In a long-term prospective study 46 unselected infants born before 35 weeks of gestational age were followed up, which included repeated neurological and psychological examination. 40 of them were neurologically evaluated at term conceptional age and compared with 26 full-term newborns, all the infants being considered healthy. There was no difference in mean weight or length between the two groups, but mean head circumference was greater in the preterm group. The preterm infants had lower muscle tone as judged by spontaneous posture of arms and legs, and poor resistance to passive movements and slow arm recoil. In the traction test they had more head lag. The withdrawal and Moro reflexes were weaker, while asymmetric tonic neck reflex (ATNR) was easier to elicit. Head control in the sitting position was better. All infants were evaluated with a prenatal and perinatal optimality score, which was not however significantly correlated with the neurological findings. The preterm infants with only slightly reduced optimality score (low risk group) had approximately the same birth weight and gestational age as the other preterms. They more resembled the full-term infants with good resistance to passive movements, fast arm recoil and good responses concerning the withdrawal and Moro reflexes. However, they had a semiflexed position in supine and thus in this respect were more like the other preterms.