42 



OLD RED SANDSTONE FISHES. 



is more or less that of a Cepbalaspid with head and body. The spmous tubercles on some 

 species of Cephalaspis (PI. X, fig. 5) suggested the idea that this might belong to 

 the lower surface of one of these fishes, covering in the ventral aspect of the head and the 

 ventral series of scales ; being detached, as it v^^ere, in this part of the animal, but adhering 

 closely to the deeper aponeurotic layers of the exoskeleton in the superior parts, and form- 

 ing the characteristic tuberculated surface of the shield and scales. There is, however, no 

 proof of its connection with these fishes, and it seems more probable that it belongs to some 

 early representative of the Sharks and Rays (representative, perhaps, only in the character of 

 its dermal ossifications) than that it is the ventral covering of a Cepbalaspid. Its mention 

 here may incite investigation and lead to the discovery of specimens showing clearly a, 

 connection, but as matters at present stand there is not sufficient evidence to justify the 

 introduction further of the specimen into these pages. 



Value of Specific and Generic Distinctions. — Before proceeding to describe the genera 

 and species of Osteostraci, I would allude to the principles Avhich have determined me in 

 forming genera and species. In the first place, as to species. It appears to be better for the 

 end of advancing knowledge to signalise a doubtful form by a name, than to pass it over 

 as a possible variety or questionable species. The error of associating under one head what 

 are really distinct species is undeniably as great as that of founding new species on 

 specimens which may eventually prove to be imperfect conditions or locally distorted 

 states of species already known. The object to be held in view is to attract attention to 

 the matter in doubt, and this can be best done by the use of a name, which has only to 

 be rescinded in the event of further discovery rendering it desirable so to do. The 

 object of a descriptive work such as this, in which the material is fully placed before the 

 reader, is to point out distinctions and peculiarities in the various specimens studied, and 

 it seems better in this case to err on the side of division than on that of fusion. 



The only use that can be made of generic divisions in the case of fragmentary fossil 

 remains is to place those things which are ahke in the same group, and those which differ 

 in different groups, thus pointing out the gaps in om* knowledge of the continuity of 

 forms. If our knowledge were more complete there would be greater difficulty in 

 indicating groups, a fact which is true for all groups of organisms. Grouping is 

 facilitated by the occurrence of gaps that are not real, but are due to our imperfect 

 knowledge. 



List of Sjjecies of Os,'mo?,TV.^ci, arranged in Order of their Occurrence. 



Upper Silukian. Thyestes verrucosus, Eicliwald ; Cephalaspis (?) Schrenkii, Pander. 

 DowNTON Sandstone. Eukeraspis puslidiferus. 



Passage beds. Didi/maspis Grindrodu,Auchenaspis Salteri, Av. Egertoni, CepJialaspis {Hemicijclaspu) 



Murchisnni, C. Liyfitbodii. 

 Lower Devonian (of Devonshire and Cornwall). Cephalaspis (?) Carleri. 

 CoRNSTONES (Wcst of England). Euceplialaspis Agassizii, Zenaspis Salweiji. 

 Lower Old Red Sandstone (Scotland). Eucephulaspis Li/ellii, E. Powriei,E. Pagei, E. asper. 



