1. Intracellular and extracellular recordings were made from rat olfactory bulb mitral and tufted cells during odor stimulation and during electrical stimulation of the olfactory nerve. Neurons were identified by horseradish peroxidase injections and/or antidromic activation. The presentation of multiple concentrations of at least one odorant in a cyclic artificial sniff paradigm, as reported previously (10), allowed the study of odor responses. This approach was extended to multiple odorants to compare their concentration-response profiles. This procedure avoids the problems of interpretation resulting from nonequivalence of the effective concentrations of different odorants used as stimuli that have characterized previous studies of odor quality effects. Comparisons of intracellular events and responses to electrical stimulation with the odor-induced spike train activity allow us to begin to delineate the local circuitry involved in generating odor-induced responses. 2. The concentration-response profiles of the 72 cells in the present study are comparable to those previously reported for output neurons of the olfactory bulb, showing ordered changes in the temporal patterning of spike activity with step changes in odor concentration. However, eight of the neurons exhibited inhibitory responses to lower concentrations, but excitation, at similar latency, to higher concentrations of the same odorant. These data emphasize that to study pattern changes induced by changing odor quality the influence of stimulus intensity must also be carefully examined. The data also provide evidence that the temporal pattern evoked by an odorant is probably not in itself the code for odor quality recognition. 3. Complete concentration-response profiles, including subthreshold concentrations, to more than one odorant show that, although responses to the different odorant can evolve systematically with concentration, the responses to different odorants can evolve through very different patterns. For example, in some cells, the response patterns to different odors were complementary in form. These results demonstrate that the patterned responses of olfactory bulb neurons can reflect changes in odor quality as well as intensity. 4. Intracellular recording was employed to compare the temporal patterning of spikes during odor stimulation with membrane potential changes. In some cases, the spike pattern was closely correlated with apparent postsynaptic potentials. However, there were several clear exceptions. In five cells, a prominent hyperpolarization, seen in the first sniff of a series of 10 consecutive sniffs, was associated with pauses in spike activity. In the following