The study of dental morphology involves close observation of characteristics of tooth crowns and roots. All observations from which the comparison tables were developed were made on dental casts; however, the test of the method was performed on both casts and real teeth. Observations were made of 29 dental characteristics, with each observation made on all the teeth on which the character was potentially revealed. Both antimeres were scored when present; the higher or more complex of the two scores represented the maximum expression of the trait in the individual. Results are presented as predictions based on logistic regression probability (RP) as well as a method based on Bayes' theorem. Probability analysis was done with dichotomized character states for two, three, and four trait combinations. To use these results, as many as possible of the sight traits considered should be observed in the dentition of an unknown individual. Observation should be scored with the ASU dental anthropology system, with the modification for trigonid crest described in this article, and then dichotomized according to the presented table of dental traits useful for determining African-American or European-American ancestry. 8 tables, 21 references, and an appended worked example