This study evaluated mortality rates from leukemia and other diseases during the time period 1944 through 1998 among 17,924 men employed in the synthetic rubber industry. In this group, there were 6237 deaths, which is 14% fewer than the 7242 deaths expected based on general population rates. Numbers of observed versus expected deaths (shown hereafter as observed/expected) were 1608/1741 for all cancers combined, including 71/61 for leukemia, 53/53 for non-Hodgkin lymphoma (NHL*), and 26/27 for multiple myeloma. The higher than expected number of deaths from leukemia (16% increase) was concentrated in workers paid hourly who had started work 20 to 29 years earlier, had worked 10 or more years in the industry, and had worked in subgroups employed in polymerization, coagulation, maintenance labor, and laboratory operations. The overall higher leukemia mortality rate, as well as the higher rate in the subgroup of hourly workers who had 20 or more years since hire and 10 or more years worked, was not limited to a particular form of leukemia. Cumulative exposure to 1,3-butadiene (BD) was associated positively with all leukemias, with chronic myelogenous leukemia and, to a lesser extent, with chronic lymphocytic leukemia (CLL). Exposure to styrene or to dimethyldithiocarbamate (DMDTC) also was associated positively with leukemia. Exposures to these two agents were correlated with exposure to BD; data were limited on the independent effects of each of the three chemicals on leukemia. After controlling for the effects of BD, we found no consistent exposure-response relation between either styrene or DMDTC and all leukemias, chronic myelogenous leukemia, or CLL. However, a positive association between any exposure to DMDTC and leukemia persisted. The data from this study indicate that employment in the synthetic rubber industry is related causally to leukemia. Uncertainty remains about the specific agent or agents responsible for the association. The carcinogenic mechanisms through which BD, styrene, or DMDTC could cause leukemia in humans have not been established, and epidemiologic support for a leukemogenic role is limited for these agents. Styrene and DMDTC were associated positively with NHL. External support for this relation has not been reported from other epidemiologic studies. The study did not find any clear relation between exposure to BD, styrene, or DMDTC and multiple myeloma. Some subgroups of subjects had more than the expected number deaths from colorectal cancer, prostate cancer, and other diseases. These increases did not appear to be related to occupational exposure in the industry.