218 



bulletin: museum of comparative zoology. 



studied by Dean. The two left-hand pahatal plates being removable 

 from the specimen, they were rearranged in the prescribed manner, and 

 their oral surfaces fitted against the mandibular dental plate. The 

 experiment was also repeated with the aid of plaster casts of the two 

 right-hand palatal plates, ^Yhich M'ere not removable, and in both cases 

 the new arrangement was found competent to explain all the facts in 

 thoroughly satisfactory manner. Its effectiveness will be readily under- 

 stood from inspection of Figure C, drawn from the original specimens, 

 and from the following discussion of details. 



Figure C. 



Proposed reconstruction of Mylostomid dentition, based upon the originals shown in preced- 

 ing text-figures. X J. 



We have already stated that the juxtaposition of dental plates in the 

 nearly complete example of Mylostoma is such as to admit of three 

 possible conclusions, only one of which can be true. The new arrange- 

 ment proceeds to test, and afterwards to affirm the correctness of the 

 second of these conclusions, which are here restated for sake of 

 clearness. 



1. The two right-hand palatal plates are retained in their natural 

 position with respect to each other. (Disproved.) 



2. The two left-hand palatal plates are naturally oriented with re- 



