Studies were initiated to assess the response of patients with disseminated melanoma to recombinant alpha interferon (rIFN-alpha A) and to monitor effects of rIFN-alpha A on several tests of immune function. Twenty patients were treated with rIFN-alpha A given by i.m. injection in escalating doses from 15 to 50 X 10(6) um-2. The responses of two patients were considered unevaluable. Of the remainder there was complete remission of tumour in two and stable disease in two. Subsequent progression of tumour in one of the latter patients coincided with development of antibodies to IFN. Side effects (usually fatigue) were dose rate limiting in 11 patients. Laboratory tests on samples taken 6 hours after rIFN-alpha A indicated a marked lymphopenia and a reduction in natural killer (NK) cell activity particularly against K562 target cells. Longer term changes measured in samples taken 2 days after the previous rIFN-alpha A injections consisted of neutropenia and an increase in the T4/T8 ratio due mainly to a relative increase in OKT4 positive T cells compared to OKT8 positive T cells. NK activity against the K562 target cell increased in most patients during the first week of treatment and then returned to below or near pretreatment levels thereafter against the K562 target cell. This contrasted with NK activity against the melanoma target cell which showed a more gradual increase over the duration of the treatment in 6 patients. The latter correlated with an increase in mitogen stimulated IL 2 production from their blood lymphocytes and may indicate that the cytotoxic activity resulted from lymphokine-activated killer (LAK) cells. These results confirm the activity of rIFN-alpha A against melanoma in certain patients. They suggest that further studies are needed to select patients who may respond to rIFN-alpha A and to optimize treatment regimens. Tests of IL 2 production and LAK activity may assisted in achieving these objectives.