Single jaw-muscle spindle afferent axons were characterized physiologically and intracellularly stained to determine whether particular physiological types of spindle afferent show distinctive morphologies. Microelectrodes filled with either horseradish peroxidase (HRP) or biotinamide (Neurobiotin) were advanced into the mesencephalic trigeminal nucleus (Vme) in anesthetized rats. Intracellular recordings then were characterized by their response: to palpation of the jaw muscles; when pressure was applied to the teeth and during passive ramp and hold and sinusoidal jaw movement. Seventy-one afferents were characterized physiologically and injected with HRP; an additional 61 afferents were typed and injected with biotinamide. The response of 43 stained neurons was recorded in the presence of suxamethonium. The major projection areas of these afferents were the: trigeminal motor nucleus (Vmo); region dorsal to Vmo; reticular formation, spinal trigeminal nucleus, superior cerebellar peduncle and Vme. One afferent type was modulated strongly during stretching of the jaw-elevator muscles. Based on their high sensitivity during stretching of the jaw muscles and/or their silencing during the release phase of muscle stretch, these afferents were classified as primary-like spindle afferents. These afferents projected most strongly to Vmo. A second type of afferent was modulated only modestly during stretching of the jaw-elevator muscles. These tonic afferents were classified as secondary-like spindle afferents because of their low dynamic sensitivity during ramp muscle stretch and their continued discharge during the release phase of muscle stretch. Secondary-like afferents projected most strongly to the region dorsal to Vmo. Boutons (n = 3,834) from 11 afferents were studied in detail. Secondary-like afferents had statistically larger boutons within Vmo. In both secondary- and primary-like spindle afferents, only a small number of boutons were associated closely with the somata and proximal dendrites of trigeminal motoneurons. In these cases, however, two to five boutons appeared to contact individual motoneurons, implying multiple monosynaptic inputs to a selective subset of jaw-elevator motoneurons. Some "giant" boutons were present dorsal to Vmo and in Vme. These results demonstrate that dynamically sensitive and nondynamically sensitive jaw-elevator muscle spindle afferents project preferentially to different regions. Primary-like spindle afferents are capable of providing feedback related to the dynamic phases of muscle stretch and project most heavily to Vmo. Secondary-like spindle afferents can transmit a feedback signal associated with muscle length and project most strongly to the supratrigeminal region. Both types of afferent have projections caudal to Vmo that may serve longer latency jaw-muscle stretch reflexes and/or the projection of proprioceptive information to the thalamus and cerebellum.