— i:;o 
though it does not permit the cranial sutures to be so 
Fig. 3. precisely defined as in the non- 
enamelled skulls of Saurians, 
_^f- yet manifests indicationsof those 
siitures which are not alto- 
gether devoid of signiflcance. 
The dentiferous part of the 
lower jaw, like the opposed bor- 
, . I der of the upper jaw is «en iine 
• I ligne presque droite,» as in 
' \ \ ' i Bhinosaurus^ and the teeth are 
subequal, numerous, small, slen- 
Petrophryne granuiata. der (fines ), and sharp-polnted. 
They are somewhat closer set than in BMnosaurus: I 
couDt 26 along the side of the upper jaw in which they 
are best preserved (fig, 2). There are 24 of such teeth 
in the corresponding part of the skull of Rhinosatirus 
(loc. cit. p. 6). 
The tympanic elemeut of the temporal, fig. 2, 28, is 
large, convex (*bombe'), produced backward (avancant 
en arriere), and separated above from the raastoidal de- 
ment (ib. 8) of the occiput by a large notch (ib. v). The 
resemblance of the South African fossil to the 3Ioscou 
one (loc. cit. tab. Y) is close and significant in this cha- 
racter. The occiput is similarly broad, low and truncate 
behind. But the character of most importance in deter- 
mining the ordinal affinities of Bhinosaurm could not 
be discerned in that fossil, apparently by reason of the 
close attachment to tbeir matrix of the atlas and other 
cervical vertebrae to the cranium 
*) See the figures in Tab. V and the remark: „— Les vertebres 
attachees aa cräne scnt trop empätees dans la matrice pour en in- 
