The sense of taste in humans differs substantially from that of rodents, from which a preponderance of gustatory electrophysiology derives. To establish a more appropriate neural model for human gustation, we recorded the activity of single neurons in the primary taste cortex in 11 alert cynomolgus macaques. Taste cells composed 6% of all neurons encountered. Another 24% responded during mouth and jaw movements, and 4% were sensitive to tactile stimulation of the mouth. Smaller numbers responded during olfactory or visual stimulation, or when the monkey extended his tongue. Taste cells could be divided into four statistically independent groups, corresponding to those most responsive to glucose (38%), NaCl (34%), quinine (22%), or HCI (5%). The location of a taste cell did not predict its response profile, i.e., there was no clear topographic organization of taste sensitivity. We established neural thresholds and intensity-response functions to the basic stimuli and determined that-with the exception of HCl, to which the macaque is relatively insensitive-they were similar to those reported by human subjects. We then turned to the coding of taste quality, as inferred in macaques from the patterns of neural activity elicited by each of greater than 100 stimuli. The results proved generally faithful to human reports of the perceived qualities of these same tastants. Finally, an investigation of taste mixtures revealed a degree of mixture suppression and interaction among basic qualities similar to those reported by humans. We conclude that the alert macaque offers a reliable neural model for human gustation.