Geographical variation in environmental temperatures is expected to impose clinal phenotypic selection that results in the expression of large-scale gradients of body mass variation within animal clades. Body size is predicted to increase with increasing latitude and elevation, and hence, with decreasing temperature, a pattern broadly known as Bergmann's rule. However, empirical observations are highly conflicting. Whilst most studies support this prediction in endotherms (birds and mammals), analyses conducted on ectotherms often fail to report this pattern. Does it reduce the validity of this macroecological rule? Since the original formulation of Bergmann's rule only involved endothermic organisms, I argue that the controversy is not a consequence of its predictive power, but a result of the later inclusion of ectotherms as part of the prediction. Here, I propose that the common conception of Bergmann's rule maintained for half a century is changed back to its original definition restricted to endotherms. This temperature-size relationship might therefore consolidate as a well-established macroecological rule if its original formulation is respected. Finally, I develop these claims on my initial suggestion that Bergmann's rule should be recognized as the evolutionary outcome of a general process with no phylogenetic scale distinction of species or populations, being equally applicable amongst and within species.