Metabolic effects of increased mechanical work were studied by comparing isolated pumping rat hearts perfused by the atrial-filling technique with aortic-perfused non-pumping hearts perfused by the technique of Langendorff. The initial medium usually contained glucose (11 mm) and palmitate (0.6 mm bound to 0.1 mm albumin). During increased heart work (comparing pumping with non-pumping hearts) the uptake of oxygen and glucose increased threefold, but that of free fatty acids was unchanged. Tissue contents of alpha-oxoglutarate, NH4+, malate, lactate, pyruvate and Pi rose with increased heart work, but contents of ATP, phosphocreatine and citrate fell. Ketone bodies were produced with a ratio of beta-hydroxybutyrate/acetoacetate of about 3:1 in both pumping and non-pumping hearts but with higher net production rates in non-pumping hearts. When ketone bodies were added in relatively high concentrations (total 4 mm) to a glucose (11 mm) medium the medium, ratios of beta-hydroxybutyrate/acetoacetate were not steady even after 60 min of perfusion. The validity of calculating mitochondrial free NAD+/NADH ratios from the tissue contents of the reactants of the glutamate dehydrogenase system or the beta-hydroxybutyrate dehydrogenase system is assessed. The activities of these enzymes are considerably less in the rat heart than in the rat liver, introducing reservations into the application to the heart of the principles used by Williamson et al. (1967) for calculation of mitochondrial free NAD+/NADH ratios of liver mitochondria...