300 Papers from the Department of Marine Biology. 



a 3-cell group, located just to the side of the mid-ventral line. This 

 group is still partially attached dorsally to the endothelium and may 

 be regarded as a later step in the process of endothelial hemoblast 

 formation from the yolk-sac shown in figure 4. 



In figure 12 is illustrated a 4-cell group from the ventro-lateral wall. 

 Three cells are arranged in horizontal series and jut out directly into 

 the lumen of the aorta; the three cells, moreover, are at successively 

 later stages of differentiation. The proximal cell has intermediate 

 features betM^een an endothelial cell and a hemoblast {a); the distal 

 cell has erythroblast characteristics and is very similar to some of the 

 adjacent aortic erythroblasts (c). Close to the proximal cell of this 

 group the endothelium is thickened and contains an increased number 

 of nuclei, similar to the attached pole of the multicellular clusters. For 

 purposes of comparison, both dunensional and structural, a few of the 

 adjacent cells of the lumen are shown. A series of successive stages 

 of differentiation is indicated by the letters a to d ; a is a. typical hemo- 

 blast; h, a typical megaloblast; c, typical normoblasts, and d, a late 

 erythroblast (normoblast) stage. 



From the foregoing it seems clear that the endothelium of the aorta 

 in the mesonephric portion may proliferate locally in the ventral or 

 latero- ventral or even dorsal wall, giving rise ventrally to cell-clusters 

 which grow in size, the constituent cells of which undergo a coincident 

 differentiation into hemoblasts and young erythroblasts. Similar cell- 

 clusters of larger size are nowhere to be found in the yolk-sac vessels; 

 such are apparently not formed during the yolk-sac homopoiesis; this 

 fact, if countervailing evidence were needed, would sufficiently dis- 

 credit any interpretation of the aortic cell-clusters in terms of masses 

 of hemoblasts carried to the aorta by the blood and caused to adhere 

 to the ventral wall by reason of pressure and the adhesive properties 

 of their cytoplasm. Coincident with the production of hemoblasts in the 

 aorta through a cell-cluster phase, the endothelium of the same portion 

 of the aorta at any point, though more generally ventrally, may pro- 

 duce also, through the transformation of individual endothelial cells, 

 hemoblasts in an exactly identical manner as above described for the 

 vessels of the yolk-sac. These two processes — hemoblast production 

 by endothelial cell-clusters and by transformation of individual endo- 

 thelial cells — are essentially similar; the difference involved is one of 

 the degree of endothelial proliferative and differentiative capacity, not 

 one of kind. A few of the cells of the larger cluster appear to have also 

 phagocytic properties. 



The question arises as to why the hemogenic activity of the aortic 

 endothelium is generally limited to the ventral area of the mesonephric 

 portion. Why do the larger clusters form only ventrally? This 

 portion differs from other portions in that this is the region along which 

 the larger ventral arterial blood-vessels to the abdominal viscera 

 migrate caudally to their definitive location, the process involving a 



