OBJECTIVE It is hypothesised that nutrients increase pancreatic enzyme secretion by converting cyclical interdigestive secretion to a non-cyclical pattern. This study tested the hypotheses that nutrients do not interrupt cycles and determined the relation of nutrients, calories, and osmotic load to the rate of pancreatic secretion. METHODS Twenty six healthy persons were intubated with oroduodenal and orogastric tubes. Each had one of four different solutions containing 12 to 36% of calories as protein, 24 to 48% as fat, and 40 to 64% as carbohydrate infused into the duodenum at 40, 90, or 160 kcal/h for 300 minutes. Nine g/l sodium chloride (290 mOsm) was added to 16 infusates; osmolality of the other 10 infusates was 24 to 98 mOsm. Pancreatic enzyme outputs were measured every 15 minutes and peaks of enzyme secretion were identified. RESULTS The number of enzyme peaks was similar for the different infusates and the proportion of nutrients in the infusates did not affect secretion of individual enzymes. The nadir, but not the peak of the cycles of enzyme outputs correlated with increasing the caloric load (r = 0.55, p < 0.003 for nadir:peak ratio). Increasing osmolality did not affect cycling but reduced (p < 0.001) enzyme output. CONCLUSIONS Nutrients entering the duodenum do not abolish cycles of enzyme secretion; instead they modulate cycles by increasing the nadir. Forty and 90 kcal infusions submaximally stimulate pancreatic secretion and might be used in patients with pancreatitis without producing pain; adding sodium chloride to solutions should increase this effect.