In vertebrates, maximal rates of oxygen consumption (V(O(2),max)) exceed resting rates (V(O(2),rest)) by an average factor of ten. This pattern of factorial scope has led to the hypothesis that V(O(2),rest) and V(O(2),max) are causally linked in vertebrates (aerobic capacity model, Bennett and Ruben, Science 206, 649-654, 1979). We propose an alternate theory that vertebrate resting metabolic rates are regulated at levels to optimize metabolic performance during activity, by reducing cardiovascular response times for O(2) transport. First, we argue that circulatory convection has the potential to be rate-limiting to vertebrate aerobic adjustment. We then show mathematically that incremental changes in convection requirements exhibit a nonlinear dependence on initial values. From this, a cost-benefit model is constructed, using energetics and blood-convection requirements, to predict the optimal fractional allocation to V(O(2),rest) in vertebrates as 11% of V(O(2),max). The implications of our results to vertebrate metabolic design and the evolution of endothermy are discussed.