DEVONIAN FISHES OF IOWA 



181 



though marginal serrations have disappeared,* the divided ridge 

 which is situated close to the inner margin is perhaps to be 

 regarded as a relic of one of the most persistent features of 

 Dipnoan dentition. As for the upper dental plates, had they 

 invariably been found in the detached condition, and were we 

 ignorant of their association with typical Arthrodiran man- 

 dibles, they would be unhestitatingly identified with the Cteno- 

 dipterine order of Dipnoans. That these plates were supported 

 by cartilage forming the roof of the mouth is distinctly apparent 

 from their rugose, slightly hollowed upper surface, and out- 

 wardly bevelled edges ; and the contour of the hinder pair ren- 

 ders it extremely probable, at least, that the supporting palato- 



FlG. 25 



Fig. 25. Mylostoma variabile Newt). Cleveland shale; Cleveland, Ohio. Restoration show- 

 ing two pairs of palato-pterygoid dental plates arranged in their inferred normal position, 

 and outlines of mandibular plates functioning against them, all drawn from a single, 

 nearly complete and probably young individual. 



*They are prominently retained in Diplognathus, however, a genus that belongs 

 indubitably to the same family as Mylostoma, and are developed in the form of 

 powerful denticles entirely around the margin of the dental plate proper. Judging 

 from the relative proportions of plates forming the ventral armor in the imperfectly 

 known genera Glyptaspis and Holonema, they should properly be included among 

 Mylostomids, and it is extremely likely that they were provided with a crushing 

 type of dentition, or modification of that type. Diplognathus is really only an 

 extreme modification of the Mylostomid type, and, as suggested by Newberry, 

 there is considerable reason to suppose that the mandibles known under this name, 

 and the detached abdominal plates assigned to the provisional genus Glyptaspis, 

 belonged in reality to the same kind of fish. 



