In a previous cross-sectional study of twins, the authors found evidence of a substantial genetic influence on functional abilities among elderly women. It has been suggested that rate of change in functional abilities over time could underlie such findings and that rate-of-change phenotypes may have an even larger genetic component than "level" phenotypes (e.g., functional abilities per se). If so, rate-of-change phenotypes could be more powerful than level phenotypes in studies aimed at identifying specific polymorphisms of importance for aging. In 1995, the authors assessed a population-based sample of 2,401 Danish twins aged 75 years or more. The survivors were recontacted after 2 years and again after 4 years. Consistent mean-level declines, high within-person correlations over time, and substantial heritability in the female sample were observed for functional abilities. Nonetheless, structural-equation analyses revealed only a very modest and nonsignificant heritability for rate of change in functional abilities: 16% (95% confidence interval: 0, 35) for women and 9% (95% confidence interval: 0, 44) for men. This study had a large initial sample size, high participation rates, and a valid and reliable measure of rate of change in a phenotype that had previously shown substantial heritability in cross-sectional analyses in the same twin population. Still, the present study revealed only a modest and nonsignificant genetic influence on rate of change, which suggests that detection of polymorphisms influencing rate of change in functional abilities among the elderly may prove to be difficult.