The efferent connections of the caudal pole of the globus pallidus (GP) were examined in the rat by employing the anterograde axonal transport of Phaseolus vulgaris leucoagglutinin (PHA-L), and the retrograde transport of fluorescent tracers combined with choline acetyltransferase (ChAT) or parvalbumin (PV) immunofluorescence histochemistry. Labeled fibers from the caudal GP distribute to the caudate-putamen, nucleus of the ansa lenticularis, reuniens, reticular thalamic nucleus (mainly its posterior extent), and along a thin strip of the zona incerta adjacent to the cerebral peduncle. The entopeduncular and subthalamic nuclei do not appear to receive input from the caudal GP. Descending fibers from the caudal GP course in the cerebral peduncle and project to posterior thalamic nuclei (the subparafascicular and suprageniculate nuclei, medial division of the medial geniculate nucleus, and posterior intralaminar nucleus/peripeduncular area) and to extensive brainstem territories, including the pars lateralis of the substantia nigra, lateral terminal nucleus of the accessory optic system, nucleus of the brachium of the inferior colliculus, nucleus sagulum, external cortical nucleus of the inferior colliculus, cuneiform nucleus, and periaqueductal gray. In cases with deposits of PHA-L in the ventral part of the caudal GP, labeled fibers in addition distribute to the lateral amygdaloid nucleus, amygdalostriatal transition area, cerebral cortex (mainly perirhinal, temporal, and somatosensory areas) and rostroventral part of the lateral hypothalamus. Following injections of fluorescent tracer centered in the lateral hypothalamus, posterior intralaminar nucleus, substantia nigra, pars lateralis, or lateral terminal nucleus, a substantial number of retrogradely labeled cells is observed in the caudal GP. None of these cells express ChAT immunoreactivity, but, except for the ones projecting to the lateral hypothalamus, a significant proportion is immunoreactive to PV. Our results indicate that caudal GP efferents differ from those of the rostral GP in that they project to extensive brainstem territories and appear to be less intimately related to intrinsic basal ganglia circuits. Moreover, our data suggest a possible participation of the caudal GP in feedback loops involving posterior cortical areas, posterior striatopallidal districts, and posterior thalamic nuclei. Taken as a whole, the projections of the caudal GP suggest a potential role of this pallidal district in visuomotor and auditory processes.