EARLY ONTOGENETIC PHENOMENA IN MAMMALS. 183 
calculated to confirm us in our judgment that these orders 
are more primitive, and that in them the phenomenon of 
placentation has not yet come to be normalized to a particular 
type. Still, this conclusion is only of partial value, as we 
will by and by see that the diversity here alluded to is in 
one case characterised by high specialisation, in another by 
the appearance of peculiar characteristics, which throw hght 
on certain general problems of placentation, whereas in others, 
again, types are represented which might furnish an argument 
to those who wish to subdivide the order of Insectivores in 
two or more independent orders. 
At all events, we must conclude from the facts before us 
that the really simplest and earliest form of placentation is 
no more represented in any living genus of mammals, and 
we have to attempt to disentangle out of all the numerous 
data at our disposal the phylogenetic evolution which has 
gradually brought about the numerous forms now known 
to us. 
When discussing the trophoblast on p. 18 of this treatise, 
we saw that a change of function, which must have occurred 
at a very early period, when this larval envelope contributed 
towards the retention of the blastocyst inside the genital 
ducts of the henceforth viviparous Protetrapod, in the first 
place developed adhesive qualities by which the blastocyst 
remained fixed to the uterine wall. We have supposed that 
a second parallel phenomenon was an increase in size of the 
larval trophoblast, precursory to the further development of 
the embryo proper. In consequence of this the adhesive 
surface would become of more considerable extent, and could 
be pressed more firmly against the maternal mucosa. If, at 
the same time, phagocytic properties become developed (which 
are now generally recognised to be characteristic for ever so 
many mammalian trophoblasts), then in addition the tropho- 
blast layer might serve to hand over into the cavity enclosed 
within it material elaborated by it, which might in its turn 
serve towards the growth and nutrition of the embryonic 
cells (s. str.). For it is sufficiently known that both in the 
