another of my preparations, it appears that the egg at first 

 divides into a few blastomeres, which are arranged around a 

 vestigial, or perhaps we might say virtual blastocoele, and that 

 as development proceeds this 

 blastocoele swells but is kept 

 continually filled by cells pro- 

 duced by tangential divisions of 

 the peripheral blastomeres. At 

 the conclusion of segmentation, 

 the embryo strikingly recells a 

 coelenterate planula, as shown in 

 Figure 2. 



The interior mass of rounded 

 cells, however, does not consti- 

 tute the endoderm, for, in the next stage (Fig. 3) a regular 

 invagination has made its appearance, giving rise to an archen- 

 teron, which is, however, two or three cells thick on one side. 

 The interior cells of the former stage are seen to occupy the 

 space between archenteron and ectoderm, homologous with the 

 segmentation cavity of other forms. In the completed gastrula 



a curious plug of cells is seen projecting into the archenteron 

 which is no doubt the remains of the thickened side of the arch- 

 enteron of the younger gastrula (Fig. 4). 



When the embryo has attained the age of two days, the rudi- 

 ment of the coelom appears as a vesicle at the anterior end of 

 the archenteron exactly as it does in the embryos of Asteroidea. 

 At the same time the embryo takes on a shape which may be 



