EASTMAN: MYLOSTOMID PALATAL DENTAL PLATES. 267 



halves of the inferior dentition, that is, the right and left mandibular 

 dental plates, it must needs have been both median and anterior in 

 position. This initial step in assigning the position of the plate in 

 question is clearly imperative. 



Secondly, it is obvious that whatever other elements may have been 

 concerned in the formation of the upper dentition of Dinognathus, they 

 must have been disposed so as to function against the remaining posterior 

 portion of the opposed mandibular dental plates ; in effect, against all 

 that portion of the inferior dentition which is shown by marks of wear 

 and other features to have reciprocated with the upper, but whose spatial 

 limits extend beyond the area of interplay with the compound median 

 anterior element. The approximate form and size of this second poste- 

 riorly placed pair of palatal dental plates are therefore predetermined in 

 advance of their actual discovery by our knowledge of the conformation 

 of the mandibular dental plate in typical Mylostomids, and again, in still 

 another way, by analogy with the hinder pair of palatal plates such as 

 are already known in other species and genera. 



Bearing these considerations in mind, one readily perceives that of the 

 two pairs of plates constituting the upper pavement dentition of the 

 type species of Mylo'stoma, only the narrow and elongate anterior pair 

 is capable of being homologized with the similarly situated compound 

 plate in Dinognathus, which latter, for convenience of distinction in the 

 present discussion, may be referred to as the " ferox " type of Mylosto- 

 mid palatal plate. For it is clear that only the members of this particu- 

 lar pair admit of being symmetrically arranged on either side of the 

 median axis so as to present even a remotely similar configuration, 

 whereas the combined outline of the pair chosen for comparison betrays 

 essentially the same pattern and proportions as are to be observed in the 

 "ferox" type of plate. Manifestly, the resemblance is too intimate to 

 be ascribed to a chance coincidence. And finally, any attempted inter- 

 change of the two sets of pavement plates which we have uniformly des- 

 ignated as anterior and posterior pairs, both in this and former articles 

 (Bull. Mus. Comp. ZooL, 1906, 1907, 50, Nos. 1, 7), instantly throws 

 their functional surfaces out of adjustment for interplay with the lower 

 dentition. 



Hence it follows that the new "ferox" type of Mylostoniid dental 

 plate serves as a valuable criterion for fixing the precise homologies of 

 the two pairs of pavement plates in the type species whose relations have 

 been in dispute, and for confirming the validity of that reconstruction 

 of the palatal dentition which is represented in Figure 3. To judge of 



