Environment has strong effects on avian BMR, above and beyond migratory tendency.

BMR was first modeled as a function of body mass (M), passerine/non-passerine differences (Pass/non-P) and migratory tendency (Migratory). Controlling for these three variables, subsequently single environmental predictors were tested for their effect on BMR. The table shows the fitted parameter (b), the estimate of the AIC and the P-value testing whether the parameter is significantly different from zero. Parameters were estimated singly, and we conducted two analyses for each parameter: first one in which phylogeny was ignored (�� = 0), and then one in which we estimate Pagel's ��, and set it equal to its maximum likelihood value (ML ��). We tested whether the maximum likelihood estimate of �� was different from zero, i.e. whether the data show significant phylogenetic signal.