THE SKULL. 593 



types these bones are liable to vary very greatly from the usual 

 arrangement. 



Besides these bones there is usually present in the higher 

 forms a lacrymal bone on the anterior margin of the orbit 

 derived from one of a series of periorbital membrane bones 

 frequently found in Fishes. Various supraorbital and postorbital 

 bones, etc. are also frequently found in Lacertilia, etc. which are 

 not impossibly phylogenetically independent of the membrane 

 bones inherited from Fishes; and may have been evolved as 

 bony scutes in the subdermal tissue of the papillae of the saur- 

 opsidan scales. 



The visceral arches of Fishes, especially of the Teleostei, are 

 usually provided with a series of membrane bones. In the true 

 branchial arches these take the form of dentigerous plates ; but 

 no such plates are found in the Amphibia or Amniota. 



The opercular flap attached to the hyoid arch is usually 

 supported by a series of membrane bones, which attain their 

 highest development in the Teleostei. One of these bones, the 

 praeopercular, is very constant and is primitively attached 

 along the outer edge of the hyomandibular. It seems to be 

 retained in Amphibia as a membrane bone, overlapping the 

 attachment of the quadrate and known as the squamosal ; 

 though it is not impossible that this bone may be derived from a 

 superficial membrane bone, widely distributed in Teleostei and 

 Ganoids, which is known as the supra-temporal. In Dipnoi 

 the bone which appears to be clearly homologous with the 

 squamosal would seem from its position to belong to the series of 

 dorsal plates, and therefore to be the supra-temporal ; but it is 

 regarded by Huxley (No. 446) as the praeopercular 1 . 



In the Amniota the squamosal forms an integral part, of the 

 osseous roof of the skull ; but in the Sauropsida it continues, as 

 in Amphibia, to be closely related to the quadrate. 



A larger series of persistent membrane bones are related to 

 the mandibular, and its palato-quadrate process. 



Overlying the palato-quadrate process are two rows of bones, 



1 It is not impossible that the solution of the difficulty about the praeopercular is 

 to be found by supposing that the praeopercular as it exists in Teleostei is derived 

 from a dorsal dermal plate, and that in the Dipnoi this plate retains more nearly than 

 in Teleostei its primitive position. 



B. III. 3 8 



