A prospective randomized trial of low versus high doses of human leukocyte alpha-interferon (1 X 10(6) units/day for 28 days versus 10 X 10(6) units/day for 28 days) was carried out in 30 patients with metastatic renal cell carcinoma, to test the tolerance and relative antitumor effects of these interferon doses. Both doses were tolerated well, and responses to the human leukocyte alpha-interferon were observed overall in seven individuals, including complete, partial, and minimal tumor regressions. Six of the seven responses occurred in patients who received the high dosage, and three of these responses were major responses. While not statistically significant, this result suggested a dose-response relationship. One minimal response was observed in a patient treated at low dosage. Nine individuals who were stable after 1 month of therapy at low dosage were randomized to a further month of therapy at low or high dosage, during which one of four at high dosage had a partial response, and none of five at low dosage manifested response. Regression of pulmonary disease in one individual was delayed, occurring 3 months after therapy at the high dose and enduring for a period of 28 months. Major objective responses in other patients were of 4 and 15 months duration. Human leukocyte alpha-interferon is an active agent in renal cell carcinoma at the dosage of 10 million units daily. No relationship of toxicity to response was evident in this trial. Optimum dosage and duration of treatment have yet to be established.