1. In Saffan‐anaesthetized rats, we have further investigated the mechanisms underlying the vasodilatation induced by adenosine in skeletal muscle by acute systemic hypoxia (breathing 8% O2 for 5 min). 2. In eleven rats the nitric oxide (NO) synthesis inhibitor nitro‐L‐arginine methyl ester (L‐NAME, 10 mg kg‐1, i.v.) reduced the increase in femoral vascular conductance (FVC) induced by hypoxia by approximately 50%. L‐NAME had similar effects on the increase in FVC induced by intra‐arterial (I.A.) infusion of adenosine (at 1.2 mg kg‐1 min‐1 for 5 min via the tail artery) and by ATP (I.A., 1 mg kg‐1 min‐1 for 5 min). Subsequent administration of the adenosine receptor antagonist 8‐sulphophenyl theophylline (8‐SPT, 20 mg kg‐1, i.v.) virtually abolished the adenosine‐ and ATP‐induced increase in FVC. 3. In a further nine rats, 8‐SPT reduced the increase in FVC induced by hypoxia by approximately 50%. This remaining increase in FVC was substantially reduced by L‐NAME. 4. In an additional nine rats, alpha,beta‐methyleneADP (160 micrograms kg‐1, i.v.) which inhibits the 5'‐ectonucleotidase that degrades AMP to adenosine, reduced the peripheral vasodilatation (fall in arterial blood pressure, ABP) induced by ATP infusion, but had no effect on the increase in FVC or decrease in ABP evoked by systemic hypoxia. 5. These results provide the first evidence that the muscle vasodilatation induced by adenosine during systemic hypoxia is mainly dependent on NO synthesis. They also suggest that adenosine is released as such rather than being formed extracellularly from AMP. Given evidence that extraluminal adenosine acts in an NO‐independent fashion we propose that hypoxia releases adenosine from the endothelium. Our results also indicate that hypoxia induces muscle vasodilatation that is adenosine independent but NO dependent: they allow the possibility that this is partly mediated by ATP released from the endothelium.