The rat intraperitoneal infusion procedure was used to chronically administer drugs for evaluation of the physical dependence liability of narcotic antagonist analgesics. Three methods were used to assess dependence liability: presence of withdrawal signs upon abrupt cessation of chronic infusion (primary dependence), attenuation of the withdrawal signs produced by cessation of chronic morphine infusion (morphine substitution), and production of withdrawal signs when chronically morphine-fused rats were administered the drugs (precipitated withdrawal). Butorphanol, nalbuphine and pentazocine all caused a mild withdrawal in the rat primary dependence model which agrees with the conclusions from experiments with monkey and man. None of these agents substituted for morphine in the rat and three appeared to precipitate withdrawal. Two experimental drugs, Codorphone and TR5400, did not induce primary dependence in the rat, and in chroni-morphinized rats, they precipitated a withdrawal syndrome comparable to naloxone. Another experimental drug, TR5257, substituted for morphine. The correlation between these observations in the rat and previously published data from the monkey are excellent. It is proposed that the rat could be used as a reliable indicator of potential physical dependence liability for the narcotic antagonist analgesics.