CROCODILIANS, LIZARDS, AND SNAKES. 187 



meet the petrosal; and in the Iguania, where it is decurved, it does not 

 come in contact with the petrosal, owing to the shortness of the latter. 

 In certain families where the petrosal is x)rodiiced beyond the arcade, 

 and the parietal is decurved, the two elements are in contact for a 

 short distance, as in the Varanidte. In the Teiidaj and Scincid* the 

 contact is mainly efl'ected by a short descending process of the parie- 

 tal. This process is especially elongate in the Scincidfc. The arcade 

 is the anterior border of the petrosal in the Permian Theriodouta, and 

 it marks the position of the anterior semicircular canal. The membra- 

 nous wall of the brain case, anterior to the petrosal, contains an ossifi- 

 cation which is of uncertain homology. It reaches or approaches by its 

 superior extremity the frontal, and might hence be supjjosed to be the 

 orbitosphenoidj but this homology is vitiated by the fact that its infe- 

 rior portion passes behind the optic foramen. The latter j)osition is 

 that of the alisphenoid, and so the bone is named by Parker ^ But 

 there is another element, the epipterygoid, posterior to it and immedi- 

 ately anterior to the petrosal, which has been sui^posed to be the true 

 alisphenoid. 



Leaving this question, and adopting for the bone in question the pro- 

 visional name of postoptic, I remark that it is typically triradiate, send- 

 ing two branches upward and one downward. This is its character 

 in Agamidae, Yarauidie, and Teiidae. The posterior superior branch is 

 much reduced in many Iguanid;T3 and Lacertidie and in some Agamidae 

 {Megalochilus), and it is absolutely wanting in Gerrhosaurus and 

 Ghamceleon. There is no postoptic in Heloclerma. In the Khynchoce- 

 phalian genus SpJienodon these two elements coexist with an orbito- 

 sphenoid, lying between the optic and trigeminal foramina. The two 

 together may be homologous with the mammalian alisphenoid. The 

 epipterygoid is present in all Lacertilia excepting the Ohamseleonidie 

 and Annulati ( Amphisbaenia). Its superior connections are quite char- 

 acteristic of the different families. Inferiorly it rests on the pterygoid 

 posterior to its ectopterygoid process, excepting in the Geckonidae, 

 where its point of attachment is opposite to that process. In the same 

 family it does not reach the parietal, but the superior extremity rests 

 on the apex of the supraforaminal part of the petrosal. In the remain- 

 ing families there are three modes of superior attachment. In most of 

 the Iguania and Acrodonta it reaches the parietal and does not touch 

 tlie short petrosal. In the other superfamilies it is in contact with the 

 ] tetrosal. In the Varanidie, Helodermidie, and most Anguida? it reaches 

 the parietal, which does not meet it with a conspicuous descending 

 process. In Scincidu' and Teiid;^; a conspicuous descending process 

 meets it. In a certain number of genera of various families it does not 

 quite reach the parietal. Such are Eublepharidjie, Gerrhosaurid;^, 

 Anguis, Lacerta, Phrynosoma (where it rests on the arcade of the 



'Transactious Royal Society, 1879, p. 605, on the Development of the Skull in 

 Lacertilia. 



