The etiology of mycosis fungoides is unknown. Two possible causes (an unknown retrovirus with increased prevalence among never-married men, and prior malignancies) were investigated to determine whether they are associated with the incidence of mycosis fungoides. During 1973 to 1986, 953 case patients with mycosis fungoides or Sézary syndrome were registered by the Surveillance, Epidemiology, and End Results program. Each was matched by 5-year age group, sex, ethnicity, and geographic area to four control subjects, one each with cancer of the pancreas, brain, and stomach, and non-Hodgkin's lymphoma. For never-versus ever-married men, none of the relative risks differed significantly from those for women (odd ratios, .8-1.0). For any prior malignancy, the relative risks (and 95% confidence intervals) were 1.3 (.9-2.0), 1.2 (.8-1.8), 1.0 (.7-1.5), and 1.1 (.7-1.6). These data reject the previous relative risk estimate of 3.3 with greater than 99% power, and are consistent with only a small risk, if any, attributable to prior malignancy.