Adult height has been found in some but not all studies to be associated positively with overall cancer incidence as well as several site-specific cancers. The Physicians' Health Study (PHS), a randomized trial of beta-carotene and aspirin in the primary prevention of cancer and cardiovascular disease in men, provided an opportunity to examine the association between height and total malignant neoplasms (excluding non-melanoma skin cancer), as well as site-specific cancers including prostate, colorectal, and lung cancer. The PHS is comprised of 22,071 US male physicians in the United States, a population homogeneous for adult socioeconomic status, aged 40 to 84 years in 1982. Participants were classified into five height categories at study entry. After an average follow-up of over 12 years, there were 2,566 cases of incident total malignant neoplasms, including 1,047 prostate, 341 colorectal, and 170 lung cancer cases. Height was associated positively with both total malignant neoplasms and prostate cancer. Compared with men in the shortest category (<67 inches), relative risks and 95 percent confidence intervals (CI) for total malignant neoplasms for men whose height (in inches) was 68-69, 70-71, 72, and 73+ were, respectively: 1.13 (CI = 0.99-1.28), 1.15 (CI = 1.02-1.30), 1.29 (CI = 1.12-1.49), and 1.21 (CI = 1.05-1.39), P trend 0.001, adjusted for age, randomized treatment assignments, body mass index (wt/ht2), cigarette smoking, alcohol use, and exercise frequency. For prostate cancer, the corresponding RR values were 1.23 (CI = 1.00-1.51), 1.26 (CI = 1.04-1.54), 1.59 (CI = 1.27-1.98), and 1.26 (CI = 1.00-1.59), P trend 0.005. For colorectal cancer, in some but not all height categories compared with the shortest, there were elevated RRs without a significant linear trend: RR = 1.51 (CI = 1.06-2.14), 1.14 (CI = 0.80-1.62), 1.19 (CI = 0.79-1.80), and 1.53 (CI = 1.04-2.25), P trend 0.23. In contrast, there was no evidence of an association of height with lung cancer. These data indicate a positive association between height and risk of total malignant neoplasms, as well as of prostate cancer and, possibly, colorectal cancer.