To assess the fate and efficacy of orally ingested enzymes in pancreatic insufficiency, we administered pancreatin to six patients by two schedules--eight tablets with a standard meal or two tablets hourly--and in six normal controls, quantified duodenal enzyme activity and related inactiviation of ingested enzymes to gastric and duodenal pH; in the six patients we measured malabsorption by fecal balance studies. Postprandially, gastric pH was similar in health and pancreatic insufficiency, and below 4 after 40 minutes. Duodenal pH in pancreatic insufficiency declined to approximately 4 beyond 100 minutes--lower than in health (P less than 0.05). Approximately 22 per cent and 8 per cent of trypsin and lipase activity ingested with either schedule was delivered to the ligament of Treitz. Prandial was as effective as hourly administration in decreasing steatorrhea and perhaps more effective in abolishing azotorrhea, and since it is also more convenient, we recommend it.