PURPOSE: Three commonly used classifications for thumb polydactyly are the Wassel-Flatt, Rotterdam, and Chung. The ideal classification system would have high validity and reliability and be descriptive of the thumb anomaly. The purposes of this investigation were to (1) compare the inter- and intrarater reliability of these 3 classifications when applied to a large sample of patients enrolled in the Congenital Upper Limb Differences (CoULD) Registry and (2) determine the prevalence of radial polydactyly types when using the various classifications in a North American population. METHODS: Inter- and intrarater reliability were determined using 150 cases of radial polydactyly presented in a Web-based format to 7 raters in 3 rounds, a preliminary training round and 2 observation rounds. Raters classified each case according to the Wassel-Flatt, Rotterdam, and Chung classifications. Inter- and intrarater reliability were evaluated with the intraclass correlation coefficient (ICC) calculated using 2-way random measures with perfect agreement. RESULTS: For Wassel-Flatt, both the interrater (ICC, 0.93) and the intrarater reliability (ICC, 0.91) were excellent. The Rotterdam classification had excellent reliability for both interrater reliability (ICC, 0.98) and intrarater reliability (ICC, 0.94), when considering type alone. Interrater analysis of the additional subtypes demonstrated a wide range of reliabilities. The Chung classification had good interrater (ICC, 0.88) and intrarater reliability (ICC, 0.77). Within the Wassel-Flatt classification, the most frequent unclassifiable thumb was a type IV hypoplastic thumb as classified by the Rotterdam classification. CONCLUSIONS: The Wassel-Flatt and Rotterdam classifications for radial polydactyly have excellent inter- and intrarater reliability. Despite its simplicity, the Chung classification was less reliable in comparison. The Chung and Rotterdam classification systems capture the hypoplastic subtypes that are unclassifiable in the Wassel-Flatt system. Addition of the hypoplastic subtype to the Wassel-Flatt classification (eg, Wassel-Flatt type IVh) would maintain the highest reliability and classify over 90% of thumbs deemed unclassifiable in the Wassel-Flatt system. CLINICAL RELEVANCE: The Wassel-Flatt and Rotterdam classifications have excellent inter-and intrarater reliability for the hand surgeon treating thumb polydactyly. Addition of a hypoplastic subtype to the Wassel-Flatt (Type 4h) allows classification of most previously unclassifiable thumbs.