OBJECT: The intradural contributions of the C-4 nerve rootlets have not been previously evaluated for their connections to the brachial plexus. The authors undertook a cadaveric study to evaluate the C-4 contributions to the upper trunk of the brachial plexus. METHODS: The posterior cervical triangles from 60 adult cadavers were dissected. All specimens that were found to have extradural C-4 contributions to the upper trunk of the brachial plexus were excluded from further study. In specimens found to have no extradural C-4 contributions to the brachial plexus a C1-T1 laminectomy was performed. Observations were made of any neural communications between adjacent spinal rootlets, specifically between C-4 and C-5. RESULTS: Nine (15%) of the 60 sides were found to have extradural C-4 contributions to the upper trunk of the brachial plexus. These sides were excluded from further study. No specimen was found to have a postfixed brachial plexus. Of the remaining 51 sides, 11 (21.6%) were found to have intradural neural connections between C-4 and C-5 dorsal rootlets and 1 (1.96%) had a connection between the ventral roots of C-4 and C-5. Communications between these 2 adjacent dorsal cervical cord levels were of 3 types. Type I was a vertical communication between the more horizontally traveling dorsal roots. Type II was a forked communication between adjacent C-4 and C-5 dorsal rootlets. The Type III designation was applied to connections between ventral rootlets. Although communications were slightly more frequent on left sides, this did not reach statistical significance. CONCLUSIONS: In approximately 20% of normally composed brachial plexuses (those with extradural contributions from only C5-T1) we found intradural C4-5 neural connections. Such variations may lead to misinterpretation of spinal levels in pathological conditions of the spinal axis and should be considered in surgical procedures of this region, such as rhizotomy.