I. Introduction. 



The specimens of scales and fragments of bone belonging to Fishes which liave 

 been snbmitted to me by Prof. Lindström were collected b}- Dr A. G. Nathohst 

 during the Swedish geological expedition to Spitzbergen in 1882. They fall into two 

 groups. In the first place there are numerous fragments of the shields of Cephala- 

 spidian fishes which were all obtained from the Red Sandstones of Dickson Bay. In 

 the second place are the scales of various sizes, teeth and fragments of bone which 

 all appear to belong to fishes allied to the genus Holoptychius and were all obtained 

 from the shales of the Mimers Valley. I am informed l)y Dr Nathorst that there is 

 no doubt that the Red Sandstones of Dickson Bay are older than the pisciferous shales 

 of the Mimers Valley. At the same time there appears to be some difficulty in draw- 

 ing a line between the latter and the beds of sandstone undoubtedl)' belonging to 

 the Cai-boniferous system, whicli overlie them. 



With reference to this point I may state at once that the evidence afforded 

 by the fish remains tends to associate the red sandstones of Dickson Bay with the 

 Lower Devonian and Upper Silurian of other parts of Europé, whilst the same evidence 

 renders it highly improbable that the shales of the Mimers Valley are older than the 

 middle Devonian (of Scotland) and is not in opposition to the view that they are as 

 late as the Carboniferous. 



In fact the true Cephalaspidian fishes belonging to the genera Cephalaspis, Auche- 

 naspis, Pteraspis, Scaphaspis etc. are essentially Upper Silurian forms, whilst Holoptychius 

 and its Upper Devonian allies — though not necessarily genera which extend into the Car- 

 boniferous period — yet belong to families well represented in that period. On the 

 other hand no true Cephalaspida3 extend upward to the lowest horizon of the Glypto- 

 dipterini nor indeed is there any satisfactory evidence of the association of Cephala- 

 spida3 with any of the Crossopterygian Ganoids — the former being all earlier than 

 the latter — and the true associates of the Cephala.spida3 being the Ctenodipterini and 

 the Acanthodidie. An apparent exception to the trutli of this proposition is found in the 

 occurrence in the Lower Devonian of South Devon of the remains of Pteraspis (Ste- 

 ganodictyum) associated with a reputed Holoptychian fish — Phyllolepis — of whicli 

 a single scale has been found near Torquay by Mr Pengelly. The exception vanishes 

 altogether when we enquire as to what is known witli regard to »Phyllolepis». It ap- 

 pears that there are only two specimens of Phyllolepis in existence besides that of iMr 

 Pengiolly and that they are very large deeply furrowed scutes, the structure of which 

 in no way implies any special relationship of the fish wich l)ore them — to the Glyp- 

 todipterini still less to Holoptychius. 



