A death certificate-based case-control study was performed to investigate associations between occupation and non-Hodgkin's lymphoma (NHL) in North Carolina. Cases consisted of 501 men who died of NHL (International Classification of Diseases codes 200 and 202) during the years 1968-1970, 1975-1977, and 1980-1982. Controls were selected from other noncancer deaths, and were frequency matched for age, year of death, and race. Occupation and industry were obtained from the death certificates and coded without knowledge of case-control status. An increased risk for men in professional, technical, and managerial occupations, compared with all others, was detected among whites (OR = 2.69, 1.95-3.72). Black men classified as having "low exposures" by an occupational exposure linkage system had an odds ratio of 1.74 (0.84-3.60). Because of this finding, the occupations were ranked by social class and a statistically significant linear relationship was noted in whites, with risk increasing from lower social class to upper social class. An increased risk was also detected among whites in the rubber, plastics, and synthetics industries (p = .03), and among blacks employed in machine trades occupations (OR = 3.63, 1.32-9.97) and structural work occupations (OR = 2.38, 0.93-6.05). An increased risk was also detected for black painters (p = .02), but not for whites. There was no association found between NHL and employment in the following areas: textile industry; farming; laborers; or occupations with exposures to asbestos or benzene. The association with farming was further examined in counties with high use of pesticides and herbicides, and no increased risk of NHL was detected. Cases were more likely to live in the western part of the state than the eastern. However, NHL mortality rates provided by the North Carolina State Center for Health Statistics did not confirm the relationship.