tial effect of gene-environment correlations that could affect environments that matter.
Fourth, concerns have been expressed over the fact that measures of many traits do not show a normal distribution (Capron and Vetta, 2001). This is not a valid criticism because there is no necessary assumption that the “true” distribution is normal and because there are various statistical techniques that can be used to normalize distributions and hence to make them more suitable for the usual genetic analyses. There is no reason to suppose that this creates a significant problem.
At first sight, adoptee designs might seem preferable to twin designs because they provide a cleaner separation of biological heritage and rearing environment. For that reason they have indeed proved very valuable with respect to some phenotypes, most particularly schizophrenia (Kety et al., 1994). However, there are more limitations to adoptee designs than usually appreciated by behavior geneticists. To begin with, social agencies select prospective adoptive parents on the grounds that they will provide a good rearing environment. Of course, they succeed in that goal to only a limited extent, but there is no doubt that extremely adverse environments are greatly underrepresented in samples of adoptive parents (Rutter et al., 2001; Stoolmiller, 1999). Also, adoptive parents tend to be older and better educated than biological parents in the general population. In addition, women who give up their children for adoption are far from a random sample of the general population. With respect to the study of fertility, there is also the obvious problem that, by definition, biological parents cannot include any individuals who have not had children. Finally, in most industrialized countries the number of nonhandicapped children adopted in infancy has dropped greatly. Intercountry adoptions have risen pari per su but these involve a much more complicated situation, often including serious early deprivation and adoption after infancy (Selman, 2000).
For all these reasons, it is not likely that adoptee studies will have a major place in the study of fertility, and because, so far as I am aware, there have been no adoptee studies that are relevant, these designs will not be considered further here.
If only because the theory of evolution is based on natural selection for reproductive fitness, it is necessary to consider how evolutionary theory might inform the quantitative genetic study of fertility. Indeed, as other chapters in this volume illustrate, it does include most useful pointers as to