Insulin stimulated gastric secretion was studied in 74 unoperated duodenal ulcer patients (DUs), (20 women and 54 men). Three indices of secretion were studied--observed volume, acid output, and volume of gastric juice corrected for pyloric loss and duodenal reflux (VG). These three measurements were expressed both as peak secretion and as secretion during the 1/2 to two hour period after insulin, and also both before and after standardisation for height, making 12 different indices in all. From the data a significant correlation between insulin-stimulated secretion and height in DUs was found. A method of standardising each patient's secretion for height is described. We confirm a significantly higher insulin-stimulated secretion in men than in women and show that this difference can be explained by their difference in height. For each of the 12 indices of secretion, the range of secretion for the unoperated subjects was obtained. The same indices were measured in 155 postvagotomy patients, including 33 patients with recurrent DUs, and compared with the ranges of secretion established in the unoperated patients. Responses above the lower 95% tolerance limit of the preoperative range were designated positive and those below negative. The Hollander status was determined. It was found that the least satisfactory criterion was Hollander's (7% false negative and 69% false positive). The best was 1/2-2 VG standardised for height (3% false negatives and 43% false positives). The improvement in predictably was significant at the 0-0005 level.