The current view of basal ganglia organization holds that functionally corresponding subregions of the frontal cortex, basal ganglia and thalamus form several parallel segregated basal ganglia-thalamocortical circuits. In addition, this view states that striatal output reaches the basal ganglia output nuclei (the substantia nigra pars reticulata (SNR) and the internal segment of the globus pallidus (GPi)) via a 'direct' pathway, and via an 'indirect pathway' which traverses the external segment of the globus pallidus (GPe) and the subthalamic nucleus (STN). However, the topographical relationships of GPe and STN, and their topographical relationships with the basal ganglia-thalamocortical circuits are still unclear. The present work reviewed primate data on the topographical organization of STN afferents from GPe, and STN efferents to the pallidum, striatum and SNR, and examined these data with respect to a tripartite (motor, associative and limbic) functional subdivision of the striatum and pallidum. This examination indicated the following. (1) On the basis of its efferent connections, the STN may be divided into a motor and an associative territories, as well as a smaller limbic territory, each projecting to corresponding areas in the pallidum and striatum. (2) Efferents from GPe are in a position to contact subthalamic cells projecting to GPi/SNR, thus providing anatomical support for the existence of indirect pathways. (3) Moreover, given the tripartite division of the striatum, pallidum, and STN, the available data indicate the existence of indirect pathways connecting functionally corresponding subregions of the striatum, pallidum, and STN, as well as indirect pathways connecting functionally non-corresponding subregions. On the basis of the above we suggested that there may be two types of indirect pathways, one which terminates in the same subregion in GPi/SNR as the direct pathway arising from the same striatal subregion, and another which terminates in a different GPi/SNR subregion than the direct pathway arising from the same striatal subregion. We termed the former a 'closed indirect pathway' and the latter an 'open indirect pathway'. The application of these concepts to the surveyed data suggested the existence of three closed indirect pathways, each connecting the corresponding functional (motor, associative, and limbic) regions of the striatum, pallidum, STN, and SNR, as well as of two open indirect pathways, one connecting the associative striatum to the motor subregions of the basal ganglia, and the other connecting the associative striatum to the limbic subregions of the basal ganglia. While the organization of the closed indirect pathways fits the closed segregated arrangement of basal ganglia-thalamocortical circuitry, the organization of the open indirect pathways fits the recently suggested open interconnected scheme of basal ganglia thalamocortical circuitry. The clinical implications of this scheme for Huntington's disease are discussed.