While the evidence is strong that dopamine plays some fundamental and special role in the rewarding effects of brain stimulation, psychomotor stimulants, opiates, and food, the exact nature of that role is not clear. One thing is clear: Dopamine is not the only reward transmitter, and dopaminergic neurons are not the final common path for all rewards. Dopamine antagonists and lesions of the dopamine systems appear to spare the rewarding effects of nucleus accumbens and frontal cortex brain stimulation (Simon et al 1979) and certainly spare the rewarding effects of apomorphine (Roberts & Vickers 1988). It is clear that reward circuitry is multisynaptic, and since dopamine cells do not send axons to each other or receive axons from each other, dopamine can at best serve as but a single link in this circuitry. If dopamine is not a final common path for all rewards, could it be an intermediate common path for most rewards? Some workers have argued against such a view, but at present they must do so on incomplete evidence. For example, Phillips (1984) has argued that there must be multiple reward systems, functionally independent and organized in parallel with one another. His primary evidence, however, is the fact that brain stimulation is rewarding at different levels of the nervous system. As we have seen in the case of midline mesencephalic stimulation, the location of the electrode tip in relation to the dopamine cells and fibers tells us little about the role of dopamine in brain stimulation reward. It seems clear that the ventral tegmental dopamine system plays a critical role in midline mesencephalic reward, despite the distance from the electrode tip to the dopamine cells where morphine causes its dopamine-dependent facilitory effects or to the dopamine terminals where low-dose neuroleptics presumably cause theirs. Until pharmacological challenge has been extended to the cases discussed by Phillips, we can only speculate as to the role of dopamine in each of those cases. In the cases where pharmacological challenge has been examined, only nucleus accumbens and frontal cortex have been found to have dopamine-independent reward sites. It is not consistent with the dopamine hypothesis that dopamine-independent reward sites should exist in these areas, since any reward signals carried to nucleus accumbens or frontal cortex by dopamine fibers would-unless we are to believe that reward "happens" at these sites-have to be carried to the next stage of the circuit by nondopaminergic fibers (there are no dopaminergic cell bodies in any of the dopamine terminal areas).(ABSTRACT TRUNCATED AT 400 WORDS)