The search for the physiological function of nicotinic receptors on neurons in the brain began with their discovery. It was initially assumed that, as in ganglia and at the neuromuscular junction, nicotinic receptors would gate fast synaptic transmission in the brain. The best functional evidence now, however, points to a role in modifying the release of other transmitters. This does not preclude a postsynaptic role in transmission for nicotinic receptors in the brain, but attempts to locate such a synapse have not been successful. If fast nicotinic synapses are present in the brain, they are probably low in number and may be masked by other more prevalent synapses (such as glutamatergic) so identification will not be easy. The extent of diversity of nicotinic receptors is substantial. At the molecular level this is reflected in the number of different genes that encode receptor subunits and the multiple possible combinations of subunits that function in expression systems. From the cellular level there is a broad diversity of properties of native receptors in neurons. Some useful pharmacological tools allow the limited identification of subunits in native receptors. For example, block by alpha-bungarotoxin identifies alpha 7, alpha 8, or alpha 9 subunits; activation of a receptor by cytisine indicates an alpha 7 or beta 4 subunit; and neuronal bungarotoxin block identifies a beta 2 subunit. Despite the clues to identity gained by careful use of these agents, we have not been able to identify all the components of any native receptor based on pharmacological properties assessed from expression studies. When both pharmacological and biophysical properties of a receptor are taken into consideration, none of the combinations tested in oocytes mimics native receptors exactly. The reason for this discrepancy has been debated at length; it is possible that oocytes do not faithfully manufacture neuronal nicotinic receptors. For example, they may not correctly modify the protein after translation or they may allow a combination of subunits that do not occur in vivo. Another possibility is that correct combinations of subunits have not yet been tested in oocytes. Data from immunoprecipitation experiments suggest that many receptors contain three or more different subunits. Results from further experiments injecting combinations of three or more subunits into oocytes may be enlightening. The diversity of receptors may allow targeting of subtypes to specific locations. Nicotinic receptors are located presynaptically, preterminally, and on the cell soma. The function of the nicotinic receptors located on innervating axons is presumably to modify the release of other neurotransmitters. It is an attractive hypothesis that nicotinic receptors might be involved in modifying the weight of central synapses; however, in none of the regions where this phenomenon has been described is there any evidence for axoaxonal contacts. The presynaptic receptors described so far are pharmacologically unique; therefore, if there are different subtypes of nicotinic receptors modifying the release of different transmitters, they may provide a means of exogenously modifying the release of a particular transmitter with drugs. There are still many basic unanswered questions about nicotinic receptors in the brain. What are the compositions of native nicotinic receptors? What is their purpose on neurons? Although there is clearly a role presynaptically, what is the function of those located on the soma? Neuronal nicotinic receptors are highly permeable to calcium, unlike muscle nicotinic receptors, and this may have important implications for roles in synaptic plasticity and development. Finally, why is there such diversity? (ABSTRACT TRANCATED)