CORRELATIONS BETWEEN INCIDENCE AND ABUNDANCE ARE EXPECTED BY CHANCE

Relationships between the distribution and the abundance of species are a current topic of interest among biogeographers. When distribution is measured by the frequency with which a species occurs in a set of samples ('incidence'), however, correlations between incidence and abundance are expected by chance. This is because rarer species are less likely to be found in samples. Plotting incidence against abundance-when-present (omitting samples with zero individuals from the calculation of mean abundance) does not eliminate this problem. Therefore the existence of a correlation between incidence and abundance cannot be used as support for any particular ecological hypothesis, since a correlation is the null expectation. The appropriate null hypothesis is never 'no correlation between incidence and abundance.' I present a statistical test that can differentiate observed patterns of incidence from the simplest random pattern: a Poisson model. More sophisticated models suggest that many ecological phenomena can shape patterns of incidence and abundance, including spatial variation in habitat quality, social behaviour, localized reproduction, and stochastic immigration/extinction processes. The relative importance of these different factors cannot be distinguished with the current methods. I suggest that simple correlations between distribution and abundance are of limited utility, and that other, more promising, avenues of analysis of this interesting area of biogeography should be investigated.