In order to determine the cost of radiation treatment and the survival rate of a cohort of patients treated in a 6-month period in 1988, estimates of the capital and recurrent costs of this service were made for the calendar year 1988, expressed as $A(1988). Data collected prospectively included workload statistics, including number of attendances, field treated and complexity of treatment. Patient and tumour-related data included tumour site, intent of treatment and survival. The survival rate of patients during this period was determined in June 1995. The cost per field in 1988 was estimated at $A44.32. The 1988 costs of courses of definitive, adjuvant and palliative radiation therapy were estimated at $A2545, $A2482 and $A929, respectively. The major contributor to the cost was salaries and consumables within the Radiation Oncology Department (81.6%), with capital costs accounting for 13.5%, overheads accounting for 4.5% of the costs and planned admissions accounting for 0.2%. The median survival time of 580 patients with malignant disease treated during this period in 1988 was 12.4 months. The overall 5-year survival rate was 27%. For 105 patients treated definitively with radiation therapy, the median and 5-year survival rate figures were 26.0 months and 40%. For 149 patients treated with adjuvant radiation therapy, the 5-year survival rate was 62% (median survival rate not reached). For 279 patients treated palliatively, median and 5-year survival rate figures were 5.2 months and 3%. The cost per month of survival for all patients with malignancy was $A67, the figures for definitive, adjuvant and palliative treatments being $A74, $A48 and $A105, respectively. A sensitivity analysis indicated that these figures were robust. The cost of radiation treatment per field was comparable to reports for other centres and emphasizes the utility of radiation therapy as a cost-effective cancer treatment modality. With increasing pressure on treatment facilities in the public sector, further cost/benefit analysis is required to optimize the use of restricted resources.