Plasma cortisol concentrations and urinary 17-OHCS excretion were measured in two groups of 15 patients receiving elective gastrectomy under either general anaesthesia or epidural blockade. In another 15 patients receiving epidural blockade, the vagus nerve was blocked by injection of local anaesthetics or by truncal vagotomy, and their cortisol response to gastrectomy was compared with the other two groups. Gastrectomy under general anaesthesia caused a marked increase in plasma cortisol concentrations and in urinary excretion of 17-OHCS. These adrenocortical responses to gastrectomy were significantly inhibited in patients operated upon under epidural blockade extending from T3-4 to L1-2 and continuing for 48 h postoperatively. In patients receiving both epidural and vagus nerve blockade, the plasma cortisol response was the same as in those receiving epidural blockade alone. The results indicated that the adrenocortical response to upper abdominal surgery was safely attenuated, though not abolished, with high spinal epidural blockade continuing for 48 h postoperatively. The vagus nerve was not likely to be playing an important role in the adrenocortical response to gastrectomy.