An explanation for unexpected population crashes in a constant environment

Unexpected population crashes are an important feature of natural systems, yet many observed crashes have not been explained. Two difficulties in explaining population crashes are their relative rarity and the multi-causal nature of ecological systems. We approach this issue with experimental microcosms, with large numbers of replicates of red flour beetle populations (Tribolium castaneum). We determined that population crashes are caused by an interaction between stochasticity and successive episodes of density dependence: demographic stochasticity in oviposition rates occasionally produces a high density of eggs; so high that there are insufficient flour resources for subsequent larvae. This mechanism can explain unexpected population crashes in more general settings: stochasticity 'pushes' population into a regime where density dependence is severely overcompensatory. The interaction between nonlinearity and stochasticity also produces chaotic population dynamics and a double-humped one-generation population map, suggesting further possibilities for unexpected behaviour in a range of systems. We discuss the generality of our proposed mechanism, which could potentially account for previously inexplicable population crashes.