.MKKKIAM Tin: Til A I, ATTOSA UKl A 11 



The ixtrictuls art- iiiurli smaller than the t'ruiitals ami an- also scjuirated i)y a 

 miMliaii suture. A large parietal foramen lies between the two elements, hut is 

 a considerable distance behind the coronal suture. The posterioi- outei- angles 

 of the parietals are produced backward around the outer side of the superior 

 temporal openings for more than half the length of the upper temporal arch. 



In tigui'e 2a, Plate i\', what appears to be the squamosal is seen to reach 

 forward over the post-fronto-orbital to touch the frontal. If this is the normal 

 relation of these (di'ments, the outer border of the parietal is separated from the 

 post-fronto-orbital. The median portion of the parietal is also })roduced back- 

 ward as a slender arm and was probably in contact with the squamosal. The 

 two posterior arms seem almost to have surrounded the upper temporal opening, 

 giving to the skull an altogt'ther peculiar character in this region. 



Tcinporal Reginu. — The skulls appear to be characterized by the presence of 

 both superior and lateral temporal openings. The ■mperior openings are con- 

 siderablv smaller than the lateral. .ludgingfrom the specimen seen in I'late iii, 

 the posterior ends of the pai'ietals wei'e very slender and were dropped some 

 distance bcdow the level of the roof of the skull, so that the superior oj)enings 

 faced backward somewhat, as in Belodaii. The anterioi- half or two thirds of this 

 opening was enclosed by the parietal. The posterior boundaries have not been 

 seen. They were probably formed by connection of the parietal and squamosal, 

 though the supraoccipital may have intervened as seems to have been the case 

 in SdH ra iindiDi. 



The rehitivtdy large hilcral fi'iiijionil iipoiiiKji^ are seen in two specimens 

 (IM. Ill, tig. 2 and 1*1. iv, tig. 'l(i). The boundaries are not perfect, even in the 

 more nearly complete specimens, owing to the absence of the quadratojugal. In 

 both instances, however, the jugal sends backward a long and strong process 

 extending more than half the distance to the ([uadrate, and in one specimen the 

 posterior end of this process is roughened as if fi'om contact with a (juailrato- 

 jugal. Under these circumstances it would lie ditHcult to believe that the 

 iiifcriar liar was incomplete. Should we assume that it was not coin])lete it would 

 l)e necessary to supfiose that it is caught, as it were, in thi' process of breaking 

 down, as no form losing the lower i)ar retains a posterior jugal process similar 

 to that shown here, lii reality there is hardly reason for supposing that the 

 lower bar was even weak, as the posterior process sent out from the jugal is not 

 stronger than it is here in many forms in which this arch is t-onsidcred w(dl 

 developed. 



The iipjX'r fciiipiirdl Ixir is certainly made u]) to a gi'eat extent (.)f the large 

 post-fronto-orbital and the s(Hiamosal. as seen in IMate IV. ligure 2(i. with the 

 addition of the outer posterior pi'ocess of the parietal lying on the me(lian side 



