1. Responses of 199 single olfactory bulb neurones in the frog were observed with odour stimulation.2. To effect control over the concentration, flow rate and time course of the stimulus application, a simple olfactometer was developed and calibrated using a gas chromatograph. The stimulus pulses were monitored by recording the electro-olfactogram from the olfactory mucosa.3. Several categories of unit activity became evident after delivering many odour applications (637). Some odour applications did not cause any change in the spontaneous activity of the unit. This was classified as type N (no response) activity. Some applications caused a suppression of ongoing activity (type S activity). Some applications caused an excitation during the rising concentration phase of the stimulus pulse (type E activity). Three temporal patterns of excitatory responses were seen and these were categorized as subgroups of activity type E.4. Since each response type was seen with about the same frequency regardless of the odour used, this suggested that the temporal patterns of response per se were not coding for odour type.5. When odour concentration was manipulated for a unit showing suppressive (type S) activity at one concentration, then the unit tended to show qualitatively similar suppression at other concentrations as long as they were above threshold.6. When peak concentrations were manipulated for units showing excitatory (type E) activity, the temporal patterns of response could change dramatically. In many cases the units fired after one level of concentration had been reached and were subsequently suppressed with increase in concentration. Such ;concentration tuning' suggested that excitatory temporal patterns, in contrast to suppressive ones, were coding odour concentration information.7. The significance of the changes in temporal response patterns seen here, for experiments examining the relationships between odours using single unit response data, is discussed.