Lettei' on the Placodermi. 279 



and I am quite uncertain whether they are quoted against my 

 view or in support of it. It reminds one of Cuvier, who, 

 when he first saw some Palaonisci from the Copper slate, said, 

 ' they should be arranged either with the sturgeons among the 

 Chondropterygiiy or with the bony pike among the Clupece, and 

 yet he missed the almost inevitable conclusion which any one 

 might have drawn from his own evidence, it being reserved 

 for his successor in those inquiries to make the most import- 

 ant improvement in his classification of fish by removing the 

 sturgeons from the remainder of the Chondropterygiij with which 

 they had no obvious affinity, and the bony pike from the ClupecBy 

 which group their presence equally disturbed, and uniting them 

 with the aforesaid Palceonisci and the like fossil types to form 

 the distinct order of Ganoids. 



The facts of the present case are these : — Mr. Lyell long ago 

 discovered the peculiar fish which Agassiz called Cephalaspis 

 and made the type of his family Cephalaspides ; it has the 

 head covered by a single bony shield, the body covered with 

 rhomboidal scales of the ordinary Ganoid construction, and has 

 a large heterocercal tail and caudal fin, such as we so commonly 

 see in the Ganoids of the old rocks : in his last book, Agassiz 

 adds to this family the genera Pterichthys, Pamphractus, Poly- 

 phractus and Coccosteus, all of which agree among themselves, 

 and differ from Cephalaspis, in, 1st, having the head covered 

 with several distinct plates instead of one single shield, as in 

 it ; 2nd, in not having the body covered with small rhomboidal 

 scales, as in it, but cased in a few large tuberculated bony plates ; 

 and finally, instead of the large heterocercal tail and distinct 

 caudal fin of that genus, having a straight simply pointed tail 

 destitute of caudal fin. In the same work (Monog. of the Old 

 Red Fish) we find a new definition of the family of Coelacanth 

 fishes, distinguishing them from the Sauroids by the sole cha- 

 racteristic of their body being covered with rounded, imbricating 

 scales ; and yet strangely enough here we find placed the genera 

 AsterolepiSj Bothriolepis and Psammosteus, the only certainly 

 known parts of which are great, irregular, tuberculated bony 

 plates, agreeing most nearly with Coccosteus, differing principally 

 in their greater size; and I find it impossible, either from the 

 examination of the fishes themselves or the writings of authors, 

 to trace their connection with the obvious types of the family, 

 Holoptychius, Phyllolepis and Glyptolepis, which strictly agree 

 with the definition. By withdrawing the discordant elements 

 of those two families, the Cephalaspides and Ccelacanthi, I have 

 left them distinctly and well defined, and by uniting those dis- 

 joined members (whose affinity Sir P. Egerton^s extracts show 

 to have been recognized by Agassiz) I have formed what seems 



