Multiple myeloma (MM) is a B-cell neoplasm of unknown etiology. We searched for etiological clues by examining the literature on geographic clusters of MM. We searched the MEDLINE database from 1966 to 1996 for spatial occurrences of MM that were significantly greater than expected (spatial "clusters"). Eight clusters with verified diagnoses of MM were identified. All of the eight clusters occurred in locations that were proximate to a body of water. Six of these bodies of water are known to have been contaminated with dioxins. We hypothesize that the observed association between MM and proximity to bodies of water is caused by exposure to dioxins in individuals who consume local fish and seafood. This hypothesis is consistent with the significantly elevated risks for MM in groups with high consumption of dioxin-contaminated fish, e.g., Baltic Sea fishermen and Alaskan Indians, and among persons accidentally exposed to dioxins in Seveso, Italy. Dioxins are immunotoxic and inhibit the differentiation of B cells. Thus, dioxins are plausible myelomagens. A dioxin hypothesis could illuminate many epidemiological features of MM and may suggest new avenues for analytic research.