Rats were equipped with chronic gastric cannulas, and the effects of intraperitoneal injections of pancreatic glucagon on sham feeding, with cannulas open, and real feeding, with cannulas closed, were compared. Glucagon (100-2,500 micrograms/kg) suppressed cumulative food intake during real feeding tests 9-33%, but had no effect during sham feeding. Despite their increased food intakes, sham feeding rats took discrete meals terminated by the behavioral satiety sequence. In addition to not affecting total intake, glucagon failed to affect meal size, latency to rest, or intermeal interval during sham feeding. To investigate the possibility that glucagon did not inhibit sham feeding because it did not elicit hyperglycemia, we measured hepatic vein blood glucose after glucagon injections in sham feeding rats: glucagon elicited marked hyperglycemia during sham feeding. Therefore, the absence of glucagon's satiety effect in sham feeding is not due to the lack of hepatic glycogenolysis and hyperglycemia. These results suggest that some mechanism other than hyperglycemia which normally accompanies food ingestion is necessary for glucagon to have a satiety effect.