Mechanisms underlying density-area relationships (correlations between population density and patch size) have rarely been tested experimentally. It is often assumed that higher density on large patches is evidence that large patches are high quality (i.e. have greater survival and reproduction), but the same pattern could result from disproportionate movement from small to large patches. Movement-based and within-patch processes must be experimentally separated to show that large patches are indeed of higher quality, but few studies have done so. We experimentally tested movement-based and within-patch hypotheses to explain the positive density-area relationship observed for a saproxylic (decayed wood-dependent) beetle, Odontotaenius disjunctus Illiger (Coleoptera: Passalidae). In separate experiments we quantified (1) immigration into and (2) finite growth rate within logs (=patches) that varied in size and density of conspecific beetles. A log was 11.7-fold [95 % confidence interval (CI) 3.4-40.8) and 10.5-fold (95 % CI 2.7-40.9) more likely to contain a new immigrant if it was large or contained a conspecific pair of beetles, respectively. Neither log size nor conspecific density was associated with changes in finite growth rate that would lead to higher density: decreased log size and increased conspecific density reduced finite growth rate in direct proportion to the loss in available resources per mating pair. We conclude that movement behavior rather than habitat quality is responsible for the positive density-area relationship for O. disjunctus. An important implication of our results is that population density is an imperfect indicator of habitat quality.