Document View

Ricklefs and Bermingham examine mitochondrial DNA sequences from 161 island populations of 37 small land bird species in the Lesser Antilles. From these, they estimated the average genetic divergence from the closest sister populations on Trinidad or the Greater Antilles.

Human destruction of populations, species, and habitats in the hot spots of biodiversity is causing an extinction crisis (1). To quantify this crisis, we must estimate how often extinctions happen in the absence of human pressures. Although aggregated studies of the fossil record reveal rates of 0.1 to 1 extinction per year per million species (2), there is not really any "normal" extinction rate. Rather, the severity of background extinction events is inversely proportional to how often they occur; major catastrophic events are rare (3). Thus, a plot of magnitude versus frequency has the form of a hollow curve (see the figure). For life to persist, the processes that generate biodiversity-speciation and, locally, colonization-must keep pace with these extinctions. The archipelago of Caribbean islands called the West Indies has always been a fertile testing ground for those who study these processes, because of the islands' complex geographic history and propensity for catastrophe. Two elegant studies in this issue (4, 5), carried out at opposite ends of the Caribbean island arc, throw new light on extinction (and hence conservation) at opposite ends of the magnitude-frequency curve.

On page 1522 of this issue, Ricklefs and Bermingham examine mitochondrial DNA sequences from 161 island populations of 37 small land bird species in the Lesser Antilles (4). From these, they estimate the average genetic divergence from the closest sister populations on Trinidad (and South America) or the Greater Antilles. To assess the pattern of colonization over time, they plot the cumulative number of lineages against increasing genetic divergence. Strikingly, whereas around half of the lineages differed only slightly from their presumed sister stock, with genetic divergence of 2% or less, the other half showed high genetic divergence values from 2% up to 15%. Assuming that genetic divergence increases uniformly over time, and calibrating this molecular clock from the literature, Ricklefs and Bermingham found that a dramatic change in the mean age of the Lesser Antillean avifauna apparently occurred a little over half a million years ago. The authors suggest two possible causes for this change. One would be a mass extinction event (perhaps due to a tsunami or the impact of a meteorite) superimposed on a background of high colonization. Alternatively, a sudden increase from a low to a high colonization rate (perhaps due to increased exposure of land during periods of lowered sea levels) would have had the same effect. They conclude that the existing Lesser Antillean avifauna is primarily a product of such historical events, with little post-catastrophic recovery.

On a much finer scale, the work of Schoener et al. on page 1525 reveals a very different story (5). They tracked the fate of a single lizard species, the Cuban brown anole Anolis sagrei, on 66 individual islands in the Bahamas, ranging in size from 10 to 10,000 m2. Initially, island area was the single determinant of presence or absence of the species. Then, in September 1999, Hurricane Floyd extirpated lizards from 37 of the 49 islands that they had inhabited before the catastrophe. Elevation proved to be the determinant of adult survival through the hurricane and especially through the associated 3-m storm surge. On islands lower than 3 in, all hatched lizards perished. As a fascinating aside, the authors discovered recently hatched lizards soon after the hurricane on 10 islands from which all adults were eliminated. Subsequent experimentation revealed that eggs of A. sagrei can in fact survive for 6 hours submerged in seawater-a surprising physiological result as well as a key factor in allowing such a rapid recovery rate. Crucially, however, Schoener et al. then found that island area progressively increased in importance as a predictor of the presence of this lizard species. Within only 2 years, a species-area relationship-similar to that before the hurricane-was reestablished on the islands.

Enlarge 200% Enlarge 400%

What might explain these apparently contradictory results? The obvious possibility is scale (see the figure). Although the time lag for recolonization has yet to be shown to be related to area, the converse situation-that of extinction after habitat loss-is certainly scale dependent (6). It therefore appears likely that the impact of catastrophes involving relatively simple communities and covering relatively small areas, such as Simberloff Is classic fumigation of islets in the Florida Keys (7), can be experimentally observed to decrease in importance over time. On the other hand, huge, very rare catastrophes affecting entire regions are likely to remain imprinted in local community structure for millennia (8).

The conservation implications of these studies are clear. The Caribbean is already one of the world's hottest hot spots, retaining only just over 10% of its original forest cover (1). Habitat across the archipelago has been reduced to tiny patches-islands within islands (9)--aggravating the vulnerability to catastrophes inherent in the region's geographic subdivision. Species extinctions resulting from human pressures have already struck the West Indies on a massive scale. Of the 197 endemic mammals and birds across the islands (1), at least 43 have become extinct over the last 500 years (10). This equates to nearly 500 extinctions per year per million species, three orders of magnitude higher than expected given species' lifetimes in the fossil record (2). Worse yet, 84 more Caribbean endemic mammals and birds are classified on the Red List as threatened with a high probability of extinction in the medium-term future (10). Seen from a gloomy perspective, these species represent an extinction debt-losses already under way after habitat destruction. Worst of all, the remaining habitat patches of the Caribbean are small (and getting smaller), and so, given that the rate of extinction after habitat loss is scale dependent (6), these extinctions will probably occur soon.

The studies of Schoener et al. and of Ricklefs and Bermingham do, however, cast one ray of hope for conservation of the Caribbean's unique biodiversity. Imagine a conservation vision across the region, with the land- and seascape of surviving habitat fragments connected within a matrix of benign land use by "corridors" (11). Such corridors would consist not only of restored habitat and zones of low-impact human activity, but also, as Schoener et al. indicate, interdependent systems of tiny, largely pristine, islands. Recall that to reconcile the two studies, we invoke scale dependence in the persistence of the impact of historical catastrophe. If this is correct, then surely the recolonization of tiny habitat fragments across the conservation landscape would be rapid, analogous to the situation illustrated by Schoener et al. Obviously, it is too late for groups such as the West Indian macaws, already forced into catastrophic extinction (12). For the large portion of Caribbean biodiversity currently threatened with extinction, though, these studies suggest that all is not yet lost-as long as conservation can be implemented on an unprecedented scale across the region.