1S92.] loJ [Ciipe. 



]iil;v and Ainplusbaniia. In the Gecconidtt; it does not exist, nor is tlie 

 petiosiil notched by tlie foramen, while the anterior border of tlie petrosal 

 forms a free crest wliich extends from above downwards and backwards. 

 In the Chamceleonidae, Agamidse, Iguanidoe and Anolidse (families with 

 papillose tongues), the supraforaminal part of the petrosal is sliort and is 

 bounded by a convex anterior border which marks the position of the an- 

 terior semicircular canal. In the Nyctisaura, Thecaglossa, Diploglossa, 

 Leptoglossa and Amphisbfenia (families with smooth or squamous 

 tongues, except Anniella, Diploglossa and Nyctisaura) the petrosal is 

 produced beyond this curved border below the parietal. In many 

 foi'ms an outline of the semicircular canal, which forms the boun- 

 dary iu the other superfamilies, may be traced, w^hence I have termed 

 this part of the petrosal posterior to it "the arched body" in my former 

 system of the Lacertilia.* The petrosal is produced furthest beyond this 

 arcade in the snakelike forms of the Anniella and Amphisbsenia, reach- 

 ing almost to the orbit in Lepidosteruum. The relation to the parie- 

 tal difiers, the differences resulting from the greater or less reduction of 

 the primitive supratemporal roof and the greater or less entrance of the 

 parietal into the lateral wall of the brain case. In most of the families 

 it is little or not decurved to meet the petrosal ; and in the Tguania, where 

 it is decurved, it does not come iu contact with the petrosal owing to the 

 shortness of the latter. In certain families where the petrosal is produced 

 beyond the arcade, and the parietal is decurved, the two elements are in 

 contact for a short distance, as in the Varanidoe. In the Tiidoe and Scin- 

 cida? the contact is mainly effected by a short descending process of the 

 parietal. This process is especially elongate in the Scincid*. The arcade 

 is the anterior border of the petrosal in the Permian Theriodonta, and it 

 marks the position of the anterior semicircular canal. The membranous 

 wall of the brain case, anterior to the petrosal, contains an ossification 

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

 extremit}^ the frontal, and might hence be supposed to be the orbito- 

 sphenoid ; but this homology is vitiated by the fact that its inferior por- 

 tion passes behind the optic foramen. The latter position is that of the 

 alisphenoid, and so the bone is named by Parker.f But there is another 

 element, the epiplerygoid, posterior to it and immediately anterior to the 

 petrosal, which has been supposed to be the true alisphenoid. Leaving 

 this question, and adopting for the bone in question the provisional name 

 of postoptic, I remark that is typically triradiate, sending two branches 

 upwards and one downwards. This is its character in Agamida;, Varanidae 

 and Tiidaj. The posterior superior branch is much reduced in many Ig- 

 uanidae and Lacertidte and in some Agamidae (Megalochilus), and it is ab- 

 solutely wanting in Gerrhosaurus and Chamajleon. There is nopostoptic 

 in Heloderma. In the Rhynchocephalian genus Sphenodon these two 

 elements coexist with an orbitosphenoid, Ijing between the optic and tri- 



* Proceeds. Academy Philadelphia, 1864, p. 224. 



tTransac. lloyal Society, 1879, p. G05,on the "Development of the Skull in Lacertilia." 



