CLASSIFICATION OF DEVONIAN FISHES. 



18 



ethmoid. If it were amalgamated with the frontals and these with 

 one another, we should have an almost exact reproduction of the 

 anterior cranial buckler of Osteolepis. In a well preserved specimen 

 of the skull of Megalichthys before me, the orbits are small circular 

 cavities, placed at about the junction of the anterior and middle thirds 

 of the head. They are bounded, in front and below, by a small tri- 

 angular bone (like a lachrymal) as in Polypterus ; below, by a small 

 part of a large suborbital bone, whose anterior margin joins the 

 premaxilla and its inferior margin the maxilla ; below and behind, 

 by another suborbital bone, fitted in between the preceding, the 

 maxilla, and a postorbital bone. The maxilla, large and long, is 

 narrow anteriorly, where it abuts upon the bone termed "pre- 

 operculum " by Agassiz ; like the premaxilla, its edges are beset 

 with small teeth. Agassiz says, " Le cote anterieur du mufle est 

 " el^gamment echancre au milieu et renfle en un bee, tres obtus, 

 " qui porte dans notre exemplaire une grosse dent canine ; " and on 

 making a transverse section of a Megalichthys snout I found a 

 median, stout, backwardly projecting ridge of bone, containing two 

 large alveoli, one on each side of the middle line. The one of 

 these alveoli exhibits the section of the base of a large tooth 

 with greatly folded dentine. 



While the exoskeleton of Megalichthys is exceedingly similar to 

 that of Diplopterus and. Osteolepis, the endoskeleton presents a 

 remarkable advance on that of any other Saurodipterine, in that 

 both the centra and the neural arches of the vertebral column are 

 thoroughly ossified. Excellent specimens of these vertebrae are to 

 be seen in the British Museum. 



The Saurodipterini and Glyptodipterini being thus separated 

 from other Palaeozoic fishes, as well-defined families, perfectly 

 distinct from one another, though closely allied by the community 

 of characters displayed in the number, structure and disposition, of 

 their fins, the absence of branchiostegal rays and their replacement 

 by jugular plates ; we have next to consider what other families of 

 fish, if any, should be ranged alongside of them, or in other words, 

 what are the limits and what the importance of the larger group, 

 formed by the association of these families. 



In the first place, I conceive there can be no doubt that the 

 Ctenododipterini, a family justly established by Professor Pander* 

 for the recep^^ion of Dipteriis and its immediate allies, must take 

 its place in close juxtaposition with the Saurodipterini and Glypto- 



* Under the name of Ctenodipterini. Sir Phillip Egerion has, I think, given good 

 reasons for the slight change I have adopted. Vide wfra, p, 55. 



