To determine which sensory nerve conduction studies (S-NCS) are helpful in detecting supraclavicular axon loss brachial plexopathies, we selected 53 cases (of 417 reviewed) in whom complicating factors were absent and which, by needle electrode examination findings, involved only a single "truncal" element (upper, middle, or lower) of the brachial plexus. Extensive S-NCS included: median, recording thumb (Med-D1), index (Med-D2), and middle fingers (Med-D3); ulnar, recording fifth finger (Uln-D5); dorsal ulnar cutaneous, recording dorsum of the hand (DUC); radial, recording base of thumb; and both medial and lateral antebrachial cutaneous (MABC, LABC), recording forearm. Except for the median sensory fibers, the "cord" elements traversed by the sensory fibers assessed during the S-NCS listed above are anatomically defined (i.e., the sensory fibers enter the brachial plexus at only one cord). In regard to the median sensory fibers, however, there are two possible pathways through the infraclavicular plexus: (1) the lateral cord and/or (2) the medial cord. Because the lower trunk is only accessible via the medial cord, any sensory fibers found to be traversing the lower trunk had to first traverse the medial cord. Similarly, those traversing the upper and middle trunks must first be a component of the lateral cord. The frequency that the various S-NCS responses were abnormal (unelicitable, below laboratory normal value, or < or = 50% of the contralateral response) for a given brachial plexus element lesion was as follows: (1) upper trunk (UT): 25 of 26 Med-D1, 25 of 26 LABC, 15 of 26 radial, 5 of 26 Med-D2, 2 of 26 Med-D3; (2) middle trunk (MT): 1 of 1 Med-D3; (3) lower trunk (LT): 25 of 26 Uln-D5, 22 of 23 DUC, 11 of 17 MABC, 3 of 23 Med-D3. With lower trunk brachial plexopathies, both "routine" (Uln-D5) and "uncommon" (DUC; MABC) S-NCS are abnormal. With upper trunk brachial plexopathies, in contrast, only the "uncommon" S-NCS (Med-D1; LABC) are consistently affected. The "routine" median S-NCS recording digit 2 (Med-D2) is far less reliable than the median S-NCS recording digit 1 (Med-D1) in detecting upper trunk axon loss brachial plexopathies. Additionally, the various pathways traversed by the fibers contributing to the individual S-NCS responses can be predicted, an important point when the full extent of a brachial plexus lesion is sought.