Mapping the geographic distribution of human T-lymphotrophic virus types 1 and 2 (HTLV-I and -II) has been complicated because conventional serologic approaches cannot distinguish between these 2 viruses. To more precisely define the epidemiology of HTLV-I and HTLV-II, we evaluated a convenience sample of 4,832 HTLV-I immunoblot results from over 140,000 samples screened for HTLV-I, collected in 16 countries for 35 separate studies. An algorithm that compares reactivity against p19 and p24, 2 gag (HTLV-I core) proteins, was employed to characterize the immunoblots: type I, p19 stronger than p24 (presumptive HTLV-I), type 2, p24 stronger than p19 (presumptive HTLV-II), or indeterminate (p19 and p24 weakly positive or p19 weakly positive in the presence of p21e). Geographic areas could be grouped into 4 patterns. Patterns A (> 75% type I) and B (> 75% type 2) were usually observed where the specific type of HTLV or its characteristic diseases had been found. Pattern C (mixed type 1 and 2 pattern) was observed predominantly in intravenous-drug-using and other populations in which both virus types have been reported. Pattern D (> 10% indeterminate), suggests the presence of non-specific reactivity, perhaps resulting from exposure to non-virus-related antigens or an HTLV-related virus. HTLV-I predominates in southern Japan, the South Pacific, parts of West Africa, and in Afro-populations of the Western hemisphere. HTLV-II clusters in Native American populations and among intravenous drug users. Pattern-D areas in Africa and Venezuela might prove to be fertile in the search for new and variant HTLV virus types.