326 DR RAMSAY H. TRAQUAIR ON FOSSIL FISH REMAINS 



necessary to complete the shield (assuming its near relationship to Pterichthys and 

 Bothriolepis, which seems reasonable) all combine to indicate the dorsal rather than the 

 ventral aspect of the fish." 



But in 1891 Dr Smith Woodward, in the second part of his " Catalogue" (pp. 314- 

 315), unhesitatingly classes the genus as Coccostean, pointing out the resemblance in 

 form between the median plate figured by Claypole and Newberry and the median 

 ventral of certain species of Coccosteus, e.g. C. disjectus ; holding also that "the recent 

 description of the complete ventral shield by Claypole proves that it agrees with that 

 of Coccosteus in every essential particular. The ' post-dorso-median ' plate of Claypole 

 is obviously the anterior median ventral, while the ' post-dorso-lateral ' and ' dorso- 

 lateral ' of the same author are the anterior and posterior ventro-lateral plates re- 

 spectively." The cuirass figured by Claypole is in effect a Coccostean plastron turned 

 the wrong way. 



In the same year Professor E. D. Cope # mentioned a specimen of the genus from 

 Mansfield, Tioga County, Pennsylvania, namely, a lateral plate of the plastron. He 

 also states : "Besides this there is a nearly complete pectoral spine, which is of much 



interest, as this part of the skeleton has not been previously known The spine 



differs from that of both Bothriolepis and Pterichthys in being without complete 

 segmentation. It is continuous throughout to the apex. This then will constitute the 

 generic distinction so far known between Holonema and Bothriolepis" Of this supposed 

 spine of Holonema a figure is given, which at once betrays to the eye of anyone 

 familiar with the structure of the Asterolepidse that it is not an entire appendage but 

 merely the distal portion or segment of the pectoral limb of Bothriolepis. 



In another paper t published in 1892 Cope figured and described as Holonema 

 rugosum a plate which he considered to belong to a new species. As regards the genus, 

 he still looked upon it as belonging to the Antiarcha, but with tivo ventral median plates, 

 thus differing from Bothriolepis. He also described without figure a plate which he 

 referred to H. rugosum, saying regarding it : " This piece, together with the pectoral 

 limb which I have already described, demonstrates the position of the genus to be with the 

 Antiarcha and not with the Arthrodira as has been suspected by Mr A. S. Woodward." 



Dr Smith Woodward,! however, in a review of the first of Cope's two papers, 

 reaffirmed that " the so-called dorsal shield of Holonema is really the ventral shield 

 turned the wrong way forwards ; and the genus belongs to the Arthrodira, not to 

 the Ostracodermi." Further, that " the limb referred by Professor Cope to Holonema 

 is the distal segment of the arm of Bothriolepis, originally named Stenacanthus by 

 Leidy." § That he was right in the second statement goes without saying ; and as 



* Proc. Am. States National Museum, vol. xiv. (1891), pp. 447-463. 



t Proc. Am. Phil. Soc., vol. xxx. (1892), pp. 221-229, pis. vii. and viii. 



t Geol. Mag. (3), vol. ix., 1892, pp. 233-235. 



§ It will also be remembered that Agassiz himself was deceived by the corresponding element in Bothriolejns 

 major, from the north of Scotland, which he figured, under the name of Placothorax paradoxus, as " un type nouveau 

 de la Eamille des Cephalaspides." (1'oi.ss. Foss. vieux gres rouge, p. 134, tab. 30 a, figs. 22, 23.) 



