Male rats were irradiated with 19 r on the day of birth, and killed at intervals ranging from 5 to 18 days. Estimates were made of the absolute and relative numbers of germ cells at different stages of spermatogenesis in 64 irradiated and 61 untreated specimens. In the normal rat, the calculated population of germ cells increased from about 160000 at 5 days to 30 million at 18 days. Only negligible numbers of primordial germ cells (gonocytes and transitional cells) persisted beyond the age of 10 days. Small numbers of spermatogonia type A appeared at 5 days (15000) and their population rose to about 1 million at 12 days, and 2 million at 18 days (7 % of all germ cells). Intermediate spermatogonia first occurred in appreciable numbers (23000 to 55000) at 8 or 9 days, when the population of type-A spermatogonia was 360000. The subsequent rise in the population of intermediate spermatogonia was more rapid than that of type A (4 million at 18 days). Spermatogonia type B and primary spermatocytes appeared at 9 to 10 days, and their numbers rose more steeply still (6.5 and 16 million at 18 days, respectively). Irradiation at birth exerted no rapid effect on the cytological appearance of primordial germ cells. Transformation from gonocytes to transitional cells appeared to proceed normally and the estimated total population of germ cells at 5 days was no smaller than in the controls. Subsequently, however, many of the transitional cells failed to divide: they enlarged to form giant cells, acquired bizarre nuclear outlines, and persisted for unusually long periods. Some degenerated at mitotic prophase or metaphase, while a few seemed to die at interphase, without entering division. The calculated total population of germ cells in irradiated rats rose from 160000 at 5 days to 9.4 million at 18 days. Small numbers of spermatogonia type A, presumably derived from such primordial germ cells as were able to complete mitosis, appeared some 2 to 3 days later than in controls. The number of type-A spermatogonia in 7-day-old irradiated rats was 44000, cf. 215000 in controls; the difference became less pronounced with time, and by the age of 18 days, the population of 1.9 million was comparable to that estimated for the controls. Small numbers of intermediate spermatogonia appeared on the 9th (8000) and 10th day (35000), when the population of type-A spermatogonia was about 110000 and 260000 respectively. By the 18th day, intermediate spermatogonia numbered 2 million. The populations of type-B spermatogonia and primary spermatocytes rose from 11000 to 13000 at 10 days to 1.6 and 3.4 million, respectively, at 18 days. The difference in the absolute and relative numbers of germ cells between normal and irradiated testes widened progressively with advance in the developmental stage of the germ cells. Analysis of the results indicates that in the reduced population of spermatogonia type A after irradiation, the pattern of spermatogonial mitoses is modified so as to favour the formation of more type-A, in preference to intermediate, spermatogonia.