A study programme was set up in Wales and the West Midlands to evaluate serum immunoreactive trypsin screening for cystic fibrosis in neonates using blood spots collected for metabolic screening. By screening half the blood spots from each area, it was hoped to generate two comparable groups of fibrocystic children; those detected by screening and those not screened who would be diagnosed clinically. Over almost three years, more than 120,000 specimens were screened and 37 infants detected with cystic fibrosis. Four additional fibrocystic patients were missed on screening: two had negative immunoreactive trypsin values, of which one had meconium ileus, and two, although giving initial positive tests, were negative on follow up. Excluding infants known to be at risk, comparison of the numbers of children detected in the screened and unscreened groups showed more than a two-fold difference in favour of the screened group. There may be a large number of undiagnosed fibrocystic patients in the general population.