The associations between soft tissue sarcoma (STS) and occupational exposures were studied in a case-referent study in the southeast of Sweden. Exposure information was obtained through mailed questionnaires to 96 cases, 450 randomly selected population referents, and 200 cancer referents. Odds ratios (OR), were calculated for various occupational groups, and particularly, for occupations with potential exposure to chlorinated phenoxy herbicides and chlorophenols. In the analyses based on population referents, increased risks for soft tissue sarcoma were seen for especially gardeners (OR = 4.1), but also railroad workers (OR = 3.1); construction workers with exposure to impregnating agents (OR = 2.3), asbestos (OR = 1.8), or pressure impregnating agents (OR = 1.7); and unspecified chemical workers with potential exposure to phenoxy herbicides and/or chlorophenols (OR = 1.6). A similar pattern appeared when cancer referents were used although the numerical values of the odds ratios became different. A grouping of jobs resulted in Mantel-Haensel OR from 1.5 to 1.9 for farmers and forestry workers, dependent on referents used and even more increased OR for railroad workers and unspecified chemical workers with potential exposure to phenoxy herbicides and chlorophenols. The results of the study confirm rather than refute that phenoxy herbicides and chlorophenols could be of etiologic importance for STS; the high risk for gardeners, although based on a small number of individuals, was unexpected and remains unclear. Also, since other cancers were used as referents, no definite problems of recall bias should obtain in this material. None of the exposed groups had a higher proportion of smokers than the unexposed group.