Of 50 normal subjects, 23 (46%) had at least one false positive electrodiagnostic test for carpal tunnel syndrome (CTS). There were 30% of the subjects who exhibited an abnormal median to ulnar sensory amplitude ratio of less than 1.1. In 7 subjects 8 extremities (14%) revealed prolonged residual latencies, and 4 extremities in 4 subjects (8%) had a difference of 0.4 msec between the median and ulnar palmar sensory latencies. The results of this study indicate that certain reported criteria for CTS are abnormal in a high percentage of normal subjects, thereby making them of limited value in the diagnosis of CTS. Of all the criteria studied, it appears that the comparison of the median to ulnar sensory latency across the carpal tunnel is of greatest potential value. However, even here a more conservative difference of 0.5 msec between median and ulnar nerves must be used to avoid false positive tests for CTS.