The records of 121 patients hospitalized in Presbyterian-University Hospital, Pittsburgh, during 1976 to 1980 for syncope of unknown origin were reviewed. The were 58 men and 63 women, whose mean age was 63.1 years. Cardiac monitoring in 67 patients showed abnormalities in seven patients, considered diagnostic of the cause of syncope. In 13 patients with electrophysiologic studies, four patients had abnormal results, suggesting a probable cause for the syncope. Cardiac catheterization in 14 patients showed significant findings that demonstrated the cause of syncope in three patients. Glucose tolerance tests in 37 patients, head computed tomographic scans in 39 patients, radionuclide brain scans in 15 patients, lumbar punctures in 22 patients, and skull roentgenograms in 46 patients did not aid in the diagnosis of the cause of syncope in any patient. In 67 patients, EEGs produced abnormal results in 26, but the role of EEGs in the diagnostic workup of syncope could not be completely defined. The definitive cause for syncope was diagnosed in only 13 of 121 patients, with an average hospitalization of nine days and an average cost of $2,463 per patient. These findings suggest that an extensive evaluation of syncope is cost--ineffective and that prospective goal-directed approaches need to be developed.