THE LOWER DEVONIAN FISHES OF GEMtJNDEN. 735 



middle line, we see a small ophiurid starfish firmly adherent to the surface of the fish. 

 Within the concavity formed by the two points of the crescent we find other two 

 curved elevations of the surface, concave towards the middle line and meeting behind. 

 In the middle of the oval space thus enclosed is an oblong elevation, which is clearly 

 of an axial nature. 



With the exception of the last-mentioned feature, it is extremely hard to put any 

 interpretation on the above-described appearances, which are also very well shown 

 in PL VII. fig. 1. The anterior broad crescentic portion may possibly represent a 

 branchiostegal or opercular flap, but what of the two curved elevations behind and 

 within the deep concavity of the horns of that crescent ? Have they to do with the 

 shoulder-girdle ? 



Immediately behind the elliptical space enclosed by the prominent lines last referred 

 to, and continuing the line of the axial elevation in its centre, is a raised band, f inch 

 broad in front, which runs straight back in the middle line for the whole remaining 

 length of the fossil. This band is transversely segmented, the segments being very 

 distinct in front, where fourteen may be counted in the extent of If inch ; they get, 

 however, more obscure further back, though still evident. The segments in front likewise 

 show a median furrow, which is well exhibited in the plate. 



I interpret this band as a vertebral column, consisting of calcified ring-vertebrae, the 

 longitudinal median groove seen in the front segments being, according to this view, due 

 to vertical crushing. But no evidence of apophyses, either neural or haemal, can be 

 discerned. 



As to the dermal structures, almost the entire surface is studded with stellate 

 tubercles, which are larger and more distantly placed in the anterior part of the fish, 

 smaller and closer together behind. Indeed, in front, each of these tubercles seems to 

 form the centre of a small polygonal plate, the rest of the area of which is minutely 

 granulated. This appearance of a mosaic of defined polygonal areas gets lost posteriorly, 

 but at the tail, for a couple of inches before it is cut off by the edge of the stone, we 

 find on each side of the vertebral column a row of small, longitudinally -ridged and 

 imbricating scutes. The fact that the segmented body, which I interpret as the 

 vertebral column, is nearer the right than the left row of scutes, seems to me to be due 

 to a slipping over of the skin to the left side, and therefore to prove that it is an 

 endoskeletal structure. 



As shown in fig. 2 of PL VII., the specimen has also been worked out from the 

 opposite, or, from my point of view, the dorsal side ; but, of course, in order that the 

 fossil might hold together, the operator has, except in the extreme front, kept well 

 within the margin, so that this aspect does not exhibit nearly such an extent of surface 

 as the other. What we do see is unfortunately quite as problematic as before. Quite in 

 front there certainly is what appears to be a wide-open mouth, close behind which are 

 two oval markings, right and left, reminding us of eyes, but I rather suspect that this, 

 at all events, is a deceptive appearance. The rest of the head and body part of the 



