I30 THE PERMO-CARBONIFEROUS RED BEDS OF 



ning stages of armor would prevent injury to the spine on many occasions. 

 The possibiUty that the armor may be considered as an evidence of a ter- 

 restrial habit has been noted above. 



Abel, in his Paleobiologie, has suggested the correlation of dorsal armor 

 with fossorial habits in extinct amphibians and reptiles, but it could not be 

 the case with Cacops, which was devoid of strong claws. 



Its habits were possibly similar to those of the living frogs, lurking in the 

 vegetation on the banks of bodies of water, sometimes wandering inland, but 

 always ready to make a quick rush to the protection of the water when danger 

 threatened. The total length did not exceed 45 centimeters. 



Aspidosaurus (fig. 18). — This genus did not develop a dermal armor, but 

 the distal ends of the netiral spines were expanded to a greater or less degree, 

 and undoubtedly furnished some protection to the animal. It is not probable 

 that the bony process projected for any distance above the line of back (un- 

 less it were in the case of Aspidosaurus crucifer). Any attempt to restore 



Fig. 18. — Restoration of Aspidosaurus chiton Broili. Probably about one-fourth natural size. 



these animals would be almost wholly conjectural. A. novomexicanus had 

 a skull similar in general proportions to that of the members of the Dissoro- 

 phidce, but nothing more can be said with certainty. We can only imagine 

 it as a stegocephalian, with the general form of Cacops or Dissorophus, devoid 

 of any armor, but perhaps with a row of projecting knobs on the back. 



Broiliellus. — In a recent paper Williston '^ has described a new amphibian, 

 Broiliellus texensis, from the Clear Fork beds of Texas. In this animal there 

 was a single row of dorsal dermal plates which were entirely free from the 

 ends of the neural spines, and the ends of the neural spines were not expanded. 

 This is the third type of armored amphibian in the fauna. Dissorophus has 

 the extremities of 'the neural spines expanded into overarching plates, and 

 this is overlain by dermal plates of similar form. Aspidosaurus has the ends 

 of the neural spines expanded without overlying plates. Broiliellus has the 

 dermal plates without any expansion of the end of the neural spines. Willis- 

 ton thinks that the expanded ends of the neural spines in Aspidosaurus are 

 really dermal plates which have become secondarily coossified with the neural 

 spines. The author has not held this idea and can not now accept it for all 

 cases. The only evidence he has that such an action may have taken place 

 is that in the Brier Creek Bone Bed he found a V-shaped dermal plate very 

 similar in appearance to the terminal process of an Aspidosaurus, but which 



» Williston, Jour. Geol., vol. xxii, No. i, pp. 49-56, 1914. 



