We have used the autoradiographic method to locate trigeminal nerve endings in monkey teeth. The nerve endings were labeled in two adult female Macaca fascicularis by 20 hours of axonal transport of radioactive protein (3H-L-proline). We found a few labeled axons in contralateral mandibular central incisors and one mandibular canine. In ipsilateral teeth, numerous myelinated and unmyelinated axons were labeled; they formed a few terminal branches in the roots but primarily branched in the crown to form the peripheral plexus of Raschkow and to terminate as free endings in the odontoblast layer, predentin, and as far as 120 micrometers into dentinal tubules. Electron microscopic autoradiography showed that the radioactive axonally transported protein was confined to sensory axons and endings; odontoblasts and dentin matrix were not significantly labeled. Labeled free nerve endings were closely apposed to odontoblasts in dentin but did not form distinctive junctions with them. Nerve endings were most numerous in the regular tubular dentin of the crown adjacent to the tip of the pulp horn, occurring in at least half of the dentinal tubules there. Reparative dentin was poorly innervated, even near the tip of the crown, and it had a different tubular structure and adjacent pulpal structure from the innervated dentin. Radicular dentin was not innervated in most areas but did contain a few labeled axons where the predentin was wide and the odontoblasts were columnar, as at the buccal and lingual poles of some roots. Our results show tha dentinal sensory nerve endings in primate teeth can be profuse, sparse, or absent depending on the location and structure of dentin and its adjacent pulp. When dentin was innervated, the tubules were straight and contained odontoblast processes, the predentin was wide, the odontoblast cell bodies were relatively columnar, and there was an adjacent cell-free zone and pulpal nerve plexus.