A cold plate apparatus was designed to test the responses of unrestrained rats to low temperature stimulation of the plantar aspect of the paw. At plate temperatures of 10 degrees C and 5 degrees C, rats with either chronic constriction injury (CCI) of the sciatic nerve or complete Freund's adjuvant (CFA) induced inflammation of the hindpaw displayed a stereotyped behavior. Brisk lifts of the treated hindpaw were recorded, while no evidence of other nociceptive behaviors could be discerned. The most consistent responses were obtained with a plate temperature of 5 degrees C in three 5-min testing periods, separated by 10-min intervals during which the animals were returned to a normal environment. Concomitantly to cold testing, the rats were evaluated for their response to heat (plantar test) and mechanical (von Frey hairs) stimuli. In both injury models, while responses to heat stimuli had normalized at 60 days post-injury, a clear lateralization of responses to cold was observed throughout the entire study period. Systemic lidocaine, clonidine, and morphine suppressed responses to cold in a dose-related fashion. At doses that did not affect motor or sensory behavior, both lidocaine and its quaternary derivative QX-314 similarly reduced paw lifts, suggesting that cold hyperalgesia is in part due to peripheral altered nociceptive processing. Clonidine was more potent in CCI then in CFA rats in reducing the response to cold. Paradoxically, clonidine increased the withdrawal latencies to heat in the CCI hindpaw at 40 days and thereafter, at a time when both hindpaws had the same withdrawal latencies in control animals. Morphine was also more potent on CCI than CFA cold responses, indicating that, chronically, CFA-induced hyperalgesia might be opiate resistant. Evidence for tonic endogenous inhibition of cold hyperalgesia was obtained for CFA rats, when systemic naltrexone significantly increased the number of paw lifts; this was not found in rats with CCI. At 60 days, neither morphine nor naltrexone affected cold-induced paw lifting in CFA rats, suggesting that the neuronal circuit mediating cold hyperalgesia in these animals had become opiate insensitive. In conclusion, the cold plate was found to be a reliable method for detecting abnormal nociceptive behavior even at long intervals after nerve or inflammatory injuries, when responses to other nociceptive stimuli have returned to near normal. The results of pharmacological studies suggest that cold hyperalgesia is in part a consequence of altered sensory processing in the periphery, and that it can be independently modulated by opiate and adrenergic systems.