Although several studies have shown a lower risk of non-Hodgkin lymphoma (NHL) in alcohol drinkers compared with nondrinkers, the dose-response relation and potential differences between former and current drinking and across beverage types and subtypes are unclear. The authors examined associations of alcohol intake with risk of NHL and NHL subtypes in the Cancer Prevention Study II Nutrition Cohort, a prospective study of US men and women aged 50-74 years. Between 1992 and 2007, there were 1,991 incident NHL cases among 143,124 participants. Multivariable-adjusted relative risks and 95% confidence intervals were computed using Cox proportional hazards regression. Compared with nondrinkers, the relative risk of NHL associated with former drinking was 0.90 (95% confidence interval (CI): 0.75, 1.10); the relative risks associated with current intakes of <1, 1-2, and >2 drinks/day were 0.93 (95% CI: 0.83, 1.03), 0.91 (95% CI: 0.78, 1.06), and 0.78 (95% CI: 0.65, 0.93), respectively. Associations did not differ by sex (P-interaction = 0.45) or beverage type (P-difference = 0.22). Alcohol intake was more strongly associated with B-cell lymphoma (P-trend = 0.005) than with T-cell lymphoma (P-trend = 0.76), and associations were similar among B-cell lymphoma subtypes. In this prospective study, current heavy alcohol intake was associated with a reduced risk of NHL. Associations did not differ by beverage type and were slightly stronger for B-cell tumors than for T-cell tumors.