It has been hypothesized that birth size is positively associated with breast cancer risk in adulthood. We studied birth length, birth weight and head circumference at birth and subsequent risk for breast cancer in a cohort of 16,016 women in Norway. Birth length was positively associated with risk (p trend = 002), and women who were 53 cm or longer had a relative risk of 1.8 (CI = 1.2–2.6) compared with women who were shorter than 50 cm, after adjustment for birth year, length of gestation, birth order, maternal age, maternal marital status and socioeconomic status at childbearing. Mutual adjustment for birth weight did not influence the results, and further adjustment for maternal height and adult factors (age at first birth and parity) in a subset of the cohort did not change the results. For birth weight, women in the highest category (≥ 3,840 g) had an adjusted relative risk (RR) of 1.5 (CI = 1.0–2.2) compared to women in the lowest (< 3,040 g), but mutual adjustment for birth length attenuated this association (RR = 1.1; CI = 0.7–1.8). Head circumference at birth showed a similar association as birth weight, with attenuation after mutual adjustment for birth length. The positive association with birth length was stronger among women whose mothers were relatively tall (median or taller, p trend = 0.001) compared to women whose mothers were relatively short (below median, p trend = 0.67) at childbearing. The results provide evidence that intrauterine factors influence future breast cancer risk. The positive association related to birth length suggests that factors that stimulate intrauterine longitudinal growth are particularly important. © 2004 Wiley-Liss, Inc.