OBJECTIVE To assess the ability of noninvasive tests of fetal wellbeing to predict hypoxic morbidity independent of fetal size. METHODS A prospectively planned, longitudinal, observational study. METHODS Fetal Surveillance Unit, King's College Hospital, London. METHODS One hundred and ninety-one pregnant women with singleton pregnancies who were delivered after 32 weeks' gestation who were seen in the Fetal Surveillance Unit within seven days of delivery. METHODS Fetal surveillance by fetal abdominal circumference, heart rate variability (mean range), biophysical profile score and umbilical artery pulsatility index measurements. METHODS Birthweight was classified as > or < or = 2.5th centile (AGA or SGA) for gestational age and sex. Morbidity was defined as at least one of the following at birth: delivery by emergency caesarean section for fetal distress, umbilical venous blood pH less than 7.15, 5 min Apgar score less than 7 or admission to the Special Care Baby Unit (SCBU). RESULTS Fetal abdominal circumference was the best indicator of which fetuses (n = 30) would be SGA. Fourteen of the 30 (47%) SGA fetuses had morbidity at birth and abnormal umbilical Doppler studies significantly predicted this (chi 2 = 2.93, P = 0.003). By contrast, fetal heart rate variability and the biophysical profile score did not. Twenty-seven of the 161 (17%) AGA fetuses had morbidity at birth, but this was not significantly predicted by heart rate variability, biophysical profile score or umbilical Doppler studies. CONCLUSIONS None of the antenatal testing techniques studied predicted morbidity in normally grown fetuses but Doppler studies indicated whether a small fetus was 'sick small' or 'normal small'.