There is concern about the long-term carcinogenic effects of psoralen and ultraviolet A radiation (PUVA) for treatment of skin disorders. Many authors have found an increased risk for cutaneous squamous cell carcinoma (SCC). Except in anecdotal reports, malignant melanoma had not been observed in patients treated with PUVA until recently. In the U.S.A., a 16-centre prospective study of 1380 patients showed for the first time that there might also be an increased risk for malignant melanoma in patients treated with high cumulative dosages of PUVA. We have therefore followed up the Swedish PUVA cohort until 1994. This cohort had previously been followed up until 1985. Information from 4799 Swedish patients (2343 men, 2456 women) who had received PUVA between 1974 and 1985 was linked to the compulsory Swedish Cancer Registry in order to identify individuals with cancer. The average follow-up period was 15.9 years for men and 16.2 for women. We did not find any increased risk for malignant melanoma in our total cohort of 4799 patients treated with PUVA or in a subcohort comprising 1867 patients followed for 15-21 years. For cutaneous SCC there was an increase in the risk: the relative risk was 5.6 (95% confidence interval, CI 4. 4-7.1) for men and 3.6 (95% CI 2.1-5.8) for women. Significant (P < 0.05) increases were also found in the incidence of respiratory cancer in men and women and of kidney cancer in women. In conclusion, we did not find any increased risk for malignant melanoma in our patients treated with high doses of PUVA and followed up for a long time. We confirm previous reports of an increase in the incidence of cutaneous SCC in patients treated with PUVA, and recommend that patients should be carefully selected for PUVA and rigorously followed up.