EARLY ONTOGENETIC PHENOMENA IN MAMMALS. 125 
and the two spots here alluded to become very marked. They 
protrude with cushion-like convexity in the uterine lumen, 
and are covered by a pallisade epithelium, against which the 
blastocyst becomes attached. The trophoblast proliferates, as 
was noted above (Hubr.,’99, Pls. 5 and 6), and as soon as the. 
blastocyst has come to adhere to the two spots just mentioned 
the maternal epithelium is destroyed, and processes of mutual 
interlocking between the subepithelial maternal proliferation 
and the trophoblastical proliferation now ensue. 
Here, again, as in Erinaceus and Sorex, the embryonic 
proliferation becomes very soon dominant when once the 
passage of maternal blood into trophoblastic spaces has been 
brought about by the aid of the maternal trophospongia, and 
now the allantoidean villi, ensheathed by trophoblast, con- 
tinue by their further mutual growth to considerably thicken 
the incipient placenta. This thickening takes place here 
again in a centripetal direction. It should be remarked as a 
very wide difference between what is observed in Tupaja and 
what occurs in the hedgehog and the shrew, that the two 
pairs (one pair for each foetus) of cushion-shaped spots of 
Tupaja, described above, first serve for an omphaloidean 
placentation, but that after a certain time the vascular area 
on the umbilical vesicle is dislodged out of its situation, its 
place being then taken by the allantois, which develops the 
villi ensheathed by trophoblast that were above alluded to. 
The two placentas right and left are of course identical. 
They seem to be rarely expelled as afterbirth in toto, but 
rather to be broken up partially, perhaps even partly to be 
subject to a resorption in loco, as was noticed for the mole’s 
placenta. These data I owe to Dr. Miss M. v. Herwerden 
(?06), who has lately looked through a series of preparations 
of puerperal Tupaja uteri. 
b. Chiroptera, Carnivora, Rodentia.—In the Chirop- 
tera the placentation has been studied by Frommel (’88), 
v. Beneden and Julin (’84), Géhre (92), Nolf (96), Duval 
(99), and others. Here, too, there is a considerable amount 
of maternal trophospongian proliferation, which in many cases 
