AbstractWe undertook a case‐control study regarding the excretions of 14 urinary steroids in gastric cancer (GC) patients. The results are as follows: the levels of androgens, progestins, and two corticosteroids were, relative to tetrahydrocortisol, significantly depressed in GC patients of both sexes compared with the corresponding normal controls. The deviation profile of urinary steroids was not affected by radical gastrectomy. Evidence indicated that observed changes of GC urines were the steroidal expression of a decrease of endogenous testosterone combined with an increase of endogenous hydrocortisone; there was also evidence that the hormonal environment of our GC patients was endocrinologically homologous to that of rice‐fed or salty rice‐fed mice.Epidemiological inquiry revealed that GC patients having more access to rice‐rich or salt‐rich diets were taller and less obese than were rural healthy controls. In agreement with the anthropometric data of those cancer patients is the finding that the specific death rate of GC (as calculated for each of 15 prefectures of northern Japan) was positively correlated for each sex with the mean heights, but not with the mean weights, of 14‐year‐old youths of those areas. This paper discusses the possible relevancy of the hormonal and epidemiological aspects of GC patients to gastrocarcinogenesis in light of steroid physiology.