42 E, RAY LANKESTER. 
diverticula, and an cesophagus, grow towards the mantle-pole 
of the embryo, where at a late period of development the tu e 
so formed meets and unites with another portion of the ali- 
mentary canal which has developed on the opposite (the 
aboral) face of the embryo. The first appearance of this part 
of the alimentary canal is seen in a small tubercle in living 
embryos, at a period when the pen-sac is still open, which 
tubercle lies between the two rudimentary gill-buds, a little 
in advance of them. A cavity develops itself in this tubercle, 
which was mistaken by Kolliker for the aortic heart. A 
section at this period gives the result shown in fig. 3. Three 
layers of cells can readily be distinguished in this section as 
in that figured by me in the October number. ‘That marked 
ep is the epidermic layer or epiblast obviously enough. The 
mass of rounded cells below it may be called ‘ mesoblast,” 
since from it the muscles and skeletal elements are deve- 
loped; and by splitting, it gives rise to primitive vascular 
spaces (see fig. 5, pv), and an inner and outer group of;meso- 
blastic elements, corresponding to ‘‘ Darmfaserplatt’? and 
“« Hautfaserplatt.” Below this we come to a single layer of 
large fusiform cells separated by membranous intervals, ym. 
Is this the hypoblast? Leaving aside speculative dis- 
cussions the question is, Does the epithelium of the middle 
portion of the alimentary canal develop from this layer? 
The cell-layer in question separates the embryo from the 
enclosed mass of yelk, and extends everywhere over the yelk 
(into the yelk-sac as within the embryonic area). It persists 
as long as there is a trace of yelk. Owsjannikow describes 
such a layer of cells in Corregonus, which he does not regard 
as hypoblast. ‘The section fig. 3 shows the space al, the first 
trace of this portion of the alimentary canal, lined by cells, 
which appear to be independent of the layer ym. They seem 
rather. to be a differentiated layer of the mass which is 
marked mes. A somewhat later stage (fig. 4) shows also no 
connection between the cell-layer ym and the lining cells of 
the cavity al. It moreover gives evidence of a passage of 
yelk-granules y into the cavity al, which may be due to a 
natural inflection of the layer ym (in which case the cells 
lining the cavity a/ might be continuous with the layer ym), 
or, as is more probable, may be due to rupture of the layer 
ym in the preparation of the section. 
Transverse sections, of which one is figured in fig. 6, cor- 
responding in age to the vertical antero-posterior section 
fig. 4, equally fail to establish the derivation of the intestinal 
epithelium from the layer ym. Fig. 7 represents a section 
which comes nearest to such a condition as would render it 
