From October 1983 until December 1988, 50 patients with asymptomatic multiple myeloma stage I were included in a prospective randomized multi-centre study comparing melphalan-prednisone (MP) therapy started at the time of diagnosis with deferred therapy where MP was started at the time of disease progression. Twenty-five patients were randomized to each group. The median time from diagnosis to start of therapy in the group with deferred therapy was 12 months. The reasons for starting therapy were increasing M-protein in 8 cases, symptomatic bone disease in 9 and anaemia in 5. In 2 cases, disease progression was complicated by vertebral fractures necessitating radiotherapy. Two patients in the group in which MP was started at the time of diagnosis developed acute leukaemia. No differences in response rate, response duration or survival were observed between the treatment groups. We conclude that in asymptomatic myeloma deferral of chemotherapy is feasible in well-informed and well-controlled patients but conveys no advantage in survival. In clinical practice the benefits of treatment deferral are to some extent outweighed by disease progression before start of treatment.