Although there are recognized risk factors for breast cancer, its cause is still unknown. It is hypothesized here that breast cancer results from late exposure to a common virus. This hypothesis is investigated by relating the epidemiology of breast cancer to the seroepidemiology of cytomegalovirus, as a surrogate for a breast cancer virus. The hypothesis is consistent with the geographical distribution of breast cancer; a correlation between breast cancer incidence and the percentage of adults who are cytomegalovirus seropositive in various countries was found (Pearson correlation coefficient -0.79). The hypothesis is also consistent with other risk factors for breast cancer, such as age at onset, family history, hormonal factors and migration.