RESUMO
OBJECTIVES: To compare the predicted tooth width measurements of permanent canine and premolars from Tanaka-Johnston regression equations and Moyers probability tables with the in situ measurements in a sample of Peruvian adolescents. DESIGN: Cross-sectional. SETTING AND SAMPLE POPULATION: Trujillo, Peru; 248 dental casts were measured using a sliding caliper with a Vernier scale rounded to 0.1 mm. RESULTS: Tanaka-Johnston regression equations were not precise, except for the upper arch in the male sample. For females, the Moyers 95th percentile in the upper arch and the 65th percentile in the lower arch predicted the sum precisely. For males, the Moyers 65th percentile for the lower arch predicted the sum precisely, but none of the Moyers percentiles provided precise prediction in the upper arch. CONCLUSIONS: Using tooth width prediction methods from a different racial origin could create an under- or overestimation of the actual combined canine and premolar tooth width, although their clinical significance is disputable.