IntroductionMorphological and morphometric features of the teeth are of interest to various clinical and academic dental and medical fields including prosthodontics, orthodontics, anatomy and anthropology, pathology, archeology, and forensic dentistry. These have been more or less researched in the case of the permanent dentition. However when it comes to the primary dentition, the literature is scarce and controversial. No study worldwide exists on the cutoff points (thresholds) for sex identification; no study exists on metric or nonmetric traits of deciduous teeth in Iranians. Hence, the aim of the study was to assess both the metric and nonmetric traits of primary molars, as well as their cut-off points for sex identification.MethodsIn this epidemiological cross-sectional study, pretreatment casts of 110 children (51 boys and 59 girls) aged 6 to 12 years were collected. Maxillary and mandibular first and second primary molars were evaluated regarding their metric traits (mesiodistal and buccolingual widths) and 9 nonmetric traits (Accessory cusp on the upper D, Accessory cusp on the lower D, Fifth cusp on the upper E, Carabelli’s cusp on the upper E, Protostylid on the lower E, Fifth cusp on the lower E, Sixth cusp on the lower E, Tuberculum intermedium [metaconulid] on the lower E, and Deflecting wrinkle on the lower E). ROC curves were used to identify cut-off points for sex determination as well as the usefulness of metric measurements for this purpose. Data were analyzed using independent-samples and paired-samples t-tests, McNemar, Fisher, and chi-square tests, plus Pearson and Spearman correlation coefficients (α = 0.05).ResultsAll the primary molars’ coronal dimensions (both mesiodistal and buccolingual) were extremely useful for sex identification (ROC curves, all P values ≤ 0.0000099). Especially, the mandibular primary molars (areas under ROC curves [AUCs] between 85.6 and 90.4%, P values ≤ 0.0000006) were more useful than the maxillary ones (AUCs between 80.4 and 83.1%, P values ≤ 0. 0000099). In the mandible, the first primary molar (maximum AUC = 90.4%) was better than the second molar (maximum AUC = 86.0%). The optimum thresholds for sex determination were reported. Sex dimorphism was significant in buccolingual and mesiodistal crown widths of all the primary molars (all P values ≤ 0.000132), but it was seen only in the case of 2 nonmetric traits: Deflecting wrinkle (P = 0.001) and Tuberculum intermedium (metaconulid, P = 0.029) on the lower Es, taking into account the unilateral and bilateral cases. The occurrence of nonmetric traits was symmetrical between the right and left sides (all P values ≥ 0.250). All mesiodistal and two buccolingual molar measurements were as well symmetrical (P > 0.1); however, two buccolingual measurements were asymmetrical: in the case of the maxillary E (P = 0.0002) and mandibular D (P = 0.019). There were three weak-to-moderate correlations between the nonmetric traits of the mandibular second molars (Spearman correlations between 22.7 and 37.5%, P values ≤ 0.045). Up to 6 concurrent nonmetric traits were observed in the sample, with 53.6% of the sample showing at least 2 concurrent nonmetric traits at the same time, without any sex dimorphism (P = 0.658).ConclusionSex dimorphism exists considerably in primary molars’ sizes, but it is not as prevalent in their nonmetric traits or abnormalities. Primary molars’ crown sizes are useful for sex identification; we calculated optimum cut-off points for this purpose, for the first time.