The relationship between gynecological history and bone mineral density (BMD) of the lumbar spine and femoral neck was studied in 3126 perimenopausal women. The study population was a random, stratified sample of participants, selected from the Kuopio Osteoporosis Study, which consisted primarily of all 14,220 women aged 47-56 years in Kuopio Province in 1989. After exclusion of 1521 women reporting past or present hormonal replacement therapy (HRT), 1605 women formed the final study population. Present HRT users had significantly higher lumbar BMD but not femoral BMD, than non-hormone users. Postmenopausal status, late menarche, and bilateral oophorectomy were risk factors for low BMD. Protective factors against low BMD were increased body weight, increased number of pregnancies, as well as hysterectomy without bilateral oophorectomy. The majority (43.8%) of these operations had been performed due to the presence of leiomyomas. No significant correlation was found between nulliparity, breast-feeding or amenorrhea before the age of 30 and BMD. In the multiple regression analysis, gynecological variables could account for only 18.4-26.8% of the variance in BMD, while time since last periods, age, age at menarche, weight and hysterectomy were the most significant variables. We conclude that reproductive history gives rise to some special risk groups, to whom BMD measurements and osteoporosis prevention efforts should be directed. However, it is impossible to predict BMD by gynecological characteristics.