An obstetric population was screened 'blind' by measuring unconjugated estriol (ucE3) and total estriol (tE3) in serum. The purpose was firstly to ascertain the predictive values of ucE3 for the prediction of low birth weight and distressed neonates (perinatal deaths included), and secondly to compare the predictive values of ucE3 and tE3. The main material consisted of 1018 singleton pregnancies with known term, who would not have undergone estriol (E3) measurement had it not been for this investigation. The study was carried out prospectively, and the obstetricians had no knowledge of the E3 values in the screened pregnancies. When at least one of the ucE3 values was low, the risk of having a small for gestational age (SGA) infant was 22%. Low tE3 implied a 27% risk. When both were low, the risk of a SGA infant was 54%. The percentages of SGA infants found were 18, 19, and 12%, when ucE3, tE3, or both respectively, were low. ucE3 does not appear to be superior to tE3. From the predictive values found in this study, screening of a total obstetric population with ucE3 and/or tE3 cannot be recommended.