€0 TRANSACTIONS OF THE [nOV. 16, 



Diplognathus and Mylostoma preserved either well marked dermal 

 tubercles or conical mandibular teeth or both. But unfortunately 

 a definite knowledge of the dermal plates of our American Cocco- 

 steids is still to be sought. And we might, therefore conclude 

 that the plastron of Mr. Eastman, if it cannot be definitely 

 identified, might well have belonged to some Arthrodiran other 

 than Dinichthys. It would certainly be more natural to assume 

 that this was the case than to believe that the ventro-median 

 plates could be both fused and separate within the limits of a 

 single genus. 



The ventro-median plate of D. terrelli deserves, in closing, a 

 brief description. As shown in PI. III., it will be seen to corre- 

 spond closely in outline to that of D. govldi, PI. II., VM., the 

 figure at the right in PI. III. showing the corresponding aspect, 

 the ventral (ectal) one. It is, however, evident that the plate is 

 relatively a heavier one in the larger species, and that its ante- 

 rior and posterior elements correspond more nearly' in size. From 

 the present fossil the plate is evidently a single one, its anterior 

 and posterior elements having fused in the narrow part. This 

 condition can be satisfactory made out, especially on the visceral 

 side of the plate where the nutritive canals can be clearly fol- 

 lowed, and no trace can be detected even of a sutural joint. 



A comparison of the ventral aspect of the plate in the two 

 species shows that in D. teyn^elli the marginal .overlap in the 

 hinder region was considerably wider and deeper, the median 

 portion, which was exposed at the surface, rising somewhat ab- 

 ruptly. This portion shows less prominently the divergence of 

 the vascular canals of the smaller species. And the margin of 

 the overlap is smoother, i. e., lacking the longitudinal strife of 

 D. gouldii. At the anterior end of the plate the beveling of the 

 rounded margin is also apparent. 



From the conditions in D. terrelli it appears that the ventro- 

 median element was overlapped on its ventral face by all mar- 

 gining plates, and the shape of the bevel of its rounded anterior 

 margin suggests closely the condition in Coccosteus, where the 

 inferior laterals (Traquair) take their support. 



The plate's visceral aspect is interesting on account of the 

 transvei'se ridge which marks the anterior enlargement of both 

 its component elements. The one in front is the better shown, 

 and from its median point a slight sagittal ridge passes back- 

 ward : this character, but fainter, can also be traced in the 

 hinder element of the plate. The writer notes that these char- 

 acters occur as well in other Coccosteids of the Cleveland shale 

 in which the ventro-median plates are separate. 



As far, finally, as the fusion of the elements of the ventre- 



