xym 
INTRODUCTION. 
occurring further backwards, and especially distinct in Cyathaspis l , 
are exactly such as might result from contact with ridges upon the 
auditory capsules, due to a great development of the upper semi¬ 
circular canals, as in Sharks. In short, all positive characters 
are rather in favour of an alliance with the class Pisces than 
otherwise ; and although these organisms cannot be defined with 
scientific precision, it seems advisable at present to regard them 
as a primitive Piscine subclass of uncertain affinities. 
The name Ostracodermi is preferred for this subclass, because 
Prof. Cope seems to be the only naturalist who has hitherto ven¬ 
tured to remove the Coccostean fishes far from the order that com¬ 
prises the Asterolepidse. So long ago as 1848, M‘Coy 2 proposed to 
institute the “ family Placodermi ” for the Asterolepidse and Cocco- 
steidse, allowing Cephalaspis to remain as the type of a distinct 
family; and all subsequent authors seem to have adopted this 
arrangement, with only slight changes in the rank allowed to the 
great divisions. Even so recently as 1888, Traquair 3 regarded the 
Asterolepidse and Coccosteidse as separated by characters merely of 
family value; and in the latest work of Newberry 4 , the same 
classification, though not systematically formulated, is implied. It 
must, however, be remarked that both Newberry, Traquair, and 
other authors have on several occasions pointed out the close 
resemblance between the dentition of the Coccosteidae and that of 
the Dipnoi ; and it is by extending this suggestion to its logical 
issue, in the light of the latest researches, that the classification 
adopted in the following Catalogue has been attained. The Cocco¬ 
steidae and their allies possess ossified jaws and a dentition that are 
far from incipient or rudimentary. Some are believed to have had 
pectoral fin-spines (e. g., Dinichthys and Brachydirus) —a circum¬ 
stance implying the presence of highly specialized paired fins; and 
even where pectorals have not been observed (e. g., Coccosteus ), 
membrane-bones identical with those of an ordinary pectoral arch 
are certainly well developed. Coccosteus, moreover, is now proved 
to exhibit highly specialized pelvic fins. These characters suffice, 
at least in the present state of knowledge, to separate the Coccosteus- 
like fishes very widely from those now termed Ostracodermi ; and 
it may be added that even detached fragmentary plates can in many 
1 See Kunth, von Alth, and Lankester's figures already quoted. 
2 F. M‘Cov, Ann. Mag. Nat. Hist. [2] vol. ii. (1848), p. 6. 
3 R. H. Traquair, G-eol. Mag. [3] vol. v. (1888), pp. 508, 511. 
4 J, S. Newberry, 1 Palaeozoic Fishes N. America’ (1889). 
