MORPHOLOGICAI, REVISION OF THE SUBORDER. 103 



exceeding 2 mm., even in the largest specimen, while the edge of the bone carrying 

 the teeth may reach a thickness of 2 and 3 centimeters. The thinness of the upper 

 portion of the maxillary is shared by the adjacent bones, the nasals, prefontal, jugal, 

 and lachrymal, so that this part of the skull is almost always shattered in the processes 

 of fossilization and lost. The lower edge of the bone is very abruptly widened into 

 a thick dentigerous border (figs. 30 and 31 and plate 10, fig. 4), which is in strong 

 contrast to the weak upper portion of the facial region. The width of this border is 

 greatest opposite the enlarged canine near the anterior end of the maxillary and 

 decreases in width toward the posterior end of the bone as the teeth become smaller. 

 In the diastemal notch there seems to be no great widening of the edge, even in the 

 fonns where teeth are present in the notch. The posterior end of the bone articu- 

 lates with the jugal, as described above. The outer surface is marked on the lower 

 edge with pits and rugosities. 



The teeth are lenticular in fonn, with distinct fore and aft cutting edges, which 

 are strongly serrate. The roots of the teeth are implanted in distinct sockets which 

 may reach a depth as great as the length of the tooth beyond the outer edge of the 

 bone (plate 14, fig. 8). The outer edge of the bone extends much farther down than 

 the inner, so that a good bit of the length of the tooth after it leaves the socket rests 

 against this edge. The root of the tooth is hollow and its inner end is open, so that 

 it is evident that the teeth were replaced by absorption of the root and continued 

 growth of new teeth. This process is seen in actual progress in several specimens. 

 One peculiarity of the deutition in the family Clepsydropidce is very striking. The 

 maxillary has always two alveoli for enlarged canine teeth, but in only one or two 

 specimens of the many preserved are there two enlarged teeth present in the maxillary ; 

 in all the others there is one tooth functional and either an empty alveolus or a tooth 

 just beginning to appear from the socket. It seems that the two large canines of each 

 side were alternately functional, one growing to full size and function as the other fell 

 or was pushed out by the successional tooth. It is as often the anterior as the posterior 

 tooth which is lost, and the anterior is often present on one side of the jaw while 

 the posterior is present on the other side. Thus the animal was never deprived of its 

 powerful tusks for any considerable interval. 



The number of the teeth seems to be variable in the different specimens, but 

 was never far from 20 in the maxillary and 28 in the lower jaw. 



The traiiruerse is seen in specimens Nos. looi and 1002 University of Chicago. 

 On the inner side of the jugal, as described above and shown in fig. 30, a strong ridge 

 extends forward and receives into its bifurcated end the upper end of the transverse. 

 From this point the transverse extends straight downward on the anterior and outer 

 face of the outer process of the pterygoid; its lower edge fuses with the pterygoid so 

 that it is impossible to describe its lower limit exactly, but it does not extend very far 

 down on the pterygoid. The anterior edge of the transverse unites with the posterior 

 end of the maxillary, so that it is held firmly in its position. 



TlxG: pterygoid^ as repeatedly described, has a distinct tripartite form, consisting 

 of an anterior horizontal portion, a medium vertical process, and a posterior portion 

 which joins the quadrate. The form of the bone is best shown in plate 1 1 , figs. 5 and 6, 

 which are from specimen i University of Chicago. 



The anterior plate is separated from the maxillary by the palatine and the trans- 

 verse. These bones join the pterygoid directly, so that there are no palatine vacuities 

 in the posterior part of the palate. The anterior processes come ver>' close together 

 in the median line, but it is impossible to say whether they are united throughout 

 their length or not. It seems probable that there was a space between the posterior 



