216 bulletin: museum of comparative zoology. 



the relations of these plates have not been disturbed ; and this inference 

 is made by him the determining factor of his restoration (Fig. B), the 

 starting-point of his theory of rotary jaw-movements, and the key-note 

 to a novel interpretation of Arthrodires. One perceives, accordingly, 

 that extremely weighty conclusions depend upon Dean's initial assump- 

 tion, the truth or error of which requires to be demonstrated. Suffi- 

 cient reason for distrusting its correctness is found in the improbability 

 of the conclusions resting upon it, the more important of which are 

 contrary to analogy. We may therefore profitably inquire into the 

 reasonableness of the author's initial assumption, and ascertain, if 

 possible, to what extent it invites confidence. 



The one clearly demonstrable feature of the single specimen of 

 Mylostoma studied by Dean is that the right and left halves of the 

 palatal dentition have become separated ; and that, although the com- 

 ponents of either half remain in association, they are dissimilarly oriented. 

 This state of affairs permits of three possible explanations, which may 

 be stated as follows : — 



1. The two right-hand palatal plates have preserved their natural 

 orientation with respect to each other, and the two left-hand plates 

 alone have become disarranged. 



2. The two left-hand palatal plates retain their natural orientation 

 with respect to each other, and the two right-hand plates have become 

 disarranged. 



3. The components of both halves of the palatal dentition have 

 become turned about, so that none of them any longer occupy their 

 original position with respect to one another. 



Circumstantial evidence is our only resource for determining which 

 one of these conclusions is correct. A strong point in favor of the first 

 is the neat adjustment between the contact edges of the plates, which 

 are preserved in direct apposition. At the same time, the close fit 

 observed between the two nearly straight edges cannot be regarded as 

 really decisive proof, owing to the possibility of its being the result 

 of chance. The two right-hand palatal plates may or may not be 

 retained in natural position ; the only test that can be absolutely relied 

 on for determining their arrangement, the real experimentum ciiicis, 

 consists in bringing the functional surfaces of the two upper dental 

 plates into harmonious adjustment with the mandibular, so that a 

 number of diverse features of both upper and lower dental plates shall 

 stand in reciprocal relations. When tubercles are found to fit into 

 grooves or pits, eminences into depressions, and marks of wear to coin- 



