Objectives: One of the important aspects of diagnosis in the mixed dentition is the determination of the tooth size arch-length relation ship. Such a determination is often made before eruption of the permanent canines and premolars. The most commonly used prediction method of Tanaka and Johnston is based on data from a sample of children of northern European descent. The purpose of this investigation is to examine the applicability of the Tanaka and Johnston method of prediction in an Iranian population. Methods: Altogether 200 high quality dental plaster casts of Iranian subjects, all of whom were under the age of 21 years, were used. The teeth measured had to be free of restoration, fractures, or proximal caries. The mesiodistal width of a tooth was obtained by measuring with a digital boley gauge. Results: Differences between the predicted widths of the canine and premolars with the Tanaka and Johnston equations and the actual widths were highly significant. Conclusion: The data illustrate the limitations of the Tanaka and Johnston method when applied to a sample population of other than European descent. Two linear regression equations were developed for tooth size prediction in Iranian children: Maxillary: Y=9.3 +0.5(X); Mandibular: Y=8.7 +0.5(X) X= mesiodistal width of the four mandibular incisors in millimeters; Y= mesiodistal width of the canines and premolars in one quadrant in millimeters.