Female registered nurses in the United States who responded to a questionnaire in 1976 that inquired about height, weight, and smoking history were followed for the development of colon or rectal cancers through May of 1984. Among the 118,404 respondents free of cancer in 1976, 191 colon cancers and 49 rectal cancers were observed during 916,170 person-years of follow-up. After omitting cases diagnosed within two years of weight report, we found little overall relation of body mass (Quetelet's) index to colon cancer risk; however there was a suggestion of elevated risk for the heaviest category of body mass index (greater than or equal to 29 kg/m2, relative risk (RR) = 1.5; 95 percent confidence interval = 0.8 - 2.7) relative to the lowest category (less than 21 kg/m2). Self-reported body mass index from adolescence had a slightly more pronounced, although not significant, association with risk of colon cancer. Increasing height was significantly associated with colon cancer (RR = 1.6, 95 percent confidence interval = 1.1 - 2.5 for the tallest category [greater than or equal to 168 cm] versus the shortest (less than 160 cm], trend, P = 0.04). Measures of current or past smoking failed to demonstrate any consistent relationship with colon cancer.