To investigate whether the association between agricultural pesticide use and the risk of non-Hodgkin's lymphoma (NHL) is modified by a family history of hematopoietic cancer, including leukemia, myeloma, and lymphoma, we analyzed pooled data on white men from three population-based, case-control studies of NHL conducted in Iowa/Minnesota, Kansas, and Nebraska. Information on the agricultural use of insecticides, fungicides, and herbicides; a family history of cancer; and other risk factors was obtained by interviewing 973 cases and 2,853 controls or, if deceased, their next-of-kin (37% of cases, 43% of controls). The NHL risk was estimated by odds ratios (ORs) and 95% confidence intervals (CIs), adjusted for age, state of residence, type of respondent, and use of hair dye. Compared to men with no family history of cancer, the ORs (95% CIs) of NHL was 1.5 (1.3-1.8) for men with a family history of nonhematopoietic cancer, and 2.7 (1.9-3.7) for those with a history of hematopoietic cancer among first-degree relatives. This positive association was noted for each group of NHL defined according to the Working Formulation, and was most pronounced for small lymphocytic NHL. Among direct respondents, farmers who used pesticides and had a positive family history of cancer or hematopoietic cancer were not at elevated risk of NHL, compared to nonfarmers who had no family cancer history. However, among proxy respondents, ORs were elevated for farmers who had a positive family history of hematopoietic cancer and used animal insecticides (OR = 4.6; 1.9-11.2), crop insecticides (OR = 4.7; 1.6-13.4), or herbicides (OR = 4.9; 1.7-14.2), although the interaction of family history of cancer and agricultural pesticide use was not statistically significant. In summary, the joint effects of the family cancer history and pesticide use were limited to proxy respondents with wide CIs and, thus, provide little evidence that a family history of cancer modifies the association of agricultural exposures with NHL.