Four mid-lactation Holstein dairy cows (mean milk yield on day of experiments 26·1 kg/d) were used in a series of experiments to establish the contribution of non-insulin-mediated glucose uptake to total glucose uptake at basal insulin concentrations. A secondary objective was to determine whether somatostatin affects the action of infused insulin. In part I of the experiment a primed continuous infusion of [6,6-2H]glucose (45·2 μg/kg per min) was begun at time 0 and continued for 5 h. After 3 h of [6,6-2H]glucose infusion (basal period) a primed continuous infusion of insulin (0·001 i.u./kg per min) was administered for 2 h. Coincident with the insulin infusion, normal glucose was also infused in order to maintain the plasma glucose concentration at euglycaemia. Part II of the experiment was the same as part I except that somatostatin was infused for 2 h (0·333 μg/kg per min) instead of insulin. In part III of the experiment both insulin and somatostatin were infused for the final 2 h. Plasma insulin levels were increased by insulin infusion (to 0·1476 and 0·1290 i.u./l for parts I and III respectively) and were reduced by somatostatin infusion in part II (to 0·006 i.u./l) relative to the basal periods (mean 0·021 i.u./l). Glucose uptake during somatostatin infusion (2·50 mg/kg per min; part II) was 92·0% of that observed in the respective basal period (2·72 mg/kg per min). Circulating insulin levels were much lower than the dose of insulin that causes a half maximal effect on glucose uptake (0·06–0·10 i.u./l for ruminants); consequently insulin-mediated glucose uptake was probably absent in part II. Secondly, glucose uptake following insulin only infusion (4·05 mg/kg per min) was significantly lower than that observed when insulin plus somatostatin was infused (4·69 mg/kg per min), indicating that somatostatin either directly or indirectly enhanced the action of insulin on glucose uptake.