160 TRANSACTIONS OF THE [May 11, 
ing in shagreen-like structures, thus giving ground for the belief 
that these plates were deeply embedded in the ventral integument. 
The Ventro- Median plate, Pl. VII., VM, has been retained in a 
fair degree of preservation. Like the neighboring plates, if ex- 
hibits its visceral aspect, and has been but little displaced. Par- 
allel to its margins are well-marked striz, and a series of 
straight vascular tubules converge toward its center. The 
plate’s anterior end terminates in an enlarged, flat, rounded 
area, which appears to be the homologue of a separate plate in 
Coccosteus, the ‘‘anterior median ventral” of Traquair. In 
Dinichthys, however, this element is relatively much smaller in 
size. In the present specimen its surface has been greatly 
weathered. 
Two additional dermal plates are shown in the fossil, Pl. VIL., 
U. They appear to represent a pair of plates hitherto unde- 
scribed in Dinichthys. Of their relations little is suggested, 
although it is possible that they represent the antero-lateral 
plates of the shoulder armoring of Coccosteus. In the neigh- 
borhood of these plates the fossil has retained a large pale- 
colored mass, amorphous, ragged in fracture, which in character 
and position may well represent a coprolite. It is lacking, how- 
ever, in fish fragments, and, if a coprolite, it affords no 
evidence of the presence in Dinichthys of an intestinal spiral 
valve. 
Comparison of the ventral armor of Dinichthys gouldi with 
that of D. terrelli, as described by Wright, gives evidence that 
the conditions in the smaller species were by far the more 
generalized. (cf. Pl. VIII., Figs. 2 and 3.) 
Thus, in D. gouldi, Pl. VIII., Fig. 2. 
1. The anterior and posterior ventro-lateral plates correspond 
closely to each other in size. 
2. The anterior ventro-laterals were simple oblong plates, 
comparatively uniform in thickness, and straight margined, 7. e., 
not crescent-shaped as in D. terrelli. 
3. The relation of the anterior ventro-laterals to the poste- 
rior ventro-laterals was a simple one,the former overlapping the 
latter broadly and apparently loosely ; there is no evidence that 
the overlapping portion of the anterior plates were fitted into 
an insunken area in the hinder plates. 
4. The posterior ventro-laterals were simple oblong plates ; 
their hinder median margins overlapping ‘(the right possibly 
overlapping the left, unlike in Coccosteus), but showing no trace 
of the peg-and-socket suture by which they are united in D. 
terrelli. 
5. The outer surface of the plates (posterior ventro-laterals, 
