by the direct implication of the contiguous tissue, and not by propagation along the branches of the central artery.
2. Occasionally I found the round cells accumulated about the peripheral veins and associated in occurrence and distribution with the peripheral vessels. Around a fork-shaped peripheral vein, which is ramifying in the lateral part of the anterior horn, such an infiltration appears in Plate II, Fig. 3. This section seems to me to demonstrate distinctly how easily a false conception may arise. The focus here is wholly confined within the anterior horn. It is therefore within the extent of the region supplied by the central artery. But serial sections showed that the infiltration was associated only with this peripheral vein and not with the central artery.
3. I found, as a rule, infiltrations occurred in connection with the distribution of the largest vessels (or of the lymphatics).
I assert not that the central artery plays no part, but that it plays not the most important part, in the pathogenesis. The round cell accumulations are sometimes associated with it; but they are also associated with the central vein, and sometimes with the peripheral vessels. I believe that these cells are directly derived not from the blood vessels, but from the lymphatics of the blood vessels. In my earliest publication on this subject, I maintained that the specific pathologic changes could best be explained as a lymphatic infection. If we regard the disease as a blood infection, it is undoubtedly difficult to account for the continuity of the changes in the long axis of the cord in fatal cases; for in a vascular condition one would expect the scattered foci of invasion to be separated by normal areas. The continuity of the lesions and the tendency of the paralysis to spread from below upward seem to me to indicate a diffusion of the virus within the spinal cord and its membranes. And this diffusion obviously follows not the blood but the lymph channels—the lymph vessels and the lymph spaces. In corroboration of this opinion, I may adduce the marked infiltration of the lymphatics of the large blood vessels in Plate II, Fig. 5, and the fact—still more convincing—that, at least in several places, the earliest infiltration appears in the lymphatics. In a blood infection, one would expect the capillary region to be the first attacked; whereas, here, initial implication of the capillaries has been shown not to occur.
