5‘-isobutylthioadenosine, but it was inhibited by 10 mMp-nitrophenylphsphate to 50 k 11% ( n = 6). The inhibition of the pigeon ventricle AMPase by other nucleoside monophosphates (with specificity for the base moeity), but not by ribose phosphate or B-glycerophosphate, indicates the presence of a 5’-nucleotidase. its preference for AMP, resistance to inhibition by P, and 5’-deoxy5‘-isobutylthioadenosine and its inhibition by p-nitrophenylphosphate suggest that it is distinct from the previously purified cytosolic 5‘-nucleotidase. Its high activity towards AMP suggests that it is responsible for ischaemiainduced adenosine production. The similar activity discovered in rat heart is too small to account for all of the absolute rate of ischaemia-induced adenosine formation (470 k 70 nmol/min per g) observed in this tissue (Meghji et ul., 1988). It may, however, contribute to this rate. In conclusion, there is extensive evidence for adenosine production in response to net ATP breakdown by a cytoplasmic pathway not involving ecto-.5‘-nucleotidase. A soluble AMP S’-nucleotidase abundant in pigeon ventricle, but also found in rat ventricle may be responsible for the formation of that adenosine which mediates metabolic vasodilatation.