For many years, bromocriptine has proven to be a useful treatment for some of the disabling motor effects seen in Parkinson's disease. As such, it has been the only commonly used directly acting D2 agonist available. But its mechanism of action has been obscure because many animal models indicated an absolute requirement for the presence of endogenous DA for bromocriptine to have any efficacy, despite its undoubted occupation of the D2 receptor with high affinity. Several scattered reports indicated, however, that bromocriptine could potentiate the effects of a number of other dopamine agonists (such as apomorphine and L-dopa) in a variety of pharmacological models and in the clinic. With the availability of SKF38393 and SCH23390, it soon became clear that bromocriptine, while a selective D2 agonist, depended in an absolute sense on the integrity of the D1 receptors. Thus, if SKF38393 was administered together with bromocriptine to rodents depleted of dopamine, marked locomotor excitation was produced, despite either drug alone being inactive. The present review explores the literature on the motor effects of bromocriptine and endeavours to integrate its behavioural, biochemical and electrophysiological effects into a coherent whole. It closes with a consideration of several remaining unsolved problems associated with the pharmacology of bromocriptine and suggests some future studies.