Intra-axonal recording and horseradish peroxidase labelling techniques were used to examine structure-function relationships for vibrissa-sensitive primary afferent fibers (N = 40) in rat trigeminal brainstem subnucleus interpolaris (SpVi). All responded at short (mean-0.42 ms) latencies to trigeminal ganglion shocks and to innocuous stimulation of an individual vibrissa in a slowly adapting type I, slowly adapting type IIa, slowly adapting type IIb, low-velocity-sensitive rapidly adapting, or high-velocity-sensitive rapidly adapting fashion. As in the medullary dorsal horn (Hayashi, '82; Jacquin et al., '86a), functionally distinct mystacial vibrissae-related fibers were morphologically indistinguishable. Each gave rise to up to ten collaterals that entered interpolaris perpendicular to the long axis of the nucleus and often overlapped to form a densely packed, highly circumscribed, and largely continuous column of terminal arbors. While some morphological variability was observed both within and between individual axons, variance within a given functional class was no greater than that between classes. Nonmystacial vibrissae afferent arbors also formed similar ovoid, dense circumscribed terminal plexuses. Presumably, these individual arbors collectively form a longitudinal tubelike vibrissae representation in this and other components of the trigeminal brainstem nuclear complex. The relative locations of each fiber's terminal field could be accurately predicted by the particular vibrissa innervated. Contrary to previous data obtained with similar methods (Hayashi, '82), but consistent with the findings of earlier transganglionic tracing (Arvidsson, '82) and histochemical (Belford and Killackey, '79) studies, the arbors of these fibers terminated throughout the mediolateral extent of SpVi. Axons innervating rostral vibrissae terminated medially, and those that supplied caudal vibrissae innervated the lateral SpVi. Dorsal vibrissae were represented in the ventral SpVi, while ventral vibrissae were represented more dorsally. This transverse topography extended largely throughout the rostrocaudal extent of interpolaris. In summary, the vibrissae primary afferent map in SpVi is inverted, faces medially, and is rostrocaudally consistent. In its most caudal aspect, as the displaced substantia gelatinosa swings laterally, the map is less orderly and incomplete. These generalizations also apply to the nonmystacial vibrissae afferents.