ABSTRACT
AbstractMany attempts at providing a single-scale exponent and mechanism to explain metabolic rate assert a monolithic selective mechanism for allometries, characterized by a universal allometric scale power (usually chosen to be 0.75). To test for the deviations from universal allometric scaling, we gathered data from previously published metabolic measurements on 903 bird species and performed regressions of log(basal metabolic rate) and log(body mass) for (1) all birds and (2) 20 monophyletic clades within birds. We constructed two Bayesian linear mixed models-one included ecological variables and the other included data for mammals from Sieg et al. (2009). Overall allometric patterns differed significantly among clades of birds, and some clades were not consistent with the 0.75 scale power. We were unable to find apparent physiological, morphological, phylogenetic, or ecological characteristics among clades, predicting a difference in allometry or consistency with any previously proposed universal allometry. The Bayesian analysis illuminated novel bivariate, clade-specific differences in scaling slope-intercept space, separating large groups of birds and mammals. While significantly related to basal metabolic rate, feeding guild and migratory tendency had small effects compared to clade and body mass. We propose that allometric hypotheses, in general, must extend beyond simple overarching mechanisms to allow for conflicting and interacting influences that produce allometric patterns at narrower taxonomic scales-perhaps including other processes whose optimization may interfere with that of the system proposed by the metabolic theory of ecology.
