The primary objectives were: (1) to examine if vasodilatory amounts of metabolites are released during moderate as well as during severe hypoxia of donor hearts; (2) to compare vasodilation to hypoxia with infusion of adenosine, and (3) to determine if metabolically induced vasodilation can be selectively reduced or abolished in order to determine the responsible metabolic factors. Pairs of isolated guinea pig hearts (n = 50) were perfused with a modified physiological saline solution by the Langendorff method. Assay hearts were also perfused with either uncorrected, anaerobic donor heart effluent or with effluent in which Po<sub>2</sub>, Pco<sub>2</sub> and pH were corrected to control perfusate levels. When donor hearts were made moderately hypoxic (effluent Po<sub>2</sub> – 45 ± 5 Torr; myocardial oxygen consumption – MVO<sub>2</sub>-unchanged) and assay hearts were perfused at constant pressure with gas-corrected donor effluent, we saw no change in assay flow. But when adenosine was infused to produce the same flow increase observed during mild hypoxia in donor hearts, the assay heart conductance rose to 38 ± 7%. In contrast, when donor hearts were made severely hypoxic (effluent Po<sub>2</sub> = 34 ± Torr; MVO<sub>2</sub> reduced by 53 ± 6% and assay hearts were perfused at constant flow with corrected donor effluent), we found a 60 ± 5 % increase in vascular conductance; this dropped to 24 ± 6% over control when adenosine deaminase was infused into the donor effluent tubing and to 5 ± 4% when theophylline was infused into assay hearts. In this bioassay study we also found that 68 ± 5% of adenosine infused into donor hearts during normoxia was removed in transit through the vasculature. Overall, our results indicate that adenosine, primarily, but perhaps also adenine nucleotides, are released in vasodilatory concentrations with severe hypoxia. However, the metabolic factors which increase coronary flow during mild hypoxia when MVO<sub>2</sub> is not compromised could not be ascertained.