High-dose chemotherapy and radiotherapy with autologous bone-marrow transplantation (ABMT) are increasingly used for the treatment of relapsed and resistant Hodgkin's disease, although there has been no randomised trial of this treatment. The British National Lymphoma Investigation therefore undertook a randomised comparison of high-dose chemotherapy (BEAM = carmustine, etoposide, cytarabine, and melphalan) plus ABMT with the same drugs at lower doses not requiring bone-marrow rescue (mini-BEAM) in patients with active Hodgkin's disease, for whom conventional therapy had failed. 20 patients were assigned treatment with BEAM plus ABMT and 20 mini-BEAM. All have been followed up for at least 12 months (median 34 months). 5 BEAM recipients have died (2 from causes related to ABMT and 3 from disease progression) compared with 9 mini-BEAM recipients (all disease progression). This difference was not significant (p = 0.318). However, both event-free survival and progression-free survival showed significant differences in favour of BEAM plus ABMT (p = 0.025 and p = 0.005, respectively). Recruitment to the trial became increasingly difficult because patients refused randomisation and requested ABMT. It was therefore closed early (40 patients rather than 66 intended). Nevertheless, we found a dose-response effect in these patients with relapsed and resistant Hodgkin's disease. High doses facilitated by ABMT can lead to better disease-free survival.