BROILIELLUS, A NEW GENUS OF AMPHIBIANS 53 



been described and figured by Watson, and before him by Broili 

 in Cochleosaurus. 



The palate has been exposed only so far as is possible in each 

 specimen with the mandibles tightly closed and the clavicular 

 girdle in position. Doubtless the structure throughout is like 

 that of Cacops,^ but the nares and all evidences of enlarged teeth are 

 concealed. The interpterygoidal vacuities are very large, though 

 less elongated than in Cacops. Sutural lines for the vomers are 

 apparent, as I have figured them. The whole surface in front, 

 between the mandibles, is covered with minute chagrin-like teeth. 



Indications of the stapes, as in Cacops, are present, as also the 

 sutural division between the parasphenoid and the exoccipitals. 



The sutures of the mandibles, in their closed condition, are not 

 distinguishable. The teeth are shown in both specimens. They 

 are small, pointed cones, of nearly uniform size, throughout. 



DORSAL CARAPACE 



The general shape of the dorsal shields is shown sufficiently 

 well in the photograph of the larger specimen. In this specimen 

 the vertebrae, with two exceptions, back of the clavicular girdle 

 had been separated and lost before fossilization, as has been demon- 

 strated by excavating the under side in the middle. In the smaller 

 specimen, the vertebrae are all in position as far as the hind end of 

 the carapace, though the last two or three are somewhat dis- 

 arranged. Furthermore, in the smaller specimen, several of the 

 shields have been cleanly removed from the matrix, proving that 

 they had no connection whatever with the underlying spines; 

 indeed they lie some distance above the vertebrae, with the 

 mat];ix intervening. The plates in this specimen correspond in 

 number with the vertebrae below, that is, each vertebra corresponds 

 to a single plate and not to two as in Cacops and Dissorophus. 

 The plates are not of uniform width anteroposteriorly; the third, 

 fifth, and seventh at least are narrower than the intervening ones, 

 which suggested at first that each vertebra had two plates, but this 

 is positively not the case; all of which goes to prove that the plates 

 were entirely distinct from the spines. Indeed the spines, so far 



' Willis ton, Bull. Geol. Soc. Am., 249, 1910. 



