PhylogeuTj of the Teleostomi. 345 



Cephalaspidse, although it has long been known that they 

 differ from them fundamentally in the microscopic structure 

 of their dermal armour, bone lacunge being entirely absent, 

 whilst there is great similarity to the tooth-structure of the 

 Elasmobranchs. Lankester has strongly maintained that the 

 Heterostraci and Osteostraci are an unnatural association, 

 and as long ago as 1867 he wrote*: — "The Heterostraci 

 are associated at present with the Osteostraci because they 

 are found in the same beds, because they have, like Cepha- 

 laspis, a large head-shield, and because there is nothing else 

 with which to associate them — the shields are not so closely 

 similar in plan, much less in histological structure f, as to 

 warrant any inference of similarity in other parts." Within 

 the last few years Traquair:]: has discovered new forms 

 which seem to place it beyond doubt that the Heterostraci 

 are armoured Chondropterygii. He has also discovered a 

 new genus, Ateleaspi's, which he considers is annectent 

 between Heterostraci and Cephalaspid;e, but this view I am 

 not prepared to accept. Ateleaspi's is certainly very closely 

 allied to Cejyhakispi's, but I cannot see that there is the least 

 reason for regarding it as allied to anything else. The 

 shield is divided superficially into hexagonal areas, which are 

 compared to those of Cephalaspis, in which genus this 

 appearance has been shown by Lankester § to be due to the 

 arrangement of the vascular canals, which may even cause 

 the shield to crack along these lines, whilst in pi. x. fig. 5, a 

 specimen of Cepkalaspis asper is figured in which the polygonal 

 areas are very strongly brought out by the great pressure and 

 the infiltration to which the shield has been subjected. If 

 Lankester is correct, and the polygonal areas of Cephalaspis 

 are due to the arrangement of the vascular canals, then they 

 are not due to the coalescence of originally separate poly- 

 gonal pieces, as suggested by Traquair, who believes he has 

 found in Ateleaspis a stage in this development. Traquair's 

 idea that the superficial tubercles of the shield of Ateleaspis 

 represent originally separate Ccelolepid denticles appears to me 



* Mon. Palseont. Soc, Cephalaspidae, p. 62 (1867). 



t The difference in structure of the dermal armour of Fterafipis and 

 Cephalaspis is essentially that between a ■' placoid " and a " ganoid '' scale. 

 There is no reason why the former should not have given rise to the 

 latter and to membrane-bones, by fusion and by the development of a bony 

 substratum, more than once. On the other hand, the evidence shows that 

 the Teleostomi, as here understood, are monophyletic. 



X Trans. Roy. Soc. Edinburgh, xxxix. 1899, p. 827 et seq., and Rep. 

 Brit. Assoc. 1900, p. 773. 



§ ' Cephalaspida3,' p. 10. 



Ann. d: j\[ag. N. Hist. Scr. 7. Vol. xiii. 23 



