Environmental exposure to polychlorinated biphenyls (PCBs) and p,p'-dichlorodiphenyldichloroethylene (p, p'-DDE) has been associated with the risk of non-Hodgkin lymphoma. We conducted a case-control study nested within the Physicians' Health Study, a prospective cohort established in 1982. We measured concentrations of PCBs and p,p'-DDE in baseline blood samples from 205 men later diagnosed with non-Hodgkin lymphoma and 409 age- and race-matched controls. Lipid-adjusted organochlorine concentrations were categorized into quintiles based on the distribution among controls. We used conditional logistic regression to estimate the odds ratios (ORs) and 95% confidence intervals (CIs) for each quintile relative to the lowest quintile. We also evaluated these associations for major histologic subtypes of non-Hodgkin lymphoma. The risk of non-Hodgkin lymphoma was positively associated with the sum of 51 PCB congeners assayed (SigmaPCB); the group of immunotoxic congeners; the individual congeners 118, 138, 153, and 180; and the sum of these 4 congeners. The simple OR for the highest quintile of lipid-adjusted SigmaPCB versus the lowest was 1.9 (95% CI = 1.1-3.2; test for trend, P = 0.001), with similar trends for individual congeners and groups defined as above. Adjustment for height, body mass index, alcohol intake, smoking, and fish intake did not substantially change the effect estimates. No association was observed for p,p'-DDE. There was no evidence of statistical heterogeneity in effects by histologic subtype of lymphoma; however, this analysis was underpowered. These results support the hypothesis of a positive association between PCB exposure and development of NHL in men.