Fifteen cases in which the changes in the cerebellar neuroglia were specially studied are given in some detail. The results correspond generally to those of Weigert. In the molecular layer, Bergman's fibres are increased in numbers, but unevenly. Most of the new fibres run vertically, but some obliquely or transversely, the last often forming bands at two levels, viz. along the outer margin of the cortex and at the boundary of the granular layer. In the latter position they form basketworks enclosing Purkinje's cells. The transverse fibres are mostly delicate, but a number of large spider-cells at the border of the granule layer give off coarse fibres, running to the surface. The largest collection of glia-nuclei is in the same situation. In the granule layer also the changes are of unequal degree. They consist in loss of granules, the place of which is taken by hypertrophied neuroglia fibre and nuclei. In the medulla the hypertrophy is rarely of great extent, and appears to prefer the immediate neighbourhood of the vessels, where large, coarse-fibred spider-cells are also found. Fibres and nests of glia-nuclei occur, however, between the nerve-bundles. Jn general, the rule is that in progressive paralysis the molecular layer is most involved, then the granule layer, more in spots, and last and least the medulla. Hence it is the dendrites of Purkinje's cells which appear to be chiefly affected in this disease, and their bodies also vanish in advanced cases. The morbid process thus seems to advance from without inwards. Little clinical value is claimed for these results, owing to the irregular distribution and frequently slight degree of the foci of disease; but it may be supposed that the changes contribute to the ataxy and incoordination. The paralytic seizures are more likely, from these cases, to be connected with diseased foci in the thalamus. No relation could be detected between the cerebellar changes and absence or increase of the reflexes.
