1. Monoarthritis was induced in Lewis rats by interdermal injection in the left hind paw of a suspension of Mycobacterium tubercolusis in mineral oil (500 micrograms 100 microliters-1). Controls were injected with 100 microliters mineral oil. 2. Withdrawal latencies to thermal stimuli of the inflamed paw, the contralateral and both paws of control rats were measured at daily intervals after injection by the plantar test. 3. After detection of the pain threshold, rat spinal cords were removed and horizontal dorsal slices were mounted in a 3-compartment bath to measure electrically-evoked release of substance P-like immunoreactivity (SP-LI). 4. The inflamed paw of monoarthritic rats exhibited a lower pain threshold to thermal stimuli than the contralateral paw of the same animals and both paws of control rats. Inflamed paw hyperalgesia was maximal two days after injection, and declined gradually between 7 to 21 days with no evidence of excitability of withdrawal reflexes after 28 days. 5. During the 28 days study, monoarthritic rats gained less weight than control rats. 6. Electrical stimulation of the dorsal roots attached to rat isolated spinal cord slices induced a significant increase (174 +/- 18% of basal outflow which was 30.3 fmol 8 ml-1, n = 5) in SP-LI release. 7. One-week after induction of inflammation no differences in the amount of SP-LI released from the spinal cord of incomplete Freund's adjuvant-treated rats (IFA) and Freund's adjuvant-treated rats (CFA) were detected. Two weeks after, CFA spinal cord tended to release more SP-LI than IFA cords and, 21 days after injection, the spinal cord of CFA rats released significantly more peptide than IFA rats (17.8 +/- 2.8 fmol ml-1, n = 12 and 6.9 +/- 3.2 fmol ml-1, n = 9, respectively).8. Twenty-one days after treatment, the evoked release from monoarthritic rat spinal cords was increased by 263 + 42% (n = 3) in the presence of the GABAB receptor antagonist, CGP 36742 (100 micro M)which also significantly potentiated monoarthritis-induced hyperalgesia up to 45 min after injection(100 mgkg-1, i.p.).9. These findings may provide a basis for a novel approach to chronic pain therapy but also an explanation for the lack of analgesia produced by the GABAB agonist, baclofen, in chronic as compared to acute pain.