For most cancer sites there is a linear log-log relationship between incidence and age. This relationship does not hold for breast cancer, and certain 'key' breast cancer risk factors suggest that breast tissue does not 'age' in step with calendar time. A quantitative description of 'breast tissue age' is suggested which brings the age-incidence curve of breast cancer into line with the common log-log cancers and explains quantitatively the known key risk factors. The model also explains the 'anomalous' finding that although early first birth is protective, late first birth carries a higher risk than nulliparity. US breast cancer rates are some four to six times the rates in Japan--the model suggests that the key risk factors, when considered jointly with weight, can explain about 85% of the difference.