Locomotor stepping can be elicited by brain stimulation at various diencephalic sites under moderate levels of Nembutal. This study determined if locomotor initiation measured under anesthesia provides a valid measure of the intersite factors which determine initiation in the awake condition. We compared the latencies to initiate locomotor stepping elicited by electrical stimulation (50 microA, 0.5-msec pulses, 10 to 160 Hz) by rats tested while awake and unrestrained in a rotary runway or anesthetized and held in a stereotaxic apparatus. In the latter tests, initial anesthesia was provided by Nembutal (25 mg/kg) and 2% halothane and maintenance anesthesia was provided by 7 mg/kg as needed and local injections of lidocaine. For 30 sites in 16 rats, average locomotor initiation latency in the awake condition and the shortest latencies in the anesthetized condition were positively correlated (r = .78). Locomotion at sites with long latencies in the awake condition was frequently blocked in the anesthetized condition, but sites with short latencies were rarely blocked. The results indicate that the shortest locomotor latencies in the anesthetized condition approximate the latencies measured in the awake condition. It is concluded that the anesthetized condition can provide valid initiation measures, but sites with long latencies in the awake condition are prone to depression under anesthesia.