Testing Theories of Island Biogeography

Islands in the Sea of Cortez suggest that not one but many models may be required to explain patterns of colonization and extinction.

The answers to these questions might depend on historical factors such as whether an island was ever connected to the mainland by a landbridge. But if the turnover of species on an island is high-that is, if colonization and extinction events are rapid there-then historical legacies might be quickly erased, making interpretation of the island biota dependent on current physical attributes of the island and on the biological attributes controlling the dispersal and persistence of its potential colonists.

Before much was known of continental drift and plate tectonics, former landbridges were often invoked to account for discontinuous distributions of species on widely separated land masses. It seemed implausible that many island-dwelling taxa, such as lizards or mammals, could colonize distant islands by oversea dispersal. "I do not deny that there are many and grave difficulties in understanding how several of the inhabitants of the more remote islands, whether still retaining the same specific form or modified since their arrival, should have reached their present home," Darwin commented in Origin of Species (1859, p. 396). It was commonly believed that past geological events had provided temporary terrestrial avenues for colonization, and that once these avenues disappeared all colonization ceased. While a few isolated forms might subsequently become extinct, islands were generally thought to serve as a kind of "museum" of archaic taxa, preserving such species as the giant tortoises of the Galapagos Islands, the tuatara of New Zealand, the Cuban Solenodon, the Tasmanian wolf, the laurel-dominated woodlands of the Canary Islands, and (until recently) the dodo of Mauritius. Moreover, it was believed that island extinctions were more than compensated for by the internal production of new species-and indeed, a large part of the floras and faunas of many archipelagos is comprised of genera and even families that occur nowhere else in the world. The Galapagos finches, the Hawaiian honeycreepers and silverswords, and the Malagassy vangashrikes and tenrecs are just a few examples.

This view sees island biotas as made up largely of evolutionarily modified species that are relics of past historical events. It contrasts with views that emphasize the role of ongoing dispersal to islands from continental areas, a phenomenon for which there is a considerable body of evidence. The colonization of new or recently active (and sterilized) volcanic islands such as Surtsey near Iceland or Krakatoa in the Java Straits has been well documented. Long Island, off New Guinea, was devastated about two centuries ago by a volcanic explosion that deposited a layer of ash 30 m thick, but today the island has a diverse flora with nearly as many resident species of land birds as islands without any such recent cataclysm (Diamond 1974). Other examples of unaided dispersal within and beyond continents are the recent spread of the collared dove, Cetti's warbler, and the firecrest northward through Europe to Britain, the 1975 colonization of Minorca by Dartford warblers, the colonization of Norfolk Island by blackbirds from New Zealand, some 250 km away (Turner et al. 1968), and the recent spread of African cattle egrets throughout much of the world (see Cody 1985 for details of several of these examples).

Figure 1. The islands of the Sea of Cortez, located between the peninsula of Baja California and mainland Mexico, offer a varied and relatively undisturbed setting in which to observe biogeographical patterns. The some two dozen larger islands and numerous smaller islets differ not only in size and isolation but in age and geological formation. Shown here is Partida Norte, a small island of about 1.2 km2 off the east coast of Baja, which is seen in the distance. In spite of its barrenness, Partida Norte supports four species of reptiles and seven species of land birds. Larger islands in the group are characterized by richer vegetation and many more species of vertebrates. (Photo by Martin L. Cody.)

Figure 2. The varied geological history of the islands in the Sea of Cortez, also known as the Gulf of California, aids biogeographical studies. The islands within the 110 m depth contour (light gray) were connected to the mainland within the last 13,000 years. By contrast, the islands of Raza and Tortuga, although about the same age, arose as oceanic volcanoes without any landbridge connection. The remaining islands are at least a million years old.

The debate over the relative importance of dispersal over water versus dispersal over land across temporary landbridges was heated at a 1961 symposium (Gressitt 1963), but a 1982 symposium on the same subject proceeded without polemics (see Diamond 1982). Three scientific developments in the intervening period had dramatically altered the nature of the debate. The first was the analysis of new geological data on continental drift and seafloor spreading, which revealed that some isolated and deepwater islands such as New Zealand, the Seychelles, and New Caledonia were formerly parts of much larger landmasses. At the same time, it became clear that many other volcanic islands had never been connected to any other landmass (Kennett 1982; Brown and Gibson 1983). The ancestors of present-day plants and animals on such islands as Hawaii, Tahiti, the Marquesas, the Galapagos, the Canary Islands, and the Azores all must have arrived by oversea dispersal.

Second, major advances have been made in applying molecular techniques to systematics and in calibrating the ages and relationships of species when the fossil record is insufficient. The use of these techniques has solved some long-standing problems. For example, molecular dating reveals that the endemic Hawaiian honeycreepers, Drepanididae, are descended from fringillid finches, and that they diverged from their mainland relatives 15 to 42 million years ago-long before the present-day Hawaiian Islands even existed (Sibley and Ahlquist 1982). We now know with reasonable certainty that the two species (and genera) of endemic Galapagos iguanas diverged around 20 million years ago on the South American mainland, each probably colonizing the Galapagos much later (Wyles and Sarich 1983).

The third major development, more conceptual in nature, was MacArthur and Wilson's equilibrium theory of island bigeography (1963, 1967), which realigned in a single sweep much of the history-oriented thinking on the ecology and biogeography of island systems. In the following sections we describe this theory, relating it to other current theories and especially to the view that historical factors are paramount. The remainder of the paper will compare the theories using a recent synthesis of data primarily from one source, the islands in the Sea of Cortez (Fig. 1; Case and Cody 1983).

The equilibrium theory

Very few assumptions are needed to show that the existence of such an equilibrium per se is virtually guaranteed. Whether or not such equilibriums are reached or even approached with any frequency in nature is, however, quite another matter. Various arguments shape the immigration and extinction curves, but the most parsimonious of them will produce curves continuously decreasing and increasing respectively over much of their range, and will thus produce an intersection and a stable equilibrium number of species S.

For ecologists, the question is not whether an equilibrium species number S exists, but rather whether the circumstances that determine it are constant over a sufficient time to render the equilibrium attainable or even approachable. For this to be so, the sizes and other characteristics of source biotas, the magnitude and shapes of immigration and extinction curves, and the sizes and positions of the islands must remain relatively constant over the "ecological" time required for population establishment, growth, and decline, even though we know that over geological time all of these factors will vary.

Some of this variation can be examined within the premises of the equilibrium model. The equilibrium is characterized by both the number of species and the rate of turnover; islands similar in the former may differ in the latter, since a small island near a colonization source might support the same number of species as a large, isolated island, but with a much higher rate of turnover. Further, more recent landbridge islands created by rising post-Pleistocene sea levels may still be losing species as a result of reduced immigration rates and increased extinction rates, and may be "supersaturated" relative to their size and isolation; such islands may require more time for the process of equilibration than the few thousands of years since their island status was achieved.

On the other hand, new islands that arise from beneath the sea following dramatic geological events will approach equilibrium conditions from below (S increasing) rather than from above (S decreasing), and again the time necessary for the number of species to increase through colonization might not yet have elapsed. More important, islands move on the backs of oceanic plates, with changing areas and degrees of isolation, through changing climates and habitats over time. Can an island's biota constantly remain "on track," in continuous equilibrium with its changing circumstances, or will there be a serious time lag between the physical changes and biotic responses? It is not difficult to envision a variety of circumstances in which the historical legacy might prevail.

Other biogeographical models

For mobile species with rapid colonization rates-most birds, bats, and perhaps winged insects-an equilibrium should be reached relatively quickly, and it is unlikely to be coincidental that these taxa have provided some of the best evidence for the equilibrium theory. For terrestrial mammals, amphibians, freshwater fish, and many woody plants with limited dispersal powers, such an equilibrium may require time periods equalling those of major geological or evolutionary events. Accordingly, examples of historically-determined island biotas seem to predominate in the literature of these taxa (see, for example, Raven and Axelrod 1974; Rosen 1975, 1978; MacFadden 1980).

The discovery of extinctions on islands is consistent with any number of island biogeographical theories. What is critical is whether the extinctions appear relatively constant over time (as predicted by equilibrium theory) or are episodic and associated with major geological or climatic changes. Monitoring the turnover rates of species might decide the issue, but many human lifetimes would be needed to distinguish regularity from spurts related, for instance, to climate. The fossil record could be brought to bear, but often the geological circumstances on oceanic islands are poorly suited to preserving fossils. For most small islands, only the relatively narrow time window provided by Holocene subfossils is available, and even this evidence is sporadic.

Subfossil documentation of extinctions has, however, been found on a number of islands, including Puerto Rico, the Bahamas, Barbuda, Antigua, Henderson, New Zealand, and the Canary and Hawaiian islands (Arnold 1976, 1980; Pregill 1981; Olson and James 1982; Steadman 1982). Although vertebrate extinctions are common, in many cases the temporal pattern and the interpretation of their causes are ambiguous. Man and his entourage of introduced mammals are strongly implicated in many instances (Steadman et al. 1984; Steadman and Martin 1984; Steadman and Olson 1985), as are climatic changes in others (Pregill 1982). Does any "equilibrium turnover" remain after the effects of such agencies are removed? As we explain below, the islands in the Sea of Cortez provide an important clue to this puzzle, since significant extinction has occurred there over the last 10,000 years in the absence of much human or climatically induced environmental change.

Besides the possibility that immigration and extinction rates are dominated by long-term historical effects, they may be poor predictors for other reasons. If such curves cannot be drawn as quasi-constant functions of the number of species on islands but are wholly stochastic, sporadic, and episodic, or if they are in general unrelated to island size and isolation, then an equilibrium species number either will not exist or will not be predictable from a simple description of the island. In this case patterns of diversity will appear vague and will be essentially uninterpretable by the equilibrium model.

A second group of models augments the equilibrium theory by focusing on species interaction and coevolution. One way in which immigration and extinction curves might be functionally dependent is by dint of interspecific interactions such as competition. Thus established species might reduce the chances that ecologically similar species can invade successfully, and may allow access only to ecologically compatible species in diminishing numbers over time. This is precisely what happens with laboratory "islands" in experimental aquariums, where a controlled sequence of colonization in a species pool of various bacteria, algae, and small metazoans results in a particular community which is stable and resistant to colonization by any of the missing species (Drake 1984).

Over longer periods of time, island communities might be altered by coevolutionary adjustments such that eventually the island becomes resistant to invasion, immigrations and extinctions approach zero, and the turnover of species declines steadily with time as island assemblages become more stable and finely tuned. This view was espoused by Lack (1976), who found evidence for it in the birds of the Canary and Caribbean islands, and who argued that short-term competitive exclusion and long-term coevolutionary adjustment of niches would produce balanced and stable sets of species. According to Lack, small islands have fewer and more generalized species because of the absence of richness of ecological opportunities, not because of higher extinction rates.

Coevolutionary models seem to find their best application to old islands that are intermediate in size and isolation. Random extinction must not be so great as to muddle the deterministic extinction predicted from interspecific competition, a requirement that excludes smaller islands. On the other hand, colonization rates must be both low enough that gene flow will not impede local evolutionary adjustments yet high enough that colonists are available, a requirement that points to combinations of taxa and islands characterized by intermediate isolation, such as lizards on the Canary Islands, the Lesser Antilles, and the islands in the Sea of Cortez (Case 1979; Roughgarden et al. 1983; Case and Bolger, unpubl.).

A third model hypothesizes that, as in Lack's theory, island communities are closely related to the diversity and abundance of the island's resources and have low to zero rates of species turnover, but (in contrast to Lack's hypothesis) are composed of species that do not interact. If resources are distinct and not shared among species there will be no interspecific competition for them, no increased densities where some species are absent, and no negative effect on immigration curves from existing island residents. There will be, however, a tight coupling of the island's resources with its consumer species that is free of species interaction. In such a situation, similar islands with similar resources would support not just similar numbers of species but the same kinds of species or even precisely the same species. Further, the densities of individual species would be predictable from a knowledge of the island's resources, which is not possible when species interactions permit complementary relationships and density compensation. With such a non-overlapping allocation of resources, chance extinctions would allow neither community reorganization nor easier access to new species but only a chance for recolonization by the same (recently extinct) species from mainland sources.

This noninteractive, resource-coupled model might be most appropriate where the diversity of the mainland species pool is low and its members ecologically distinct. Islands drawing their colonists from this pool should support various subsets of the mainland communities, subsets dependent on the degree to which the resources of the mainland habitats are represented on the islands. Islands of increasing size would presumably support increasingly large subsets of the mainland community, while similarly sized islands, given their similar resource base, would support nearly identical communities. If isolation is a factor in colonization, then more distant but similarly sized islands would be reached by fewer colonists. This model is likely to be supported where three conditions exist: a dearth of species that are ecologically similar, a single source of colonizing species (rather than several sources with ecologically equivalent species), and source habitats in which ecological segregation arises from interspecific differences in resource use and associated morphological or physiological differences (rather than from subtle behavioral differences, such as foraging in different subhabitats).

From this discussion it is clear that to expect a single biogeographical model to account for distribution and density patterns on all islands is hardly reasonable. Before backing a specific model, one would like to know the relative time frames both of biological events such as colonization, population establishment, and extinction and of geological events governing the island's origin, age, history, and climate. One would need to know, also, the way in which climate, substrate, and topography combine to produce the island's particular resources, as well how biological interactions affect the consumers of these resources.

The biogeography of the Gulf islands

The islands in the Sea of Cortez are exceedingly dry, hot, and rocky. Like the Baja peninsula to the west and the mainland Mexican states of Sonora and Sinaloa to the east, they support a predominantly Sonoran Desert vegetation, sparse, low, and open except for the few trees and taller cacti in and around the deeper arroyos. Plants represent the largest of the reasonably well-known taxonomic groups on the islands, with over 570 species recorded. About 100 of these are limited to the large island of Tiburon, only 2 km from the Sonoran mainland across a 3-fathom channel (Cody et al. 1983; Moran 1983). Approximately 300 plant species occur on Tiburon and nearly 200 on Angel de la Guarda, but totals are much more modest on the smaller islands. By contrast, the Sonoran Desert as a whole contains about 3,500 plant species.

Over 100 species of fish are primary residents of the shallow reefs around the islands, with up to two dozen occurring at a given collection site (Thomson and Gilligan 1983). Reptiles are diverse and conspicuous in deserts, and 130 taxa (including subspecies) have been recorded on the islands (Fig. 3). Twenty or more species are found on four of the larger islands, some of which were once connected to the mainland by landbridges (Case 1983a; Murphy 1983). More conspicuous, perhaps, but less diverse are the islands' land birds (Cody 1983). Thirty species breed on the islands, with 18 to 21 species found on six larger islands, as opposed to 21 to 31 species recorded at mainland census sites, comparable to Tiburon's total of 26 species. Finally, mammals are the least conspicuous and the least diverse, with scattered records of some larger species-deer, coyote, fox, ringtail cat, rabbits-for a few landbridge islands only (Lawlor 1983). Most island records show only pocket mice (Perognathus), white-footed mice (Peromyscus), or woodrats (Neotoma). Twenty-four mammal species in all are found in the islands, but most islands support only one to three species, with four to seven species recorded on landbridge islands and thirteen counted on Tiburon.

In both their geological variety and their relative freedom from the impact of man's activities, the islands of the Sea of Cortez offer an ideal opportunity to compare island biogeographical theories. Recent studies of the biogeography of these islands, summarized in Table 1, have given us an index of the overall diversity of six taxonomic groups (the fifth group, lizards, is a subset of the fourth, land reptiles). According to equilibrium theory, islands formerly connected to the mainland by way of landbridges approach equilibrium from above, with extinction exceeding immigration until the two rates eventually balance. As seen in the first column of Table 1, however, after roughly 10,000 years this historical difference between recent landbridge and old oceanic islands has been erased for the highly mobile taxa of plants, shorefish, and land birds. For the less mobile groups, on the other hand-reptiles and especially mammals-landbridge islands are currently richer in species, dramatically so in the case of mammals, and now appear supersaturated compared to oceanic islands of similar size.

Table 1. Biogeographical Comparison of taxonomic groups on islands in the Sea Of Cortez Taxa Average no. of species on 50 km2 oceanic islands Are landbridge islands species-rich relative to oceanic islands? Are islands species-poor relative to the mainland? Are Holocene oceanic islands species-poor compared to old oceanic islands? Does number of species decrease with distance? Endemism (%) Landbridge islands All oceanic islands Landbridge islands All oceanic islands Land plants 105 no* no, for islands >3 km2; yes, for all others no no no 0 ˜2 Shorefish - no no no no no 0 0 Land birds 13 no slightly no no no 0 0 Land reptiles 9.3 yes yes no no no 5 35 Lizards 3.5 yes yes no no yes 0 47 Land mammals 1.3 extremely extremely slightly slightly yes 16 69 *The answers "yes" and "no" imply statistical significance or its absence at p < 0.05. "Slightly" means that an effect exists but its magnitude is small, even though it may be statistically significant; "extremely" implies an effect of large magnitude.

Figure 4. Estimated extinction rates for reptiles and mammals on various landbridge islands off both sides of the Baja peninsula suggest that small islands lose reptile and mammal species (gray and black dots, respectively) at rates substantially greater than large islands, and that mammals become extinct faster than reptiles. The presumed initial number of species on each island was calculated from the observed relationship between area and number of species on the mainland today. The lines show the best fit through the points, which represent taxa on individual islands. The extinction rate is determined as K from the model ds/dt = -ks² (see Richman et al., in press).

Is there an "island effect" at all-that is, does the isolation that characterizes islands in general affect the diversity of different groups? Column three of the table shows that in plants this effect is seen only on very small islands, which with their small catchment areas and lack of drainage channels are qualitatively different from mainland areas, supporting less diversity; in larger islands the effect disappears. It is absent also in shorefish and is minor in birds, which are nearly as diverse on larger islands with vegetation comparable to mainland sites as on the mainland itself, regardless of the degree of isolation of the island. Only on the smaller islands does the diversity of birds fall, most likely because of the impoverishment of vegetation. In the case of reptiles, however, the paucity of species on islands is immediately apparent, and with mammals the island effect is even more pronounced. Here the fieldworker immediately recognizes differences between island and mainland, whereas with plants and birds a lower diversity is not obvious in a superficial survey. In these more mobile taxa, recolonizations from the mainland must be sufficiently frequent that local extinctions are compensated for either before or shortly after they occur. Column four of the table indicates that the two newest oceanic islands, Raza and Tortuga, have received ample immigration; only in the case of mammals are there fewer species than would be expected for islands of their size.

The phenomenon of frequent recolonization can produce lower extinction rates on islands that are closer to the mainland as opposed to those that are more isolated (cf. Schoener 1983), resulting in a "distance effect"-that is, on more isolated islands there will be fewer species in taxa with more limited mobility. Column five of the table confirms that on landbridge islands this effect is detectable only in mammals; on oceanic islands it is more pronounced, but only in the least mobile taxa, lizards and mammals.

Landbridge islands that have retained their historical legacy and still remain supersaturated thousands of years after gaining (or regaining) island status can be used to calculate the relationship between extinction rates and island size (see, for example, Diamond 1972; Wilcox 1980). The necessary condition of low to zero immigration is most closely satisfied in reptiles and mammals on Gulf landbridge islands; the relationship of extinction rates to island size, calculated in Figure 4, is similar in the two taxa, although reptiles persist much better over time than do mammals.

These results are important because of the controversy surrounding the cause of island extinctions. While many biogeographical theories predict extinctions with changes in the physical and climatic setting, equilibrium theory predicts the turnover of species even in the absence of such physical changes. In theory this essential difference provides a way of testing the various models, but in practice it is next to impossible to find a set of islands without environmental change or evidence of the severe impact of man's activities. The islands in the Sea of Cortez probably come closer than any to fulfilling the requisite conditions. There was never any permanent aboriginal population except on Tiburon, and the presence of modern man is restricted to a few small settlements on three of the larger islands. Relatively few islands have any plants or animals introduced from elsewhere, and the effect of such imports has been severely limited by the inhospitable climate and the absence of standing water.

Richman and his colleagues (in press) have shown that extinction rates for reptiles of the Gulf islands are nearly identical to those calculated for relatively undisturbed landbridge islands off South Australia during the same period. The extinction rates for Gulf reptiles are, moreover, not very different on average from rates for other islands based on the direct evidence of subfossil deposits. The important difference is that small oceanic islands with substantial human disturbance exhibit significantly elevated extinction rates compared to undisturbed islands. Calculating probable separation times ranging from 6,000 to 14,000 years for the landbridge islands in the Sea of Cortez, Wilcox (1978) showed that the number of lizard species "relaxed" over this interval: older landbridge island have lost relatively more lizard species than have islands isolated for shorter periods of time. While there were some procedural and statistical problems with Wilcox's original methods (Faeth and Connor 1979), a "time effect" still emerges after correcting for these. This effect is not seen in mammals (Lawlor 1983), however, perhaps because mammalian extinctions occur more rapidly, so that the number of species on even the youngest landbridge islands has already declined to levels appropriate for their present areas.

The question of the emergence of new species on islands via evolution rather than colonization is examined in the last column of Table 1. For endemic populations to form on islands, colonization rates must be sufficiently low that island populations are not genetically swamped by immigrants with mainland genes, and in addition island populations must persist long enough for evolution to act upon them. If these two conditions are satisfied, a third is necessarily met: since islands for which colonization rates are low for a given taxon will be species-poor, selective pressures for genetic change will be enhanced for the successful colonists. Table 1 shows that endemism at the level of species or above is zero for shorefish and birds, and zero on landbridge islands and extremely low (˜2%) on oceanic islands for plants. It reaches appreciable levels only in the least mobile taxa, reptiles and mammals. On landbridge islands, where swamping by mainland genomes is most recent, levels of endemism are modest, but on oceanic islands levels are high, increasing from 35% for reptiles to 46% for lizards to 69% for mammals.

The effect of island size on the level of endemism is equivocal. Although larger islands will support the longer persistence times necessary for endemics to form, they will also have higher colonization rates and a community structure more similar to the mainland, and thus more swamping by incoming genes and less impetus for evolutionary change. In fact, in lizards, the only group large enough to test for effects of island size on endemism, there is no such effect (Case 1983a). The effect of island age on endemism can be studied by comparing the two Holocene islands, Raza and Tortuga, with other, much older islands. Again, the expected degree of endemism is equivocal. Endemism might be expected to be low because Raza and Tortuga are young oceanic islands but high because they are well-isolated, with low colonization rates. In fact, levels of endemism in lizards on these two islands-50% and 25%, respectively-are typical of those on much older oceanic islands and significantly greater than those of equally young landbridge islands, a result which suggests that isolation is more important than time in producing speciation.

Finally, let us examine evidence for species interactions and ask whether such interactions have contributed to the observed patterns of distribution and density on the Gulf islands. When records on land birds on the Gulf islands are compared, going from the smallest to the largest islands, the sets of species are simply subsets of the species occurring on the next larger island. This is just the sort of pattern that would be expected from the non-interactive species model and from the Sonoran Desert vegetation, with its collection of relatively specialized bird species showing dissimilar morphologies and little ecological overlap in foraging habits. One of the few possibilities for interspecific competition occurs between two kinds of small foliage insectivores, verdins (Auriparus flaviceps) and gnatcatchers (Poliopthus spp.), but in fact these insectivores show either identical (on the southern islands) or nearly identical (on the northern islands) distributions as well as similar variations in density over island size, with verdins about twice as dense as gnatcatchers. Another possible instance of interaction is found in three species of the thrasher Toxostoma. Out of ten islands apparently large enough to support the taxon, six have only a single species--a result that does not differ significantly from what would be expected from chance if no interaction were present.

Figure 5. In this chart of elements governing diversity, the axes of colonization ability and persistence ability are paralleled by axes of increasing distance and increasing island area, respectively. Levels of endemism are shown as a joint function of these axes, as is the "island effect"-the effect of insularity on diversity. The biological attributes of colonization ability and persistence tell most of the story. High colonization ability and low persistence result in minimal endemism (birds), whereas the reverse pattern produces maximal endemism (reptiles and mammals). Where both colonization ability and persistence are high, minimal island effect is found (plants); where both elements are low, the island effect is at its maximum (mammals). As the chart suggests, islands of similar size should be relatively poor in mammals but rich in plant species; lizards will have high levels of endemism and birds relatively low levels. For all taxa, increasing island area will increase both the richness of species and the level of endemism, whereas increasing distance from the mainland will decrease the richness of species but increase endemism.

The same high degree of predictability of particular species from island size that occurs among the land birds (Murphy 1983) is found also in lizards, indicating that many of the lizard species are likewise noninteractive and resource-coupled. But this does not preclude the possibility that some lizard species are competitively interactive, with overlapping resource requirements. For instance, body sizes and population densities of Uta stansburiana, the most ubiquitous species, decline with increasing numbers of other related species with which it coexists (Soule 1966; Case 1983a). Moreover, the overall co-occurrence of lizard species on the islands cannot be interpreted simply as random subsets from the mainland species pool; certain lizard species with low ecological overlap occur together on islands statistically more often than chance would dictate (Case 1983a, 1983b). Particular pairs of lizard species reveal strong patterns of negative distribution, most obviously in the case of Uta and Sator, which occur together on no islands-a result that would be expected by chance less often than once in a million cases. In laboratory cages, the larger Sator dominates and frequently kills and eats the smaller Uta.

Another example of interspecific interaction, of a milder nature, is seen in the whiptail lizards (genus Cnemidophorus), which are diurnal, actively foraging insectivores. On the peninsula two species fill this niche, both occupying the same habitats, but C. hyperythrus is small and C. tigris is substantially larger; unlike Sator and Uta, these two species rarely fight. Their ecological segregation depends on the larger lizards eating larger prey (see Case 1979). Knowing the size distribution of available prey, it is possible to predict the optimal sizes of lizards in one- or two-species sets at a given locality. Some islands support just one Cnemidophorus species, on the six oceanic islands where the single species is C. tigris, it is smaller in size than on the mainland, and on two islands where C. hyperythrus or its derivatives occur alone, it is larger. On some landbridge islands C. tigris occurs alone, but no reduction in body size is apparent. Perhaps C. hyperythrus has only recently become extinct there, or perhaps the evolution of smaller body size requires more time than has been available on these recent islands; alternatively, gene flow from the mainland might have precluded a shift to smaller body size in these lone C. tigris populations.

A diversity of models

As work continues on these islands, it will be interesting to see whether the generalities of Figure 5 survive, and how other taxa might fit into the scheme. Ongoing work by Gary Polis on the scorpions of the islands and the mainland suggests that their colonization abilities are generally better than those of reptiles but not as good as birds, and that their extinction rates are similar to those of reptiles. The average Gulf island of 50 km² has about six scorpion species, but (unlike any other taxa studied so far) there is a sharp geographic increase in the richness of species from north to south on the islands, a latitudinal trend not paralleled on the peninsula. Endemism is not quite as high for scorpions as for reptiles. Similar work is needed on other invertebrates; ants and land snails would make exciting candidates, because the peninsular fauna is rich and highly endemic. The time is also ripe for the use of molecular techniques to clarify the ages and relationships of the many endemic species of the region. From what we have learned so far, one warning seems dear: future island biogeographers must be very cautious in generalizing beyond the bounds of their taxon and the island system in which it has been studied.

Ted J. Case is professor of biology at the University of California at San Diego, where he has been since 1978. He received his Ph.D. from the University of California at Irvine in 1973. He is associate editor for the journals Evolution and Oecologia, and has been investigating reptiles in the Sea of Cortez since 1970. Martin L. Cody is professor of biology at the University of California at Los Angeles. He received his Ph.D. from the University of Pennsylvania in 1966. His research in the Sea of Cortez has focused on the ecology of land birds and plants. The work reported here was sponsored by the NSF and the National Geographic Society. Address for Professor Case: Department of Biology, C-016, University of California at San Diego, La Jolla, CA 92093.

References

Arnold, E. N. 1980. Recently extinct populations from Mauritius and Reunion, Indian Ocean. I. Zool. 191:33-47.

Brown, J. H. 1971. Mammals on mountaintops: Non-equilibrium insular biogeography. Am. Nat. 105:467-78.

Brown, J. H., and A. C. Gibson. 1983. Biogeography. St. Louis: Mosby.

Case, T. J. 1979. Character displacement and coevolution in some Cnemidophorus lizards. Fortschr. Zool. 25:235-82.

Case, T. J. 1983a. The reptiles: Ecology. In Island Biogeography in the Sea of Cortez, ed. T. J. Case and M. L. Cody, pp. 159-209. Univ. of California Press.

Case, T. J. 1983b. Assembly of lizard communities on islands in the Sea of Cortez. Oikos 41:427-33.

Case, T. J., and D. Bolger. Unpubl. The role of species interactions in the island biogeography of reptiles.

Case, T. J., and M. L. Cody, eds. 1983. Island Biogeography in the Sea of Cortez. Univ. of California Press.

Cody, M. L. 1983. The land birds. In Island Biogeography in the Sea of Cortez, ed. T. J. Case and M. L. Cody, pp. 210-45. Univ. of California Press.

Cody, M. L., ed. 1985. Habitat Selection in Birds. Academic Press.

Cody, M. L., R. Moran, and H. Thompson. 1983. The plants. In Island Biogeography in the Sea of Cortez, ed. T. J. Case and M. L. Cody, pp. 49-97. Univ. of California Press.

Darwin, C. R. 1859. On the Origin of Species by Means of Natural Selection. London: Murray.

Diamond, J. M. 1972. Biogeographic kinetics: Estimation of relaxation times for avifaunas of Southwest Pacific Islands. PNAS 69:3199-203.

Diamond, J. M. 1974. Colonization of exploded volcanic islands by birds: The supertramp strategy. Science 184:803-05.

Diamond, J. M. 1982. The biogeography of the Pacific Basin. Nature 298:604-05.

Drake, S. A. 1984. Ecological communities as assembled structures. Ph.D. diss., Purdue University.

Faeth, S. H., and E. F. Connor. 1979. Supersaturated and relaxing island faunas: A critique of the species-age relationship. J. Biogeography 6:311-16.

Gressitt, J. L., ed. 1963. Pacific Basin Biogeography. Honolulu: Bernice P. Bishop Museum Press.

Kennett, J. P. 1982. Marine Geology. Prentice-Hall.

Lack, D. 1976. Island Biology Illustrated by the Land Birds of Jamaica. Univ. Of California Press.

Lawlor, T. 1983. The mammals. In Island Biogeography in the Sea of Cortez, ed. T. J. Case and M. L. Cody, pp. 265 - 89. Univ. of California Press.

MacArthur, R. H., and E. O. Wilson. 1963. An equilibrium theory of insular zoogeography. Evolution 17:373-87.

MacArthur, R. H. 1967. The Theory of Island Biogeography. Princeton Univ. Press.

MacFadden, B. J. 1980. Rafting mammals or drifting islands?: Biogeography of the Greater Antilles insectivores Nesophontes and Solenodon. J. Biogeography 7:11-22.

Moran, R. 1983. Vascular plants of the Gulf islands. In Island Biogeography in the Sea of Cortez, ed. T. J. Case and M. L. Cody, pp. 348-81. Univ. of California Press.

Murphy, R. W. 1983. The reptiles: Origins and evolution. In Island Biogeography in the Sea of Cortez, ed. T. J. Case and M. L. Cody, pp. 130-58. Univ. of California Press.

Olson, S. L., and H. F. James. 1982. Prodromus of the Fossil Avifauna of the Hawaiian Islands. Smithsonian Contributions in Zoology, no. 365: 1-79.

Pregill, G. K. 1981. An appraisal of the vicariance hypothesis of Caribbean biogeography and its application to West Indian terrestrial vertebrates. Syst. Zool. 30:147-55.

Pregill, G. K. 1982. Fossil amphibians and reptiles from New Providence Island, Bahamas. Smithsonian Contributions in Paleobiology, no. 48:8-21

Raven, P. H., and D. I. Axelrod. 1974. Angiosperm biogeography and past continental movements. Annals of the Missouri Botanical Garden 61:539-673.

Richman, A., T. J. Case, and T. Schwaner. In press. Natural and unnatural rates of extinction. Am. Nat.

Rosen, E. D. 1975. A vicariance model of Caribbean biogeography. Syst. Zool. 24:431-64.

Rosen, E. D. 1978. Vicariant patterns and historical explanations in biogeography. Syst. Zool. 27:159-88.

Roughgarden, J., D. Heckel, and E. Fuentes. 1983. Coevolutionary theory and the biogeography and community structure of Anolis. In Lizard Ecology: Studies of a Model Organism, ed. R. B. Huey, E. R. Pianka, and T. W. Schoener, pp. 371-410. Harvard Univ. Press.

Schoener, T. W. 1983. Rate of species turnover decreases from lower to higher organisms: A review of the data. Oikos 41:372-77.

Sibley, L. G., and J. E. Ahlquist. 1982. The relationships of the Hawaiian honeycreepers (Drepaninae) as indicated by DNA-DNA hybridization. Auk 99:130-40.

Soule, M. 1966. Trends in the insular radiation of a lizard. Am. Nat. 100:47-64.

Steadman, D. W. 1982. Fossil birds, reptiles, and mammals from Isla Floreana, Galapagos Archipelago. Ph.D. diss., Univ. of Arizona.

Steadman, D. W., and P. S. Martin. 1984. Extinction of birds in the late Pleistocene of North America. In Quaternary Extinctions: A Prehistoric Revolution, ed. P. S. Martin and R. G. Klein, pp. 466 77. Univ. of Arizona Press.

Steadman, D. W., and S. L. Olson. 1985. Bird remains from an archaeological site on Henderson Island, South Pacific: Man-caused extinctions on an "uninhabited" island. PNAS 82:6191-95.

Steadman, D. W., G. K. Pregill, and S. L. Olson. 1984. Fossil vertebrates from Antigua Lesser Antilles: Evidence for late Holocene human-caused extinctions in the West Indies. PNAS 81:4448-51.

Thomson, D., and M. R. Gilligan. 1983. The rocky-shore fishes. In Island Biogeography in the Sea of Cortez, ed. T. J. Case and M. L. Cody, pp. 98-129. Univ. of California Press.

Turner, J. S., C. N. Smithers, and R. D. Hoogland. 1968. The Conservation of Norfolk Island. Australian Conservation Foundation Special Publ. 1.

Wilcox, B. A. 1978. Supersaturated island faunas: A species-age relationship for lizards on post-Pleistocene land-bridge islands. Science 199:996-98.

Wilcox, B. A. 1980. Insular ecology and conservation. In Conservation Biology, ed. M. E. Soule and B. A. Wilcox, pp. 95-118. Sinauer.

Wyles, J. S., and V. M. Sarich. 1983. Are the Galapagos iguanas older than the Galapagos? Molecular evolution and colonization models for the archipelago. In Patterns of Evolution in Galapagos Organisms, ed. R. I. Bowman and A. E. Leviton, pp. 177-86. San Francisco: AAAS.

This page © Ross E. Koning 1994.

Go to the Course Schedule Page.

Go to the Plant Physiology Information Homepage.

Send comments and bug reports to Ross Koning at rkoning@snet.net.