196 



F. H. EDGEWORTH. 



is not certain — it might be supposed that a muscle strip 

 Avhich formerly divided into upper and lower portions now by- 

 some atavistic process no longer does so. On the other hand, 

 the fact that, in all the animals of the second class, the myo- 

 tome, undivided, lies at first across and unattached to the 

 palato-quadrate, i.e. shows a condition which is the perma- 

 nent one in Amphibia and Ceratodus, suggests that the con- 



Text-fig. 22a. 



.•^[k I. 



kv^oul \jxi^ 



KuO«T\<V.|[. 1^^ 



2 2 a. 



Acipenser, embryo 9 mm. 



dition in Amphibia, Ceratodus, and Mammalia is the primary 

 one, and tliat the one present in Selachii, Teleostomi, and 

 Sauropsida is a secondary one. It would follow that the 

 palatine or ptei'ygoid process of the quadrate was not primarily 

 a process for attachment of muscles nor an upper jaw. 



Fiirbrin<rer divided Vertebrates into two classes with 

 regard to the connection of the quadrate with the skull — 

 those with movable quadrates (streptostylic), and those with 

 immovable quadrates (monimostylic). The latter condition, 



