Neonatal skeletal maturation was assessed by different methods based on the bone centres in the knee and ankle region in 46 infants with true-positive (patients) and 17 infants with false-positive screening tests (controls). The patients and controls did not differ in mean age at X-ray examination and age at the start of treatment (14.5 ± 5.7 days). One of the methods used to score the size of femoral and tibial epiphyses was just as good as the other ones tested, but has the advantage of being the easiest to use and therefore clinically most suitable. Skeletal maturity assessed by this method correlated best with serum T<sub>4</sub> (r = 0.62, p < 0.01). Griffiths tests were performed in 37 of the 46 patients at 28-48 months of life. The best correlation obtained between neonatal skeletal maturity and Griffiths global developmental quotient at 3 years of age was r = 0.58 (p < 0.001). Although statistically significant, it was too weak to be of clinical value in identifying individual patients at risk. We conclude that an assessment of skeletal maturation is not useful for the prognosis of psychomotor development in individual patients with congenital hypothyroidism receiving treatment within the first 2 weeks of life.