The detection and management of small-for-gestational age babies (SGA) was assessed in a case record review of 1302 randomly selected pregnancies. Of the 129 babies with birthweights below the 10th centile for gestational age, 34 (26%) were identified antenatally. For every two correctly identified SGA babies there were three false positive predictions. In-patient monitoring and early elective delivery occurred both in the correctly identified pregnancies (24%) and in the false positives (12%). The management of suspected pregnancies bore no apparent relation to test results and appeared arbitrary. Mortality and morbidity, as measured by nursery admission for greater than 48 h and retention in the nursery after the mother's discharge, were higher in SGA babies than in the hospital population as a whole. The number of ill babies was small, however, reflecting the heterogeneous aetiology of small size for gestational age. Moreover, antenatal detection had little influence on these measures of outcome. It is concluded that tests for detection of SGA babies remain imprecise in practice, gestational weight alone correlates poorly with fetal well-being, and the need remains for sensitive tests to detect babies with genuine morbidity.