AND ON THE CLASSIFICATION OF FISH. 519 
Rhinellus, are Ganoids or not, this question has no more influence 
on the geognostic deductions than the separation of the false 
Ganoids, viz. the Plectognathi and Lophobranchii; for, after all, 
it refers to fishes which are subsequent to the Jura formation: 
the hitherto assumed relations of the fish to the age of the 
formations are not thereby affected. Agassiz has advanced the 
opinion that the Ganoids predominate in the older formations, 
that, leaving out the Placoids, all the fish prior to the Chalk 
formation are Ganoids, and that true osseous fishes commence 
with the Chalk. This conclusion is not in the least shaken, and 
must be considered as approximating to the truth. But the 
state of preservation of fossils makes us feel that much is want- 
ing, in particular instances, for the proof to be absolute. The 
deductions as to the relation of the Ganoids to the formations 
are merely changed by our investigations in reference to the 
formations beginning with the Chalk, and the development of 
the Ganoids in all newer formations, as well as in the existing 
creation, is considerably reduced by the removal of the foreign 
families. 
With existing fish we can decide with absolute certainty, from 
their anatomy, whether they are Ganoids or not. But what 
characters are to guide us with respect to fossil fish? In the 
first place, we have enamelled rhomboidal scales articulated with 
each other by appendices and arranged in oblique rows, spinous 
plates (fulcra, Agass.) on the anterior border of one or more fins, 
heterocercy in a fish furnished with an operculum, an abdomi- 
nal position of the ventral fins and soft articulated fin-rays. 
Where fulcra exist on the borders of the fins, I consider the 
ganoid nature of a fossil as decided, whatever the form of the 
scales, for this character is not found in any other fish. The 
perfect heterocercy in a fish having opercular and cephalic bones 
is equally decisive, for these are not found except in the Plagio- 
stomi. The covering of the margin of the fin with fulcra is widely 
diffused, and may sometimes be overlooked, although present. I 
find it among several specimens of the great Pachycormus macro- 
pterus of the Lias slate, in one instance perfectly distinct, both on 
the dorsal fin and the anal fin; whilst on the tail fin it is absent 
in all. However, in many genera the fulcra appear to be en- 
tirely wanting; and that this is both possible, and does actually 
occur, we find a distinct proof among those at present existing 
in Polypterus and Polyodon. Although the vertebral column of 
