428 Metschnikoff on Germ-Layers. [May 
must be regarded as together forming a phagocytoblast. This 
conception is not in the least weakened by the fact that the cells 
forming these layers do not appear all at once, but often gradually 
migrate from the ectoderm (for instance, in Halisarca). The 
organization of the Sponges presents no peculiarities such as 
to justify us, along with Balfour and Bütschli, in separating the 
group from the other Metazoa. In this respect, therefore, I fully 
agree with most students of the Sponges, more especially with 
K. Heider (30). The lack of a mouth-opening, or in other words, 
the presence of numerous pores of entrance, can only be accounted 
as an important distinction, when an unwarranted genealogical 
significance is ascribed to these structures. From the stand-point 
occupied by the Phagocytella theory, these peculiarities of the 
sponge body are easily explicable. Again, the predominant part 
_ played by the ameeboid cells in taking in food would only give 
cause for surprise in case it were possible to speak of a differ- 
entiated mesoderm in the Sponges. But in this group there 
still endures a phagocytoblast, which must be regarded as the 
common foundation of endoderm and mesoderm; and conse- 
quently the rôle of the amceboid cells presents no difficulties. 
About a year after I had given in my “ Studies on the Sponges”? 
a general sketch of my views, Balfour arrived at conclusions 
regarding the phylogeny of the Coelenterates which were in per- 
fect accord with the principles of the Phagocytella theory. 
“ Paradoxical as it may seem,” says the English embryologist 
(29) (p. 147, vol. i.), “it appears to me not impossible that the 
Cæœlenterata may have had an ancestor in which a digestive tract 
was physiologically replaced by a solid mass of amceboid cells. 
This ancestor was perhaps common to the Turbellarians also.” 
It is very surprising that Balfour, believing this, was so strongly 
in favor of the gastræa theory, and regarded the parenchymella 
theory .as oe in itself. More recently Gotte (34) has . 
_* As early as 1877, in a paper on the digestive organs of fresh-water Turbellaria 
(“ Memoirs of Natural mae J Suey of New N vol. v.), I wrote as fol- 
lo th ł animals, it is evident that in 
this respect there is a , fundamental similórity between the Wii representatives of 
two types of the Metazoa reen and Vermes); that is, between Sponges and 
Turbellaria. If we compare the larvæ of the first group, particularly the Amor- 
aes sited deicribed = Schmidt, with the lower Turbellaria, we are at once 
robability et two classes are much more closely related 
