26 rUOCKKUlNCiS OF THE 



As ,111 illustration of tlui dilliculties and limitations in the study 

 ol' extinct animals, 1 would refer to tlie earliest f^eological evi- 

 dence of the Vertebrata whicli 1 have long had the opportunity 

 of investigating. The oldest fossils which are comparable with 

 \ertebrate skeletons can scarcely belong to animals lower in i;rade 

 than the existing Cyclostomes, and they are all either too viiguely 

 shown or loo highly specialised to give an}' satisfnctory clue to the 

 invertebrate group from which they were descended. In outward 

 shape n}any of the armoured forms much reseudjle the contem- 

 porary Merostouiata and the later marine Arachnids, but to 

 recognize any genetic connection, such as is advocated by Gaskell* 

 and fatten t, involves more assumptions than are justifiable. 

 The most generalised forms found as fossils are all distinctively 

 fish-shaped, and there can be no doubt that the really annectant 

 types which j)receded them were soft-bodied and not likely to be 

 preserved. 



These earliest vertebrates occur abundantly in some of the 

 uppermost Silurian deposits of western and northern I'^urope, but 

 most of them have so slight a skeleton that they ap[)ear as mere 

 stains on the surface of the rock. Interesting,' though tantalis- 

 ing new specimens have lately been collected by .Mr. William 

 Md'herson for the liritish Museum from the Downtonian shales 

 of Ayrshire. Even those genera in which the skin is provided 

 with well-calcitied shagreen or scales show very little beyond their 

 general contour. Fortunately, however, they are followed in the 

 overlying Devonian formations both of Europe and North America, 

 and even of the southern hemisphere, by numerous nu)re specialised 

 members of the same group, in w hich many of the dermal tubercles 

 have coalesced with a deeper-seated calcification into symmetrically- 

 arranged plates, which often bear marks of subjacent internal 

 organs, it is thus possible to make some attempt at their 

 interpretation. 



JSome have doubted whether all these primitive organisms 

 belong to a single group, but 1 still think Cope was probably 

 right when he included all those known to him in his subclass 

 Ostracodermi (or Ostracophorj). The genera without armour- 

 plates discovered in more recent years seem to pass by gradations 

 into the others, and therefore presumably had the same funda- 

 mental characters. The anterior visceral arches are not modified 

 into ordinary jaws — at least, if they were so, we should expect to 

 find them either calcified or covered with a corresponding exo- 

 skeleton. The gill-arches, in an extensive gill-chamber, are far 

 forwards. There are no paired iins; while in the median fins 

 there are )io ordinary lin-rays. but rows of scales instead. The 

 dermal plates, when present, are highly vascular, and tiiey always 

 retain as a superficial layer the tubercles of dentine which are the 

 sole covering of the more generalised fornis. 



* ^Y. ir. Gaskell, 'The Origin of Vertebrates ' (London. 1908). 

 t W. Piitten, 'The Evohition of the Yeitebratcs and their Kin' (Phihi- 

 delphiii, iyi2). 



