A SOUTH AMERICAN SPECIES OF PEKIPATUS. 349 



which is very useful in later stages for determining the 

 boundary between the embryo and the uterus. The size of the 

 segmenting ova of Peripatus imthurni varies from "04 

 mm. to "07 mm. in diameter, and that of the blastosphere is 

 "105 mm. in diameter. 



The next stage, measuring "112 mm. x *190 mm., represented 

 in figures 7 a and 7 b, is a most important stage ; during it the 

 blastosphere, which was described above, is invaginated so as to 

 form what appears to be a gastrula. I have several series of 

 sections of embryos in this stage, but those figured, which are 

 both from the same embryo, are by far the best ; fig. 7 b is a 

 transverse section through the embryo in the middle of its 

 length, showing the invagination ; fig. 7 a is a section through 

 one end of the embryo beyond the point of invagination. 



The cells are large and generally fairly well defined, more 

 especially the outer layer ; in the inner layer it is more difficult 

 to distinguish cell-outlines; the nuclei of the outer layer show 

 a distinct reticulum, and those of the inner layer are rather 

 more chromophilous. The opening of the gastrula is situated 

 on one side of the vesicle, that is, it opens at right to the long 

 axis of the uterus. 



This stage, which may be called the pseudogastrula stage, 

 seems to me of great interest, and to be really the key of the 

 whole matter. 



The outer layer of the pseudogastrula forms in later stages 

 the wall of the embryonic vesicle ; the embryo proper is formed 

 solely from the inner layer of the pseudogastrula. 



Of the significance of this stage, and of its relations to the 

 mammalian pseudogastrula, I will say more later on in the final 

 part of the paper. 



This stage doubtless corresponds to Kennel's stage, figured 

 in Part I, PL viii, fig. 56, where he makes the outer wall 

 of the gastrula u. e., which he iuterpretes to be uterine epi- 

 thelium ; he also letters the point of invagination as o., but 

 neglects to give, as far as 1 can see, any explanation of o. 



The result of this is that Kennel throughout his paper con- 

 siders what I have termed the wall of the embrvonic vesicle to 



