Sibly, R. M.,Baker, J.,Grady, J. M.,Luna, S. M.,Kodric-Brown, A.,Venditti, C.,Brown, J. H.

The fundamental features of growth may be universal, because growth trajectories of most animals are very similar, but a unified mechanistic theory of growth remains elusive. Still needed is a synthetic explanation for how and why growth rates vary as body size changes, both within individuals over their ontogeny and between populations and species over their evolution. Here, we use Bertalanffy growth equations to characterize growth of ray-finned fishes in terms of two parameters, the growth rate coefficient, K, and final body mass, m(infinity). We derive two alternative empirically testable hypotheses and test them by analyzing data from FishBase. Across 576 species, which vary in size at maturity by almost nine orders of magnitude, K scaled as m(infinity)(-0.23). This supports our first hypothesis that growth rate scales as m(infinity)(-0.25) as predicted by metabolic scaling theory; it implies that species that grow to larger mature sizes grow faster as juveniles. Within fish species, however, K scaled as m(infinity)(-0.35). This supports our second hypothesis, which predicts that growth rate scales as m(infinity)(-0.33) when all juveniles grow at the same rate. The unexpected disparity between across-and within-species scaling challenges existing theoretical interpretations. We suggest that the similar ontogenetic programs of closely related populations constrain growth to m(infinity)(-0.33) scaling, but as species diverge over evolutionary time they evolve the near-optimal m(infinity)(-0.25) scaling predicted by metabolic scaling theory. Our findings have important practical implications because fish supply essential protein in human diets, and sustainable yields from wild harvests and aquaculture depend on growth rates.