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



Specialisation from the most primitive form (Lanarkia) to the most specialised 

 (Pteraspidce) has been accompanied by : — 



1. Fusion of the spinelets (Lanarkia) or shagreen grains (Thelodus) into plates, 

 scutes, and rhombic scales, supported by hard matter, developed in a deeper layer of 

 the skin. 



2. Alterations in the pectoral fin-flaps, which becoming covered up by the postero- 

 lateral plates in Drepanaspis, are finally no longer recognisable in the Pteraspidse. 



Now, if it be admitted that the Ccelolepidas are of Elasmobranch origin — they have 

 hitherto, from the structure of their shagreen-bodies, been looked upon as actual sharks 

 — then it follows, if my views are correct, that the entire group of Heterostraci owes 

 its origin to an Elasmobranch source. This idea has been already foreshadowed, at 

 least as regards the Pteraspidse and Psammosteidse, by Dr 0. M. Reis of Munich, in his 

 remark — " dass Pteraspiden und Psammosteiden sehr nahe mit einander verwandt eine 

 einheitliche Degenerationsgruppe der Elasmobranchier bilden, fiir welche ich den 

 Namen Psammacanthiden vorschlage " (xxv. p. 64). Again, in another and later 

 publication, he says — " Pteraspiden und Psammosteiden gehoren zusammen auf Grund 

 der microscopischen Structur, welche zwar placoid ist, aber die Eigenheit zeigt, dass 

 das Dentin auf die ausserste Schicht beschrankt wird ; es ist dies aber eine Structur- 

 differenzirung zu gross plattiger und massiver Stachel- und Hautplattenentfaltung 

 deren Beginn audi bei den Holocephalenzahnen zeigt " (xxvi. pp. 213-214). However, 

 he does not seem to have suspected any special affinity between his Psammacanthiden 

 (Heterostraci) and the Ccelolepidse, for, on a previous page of the same paper (p. 211), 

 he adheres to the old view as to the correlation of Thelodus and Onchus, — "mit 

 mehreren andern Forschern halte auch ich es fiir sehr wahrscheinlich, dem die zusam- 

 menvorkommenden Thelodus-Schu^'peii und Onc/ms-Flossenstachel einer und derselben 

 Haifischgattung angehoren." 



It follows now that, if the views which I have here supported as to the derivation 

 of the Heterostraci be accepted, two other theories which have been propounded 

 regarding their affinities must fall to the ground. 



The first is that originated by the late Prof. E. D. Cope, who, placing the Pteraspids, 

 Cephalaspids, and afterwards also the Asterolepids,* in one sub-class of " Ostracodermi," 

 associated this sub-class with the Marsipobranchii or Cyclostomes (Lampreys and Hags) 

 in one class of Agnatha, apart altogether with the Pisces or Fishes. This idea, founded 

 on the apparent absence of lower jaw and shoulder-girdle in the Ostracodermi, has 

 been accepted by Smith Woodward in his obituary notice of Cope(x1. p. 379), as well 

 as in his recently published Manual of Vertebrate Palceontology (xli. p. 1). It also 

 appears in Dr Bashford Dean's Fishes, Living and Fossil, where, in his table of Classi- 

 fication of Fishes, he boldly adopts, in place of " Agnatha," the term " Marsipobranchii," 

 previously used only for the Lampreys and Hagfishes themselves. In the table of 



* The Asterolepids or " Antiarcha " were, however, first added to the Ostracodermi by Mr Smith Woodward 

 (xxxviii. p. 17). 



