Introduction

Recent increases in the number of time-series long enough to provide an adequate description of population fluctuations clearly show that population stability varies widely among animals with similar longevities and rates of reproduction, as well as between species with contrasting life histories (Caughley & Krebs 1983; Gaillard et al. 2000). For example, among grazing ungulates, populations may either show little variation in size across years, irregular oscillations, semi-regular oscillations resembling the stable limit cycles found in some smaller mammals or dramatic oscillations occasionally leading to extinction (Peterson et al. 1984; Fowler 1987b; Coulson et al. 2000). While many ecological differences probably contribute to these differences (including predation, disease and human interference), the fact that stability varies widely among naturally regulated ungulate populations living in environments where human intervention is minimal and predators are absent (Boyd 1981a,b; Boussès et al. 1991; Clutton-Brock et al. 1997a), suggests that variation in population dynamics may often be caused by interactions between populations and their food supplies.

Theoreticians have explored the possibility that contrasts in population dynamics may be consistently related to differences in life histories or in the temporal or spatial distribution of resources (e.g. Peterson et al. 1984; Sinclair 1989; Sæther 1997; Illius & Gordon 2000; Owen-Smith 2002). While it is likely that both these differences may contribute to variation in dynamics, attempts to explain observed variation mostly assume that the causes of contrasts are sufficiently simple to be explained by general models derived from first principles (Caughley 1977).