The roles of nonagricultural occupations, tobacco use, beverage consumption, medical history, and other factors in the development of soft-tissue sarcoma were examined in a population-based case-control study in Kansas. Based on 133 cases diagnosed between 1976-1982 and 948 controls, there were significant excesses associated with use of the drug chloramphenicol (odds ratio (OR) = 5.4, 95% confidence interval (Cl) 1.2-23.9) and chewing tobacco or snuff (OR = 1.8, 95% Cl 1.1-2.9). The risk associated with smokeless tobacco varied with the location of the tumors; greater risks were observed for tumors of the upper gastrointestinal tract (OR = 3.3), the lung, pleura, and thorax (OR = 3.1), and the head, neck, and face region (OR = 2.4) than other regions of the body (OR = 1.4). A nonsignificant excess was seen with the use of cholesterol-lowering drugs, such as clofibrate (OR = 1.7). Four cases reported histories of prior radiation treatment to the same area of their bodies as their tumors. Soft-tissue sarcoma was also associated with employment in woodworking occupations (OR = 1.7, 95% Cl 0.9-3.2) and risk increased with increasing duration of employment. Persons with first-degree blood relatives with a history of Hodgkin's disease, lymphoma, or cancers of the pancreas, prostate, brain, or skin were at increased risk. Many of the associations observed in this study, notably the risk of soft-tissue sarcoma with smokeless tobacco and medications such as chloramphenicol, deserve further evaluation.