Incidence date from the Third National Cancer Survey were used to study geographic variation in the occurrence of nine common cancers of nonsexual sites and five sexual sites. It was hypothesized that cancers which varied together across the nine survey communities might have common etiologic factors. The variation was measured by a series of rank and product-moment correlation coefficients in order to summarize the association between pairs of cancers in white males and in white females. The most notable findings were the high correlations between the incidence rates for the three gastrointestinal sites (stomach, colon, and rectum) and bladder cancer in both men and women and the high correlation between three female sexual sites (breast, corpus, and ovary). These two clusters are also correlated with one another. A number of other substantial correlations between pairs of cancers were observed and discussed. These associations suggest possible common etiologic agents, despite the fact that the individual secular trends for some of these cancers differ.