To estimate the region from which phase III of the migrating myoelectric complex (MMC) originates and to establish the role of nicotinic and muscarinic receptors in the initiation of this phase, chronic experiments were performed in eight sheep with bipolar platinum electrodes attached to the antral and small intestinal wall. Myoelectric activity was recorded by means of a multichannel electroencephalograph before and after slow intravenous injections of cholinergic agonists and antagonists as well as erythromycin, an agonist to the hormone motilin. MMC phases were identified according to the criteria proposed by Code and Marlett. A total of 263 MMC cycles was observed with a duration of 103±24 and 91±26 min in fasted and non-fasted animals, respectively. In fasted sheep, 8.7% of phases III MMC originated from the pyloric antrum, 50.5% from the duodenal bulb, 29.7% from the duodenum, 8.7% from the jejunum and 2.3% from the ileum. In non-fasted sheep these values were: 10.3, 31.3, 39.4, 15.5 and 3.5%, respectively. Neither cholinergic agonists nor erythromycin markedly changed the site of phase III MMC start. Following administration of anticholinergic drugs phase III MMC was not observed in the antrum. In fasted sheep following hexamethonium administration, 33% of phases III MMC originated from the duodenal bulb, 33% from the duodenum and 33% from the jejunum; in non-fasted sheep these values were equaled 50% each in the duodenal bulb and jejunum. When atropine was injected into fasted sheep, 20% of phases III MMC originated from the duodenal bulb and 80% from the duodenum while in non-fasted animals 37.5, 37.5, and 25% of phases III started from the duodenal bulb, duodenum and jejunum, respectively. Following pirenzepine injection, 20% of recorded phases III MMC were induced from the duodenal bulb and 80% from the duodenum; in non-fasted animals 40% started from the duodenal bulb, 40% from the duodenum and 20% from the jejunum. The time lag from anticholinergic drug administration until the start of phase III MMC was dose-dependent and was the longest following hexamethonium administration. Thus, the ectopic pacemaker appears to be the most important in generation of phase III MMC and cholinergic control of phase III origin from the gastrointestinal region is important in sheep. The role of nicotinic receptors seem to be greater than that of muscarinic receptors in this area.