110 A. A. W. HUBRECHT. 
‘Morph. Jahrb., vol. 15, p. 165), does not yet sufficiently 
express the fundamental difference by which the mammalian 
segmentation process is characterised. 
There is no shred of evidence that with the disappearance of 
the yolk, which took place at a comparatively late stage when 
the mammalian character had already become predominant, the 
process of segmentation immediately fell back into the lines of 
-the so infinitely more distant alecithal ancestral forms. 
A further reason for the distrust with which we may look at 
this apparently holoblastic segmentation process is the fact that 
it finally results in the appearance of a tridermic blastocyst 
with elliptical blastoderm, primitive streak, &c., entirely 
corresponding to the arrangement of the Sauropsida. So these 
later stages have not returned to the earlier modes of develop- 
ment, but have continued along the lines laid down by the 
hereditary transmission of characters that were peculiar to-those 
ancestral forms that possessed a considerable amount of food- 
yolk. 
If the mammalian (pseudo) morula and (pseudo) blastula 
were indeed comparable with the same stages, 7. e. with the 
true morula and true blastulain Amphioxus and the Amphibians, 
about one half of the segmentation spheres would represent 
potential epiblast, and the other half potential hypoblast. Now 
this is evidently not the case. By far the greater portion of 
these segmentation spheres gives rise to what will afterwards 
be not any integral portion of the embryo, but a part of the 
foetal envelopes and of the membranous expansion by which 
the embryo is connected with the mother. Suppose we were 
able to repeat Roux’s or Chabry’s most important experiments 
on the partial destruction of one or more of the segmentation 
spheres with the earliest mammalian stages, then we might 
predict with great certainty from the data that Rauber, van 
Beneden, Heape, Selenka, and others have brought to light that 
only if the mother-cell of the inner cell-mass were attained, the 
normal development would be interfered with ; and that in the 
case of other segmentation cells being punctured, only a local 
defect in the foetal membranes would ensue. This hypothetical 
