Several methods have been described to estimate inter-arch tooth size relationship, such as Bolton?s ratios. The aims of this study were to verify the validity of Bolton indexes in a sample of untreated Portuguese subjects based on Angle classification and to evaluate the gender difference. 168 pre-treatment dental casts of orthodontics Portuguese subjects (59 males and 109 females) with different occlusions were used, which were selected randomly from 541 consecutively treated orthodontic patients. The mesiodistal widths from first molar to first molar were measured on each pre-treatment cast to the nearest 0.01 mm using digital caliper, and Bolton?s anterior and overall ratios were calculated. Descriptive statistics as mean, standard deviation and range were calculated. Moreover, the results were compared to Bolton?s ratios and differences based on gender and occlusion groups were evaluated by statistical inference methods. The results reveal that the mean values, standard deviation and range were larger than Bolton?s in normal occlusion group (78.3±3.5% in anterior ratio and 92.1±2.2% in overall ratio) and there were no differences between genders (p >0.05). Class I (anterior and overall ratios, p=0.001 and p<0.001, respectively), Class II/2 (anterior ratio, p=0.032) and Class III (overall ratio, p=0.041) were significantly different from Bolton?s reference data. The results showed no differences between gender and no difference between normal occlusion and malocclusion groups. Moreover, in normal occlusion group, the anterior and overall tooth size ratios was equivalent to the original Bolton?s ratios, although the mean and standard deviation were large.