858 DR RAMSAY H. TRAQUAIR ON FOSSIL FISHES COLLECTED BY THE 



one continuous structure in the forms with which we are acquainted, we are, I think, 

 justified in supposing that they were originally distinct pieces. 



If we now turn to the lappet-like expansions behind the head in the Coelolepidse and 

 Ateleaspidse, which I have interpreted as pectoral fins, we. shall see that they bear a 

 strong resemblance to the flap-like structures in Cephalaspis, which are organically 

 continuous with the head shield, and placed immediately internal to the cornua and 

 behind the head. It will be remembered that these were originally described by 

 Lankester as pectoral fins (xiv. p. 41), though Smith Woodward designates them 

 with a query as opercula (xxxvi. p. 176). If the homology of these parts is accepted, 

 it follows that if they represent pectoral fins in Thelodus, as such they must also be 

 looked upon in Cephalaspis, and consequently Lankester was right in his original 

 interpretation. 



I am therefore meanwhile of opinion, that the association of Heterostraci and Osteo- 

 straci in one great subclass of Ostracodermi is not a mere delusion founded on the occur- 

 rence together, geologically, of their fossil remains, and on the presence in both of a 

 cephalic shield, but is supported by the facts brought forward in this paper. The 

 position of the Asterolepidse does not come within the scope of the present observations. 



But unless the Ostracodermi are to be maintained, though it might be only as a 

 " lumber room," we have no place in the system for the two remarkable genera of fishes, 

 with the consideration of which I may now conclude this report. 



The Anaspida. 



I have placed the genera Birkenia and Lasanius together in one family, because 

 both possess a fusiform body with bluntly rounded head, a bilobate heterocercal tail, 

 and a median row of aculeated scutes on the ventral margin. In neither do we find 

 any jaws, teeth, paired fins, shoulder-girdle, or ossified internal skeleton. 



Birkenia has the head and body completely covered with scutes. Lasanius, though 

 nearly naked, has a row of median ventral aculeated plates resembling those of Birkenia, 

 and the only other dermal hard parts which it possesses — the parallel rods behind the 

 head — are directed, for the greater part of their extent, downwards and forwards like the 

 elongated lateral scutes of Birkenia. 



So I am inclined to look upon Lasanius as standing much in the same relation 

 to Birkenia as the nearly naked Phanerosteon does to the other genera of Palseoniscidse, 

 whose bodies are covered with osseous scales. Lasanius seems to have lost its scutes, 

 and is consequently a more specialised form. 



As to Birkenia itself, its heterocercal tail, its small posteriorly situated dorsal fin. 

 and its narrow tuberculated body-scutes, do remind us strangely of Cephalaspis, in 

 which genus it must also be noted that the lateral scutes, if not directed actually down- 

 wards and forwards like those of the former fish, are at least vertical, and do not pass 

 obliquely backwards like the bands of scales on the sides of the ordinary "Ganoids. ' 



The under surface of the head of Cephalaspis is as yet imperfectly known, and no 



