Objective To compare spinal bone mineral density (BMD) and 1-year BMD change between premenopausal vegetarian and nonvegetarian women.Design Cross-sectional comparison of spinal BMD at baseline and prospective comparison of a subsample.Setting A western Canadian metropolitan area.Subjects/samples Healthy vegetarian (n=15 lacto-ovo-vegetarian, n=8 vegan) and nonvegetarian (n=22) women aged 20 to 40 years, with regular menstrual cycles and stable body weight completed baseline measurements. Twenty of these women (6 lacto-ovo-vegetarian, 5 vegan, 9 non-vegetarian) participated in repeat measurements at approximately 13 months.Statistical analyses performed Descriptive statistics, independent sample and paired t tests, 1-way analysis of variance, correlation analysis, and stepwise multiple regression were used to compare groups and to assess associations with BMD.Results At baseline, subjects were 27.2+/-5.1 years old. Vegetarians had lower body mass index (21.1+/-2.3 vs 22.7+/-1.9, P<.05) and percent body fat (24.0+/-5.5% vs 27.4+/-5.1%, P<.05); they also tended to have lower BMD (1.148+/-0.111 g/cm 2 vs 1.216+/-0.132 g/cm 2 , P=.06), although this was not apparent with weight as a covariate (P=.14). Baseline BMD was predicted by vitamin B-12 intake and total body fat (R 2 =.24, P=.001). Participants in the follow-up differed only in their being older than nonparticipants. Over 1 year, mean BMD increased significantly (1.1%): by diet group, non-vegetarians' BMD increased but vegetarians' BMD was unchanged. No other monitored variables were associated with BMD change.Applications/conclusions Vegetarian women should be aware of links between low BMD and low body weight/body fat, and should maintain adequate intakes of nutrients believed to affect BMD. J Am Diet Assoc. 1998;98:760-765.