1. Recently, fresh attempts have been made to understand the mechanisms structuring species–body size distributions. Of these, the model developed by Brown, Marquet & Taper (1993) (BMT), which uses measures of resource acquisition and conversion to determine an optimal body size (M*) for a given assemblage, is potentially the most significant. Here, we examine the novelty of the model and some of its assumptions, and test its empirical predictions.
2. The BMT model is one of a suite of physiological/life‐history models examining size at maturity. Such energetic or ‘physical’ approaches to body size evolution have a considerable modern history and continue to enjoy attention in physiological ecology and life‐history theory.
3. Although mortality significantly influences life‐history evolution and mostly precludes the evolution of a body size that maximizes reproductive power output, it is excluded from the BMT model. Likewise, the model assumes that power and not efficiency is maximized, although there are conditions where this is not likely to be the case. Furthermore, the BMT model assumes that resource acquisition and conversion are physiologically limited, although the importance of physiological limitation in ecology remains unclear.
4. Additional assumptions of the model include coincidence of the optimum body size of an individual and a species, when in many species this is not the case, and coincidence of the most speciose size class in an assemblage and the optimal body size. Similarly, the probable influence of differences in the scaling of various parameters at the intra‐ and interspecific levels are not addressed, nor are the impacts of discrepancies between phenotypic and genotypic optima.
5. The measures of resource acquisition and conversion used in the BMT model are not only problematic, but also limit the utility of the model. If the scaling constants and exponents of field metabolic rate minus basal metabolic are used as a measure of resource acquisition beyond maintenance needs, and those for intrinsic rate of increase (rmax) as a measure of resource conversion, the applicability of the model to other taxa can be extended.
6. Using these measures we show that the model continues to provide a reasonable prediction of M* for the terrestrial mammal assemblages of both North America and Australia. However, when taxonomic inclusiveness is reduced by removing eutherians from the Australian data set and by examining the predictions of the model with regard to diprotodonts only, the model fails to provide reasonable predictions of M*.
7. Given problems with the BMT model, but its clear ability to predict the modal body size of at least two terrestrial mammal assemblages, we suggest that there is considerable scope for exploring the relationships between resource acquisition and conversion at the level of the individual and energy partitioning between individuals in multispecies communities.
