Testing metabolic theory with models of tree growth that include light competition

1. Metabolic scaling theory predicts that diameter growth rates of tree species are related to tree diameter by a universal scaling law. This model has been criticised because it ignores the influence of competition for resources such as light on the scaling of demographic rates with size. 2. We here test whether scaling exponents of abundant tropical tree species comply with the prediction of metabolic scaling theory and evaluate whether the scaling of growth with size depends on light availability. Light reaching each individual tree was estimated from yearly vertical censuses of canopy density, and a hierarchical Bayesian approach allowed quantifying confidence intervals for scaling exponents and accounting for different sources of error. 3. We found no universal scaling relationship, and 5070% of the species had scaling exponents that significantly differed from the predicted value of 1/3. As would be expected if competition for light were important, scaling exponents were >1/3 for the majority of species when all trees were combined. However, the community average of scaling exponents was not significantly different from the predicted value of 1/3 when only considering individuals that grew under high-light conditions. 4. These results support the hypothesis that the prediction of metabolic ecology for the scaling of tree growth with size is only valid when competition for light is unimportant.