In a previous prospective study we showed elevated risks for breast cancer in nulliparous women compared to parous women, in those having their first pregnancy at a higher age, and those with few children. This was based on 216 women diagnosed with breast cancer during 1965 to 1975 among 34,525 women having attended the cervix cancer detection clinic in Iceland by the end of 1974, and born between 1906 and 1945. The present investigation on 848 cases, diagnosed among 61,040 women attending the cervix cancer detection clinic during 1964 to 1984 and born between 1901 and 1960, shows the same risk factors to be significant. The relative risks are, however, smaller. The reasons for the difference in relative risks are discussed. We find that the effect of age at first birth is significant for women up to the age of 65 and not for older women. In both cohorts, women older than 55 are underrepresented and more so in the earlier report. In addition, the small number of cases in the reference group with age at first birth below 20 appears to have made the figures of our earlier report unreliable.