EASTMAN: THE DINICHTHYIDS. 25 



associated or not with other remaius cannot now be ascertained. Theo- 

 retical considerations are certainly opposed to the idea that this plate 

 pertained to either D. terreUi or D. ingens ; and we are compelled to 

 regard it as indicating an as yet unknown Dinichthyid species. 



Fused Median Ventrals. — Under this head must be placed the two 

 examples already referred to, that have been described by Newberry ^ 

 and Dean.^ The originals are preserved in the School of Mines Cabinet 

 at Columbia University, and have been determined as D. terrelli and 

 D. (1) gouldi. Whether fusion took place as a strictly adaptive chai-ac- 

 ter in forms iiaving a thin plastron, whether it occurred only in adult 

 individuals, or whether it characterized all the individuals belonging to 

 particular species, are questions for future discoveries to determine. 

 That fusion did not exist in all species of Dinichthys appears, however, 

 extremely probable. 



Overlapping Median Ventrals. — Species which have the postero-ventro- 

 median overlapped by the anterior element represent the normal or 

 primitive condition, as exemplified by the genus Ooccosteus. Three 

 specimens of the detached antero-ventro-median and two of the postero- 

 ventro-median are preserved in the Cambridge collection, whose relations 

 to contiguous plates were plainly those of overlap and underlap. The 

 bone shown in Plate 2, Figs. 5, 6, exhibits such a striking resemblance 

 to its homologue in D. terrelli, that there can be no doubt as to its iden- 

 tity. It is evident that the plate under discussion is entire, since its mar- 

 gins taper gradually to a thin edge, and show no signs of having been 

 broken away from a lower portion. Hence, the only important difference 

 that is to be observed between this specimen and D. terrelli relates to the 

 mode of union with the posterior element ; in the present case it overlaps, 

 iu D. terrelli it is fused with the hinder piece. As we know of no other 

 species to which it can be referred, we must include it, provisionally at 

 least, under the last named species. 



The special characters of this plate have been described elsewhere, 

 although at that time the specimen was supposed to belong to Titan- 

 icJithys. It may be remarked in passing that the semicircular flange 

 forming the anterior margin (seen best on the ventral aspect) is contin- 

 uous with similar compressed borders on the antero-ventro-laterals. 

 None of these margins reveal any trace of plates overlapping them in 



1 Report Geol. Survey of Ohio, Vol. II. Part. II. (Palaeontology), pp. 10, 31, 

 and Chart VI. Fig. A. 



2 Trans. N. Y. Acad. Science, Vol. XV. pp. 157-163, 1896; Ibid., Vol. XVI. 

 pp. 57-60, 1897. 



