1. Opioids, though widely used as analgesics, have not been seriously considered as therapy for rheumatoid arthritis. The present study evaluated the dose-effect and time-dependence relationships of a new peripherally selective kappa agonist, asimadoline, in rats with adjuvant arthritis. 2. The arthritis was assessed by a pooled severity index combining the comprehensive criteria of oedema, radiography and histological changes, in the hind limbs. Asimadoline was extremely effective in attenuating joint damage (by up to 80%) when administered parenterally (0.5 to 10 mg kg(-1) day(-1), i.p.) throughout the disease or during its early phase; treatment was less successful if confined to the latter stages. Ten fold higher doses were effective orally. 3. Equimolar doses of a peripherally-selective antagonist, naloxone methiodide, and the kappa-selective antagonist, MR2266, fully reversed the peripheral anti-arthritic effects of asimadoline (5 mg kg(-1) day(-1)), indicating that asimadoline acts through peripheral kappa-opioid receptors. However, an equivalent dose of MR2266 did not fully reverse the anti-arthritic effects of the highest dose of asimadoline (40 mg kg(-1) day(-1)), suggesting a loss of kappa-selectivity at this dose. 4. Asimadoline also exhibited analgesic effects (mechanical nociceptive thresholds) in arthritic but not non-arthritic rats, indicating that inflammation is necessary for asimadoline-induced analgesia. 5. These data confirm our previous findings that kappa-opioids possess anti-arthritic properties and that these effects are mediated via peripheral kappa-receptors. The present results are new in showing that the peripherally acting kappa-opioid agonist, asimadoline, is a potent anti-arthritic agent. Such novel drugs, essentially lacking central side effects, herald new treatments for rheumatoid arthritis.