THE HISTOLOGY OF DISSEMINATED SCLEROSIS. 671 



surrounding the vessels of the cord are, anatomically, processes of the arachnoid 

 space, and the lymph in them is in connection with the cerebro-spinal fluid. With 

 the advance of the pathological process this peri-vascular infiltration follows along 

 the vessels as they enter from the meninges and is most marked around the central 

 vessels. The evidence, first pointed out by Flexner, that the respiratory mucous 

 membrane provides for both the ingress and egress of the virus has much to support 

 it. Romer and Wickman, on the other hand, on clinical and experimental grounds, 

 think that the virus has its habitat in the intestinal tract and thence finds its way 

 along the lymphatic sheaths of the sympathetic nerves to the central nervous system. 

 Once this is reached an infection of the lymph spaces in the adventitial sheath of the 

 veins in the pia mater and spinal cord immediately follows. Wickman supports his 

 view of the lymphogenous origin of acute polio-myelitis on the following histological 

 grounds : that in many parts the chief and only change consisted in an infiltration 

 of the larger vessels, while the capillary region of such vessels was quite free ; that 

 the changes in the longitudinal axis were continuous — a continuity which reaches its 

 maximum intensity in Landry's paralysis, and that the infiltration of the adventitial 

 sheath of the vessels argued for the causal agent circulating in the lymphatic spaces. 



It has been assumed by most writers that the causal agent in disseminated sclerosis 

 circulates in the blood, but the possible lymphogenous source of the infection has been 

 supported on the following grounds : (l) the endo-vascular changes are much less 

 marked than the peri-vascular ; and (2) the pathogenic significance of the peri- 

 ventricular sclerosis. With reference to the former argument, it is again necessary 

 to point to the fact that in this investigation peri-arteritic changes, in the sense of 

 adventitial nuclear proliferation, were completely absent in the early areas before the 

 onset of a secondary cell infiltration due to resorptive processes had occurred. There 

 was thus no evidence in the cell proliferation of the adventitial sheath that a "noxa" 

 was circulating in the adventitial spaces. In view, however, of the findings of Orr 

 and Rows, that in infection of the cord by the passage of toxins along the sym- 

 pathetic nerves no adventitial proliferation was found, such a possible source of 

 lymphogenous infection cannot be denied. 



The peri-ventricular localisation was such a striking feature in several of the cases 

 that early in the investigation it was recognised that ependymal and peri-ependymal 

 lesions lead to important considerations in reference to the toxicity of the cerebro- 

 spinal fluid. Bullock's recent experiments point to this toxicity, and Lhermitte and 

 Guccione found that when carmine was injected into the lateral ventricles of a dog, it 

 was found, ten days later, almost wholly in the sheath of the sub-ependymal veins and 

 in the sub-ependymal tissue. It is logical, therefore, to assume that toxi-infective 

 agents in the cerebro-spinal fluid might follow the same route. The absence, however, 

 of any change in the ependymal epithelium seemed to contra-indicate the possibility of 

 a simple soakage of the cerebro-spinal fluid into the peri-ventricular tissue and also to 

 contra-indicate the irritating character of the fluid, for a granular condition of the ven- 



