In a new method of testing stool samples from newborn babies for cystic fibrosis (C.F.), a colourless substrate, benzoyl-arginine-p-nitroanilide (B.A.P.N.A.), releases yellow p-nitroaniline when hydrolysed by trypsin. Samples from infants with C.F., who lack trypsin, give negligible colour. 2 infants with C.F. were detected among 2500 consecutive newborn babies tested. The incidence of false-positive results was 1.2% after the first specimen and 0.05% after the second specimen. A further refinement has reduced the positive rate to 0.1% after the first specimen (2000 samples). Tests on samples from 5 other older patients with untreated C.F. have yielded no evidence for false-negative results.