BACKGROUND Upper extremity function in children with brachial plexus birth palsy (BPBP) is assessed with clinical tests such as the Mallet classification, which uses a hand to spine position to assess shoulder internal rotation, or the modified Mallet classification, which adds an additional internal rotation task (hand to belly). Children with BPBP frequently have difficulty performing the hand to spine task. This study compared scapulothoracic and glenohumeral (GH) parameters associated with successful completion of the hand to spine and hand to belly modified Mallet positions. METHODS Motion capture measurement of 32 children with BPBP was performed in hand on spine, internal rotation (hand to belly), hand to mouth, and maximal humerothoracic extension positions. Modified Mallet scores were determined by a hand surgeon. RESULTS Children with better hand to spine performance demonstrated significantly greater GH extension and a nonsignificant trend toward increased GH internal rotation compared with children with scores <3. Children with better internal rotation position performance demonstrated significantly greater GH internal rotation and no significant difference in GH extension. Hand on spine and internal rotation Mallet scores moderately correlated (Pearson r=0.469); however, 54% of children who could place their palms flat on their bellies could not reach behind their backs. CONCLUSIONS Successfully reaching behind one's back requires both internal rotation and extension, representing a multiplanar motion. The hand to belly performance is less affected by extension and should be considered for internal rotation assessment, particularly for children undergoing surgical intervention that may affect internal rotation. METHODS Level II.