In a retrospective analysis of 24 cases of congenital adrenal hyperplasia (CAH) in neonates born in the province of Manitoba during the last 20 years, we set out to determine whether patients, in particular male infants with salt-losing CAH, were being missed by the usual forms of clinical ascertainment. Although the overall incidence of 1/14,500 live births was similar to that found in several screening surveys, a skewed female/male sex ratio of 2.2:1 suggested probable death among male infants with unrecognized adrenal insufficiency. These results led to a prospective analysis of 17 alpha-hydroxyprogesterone (17-OHP) levels in 1194 neonatal blood specimens by a solid-phase direct radioimmunoassay procedure to determine whether this method would be suitable for CAH screening. In 1103 neonates weighing greater than 2500 gm at birth, all 17-OHP values were less than 30 nmol/L (approximately 1000 ng/dl), with a mean of 8.2 nmol/L; values in male infants were slightly higher than in female infants. In 89 neonates with a birth weight less than 2500 gm, 17-OHP values were skewed, with nine having levels greater than 30 nmol/L and two greater than 50 nmol/L. Postnatal age (1 to 24 days) at the time of specimen collection had no effect on 17-OHP levels, although higher values occur during the first 24 hours. One unsuspected case of CAH in a male infant was discovered during the trial period. We conclude that neonatal CAH screening can permit diagnosis and therapy of affected male infants who are being missed by normal clinical evaluation. This radioimmunoassay method is relatively simple and inexpensive, and it has the specificity and sensitivity necessary to provide such mass screening.