Metabolic theory predicts animal self‐thinning

Summary The metabolic theory of ecology (MTE) predicts observed patterns in ecology based on metabolic rates of individuals. The theory is influential but also criticized for a lack of firm empirical evidence confirming MTE's quantitative predictions of processes, e.g. outcome of competition, at population or community level. Self‐thinning is a well‐known population level phenomenon among plants, but a much less studied phenomenon in animal populations and no consensus exists on what a universal thinning slope for animal populations might be, or if it exists. The goal of this study was to use animal self‐thinning as a tool to test population‐level predictions from MTE, by analysing (i) if self‐thinning can be induced in populations of house crickets (Acheta domesticus) and (ii) if the resulting thinning trajectories can be predicted from metabolic theory, using estimates of the species‐specific metabolic rate of A. domesticus. I performed a laboratory study where the growth of A. domesticus was followed, from hatching until emergence as adults, in 71 cohorts of five different starting densities. Ninety‐six per cent of all cohorts in the three highest starting densities showed evidence of self‐thinning, with estimated thinning slopes in general being remarkably close to that expected under metabolic constraints: A cross‐sectional analysis of all data showing evidence of self‐thinning produced an ordinary least square (OLS) slope of −1·11, exactly that predicted from specific metabolic allometry of A. domesticus. This result is furthermore supported by longitudinal analyses, allowing for independent responses within cohorts, producing a mean OLS slope across cohorts of −1·13 and a fixed effect linear mixed effects models slope of −1·09. Sensitivity analysis showed that these results are robust to how the criterion for on‐going self‐thinning was defined. Finally, also as predicted by metabolic theory, temperature had a negative effect on the thinning intercept, producing an estimate of the activation energy identical to that suggested by MTE. This study demonstrates a direct link between the metabolic rate of individuals and a population‐level ecological process and as such provides strong support for research that aims to integrate body mass, via its effect on metabolism, consumption and competition, into models of populations and communities.