On Expected Lifetimes of Small-Bodied and Large-Bodied Species of Birds on Islands

"A major problem in population biology is to understand what determines the expected lifetime of a population" (Pimm et al. 1988, p. 757). Theoretical models and a score of empirical studies agree that large populations tend to have a lower risk of extinction and, hence, longer expected lifetimes than small ones (reviewed by Diamond [1984a]). The consequences of several other factors for the risk of extinction are much more controversial. These include generation or individual lifetime (Diamond 1984a), intrinsic rate of population increase (Leigh 1981), ratio of birth to death rates (MacArthur and Wilson 1967; Richter-Dyn and Goel 1972), variability in population size (Leigh 1981; Goodman 1987; Schoener and Spiller 1992), and body size (Brown 1978; Diamond 1984a; Patterson 1984). Pimm et al. (1988) analyzed the risk of extinction in bird populations on 16 small islands off the coast of Britain. Their main new empirical finding was that the risk of extinction decreases more rapidly with increasing population size in small-bodied than in large-bodied species of birds. They also predicted this result on two grounds. First, when population size is very small, at the limit of only one individual, large-bodied species have a lower per-year extinction rate than small-bodied ones because of their greater longevity. Second, when population size is greater, "the large-bodied species is at a disadvantage because, following some severe reduction in numbers, its lower intrinsic rate of increase makes it take longer than the small-bodied species to climb from low levels to higher numbers" (Pimm et al. 1988, p. 761). One of us has recently analyzed the risk of extinction in shrew populations on small islands in lakes (Peltonen and Hanski 1991; Hanski 1992, 1993). In shrews, we found a pattern opposite to that reported by Pimm et al. (1988). Notwithstanding that data were available for only four species, extinction risk decreased rapidly with increasing population size in the largest-bodied species but much more gradually in the smaller-bodied species. This was expected because the smallerbodied species are more sensitive than the larger-bodied ones to environmental stochasticity, especially to temporal variation in food availability (Hanski 1992). We have thus two contradictory sets of empirical results and two different supporting arguments to demonstrate that the expected lifetime of a population increases more, or less, rapidly with population size in small-bodied than in large-