On the Risk of Extinction

Well-known theoretical predictions are that the risk of extinction should decrease with maximum population size (K) and should increase with the temporal coefficient of variation in population size (CV). In an unvarying environment, where extinction is caused solely by demographic accidents, the risk of extinction should decrease steeply with K; the greater the contribution of environmental variability to the risk of extinction, the less steep should be the dependence on K. Large-bodied species tend to have long lifetimes but low rates of increase, which have opposite effects on the risk of extinction per year. We show that in comparisons of a large- and small-bodied species at the same average population size (N), the large-bodied species should be at less risk at low N but at greater risk at high N. We test these predictions using a data base of short-term survivals (up to a few decades) of 355 populations belonging to 100 species of British land birds on 16 islands. The mean N and risk of extinction are known for these populations, and we can calculate CV's for 39 of the species. To identify how factors other than N affect the risk of extinction, we devise a means of correcting that risk for much of the effect of N. We make the following observations. (1) Risk of extinction does decrease sharply with N. (2) After correcting for much of the effect of N, we confirm the theoretical prediction that the relative susceptibility to extinction of large- and small-bodied species reverses with increasing population size. Above seven pairs, larger-bodied species are at greater risk than smaller-bodied species; the reverse is true below seven pairs. (3) Migratory species are at greater risk of extinction than resident species. (4) Finally, after accounting for the effects of N, body size, and migratory status, we show that the risk of extinction does increase with the CV.