Few studies distinguish between the effects of different forms of social capital on health and of those that do none use physical indicators of health as an outcome variable. The objective of this study was to determine whether vertical and horizontal social capital had different associations with dental status of elderly Japanese. In this cross-sectional study, self-administered questionnaires were mailed to elderly individuals living in 25 Japanese communities in 2003. Data from 5560 respondents (49.9%, 72.9+/-6.0 years) included information on social capital, numbers of remaining teeth, health behaviors, physical and mental health, and socioeconomic status. Vertical social capital was defined as participating in groups which encouraged hierarchical relations and horizontal social capital as participating in groups of equals. A multilevel logistic regression analysis was conducted to assess the association between social capital and number of remaining teeth (< or = 19 teeth or > or = 20 teeth). The prevalence of people with 19 or less teeth was 70.7%. Univariate analysis showed significant beneficial associations between individual horizontal social capital and number of remaining teeth. After adjusting for individual- and community-level covariates in multilevel analysis, community-level horizontal social capital still showed beneficial association. Compared to respondents living in highest horizontal social capital areas, those living in lowest horizontal social capital areas had a 1.25 times higher odds ratio for having 19 or less teeth. Individual-level horizontal social capital also showed beneficial effects on number of remaining teeth. Community- and individual-level vertical social capital did not show significant associations with dental status. The results suggest that horizontal social capital, not vertical social capital, has beneficial effects on numbers of remaining teeth in older Japanese adults.