With an inhibitor assay technique rates of passage of salivary and pancreatic isoamylase through the jejunum were measured in six healthy volunteers after different liquid, intragastric meals. In all subjects and in 13/17 experiments, more than 2500 units of salivary amylase were passed over 200 postcibal minutes. Salivary amylase comprised 13.8 +/- 3.9% (mean +/- SEM) of the total amylase and appeared predominantly as single, distinct peak. The inhibitor method was validated by isoelectric focusing (r = 0.988; P less than 0.001; N = 7). The frequency of detection of salivary amylase in gastric or jejunal samples fell as gastric pH fell below 3.0. In vitro, amylase was inactivated in gastric juice as pH fell between 3.8 and 3.3. Salivary amylase accounted for 11% of total amylase output in a normal and 27% in an achlorhydric subject after a hamburger meal. We conclude that amylase should not be measured in postprandial studies of pancreatic secretion in humans without correction for salivary amylase.