Gigantopyramidal neurons, referred to as Betz cells in primates, are characterized by large somata and extensive basilar dendrites. Although there have been morphological descriptions and drawings of gigantopyramidal neurons in a limited number of species, quantitative investigations have typically been limited to measures of soma size. The current study thus employed two separate analytical approaches: a morphological investigation using the Golgi technique to provide qualitative and quantitative somatodendritic measures of gigantopyramidal neurons across 19 mammalian species from 7 orders; and unbiased stereology to compare the soma volume of layer V pyramidal and gigantopyramidal neurons in primary motor cortex between 11 carnivore and 9 primate species. Of the 617 neurons traced in the morphological analysis, 181 were gigantopyramidal neurons, with deep (primarily layer V) pyramidal (n = 203) and superficial (primarily layer III) pyramidal (n = 233) neurons quantified for comparative purposes. Qualitatively, dendritic morphology varied considerably across species, with some (sub)orders (e.g., artiodactyls, perissodactyls, feliforms) exhibiting bifurcating, V-shaped apical dendrites. Basilar dendrites exhibited idiosyncratic geometry across and within taxonomic groups. Quantitatively, most dendritic measures were significantly greater in gigantopyramidal neurons than in superficial and deep pyramidal neurons. Cluster analyses revealed that most taxonomic groups could be discriminated based on somatodendritic morphology for both superficial and gigantopyramidal neurons. Finally, in agreement with Brodmann, gigantopyramidal neurons in both the morphological and stereological analyses were larger in feliforms (especially in the Panthera species) than in other (sub)orders, possibly due to specializations in muscle fiber composition and musculoskeletal systems.

In the parietal cortex of the rat, sections cut tangentially show that profiles of medium and large apical dendrites are grouped into clusters. The number of apical dendrites in each cluster is variable and the usual separation between individual clusters is about 50 p. Despite these variations the pattern does not appear to be randgm. Reconstructions from one micron serial sections show that neurons giving rise to the ascending dendrites forming clusters are located at different levels in layer V. The cell bodies of these neurons are arranged vertically below their dendrites and show a tendency to form groups. All of the neurons have apical dendrites that ascend through the cortex with a few secondary branches occurring close to the base. The principal secondary branching begins in layer I11 and spreads obliquely up through layer I. Furthermore, beginning in the inferior region of layer I11 apical dendrites are added to the clusters at their peripheries. These are from layer I11 pyramids. It is clear that the superior aspects of the cluster arrangements must intermingle with those of the neurons in adjacent clusters. The neuropil surrounding the dendrites forming clusters appears to contain a few smaller dendrites. Small unmyelinated axons are the most frequent component of the surrounding neuropil and these form terminals which synapse on the spines and trunks of the clustered dendrites. There is no obvious function that can be ascribed to the clusters other than they may form a component of the columnar organization in cortex described in part by physiological techniques.

The distributions of cerebellar and somatic lemniscal projections to the ventral nucleus of the thalamus were compared in the opossum to determine the extent of overlap in the terminal field of these two fiber systems. Following lesions of these structures, the degenerated fibers were traced to the thalamus using the Fink-Heimer technique. The results indicate that the cerebellum projects only to the rostra1 portion of the ventral nucleus, while the gracile and cuneate nuclei project to the caudal portion of the ventral nucleus. We conclude from comparing these two afferent fiber systems that there is no detectable overlap in the cerebellar and lemniscal projections to the ventral thalamic nucleus of the opossum. These results support the hypothesis that the single somatic sensory-motor area of the opossum's cortex receives afferent fibers from at least two separate subdivisions of the ventral nucleus. Outside the border of the ventral nucleus, cerebellar and lemniscal projection fields do overlap, especially in a cell group just dorsal to the ventral posterior nucleus. This is a distinctly different type of organization of these afferent fiber systems. This cell group has recently been shown likewise to project to the parietal somatic sensory-motor cortex. Since the mode of intracortical termination of this projection differs markedly from that of the VP and VAL projections, the somatic sensorymotor cortex of the opossum can be said to receive fundamentally different projections from two thalamic regions both of which are recipients of both cerebellar and somatic-lemniscal information. C 1 Figure 9