CEPHALASPID.E. 



45 



of the body can be removed, and the view given in fig. 2, PI. XI, is obtained. This 

 demonstrates, firstly, a series of ventrally placed scales, which, though rather confused, 

 covered in the ventral surface, and appear to have consisted of several rows. The 

 great thickness of these scales, and indeed of those of the upper part also, is noticeable. 

 The calcified exoskelcton in these fishes bore some resemblance to that of Lophobranchs, 

 and, no doubt, formed a rather rigid and tough kind of armature. Secondly, the fractured 

 piece of the Arbroath specimen exhibits the body in vertical section, and shows its very 

 small calibre and triangular form. 



Fig. 17. 



Ideal vertical transverse section of the body of Eucephalaspis. 



The number of rows in each series of scales is, no doubt, an important matter — pos- 

 sibly differing in species. Both specimens of C. Li/cUii are unsatisfactory as regards this 

 point, since the scales are fractured ; but there appears to have been in front of the dorsal fin 

 but one dorsal row of scales on each side the middle line, attached by their lower borders to 

 the single series of great flanking or lateral scales : posteriorly to the fin the series became 

 broken up into more numerous scales. The number of scales in a row and the relative 

 length of head and body are important for specific distinction ; and though it does not 

 bear on the question of distinction between C. Lyellii and C. Agassiziiy since the body of 

 the latter is unknown, yet a very clear difference in this respect is exhibited between 

 C. Lyellii and another Scotch species, C. Powriei. 



The two specimens of C. Lyellii agree, as nearly as their rough condition will 

 permit one to estimate, in having about twenty-five scales of the lateral series, placed 

 praedorsally, i. e. which are in front of the anterior origin of the dorsal fin. In the British 

 Museum specimen of C. Lyellii the length of the head is 3 inches, of the whole fish 

 inches, allowing for the tail, which gives a ratio of 1 to 2-4. In the Arbroath specimen 

 the head is 2| inches in length, and the whole fish 7^ inches, giving a ratio of 1 to 2 0. 

 In C. Powriei, on the other hand, the ratio appears to be about 1 to 3-2. Allowance is 

 made in these estimates for the parts which are broken. 



The details of structure of the fins shown in the specimen at Arbroath do not allow of 

 much being said, since they are not very clear. Small scales api)ear to extend along the 

 fin-rays in both dorsal and caudal fin. It is, however, very important to notice that the 



