We evaluated the hypothesis that the noncaffeine gastric acid stimulant effect of coffee might be by way of serum gastrin release. After 10 healthy volunteers drank 50 ml of coffee solution corresponding to one cup of home-made regular coffee containing 10 g of sugar and 240 mg/100 ml of caffeine, serum total gastrin levels peaked at 10 min and returned to basal values within 30 min; the response was of little significance (1.24 times the median basal value). Drinking 100 ml of sugared water (as control) resulted in occasional random elevations of serum gastrin which were not statistically significant. Drinking 100 ml of regular or decaffeinated coffee resulted in a prompt and lasting elevation of total gastrin; mean integrated outputs after regular or decaffeinated coffee were, respectively, 2.3 and 1.7 times the values in the control test. Regular and decaffeinated coffees share a strong gastrin-releasing property. Neither distension, osmolarity, calcium, nor amino acid content of the coffee solution can account for this property, which should be ascribed to some other unidentified ingredient. This property is at least partially lost during the process of caffeine removal.