374 A. A. W. HUBRECHT., 
“At a very early period, when the hindgut commences to 
be formed, there arises at its ventral surface as the first indica- 
tion of the allantois a cellular knob, which encloses but a small 
outgrowth of the hypoblast. The allantoidean knob, however, 
does not develop freely into the coelomic cavity as it does in 
other mammals, but it proliferates along the ventral wall of 
the pelvic region. At the point, where the latter passes into 
the amnion fold, it continues to proliferate along the ventral 
wall of the amnion (see fig. 121) up to the point where the 
latter has remained connected with the chorion. The out- 
growth of the hypoblast elongates and becomes the narrow 
allantoic duct; the much more considerable proliferation of 
connective-tissue conducts the umbilical vessels towards the 
chorion, spreads out in the well-known way against its inner 
surface, and penetrates into the villi of the serous membrane. 
The allantois thusutilises during its development the connection 
which was already present between the serous envelope and the 
embryo, in the pointed prolongation of the amnion, and does 
not grow up towards the serous membrane independently as a 
free outgrowth.” 
I will now explain in how far the hedgehog’s embryology 
tends to substantiate or to invalidate any of these assumptions. 
We have seen that after the completion of the germinal area 
its epiblast remains attached to the trophoblast (from which 
it has split off) along a circular line of insertion (figs. 15, 
20, 27, 29, 31, p. 289). 
At the time of the formation of the amnion this line of 
attachment travels upwards (figs. 82 and 51), the circle becom- 
ing smaller and smaller until at the completion of the amnion 
no connection between embryonic epiblast and trophoblast any 
longer exists, and the secondary connection is brought about in 
the usual way above referred to by the budding forth of the 
allantois and its juxtaposition against the serous envelope. A 
similar primitive circular connection of the germinal area with 
the wall of the blastocyst can be found back in the mole 
(Heape), in many rodents with partly inverted germinal layers 
(Selenka), and (if I understand v. Beneden’s modified inter- 
