The estrogen receptor (ER) gene polymorphisms of PvuⅡ and XbaⅠ restriction were examined in parallel with bone mineral density of the lumbar vertebrae (BMD) and bone metabolic markers, such as serum bone Gla protein (BGP), serum bone alkaline phosphatase (B-ALP) and urinary hydroxyproline (Hyp), in 178 healthy Japanese women, 30 of whom were premenopausal and 148 postmenopausal. The allele was defined as capital P or X, when the ER gene was not digested with Pvu Ⅱ or Xba Ⅰ ; otherwise, it was defined as small letter p or x. The distribution of the Pvu Ⅱ and Xba Ⅰ restriction fragment length polymorphism (RFLP) was as follows:PP, 32 (16.9%):Pp, 96 (50.8%); pp. 61 (32.3%) and XX, 5 (2.6%);Xx, 74 (39.2%);and xx, 110 (58.2%);respectively. The allele frequencies were 42.3% for P, 57.7% for p, 22.2% for X and 77.8% for x. In PX haplotype distribution, 5 (2.6%), 19 (10.1%), 8 (4.2%), 51 (27.0%), 45 (23.8%), 4 (2.1%) and 57 (30.2%) of the subjects had PPXX, PPXx, PPxx, PpXx, Ppxx, ppXx and ppxx types, respectively. Neither PpXX nor ppXX haplotype was found in the study subjects. Pvu Ⅱ genotypes had no correlation with either BMD or biochemical markers for the bone metabolism. Although the subjects with heterozygous Xx genotype had higher BMD than those with homozygous XX or xx, Xba Ⅰ genotypes did not have any statisti- cally significant correlation with the bone metabolic markers. These results suggest that ER gene polymorphisms of Pvu Ⅱ and Xba Ⅰ restriction may not have any effect on either osteogenetic or osteoclastic steps of the bone metabolism, at least in Japanese middle-and higher-aged women.