Natural and Unnatural Extinction Rates of Reptiles on Islands

Abstract

When faced with the problem of interpreting the significance of an extinction event, it is easiest to grasp the explanation closest to hand in the undeniable possibility that conditions changed enough that extinction was the inevitable result. In many instances, however, local extinctions may have been driven by an inconstant environment rather than by a permanently altered one. Reestablishment of a population is expected in this case, unless the taxon is unable to recolonize because of poor dispersal ability or the lack of a nearby source of colonists. The reptile faunas of land-bridge islands are an excellent example of this scenario. For the land-bridge faunas of South Australia and Baja California [Mexico], we find that a background extinction rate exists even in the absence of the kinds of disturbance usually invoked to explain extinction events. There is little evidence for large-scale alteration of environment since the end of the Pleistocene for landbridge islands of the Baja Peninsula or offshore South Australia. Moreover, the basic contention that climatic change inevitably results in extinction is not supported by available evidence; terrestrial reptiles in the western United States often did not respond to large changes in habitat with consequent shifts in species ranges, let alone extinction. The possibility that human disturbance in the recent or distant past has affected local extinctions on the two groups of land-bridge islands is similarly without evidence. Modeling extinction rate as an extra-linear function of the number of species, we find a significant, negative correlation between the extinction parameter and area, consistent with a model of stochastic extinctions in which reduced population size increases vulnerability to local extinction. Lastly, the extinction risk on islands with a history of extensive man-made disturbance is elevated above that of relatively pristine land-bridge systems. The magnitude of extinction risk declines with area for the disturbed islands, just as was found for land-bridge islands, suggesting that human disturbance simply exacerbates the process of island extinction already in place.