The endothelium of these pericerebral vascular channels has con siderable proliferative capacity, both internal and external (fig. 17). Extravascularly dividing endothelial cells round up more like hemo blasts and may indeed be progenitors of extravascular hemoblasts; but certain minor morphologic differential characters suggest that we may here be possibly dealing with simply slightly modified proliferating mesenchyma.
It appears, then, that young endothelium, whether in the yolk-sac, ventral area of the mesonephric portion of the aorta, or in the intra embryonic mesenchyma (pericerebral), functions in the formation of hemoblasts, and in an essentially similar fashion.' From the above there seemed to be indicated a hemogenic capacity also on the part of the endothelium of the mesonephric glomerulus. This region was carefully studied. But there is apparently no intra vascular differentiation of endothelium into hemoblasts. An occa sional cell has the appearance of being at an early phase of separation from the vascular endothelium; but in consequence of the irregularity of the plane of section, due to the delicate nature and the contorted condition of the glomerular capillaries, a confident interpretation seems impossible. The normal hemogenic capacity of the glomerular endo thelium is certainly very meager, if not actually nil.
In an earlier paper (a) I stated the conclusion that in the pig embryo the glomerular endothelium liberated cellular elements (hemoblasts) extravascularly. Study of the sections of the mongoose embryo (of relatively younger stages of development) compels a revision of this conclusion. In the mongoose embryo the visceral layer of the Bow man's capsule of the mesonephric tubule, while closely applied to the glomerulus, is nevertheless easily distinguishable from the endothelium of the glomerular capillaries. The endothelial cells are flat plates; the cells of the capsule are cuboidal or pyriform in shape. An occasional cell appears between the two layers. This may be an endothelial cell separating extravascularly, but such interpretation must remain uncertain.
In the pig embryo the capsular epithelium is still more closely applied to the glomerulus, and the cells of both layers are very similarly flattened. Generally the endothelial nucleus is more vesicular than that of the capsule cell. The cells of the capsule are frequently rounded up and appear to be in early stages of separation. These are the cells which I interpreted as endothelial elements separating extra vascularly; but in the light of conditions in the mongoose embryos, this conclusion does not seem warranted. In the mongoose embryo the vascular endothelium of the mesonephric glomerulus is apparently hemogenically inactive; but the presence of the pyriform cells in the capsular membrane (fig. 22), certain of which remain attached by the merest thread of protoplasm—coupled with the fact of an occasional free cell within the lumen of the capsular portion of the mesonephric tubule—suggests that certain of these cells may separate in the embryo to become macrophages. A similar process is described by Mallory in the case of certain infections of the human kidney (e. g., acute capsular glomerulo-nephritis).
It is probably incorrect to regard the endothelium of the mesonephric glomerular capillaries as "embryonic" and relatively undifferentiated. This endothelium has to perform a specialized secretion process in connection with the nephric function of the mesonephros. This may explain the inability normally to produce hemoblasts. On the other hand, the capsular portion of the tubule plays a lesser role in this were tory process, and in the mongoose embryo is relatively much less highly differentiated, as indicated by the still cuboidal character of its epithe lium. This fact, in view of the mesenchymal origin of this portion of the tubule, may underlie the possibility of a capsular cell to separate from its epithelium and become a free intratubular element, perhaps a macrophage.