2B0 
BULLETIN OE THE UNITED STATES FISII COMMISSION. 
to come slowly into existence througli a process of folding, but is at once cut out of a i 
preformed mass of material, and, as I have said, I am inclined to believe the same is i 
true of the proximal parts of the mesoblast plates. Indeed, Hertwig seems to enter- ;; 
tain much the same idea when he speaks of the way in which the chorda and chorda- i 
entoblast are separated from the mesoblast. Only he views the process from a differ- , 
ent standpoint, and so compares the delamination fissure {a to h, Fig. 5) directly with 
the dorsal fold which lies between the outfolding notochord and mesoblast sheet of , 
Triton. 
If, now, the invaginate entoderm, which in the frog forms only a part of the roof J 
of the alimentary canal, should take upon itself to form gradually more and more of D 
the alimentary canal, it would probably from the very beginning become wider : the 
space from x to x would increase in width. The increase in width would naturally 
bring it about that a much larger j^art of the mesoderm jilate would be delaminated • 
than was formerly the case ; there would constantly be less and less of the mesoderm . 
which would have to arise as an outgrowth. Finally the condition in the Teleost would , 
be reached, in which the mesoderm plate is split off' from the primitive hypoblast along '| 
its whole width. j 
V. FORMATION OF THE ALIMENTARY CANAL. ! 
The alimentary canal is formed from the simple entoderm lamella by a process of j 
folding along the median line (Fig. 76, PI. xcvii). The fold is converted into a tube by ;i 
the meeting of its lower edges. There is a solid postanal gnt formed as a thickening of ' 
the entoderm lamella, not as a fold. At the end of the xiostanal gut is Kupff'er’s vesicle, 
which is formed in essentially the same way as the iiermanent alim entary tube. It is ij 
scarcely necessary to say that Kupff'er’s vesicle and the entire iiostanal gut atrophy. 
It seems worth while to give a detailed descrixition of the formation of the ali- 
mentary canal, and this will best be done by going methodically through the several 
stages, taking ux) each series of sections at the x^osterior end of the embryo and passing , 
forward. 
The formation of the alimentary canal begins fir stin the region of Kuxiffer’s vesicle 
and very shortly afterwards in the bra nchial region. Some hours before the closure of ij 
the blastoxiore the entoderm cells of the neurenteric streak {n. str.) become columnar 
(Fig. 60, PI. xcv). These cells subsequently inclose Kupff'er’s vesicle. The rest of | 
the entoderm lamella in this stage is made uxi of flattened cells. 
Stage of 35 hours (Fig. 65, PI. xcvi, and Figs. 66 to 72, Pis. xcvi and xcvii, series 
of sections from the same embryo). At the time when the blastoxiore closes sections j 
through the neurenteric streak show an early stage in the formation of Kiixiff'er’s vesi^ 
cle. The columnar entoderm cells have been xmshed uxi in the middle line (Fig. 66),,;, 
and arch over the cavity of the vesicle {1c. v.), the floor of which is formed by the x^eri- i 
blast. The vesicle is in this stage broad and shallow. Posteriorly the cavity fades ! 
away (longitudinal section. Fig. 65), and anteriorly likewise. The lining cells of thej 
vesicle are directly continuous with the entoderm lamella. Passing forward (Figs. ; 
67 and 69) the entoderm of the trunk is found to be thicker in the median line than 
laterally, the cells under the notochord axiproaching a columnar shape. Another | 
feature will be noticed on comxiaring Fig. 67 with a section through an earlier stage! 
(Fig.61,Pl. xcv). The whole embryo is growing deexier and narrower, and coincideutlyj 
