Autologous bone marrow transplantation (ABMT) is now widely used as salvage therapy in Hodgkin's disease but its value will have to be finally proved by the use of randomised trials. However ABMT is now being used in first remission. The justification of this is based on the view that patients can be identified once remission is achieved who are at high risk of relapse and that these patients will be prevented from relapsing by high dose therapy and ABMT in first remission with an 'acceptable' procedure related toxicity. The choice of patients judged to be suitably poor risk is critical but unproven. The only report of ABMT in CR1 yet published [1] is encouraging but can be criticised on a number of grounds. The incidence of procedure related mortality is a significant factor in determining the balance of any risk/benefit analysis. Randomised trials need to be undertaken but must identify the correct group of patients to avoid treating with high dose therapy those who may already have been cured by initial therapy.