Density Dependence: Are We Searching at the Wrong Spatial Scale?

Regulation of populations by density-dependent mechanisms is one of the basic tenets of theory in population biology. Yet, it has been argued that population studies rarely reveal density dependence. We show that the detection of density-dependent processes is a function of study spatial scale, so that appropriately scaled studies commonly reveal density dependence, while inappropriately scaled studies do not. The detection of density dependence and the potential for spatial structure in population density were examined in 79 insect population studies. The potential for spatial structure within the study area was estimated from population movement relative to the area over which density was measured. Relative movement was classified as 'high' (and potential for spatial population structure 'low') if more than 10% of the population moved across the boundary of this area each generation. Analyses were stage-based, addressing the prevalence of density-dependent processes, rather than population regulation. In mobile stages (small and large instars and adults), density dependence was found in 69-73% of studies with low potential spatial structure ('high' relative population movement), and in 23-35% of studies with high potential structure ('low' movement). In immobile stages (eggs and pupae), density dependence was rarely detected in studies larger than 0.1 hectare (7-16%), but often detected in smaller-scale studies (65-69%). One-third of studies reporting both density relationships and 'key factors' (sources of mortality that are most correlated with population fluctuation) found the key factor to be directly density-dependent. Density-dependent growth was detected in at least one stage in 74-76% of studies that tested for density relationships. Study length (in generations) was not related to the frequency of detection of density dependence in studies with low potential spatial structure in population density. This result calls into question the need for long-term studies to detect density relationships. The detection of population regulation may be hampered by scale effects similar to those demonstrated here.