In a multicentre randomised clinical trial 364 children with biopsy proven medulloblastoma were randomly assigned to receive or not pre-radiotherapy chemotherapy. Children with total or subtotal removal of the tumour, no evidence of invasive brain stem involvement, and no evidence of metastatic disease either within or without the cranium were designated "low risk", those with gross residual tumour, evidence of invasive brain stem involvement or metastases in the central nervous system were designated "high risk". All children were prescribed 55 Gy to the tumour bearing area. "Low risk" children could be randomised to "standard" radiotherapy 35 Gy to the craniospinal axis or "reduced" dose 25 Gy to the craniospinal axis. Chemotherapy consisted of vincristine, procarbazine, and methotrexate given in a 6-week module before radio-therapy, and for "high risk" children, vincristine and CCNU given after radiotherapy. No benefit for the receipt of pre-radiotherapy chemotherapy could be demonstrated for any group. In addition, a negative interaction was observed between the receipt of the chemotherapy and reduced dose radio-therapy with a particularly poor outcome being observed in this group of children.