740 TRANSACTIONS OF SECTION D. 



whole of the invaginated layer is endoderm, from which notochord and mesoderm 

 take their origin by a process of folding. (2) That the invagination is a double pro- 

 cess ; on the ventral side of the blastopore the unfolded cells are true endoderm, 

 whilst on the dorsal side they are ectoderm ; the ectodermic roof of the archenteron 

 becomes completely used up in the formation of notochord and mesoderm, which 

 become cut out of the wall of the archenteron by the upgrowth of the true endo- 

 derm cells at the side. 



The first view was put forward by Kowalevsky and Hatschek and was sup- 

 ported by Samassa and myself, while the second view was advocated by Livofif 

 and has recently been strongly supported by Cerfontaine. Morgan and Hagen 

 take an intermediate view, asserting that there is a difference between the cells 

 on the dorsal and ventral sides of the archenteron, but they do not assert that the 

 dorsal cells are ectoderm. The object of the present communication was to 

 present the results of a careful re-examination of the whole subject. 



The starting-point may be made from a flattened blastula. On the flat side are 

 some larger cells, which have been rashly identified with the whole endoderm, 

 but which pass by imperceptible gradations into the cells of the rounded side, 

 and which have yolk-granules of the same size and character as those of the 

 other cells. Invagination commences at one edge of the flat surface by the 

 establishment of a growing point, where cells are added to both inner and outer 

 layers, accompanied by an invagination of the newly formed inner cells, which 

 gradually involves the larger cells mentioned above. The impression is given 

 that these larger cells are passively dragged in by the movement of the active 

 cells. The growing point becomes the dorsal lip of the blastopore, and eventually 

 a cylindrical gastrula, with a wide terminal blastopore, is formed. The future 

 neural plate is indicated by a flattening, and in this way we are enabled to identify 

 the growing point mentioned above with the ^orsal lip of the blastopore. A 

 difference between the nuclei of the outer ectodermal cells and those of the invagi- 

 nated cells now becomes observable, the passage from one to the other kind 

 taking place sharply at the blastoporal lips. No difference whatever, either in size 

 orquality of yolk-granule, is observable between the cells forming thereof and those 

 forming the floor of the archenteron. 



Subsequently a new growing point is formed at the ventral lip of the blasto- 

 pore, and by the activity of this the blastopore becomes narrowed and forced up 

 on to the dorsal surface. Then the first indication of the folds to form mesoderm 

 and notochord are seen, and then, and then only, in consequence of a disappearance 

 of yolk-granules, some difference becomes observable between floor and roof of 

 the archenteron ; but this difference is a secondary, not a primary, histological 

 differentiation. The mesoderm originates from a dorsal lateral fold of the endo- 

 dermic wall, which becomes cut into anterior and posterior halves by the growth 

 of a septum. Both halves of the fold remain open into the gut. The front half 

 becomes, however, eventually closed oft" and forms the first myotome on each side, 

 which is different from the rest, both in its shape and history, and corresponds to 

 the mandibular head-cavity of other vertebrate embryos and to the collar cavity of 

 Balanoglossus. The posterior division corresponding to the lateral plate of meso- 

 derm of other vertebrate embryos, and to the trunk cavity of Balanoglossus, retains 

 its connection with the gut for a longer time. From its front end the somites of 

 the body are cut off. The head cavities arise later still as a single median evagina- 

 tion of the anterior gut-wall, agreeing in this respect with Balanoglossus and the 

 Elasmobranch embryo. Before this outgrowth separates from the gut, it begins to 

 be divided into right and left halves of different character. 



With some slight modifications, therefore, the simple view of the development 

 of Amphioxus taught by Kowalevsky and Hatschek is to be maintained, and the 

 development of other vertebrates ought to be interpreted in terms of the develop- 

 ' raent of Amphioxus, not vice versa. 



3. The Evolution nf Fishes, By Dr. A. Smith Woodward, F.R.S, 



