610 DR JAMES W. DAWSON ON 



probably as a reaction to the irritation caused by the products of degeneration. At 

 the same time it may happen that one and the same " noxa " destroys the specific 

 nerve tissue, and in an equally primary manner causes the glia to proliferate. Again 

 it is possible that we may have isolated primary proliferative processes in the glia — 

 which in their turn cause secondary degeneration of nerve cells and fibres. Storch 

 has pointed out that in chronic diseases, in which the plan of the nerve tissue remains 

 unchanged, the newly-formed glia fibrils show exactly the same arrangement as the 

 original fibres, whereas in cases of acute destruction of tissue, this regularity does not 

 hold good. He, therefore, distinguishes between an isomorphous and a reparatory 

 sclerosis. The glia proliferation may, therefore, be merely a substitution process, or 

 an inflammatory process, or, more rarely, a primary glia proliferation. 



It is impossible to discuss here the question of the spatial relation of the glia 

 fibres to the glia cell protoplasm. All that can be done is to indicate the stages in 

 the elaboration of the glia fibrils (figs. 382-384). These can be followed very 

 beautifully in the evolution of an early area into a sclerotic area by means of 

 Heidenhain's iron-hsematoxylin method and Ford-Robertson's methyl-violet stain. 

 In the rapidly proliferating glia cells, the first stage in their transformation is an 

 enlargement of the nucleus (fig. 8), by which its chromatin structure becomes clearer. 

 This is followed by the development of a considerable amount of deeply-staining 

 protoplasm around the nucleus and the further development of large, branching, 

 protoplasmic processes (figs. 9, 349), till forms of very varying size and shape are 

 produced. In many of these large glia cells, two or more nuclei may be found 

 (fig. 379) : this may represent a karyokinesis which has remained incomplete. After 

 cell division the new cells rapidly increase in size, and form protoplasmic processes, 

 and are potential fibril-forming cells (figs. 379, 380). From a close consideration of 

 the specimens the conclusion was reached that the formation of fibrils can take place 

 in the enlarged pre-existing glia cells without cell division. The first indication of 

 the formation of fibrils consists in a definition of the edge of the protoplasmic pro- 

 cesses (fig. 382) : when this can be followed throughout the concave border of two 

 adjoining processes (fig. 383), it gives to the fibril formation the appearance of 

 recurving fibres (fig. 384) with their convexity near the cell nucleus. The general 

 arrangement of the fibrils corresponds at first to the general outline of the borders 

 of the protoplasmic processes. At a later stage the relation to individual nuclei is 

 less easy to determine. Ford-Robertson thinks that each branching process becomes 

 converted into several plain processes by gradual splitting at the forks down to the 

 close vicinity of the nucleus. His methyl-violet method shows very clearly that 

 many of the fibres are attached to the walls of a vessel by an expanded " foot," but 

 frequently, especially around the capillaries in the cortex, the new formed fibrils of 

 adjacent cells were found to form a network, with elongated meshes, around the 

 capillary wall (figs. 21, 391). 



l',\\;\ cells, that have produced fibrils, undergo slow and gradually regressive 



