We have assessed the feasibility of screening newborn babies for congenital adrenal hyperplasia (CAH) by the direct measurement of 17-hydroxyprogesterone (17-OHP) in blood spots collected on filter paper (Guthrie cards) for the phenylketonuria, hypothyroidism and galactosaemia screening programmes run in Scotland. The procedure described for CAH uses an iodinated 17-OHP tracer and a specific 17-OHP antiserum sheathed within semipermeable nylon microcapsules. The method does not require a solvent extraction step, is inexpensive, precise, efficient and, therefore, practical for large-scale use. With this system the value of a neonatal screening programme was assessed in a retrospective analysis and a prospective trial. The retrospective study of 15 paediatric cases of CAH illustrated that at least half were not diagnosed within 3 weeks of birth. Analysis of the original Guthrie card samples revealed increased levels of 17-OHP in all cases. The prevalence of CAH as calculated in the retrospective study was 1 in 20 907 with a range (within 95% confidence limits) of from 1 in 12 675 to 1 in 32 604 (n = 301 450). In the prospective trial a total of 92 051 consecutive samples was screened. Five cases of CAH were correctly identified with a current false positive rate of 0.042%. Analysis of urinary steroids confirmed defective adrenal 21-hydroxylase activity in all positive cases. In the prospective trial the prevalence was 1 in 18 401 with a range of from 1 in 7422 to 1 in 50 006. We conclude that mass screening for CAH is both feasible and desirable.