Devonian Fish-Fauna of Spitzhergen. 5 



original tissue exliibit an external ornament of fine tubcrcii- 

 lations. The first shield, indeed^ may be regarded witliout 

 hesitation as having enveloped the dorsal aspect of a chordate 

 animal, vfhile the second shield can be determined with equal 

 certainty as originally ventral in position. 



The dorsal shield is unfortunately much less nearly com- 

 ])lete than the ventral, and, so far as preserved, seems to consist 

 of a single continuous piece. As shown by the fractured 

 margin the substance of the shield comprises a thin outer and 

 inner layer, very dense, separated by a thick layer of cancellee 

 with delicate septa. The section also proves that there was a 

 thickening of the middle layer, producing on the surface a 

 sharp longitudinal keel in the middle of the back (PI. I. fig. 2). 

 The posterior portion of the shield is obviously broken away, 

 and if any part of its anterior border remains this is confined 

 to the bilaterally symmetrical reentering angle, from which 

 there proceeds backwards on the visceral aspect a feebly 

 marked ridge, and in relation to which a great pair of processes 

 (x) with two pairs of linear impressions are symmetrically 

 disposed. While the downwardly turned border on tlie right 

 side is well shown, a considerable portion of the left side is 

 thus proved to be missing, and the amount is indicated by 

 the dotted line in fig. 2. The anteriorly and downwardly 

 directed processes {.v) are unlike any structure hitherto observed 

 by the present writer in an Ostracoderm dorsal shield ; they 

 are bluntly rounded, but apparently not much thickened, and 

 are most suggestive of an arrangement for complex articula- 

 tion with a shield originally occurring in front. The outer 

 of the two divergent lines extending backwards from the base 

 of each process {m) corresponds with a sharp longitudinal 

 angulation of the shield, as indicated by the transverse section 

 (fig. 2) ; and this may perhaps represent an obliterated 

 suture, though the evidence is not conclusive. This line is 

 directed outwards, but the inner line, wliicii has more the 

 appearance of a fold than an indentation, trends gradually 

 towards its fellow of the opposite side behind. 



The ventral shield is imperfect on all sides except the right, 

 but in general contour it seems to have closely resembled the 

 corresponding armour of the typical Pterichthys Milleri from 

 the Old Red Sandstone of northern Scotland. It is, indeed, 

 broadly ovate^ tapering behind. The sutures between the 

 component plates are indicated partly by impressions upon 

 the matrix and partly by the arrangement of the fibres in the 

 fragments of the exoskeletal tissue that remain. There are 

 also some traces of the superficial tubercular ornament, fine 

 towards the middle of the shield, somewhat coarser on the 



