Epidemiological studies have consistently shown a positive association between size at birth (i.e. birth weight or birth length) and height in children, adolescents and adults. To examine whether this association is explained by genetic or nongenetic (intra-uterine) factors, we investigated birth weight, birth length and height in 60 dizygotic and 68 monozygotic adolescent twin pairs still living with their parents. Birth weight of the twins was obtained from their mothers. Height was measured in a standardised way. The mean age was 17±1.7 years for the dizygotic twins and 16±1.8 years for the monozygotic twins. Both dizygotic and monozygotic twins with the lowest birth weight from each pair had a height that was lower compared to their co-twins with the highest birth weight (dizygotic twins: 172.2±7.9 vs. 173.8±9.4 cm [p = 0.05]; monozygotic twins: 171.1±9.4 vs. 171.8±9.5 cm [p = 0.01]). Similarly, both dizygotic and monozygotic twins with the shortest birth length from each pair had a height that was lower compared to their co-twins with the longest birth length (dizygotic twins: 172.3±7.9 vs. 174.9±9.7 cm [p < 0.05]; monozygotic twins: 168.9±10.6 vs. 169.9±10.2 cm [p < 0.01]). In addition, intra-pair differences in birth weight and birth length were significantly associated with differences in height in both dizygotic twins (regression coefficient: 4.3 cm/kg [95% confidence interval: 1.0 to 7.5] and 0.96 cm/cm [0.17 to 1.74], respectively) and monozygotic twins (2.8 cm/kg [1.4 to 4.1] and 0.73 cm/cm [0.40 to 1.06], respectively). These associations were stronger in dizygotic than in monozygotic twins, but this difference was not statistically significant (for birth weight p = 0.4; and for birth length p = 0.6). However, genetic model fitting indicated that models incorporating a genetic source of the covariance gave a better description of the observed association of birth weight and length with height in later life than models not incorporating this genetic source. The results were similar for data on adult height after 12 years of follow-up in a subgroup of these twin pairs. These data suggest that the association between size at birth and height in later life is influenced by non-genetic intra-uterine and by genetic factors.