In recent years, some behavior geneticists have applied the statistical procedures of the population-genetic approach within evolutionary biology to the study of psychological development. A widely recognized deficiency in the population-genetic approach is the fact that it does not deal with individual development. Thus, the application of the statistical methods of population genetics – primarily the analysis of variance – to the study of the causes of psychological development is bound to result in a nondevelop-mental understanding of developmental outcomes. Further, the overly narrow reaction-range concept of behavior genetics is shown here to be invalid for biological as well as psychological phenomena and the open-ended norm of reaction is empirically demonstrated to be a more appropriate developmental conception within both biology and psychology. The present critique impugns the population approach to developmental behavior-genetic analysis and offers a basis for the practice of a truly developmental behavior genetics. To that end, a multilevel, bidirectional, developmental systems model is presented as a way of proceeding analytically from genetic activity to behavior (and back again).