequalities in the evolution of
eusociality? J Theor Biol 166:113–115
Saito Y (1995a) Clinal variation in male-to-male antagonism and weaponry in a subsocial mite.
Evolution 49:413–417
Saito Y (1995b) Sociobiological aspects of spider mite life types. J Acarol Soc Jpn 4:55–67
Saito Y (1995c) “Altruistic gene” a priori involves two different deleterious gene effects. A reply
to Keller. J Theor Biol 174:471–472
Saito Y (1997) Sociality and kin selection in Acari. In: Choe JC, Crespi BJ (eds) The evolution of
social behaviour in insects and arachnids. Cambridge University Press, Cambridge, pp 443–457
Saito Y (2000) Do kin selection and intra-sexual selection operate in spider mites? Exp Appl
Acarol 24:351–363
Saito Y, Chittenden AR, Mori K, Ito K, Yamauchi A (2008a) An overlooked side effect of nest
scattering behavior to decrease predation risk (Acari; Tetranychidae, Stigmaeidae). Behav Ecol
Sciobiol 63:33–42
Saito Y, Ito K, Sakagami T (2005) Imaginal induction of diapause in several ‘adult-female dia-
pausing’ spider mites. Physiol Entomol 30:96–101
Saito Y, Mori H (1975) The effects of pollen as an alternative food for three species of phytoseiid
mites (Acarina: Phytoseiidae). Mem Fac Agric Hokkaido Univ 9:236–246
Saito Y, Mori K (2005) Where does male-to-male “aggression” compromise “cooperation”? Popul
Ecol 47:159–166
Saito Y, Mori K, Chittenden AR, Sato Y (2000a) Correspondence of male-to-male aggression to spatial
distribution of individuals in field populations of a subsocial spider mite. J Ethol 18:79–83
Saito Y, Mori K, Sakagami T, Lin J (2004) Reinstatement of the genus Stigmaeopsis Banks, with
descriptions of two new species (Acari, Tetranychidae). Ann Entomol Soc Am 97:635–646
Saito Y, Osakabe MH, Yasui Y (1993) A method for preparing permanent specimens of mites with
Canada balsam. Appl Entomol Zool 28:593–597
Saito Y, Sahara K (1999) Two clinal trends in male-male aggressiveness in a subsocial spider mite
(Schizotetranychus miscanthi). Behav Ecol Sociobiol 46:25–29
Saito Y, Sahara K, Mori K (2000b) Inbreeding depression by recessive deleterious genes affecting
female fecundity of a haplo-diploid mite. J Evol Biol 13:668–678
Saito Y, Sakagami T, Sahara K (2002) Differences in diapause attributes between two forms dis-
tinguished by male-to-male aggression in a subsocial spider mite, Schizotetranychus miscanthi
Saito. Ecol Res 17:645–653
Saito Y, Suzuki R (1987) Reexamination of several rearing methods for studying the life history
of spider mites (Acari: Tetranychidae). Appl Entomol Zool 22:570–576
Saito Y, Takada T (2009) Conditions under which male-to-male aggression compromises coopera-
tion in a game without learning. Open Evol J 3:1–8
Saito Y, Takahashi K (1980) Study on variation of Schizotetranychus celarius (Banks) I.
Preliminary descriptions of morphological and life type variation. Jpn J Appl Entomol Zool
24:62–70 (in Japanese with English summary)
Saito Y, Takahashi K (1982) Study on variation of Schizotetranychus celarius (Banks). II.
Comparison of mode of life between two sympatric forms (Acarina: Tetranychidae). Jpn J Ecol
32:69–78 (in Japanese with English summary)
172 References
Saito Y, Tsuji N, Chittenden AR (2008b) Sustainable pest management and moso bamboo forest
conservation: a preliminary report. Proceedings of ICSA 2008, pp 309–316
Saito Y, Ueno J (1979) Life history studies on Schizotetranychus celarius (Banks) and Aponychus
corpuzae Rimando as compared with other tetranychid mite species (Acarina: Tetranychidae).
Appl Entomol Zool 14:445–452
Saito Y, Urano S, Nakao H, Amimoto K, Mori H (1996) A simulation model for predicting the
efficiency of biological control of spider mites by phytoseiid predators. 1. The model. Jpn J
Appl Entomol Zool 40:103–111 (in Japanese with English summary)
Sakagami T (2002) Phylogenetic analysis of subfamily Tetranychinae (Acari: Tetranychidae) in
Japan, based on 28S ribosomal DNA sequences: does the Celarius species group belong to the
genus Schizotetranychus? (in Japanese). PhD dissertation, Hokkaido Univ, Sapporo
Sakagami T, Saito Y, Kongchuensin M, Sahara K (2009) Molecular phylogeny of Stigmaeopsis with
special reference to speciation through host-plant shift. Ann Entomol Soc Am 102:360–366
Santos MA, Laing JE (1985) Stigmaeid predators. In: Helle W, Sabelis MW (eds) Spider mites.
Their biology, natural enemies and control, vol 1B. Elsevier, Amsterdam, pp 197–203
Sato Y, Saito Y (2006) Nest sanitation in social spider mites: interspecific differences in defeca-
tion behavior. Ethology 112:664–669
Sato Y, Saito Y (2007a) Can the extremely female-biased sex ratio of the social spider mites be
explained by Hamilton’s LMC? Ecol Entomol 32:597–602
Sato Y, Saito Y (2007b) Evolutionary view of waste management behavior using volatile chemical
cues in social spider mites. J Ethol 26:267–272
Sato Y, Saito Y, Chittenden AR (2008) The parapatric distribution and contact zone of two forms
showing different male-to-male aggressiveness in a social spider mite, Stigmaeopsis miscanthi
(Acari: Tetranychidae). Exp Appl Acarol 44:265–276
Sato Y, Saito Y, Mori K (2000) Patterns of reproductive isolation between two groups of
Schizotetranychus miscanthi Saito (Tetranychidae: Acari) showing different male aggression
traits. Appl Entomol Zool 35:611–618
Sato Y, Saito Y, Sakagami T (2003) Rules for nest sanitation in a social spider mite,
Schizotetranychus miscanthi Saito (Acari: Tetranychidae). Ethology 109:713–724
Savory T (1964) Arachinidae. Academic Press, London
Schausberger P (1999) Juvenile survival and development in Euseius finlandicus, Typhlodromus
pyri and Kampimodromus aberrans (Acari: Phytoseiidae) feeding on con- and heterospecific
immature life stages. Exp Appl Acarol 23:389–398
Schmidt G (1976) Der einfluss der von den beutetieren hinterlassenen spuren auf suchverhalten
und suchherfolg von Phytoseiulus persimilis A.H. (Acarina: Phytoseiidae). Z Angew Entomol
82:216–218 (in German)
Schmidt JO (1990) Evolution of major defensive ensembles. In: Evans DL, Schmidt JO (eds)
Insect defences. State University of New York Press, Albany, pp 1–2
Schmidt-Nielsen K (1984) Scaling: why is animal size so important? Cambridge University Press,
Cambridge
Shimoda T, Kishimoto H, Takabayashi J, Amano H, Dicke M (2008) Comparison of thread-cut-
ting behavior in three specialist predatory mites to cope with complex webs of Tetranychus
spider mites. Exp Appl Acarol (doi 10.1007/s10493-008-9205-3)
Shinkaji N (1979) Geographical distribution of the citrus red mite, Panonychus citri and European
red mite, P. ulmi in Japan. In: Rodriguez JG (ed) Recent advances in acarology, vol 1.
Academic Press, New York
Smith RH, Shaw MR (1980) Haploid sex ratios and the mutation rate. Nature (Lond) 287:728–729
Southwood TRE (1977) Habitat, the template for ecological strategies? J Anim Ecol 46:337–365
Stouthamer R, Breeuwer JAJ, Hurst GDD (1999) Wolbachia pipientis: microbial manipulator of
arthropod reproduction. Annu Rev Microbiol 53:71–102
Sugden R (1986) The economics of rights, co-operation and welfare. Blackwell, London
Suzuki S (1978) Index to Japanese Bambusaceae. Gakken, Tokyo, Japan
Takabayashi J, Dicke M (1996) Plant-carnivore mutualism through herbivore-induced carnivore
attractants. Trends Plant Sci 1:109–113
References 173
Takafuji A, So PM, Tuno N (1991) Inter- and intra-population variations in diapause attribute of
the two-spotted spider mite, Tetranychus urticae Koch, in Japan. Res Popul Ecol 33:331–344
Takahashi K (1987) Studies on morphological and ecological variations of Schizotetranychus
celarius (Banks). PhD dissertation, Hokkaido University, Sapporo
Tallamy DW, Wood TK (1986) Convergence patterns in subsocial insects. Annu Rev Entomol
31:369–390
Tansky M (1978) Switching effect in prey-predator system. J Theor Biol 70:263–271
Taylor PD (1992) Altruism in viscous populations – an inclusive fitness model. Evol Ecol
6:352–356
Taylor PD, Bulmer MG (1980) Local mate competition and the sex ratio. J Theor Biol 86:409–419
Taylor RJ (1977) The value of clumping to prey. Oecologia (Berl) 30:285–294
Tilszer M, Antoszczyk K, Saiek N, Zajac E, Radwan J (2006) Evolution under relaxed sexual
conflict in the bulb mite Rhizoglyphus robini. Evolution 60:1868–1873
Tomkins JL, Lebas NR, Unrug J, Radwan J (2004) Testing the status-dependent ESS model: popu-
lation variation in fighter expression in the mite Sancassania berlesei. J Evol Biol
17:1377–1388
Treat AE (1958) Social organization in the moth ear mite (Myrmonyssus phalaenodectes). Proc
10th Int Cong Entomol 2:475–480
Treat AE (1975) Mites of moths and butterflies. Cornell University Press/Comstock, Ithaca, NY
Trivers R (1985) Social evolution. Benjamin & Cummings, Menlo Park, CA
Tsubaki Y, Siva Jothy MT, Ono T (1994) Re-copulation and post-copulatory mate guarding
increase immediate female reproductive output in the dragonfly Nannophya pygmeae Rambur.
Behav Ecol Sociobiol 35:219–225
Tsuji K (1990) Reproductive division of labor related to age in the Japanese queenless ant,
Pristomymex pungens. Anim Behav 39:843–849
Uetz GW, Hieber CS (1997) Colonial web-building spiders: balancing the costs and benefits of
group-living. In: Choe JC, Crespi BJ (eds) The evolution of social behavior in insects and
arachnids. Cambridge University Press, Cambridge, pp 476–498
Usher MB, Davis PR (1983) The biology of Hypoaspis aculeifer (Canestrini) (Mesostigmata): is
there a tendency towards social behaviour? Acarologia 24:243–250
Van Eydhoven GL, Helle W (1966) Sex abnormalities in the common spider mite (Tetranychus
urticae) (Acari, Tetranychidae). Entomol Ber 26:204–208
Van Hoff JARAM (1990) Intergroup competition and conflict in animals and man. In: Van der
Dennen J, Falger V (eds) Sociobiology and conflict. Evolutionary perspectives on competition,
cooperation, violence and warfare. Chapman & Hall, London, pp 23–54
Van de Vrie M, Boersma A (1970) The influence of the predacious mite Typhlodromus (A.) poten-
tillae (Garman) on the development of Panonychus ulmi (Koch) on apple grown under various
nitrogen conditions. Entomophaga 15:291–304
Van de Vrie M, McMurtry JA, Huffaker CB (1972) Ecology of tetranychid mites and their natural
enemies: a review. III. Biology, ecology, and pest status, and host-plant relations of tetrany-
chids. Hilgardia 41:343–432
Van Wilgenburg E, Driessen G, Beukeboom LW (2006) Single locus complementary sex determi-
nation in Hymenoptera: an “unintelligent” design? Front Zool 3:1–15
Veerman A (1977) Photoperiodic termination of diapause in spider mites. Nature (Lond)
266:526–527
Veerman A (1985) Diapause. In: Helle W, Sabelis MW (eds) Spider mites. Their biology, natural
enemies and control, vol 1A. Elsevier, Amsterdam, pp 279–316
Volterra V (1926) Variazioni e fluttuazioni del numero d’individui in specie animali conviventi.
Mem R Accad Naz dei Lincei Ser VI 2:657–659
Vulinec K (1990) Collective security: aggregation by insects as a defense. In: Evans DL, Schmidt
JO (eds) Insect defenses: adaptive mechanisms and strategies of prey and predators. State
University of New York Press, Albany, pp 251–288
Waage JK (1979) Adaptive significance of postcopulatory guarding of mates and nonmates by
male Calopteryx maculata (Odonata). Behav Ecol Sociobiol 6:147–154
174 References
Walter DE (1996) Living on leaves: mites, tomenta, and leaf domatia. Annu Rev Entomol
41:101–114
Walter DE, Proctor HC (1999) Mites: ecology, evolution and behaviour. CABI Publishing, Oxon
Wanibuchi K, Saito Y (1983) The process of population increase and patterns of resource utiliza-
tion of two spider mites, Oligonychus ununguis (Jacobi) and Panonychus citri (McGregor),
under experimental conditions (Acari: Tetranychidae). Res Popul Ecol 25:116–129
Weeks A R, Marec F, Breeuwer JAJ. (2001) A mite species that consists entirely of haploid
females. Science 292:2479–2482
Werren JH, Colwell RK (1987) Labile sex ratios in wasps and bees. Bioscience 37:498–506
West SA, Murray MG, Machado CA, Griffin AS, Herre EA (2001) Testing Hamilton’s rule with
competition between relatives. Nature (Lond) 409:510–513
West SA, Pen I, Griffin AS (2002) Cooperation and competition between relatives. Science 296:72–75
Williams GC (1966) Adaptation and natural selection. A critique of some current evolutionary
thought. Princeton University Press, Princeton
Wilson EO (1975) Sociobiology. Belknap Press of Harvard University Press, Cambridge
Wrensch DL (1993) Evolutionary flexibility through haploid males or how chance favors the
prepared genome. In: Wrensch DL, Ebbert MA (eds) Evolution and diversity of sex ratio in
insects and mites. Chapman & Hall, New York, pp 118–149
Wrensch DL, Kethley JB, Norton RA (1994) Cytogenetics of holokinetic chromosomes and
inverted meiosis: keys to the evolutionary success of mites, with generalization on eukaryotes.
In: Houck MA (ed) Mites. Chapman & Hall, New York, pp 282–343
Wright S (1922) Coefficient of inbreeding and relationship. Am Nat 56:330–338
Wynne-Edwards VC (1962) Animal dispersion in relation to social behaviour. Oliver & Boyd,
Edinburgh & London
Xu L (1989) The biology of Rhizomys pruinosus. Acta Theriol Sin 4:100–105 (in Chinese: translated
into English by W. Downs, Bilby Research Centre, Northern Arizona University, June 1989)
Yamamura N (2007) Conditions under which plants help herbivores and benefit from predators
through apparent competition. Ecology 88:1593–1599
Yanagida H, Saito Y, Mori K, Chittenden AR (2001) Egg-depositing behavior as a predator avoid-
ance tactic of Yezonychus sapporoensis. J Ethol 19:63–66
Yano S (2008) Collective and solitary behaviors of two-spotted spider mite (Acari: Tetranychidae)
are induced by trail following. Ann Entomol Soc Am 101:247–252
Yano S, Takafuji A (2002) Variation in the life history pattern of Tetranychus urticae (Acari:
Tetranychidae) after selection for dispersal. Exp Appl Acarol 27:1–10
Yasui Y (1988) Sperm competition of Macrocheles muscaedomesticae (Scopoli) (Acarina:
Mesostigmata: Macrochelidae), with special reference to precopulatory mate guarding behavior.
J Ethol 6:83–90
Yokoyama K (1932) New tetranychid mites attacking the mulberry leaf. Contribution I. Bionomics
and external structures of Tetranychus suginamensisi n. sp. Bull Imp Sericult Exp Sta
8:229–287, 2 pls (in Japanese with English summary)
Young SS, Wrensch DL, Kongchuensin M (1986) Control of sex ratio by female spider mite.
Entomol Exp Appl 40:53–60
Zhang YX (2002) Studies on factors causing outbreaks of pest mites in bamboo forests in Fujian,
China. PhD dissertation, Hokkaido University, Sapporo
Zhang YX, Zhang QZ (eds) (2000) Biology and control of bamboo mites in Fujian. Special pub-
lication. Systematics & Applied Acarology Society, London
Zhang YX, Zhang ZQ, Saito Y, Liu QY, Ji J (2004) On the causes of mite pest outbreaks in mono-
and poly-cultured moso bamboo forests. Chin J Appl Ecol 15:1161–1165 (in Chinese with
English abstract)
Zhang YX, Zhang ZQ, Tong LX, Liu QY, Song MG (2000) Causes of pest mite outbreaks in
bamboo forests in Fujian, China: analyses of mite damage in monoculture versus polyculture
stands. Syst Appl Acarol Spec Publ 4:93–108
Zhang ZQ, BA Croft (1994) A comparative life history study of immature Amblyseius fallacis,
Amblyseius andersoni, Typhlodromus occidentalis, and Typhlodromus pyri (Acari: Phytoseiidae)
with a review of larval feeding patterns. Exp Appl Acarol 18:631–657
Index
A Amblyseius hibisci, 31
Acari, 17, 73, 77, 98 Amblyseius orientalis (Am. orientalis), 6, 29,
Acaricide, 19 35, 36, 40, 42, 43, 46, 47
Acaridae, 6 Amblyseius sp., 2, 6, 35, 36, 40, 46, 54
Acarid mite, 77, 93 Ambulatory movement, 18
Acariformes, 1 Amicability, 111
Acquisition of cold-hardiness, 28 Amphipods, 74
Actinedida (Prostigmata), 1, 94 Amphitetranychus, 8, 9
Actor male, 128 Amphitetranychus quercivorus, 13
Adult cooperation, 14, 15 Amphitetranychus viennensis, 13, 25
Advantage of male cooperation, 121 Ancestor, 91, 92
Aedeagus, 77, 78 Animal parasite, 98
Aggregated life, 12 Annual mean temperature, 115
Aggregation, 93, 94 Antagonistic behavior, 125
Aggression, 117, 126–128, 130–134, 160 Antiherbivore device, 37
behavior, 109, 133 Antipredator
-contest, 103 adaptations, 11
female, 103 apparatus, 46
populations, 119 behavior, 44, 46, 63
strategy, 136, 137 functions, 15, 44, 46–48
traits, 113 strategy, 42–44, 52–54, 66
Aggressiveness, 111, 113–115, 117, 118, 120, traits, 42, 47
125, 130, 133, 134 Antithesis against the functions of
Agistemus iburiensis (Ag. iburiensis), 6, 40, aggregation, 60
46, 47, 54, 57–59, 61, 97, 150 Anystidae, 6, 17, 38
Agistemus summersi (Ag. summersi), 6, 38, 40, Apomixy, 71
43, 46, 47, 54, 67 Aponychus, 8, 9
Allele, 72 Aponychus corpuzae (Ap. corpuzae), 6, 12, 13,
Allodynerus delphinalis, 37 23, 25, 40–44, 47, 53, 68, 74, 79
Alnus hirsuta, 14 Aponychus firmianae, 12, 13
Alnus japonica, 14 Apparent haploid-thelytoky, 71
Alternative models, 103 Appendage, 96
Altruism, 96 Arboreal Oribatidae, 6
behavior, 98, 103, 133 Arms race, 34
trait, 85 Arrhenotoky, 71, 72
Ambient temperature, 121 Arthropod diversity, 149
Amblypygi, 98 Arthropods, 96
Amblyseius ainu, 40 Asciid, 94
Amblyseius eharai, 35, 36 Astigmata, 2
175
176 Index
H I
Hamilton’s local mate competition, 80 Immigration success, 146
Hamilton’s rule, 128 Inbred females, 71
Hamilton’s rule of altruism, 127 Inbreeding, 105, 108, 112, 122
Haplo-diploid, 71, 80, 100, 105, 107, 111 depression, 73, 100, 105–108
mite, 73 effects, 105–108
organism, 108, 137 model, 84
species, 107 nature, 103
spider mite, 73 tolerance to, 108
Haplo-diploidy, 71, 72, 74, 80, 81, 105–108 Wright’s coefficient, 139
Haploid, 71, 72, 105 Inclusive fitness matrix, 131
Haploid-thelytoky, 72 Inclusive fitness theory, 34, 64, 99, 118,
Harems, 109, 111 127–129, 134
Hawk, 126 Incomplete dominance gene, 67
Haystack model, 80–82, 102–104 Individual-based (1-locus and 2-allele) model,
Haystack with unlimited growth model, 82 82
Hemicheyletia morii (Hm. morii), 6, 94–96, Individual fitness, 127, 128
100, 102 Individual-level selection, 83
Hemizygote, 72, 73, 83, 105 Individual selections, 126, 127, 134
Herbivorous mites, 15 Inequality, 129
Heterosis, 107 Insect-mite interaction, 34
Heterozygote, 72, 73 Insemination ability, 107
High aggression (HG) form/group, 63, 114, Intensity of intrasexual selection, 121
115, 117, 119, 120, 122–125, 135, 161, Interguild prey, 33
males, 160 Interlineage crossing, 108
High longevity, 38 Intersexes, 73
High probability of predator accumulation, 68 Interspecies variation, 113
Hindering effect on predator, 56 Interspecific predation, 34
Holotype, 158 Intrasexual selection, 111, 125, 134
Homogeneous nature, 134 Intrasexual traits, 111
Homohetero relationship, 73 Intrinsic rate of natural increase, 82
Homo-hetero zygotic relationship, 73 Introgenesis, 154
Homozygote, 72 Inverted meiosis, 71
Honeybee, 96 Iphiseius degenerans, 34–36
Honeybee’s brood, 96 Iteroparity, 100
Honeydew, 34 Ixodida, 72
Hooked claws, 23
Horned beetles, 125
Host plant, 9 K
physical feature, 40 Kairomone odor, 33
shift, 92 Karyotype, 70
spatiotemporal stability, 39 Karyotype-doubling, 73
stability hypothesis, 11, 68 Kin competition, 133
Host spatiotemporal stability, 69 Kin discriminative behavior, 112, 119
Hoyer’s solution, 158 Kin male, 112
Hunting behavior, 96 Kin recognition, 112
180 Index
L
L. See Long seta forms M
Laelapidae, 1 M. See Medium seta forms
Large nest builders, 56, 60–65, 67 Macrocheles muscaedomesticae (M.
Leaf depressions, 12 muscaedomesticae), 6, 74, 77
Leaf disc method, 10 Macrocheles superbus, 97
Leaf domatia, 5, 37, 48 Male aggression, 109, 111, 118, 119, 113–139,
Leaf hair density, 40 154. See also High aggression (HG)
Leaf hairs, 44–46 form/group.
Leaf physical features, 15 Male antagonism, 118
Leaf quality, 28 Male-biased sex ratio, 83
Leg-I length, 125 Male cooperation, 119
Leg-III length, 125 advantage of, 121
Life cycle, 25, 81, 103 Male-haploidy, 96, 97, 100
Life-dinner principle, 146 Male overwintering survival, 121
Life history, 11, 68 Male populations, 119
diversity, 69 Male precedence, 74
evolution, 69 Male pugnacity, 109, 111
parameters, 11 Male relatedness, 113
variation, 39, 68–69 Male relative aggressiveness, 113, 115, 121,
Lifelines, 21 125
Lifeline thread, 68 Male sperm precedence, 77
Life style, 11 Male strategy, 75
Life subtypes, 35, 41 Male-territorial birds, 134
Lifetime fitness, 134 Male-to-male
Lifetime success, 153 aggression, 130, 160
Life type, 39, 68, 92, 3, 11–17 antagonism, 77, 120
Life type variation, 92 behavior, 111
Lion, 133 combat, 75, 76, 111, 125
Little web (LW) type, 11–14, 23, 31,68, 69, competition, 75, 84, 119, 127, 128
114 interaction, 109
Local mate competition (LMC), 78–80, Mate guarding, 73
83, 84 Maternal care, 97
model, 79, 80, 82 Mathematical model, 37
theory, 79 Mating competition, 130
Locus, 72 Mating opportunities, 117, 127–129, 133
Long-daylength, 123 Mating strategy, 78
Longevity, 107 Mating structure, 79, 82
Long-medium (LM) seta forms, 157 Mean lowest daily temperature for winter, 115
Long seta forms, 49, 157 Mean maximum daily temperature, 115
Index 181
S Scrambler male, 75
S. See Short seta forms Scramblers, 77
28 S, 91 Searching efficiency of the predator, 56
Saito–Takada model, 134 Searching time, 159
Saliva, 75 Selection intensities, 134
Same-garden analysis, 113 Selection pressure, 128
Sancassania (Caloglyphus) berlesei, 75 Selective predator, 31
Sasa, 14, 39, 92 Self-defense, 85
Sasa brevisetosus, 70 Self-fertilization, 100
Sasa celarius, 49 Selfish-herd, 60
Sasa-inhabiting mite, 39–40, 70 Semi-isolated natural population, 107
Sasa-inhabiting species, 68 Semi-sterile castes, 108
Sasa-inhabiting spider mite, 40, 69, 70 Semi-sterile female, 108
Sasa kurilensi, 39 Seta length, 49
Sasa nipponica, 39 Sex aberrant, 73
Sasanychus, 9 Sex aberration, 73
Sasanychus akitanus, 13 Sex control, 84
Sasa senanensis, 14, 24, 34, 38, 39, 41 Sex determination, 73
Sato-yama, 150, 151 gene, 73
Sc. See Schizotetranychus celarius mechanism, 71, 73
Scattered nests, 64 Sex gamete, 73
Schizotetranychus, 8, 9, 12, 16, 26, 48, 49, 70, Sex limited, 108
157 Sex mosaic, 73
Schizotetranychus approximatus (S. Sex pheromone, 21, 93
approximatus), 15, 92, 158 Sex ratio, 73, 78–84, 98, 117
Schizotetranychus bambusae, 15, 41, 79 at maturity, 80, 117
Schizotetranychus brevisetosus, 12, 15, 16, 26, primary, 80
92 strategy, 82
Schizotetranychus celarius (Sz. celarius), 39, trait, 117
48–50, 52, 85, 157, 158 Sex-related phenomena, 71
Schizotetranychus celarius long seta form. See Sexual behavior, 71, 73–78, 98
Stigmaeopsis longus (Saito) Sexual reproduction, 82
Schizoteiranychus celarius medium seta form, Sexual selection, 125–127, 153
See Stigmaeopsis takahashii Shelter-like structure, 98
Schizotetranychus celarius short seta form. Short-daylength, 123
See Stigmaeopsis saharai Saito et Mori Short seta forms (S), 49, 157
Schizotetranychus cercidiphylli, 8, 15, 26, 79 Short-term egg care, 93
Schizotetranychus colocasiae, 15 Short-term survival, 56
Schizotetranychus laevidorsatus (S. Sib-cannibalism, 34, 36, 97
laevidorsatus), 15, 92, 158 Sib-individual, 34
Schizotetranychus lespedezae, 8, 26, 79 Sibling species, 52
Schizotetranychus lespedezae (leguminosus), Silk, 3
15 Silken thread, 3, 17, 18, 19, 21, 68, 94
Schizotetranychus longus. See Stigmaeopsis Silk producing organs, 18
longus (Saito) Silk production, 19
Schizotetranychus miscanthi, 7, 52 Simple probability exercise, 58
Schizotetranychus nanjingensis, 158 Sit-and-wait group hunting, 96
Schizotetranychus recki (Sz. recki), 7, 15, 26, Sm, 83
40, 41, 46, 47, 52, 53, 68, 92 Small arthropod, 37
Schizotetranychus saitoi, 15, 92, 93, 158 Smaller nest builder, 56, 59, 65, 67, 89
Schizotetranychus schizopus, 8, 15, 26, 79, 92 Small nests, 54–61
Schizotetranychus tenuinidus, 158 Small, scattered nest, 66
Schooling, 60 Snowfall, 121
Scorpionida, 98 Social behavior, 94, 96
Index 185
Tetranychidae, 1, 6, 7, 9, 13, 14, 17, 34 Typhlodromus bambusae, 7, 32–36, 40, 42, 43,
Tetranychinae, 7, 8, 10, 69, 70, 72, 111 46, 47, 54, 62, 63, 69, 84–86, 97, 150
Tetranychini, 7, 8, 12, 17, 157 Typhlodromus occidentalis, 30, 31
Tetranychus, 8, 9, 12, 22, 23, 26, 27, 29, 32,
70, 141
Tetranychus bambusae, 144–146, 149 U
Tetranychus bambusicola, 13 Ultimate, 34
Tetranychus canadensis, 26 Ultimate factor, 154
Tetranychus cinnabarinus, 79 Unbeatable (optimal) sex ratio, 82
Tetranychus desertorum, 13, 26, 79 Under-snow temperature, 121
Tetranychus ezoensis, 26 Unfertilized female, 100, 101, 112, 115, 138,
Tetranychus fijiensis, 13 139
Tetranychus kanzawai, 13, 22, 26, 28 Unfertilized (unmated) females, 115
Tetranychus lintearius, 21 United nest, 62
Tetranychus mcdanieli, 26 Unpredictable habitat, 83
Tetranychus pacificus, 26
Tetranychus piercei, 13
Tetranychus schoenei, 26 V
Tetranychus turkestani, 26 Valid nests, 56, 57, 59
Tetranychus urticae, 3, 7, 9, 10, 12, 13, 17–24, Variation
26–28, 33, 69, 73–77, 79–81, 83, 111, in life history, 40
115, 122, 123 in overwintering, 25
Tetranychus urticae (cinnabarinus), 23 in Stigmaeopsis species nesting
Teytranychidae Bryobiinae, 72 pattern, 56
Thelytokous, 94 in the aggressiveness, 128
Thelytokous reproduction, 94, 96 Varroa jacobsoni, 7, 96, 100, 102
Thelytoky, 71, 72 Varroidae, 1
Theoretical biology, 155 Viscosity, 93, 133
Third quiescent females, 118 Void nest, 57–59, 98
Third quiescent males, 118 Volatile chemical, 33, 91
Third quiescent stage, 74
Thrips, 96
Ticks, 1 W
Tillia, 8 Walking apparatus (Ambulacrum), 23
Time-limited resources, 83 Walking path, 19
Timid female, 103 Waste management, 91, 92
Timid male, 103 Waste management behavior, 89
Timid-scramble, 103 Water evaporation, 10
Timid trait, 103 Weapon, 111
Toads, 74 Weaponry, 125
Trade-off, 9, 69 Web, 3
Trichogramma pretiosum, 139 as a shelter, 13
Trioza cinnamomi, 17 functions, 14, 15
Tritrophic system, 37 intact, 53
Trombidiidae, 2 mat, 94
Trophic level, 38 removal, 53
Two-dimensional electrophoresis, 120 structures, 35
Two-male defence, 161 Willows, 8
Tychothelytoky, 72 Wingless fig wasp, 127, 133
Tydeidae, 5, 6 Winter harshness, 136, 137
Tydeiid, 94 Winter minimum temperature, 115, 119, 121,
Tydeus sp., 94 125, 135
Typhlodromus, 36 Winter severity, 115
Index 187