The validity of the somatosensory evoked potential for serial neurological evaluation requires an understanding of normal variations in the responses Studies were conducted in 15 normal adult volunteers during two nights of sleep and in five normal adult volunteers during waking hours. The response amplitudes and latencies obtained during the day were demonstrated to be stationary by use of the runs or sign test Similarly, the earliest portions of the scalp recorded evoked potential were stationary during sleep Later portions of the responses had significantly lower amplitudes and demonstrated increased latency during deep sleep. The stability of the early portions of the SSEP suggest that it is a reliable tool for serial neurological evaluation.