Unlike past breast cancer survival comparisons between Japan and the United States, a recent study in Hawaii showed that Japanese women did not retain their survival advantage over Caucasian women after adjustment was made for stage at diagnosis. To test whether this finding in Hawaii was due to the limited duration of the follow-up (five years) or to the effects of migration, the survival experience of 1,357 Caucasian and 1,029 Japanese women with invasive breast carcinoma diagnosed in Hawaii between 1960 and 1979 was examined over a 10-year period as well as by place of birth. Multivariate adjustment by the proportional hazards regression model confirmed that the survival advantage of Japanese women in Hawaii is fully explained by their earlier stage of disease at diagnosis and suggested that, after recognition, the disease progresses at a similar pace in the two races. The survival comparison by place of birth revealed that second generation Hawaii Japanese women had better breast cancer survival rates than Japanese migrants from Japan, even after adjusting for stage, and that for Caucasian women, nativity was not associated with survival. These findings suggest that westernization, genetic constitution, or early life exposures cannot explain the overall or stage-adjusted breast cancer survival patterns observed among Caucasian and Japanese women in Hawaii.