1. The cytoarchitecture and the exact borders of the thalamic ventralis intermedius (Vim) nucleus of humans as originally delineated by Hassler (17) have been studied on the basis of stereotaxic coordinates correlated with Nissl- and Golgi-impregnated sections, using a microscopic image analyzer. 2. The Vim nucleus forms part of a relatively "cell-sparse zone" which includes the other ventrolateral thalamic subnuclei. It is distinguished by the presence of darkly stained, large and medium sized, angular cells with areas of approximately 500-1,000 microns 2 and 300-400 microns 2, respectively, and a cell density of approximately 50-90 (mean 65)/mm2 in 50-microns-thick sections. 3. Both sets of neurons have the characteristics of thalamocortical relay neurons in Golgi preparations. Large neurons have rectangular or square somata 30-50 microns diam and are concentrated mainly in the lateral and ventral two-thirds of the nucleus. The medium neurons have square to round somata, 15-25 microns diam, and are distributed homogeneously through the nucleus. The total dendritic arborization of both types is usually symmetrical in all directions and at least 500-600 microns diam. 4. The borders between the Vim nucleus and the Nucleus ventrooralis (Vo) and between the Vim nucleus and the Nucleus ventrocaudalis internus (Vci) are clearly identified by clearcut differences in cell size and cell density. The borders between the Vim nucleus and the Nucleus ventrooralis internus (Voi) and between the Vim nucleus and the Nucleus zentrolateralis intermedius (Zim) are quite obscure, and these nuclei, with Vim, seem to be parts of the large cell sparse zone comparable to that described in monkeys as VLp or VL. The border between the Vim nucleus and the Nucleus ventrocaudalis externus anterior (Vcea) is also unclear but the increased cell density and intermingling of small and medium-to-small neurons with large neurons are the major features that distinguish the Vcea nucleus from the Vim nucleus cytometrically. 5. The position and anatomic organization of the human Vim nucleus make it likely that it is the region in which kinesthetic response were recorded in the accompanying paper but extension of the recording sites into the Vcea nucleus cannot be ruled out.