470 W. B. SCOTT AND HENRY F. OSBORN. 



Elasmobranch Fishes/ the cavity in the head is formed at 

 ^ period considerably before that at which it appears in the 

 body. These two head cavities have no caramunication 

 with each other, as the mesoblast in the head is in two 

 separate masses. A longitudinal horizontal section (PI. XXI, 

 fig. 10) through an embryo slightly older than f shows this 

 cavity (pp.) as an undivided slit bounded by columnar 

 mesoblast cells. But when the first visceral cleft appears as 

 an outgrowth from the hypoblast of the throat to join the 

 external skin, this cavity is necessarily separated into two 

 portionsj one behind and one in front of the cleft. This 

 cleft in the latest stages we have been able to observe never 

 pierces the skin, but it lies close to it and so divides the me- 

 soblast. The second cleft divides the cavity into three sec- 

 tions, and each succeeding one adds a fresh segment to the 

 number. Of course this number is not so great as it is in 

 the Elasmobranch Fishes. 



The section in front of the first cleft presents some features 

 which demand attention. It grows forv/ard and becomes 

 divided spontaneously into two portions, one of which lies 

 close to the optic vesicle (PI. XXI, fig. 11), and entirely in 

 front of the mouth, while the second (2/>^.) is enclosed alto- 

 gether in the mandibular arch. The first aortic arch {\aa) 

 runs between these two sections and somewhat dorsal to 

 them. We have not been able to make any satisfactory 

 observations upon their relation to the branches of the 

 fifth nerve, but from their position it seems in every way 

 probable that they have much the same relations as those de- 

 scribed by Mr. Balfour in the Elasmobranch Fishes. The 

 first division shows a small lumen surrounded by a layer of 

 short columnar cells ; in longitudinal vertical sections (PI. 

 XXI, fig. 11, 1 pi^'), it has a somewhat oval shape; in trans- 

 verse sections (tig. 13, pp.) it has a trans versally elongated 

 shape, and the cavity in these sections is seen to be largest 

 toward the middle line. During no period as late as Stage l 

 could we find any trace of a ventral union between the ante- 

 rior segments of each side, such as occurs in the Elasmo- 

 branchs. It may, however; occur later, as during Stage l 

 they approach very closely. The second segment (PI. XXI, 

 fig. 11, ^ pp)-) is considerably smaller than the first, and has 

 a very small lumen. Its cavity also is lined with columnar 

 cells, and forms a narrow slit running parallel to the first 

 visceral cleft. The mandibular aortic arch lies just anterior 

 to it instead of close to its inner side as in the Elasmobranchs. 



The other segments of the head cavity lie in the visceral 

 ' Balfour, loc. cit., p. 86. 



