In a review of his experience with thoracic outlet compression, the author attempts to define factors that will help select patients for operation. Fifty-three patients underwent 59 transaxillary resections of the first rib. In almost all cases the diagnosis was clinical, based mainly on the history. Aortography was useful only when intrinsic vascular disease was suspected. Nerve-conduction studies were not helpful and were used only to confirm a diagnosis of peripheral nerve compression. Of 44 patients followed up, 33 had good or excellent results. Of the 11 patients with poor results, 8 complained of persistent neck and shoulder pain. The author concludes that the diagnosis of thoracic outlet compression is clinical and is based mainly on the patient's history. Transaxillary resection is safe and effective, particularly for patients whose symptoms are mainly in the hand and arm. It is not effective for those whose predominant complaint is neck and shoulder pain.