Bone marrow transplantation (BMT) has been recommended for children with high-risk acute lymphoblastic leukaemia (ALL) in first remission. The recent MRC UKALL X trial was designed to facilitate a non-randomised comparison between BMT and chemotherapy in children deemed to be at high risk of treatment failure. 198 children aged 1-15 had a presenting leucocyte count of more than 100 x 10(9)/l. All received induction and early intensification therapy. Children with an HLA-compatible sibling donor were eligible for BMT in first remission. All other children received cranial irradiation at 24Gy, late intensification, and two years of continuous treatment. 183 children achieved a stable remission of whom 111 were HLA typed; these tended to be older and to have T-cell ALL. A donor was identified in 41 cases, of whom 34 proceeded to BMT at a median time of 17 weeks; there was no difference in distribution of age, sex, or leucocyte count between the groups receiving BMT and chemotherapy. Comparison of the 144 children who were in remission at 17 weeks and received chemotherapy with the 34 proceeding to BMT showed no significant difference in event-free survival at five years (69% for BMT and 52% for chemotherapy). There were significantly more treatment-related deaths in the marrow transplant group (6 vs 4) and more relapses in the chemotherapy group (59 vs 4). There was no significant difference in event-free survival between children who were HLA typed and had a donor and those without a donor, although there were fewer relapses among the former. BMT can be evaluated in the context of a multicentre trial for paediatric ALL but the number of children with donors is too small to make a significant impact on overall survival. However, marrow transplantation was associated with a much lower relapse rate than that with the UK ALL protocol, and with better definition of higher risk patients BMT may be of benefit in some children with high-risk ALL in first remission.