648 AGNES T. M. ELLIOT. 



(Torpedo) to five (Hexauclius), and only retain their ventral 

 roots ; but in the embryo as many as seven may be indicated, 

 and all provided with the rudiment of a dorsal ganglion 

 (embryo Torpedo). More usually, however, there are only 

 five or six (Spinax) of these nerves in the embryo, and only 

 the last two or three show dorsal ganglia. But even in these 

 cases we may still find seven post-otic somites (Spinax). 



The skull itself is not usually segmented, but the myo- 

 comata are attached to the parachordal masses in segmental 

 fashion. Rosenberg has shown in some sharks (Carcharias) 

 distinct signs of extra vertebrre joined to the post-otic region 

 of the skull, but leaves it uncertain whether these represent 

 the spino-occipital segments of other sharks or whether they 

 are additional trunk vertebras which are being drawn into the 

 skull. 



In the Selachians we may therefore conclude that at least 

 seven trunk segments have been fused with the head in the 

 post-otic region, though in some forms (Scyllium and Pris- 

 tiurus) traces of no more than five are still visible during 

 development. 



In other fishes there has been a further pushing up of 

 trunk segments into the head. In Accipenser as many as six 

 more segments have been fused, and in the Teleostei probably 

 only three more. 



In Amphibia the line of division between head and trunk 

 is supposed to lie in the same place as in Selachians, but if so 

 a number of the occipital segments have entirely disappeared, 

 as at most three spino-occipital segments can be demonstrated 

 in the embryo, and in the adult there are no spino-occipital 

 nerves. In the Anura, indeed, the first spinal nerve and its 

 ganglion also disappear. 



In the Urodeles, Miss Piatt has shown for Necturus that 

 there are three post-otic segments in the head of the embryo, 

 the first not giving rise to any muscles, the second partly 

 uniting with the third, and the muscles from the second and 

 third segments being hard to distinguish in the adnlt. She 

 has also shown that there are two rndimcntary dorsal arches 



