The equilibrium model of the theory of island biogeography (MacArthur & Wilson 1967) postulates that the number of species on an island is the result of an equilibrium between two counterbalancing forces: extinction and immigration. Because populations fluctuate in size, below a certain level the population -- and possibly the species in a localized area -- will become extinct. The smaller the island, the more likely extinction becomes, due to smaller population size. Reestablishment of the species depends on immigration of new individuals into the area, which varies with how isolated the island is. The data collected in this study can be used to test two predictions derived from the equilibrium model.

Species-Isolation Effect
The first prediction is that for a set of kipukas of similar size, the more isolated kipukas will have lower rates of colonization, and therefore fewer species, than kipukas closer to the mainland.

This prediction is not true for plant species on kipukas. This is probably because the persistence of individuals of perennial species is long relative to immigration rates. The isolation effect would be important only in systems where short-lived plants predominate. In addition, the lava apparently does not provide a broad enough barrier to prevent plant dispersal.

This prediction is also not true for animals species. The location and numbers of small mammals captured permits a rough evaluation of how significant a barrier to dispersal the lava is. Four species made up 96% of all small mammal captures. The deer mouse, by far the most common species trapped in any habitat, comprised 83% of all captures. Bushy-tailed woodrats nest and breed on the lava and their distinctive urine markings were seen at the perimeter of every kipuka. Mountain cottontail ranges widely over lava, for we saw cottontails and rabbit scat on lava far from any kipuka. I observed least chipmunk a number of times on lava and trapped them more often than any species except deer mice. Although the lack of vegetation on the lava may limit population sizes of these animals, it apparently does not restrict their presence or ability to disperse. These four species have a very low probability of extinction on kipukas, because there is a permanent pool of potential colonists in the lava nearby to move into vacancies in the favorable kipuka habitats. Obviously, there will be no isolation effect if lava is not a barrier to the animals, and none was observed.

It is much less likely that the other species trapped--Great Basin pocket mouse, Townsend's ground squirrel, montane vole, and western harvest mouse -- actually reside on lava. In no instance did I see or capture any of these species on lava. On the other hand, they were not very common on kipukas either. I captured only 24 individuals of these four species, in most cases one per kipuka. Given the high trapping intensity on kipukas, it is fair to assume that single individuals are immigrants to the kipukas rather than representatives of established populations.

Although seven individuals are not enough for a firm conclusion, a weak isolation effect was apparent for these immigrants. It is suggestive that four of the immigrants were captured in an area sandwiched between two major species sources and another less than 100m from the mainland. Both of the immigrants caught on isolated kipukas were ground squirrels, which appear to disperse better than the other three rare species.

Isolation seemed to play no role in the number of reptile species present on a kipuka. But, in light of the abundance of the desert short-horned lizard on mainland close to the lava's edge, its total absence on kipukas suggests that lava is more of a barrier to the dispersal of this slow-moving reptile than for its quicker relatives.

Density-Isolation Effect
One interesting and unexpected isolation effect does exist: the density of animals increases, rather than decreases, with isolation. There are several possible explanations for this phenomenon. Perhaps animals nest in lava but forage communally on a the richer food sources of a nearby kipuka. Trapping the kipuka but not the lava would produce the impression of high density, when, in reality, animals that resided elsewhere were being counted. The fact that the isolated kipukas show almost no trapping-out effect supports this hypothesis. Generally, when trapping small mammals, the number of individuals captured declines steeply from the first to succeeding days. If animals trapped were coming from outside of the trapping area this decline wouldn't occur.

Another possible mechanism for increased density on isolated kipukas involves predation. Predation pressure on kipukas is likely to be intermittent, particularly on small kipukas. These kipukas are probably not large enough to support a snake population, individual coyote, or other large predators. So, in between periods of intense predation, kipukas may enjoy periods of total release from predation pressure. During these lulls, small mammal populations may build up to densities that greatly exceed densities on less isolated kipukas, where predators may forage more frequently.

Species-Area Effect
For most species, the larger a kipuka the higher the number of individuals of any species that can live there. If the MacArthur-Wilson model is correct, larger populations sizes will result in noticeably lower rates of extinction, and therefore higher numbers of species. The second prediction of the MacArthur-Wilson model, then, is that larger kipukas will have larger numbers of species. This prediction is true for both plant and animals species on kipukas.

The MacArthur-Wilson model is also relevant in accounting for the number of small mammal species on kipukas. The relevance of the Mac Arthur-Wilson model is apparent in an analysis of minimum area requirements necessary for maintenance of populations. Of the small mammal species observed on kipukas for which lava poses an impediment to dispersal, three species had established populations. None of these was on a kipuka less than three hectares in size. The species involved did not appear to be restricted to a particular type of habitat unavailable on smaller kipukas, nor was the increase in environmental heterogeneity on kipukas of larger size important to the individual species. Instead it is possible that these species need an area of at least three hectares in this ecosystem to support a large enough number of individuals to maintain a population. Perhaps those species without populations on the study kipukas have minimum area requirements greater than three hectares.

There are alternative explanations for the absence of certain small mammal species. There might be a lack of suitable habitat on kipukas for some species. For instance, Ord's kangaroo rat is generally associated with sandy soils for digging, and only one trapped kipuka met this criterion. If a species were absent from the mainland immediately adjacent to the lava flow, there would be no source of immigrants to colonize kipukas. Lava might be a very effective, perhaps impenetrable barrier to some species. Probably the observed distribution of small mammal species is the result of a number of such mechanisms operating together.

Summary of Results

1. Contrary to the prediction of the MacArthur-Wilson model, the number of plant species on a kipuka did not depend on how isolated it was. Lava probably poses a scant barrier to plant dispersal, so that the persistence of individuals of perennial species is long relative to immigration rates.

2. Contrary to prediction, the number of small mammal species on a kipuka did not depend on how isolated the kipuka was. Data indicate that, although lava restricted the dispersal of some species, it did not restrict others. These other species made up over 96% of all small mammal captures. Considering only the 24 captures of individuals of those species that lava did restrict, there was an isolation effect.

3. The number of plant species on a kipuka depended on the size of the kipuka, as predicted by the MacArthur-Wilson model.

4. The number of small mammal species on a kipuka depended on the size of the kipuka, as predicted by the MacArthur-Wilson model. Considering only those small mammal species restricted by lava heightened the species/area effect.

6. More isolated kipukas had higher densities of small mammals than less isolated kipukas. There are two plausible explanations for this result. Faunas of small, more isolated kipukas may have had longer periods of release from predation pressure. Or small mammals on isolated kipukas may nest in the lava but forage on kipukas, so that trapping just the kipuka reveals an artificially high density.