TRIONYCHIDJE. 535 



sixth costal plate nearly complete, fragments of several other costals, nearly the whole of the 

 right hyoplastron, the mesial two-thirds of the right hypoplastron, one cervical vertebra, the 

 proximal portion of the right femur, the distal end of the right humerus, and some other, but 

 unimportant, fragments. Whether or not there was a preneural can not yet be determined. 



The individual was a very large one, the plastral bones indicating that the width of the 

 plastron was little short of 600 mm. An estimate, not wholly reliable, made from the costals 

 present, indicates that the length of the carapace was about 500 mm. It is quite probable that 

 the carapace was at least as broad as long. 



The skull was correspondingly long. A comparison of the united quadrate and prootic 

 with those of Platypeltis ferox indicates a length of skull of 225 mm. from snout to occipital 

 condyle, but little less than nine inches. The fragment of maxilla indicates a shorter skull, 

 about 7 inches. It is not improbable that the front of the skull was relatively short. The 

 quadrate and prootic offer no peculiarities when compared with those of Platypeltis ferox, 

 except that in the latter the anterior ends of the borders wh'ich articulate with the parietal and 

 the quadrate respectively rise considerably as they approach these bones, thus making the 

 anterior end of the prootic somewhat gutter-like; whereas, in T. manducans the rise of the 

 prootic on each side is slight and the front border of the bone is quite flat. 



The fragment of the maxillary offers the most interesting modification. Fig. 2 of plate 105 

 represents the bone from the outside; fig. 3 from below. The lower border of the bone forms 

 a prominent cutting-edge, such as we find in many other trionychids. Inside of this cutting- 

 edge there is, in other trionychids, a horizontal masticatory surface which may be very narrow 



FIG. 690. Temnotrwnyx manducans. Plastron of type. X & 



or of moderate width, or very broad; but it is always nearly plane. In T. manducans, on the 

 contrary, this masticatory surface is traverst from one end to the other by a sharp ridge, such 

 as we find in various Emydidae and Testudinidae. The ridge is not at all tootht. Mesiad of it 

 the palatal surface of the bone past inward and somewhat upward to the choanae, and to meet 

 the vomer in front and the palatine behind. The structure of the jaw indicates that the animal 

 was accustomed neither to seizing and swallowing whole a living prey nor to crushing shells 

 and other hard objects; but rather to masticating animal and vegetable substances of a mod- 

 erate hardness. 



The nuchal bone offers one peculiarity. In the midline, on the under side equally removed 

 from the anterior and the posterior borders, is an oval depression 14 mm. long and 10 mm. 

 wide. It probably served as the attachment of a ligament. The bone does not come to an edge 

 in front, but is truncated. The thickness is 16 mm. The antero-posterior extent of the bone 

 was 55 mm. The lateral extent is unknown. There appears to have been a, smooth band 

 along the anterior border, narrow at the midline, while the remainder is covered with shallow 

 pits, 4 or 5 in a 20 mm. line. 



Some of the costals indicate a considerable convexity of the carapace. One fragment from 

 the middle of the length of the shell has a width of 65 mm. and a thickness of 10 mm., near the 

 sutural border. The sixth costal is 51 mm. wide proximally and 95 mm. wide at the free border. 

 Its thickness is 12 mm. at the posterior sutural border. The upper layer of the bone of the 

 costals overhung somewhat the middle layer, so that there was a sharp channel running around 

 the border of the carapace. Notwithstanding the age of the individual the ribs projected con- 

 siderably beyond the borders of the disk, in one case 42 mm. 



