A total of 22 monozygotic (MZ) and 41 dizygotic (DZ) twin pairs were given 72 trials on a stabilometer balance task over six days to study the extent of the genetic contribution to learning and performance of a gross motor skill. The expectations that interindividual differences would be less for the MZ than for the DZ twins and that intraindividual variability would not be different between the two groups were supported. Intraclass correlations were used to provide estimates for the proportions of total phenotypic variance accounted for by heritability (h2), systematic environmental variance (E2), and nonsystematic environmental effects (e2). Heritability was found to be low during the early stages of learning, before it increased to stabilize at approximately 65% for the remaining practice. E2 was highest during these early stages (24%), then declined quickly to stabilize at half that level. Error variance (e2) constituted the remaining variance. Learning profiles of the twin pairs were also analyzed, with a greater intrapair resemblance being found for the MZ twins. The present findings indicate that, for gross motor skills, there is considerable potential for influencing both the levels of performance (and learning) and the differences between individuals by judicious use of systematic environmental effects.