THE HISTOLOGY OF DISSEMINATED SCLEROSIS. 579 



Such a cell is seen in fig. 9 lying in the bay formed by two protoplasmic processes of 

 an enlarged, multi-nucleated, spicier cell. From such a cell all transitions can be 

 traced to the fully-developed fat granule cell, in which the protoplasm between the 

 vacuoles stains light purple and its outer rim forms a distinct membrane. When 

 this stage is reached, it is found that numerous glia spaces appear empty or are 

 occupied by fat granule cells in process of development ; that the long ramifying 

 processes of the enlarged glia cells extend for long distances and frequently envelop 

 the fat granule cells ; that numerous naked, swollen axis cylinders are found attached 

 to the original glia meshes, which are now only faintly visible ; that the nuclear 

 increase is largely related to the presence of deeply-staining nuclei with a small 

 amount of protoplasm ; and that all the blood-vessels in this area are dilated and 

 engorged with blood. 



Slightly later in the development of the process, it is found that as the fat granule 

 cells increase in number, there is often a definite reaction both in the endothelium 

 of the small vessels and in the adventitial wall of the pre-capillary vessels. A special 

 study was made of the recent areas to endeavour to trace proliferating endothelial 

 cells in the vessel walls and their possible migration into the tissue. In many 

 capillaries it seemed that the endothelial cells had proliferated and detached them- 

 selves from connection with the vessel. Some of the nuclei were perpendicular and 

 oblique to the vessel wall, and appeared as if passing into the surrounding tissue. 

 In the immediate neighbourhood of the vessels were frequently found, especially in 

 cerebral areas, small groups of cells, similar in their nuclear structure to the cells in 

 the vessel walls. Stages could be traced in the further development of these into 

 cells with distinct zone of protoplasm, which in its turn appeared vacuolated till true 

 fat granule cells were formed. It is thus seen that at this stage of maximum 

 development of fat granule cells, the cells of the blood-vessel walls take a share in 

 their formation, while at an earlier date they seem to arise from the proliferation of 

 the small glia nuclei. Everywhere in the area were numerous small vessels, which 

 give the impression of a new vessel-formation, but this might well be only an 

 apparent increase, because all are so dilated and perceptible. In the cerebral areas 

 this impression was much more marked than in the cord area, where the general 

 architecture of the tissue seemed retained, in spite of the large increase in the 

 cell elements. 



(3) Following this is a stage (figs. 10 and 350) in which the formation of fat 

 granule cells has reached its maximum. The whole affected tissue seems permeated 

 with these characteristic cells, which lie not only in every possible tissue space, but 

 fill the adventitial lymph spaces of all the vessels within the area — even the smallest 

 capillaries being surrounded by a uniform, cellular ring, which gives the cross-section 

 a very characteristic appearance. This stage may be termed that of a granular cell 

 myelitis, and is the type of area we have described under heading 2 as representing 

 the so-called acute multiple sclerosis. The presence of the fat granule cells in the 



