Objectives: To (a) describe the incidence of use of specific dental services; (b) test the hypothesis that certain predisposing, enabling, and need (PEN) factors are differentially predictive of service use; and (c) test the hypothesis that even with other PEN factors taken into account, race and household income are differentially predictive of certain dental services. Previously, this study identified PEN factors that predicted use of any care; herein we identify whether these same factors were differentially predictive of specific service use among users of at least one service. Methods: The Florida Dental Care Study was a longitudinal study of persons aged 45 years or older who had at least one tooth. Subjects participated for interviews and clinical examinations at baseline and 24 months later, with 6-monthly telephone interviews between those times. Results: Seventy-seven percent of subjects reported one or more visits. Results from a single multivariate multiple logistic regression suggested that even once analysis was limited to persons who used at least one dental service, at least one measure from each of the PEN domains was predictive of specific dental service use. Conclusions: Each PEN domain was predictive of service use, even once limited to persons with at least one visit. Even with differences in other PEN variables taken into account, African-Americans were much less likely to receive dental cleanings, restorative dentistry and fixed prosthodontic services, and were much more likely to have a tooth extracted. Household income was predictive of receipt of fixed prosthodontic services, but not other service categories.