STUDIES IN MAMMALIAN EMBRYOLOGY. 331 
the connective-tissue cells forming the perivascular prolifera- 
tion are seen to be more or less arranged in concentric circles, 
the long axes of the elongated nuclei being perpendicularly 
situated to the radius of vessels and proliferation. Thereby 
the vessel and its accessory sheath, called forward by the pro- 
cess, obtains a certain distinctness, a certain independence 
from the rest of the decidua, and this is seen to be indeed to 
some extent the case, when at the edge of the section (the 
muscularis having first been removed) the vessel with its 
proliferation sheath is seen to separate comparatively easily 
from the stroma. The same is seen in the section of the fresh 
after-birth towards its convexity, i. e. the surface where it has 
been torn off from the remaining decidual tissue. Here the torn 
ends of the vessels that have proliferation sheaths often tend 
to separate by rents or fissures from the rest of the stroma. 
Inside of the proliferated cavity the endothelium bordering the 
lumen is always very distinct, and the same as in earlier stages, 
thus proving that here the proliferation is interstitial and not 
endothelial, as was the case in the formation of the tropho- 
spongia. 
It would be natural to suppose that this proliferation indi- 
cated either a venous or an arterial space, i. e. to assume that 
the distinction noticed between those blood-cavities that have 
it and those that have not, coincides with the different course of 
the blood-stream towards or away from the trophospongia. I 
am prepared to show that this is not the case, and that both 
amongst the afferent and amongst the efferent vessels there 
may be such that possess and others that are devoid of a pro- 
liferation sheath. I was able to convince myself of this in an 
injected specimen, where the injection mass had not sufficiently 
penetrated to reach the regions of the trophosphere. It was 
restricted to the larger afferent vessels of the mesometrium, 
thence it could be traced to the region opposite the meso- 
metrium, where the decidual swelling is attached, and here the 
injection mass was seen to have reached as far as a certain 
number of the decidual spaces, some of which had, whereas 
others had not, the perivascular proliferation. At the point 
