toward the left side (Fig. 7). Subsequent development shows that the more anterior of these two masses is the rudiment of the hydrocoel (i.e., the left anterior coelom), and the other one is the left posterior coelomic rudiment. These coelomic organs are derived therefore from mesodermal mesenchyme in the region lateral to the archenteron, and are in no way pinched off as pouches from the archenteron. It is to be noted that the origin of the coelomic vesicles in this way as bodies at first solid can be regarded as an extension of the tendency to which Narasimhamurti (1933) drew attention in Ophiocomina nigra, where the coelomic vesicles arise initially as thick-walled bodies with very small cavities. The point is an important one, and it is discussed more fully elsewhere (Fell, 1945).

At this time the oesophageal rudiment, or stomodaeum, is still solid. The archenteron, meantime, is taking on more of an epithelial arrangement of its component cells, the nuclei in sections showing up in regular series, one deep. Its cells begin to stain more intensely with cytoplasmic dyes than do the surrounding tissues, a fact probably related to some chemical change accompanying the absorption of yolk in this region.

All this time the mesenchyme continues to proliferate actively, and it becomes very prominent in sections of these and succeeding stages. Indeed, mesenchyme plays a most important part in the development of Amphipholis squamata, as it does also in Kirk's ophiuroid. A section through a somewhat older stage (Fig. 8) shows that a cavity begins to develop by splitting within the oesophageal mass, and almost at the same time the hydrocoel rudiment begins to form a small internal cavity, also by splitting. The left posterior coelomic rudiment remains for a while without any trace of an internal lumen. Meanwhile, on the right side of the archenteron,

Figure 8.
Ect., ectoderm; Arch., archenteron; Lft.P.Coel., left posterior coelomic vesicle; Hydr., hydrocoel; Ant.P., anterior pole; Oes., oesophagus; Rt, Coel., right coelomic vesicle; Mesch., mesenchyme.

some more of the mesenchyme cells have rounded up to form a small body which from its position undoubtedly represents the rudiment of the right coelomic vesicle. There is no trace in any of my specimens of any tendency of this body to separate into anterior and posterior portions, as Metschnikoff (1869) had claimed it does.

The embryo is still very opaque at this stage of development, even after treatment with clearing reagents, and the only satisfactory way of gaining a comprehensive picture of its structure is by means of sections and reconstructions made from them by Pusey's projection method. In Fig. 9 is drawn an embryo—or young larva as it may now be called—of slightly more advanced development. It is represented as a transparent object, having been reconstructed from sections by means of projection drawings, as described by Pusey (1939). The left posterior coelomic vesicle, and the right coelomic rudiment have each by this time developed a small internal cavity, by splitting as in the case of the other vesicles. With the exception of the hydrocoel, the coelomic vesicles have now reached the greatest degree of development that they ever attain, being, like the larva as a whole, quite vestigial. The hydrocoel now measures over 100 μ in length, and is about 50 μ broad. The left posterior coelomic vesicle is more rounded, and about 50 μ in diameter. The right coelomic rudiment is more elongated, about 80 μ in length by 20 μ across. The entire embryo has a somewhat oval outline, the major axis of length corresponding with the antero-posterior axis of the larva, and measuring about 250 μ, with a breadth of 150 μ.

Fig. 9.—The early larva when all the coelomic vesicles are present. Reconstructed from sections.
Emb.At., embryonic attachment; Lt.Sk.P., left skeletal plate; Rt.Sk.P., right skeletal plate; Lt.P.Coel., left posterior coelom; Hydr., hydrocoel; Mesench., mesenchyme; Ant.P., anterior pole; Oes., oesophageal sac; Rt.Coel., right coelomic vesicle; Arch., archenteron; Post.P., posterior pole.

At the posterior pole of the larva some of the mesenchyme cells have taken on the character of spiculoblasts, and have secreted two meshes of calcareous matter which are the vestiges of the larval skeletal plates. These will be described more fully below. While this differentiation of the coelomic vesicles from the mesenchyme has been occurring, the embryo has achieved an organic attachment to the wall of the bursa of the parent, and the cells of the wall of the abradial horn of the bursa begin to grow out to form the embryonic attachment (called “umbilical cord” by the early workers). This structure is more fully described below.

Certain changes now begin to take place at the anterior end of the larva. The cavity of the oesophageal rudiment has been steadily extending, and its cells begin to take on the appearance of an epithelium, as in the section shown in Fig. 10. The hydrocoel has also begun to differentiate, and in the same section it can be seen that a division into two lobes has occurred, these being cut across. In the right coelomic vesicle, and the left posterior coelomic vesicle, no changes or growth in size have occurred, and soon these structures are seen to undergo a process of degeneration, for they are vestigial, like so many structures of the peculiar reduced larva which forms. The vestigial larval skeleton, however, is still increasing in size and degree of differentiation. Soon the right and left posterior coelomic rudiments become disorganised and no longer distinguishable from the surrounding mesenchyme, into which their component cells come to be merged. Thus not all of the organs of the vestigial larva reach

Fig. 10.—Vertical section of early Larva.
Ect., ectoderm; Arch., archenteron; Lft.P.Coel., left posterior coelomic vesicle; Ant.P., anterior pole; Hydr., hydrocoel; Oes., oesophageal sac; Rt.Coel., right coelomic vesicle; Mesench., mesenchyme; Post.P., posterior pole; Emb.At., embryonic attachment.