SYSTEMATIC RELATIONS OF FISHES. 583 
correct who regard the fishes as representing variously, from two to 
four classes. One of these (the Ganoidea), having been already 
disposed of, it remains to consider the claims of the remainder, 
viz: The Elasmobranchii (sharks), Dipnoi and typical fishes. 
If we examine the points in which the whole taken together dif- 
fer from the Batrachia and other classes above it, we find that it 
is confined chiefly to the structure of the limbs and the hyoid ap- 
paratus. The typical fishes present, however, other important 
peculiarities, viz.: 1st. The existence of two or three distinct 
bones in the suspensor of the mandible, instead of one. a 
attachment to these of the opercular bones. 3d. The i of 
pelvic bones. 4th. The suspension of the scapular arch to the 
cranium. th. The large development of the pterotic (Parker, 
mastoid of Cuvier and Owen) is characteristic of bony fishes. 
The types of variation, in the first point, only distinguish groups 
of subordinate rank. Thus: the suspensor of the mandible in the 
typical fishes consists of the hyomandibular stapes, quadrate (met- 
apterygoid or incus), symplectic and mesopterygoid (quadratoju- 
gal Müller, quadrate Huxley, Elm. Comp. Anat.). In the Mormy- 
ride, Siluride, Polypteride, and others, the symplectic is absent ; 
in the eels of several families both it and the metapterygoid are 
wanting, reducing the suspensorium to arod of two pieces. This 
condition exists in many of the rays; in others, and in the sharks, 
the inferior element is wanting (Miiller, Stannius). An important 
modification is exhibited by Chimera, where the hyomandibular, 
which alone exists, is continuous with the cartilaginous cranium, 
not being separated by the usual articulation. 
As to the opercular bones, all are wanting in the Elasmobranchs 
(sharks and rays) while the typical fishes possess four, viz: preo- 
perculum, operculum, suboperculum and interoperculum. In 
many of these, however, the suboperculum is wanting, and in the 
sturgeons and many eels there is no preoperculum. In Polyodon 
the interoperculum is also wanting. In Lepidosiren the operculum 
and interoperculum are rudimental. In respect to this point also 
the divisions indicated are of subordinate value. As regards the 
development of the pterotic bone, its history is not yet sufficiently 
made out to enable us to understand its value. It does not exist 
in those with cartilaginous cranium (Elasmobranchii). The Elas- 
mobranchs are well known to have the scapular arch suspended 
freely behind the cranium as in higher vertebrates. It is not 
