A study of sex ratios according to tumor sites in 65,000 cases of cancer is reported. Where data on etiological factors were available, it was found that the higher the correlation between a site or group of sites and the use of tobacco and alcohol, the higher the sex ratio. Where etiological data were not available, it was hypothesized that where sex ratios vary for different sites, especially within the same organ, differing etiological factors are likely to be found. For many primary sites and subsites, these ideas fit in well with usual clinical findings or anatomical and physiological considerations. A striking finding emerged especially for gastric cancer, in which no association usually appears with the use of tobacco or alcohol: For the fundus cardia area, the sex ratio was clearly higher than it was for all other regions of the stomach. Thus, this is an area that could be affected by the use of tobacco or, more likely, alcohol, just as are certain areas of the upper respiratory and digestive tracts.