The presence of membrane-enclosed vesicles, 50-100 nm in diameter (cf. Fig. 1), has been observed in the taste pores of rats, mice, and rabbits, although little attention has been devoted to their importance. Murray has noted that fungiform papilla taste pores contained more vesicles than foliate papilla pores. In a recent paper we showed that thaumatin, an intensely sweet, basic protein (pl = 12), binds to the vesicles and to microvilli in taste pores. We suggested that the vesicles were shed from the microvilli as a kind of apocrine secretion, and proposed that the shedding of these vesicles may be an important means by which taste bud cells rid themselves of certain stimulus/receptor complexes, particularly when the stimulus is a large and/or highly charged molecule, such as thaumatin. To investigate this hypothesis further, we used electron microscopy to examine taste pores of both vallate and foliate papillae from Rhesus monkeys, before and after stimulation with thaumatin. We also recorded neural activity from the glossopharyngeal and chorda tympani nerves during stimulation with thaumatin and other tastants.Rhesus monkeys were anesthetized with ketamine and given glycopyrrolate to inhibit salivary secretion. Tongues were thoroughly rinsed and the region of the foliate or vallate papilla treated with thaumatin (33 mg/1) or sucrose (0.3M) for 5-10 min. After a brief rinse, papillae were removed surgically. Control papillae were biopsied with no stimulation. Specimens were fixed for 2 h in: 2% paraformaldehyde, 2% glutaraldehyde in phosphate buffer, pH 7.2, rinsed and post-fixed in phosphate-buffered 1% OsO4,dehydrated in ethanols, and embedded in Epon-Araldite. Thin sections were examined in a JEOL-100 CX electron microscope with particular attention to the contents of the taste pores. For neurophysiology, the glossopharyngeal or chorda tympani nerve was exposed, in anesthetized monkeys, by dissection, and electrodes were placed on the nerve. Impulse activity was recorded with a PAR 113 amplifier, monitored over a loudspeaker and an oscilloscope, and fed into a recorder together with the output from an integrator which indicated the type and time of stimulation. The tongue was stimulated with a system that delivers solutions at programmed intervals under conditions of constant flow and temperature. Each stimulation lasted 10 sec, followed by a 30 or 50 sec rinse before the next stimulus. Stimuli were 0.02M citric acid, 0.1 M NaCl, 0.3M sucrose and 33 mg/l thaumatin.