Résumé
The risk of extinction of populations has not previously been empirically related to parameters characterizing their population dynamics. To analyze this relationship, we simulated how the distribution of population dynamical characters changed as a function of time, in both the remaining and the extinct populations. We found for a set of 38 bird populations that environmental stochasticity had the most immediate effect on the risk of extinction, whereas the long-term persistence of the population was most strongly affected by the specific population growth rate. This illustrates the importance of including information on temporal trends in population size when assessing the viability of a population. We used these relationships to examine whether time to extinction can be predicted from interspecific life history variation. Two alternative hypotheses were examined. (1) Time to extinction should decrease with increasing clutch size or decreasing survival rate because of the larger stochastic components in the population dynamics of such species. (2) Time to extinction should increase with decreasing clutch size or longer life expectancy if extinction rates are most strongly influenced by variation in the specific population growth rate. In the present data set, time to extinction increased with decreasing clutch size because of larger stochastic influences on the population dynamics of species with large clutch sizes located toward the fast end of the “slow–fast continuum” of life history variation. This demonstrates that interspecific variation in extinction risk can be predicted from knowledge of general life history characteristics. Such information can therefore be useful for assessing minimum sizes of viable populations of birds.