- Summary 1 Freshwater temporary pools constitute insular habitats that can vary extensively in the richness of their macro-invertebrate fauna, even within small spatial scales. While strictly statistical analyses may identify variables that account partially for this variation, they do not address explicitly the mechanisms that generate it. 2 Here we show how a simple island-biogeographical model explains 62% of the variation in local species richness of 47 neighbouring pools, based solely on pool size and permanence. A regression of the observed richness on that predicted by the model did not indicate any systematic biases, and left residuals that were distributed according to the expected. 3 The model provides a robust working hypothesis, to which we apply a novel approach of quantifying the relative impact of local and regional factors on the species richness of individual pools. Using elasticity analysis of model parameters and invoking elasticity as a measure of parameter ‘importance’ we show, for example, that the impact of regional species richness on the richness of noninteractive communities should always exceed that of the probability of local extinction. 4 The analysis exposes potential difficulties in generalizing which factor – local or regional – has the greater relative importance, and to what extent, as the answer can depend on: (i) the specific factors being compared, and (ii) particular habitat characteristics of individual localities.