The evolution of endothermy in birds and mammals was one of the most important events in the evolution of the vertebrates. Past tests of hypotheses on the evolution of endothermy in mammals have relied largely on analyses of the relationship between basal and maximum metabolic rate, and artificial selection experiments. I argue that components of existing hypotheses, as well as new hypotheses, can be tested using an alternative macrophysiological modeling approach by examining the development of endothermy during the Cenozoic. Recent mammals display a 10°C range in body temperature which is sufficiently large to identify the selective forces that have driven the development of endothermy from a plesiomorphic (ancestral) Cretaceous or Jurassic condition. A model is presented (the Plesiomorphic-Apomorphic Endothermy Model, PAE Model) which proposes that heterothermy, i.e. bouts of normothermy (constant body temperature) interspersed with adaptive heterothermy (e.g. daily torpor and/or hibernation), was the ancestral condition from which apomorphic (derived), rigid homeothermy evolved. All terrestrial mammal lineages are examined for existing data to test the model, as well as for missing data that could be used to test the model. With the exception of Scandentia and Dermoptera, about which little is known, all mammalian orders that include small-sized mammals (<500 g), have species which are heterothermic and display characteristics of endothermy which fall somewhere along a plesiomorphic-apomorphic continuum. Orders which do not have heterothermic representatives (Cetartiodactyla, Perissodactyla, Pholidota, and Lagomorpha) are comprised of medium- to large-sized mammals that have either lost the capacity for heterothermy, or in which heterothermy has yet to be measured. Mammalian heterothermy seems to be plesiomorphic and probably evolved once in the mammalian lineage. Several categories of endothermy are identified (protoendothermy, plesioendothermy, apoendothermy, basoendothermy, mesoendothermy, supraendothermy, and reversed mesoendothermy) to describe the evolution of endothermy during the Cenozoic. The PAE Model should facilitate the testing of hypotheses using a range of macrophysiological methods (e.g. the comparative method and the reconstruction of ancestral states).