ON THE EMBRYOLOGY OF THE TELEOSTS. 205 



thickening away from the hypoblast (fig. 25) and succeeding this the hypoblast was seen 

 to extend itself across beneath the now formed notochord (figs. 26 — 28 in just the man- 

 ner which we were led to expect from our observations made from surface views. 



It is always difficult to make out cell limits in optical sections of eggs as small as those 

 on which we worked and hence they are omitted in our drawings, but it was seen that in 

 this notochordal swelling of the hypoblast that the cells were smaller than the adjacent 

 portion of the layer just as was seen in the surface views. 



We would repeat and lay especial stress upon the fact that we witnessed, by constant 

 and repeated observations on living specimens, every step in the formation of the notochord, 

 from a hypoblast of but one cell in thickness until the chord was segmented off from its 

 parent layer and eventually entirely enveloped by the mesoblastic tissues, and at no time 

 did we witness the slightest appearance that could be regarded as evidence that any por- 

 tion of the notochord was other than hypoblastic in its origin. The later history of the 

 notochord will be treated under the respective sections of the development of the fish. 



Previous to Balfour's first paper (Quarterly Journal of Micr. Science, 1874) all writers on 

 vertebrate development had regarded the notochord as belonging not only in position but 

 in -its method of origin to the mesoblastic tissues. Dr. Balfour there and also in his com- 

 plete monograph ('76 b ) showed that in the Elasmobranchs, at least, the notochord is a 

 hypoblastic structure in its origin. Following him, Hensen, in a paper on the development 

 of the rabbit, gives the same general account of the derivation of the chorda dorsalis with 

 only such variations as might.be expected in two diverse classes of Vertebrata. Calberla 

 ('77) studying Petromyzon, Sygnathus anclRana arrives at similar results and was the first 

 to show that in the Teleosts the notochord is an endodermal structure. Rudwaner ('76) in 

 a short paper also treats of the origin of the chorda dorsalis, but his article is of little 

 value and his figures are evidently diagrammatic and do not represent the true state of 

 affairs. He derives the notochord (p. 161) from the outer germ layer or epiblast. More 

 recently Braun, in his paper on the development of Parrots in the Arbeiten a. d. zool. zoot. 

 Inst, zu Wiirzburg (the exact reference to which I have not at hand), still considers the 

 notochord as of mesoblastic origin. All of these observers have worked with and studied 

 sections and hence the discrepancy in their results. On the other hand our observations 

 were made on the living embryo, and hence we saio the organ formed and did not have to 

 call upon the imagination to fill up any gaps and also the sources of error in interpretation 

 were eliminated. 



In the development of the neural canal the Teleosts present a marked contrast according 

 to all observers with the other vertebrates in that it is first formed as a solid cord in 

 which a lumen afterwards appears. 



At about the same time that the notochord was first seen the first appearances of the 

 medullary folds were witnessed. Coincident with the invagination of the hypoblast the 

 edge of the blastoderm increases in thickness and this thickening is most marked at the 

 posterior margin in the median line forming what Balfour has called the tail swelling. When 

 first noticed it presented much the appearance of the " stage A " of Dr. Balfour's Elasmo- 

 branch ('76 b pi. vi, fig. A). This soon became elongate and more and more prominent until 

 a broad shallow longitudinal furrow finally made its appearance dividing it into two lateral 

 halves. These halves are the medullary folds and are low rounded ridges. With the 



