In the context of a search for testable etiological theories of pedophilia, the relationship of pedophilia to partner sex preference was investigated. The proportional prevalences of gynephilia and androphilia were compared with the proportional prevalences of sexual offenders victimizing female children and of such offenders against male children. Since pedophilia either does not exist at all in women, or is extremely rare, only men were included in the study. We derived the proportional prevalence of androphilia from a review of the main pertinent studies, including Gebhard's reassessment of the study by Kinsey et al. particularly of the section on gynephilia vs. androphilia. The numbers of heterosexual vs. homosexual offenders against children were derived from the studies by Mohr et al., by Gebhard et al., and from a group of 457 sex offenders against children seen in the course of several years at the Department of Behavioural Sexology of the Clarke Institute of Psychiatry. There was a large difference between the proportion of prevalences of heterosexual vs. homosexual offenders against children on the one hand and the proportional prevalences of gynephilia vs. androphilia, on the other. This difference suggests that the development of erotically preferred partner sex and partner age are not independent of each other and that in pedophilia, the development of heterosexuality or homosexuality is brought about by factors different from those operative in the development of androphilia or gynephilia.