DEVELOPMENT OF CEPHALODISCU S 197 



a line of yolk granules after the cell outlines are broken up. The protoplasm also carries 

 numerous small pigmented bodies, which are probably excretor}' matter accumulated in 

 the tissue. They first appear as small black or brownish bodies, which increase in size 

 and fuse together into elongated large masses. It is probable that in C. nigrescens and 

 C. fumosus, in which the body is coloured brown or black, this pigment remains in the 

 tissue permanently, while in others it is finally extruded from the surface. 



One of the sides of the embr^'o, probably the ventral, is comparatively thick and is 

 formed of elongated cells, which closely resemble those forming the thickened wall of 

 the proboscis. This has been termed the ventral thickening, and Harmer (1905) believed 

 that it somehow becomes converted into the proboscis of the adult. This comparison 

 is strongly emphasized in my own figures of the embryo and the later larval stages, 

 though I have not been able to show how the transformation takes place. 



The inner part of the embryo, which represents a wide blastocoel, is filled with a solid 

 mass of yolk granules. Gilchrist (1917) maintains that below the external layer a few 

 cells appear, which become closely applied to the yolk granules. As the yolk granules in 

 this layer are used up, each of the cells sends out a long protoplasmic process towards the 

 adjacent ones on each side, so that ultimately they form a chain of cells which gives rise 

 to the endoderm. If any such cells are present on the outer surface of the yolk mass, they 

 must be clearly distinguishable from this mass ; but his own drawings do not throw much 

 light on this point, nor does he seem to be certain of their exact place of origin. According 

 to this view the endoderm originates by a process of unipolar proliferation of the cells 

 into the blastocoel. The cellular structure of these cells breaks down, and while some of 

 the nuclei with the associated protoplasm go to form the endoderm, others pass towards 

 the centre and become vitellophags. The same opinion was maintained by Harmer 

 (1905), but Andersson (1907) questioned the accuracy of this statement and said that at 

 least in his species the endoderm originates by typical invagination. Unfortunately he 

 was not able to carry out any study of the cellular structure. Owing to the imperfect 

 preservation of the material the ectoderm and endoderm appeared uniformly filled with 

 yolk granules, which were, however, absent from the immediate neighbourhood of the 

 central lumen. Schepotieff (1909), who found gastrula stages in C. indictis, also main- 

 tains that the endoderm is formed by invagination. 



Between these divided opinions it is difficult to say which represents the correct mode 

 of origin of the endoderm. I have not been able to obtain any of the intermediate stages 

 between the solid blastula and the early larval stage which shows the origin of the 

 coelom ; but from a consideration of the structure of the latter I am more inclined to 

 agree with Andersson. 



Judging from the quantity of yolk present inside the blastula one would infer that a 

 considerable time elapses before the commencement of further changes. During this 

 interval cilia develop on the outer surface of the body and the embryo leaves the parent 

 colony as a planula-like larva, which has been observed by Andersson (1903). During 

 this free-swimming planula stage the inner mass of yolk is slowly absorbed, leaving a 

 homogeneous detritus in which irregular cavities appear. The excretory matter accumu- 



