To determine whether the idiosyncratic distribution of transduction mechanisms for bitter tastants in rat taste receptor cells (TRCs) could be inferred from the neural activity they evoke, single neuron responses to ten bitter-tasting compounds were recorded from rat glossopharyngeal (n = 30) and chorda tympani (n = 22) neurons. Responses to several 'bitter' alkaloids were obtained: 10 mM quinine-HCl, 50 mM caffeine, and 1 mM each nicotine, yohimbine, and strychnine, plus a number of non-alkaloid bitter-tasting compounds: 0.1 M KCl, 0.01 M MgCl2, and 1 mM each phenylthiocarbamide (PTC), L-tyrosine, and denatonium benzoate. To obtain some distinctions with other stimuli NaCl (0.1 M), HCl (pH 2.0), and capsaicin (10 microM) were also tested. It was found that individual neurons in both glossopharyngeal and chorda tympani nerves differed in their relative sensitivities to the various bitter stimuli. To determine relationships among these stimuli, the differences in the evoked responses between each stimulus pair were summarized in a multi-dimensional scaling space. In these analyses neither nerve showed any obvious similarity between the placements of quinine and the other bitter stimuli. Such data suggest that first-order gustatory neurons can discriminate among the above bitter stimuli. For glossopharyngeal neurons, some similarity to quinine was found only for nicotine and denatonium, and for chorda tympani neurons, some similarity to quinine was found only for KCl and MgCl2. Of the bitter compounds tested, quinine evoked the greatest response from glossopharyngeal neurons. We propose this arises because quinine can activate TRCs by more transduction mechanisms than other bitter stimuli. The results from these studies were summarized in a qualitative model for the coding of bitter tastants where the variety of transduction mechanisms for bitters are distributed among various TRCs to account for the heterogeneous responses among the neurons.