The clinical value of estimation of serum concentrations of immunoreactive trypsin was evaluated by studying 46 healthy controls, 23 controls in hospital, 44 patients with chronic pancreatic disease, and 184 patients with non-pancreatic conditions in which pancreatic disease commonly enters into the differential diagnosis. Serum trypsin concentration had a log normal distribution in the controls, and the calculated normal range was considerably wider than that previously reported. The concentration was abnormal in only 13 out of 27 patients with chronic pancreatitis and was extremely variable in patients with pancreatic cancer. Abnormal results occurred in 11% of the patients with non-pancreatic disease. Eighteen patients had a subnormal trypsin concentration, of whom six did not have pancreatic disease and 12 had either chronic pancreatitis or pancreatic cancer. There was no correlation between serum trypsin concentration and mean tryptic activity as measured by the Lundh test. Of 11 patients with pancreatic steatorrhoea, only seven had subnormal trypsin concentrations. There results suggest that the serum concentrations of immunoreactive trypsin has a low specificity and sensitivity for pancreatic disease and does no reflect the degree exocrine insufficiency in patients with proved chronic pancreatitis.