This review article delineates the physiology and methodological principles of somatosensory (SEP) and motor evoked potentials (MEP), as well as our own results in 40 patients monitored during spinal surgery. In 29 patients an intraoperative SEP and in 15 patients a MEP monitoring was performed. Both modalities were applied in 4 patients. 19 patients had an intramedullary tumor, 15 patients had an intradural extramedullary tumor, 4 patients had an extradural mass lesion, and 2 patients had a spinal arteriovenous malformation. Technical problems with SEP monitoring occurred in 3 of 29 cases, problems with MEP monitoring occurred in 4 of 15 cases. Whereas anesthesia showed only little influence on SEP, an appropriate anesthesiological management was of major importance for MEP monitoring. Other factors, e.g. body temperature and blood pressure, also affected the evoked potentials. In all 35 patients in whom intraoperative SEP and/or MEP monitoring was successfully performed, evoked potentials showed a clear correlation with the initial postoperative neurological findings i.e. there were only cases of correct positive or correct negative monitoring. Transient evoked potential changes could always be attributed to surgical maneuvers. Our results show that intraoperative spinal cord monitoring with both SEP and MEP can supply helpful information on neural integrity. The choice of the evoked potential modality to be used and the choice of the sites of stimulation and recording depends on individual pathoanatomical findings and on the operative procedure required. Intraoperative evoked potential monitoring is indispensable during high risk spinal surgery such as surgery for intramedullary tumor or for mass lesions above C5.