beta 3-Adrenoceptor (beta 3-AR) agonists were first identified more than twelve years ago and were found to be remarkably effective in animal models of obesity and Type II (non-insulin dependent) diabetes. Those that have been taken forward to clinical studies have not, however, proved so effective in humans: they have either been of limited efficacy, or their activities have been accompanied by significant side-effects. Reasons for the failure of some beta 3-AR agonists in humans have included a poor pharmacokinetic profile and, possibly, a failure of prodrugs to be metabolised to selective beta 3-AR agonists. A more fundamental problem, however, is that the human and rat beta 3-AR differ pharmacologically, and those compounds that have been evaluated in humans have much lower efficacy at the human than the rat receptor. This problem may be compounded by there being a low number of beta 3-AR relative to beta 1-AR and beta 2-AR in those tissues that mediated thermogenesis in humans, so that low efficacy compounds tend to exhibit mainly beta 1-AR or beta 2-AR-, rather than beta 3-AR-mediated effects, despite their having selective affinity for human beta 3-AR. Nevertheless, studies using CGP 12177, which is an agonist at beta 3-AR but an antagonist at beta 1-AR and beta 2-AR, demonstrate that functional beta 3-AR are present in human adipose tissue. Moreover, the association of a polymorphism in the human beta 3-AR with obesity and diabetes demonstrates that this receptor is relevant to these diseases in humans. Thus the true potential of beta 3-AR agonists in humans can only be evaluated when a compound with good selectivity and efficacy at the human beta 3-AR, coupled with a long duration of action in vivo, has been identified.