CLASSIFICATION OF DEVONIAN FISHES 449 
5. The TELEOSTEI have hitherto been supposed to be entirely ab- 
sent from formations of Paleozoic age, and no doubt they do not exist 
under those forms which are most familiar to ichthyologists acquainted 
with marine fishes, or with the fresh-water fishes of temperate climates ; 
but, nevertheless, I shall now endeavour to show that there are grounds 
for something more than a suspension of judgment, as to the validity 
of the ordinary doctrines held upon this subject. 
The remarkable genera Coccosteus and Pterichthys are those which, 
among all Devonian fishes, have been by common consent regarded 
as the most heteroclite and anomalous, some writers having gone so 
far, in fact, as to imagine that these hard cased vertebrates offered us 
a transition to the shelled Invertebrata. 
Nevertheless, I trust I shall be able to show that the one of these 
two closely allied genera—Coccosteus—is best, indeed, I may say only, 
to be understood, by comparing its bony shields with those which 
cover the roof of the cranium and the anterior part of the body of 
certain existing Siluroid Teleosteans, 
To this end, however, I must first give the conception of the 
structure of Coccosteus which my own investigations, guided by those 
of my predecessors, Agassiz, Miller, Egerton, and Pander,! have led 
me to form. 
The superior wall of the skull only, seems to have been ossified in 
this fish, and forms a great shield, which may be roughly said to have 
a hexagonal figure. The posterior and postero-lateral sides of the 
hexagon are pretty nearly straight lines, while the anterior side is 
rounded off, to form the snout, and the antero-lateral sides, the long- 
est of all, have their anterior moieties deeply excavated, to constitute 
the upper part of the walls of the orbit. From before backwards, in 
the median line, the contour of the cranial shield is nearly straight, 
but from side to side it is convex, in consequence, more particularly, 
of the downward inflexion of its postero-lateral angles. The sutures, 
which separate the various constituent bones of the skull, may readily 
be confounded with certain superficial grooves of a totally different 
import, but, by grinding away the outermost layer of bone, this source 
of error is avoided ; and it is then seen that the cranial sutures have 
the arrangement represented in the woodcut, fig. 19, and define the 
several bones from one another with great sharpness. 
In the middle line, behind, they mark off a symmetrical, trape- 
1 Compare Agassiz, ‘‘ Monog. des Poissons Fossiles du Vieux Grés Rouge ;” H. Miller, 
“Old Red Sandstone” and Quart. Journ. Geol. Soc. 1859 ; Pander, ‘‘ Ueber die Placodermen 
des Devonischen Systems, 1857 ;°? Sir P. Egerton, ‘‘ Remarks on the Nomenclature of the 
Devonian Fishes,” Quart. Journ. Geol. Soc. 1859. 
VOL. II GG 
