Few prospective studies have examined the associations between blood levels of folate, in conjunction with methylenetetrahydrofolate reductase (MTHFR) polymorphisms, and colorectal cancer. We evaluated the associations between plasma folate, MTHFR C677T, and A1298C, and colorectal cancer in three large prospective studies: the Nurses' Health Study, the Health Professionals Follow-up Study, and the Physicians' Health Study. A total of 602 incident cases were identified and individually matched to controls who provided blood specimens. We used conditional logistic regression to calculate the relative risk (RR) and 95% confidence interval (95% CI) and then pooled the estimates using a random effects model. We found a lower risk of colorectal cancer among participants with low plasma folate levels: compared with the lowest quartile, RRs (95% CIs) for each successively higher quartile of plasma folate levels were 1.55 (1.14-2.11), 1.37 (1.00-1.88), and 1.47 (1.07-2.01; P for trend = 0.10). For the MTHFR polymorphisms, RRs (95% CIs) were 0.62 (0.44-0.90) for 677TT versus CC/CT and 0.68 (0.31-1.51) for 1298CC versus AC/AA, and these lower-risk genotypes were associated with lower circulating plasma folate levels. When we partitioned the variation in plasma folate levels, variation due to folate intake was not positively associated with colorectal cancer risk. We found that low plasma folate levels were associated with lower risk of colorectal cancer. The reasons underlying a lower risk of colorectal cancer with low plasma folate levels require elucidation because plasma folate levels can reflect dietary intake, genetic influences, and other factors.