SYNAPTA VIVIPARA. 61 



5. GASTRULATION AND FORMATION or THE HYDROCOEL AND COELOMIC VESICLES. 



Invagination of one of the poles of the blastula soon forms the archenteron from the 

 blind end of which the mesenchyme cells now begin to arise, some of the endodermal 

 cells being simply crowded out into the segmentation-cavity. I could find no evidence at 

 all of Selenka's ('83) two primitive mesenchyme cells, but on the contrary, I found cells 

 all over the archenteron which were in various stages of passing into the segmentation 

 cavity. On the other hand in not a single case were there found any of the ectoderma- 

 cells forming mesenchyme, and I feel no hesitation in affirming that the mesenchymel 

 arises exclusively from the endodermal cells, contrary to Ludwig's ('91) observations on 

 Cucumaria. The number of mesenchyme cells is comparatively small, and they never 

 become so numerous or play so important a part in larval structures as they do in S. 

 digitata, judging from Semon's ('88) account and figures. The completed gastrula (Fig. 

 9) is covered with cilia which are easily seen in the living specimens, but in none of my 

 preserved material has it been possible to demonstrate them. In a few cases gastrula 

 were found free from the egg-membrane and moving about actively in the fluid of the 

 body-cavity, but the great majority are still enclosed in the membrane within which they 

 rotate by means of the cilia. The membrane may be retained, as preserved specimens 

 show, until long after the coelomic vesicles are formed, and I am inclined to believe that 

 there is no definite time when it is cast off, but that it ruptures and is lost whenever the 

 larva has grown too big for it. As the archenteron increases in length it bends to one side 

 and unites with the wall of what subsequently becomes the dorsal surface of the larva. 

 Its lumen breaks through the surface, and thus the water-pore is formed, as described by 

 Selenka ('83) for S. digitata. Meanwhile the gastrula loses its spherical shape and 

 becomes more or less elongated (Fig. 10). The cells at the end opposite the blastopore 

 are already somewhat different in form from those elsewhere and make a sort of plate of 

 thickened ectoderm (Fig. 21) which may correspond to the so-called "neural plate" of 

 other echinoderm larvae. This plate, however, does not lie exactly opposite the blastopore, 

 but somewhat toward the ventral side of the larva, and as the gastrula increases in length, 

 it comes to lie more and more on that side. At the same time, the archenteron continues 

 to increase in length and grows forward and at the same time ventralward, thus drawing 

 away from the water-pore. In so doing, that part of it which grows forward pushes by 

 the part opening through the water-pore, on the right-hand side (looked at from the 

 dorsal surface) so that the latter comes to lie on the left side of the larva (Figs. Hand 

 12). As the archenteron grows it completely severs its connection with this vesicle, and 



