THE MALAYAN SPEOTES OF ONYCHOPHORA. 49 



have already appeared, but their nuclei present no particular 

 arrangement (PJ. 6, fig. 9 d). Even at this early stage in 

 the development the first somite, preparatory to advancing 

 along the side of the embryo towards its anterior end, is far 

 removed from the median plane. 



The endoderm is also forming, especially towards the 

 posterior end of the embryo (PI. 6, fig. 9 c, en.). In front of 

 the blastopore there are no endodermal nuclei, but at its 

 sides a few have already appeared. Towards the posterior 

 end they are more numerous, and in the region in question 

 an occasional nucleus may be seen halfway up the sides 

 (comp. figs. 9 a, 9c, 9d). There seems to be no doubt that 

 the endodermal elements in Eoperipatus are derived from 

 the lips of the blastopore, with which the endodermal layer 

 is continuous, and that they pass from that position through 

 the outer layer of the yolk. While this process of invagina- 

 tion is going on, the layer of yolk in question, which is 

 being gradually invaded by the endodermal elements, loses- 

 its dentoplasmic character and becomes more cytoplasmic. 

 The central mass of yolk presents the appearance of contain- 

 ing compound systems as well as separate yolk bodies, 

 between which there seems to exist a certain amount of 

 cytoplasm. Simultaneously with the increase in number of 

 the endodermal elements the cytoplasm grows at the expense 

 of the deutoplasm. 



Unfortunately the quantity of material at my disposal was 

 not enough to enable me to form very decided conclusions 

 on the formation of the blastula stage in Eoperipatus, 

 and I do not wish, in any way, to question the accuracy of 

 already published accounts of its formation in other genera 

 of the Peripatida3 ; but it is necessary to point out that the 

 incomplete condition of the ectodermal layer on the dorsal 

 surface of the very early gastrula stage, already described, 

 tends to show that in Eoperipatus the circuracrescence of 

 the yolk takes place after the same plan as in so many other 

 Arthi'opoda ; that is, by overgrowth from the future ventral 

 surface towards the dorsal. It is well known that the 



VOL. 45, PART 1. — NEW SERIES, JJ 



