Altogether 540 succesful insulin tests were performed on 377 patients after vagotomy, on 316 patients less than three months after the operation (early phase) and on 182 patients between three months and nine years after (late phase). The proportion of positive insulin test responses after vagotomy and antral resection was similar in the selective gastric (SV) and truncal vagotomy (TV) groups (10.7% and 11.8% respectively in the early phase) and remained the same during follow-up (12.5% and 9.3% in the late phase). In the vagotomy and drainage group there were significantly fewer positive test responses in the early phase after SV (15.6%) than after TV (36.4%), but in the late phase the incidence was similar (56.7% and 55.6%). There was a significant rise in the number of positive responses after SV and drainage during follow-up (from 15.6% to 56.7%). Insulin-stimulated peak acid output (I-PAO) was about the same in the vagotomy and drainage group after SV and TV (3.52 and 5.03 mmol/h in the early phase and 4.20 and 4.93 mmol/h in the late phase). Although the number of Hollander positive insulin tests after vagotomy and drainage increased during follow-up, no corresponding changes were seen in I-PAO. There was no evidence of vagal reinnervation.