This study first examines the hypothesis that significant intergeneration increases in stature within 85 Taiwanese families in two study locations were primarily the result of increases in subischial leg length. It then evaluates a second hypothesis that independent assessments of the extent of intergenerational change in childhood environments within these families helped account for parent-offspring differences in relative leg length. Childhood environments were assessed using two criteria: developmental environment scores derived from life history data and evidence of linear enamel hypoplasia (LEH). Relative leg length was represented in two ways, as either the difference between internally estimated sex-specific z-scores of sitting height and subischial leg length (z(D) = z(SH) - z(LL)) or as relative sitting height (RSH; sitting height/height x 100). Paired Student t tests indicated intergenerational increases in relative leg length were significantly greater than zero in both study locations (P < or = 0.012). The second hypothesis, tested using a hierarchical model with maximum likelihood estimation that allowed for nesting of more than one offspring per family, received support as midparent-offspring differences in composite scores were significantly positively associated with midparent-offspring differences in relative leg length. This was true alone (P = 0.018), and when significant associations with LEH prevalence among mothers and offspring were statistically controlled for (P = 0.010). Evidence also indicated that while the large majority of offspring were taller than their midparental average height (84.3%; 91/108), offspring who were taller were also more likely to have relatively longer legs than by chance alone (Fishers exact, P = 0.027).