221 small-for-dates (SFD) and 244 average-for-dates (AFD) children were personally examined at birth, and seen thereafter at regular intervals up to 4 years, when a thorough assessment was made. There was an excess of SFD children with major congenital abnormalities; but no difference between the groups in the number of children who were handicapped without congenital abnormality. Their general health was good, and the prevalence of specific disease or major ill-health did not differ. SFD girls seemed to have less acute hearing than AFD girls; but there was not difference for boys. The prevalence of squint, impaired vision, speech defects, and abnormalities of gross and fine motor movements were equally low in both groups. At 4 years the mean scores for five sectors of development were significantly lower in the SFD than the AFD group. Within group analyses of 16 variables and their effects on developmental scores showed no associations in either group according to maternal height and weight, birthweights of previous siblings, the presence or not of hypertension and pre-eclampsia, bleeding during pregnancy, asphyxia and injury at birth, or birth order. Within each group the net effect of the remaining eight factors was assessed; adjustment being made for the other seven variables. In the SFD group social class and method of delivery made a significant contribution to total scores. In the AFD group significant differences were found according to social class, sex, gestational age at birth and smoking. Method of infant feeding just failed to achieve significance. When all the children were considered together, and birthweight group included as an additional variable, no significant differences remained between the SFD and AFD groups for any sector of development.