The differential levels of sensory blockade of pinprick, cold, and touch were monitored throughout the course of spinal anesthesia administered to 50 patients to determine variations in the degree of spread during onset, plateau, and regression, and to establish the effects of epinephrine and the effect of posture during injection. A significant difference was observed between the dermatomal level of sensory loss of touch and the dermatomal level of loss of either pinprick or cold during onset, at 5 min in patients given tetracaine with epinephrine, at time of maximum spread in patients given tetracaine with epinephrine or in the sitting position, and in all groups during regression. Loss of touch began later, never extended as far cephalad, and regressed sooner. The extent of this difference was greatest during regression, when the anesthetic was given to patients in the sitting position, after epinephrine. The level at which the sense of touch was lost seemed to mark the limits of the zone of solid spinal anesthesia; these limits could not be assessed effectively using pinprick. We propose that loss of touch sensation be used to assess whether anesthesia is adequate to avoid tourniquet pain. If there is loss of touch sensation above the L1 dermatome, it is unlikely that tourniquet pain will occur.