Transmission of avian leukosis virus, subgroup J (ALV-J), from donor chickens inoculated as embryos to simulate congenital infection to uninfected hatchmates was studied in two strains of commercial broiler breeder chickens. Chicks of two commercial lines free of ALV-J became infected when hatched (1/2 lots positive) or reared (8/8 lots positive) in direct physical contact with ALV-J-infected donors. Infection also occurred when chicks were exposed in the hatchery to ALV-J-infected donors by cloacal swab transfer (2/2 lots positive), needle transfer during subcutaneous inoculation (2/2 lots positive), or ingestion of infected meconium (2/2 lots positive). However, transmission was delayed or prevented by wire partitions in the hatcher and rearing of small groups in cubicles, and rarely (1/10 lots positive) resulted from short-term direct or indirect contact. In a simulated field test, a flock of 503 broiler breeder chickens with an initial embryo infection rate of 4.6% was hatched and reared as 48 small groups to 4 weeks of age. Groups were tested at hatch and at 3 weeks, and 14 infected groups were eliminated. This flock tested negative for ALV-J infection from 4 to 32 weeks and did not transmit infection to progeny or develop tumours. A control group of 377 chickens with a similar initial infection rate was hatched and reared as a single group. This control flock transmitted virus to 5.7% of its progeny and about 5% of the hens developed tumours. The small-group hatching and rearing practices employed in these studies allowed for the accurate identification and removal of groups containing chickens infected prior to hatching and prevented horizontal transmission of ALV-J between uninfected and infected groups for at least 4 weeks. More importantly, application of these procedures successfully eradicated ALV-J in a single generation under laboratory conditions. This suggests that similar procedures could be a valuable adjunct to virus eradication programmes in the field.