The presence of the PvuII or the XbaI polymorphism in the estrogen receptor alpha gene (ESR1, 6q25) has been related to breast cancer risk; however, results are not fully consistent. To further elucidate this relation, we examined these polymorphisms in relation with mammographic density, a measure of dense tissue in the breast, which is strongly associated with breast cancer risk. For this study, 620 participants aged 49 to 68 years were selected from the Prospect-European Prospective Investigation into Cancer and Nutrition cohort. Blood samples, lifestyle- and medical questionnaire data and mammograms were available for these women. Genotyping was done using the TaqMan PCR assay and mammographic density was assessed using a computer-assisted method. Means of mammographic density were compared by ESR1 genotypes and haplotypes. The percentage density was higher in women with one or two copies of the PvuII p allele (means for Pp and pp are 37% and 36%, respectively) than in those with the PP genotype (32%, P(trend) = 0.09). Women with one or two copies of the XbaI x allele had higher mean percentage density (Xx and xx, 36% and 37%, respectively) than those with the XX genotype (31%, P(trend) < 0.01). Haplotype 1 (px) was associated with increased density, whereas haplotype 2 (PX) was associated with decreased density, both suggesting an allele-dose effect (P(trend) = 0.08 and <0.01, respectively). Similar associations were found with absolute density (P(trend) < 0.01). The findings of this study support the view that ESR1 polymorphisms may affect breast cancer risk through differences in breast density.