Even among vertebrate species of the same body mass and higherlevel taxonomic group, metabolic rates exhibit substantial differences, for which diverse explanatory factors-such as dietary energy content, latitude, altitude, temperature, and rainfall-have been postulated. A unifying underlying factor could be food availability, in turn controlled by net primary productivity (NPP) of the animal's natural environment. We tested this possibility by studying five North American species of Peromyscus mice, all of them similar in diet (generalist omnivores) and in gut morphology but differing by factors of up to 13 in NPP of their habitat of origin. We maintained breeding colonies of all five species in the laboratory under identical conditions and consuming identical diets. Basal metabolic rate (BMR) and daily ad libitum food intake both increased with NPP, which explained 88% and 90% of their variances, respectively. High-metabolism mouse species from high-NPP environments were behaviorally more active than were lowmetabolism species from low-NPP environments. Intestinal glucose uptake capacity also increased with NPP (and with BMR and food intake), because species of high-NPP environments had larger small intestines and higher uptake rates. For metabolic rates of our five species, the driving environmental variable is environmental productivity itself (and hence food availability), rather than temporal variability of productivity. Thus, species that have evolved in the presence of abundant food run their metabolism ''fast,'' both while active and while idling, as compared with species of less productive environments, even when all species are given access to unlimited food.M etabolic rate means the rate at which an animal burns calories to produce energy. Among vertebrate species, there is a 10 7 -fold range of metabolic rates. Much research has aimed at understanding the reasons behind this variation. Why are some animals seemingly extravagant, consuming and expending calories rapidly, whereas others are frugal, consuming very little and metering their expenditure accordingly so as to remain in energy balance?The two factors underlying most of this variation are well known: body mass and higher-level taxonomic affiliation. First, within the same higher-level taxonomic group (e.g., mammals), metabolic rate increases with body mass (the so-called mouseto-elephant curve) according to approximately the 0.75 power of body mass (1). (Metabolic rate per g of body mass decreases with body mass but by a power smaller than 1.0, so the absolute metabolic rate of a whole animal, the subject of this paper, increases with mass). Second, for species of the same mass there are some differences between higher-level taxonomic groups, notably the ca. 8-fold difference between endotherms (''warmblooded'' birds and mammals) and ectotherms (''cold-blooded'' reptiles, amphibia, and fish) (2), and also the ca. 1.3-fold difference between marsupials and placentals (3) and the ca. 2-fold difference between sloths and other placentals (4).Ho...