Intrathecal injection of dynorphin produced a loss of the tail-flick reflex that lasted throughout the 14-day experimental period whereas, the inclined plane test of motor function and tail-shock vocalization recovered within an hour. An important aspect of the loss of the tail-flick reflex was that it was an all-or-none event. At any dose of tail-flick latency either remained unchanged when compared with pre-injection latencies or the latency was elevated to the cut off time of 14 s. The ED50's +/- S.E.M. for tail-flick, inclined plane and tail-shock vocalization were 65.4 +/- 5.0, 67.7 +/- 5.0 and 68.0 +/- 3.9 nmol respectively. Results from the hot-plate test revealed no statistical difference between saline and dynorphin injected animals one day following the injection. Animals injected with morphine sulphate s.c. lost the tail-flick reflex but completely recovered by 24 h. Histology of the spinal cord of animals treated with dynorphin 24 h prior to sacrifice revealed dead neurons primarily in the ventral horn with little or no damage in the dorsal horn. These data demonstrate that dynorphin(1-13) injected intrathecally results in a rather specific neurotoxic action in the spinal cord.