In general, mammalian taste neurons are broadly responsive to stimuli representing different taste qualities. In the hamster, this breadth of tuning increases systematically from peripheral to successively higher brain stem neurons. Some investigators have classified taste-responsive neurons into "best-stimulus" categories on the basis of which of the four basic stimuli (sucrose, NaCl, HCl, or quinine hydrochloride) elicits the maximum response. However, attempts by others to demonstrate the existence of taste neuron types in the chorda tympani nerve and medulla of the rat using hierarchical cluster analysis have not been successful, resulting in the conclusion that there are no neuron types in the rat gustatory system. The present study was designed to look at the question of neuron types in the hamster, a species with a broader range of gustatory sensitivities to anterior tongue stimulation. Responses of 30 neurons in the nucleus tractus solitarius (NTS) and 31 neurons in the parabrachial nuclei (PbN) of the hamster to an array of 18 stimulus compounds were recorded extracellularly. The similarities of the neural response profiles of these cells at each synaptic level were compared using multivariate statistical techniques. The possiblee grouping of cells on the basis of similarities in their response functions was examined with hierarchical cluster analysis, and the relationships among these response functions were examined with multidimensional scaling. The results of the cluster analysis suggested that at both the NTS and PbN, there are three clusters of neural response profiles. These three clusters of response profiles are characterized at both synaptic levels by their predominant sensitivity to 1) sucrose and other sweet-tasting compounds, 2) sodium salts, and 3) nonsodium salts and acids. Representation of these neurons in a two-dimensional space yielded three nonoverlapping groups of cells in both the NTS and PbN, corresponding to the three groups identified by the hierarchical cluster solution. Classification of taste neurons either by their best stimulus or by other criteria has been criticized on the grounds that it may constitute an arbitrary division of a continuous population of neurons. The techniques of numerical taxonomy, which take the cells' variability into account, also result in a grouping of taste cells into classes. These taxonomic classes agree in most instances (80% in NTS and 80.6% in PbN) to a best-stimulus classification. The failure of some investigators to find types of neural response profiles in the rat gustatory system may be the result of species differences in taste sensitivity as well as differences in the statistical procedures employed.