During spermatogenesis in human adults, testicular germ cells proliferate, differentiate, and die by apoptosis. However, little is known about the temporal or spatial nature of this programmed cell death. Such information may be useful for understanding prenatal developmental biology as well as spermatogenesis during adulthood, particularly in the context of germ cell disorders. We undertook this study to determine 1) whether apoptosis occurred in a cell-specific fashion in the germ cell population and the supporting somatic cells; and 2) whether apoptosis varied with gestational age. We examined human fetal testicular tissues obtained from 17 karyotypically and structurally normal fetuses of mothers who underwent spontaneous or induced abortions. Three gestational ages were defined as follows: group A, 12-13 wk gestation (n = 5); group B, 20-22 wk gestation (n = 7); and group C, 37-40 wk gestation (n = 5). Morphology in conjunction with in situ end labeling was used to identify and quantify apoptotic nuclei in fetal gonadal tissues. The results of this study suggest that gonadal apoptosis occurred in germ cells, Sertoli cells, and Leydig cells at all gestational ages. Apoptotic death was highest in the Leydig cells, followed by germ cells and Sertoli cells. There was a significant positive correlation between the apoptosis of germ cells and Sertoli cells (P <0.01) and a negative correlation between healthy germ cells and Sertoli cells (P <0.001). There was also a negative correlation between the intratubular cell number and the gestational age. Specifically, the proportion of Sertoli cells decreased with gestational age, although there was no significant change in the germ cell in relation to gestational age. No such relationship was found in the Leydig cell population, all of which reside outside the seminiferous tubules. These results are the first to suggest that fetal testicular apoptosis begins in the first trimester, occurs in the three major cell types, and continues throughout pregnancy. Our data also suggest that in the fetal gonad, germ and Sertoli cell proliferation and death may be controlled by a genetic program distinct from that of the Leydig cells. This information is relevant to the understanding of abnormal spermatogenesis associated with infertility and to germ cell tumors in adult life.