Studies of the relationship between body weight and population abundance for animal species based on pooling data from many taxa and assemblages suggest that abundance scales with weight to the -0.75 power. Since metabolic rate scales with weight as (plus)0.75, this result has been taken as evidence that all species in assemblages used equal amounts of energy. The evidence for equivalence' is, however, equivocal, because within many individual assemblages the scaling of abundance on weight differs significantly from 0.75. The authors here examine the relationship between body size (weight and/or length) and abundance in nine previously unpublished animal assemblages, and five previously published assemblages. Twelve of the 14 assemblages show a negative relationship between log body size and log population abundance, but the proportion of variance in abundance which body size explains is always low. Plots of the relationship tend to be polygonal. Regression slopes for most assemblages differ significantly from the predictions of energetic constraint models: most species in these assemblages cannot be energy limited. However, the most abundant species near the upper bound slopes of assemblages may be energy limited, because in only one assemblage does the upper bound slope differ from energetic model predictions. Within natural assemblages, then, the relationship between body size and abundance is usually polygonal, with species body size being a very poor predictor of species abundance. Abundances of most species in assemblages are not, apparently, constrained by energetic requirements, but the commonest species in assemblages may be. -from Authors