In order to study the possible explanation of the marked difference in the incidences of breast cancer between Chinese and Australian women, the authors have compared and analysed the results of two case-control studies completed recently in Tianjin, China and Adelaide, Australia. Of 9 potential risk factors studied, 8 are significantly higher in Adelaide (Ad) women than in Tianjin women. Women in Ad were much taller, heavier, more obese, earlier at menarche, later at first full-term pregnancy, more nulliparous, less parity, more in history of breast cancer in first degree relatives. In addition, they were well educated. The findings obtained by analysis of logistic regression model indicated that increased risk for breast cancer was associated with early menarche, late first full-term pregnancy, less parity, nulliparity, histories of benign breast diseases and breast cancer in first degree relatives in the Tianjin study, but not in Ad study. Late menopause and history of oral contraceptive were not associated with the increased risk in both studies. These factors being not associated with breast cancer in Ad women was unexpected. The explanation of the indefinite findings in Ad study was due to the fact that the level of the risk factor is higher, more uniform, and lack of stratum. The difference in the level of the risk factor will not appear among cases and controls in case-control study, and significance of the risk factor therefore may not be found. According to data on diet survey in these two cities, the amount of fat intake was significantly higher in Ad women than in Tianjin women.(ABSTRACT TRUNCATED AT 250 WORDS)