Giant cells are frequently met with. These differ from the hemo blasts apparently only in their greater size and the irregular character (fig. le) or the multiple condition of the nucleus. The size relation between nucleus and cytoplasm is slightly altered in favor of the latter. The cytoplasm is basophilic and the nucleus generally stains as intensely as that of the mononucleated hemoblast. The giant cell is most prob ably a hypertrophied hemoblast. Its multinucleated condition re sults from amitotic division of the hemoblast nucleus. The multi nucleated forms are much less abundant in the 5-mm. mongoose embryo than in the 10-mm. pig embryo. Moreover, in the latter they have attained a much larger size and may contain many more nuclei than in the former. The evidence from the pig embyro suggested that giant cells were multiple erythroblasts, eventually differentiating into definitive erythrocytes. The giant cells of the mongoose include only the earliest stages of those described for the pig, hence no confirm atory evidence of the above conclusions regarding their erythro poietic function accrues from a study of these cells; but the evidence, so far as it goes, accords with that derived from the study of giant cells of the pig, and the tentative conclusion that they are erythro blasts rather than erythrophages may remain unaltered.
A careful study of the mesenchyma shows that numerous cells, both singly and in groups, take on nuclear and cytoplasmic hemoblast characteristics, and eventually round up more or less and separate from the parent mesenchyme. Exactly similar conditions were de scribed and illustrated for the yolk-sac of the 10-mm. pig embryo and need not be further considered. As single cells these hemoblasts may wander into adjacent vessels; or as groups (blood-islands) they may become inclosed in endothelium forming in the surrounding mesen chyma, to produce an "angiocyst" of the growing vascular net. Similar conditions have frequently been described by various authors (Maximow, Dantschakoff, and others) in sections of the yolk-sac of various forms. Since Stockard has more recently described compar able processes in the yolk-sac of the living Fundulus embryo, there need remain no further doubt that mesenchyma does actually differ entiate directly into hemoblasts and into enveloping endothelium.
The point of controversy now centers on the question whether the endothelium of the vascular net of the yolk-sac and elsewhere can trans form into hemoblasts; for in the Fundulus embryo Stockard claims that the endothelium has no hemogenic capacity; and on the basis of this fact he casts doubt upon evidence contributed as a demonstration for blood cell origin from endothelium in other forms (Schridde, Maximow, Jordan and Flippin, Jordan, Dandy, and others). However, Reagan's (18) more recent findings in Fundulus embryos contradict Stockard's con clusions on this point.
Appearances in the yolk-sac of the pig seemed to permit of no escape from the conclusion that endothelium did in fact transform in part into hemoblasts (s). The yolk-sac of the mongoose gives exactly the same evidence. This may be seen by a glance at figure 2a, 2b, or 2c. The cell b is at the crucial stage of metamorphosis. It is still contin uous with the endothelium, and directly continuous with the endothe lial cell f; but it has all the nuclear and cytoplasmic characteristics of a hemoblast (compare with la and 1b). Cell c represents a further stage in the same process and is about to separate from the endothe lium as a mononculeated giant cell. A final stage is represented in figure 3, where a metamorphosed endothelial cell has just become sepa rated, and has not yet fully rounded up into the typical hemoblast. Similar examples might be multiplied indefinitely; more are given in the paper dealing with this same phenomenon in the pig embryo; but enough has been shown, especially in the case of cell b, to leave no further doubt, I believe, that young endothelium can indeed trans form into cellular blood elements.
It is of interest and importance to note that the origin of hemoblasts from endothelial cells also in the bone marrow, both under normal and certain pathological conditions, is described by some of the leading pathologists, e. g., Aschoff (see Mallory's "Principles of Pathologic His tology"). Such a "lining cell" of the marrow blood spaces is believed by Aschoff and by Mallory to be capable of differentiation either into an erythrocyte, a granuloblast, or by hypertrophy into a megakaryo cyte, which accords with the evidence derived from my study of the yolk-sac vessels of the pig and the mongoose. Moreover, Mallory, in his list of normal cellular blood elements (p. 21), substitutes for the commonly described mononuclear leucocyte (transitional leucocyte) his "endothelial leucocyte." This cell he derives from "the endothelial cells lining blood, and to a less extent lymph, vessels by proliferation and desquamation. They also multiply by mitosis after emigration from vessels into the lesions." Since few any longer doubt that young mesenchymal cells can trans form on the one hand into certain blood-cells and on the other into endothelium, it seems a priori reasonable to suppose that the only slightly altered mesenchyma, the embryonic endothelium, can also occasionally transform directly into hemoblasts. The same thing should be true also, perhaps to a lesser extent, with regard to meso thelitun, and for the same reason. According to Bremer (1) the meso thelium covering the body-stalk of a 1 mm. human embryo does in fact give rise to some extent to blood-cells.