Jan.,i8 95 . THE PROBLEM OF PRIMJEVAL SHARKS. 39 



preserved, with a smooth margin which was evidently overlapped either 

 by skin or by another plate. The inner surface is only marked by 

 indications of blood-vessels ; and the external surface is not traversed 

 by any grooves for sensory canals, unless the network of impressions 

 happens to have served this purpose. The plate is bilaterally 

 symmetrical and a little tumid on each side ; and it is quite likely that 

 the original outline of the anterior and posterior ends is not correctly 

 given in our illustration of the actual fossil. Another fragmentary 

 example of the same plate may be seen in the School of Mines, 

 St. Petersburg ; and it is a homologous piece of armour which has 

 lately been discovered in Scotland. 



Another form of dermal plate at Dorpat, also ornamented only on 

 the convex side, but this ornament being the granulation charac- 

 teristic of Psammosteus arenatus, is more difficult to describe. There are 

 several specimens, but that best preserved measures about thirty 

 centimetres in length, while its transverse arching is so great that the 

 narrow dome it forms is about fifteen centimetres in depth. It looks, 



/—■*";' :: > \ 



■ -» 



Fig. 1. — Median shield of Psammosteus paradoxus, wanting external ornamenta- 

 tion ; one-quarter nat. size. — Upper Old Red Sandstone, Neuhausen, Livonia. 



indeed, like one end of a keel-less boat with the maximum bendings 

 thickened ; and where it tapers towards what would be the middle of 

 the boat, the centre of the floor is excavated by a long re-entering 

 angle. The ornament is abraded on the thickened bends. 



The fragments of Psammosteus arenatus from Spitzbergen, now in 

 the State Museum, Stockholm, are less satisfactory than those from 

 Russia ; but here, too, it is clear that the fish was provided with a 

 bilaterally symmetrical armature, while some of the plates were 

 overlapped (8). 



The spines of Psammosteus are of two kinds, and these, so far as 

 known, are always ornamented on the plan either of P. arenatus or P. 

 mceandrinus, never of P. paradoxus. The bilaterally symmetrical spines 

 are the smallest, and shaped much like the rostrum of Pteraspis. 1 They 



1 One of these spines is figured in Pander's " Placodermen des devonischen 

 Systems " (1857), pi. vii., fig. 16. 



