1 General methods (chronic, subacute and acute) for assessing physical dependence potential, abstinence precipitating capacity and abstinence preventing activity are briefly presented. 2 Direct dependence experiments indicate that, in general, mixed agonist‐antagonist analgesics have relatively lower physical dependence potentials than pure agonist analgesics. That of buprenorphine seems to be particularly low in various animal species. 3 When substitution techniques are used, the dependence potential of buprenorphine seems to be somewhat more developed than with direct dependence techniques. 4 Among various agonists (morphine, methadone and etorphine), antagonists (naloxone, naltrexone and diprenorphine) and mixed agonist‐antagonists (pentazocine, butorphanol and buprenorphine), buprenorphine is the most potent and the longest acting drug in preventing precipitated abstinence in mice, rats and dogs. 5 The low physical dependence potential of buprenorphine may result in part from the very slow dissociation of the complex it forms with opiate receptors. This potential might be underestimated when precipitated abstinence methods are used, as naloxone would displace buprenorphine from its receptors only to a very limited extent. New means of evaluating dependence by more direct means need to be developed. 6 Overall, the properties of mixed agonists in general justify their use as analgesics with lower physical dependence potential than the pure opiates and further, those of buprenorphine seem to indicate its possible utility for the treatment of opiate addiction.