TEMPORAL ARCHES OF THE REPTILIA. 25 



that it is impossible to get over the fact that the pteiyoo-quadrato 

 arch in the lizard is developed exactly as in 8'phenodo^i., as 1 showed 

 many years ngo, and that the immediate ancestor of the lizard mvist 

 have had a fixed quadrate. The lizard still has a lower temporal 

 fossa, which only differs from that of Sphcnodon or Youngina 

 in having become ligamentous when the quadrato-jugnl was 

 lost and the quadrate became movable. The two post-temporal 

 bones are manifestly homologous with the two in YoK.ngina, and 

 are thus the squamosal and the tabular. Elsewhere * I am giving 

 further evidence in confirmation of this view from the consider- 

 ation of the facts revealed by the study of the skeletogenesis. 



The lizards are interesting, as showing in many cases how the 

 upper temporal fossa of the ancestor may disappear by the 

 approximation of the bones, but outside the Lacei'tilia I know of 

 no forms where this has manifestly happened. 



If the interpretations of the temporal regions given above be 

 agreed to, we can very satisfactorily divide the Reptilia into 

 four subclasses : — 



I. We have the piimitive types wdth a roofed temporal region 

 wdiich are usually grouped as Cotylosaurs. The Cotylosauria. can 

 hardly be considered as an order. For convenience, every reptile 

 with a roofed skull is at present placed in it, but such a type as 

 Frocolophon has really very little in common with, say, Diadectes 

 or Paniylus, and still less with Seymouria. If we still wish to 

 keep the roofed forms together Ave must regard them as forming 

 a subclass and not an order, and Williston's term Anapsida 

 seems a very appropriate one, excluding from it tlie Chelonia. 



II. The DiAPSiDA of Osborn is certainly a very natui'al group. 

 Starting in the Up];)er Permian w^e have the primitive Eosiochia, 

 from which we can easily derive all the later two-arched forms, 

 and also the Lacertilia and the Thalattosauria . 



III. The Synapstda of Osborn, restricted to the Mammal-like 

 reptiles with a lower temporal fossa, also forms a thoroughly 

 natural third subclass. 



lY. We are then left with the Mesosauria, the Ichthyosauria, 

 the Plesiosauria, the Placodontia, and the Chelonia. Can these 

 be placed in a fourth natural subclass ? Yery early the reptiles 

 divided into two distinct lines — those that by a lizard- like habit 

 early lost the true coracoid and became possessed of a supra- 

 temporal fossa, and those that by a more Mammal-like habit 

 retained both coracoid and precoracoid and developed an infra- 

 temporal fossa. Until the Upper Permian times, when the 

 two-ai^ched reptiles make their appearance, all reptiles belonged 

 to the Synapsida, the Anapsida, or to this other proposed sub- 

 class. Unfortunately, very few of the land-types are known. 

 Doubtless they became early extinct in the struggle with the 



* Williston Memorial Volume. 



