In an effort to assess the relative importance of age at first birth, age at subsequent births, and total parity to the occurrence of breast cancer, reproductive data from 4,225 women with breast cancer and 12,307 hospitalized women without breast cancer were analyzed by a multiple logistic regression model. Age at first birth was confirmed to be the most important reproductive risk indicator; it was associated with a 3.5% increase of relative risk for every year of increase in age at first birth (the 95% confidence interval of this estimate was 2.3 to 4.7% increase per year). However, age at any birth after the first was also an independent and statistically significant risk indicator; it was associated with a 0.9% increase of relative risk for every year of increase in age at any (and every) birth (the 95% confidence interval of this estimate was 0.4 to 1.5% increase per year). There is evidence that the age of approximately 35 years represents for every birth a critical point; before this age any full-term pregnancy confers some degree of protection; after this age any full-term pregnancy appears to be associated with increase in breast cancer risk. The effect of parity is determined by the age of occurrence of the component pregnancies. While most pregnancies occur under the age of 35, the distribution varies from population to population, and this may account for the differences between populations in whether or not a protective effect is seen for births after the first, and if it is seen, its extent.