This study was designed to evaluate whether the H reflex latency obtained from the triceps surae following tibial-nerve stimulation was of value in detecting and differentiating S1 from L5 radiculopathy. Thirty-two patients were studied, 15 with a definite unilateral L5 radiculopathy and 17 with a definite unilateral S1 radiculopathy. Data revealed a mean H reflex latency difference of 0.03 msec (range--0.1 to 0.2 msec) between the affected and unaffected extremities for the 15 patients with L5 radiculopathy and of 2.9 msec (range 1.3 to 4.1 msec) for the 13 patients with S1 radiculopathy where an H reflex could be recorded. The H reflex was unobtainable in the affected extremity in four of the patients with S1 radiculopathy. It is concluded that H reflex latency testing is a valuable tool in helping to differentiate S1 from L5 radiculopathy.