AMER. ZOOL.

, 34:134-144 (1994)

Biodiversity on Oceanic Islands: Its Origin and Extinction1

GUSTAV PAULAY
Marine Laboratory, University of Guam, Mangilao, Guam 96923

SYNOPSIS. The isolation and small size of oceanic islands make them
attractive models for studies of diversification; the sensitivity of their
biota makes them important subjects for studies of extinction. I explore
the origin of island biotas through dispersal and in situ diversification,
and examine the fate of these biotas since human contact. Island biotas

Downloaded from https://academic.oup.com/icb/article/34/1/134/111610 by guest on 01 October 2022

start out depauperate and disharmonic, facilitating the survival of relict
taxa and stimulating adaptive radiations. The often highly restricted range
and small population size of insular species, together with their limited
diversity of defenses, make island biotas particularly vulnerable to extinc-
tion, largely through habitat loss or interactions with introduced species.
INTRODUCTION to other continental land masses in the past,
Islands, due to their discrete, isolated having since become separated due to tec-
nature, provide excellent opportunities for tonics or sea level rise.
understanding the origin, diversification and Many oceanic islands are the products of
extinction of terrestrial biotas. Here I exam- volcanism from stationary, "hotspot,"
ine these processes on oceanic islands, magma sources. As a tectonic plate drifts
focusing on those remote from continental over such a source, volcanoes incorporated
source areas. Such islands can serve as model into the plate are carried away from the hot-
systems for addressing fundamental ques- spot, and a linear chain of islands or sea-
tions about biodiversity and conservation: mounts results, with the youngest ones lying
what areas are most likely to develop high nearest the hotspot {e.g., Hawaiian, Society
species diversity and endemicity, and what Islands) (see Menard, 1986 and Williamson,
makes particular species or biotas vulner- 1981 for reviews of island geology). Such
able to extinction? hotspot islands usually have a well defined
In part, the biotas of oceanic islands origin, and undergo predictable develop-
diversify in situ, yielding insight into the ment with age. They change from high,
evolutionary origin of diversity. Their bio- undissected, shield volcanoes to smaller,
tas are usually endemic and among the most highly eroded and topographically complex
vulnerable in the world, and are currently islands before final submergence of their
undergoing unprecedented rates of extinc- volcanic portions, usually at 3-15 million
tion. An understanding of the processes years of age. Reef buildup around tropical
leading to these extinctions is important in islands may ultimately convert high islands
itself, and also serves as a model for con- into atolls, with low lying coral sand cays
servation problems worldwide. projecting a few meters above sea level.
Although atolls may persist for tens of mil-
lions of years, they have a propensity for
ORIGIN OF ISLANDS repeated submergence as sea levels rise and
Any isolated habitat can be considered fall; thus their terrestrial biotas are ephem-
insular. Because of their simplicity, oceanic eral (Paulay, 1991a). Continued activity of
islands are especially illuminating and are a hotspot can yield an archipelago that is
the focus of this review. By definition, oce- persistent through time, with new islands
anic islands never had connections to con- arising as older ones subside. The positions
tinental land masses. They are the products of islands on a single tectonic plate do not
of volcanism or tectonic uplift, or the result change relative to each other, although some
of organic reef growth upon foundations islands may drown while others arise.
formed by the first two processes. In con- Other oceanic islands are the result of
trast, most continental islands were joined island arc volcanism, formed where one tec-
134
 BIODIVERSITY ON ISLANDS 135

tonic plate is subducting under another (e.g., butes, and this is reflected in the composi-
Solomon, Tongan Islands). Such islands tion of island faunas. Thus land snail faunas
generally undergo a variety of changes due on remote oceanic islands are generally lim-
to recurrent volcanism, subsidence, uplift ited to minute species, a legacy of their orig-
and erosion in these tectonically complex inating largely through aerial dispersal
regions. In the Pacific, they often attain larger (Vagvolgyi, 1975). Of those species which
sizes than hotspot islands. The age(s) of can disperse to an island, a subset will be
emergence of an arc island is often difficult able to establish and maintain populations,
to define, and the size and degree of emer- depending on attributes of the species and
gence can vary greatly through time. Fur- the island. An island's size, its habitats, and

Downloaded from https://academic.oup.com/icb/article/34/1/134/111610 by guest on 01 October 2022

thermore, because such islands lie in tec- its preexisting biota are all important in
tonically active areas, their positions are determining the success of potential colo-
continually changing relative to other land nists; preexisting species can place impor-
masses. Thus while the age, position, and tant limits on colonization via interactions
physiographic evolution of hotspot islands such as competition, predation, and sym-
are simple and readily predictable, those of biosis.
arc islands are complex. This makes hotspot While the biodiversities of solitary, young
islands more attractive as model systems for islands near continents can be solely attrib-
the study of diversification. uted to colonization events (e.g., Krakatau
Island-Thornton et al, 1990), those of
ORIGIN OF BIOTA—COLONIZATION remote, older archipelagoes often result from
An island's biota is the product of oversea extensive radiations of a handful of original
colonizations and local diversifications. colonists. Thus the estimated 10,000
Oceanic islands receive their biotas solely Hawaiian insect species evolved from ca.
through dispersal. The probability of colo- 350-400 successful colonists (Howarth,
nization is influenced by the island's geo- 1990), and the 1,138 indigenous Hawaiian
graphical and environmental settings as well vascular plants are derived from 387-396
as by organismal attributes (MacArthur and founders (Wagner, 1991).
Wilson, 1967). Because the probability of Although the immediate effects of dis-
dispersal is inversely related to distance, an persal limitations are depauperate and dis-
island may receive more of its propagules harmonic biotas on remote islands, this
from relatively depauperate nearby islands biotic poverty can lead to the survival of
than from very diverse but remote conti- unusual, relict taxa, and stimulate the devel-
nents. opment of adaptive radiations in successful
Organisms vary greatly in dispersal abil- dispersers, so that high, remote oceanic
ity, both with regards to distance traveled islands can become fairly species rich
and method of transport (Carlquist, 1974). through time.
The geographic range of a taxon is often
correlated with its dispersal ability (Kohn RELICT TAXA
and Perron, 1993), and the biotas of remote The view that evolution is an arms race
islands are largely limited to species with among predators sensu lato (including par-
excellent dispersal abilities and their de- asites, pathogens, herbivores) and their prey,
scendents. In the island-rich oceanic Pacific, or among competitors (Van Valen, 1973),
where the far majority of organisms origi- may explain the temporal or spatial restric-
nates from western (Austro-Asian) source tion of certain groups in the history of life
areas, fewer and fewer taxonomic groups are (Vermeij, 1987). Species that cannot cope
represented as one moves eastward through with escalating biotic pressures in most
the basin (e.g., Zimmerman 1942), with environments become globally extinct or
poorly dispersing taxa (e.g., mammals, may survive in unusual habitats or areas
amphibians, gymnosperms) being especially that their adversaries have not penetrated,
limited to western island groups. such as the deep sea, caves, or Australia
Certain methods of dispersal are avail- (Vermeij, 1987). Islands frequently offer
able only to organisms with particular attri- such refuge for relicts, and may also pro-
136 GUSTAV PAULAY

Downloaded from https://academic.oup.com/icb/article/34/1/134/111610 by guest on 01 October 2022

FIG. 1. Distributions of the land snail families Endodontidae (solid line, based on Solem, 1976), Partulidae
(dashed line, based on Cowie, 1992) and Amastridae (dotted line).

mote the secondary, in situ evolution of spe- ically most generalized family (Endodonti-
cies with limited defensive or competitive dae—Solem, 1976) of higher landsnails
abilities (Solem, 1979; Carlquist, 1974). (Sigmurethra), are restricted to islands
Thus many land crab and seabird species (Solem, 1979). The fact that these land snails
are widespread but strictly insular, restricted have reached even the most remote Pacific
largely by mammalian predators (Hartnoll, islands, yet are virtually unknown from
1988; Steadman, 1989). Numerous benthic continents, strongly implicates biotic fac-
marine species are restricted to, but wide- tors rather than dispersal limitations as set-
spread among, the small oceanic islands of ting distributional boundaries (Fig. 1). The
the Pacific tectonic plate (Springer, 1982), past continental occurrence of at least acha-
an enigmatic distribution that similarly tinellids (Solem, 1979) and the extreme vul-
could be forced by biotic exclusion from nerability of all four families to certain
more diverse western areas. introduced predators (see below) support the
Relict taxa are well known on large, old, hypothesis that their relict distribution is
continental islands, the lemurs of Mada- the result of escalation.
gascar and the monotremes and marsupials The primitive weevil family Aglycyderi-
of Australia being classic (if not indisput- dae is similarly restricted to islands in the
able) examples. Although short-lived oce- Pacific and Atlantic oceans, with one du-
anic islands appear at first ill-suited for the bious exception (Zimmerman, 1948).
long-term survival of relict lineages, groups Although aglycyderids occur on most cen-
with good dispersal abilities can survive by tral Pacific high islands, they are rare and
island hopping. species poor except on the Hawaiian Islands,
Three (Partulidae, Achatinellidae, Amas- where they have undergone an explosive
tridae) of the six families that comprise the diversification, resulting in over 170
most primitive order (Orthurethra) of pul- described species (Nishida, 1992). Zim-
monate land snails, as well as the anatom- merman (1970) argued that this Hawaiian
 BIODIVERSITY ON ISLANDS 137

radiation was possible because of the absence mass extinctions or in remote insular set-
of the very diverse and ecologically domi- tings, the potential of organisms to diversify
nant Miocalles weevils that cooccur with the and enter novel adaptive zones can be better
family elsewhere. realized (Simpson, 1953). Radiations are
The restriction to and survival on ephem- especially prevalent on remote islands lying
eral islands by such primitive and presum- at the edge of a group's dispersal range, in
ably ancient groups depends upon their peripheral areas termed the "radiation zone"
ability to continually disperse to new islands by MacArthur and Wilson (1967), where the
as older ones sink below sea level. For low diversity of colonists facilitates in situ
example, endodontoid land snails are known diversification (Diamond, 1977).

Downloaded from https://academic.oup.com/icb/article/34/1/134/111610 by guest on 01 October 2022

from Lower Miocene to Pleistocene depos- The effects of certain "keystone" taxa in
its on three islands (Midway, Bikini, and limiting the evolutionary options of others
Enewetak atolls) that have since subsided are especially striking, as demonstrated
greatly and no longer support the snails where such taxa are absent. Ants, for exam-
(Solem, 1982); they persist on other, mostly ple, dominate arthropod communities
younger islands throughout the Pacific. through most of the world; their absence in
Similar dispersal from drowning to new Hawaii and SE Polynesia appears to be
islands has been demonstrated within dro- largely responsible for the great radiations
sophilids and honeycreepers in the Hawai- of carabid beetles and spiders, and the
ian island-chain (Sibley and Ahlquist, 1982; exploitation of predatory niches by such
Beverley and Wilson, 1985; see below). unusual organisms as caterpillars, in those
Such relict taxa not only survive on Pacific island groups (Wilson, 1990).
islands, but can form a dominant portion That diversification can be very rapid is
of the indigenous fauna, especially among immediately suggested by the sheer diver-
mollusks. The strictly insular land snail sity and high levels of single-island ende-
families discussed above constitute 79% of mism exhibited by certain clades on islands
the Hawaiian and 70% of the Rapan (SE a few million years old. Thus at least 70
Polynesia) land snail fauna (Solem, 1982, species of Mecyclothorax beetles are
1990). endemic to 1 My old Tahiti (Perrault, 1987),
25 of the 26 species of picture-wing Dro-
ECOLOGICAL RELEASE AND LOCAL sophila on 600,000 yr old Hawaii are
DIVERSIFICATION restricted to that island (Carson, 1983), and
Island biotas are famous for large evo- at least 11 lineages of insects and land snails
lutionary radiations such as that of many have radiated to over 100 species each in
hundreds of species of Hawaiian drosoph- the Hawaiian Islands. That morphological
ilids (Carson, 1983; Nishida, 1992), and differentiation as well as speciation can be
bizarre adaptations, like the boid snakes with very rapid is further attested by the minimal
jointed upper jaws in the Mascarenes (Fra- genetic differentiation found within some
zetta, 1970). Studies of adaptive radiations striking adaptive radiations {e.g., Helenurm
on islands indicate that both diversification and Ganders, 1985; Lowrey and Crawford,
and adaptive evolution are favored by a low 1985). The 19 species of Hawaiian Bidens
diversity of initial colonists, that both can exhibit much more morphological and eco-
be extremely rapid, and that diversification logical diversity than this diverse genus does
is stimulated by opportunities for isolation in the Americas, yet the genetic diversity of
on topographically complex archipelagoes. this entire insular radiation is comparable
to that found among populations within
The potential for taxa to diversify, to American species (Helenurm and Ganders,
evolve unusual adaptations and to expand 1985).
into novel adaptive zones can be thwarted
in diverse biotas by the high degree of uti- Within an oceanic archipelago, local
lization of existing niche space and high lev- diversification can occur by inter- or in-
els of escalation that limit evolutionary tra-island speciation. Only the former is
experimentation. It has long been recog- generally available to organisms with good
nized that in low diversity settings, as after over-land dispersion abilities (especially
138 GUSTAV PAULAY

vertebrates and plants). Diversifications topographic or environmental gradients

among such organisms are relatively {e.g., Tweedie, 1961).
uncommon, with only a handful of isolated Both island size and age may limit diver-
archipelagoes showing large endemic radi- sification. The frequent correlation between
ations of birds or plants. The number of species richness and island size may be the
adjacent islands appears to be more impor- result of increasing habitat heterogeneity
tant than island size in facilitating bird radi- with island size, or an ecological or evolu-
ations: the Hawaiian and Galapagos Archi- tionary equilibrium between origination
pelagoes have significant radiations in birds, (colonization and local speciation) and
while the larger and more ancient, conti- extinction rates (MacArthur and Wilson,

Downloaded from https://academic.oup.com/icb/article/34/1/134/111610 by guest on 01 October 2022

nental islands of New Caledonia and New 1967; Williamson, 1981). In remote islands,
Zealand show little in situ diversification where much of the diversity is the result of
(Olson, 1991). Inter-island speciation can local diversification, island age (the time
nevertheless lead to large radiations, as available for diversification) may be an
attested to by honeycreepers, several lin- important control (Wagner, 1991). Because
eages of plants, and numerous lineages of hotspot islands subside and shrink with age,
highly volant insects in the Hawaiian Islands older islands within an archipelago tend to
(Wagner, 1991; James and Olson, 1991; be smaller. This may give rise to opposite
Nishida, 1992). diversity patterns, depending on whether
Intra-island, or "continental," speciation time or area is limiting diversification.
is restricted to islands that are large enough Groups capable of rapid rates of specia-
to allow for effective isolation of popula- tion (relative to island age and available
tions. While intra-island speciation in birds niche space) may so rapidly saturate even
requires an island the size of Madagascar, young islands with species that they yield
for small land snails and flightless insects a the expected positive correlation between
few square kilometers can be sufficient (Dia- species richness and island area. Such is the
mond, 1977; Paulay, 1985). Extreme geo- case for the highly volant, Hawaiian Pla-
graphic localization is especially prevalent githmysus beetles, which exhibit high rates
on topographically complex, highly dis- of inter-island speciation (Fig. 2). In con-
sected volcanic islands, a reflection of the trast, if diversification is relatively slow, then
very steep topographic, climatic, and envi- diversity may strongly correlate with island
ronmental gradients occurring across islands age and therefore be negatively correlated
only a few kilometers in diameter yet hun- with island area. Such is the case for the
dreds or thousands of meters high. Such flightless Hawaiian Rhyncogonus weevils,
geographic localization can lead to remark- which appear to diversify slowly via intra-
able radiations through intra-island speci- island speciation (Fig. 2).
ation. Thus on isolated Rapa, a 40 km2, 650 Age is important from both an evolu-
m high, 5 My old island in SE Polynesia, tionary and conservation perspective, as old
local radiations resulted in 67 flightless islands often have the most divergent spe-
Miocalles weevil, and 45 achatinellid and cies and for some lineages, the most diverse
24 endodontid land snail species (Solem, fauna. The largest land snail radiations
1982; Paulay, 1985). Many of these species among SE Polynesian and Micronesian hot-
have restricted ranges on Rapa, some spot islands, both in numbers of species and
remarkably so {i.e., <1 km2). In contrast, the development of endemic genera, occur
islands with low topographic differentia- on the relatively old islands of Gambier,
tion, such as the flat-topped, uplifted lime- Rapa and Pohnpei (all 5-7.5 My old) (Solem,
stone islands of Henderson and Niue (S 1982). The most interesting and divergent
Polynesia), offer few opportunities for geo- insect fauna among the major islands of the
graphic restriction and generally lack intra- Hawaiian archipelago is on Kauai, the old-
island diversifications (Paulay, 19916). The est island in the chain (Zimmerman, 1948).
importance of topographic, climatic and The time available on an archipelago for
environmental variation in limiting species evolutionary divergence and diversification
ranges and facilitating diversification is also of a lineage can be greater than the age of
apparent in continental areas with steep the oldest emergent island. Thus the Hawai-
 BIODIVERSITY ON ISLANDS 139

O Plagithmysus
O • Rhyncogonus
50
to
O
o 40 - - o
<D a>
Q.
en in
30 -

0)
-Q -Q

Downloaded from https://academic.oup.com/icb/article/34/1/134/111610 by guest on 01 October 2022

20
E E
3
10 O

n • i i 1 1 1

10° 10"
Island age (My) Island area (km2)
FIG. 2. Diversification of Plagithmysus (Cerambycidae) and Rhyncogonus (Curculionidae) among the major
Hawaiian Islands. Species richness of Plagithmysus is strongly correlated with log island size (r = 0.93), but not
island age (r = -0.29); that of Rhyncogonus with island age (r = 0.96) but not log island size (r = -0.19). Data
from Nishida (1992) and Howarth and Mull (1992).

ian chain, in existence for at least 70 My, cies of Amastridae, a family endemic to the
was colonized by drosophilids and drepan- Hawaiian group, are extinct, with only a
idine honeycreepers well before the birth of handful of arboreal species of this primarily
the present major high islands (Sibley and ground-dwelling family surviving (M. G.
Ahlquist, 1982; Beverley and Wilson, 1985). Hadfield, personal communication, 1992).
The spectacular radiations of endodontoid
VULNERABILITY, EXTINCTION, land snails are known to be largely or com-
CONSERVATION pletely extinct in the Hawaiian Islands
The biodiversity crisis is nowhere more (195+ species), the Gambiers (25 species),
apparent and in need of urgent attention Rarotonga (11 species), and St. Helena (12
than on islands. Approximately 90% of all species), these being the islands which have
bird extinctions during historic times have been recently searched for the snails (Solem,
occurred on islands (Vitousek, 1988). More 1976, 1977, 1982). Achatinellid and par-
species of Polynesian landbirds appear to tulid land snails in the Pacific are losing
have gone extinct due to human agencies ground at a similarly alarming rate (Had-
than survive today, and the survivors have field et al, 1993; Murray et al, 1988; Hop-
greatly reduced ranges (Olson, 1989; Stead- per and Smith, 1992).
man, 1989). Among the historically known,
indigenous plants of Hawaii, 10% are extinct Direct habitat destruction
and almost 40% threatened (Wagner, 1991). The impact of humans on island habitats
Comparable quantitative data on the is usually devastating, with the most impor-
extinction rates of terrestrial arthropods are tant causes of human-induced extinction on
not available, but indications such as the islands being habitat destruction and spe-
virtual absence of native insects below 600 cies introductions.
m altitudes and the extinction of land crabs All Polynesian islands were entirely or
in Hawaii indicate similar tragedies (Zim- largely forested before man's arrival; fossil
merman, 1948; Howarth, 1990). pollen and mollusks provide evidence for
Land snails are perhaps the most vulner- this even on islands or areas where no indig-
able members of insular biotas, and the four enous forests remain (e.g., Zimmerman,
relict families discussed above are at the 1948;Paulay, 1985; Bahn and Flenley, 1992;
greatest risk. Most of the 331 described spe- Kirch et al, 1992). Humans have been
140 GUSTAV PAULAY

responsible for the partial or entire loss of islands, because of their spectacular biotas,
native forests from virtually all of these should be among the highest conservation
islands (Kirch, 1984; Kirch et al., 1992). priorities.
The conversion of land for agriculture, con-
struction, etc., and perturbations accom- Introduced species
panying this process (pollution, erosion, Introduced species threaten island bio-
etc.), directly destroy forests, while species diversity both indirectly through habitat
introductions are further, indirect causes of alteration and directly through interactions
habitat destruction (see below). An especially with native species. Their effects can be even
striking example of the environmental havoc greater than those of anthropogenic habitat

Downloaded from https://academic.oup.com/icb/article/34/1/134/111610 by guest on 01 October 2022

created by complete deforestation is Easter destruction; in this islands differ from con-
Island, where destruction of the forests not tinental environments, where exotics appear
only caused untold extinctions, but was to be much less often the cause of extinction
apparently responsible for the collapse of (Vitousek, 1988).
the island's megalithic culture (Bahn and Introduced species, especially certain her-
Flenley, 1992). bivorous mammals and plants, are among
The vulnerability of habitats is largely the worst culprits in causing large-scale hab-
determined by their accessibility and suit- itat destruction and alteration. Insular floras
ability to humans, and by their resilience to have few defenses against herbivorous
disturbance. Dry, gentle sloping, lowland mammals which, unchecked by predators,
areas on volcanic terrain are among the most can destroy most of the vegetation and lead
vulnerable, while steep, wet, mountainous, to massive erosion. Paradoxically, uninhab-
karstic areas are the least susceptible to ited islands can be among the most affected
human disturbance. In the Hawaiian Islands, by introduced feral mammals, due to lack
humans had extensively altered 80% of all of human predation. The large-scale deveg-
native habitats below 450 m in elevation etation by mammalian herbivores of unin-
by 1600 A.D. (Simon et al, 1984). Dry for- habited Laysan (Hawaiian Is.) and Eiao
ests, which may host an even more diverse (Marquesas), and of St. Helena prior to
and endemic biota than rain forests on human settlement, led to the loss of much
islands, are especially vulnerable, with but oftheir endemic biotas (Ely and Clapp, 1973;
fractions of their former cover remaining Montgomery et al., 1979; Cronk, 1989).
on all but the most pristine islands (Olson, Another serious threat to native vegetation
1989). Olson's (1989) discovery, based on is certain introduced plants, such as Miconia
subfossil remains, that bird faunas of dry in Tahiti, Psidium in Tubuai, Leucaena in
forests were much more diverse before the Marquesas, which have the ability to
human occupation, led him to question crowd out native vegetation with their
whether we are similarly underestimating monospecific stands.
the diversity of mesic and arid habitats on In addition to destroying habitat, intro-
continents, an idea which has gained recent duced species can wreak havoc in direct
support in the Americas (Mares, 1992). interactions (competition, predation) with
As a result of millions of years of erosion native species. Low diversity biotas such as
and subsidence, the oldest oceanic hotspot those on islands, whose species tend to have
islands tend to have low relief, elevation less highly evolved defensive, competitive
and (therefore) precipitation, traits which and reproductive capabilities, are especially
make them especially vulnerable to human vulnerable to biological invasion (Vermeij,
disturbance. Thus these islands, which likely 1991). As they evolved, insular organisms
held the most interesting products of evo- were exposed to but a fraction of the diver-
lution, are also among the most disturbed sity of predators sensu lato and competitors
(e.g., Eiao, Chuuk, Tubuai, Gambier). experienced by mainland species; thus the
Among Pacific hotspot islands > 5 My old, modern arrival of certain exotics, against
only Pohnpei (Carolines), Rapa (Australs), which the indigenous biota lack defenses,
Kauai (Hawaii) and the small rock-like, lee- can be devastating. As closely related spe-
ward Hawaiian high islands retain a rea- cies can all be at risk from a single exotic,
sonable cover of native vegetation; these considerable portions of the biota can
 BIODIVERSITY ON ISLANDS 141

become rapidly endangered on islands where 1991). The purposeful and thoughtless
large radiations occurred (Simon et ai, introduction of the "cannibal snail," Eu-
1984). glandina rosea, has led to the extinction of
The accidental introduction of the snake large suites of endemic land snails on one
Bioga irregularis to Guam has led to the island after another, providing some of the
extermination of practically the entire native most detailed case-studies of insular extinc-
bird fauna (Savidge, 1987); avian malaria tion (Hadfield et al, 1993; Murray et ah,
and avian pox have decimated Hawaiian 1988). Apparent relict groups, such as many
birds (Riper, 1991). Among insects, social land snails, are among the most sensitive to
Hymenoptera are an especially serious introduced predators, and the malacofauna

Downloaded from https://academic.oup.com/icb/article/34/1/134/111610 by guest on 01 October 2022

threat, and certain introduced ant species of Pacific islands is undergoing a veritable
have been implicated in the extinction of mass extinction, with only an estimated 25-
large portions of the invertebrate faunas of 35% of Hawaii's 1,461 described land snails
several Pacific island groups, especially on remaining extant (Solem, 1990, see also
some remote islands (see above) which above).
lacked indigenous ants {e.g., Cole et ai,
1992; Lubin, 1984; Howarth, 1990). CONCLUSIONS
Humans themselves are a predatory The isolation and small size of oceanic
threat. Recent work on Holocene extinc- islands yield differences in the dynamics of
tions clearly implicates Homo sapiens in the diversification and extinction when com-
demise of insular birds. The arrival of pared to continents. These attributes, how-
humans to Polynesia led to the extinction ever, also allow a more intimate under-
of generally more than half, and in some standing of insular biotas: diversity on
cases up to 80%, of each island's avifauna, islands is limited and the geographic and
with large, predatory and flightless birds ecologic ranges of species can be precisely
being most vulnerable (Olson, 1989; Stead- determined, providing excellent grounds for
man, 1989). Similarities between insular studies on the distribution and origin of bio-
bird extinctions and the late Quaternary logical diversity. The extreme vulnerability
demise of continental vertebrate mega- of insular biotas not only makes their study
faunas lends strong support to the interpre- and conservation ever more urgent, but the
tation that the latter were also human-caused mass extinctions occurring on many islands
{e.g., Olson, 1989; Martin, 1990). facilitate analysis of numerous well defined
Although it may be impossible to arrest examples of human-caused extinctions,
the arrival of all introduced species to islands studies which can lead to better conserva-
with large human populations without caus- tion policies.
ing extreme restrictions to travel and ship- Low primary (founding) diversity pro-
ping, it is critical that we stem this flow as motes in situ diversification and the evo-
much as possible. The scale of this ongoing lution of unusual adaptations. The most iso-
biotic interchange is astounding. Almost lated islands, those with the lowest number
3,000 species of terrestrial arthropods are of original colonists, often have the largest
known to have been introduced to Hawaii, adaptive radiations and most peculiar spe-
the vast majority since European contact cies. Whether intra-island and inter-island
(Nishida, 1992). Control can be more easily speciation is more important depends on
implemented for islands with few or no the dispersion ability of the taxon and
human inhabitants, by restricting and care- opportunities for isolation. Evidence from
fully screening access. Control of especially islands shows the importance of topo-
harmful exotics should be a priority, and is graphic and environmental variability in
increasingly being pursued (Gillis, 1992). limiting species ranges and promoting in situ
One aspect of this biotic interchange that diversification. Diversification can be very
can be controlled is the purposeful intro- rapid and result in large speciesflockswithin
duction of scores of species for biological the normally 3-15 My lifespan of oceanic
control. Many of these are fairly generalist hotspot islands.
predators and parasites that can lead to the The small size and isolation of islands
extinction of numerous endemics (Howarth, result in relatively small and localized pop-
142 GUSTAV PAULAY

ulations, and low primary diversity. Small REFERENCES

population size, especially prevalent among Bahn, P. and J. Flenley. 1992. Easter Island, Earth
species with large body size or those at high Island. Thames and Hudson, London.
trophic levels, makes many island species Beverley, S. M. and C. A. Wilson. 1985. Ancient
especially vulnerable to extinction. The often origin for Hawaiian Drosophilinae inferred from
highly restricted ranges of flightless inver- protein comparisons. Proc. Nat. Acad. Sci. U.S.A.
82:4753-4757.
tebrates exacerbate this problem, because Carlquist, S. 1974. Island biology. Columbia Uni-
entire species can be vulnerable to even lim- versity Press, New York.
ited habitat destruction. Data from islands Carson, H. L. 1983. Chromosomal sequences and
also imply that mesic and xeric environ- interisland colonizations in Hawaiian Drosophila.

Downloaded from https://academic.oup.com/icb/article/34/1/134/111610 by guest on 01 October 2022

ments may have been equally or more Genetics 103:465-482.
diverse, and certainly much more vulner- Cole, F. R., A. C. Medeiros, L. L. Loope, and W. W.
Zuehlke. 1992. Effects of the argentine ant on
able, than moist habitats. arthropod fauna of Hawaiian high-elevation
The relatively low levels of escalation shrubland. Ecology 73:1313-1322.
resulting from limited diversity of founders Cowie, R. H. 1992. Evolution and extinction of Par-
allow for the survival of relicts and the tulidae, endemic Pacific island land snails. Phil.
Trans. Roy. Soc. Lond. B 335:167-191.
development of species with limited rep- Cronk, Q. C. B. 1989. The past and present vegeta-
ertoires of defenses; thus island ecosystems tion of St. Helena. J. Biogeogr. 16:47-64.
are especially vulnerable to invasion and Diamond, J. M. 1977. Continental and insular spe-
extinctions caused by introduced species. ciation in Pacific land birds. Syst. Zool. 26:263-
Introduced species are much more devas- 268.
tating on islands than continents, and are Ely, C. A. and R. B. Clapp. 1973. The natural history
of Laysan Island, Northwestern Hawaiian Islands.
presently the greatest threat on many islands. Atoll Res. Bull. 171:1-361.
The most devastating exotics include plants Frazetta, T. H. 1970. From hopeful monsters to
that can literally push out native forests with bolyerine snakes? Am. Nat. 104:55-72.
their monospecific stands, social insects Gillis, A. M. 1992. Keeping aliens out of paradise.
(especially certain ants) and herbivorous BioScience 42:482-485.
mammals. Studies of vertebrate extinctions Hadfield, M. G., S. E. Miller, and A. H. Carwisle.
1993. The decimation of endemic Hawaiian tree
on islands strongly implicate humans as the snails by alien predators. Am. Zool. 33:610-622.
cause of late Quaternary extinctions world- Hartnoll, R. G. 1988. Evolution, systematics, and
wide. Purposeful introductions of generalist geographical distribution. In W. W. Burggren and
predators for biological control have also B. R. McMahon (eds.), Biology of the land crabs,
led to large scale extinction. pp. 6-54. Cambridge University Press, Cam-
bridge.
The 50-95 + % levels of extinctions Helenurm, K. and F. R. Ganders. 1985. Adaptive
exhibited by taxa with a fossil record, like radiation and genetic differentiation in Hawaiian
birds and land snails, on oceanic islands, Bidens. Evolution 39:753-765.
are comparable to the most severe mass Hopper, D. R. and B. D. Smith. 1992. Status of tree
snails (Gastropoda: Partulidae) on Guam, with a
extinctions in earth history (Jablonski, resurvey of sites studied by H. E. Crampton in
1991). Arresting habitat destruction, reduc- 1920. Pacific Sci. 46:77-85.
ing chances of future introductions, and Howarth, F. G. 1990. Hawaiian terrestrial arthro-
controlling harmful exotic species are pods: An overview. Bishop Museum Occ. Pap. 30:
urgently needed to prevent the annihilation 4-26.
of these most spectacular showcases of evo- Howarth, F. G. 1991. Environmental impacts of clas-
sical biological control. Ann. Rev. Entomol. 36:
lution. 485-509.
Howarth, F. G. and W. P. Mull. 1992. Hawaiian
insects and their kin. University of Hawaii Press,
ACKNOWLEDGMENTS Honolulu.
I thank Bern Holthuis, Frank Howarth, Jablonski, D. 1991. Extinctions: A paleontological
perspective. Science 253:754-757.
Alex Kerr, Alan Kohn, Scott Miller, and James, H. F. and S. L. Olson. 1991. Descriptions of
Barry Smith for comments on the manu- thirty-two new species of birds from the Hawaiian
script and Trish Morse for organizing the Islands: Part II. Passeriformes. Ornithological
symposium. This is contribution 331 from Monographs 46:1-88.
the University of Guam Marine Labora- Kirch, P. V. 1984. The evolution of Polynesian chief-
doms. Cambridge Univ. Press, Cambridge.
tory. Kirch, P. V., J. R. Flenley, D. W. Steadman, F. Lamont,
 BIODIVERSITY ON ISLANDS 143

and S. Dawson. 1992. Ancient environmental Psydrini) a Tahiti (Polynesie Francaise). Bulletin
degradation. Nat. Geogr. Res. Explor. 8:166-179. de la Societe Zoologique de France 112:419-427.
Kohn.A. J. and F.E. Perron. 1993. Development and Riper, C. van III. 1991. The impact of introduced
distribution: Comparative life history and bioge- vectors and avian malaria on insular passeriform
ography ofIndo-Pacific Conus. Oxford Univ. Press, bird populations in Hawaii. Bull. Soc. Vector Ecol.
Oxford. 16:59-83.
Lowrey, T. K. and D. J. Crawford. 1985. Allozyme Savidge, J. A. 1987. Extinction of an island forest
divergence and evolution in Tetramolopium avifauna by an introduced snake. Ecology 68:660-
(Compositae: Astereae) on the Hawaiian Islands. 668.
Syst. Bot. 10:64-72. Sibley, C. G. and J. E. Ahlquist. 1982. The relation-
Lubin, Y. D. 1984. Changes in the native fauna of ships of the Hawaiian honeycreepers (Drepani-
the Galapagos Islands following invasion by the nini) as indicated by DNA-DNA hybridization.

Downloaded from https://academic.oup.com/icb/article/34/1/134/111610 by guest on 01 October 2022

little red fire ant, Wassmannia auropunctata. Biol. Auk 99:130-140.
J. Linn. Sci. 21:229-242. Simon, C. M., W. C. Gagne, F. G. Howarth, and F. J.
MacArthur, R. H. andE. O. Wilson. 1967. Thetheory Radovsky. 1984. Hawai'i: A natural entomolog-
of island biogeography. Princeton University Press, ical laboratory. Bull. Entomol. Soc. Amer. 30(3):
Princeton. 8-17.
Mares, M. A. 1992. Neotropical mammals and the Simpson, G. G. 1953. The majorfeatures ofevolution.
myth of Amazonian biodiversity. Science 255:976- Columbia University Press, New York.
979. Solem, A. 1976. Endodontoid land snailsfrom Pacific
Martin, P. S. 1990. 40,000 years of extinctions on Islands (Mollusca: Pulmonata: Sigmurethra). Part
the "planet of doom." Palaeogeogr. Palaeocli- I. Family Endodontidae. Field Museum of Natural
matol. Palaeoecol. 82:187-201. History, Chicago.
Menard, H. W. 1986. Islands. W. H. Freeman, New Solem, A. 1977. La faune terrestre de Tile de Sainte-
York. Helene. Fam. Charopidae. Mus. Roy. l'Afrique
Montgomery, S. L., W. C. Gagne, and B. H. Gagne. Centrale—Tervuren, Belgique Ann., Ser. 8, Sci.
1979. Notes on birdlife and nature conservation Zool. 220:521-533.
in the Marquesas and Society Islands. 'Elepaio 40: Solem, A. 1979. Biogeographic significance of land
152-155. snails, Paleozoic to Recent. In J. Gray and A. J.
Murray, J., E. Murray, M. S. Johnson, and B. Clarke. Boucot (eds.), Historical biogeography, plate tec-
1988. The extinction ofPartula on Moorea. Pacific tonics, and the changing environment, pp. 277-
287. Oregon State University Press, Corvallis.
Sci. 42:150-153.
Solem, A. 1982. Endodontoid land snails from Pacific
Nishida, G. M. (ed.) 1992. Hawaiian terrestrial Islands (Mollusca: Pulmonata: Sigmurethra). Part
arthropod checklist. Bishop Museum Tech. Rep. II. Families Punctidae and Charopidae, zooge-
l:viii + 262. ography. Field Museum of Natural History, Chi-
Olson, S. L. 1989. Extinction on islands: Man as a cago.
catastrophe. In D. Western and M. C. Pearl (eds.), Solem, A. 1990. How many Hawaiian land snail spe-
Conservation for the twenty-first century, pp. 50- cies are left? and what we can do for them. Bishop
53. Oxford Univ. Press, New York. Museum Occ. Pap. 30:27-40.
Olson, S. L. 1991. Patterns of avian diversity and Springer, V. G. 1982. Pacific plate biogeography, with
radiation in the Pacific as seen through the fossil special reference to shore fishes. Smiths. Contr.
record. In E. C. Dudley (ed.), The unity of evolu- Zool. 367:1-182.
tionary biology, pp. 314-318. The Proceedings of Steadman, D. W. 1989. Extinction of birds in eastern
the Fourth International Congress of Systematic Polynesia: A review of the record, and compari-
and Evolutionary Biology. Dioscorides Press, sons with other Pacific island groups. J. Archaeol.
Portland, Oregon. Sci. 16:177-205.
Paulay, G. 1985. Adaptive radiation on an isolated Thornton, I. W. B., T. R. New, R. A. Zann, and P. A.
oceanic island: The Cryptorhynchinae of Rapa Rawlinson. 1990. Colonization of the Krakatau
revisited. Biol. J. Linn. Soc. 26:95-187. Islands by animals: A perspective from the 1980's.
Paulay, G. 1991a. Late Cenozoic sea level fluctua- Phil. Trans. Roy. Soc, London. B 328:131-165.
tions and the diversity and species composition of Tweedie, M. W. F. 1961. On certain Mollusca of the
insular shallow water marine faunas. In E. C. Dud- Malayan limestone hills. Bull. Raffles Mus., Sin-
ley (ed.), The unity of evolutionary biology. The gapore 26:49-65.
Proceedings of the Fourth International Congress Vagvolgyi, J. 1975. Body size, aerial dispersal, and
of Systematic and Evolutionary Biology, Vol. 1, origin of the Pacific land snail fauna. Syst. Zool.
pp. 184-193. Dioscorides Press, Portland, Oregon. 24:465-488.
Paulay, G. 199It. Henderson Island: Biogeography Van Valen, L. 1973. A new evolutionary law. Evol.
and evolution at the edge of the Pacific plate. In Theory 1:1-30.
E. C. Dudley (ed.), The unity of evolutionary biol- Vermeij, G. J. 1987. Evolution and escalation, an
ogy. The Proceedings of the Fourth International ecological history oflife. Princeton University Press,
Congress of Systematic and Evolutionary Biology, Princeton, New Jersey.
Vol. 1, pp. 304-313. Dioscorides Press, Portland, Vermeij, G. J. 1991. When biotas meet: Understand-
Oregon. ing biotic interchange. Science 253:1099-1104.
Perrault, G. G. 1987. Microendemisme et speciation Vitousek, P. M. 1988. Diversity and biological inva-
de genre Mecyclothorax (Coleoptera—Carabidae sions of oceanic islands. In E. O. Wilson (ed.),
144 GUSTAV PAULAY

Biodiversity, pp. 181-189. National Academy systems: The case of the social insects. Ecology
Press, Washington, D.C. Institute, Oldendorf/Luhe, Germany.
Wagner, W. L. 1991. Evolution of waiffloras:A com- Zimmerman, E. C. 1942. Distribution and origin of
parison of the Hawaiian and Marquesan archi- some eastern oceanic insects. Am. Natur. 76:280-
pelagoes. In E. C. Dudley (ed.), The unity of evo- 307.
lutionary biology. The Proceedings of the Fourth Zimmerman, E. C. 1948. Insects of Hawaii, Vol. 1.
International Congress of Systematic and Evolu- Introduction. University of Hawaii Press, Hono-
tionary Biology, Vol. 1, pp. 267-284. Dioscorides lulu.
Press, Portland, Oregon. Zimmerman, E. C. 1970. Adaptive radiation in
Williamson, M. 1981. Island populations. Oxford Hawaii with species reference to insects. Biotro-
University Press, Oxford. pica 2:32-38.

Downloaded from https://academic.oup.com/icb/article/34/1/134/111610 by guest on 01 October 2022

Wilson, E. O. 1990. Success and dominance in eco-