74 



CHILD. 



[VOL. I. 



Fig. 7 shows a stage of considerably over 100 cells from 

 the upper pole. The two cells which are lettered N are inter- 

 esting because at this time they become very large, but lie 

 rather deeply, so that they are partially covered by the rosette 

 cells. Each then divides equally, the spindle forming an angle 

 of 45 with the median plane, with its dorsal end nearest the 

 median plane. After a period of rest both pairs divide again, 

 but this time the more nearly median of each pair divides with 

 a vertical spindle, i.e., the ectoderm becomes two-layered here. 

 Although I have been unable to follow these cells farther, it 

 seems probable from their position and size that they may 



tr\ 



FIG. 5. 



form the apical plate and thus the supraoesophageal ganglion 

 or part of it. I do not believe that in Arcnicola the rosette 

 cells form the ganglion, though Wilson regards it as probable 

 for Nereis. 



The small cells lettered K correspond to the head-kidney 

 cells of Nereis ; I have been unable to discover a head kidney 

 in Arcnicola, and these cells remain as small, inconspicuous 

 ectoderm cells. 



The gastrulation in Arcnicola is a combination of invagina- 

 tion and epiboly. The mesoblast begins the process by slowly 

 elongating inward and gradually passing into the segmentation 

 cavity which it just fills. Its bilateral division occurs during 

 its passage. In Fig. 5 it is commencing already to pass in- 

 ward. The entomeres are not enclosed until some hours later, 

 although their superficial area is constantly decreasing. Opti- 



