Colon cancer is regarded as predominantly a cancer of affluent nutrition. A high intake of dietary fat and a lack of dietary fiber have particularly been implicated in both international correlation studies and analytic studies (1). Internationally, the endocrine-influenced cancers and colon cancer show similar variations, and breast cancer is particularly closely linked with colon cancer (2).There is little overall sex difference in the risk of colon cancer, yet for cancers of most other sex-shared sites the male risk is cleariy higher. Although this approximation of the colon cancer sex ratio (male: female) to unity in Western countries has been noted previously, little attention has been given to two additional details.First, in Western countries, the age-specific colon cance.r sex ratios vary considerably. Below age 35 years, the nsk tends to be higher in males; between ages 35 and 54 years, the risk is higher in females; and at older ages, the male risk again becomes higher. Second, a recent secular trend has occurred in the sex ratio whereby the higher RR in women 35-54 years old ha~ begun to disappear, and, simultaneously, the preponderance in males at older ages has increased further.These two characteristics of the colon cancer sex ratio are displayed in table I for ten countries representative of the three major regions of the Western World: Europe, North America, and Australasia.Any secular changes in the validity of cause-of-death certification should affect the sexes similarly. Hence real changes appear to have occurred in the RR between males and females across both age and time. This suggests, in turn, a sex difference in patterns of recent exposure to etiologic factors. Inasmuch as dietary changes probably impinge similarly on men and women, other sex-related exposures must be considered. Occupational factors appear unimportant (l). Alcohol consumption is more likely to influence rectal cancer than colon cancer, although there is some evidence for the alcohol-colon cancer association (6). The most fruitful source of explanation may therefore reside in hormonal and reproductive factors.When the data from table I are rearranged (table 2) into three. age bands (premenopause, perimenopause, and postmenopause for women) and averaged across countries, the sex ratio appears unexpectedly high at 35-44 years of age during 1970-74 and 1975-77. Within that particular generation of persons around 40 years old in the mid-1970's, the colon cancer death rates have either increased relatively suddenly in males or decreased in females. Changes in actual rates of colon cancer must therefore be examined.