[59] EMBRYOGRAPHY OF OSSEOUS FISHES. 513 



of the commeiiciDg development of the vertebral bodies, have a shape 

 somewhat like an hour-glass with a wide neck, the narrow portion rep- 

 resenting the middle of the future vertebral centrum. In Alosa of the 

 same relative age as Gambusia the sheath of the uotochord is exceed- 

 ingly thin, and cannot be made out even in sections as anything moi-e 

 than the merest film. In Sahno, on the other hand, the notochordal 

 sheath, at the time of hatching, is a thick homogeneous membrane 

 several times the thickness of that found in embryos c'' Alosa of the same 

 age, and thicker even than that of Gambusia. It follows from what we 

 have learned, from the foregoing comparisons, that we are not warranted 

 in proposing any general theory of the development of the notochordal 

 sheath, even within the limits of a group as restricted as that of Teleostei. 

 Of the development of the membrana elasiica externa, which covers the 

 notochordal sheath, I have nothing to say, not yet having been able to 

 con\ance myself of its presence in Alosa, for instance. 



It has been insisted that Teleostei "may fairly be described as pass- 

 ing through an Elasmobranch stage, or a stage like that of most pre- 

 jurassic Ganoids, or the sturgeon, as far as concerns their caudal fin" 

 (Balibur, Com.p. Embryol., II, 04). We have already noted two excep- 

 tions to this rule in the singularly modified pipe- fishes and Eippocam. 

 pus. It now remains to call attention to another type in Gambusia, 

 where the extreme tip of the notochord is bent upwards to only the 

 slightest degree; so slightly, indeed, that its extremity is not raised 

 above the level of the dorsal line of the uotochord, although the hypural, 

 the urostylar cartilages, and the rudiments of the neural and_ haemal 

 arches are developed. This is in a comparatively late stage, but when 

 we come to study still earlier stages we do not even find any evidence of 

 the dorsal prominence the same as at the margin of the tail fin of embryos 

 of Salmo, w^hich is clearly the margin of the embryonic caudal fin, where 

 the tip of the notochord grows backwards and obliquelj^ upwards into 

 this rudimental structure. Balfour is also in error when he says that 

 in Salmo the rays of the caudal fin appear sinultaneously above and 

 below the end of the notochord. This is actually the case, however, in 

 Gambusia, where at least three fin rays arise even above the end of the 

 still cartilaginous urostyle, while six develop above the level of the 

 uotochord itself. This subject has been most fully discussed by A. 

 xigassiz, to whose invaluable memoirs on the development of the cau- 

 dal fin embryologists will in future be obliged to refer lor data. 



The anterior flexure of the notochord also varies considerably in dif- 

 ferent forms. In Alosa it is slight after hatching, but in Hippocampusit 

 is excessive, and is accounted for by the extensive downward flexure of 

 the head in the region of the neck of this singularly modu^'d Teleost. 



The subuotochordal rod is developed as a strand of cells in Alosa and 

 /S'ai?«o, just below the notochord. QEllacher calls it the aorten-sirang, 

 by which he seems to imply that it shares in the development of the 

 aorta. 



S. Mis. 40 .^3 



