In early stages of embryonic life all red blood-corpuscles are nucleated, as they are permanently in birds and the other classes of vertebrates below mammals. In embryonic mammals they multiply by division whilst circulating in the blood, just as in the newt. But it is generally believed that this is not the most important source of new ones. During the earliest stages of growth they are being formed in enormous numbers. Such instances of division as can be seen in circulating blood appear to be all too infrequent to account for their rapid multiplication, and there can be no doubt but that a more complicated method of production is more important. Their formation is described as taking place “endogenously.” Certain cells termed “vaso-formative,” or “vaso-sanguiformative,” reach a considerable size, and become stellate in form, or branched. Their nuclei divide without the cell dividing. Each nucleus accumulates a little hæmoglobin round it. A space filled with fluid appears inside the cell. The nuclei project into this space. Then they drop off with their envelopes of hæmoglobin. The outer shell of the big vaso-formative cell becomes the wall of a capillary bloodvessel. By its branches it links up with other vaso-formative cells, making a network of vessels. The fluid inside it is the plasma of the blood. The nuclei and their envelopes are blood-corpuscles. This, if it be a true story, is a comprehensive way of making bloodvessels and blood at the same time. Doubts have been thrown upon its accuracy, but many leading histologists strenuously maintain that this description is correct.