In this study we investigated the site and nature of the signal responsible for the generation of the gastrocolonic response at the rectosigmoid region. Sixteen healthy subjects participated in this study. Motor activities were recorded with pressure transducers placed in the rectosigmoid colon. Balloon distention of the stomach with 100, 200, or 300 ml of water caused a volume-dependent increase in rectosigmoid motility that was abolished by atropine. To investigate the intestinal phase of the gastrocolonic response, we infused into the duodenum isocaloric (178 kcal) solutions of normal saline, lipid, glucose, or essential amino acids at 1.5 ml/min for 30 min in random order on separate days. Only lipid infusion caused an increase in rectosigmoid pressure and this was accompanied by an elevation of plasma cholecystokinin from 1.2 +/- 0.1 to 4.3 +/- 1.0 fmol/ml. The increase in motility associated with lipid infusion was antagonized (58% +/- 7%) by atropine. To further investigate the possible role of cholecystokinin as a mediator of the gastrocolonic response, we infused cholecystokinin-octapeptide intravenously at doses of 5, 10, and 20 ng/kg.h. A significant increase in rectosigmoid motility was observed only at 20 ng/kg.h and this was accompanied by an increase in plasma cholecystokinin levels to 12 +/- 2 fmol/ml, a value threefold greater than that produced by lipid infusion. These studies demonstrate that gastric distention and intestinal lipid are potent stimuli for the generation of the gastrocolonic response involving the rectosigmoid region. The gastric phase can be generated by mechanoreceptors utilizing the cholinergic pathways, whereas the intestinal phase is nutrient-specific, partially atropine-sensitive, and independent of cholecystokinin.