129 
culation. In his Comparative Embryology I, 249, BALFOUR reprints 
most of the article cited. Minor (7), 433, laid stress upon the röle 
of the chorion and upon the fact that the placenta is necessarily 
always a product of the chorion, and further upon the fact that the 
allantois in man is permanently and in the rabbit primarily merely 
a stalk of connection between the embryo and the chorion. 
The results of recent investigations strengthen greatly my view. 
It is by the chorion that the ovum is attached, except in certain 
cases, where the development has obviously been modified secon- 
darily. It is from the chorion that the foetal villi grow out. On 
the other hand it is evident that the yolk sack is primitively a pro- 
duct of the splanchnopleure and distinct from the somatopleuric chorion; 
the failure of the mesoderm and coelom to spread completely over 
the yolk (entoderm of the blastodermic vesicle) in certain mammals 
does not alter the fundamental relations. It is true that in certain 
marsupials the chorion is very imperfectly separated from the yolk 
sack, but it does not appear that this represents an ancestral stage 
of the mammalia; on the contrary it is probably a purely marsupial 
modification. I am therefore unable to recognize any reason for con- 
necting the evolution of the placenta with the yolk sack or vitelline 
circulation. The role of the allantois is secondary; it serves as a 
medium of blood supply, either, as we have seen, as a carrier of vas- 
cular trunks to supply the circulation of the chorion (unguiculates) 
or bringing its own circulation into play by growing together with 
a non-vascular chorion. 7 
The question remains whether the unguiculate or the ungulate 
type of placenta is to be regarded as the more primitive? At first 
thought the resemblance of the foetal envelopes of ungulates to those 
of sauropsida lead us to conclude that the allantoic piacenta must be 
the more primitive; the resemblance, referred to consists in the early 
complete separation of the chorion (serosa) from the other parts and 
in the development of the allantois as a large free vesicle. But the 
ungulates are highly modified mammals not related closely to the 
lower placentalia, while the unguiculates do merge into a generalized 
mammalian type. When we consider further that the lower ungui- 
culates show the typical chorionic placenta in its full perfection, the 
conclusion is unavoidable that this is the nearer type to the ancestral. 
In fact the placenta appears, in animals with the chorionic type of 
the organ, before the allantois becomes free, and the great size of 
the allantoic vessels is connected primitively not with the allantois, 
but with the already important chorionic circulation; the placenta is 
