The incidence of tourniquet pain was evaluated in 40 patients having orthopedic surgery of the lower extremities during spinal anesthesia using 15 mg of a plain solution of either 0.5% tetracaine or 0.5% bupivacaine. The drugs were administered in a randomized fashion, and measurement of the levels of sensory anesthesia to pinprick and motor blockade as well as the occurrence of tourniquet pain were made by an independent blinded observer. The onset and maximum cephalad spread of sensory anesthesia and the onset and degree of motor block were similar in both groups of patients. However, the duration of sensory anesthesia was significantly longer in patients in whom tetracaine was used. The incidence of tourniquet pain was significantly greater in patients given tetracaine (60%) than in patients given bupivacaine (25%). The occurrence of tourniquet pain was not related to the level of sensory anesthesia, because patients in the tetracaine group had a higher level of sensory anesthesia (mean T6) than did patients in the bupivacaine group (mean T10) at the onset of tourniquet pain. It is speculated that during spinal anesthesia both A and C fibers (mediating fast and slow pain, respectively) are initially equally inhibited. However, as the concentration of local anesthetic in the cerebrospinal fluid declines, C fibers may become unblocked earlier with tetracaine than A fibers, resulting in tourniquet pain in the presence of an otherwise satisfactory spinal anesthetic.