The association between fat intake and several common cancers, eg, those of the colorectum, breast, endometrium, ovary, and prostate, received its strongest support from correlation studies on populations. On an international scale, strong direct correlations were observed between fat intake and incidence or mortality from these neoplasms; several correlations were also observed on a national level and persisted after allowance for major identified covariates. Further support came from the observation of change in rates in migrant groups. This association, however, has been described as being weak in individuals as opposed to populations. Briefly, diets high in fat (and meat) have been associated with high risk of colorectal cancer in several case-control studies, with saturated fat being specifically implicated. However, the strength of the association is generally moderate, and a few disparities have to be considered. In relation to breast cancer, several case-control studies have reported associations with total fat, and there was some indication that the associations might be stronger for saturated fat. These relationships, however, were weak and inconsistent in various studies. Thus, a plausible conclusion from case-control studies is that there is indeed some association between fats and breast cancer risk, which is, however, limited and, hence, extremely difficult to prove in epidemiological terms. To further complicate the issue, the results of cohort studies do not support the association. Data from analytical studies are more limited for endometrial, ovarian, and prostate cancers, but, again, seem to indicate a possible relationship with diets high in fat.(ABSTRACT TRUNCATED AT 250 WORDS)