12 OLD RED SANDSTONE FISHES. 



laminated material is arranged round these tufts concentrically, the finer branches 

 traversing it, much in the same way as dentinal tubules traverse dentine. Indeed, each of 

 the sections of the ridges recalls very strongly the structure of a tooth or of a dermal defence 

 of a Placoid Fish. There is no trace of a 'ganoin' layer beyond the laminated material 

 forming the ridges ; but it is noticeable that the outermost laminae in specimens from the 

 Cornstones acquire a pinkish-brown colour distinct from the rest of the shield, and readily 

 scale off. In some species this outermost surface is dull ; in others it is brilliant and 

 polished. The fact of the distinct colour of this outermost part of the outermost layer, 

 together with its tendency to separate, might lead to the impression that we have here a 

 ' ganoin ' layer ; no such layer, however, is to be detected with the microscope. 



The absence of bone-lacuna3, the paucity of vascular canals, and the excavation of the 

 mid-layer of the shield into large vascular sinuses, cannot be too strongly insisted on as 

 the characteristic structure of these shields. 



Professor Huxley obtained from Sir Philip Egerton specimens of Kner's Pteraspis 

 from Gallicia, and most completely reconciled the German palseontologist's description of 

 the fancied resemblances to the Cuttle-bone with the true structure of the shield as he 

 had ascertained it from English specimens. Kner's ' innermost layer' is obviously the 

 ' nacreous layer ' described by Huxley ; the ' prismatic layer ' is formed by the polygonal 

 cavities of the middle layer, which have become infiltrated with probably carbonate of 

 lime, as in many examples from Britain; the striated layer undoubtedly corresponds 

 to the outer layer in our Fish, and the identity in structure is complete. 



The chief differences, then, between these two sets of shields in their minute structure 

 consist — first, in the absence of osseous lacunse in the Pteraspidian forms, their presence 

 in the Cephalaspidian^ forms ; secondly, in the different general character and arrange- 

 ment of the vascular sinuses ; thirdly, in the different mode of arrangement of the external 

 layer, which is, it seems, invariably marked with fine long ridges in the former, with 

 minute tubercles in the latter. 



(<?.) The Pteraspidian forms present such marked dififerences among themselves in the 

 form of their shields, that it would be inconvenient to retain them in one genus ; the 

 same is true of the Cephalaspidian forms -. hence I have adopted the Section Heterostraci 

 (ETEpoc, of another kind ; oar^aKov, a shell or dermal bone) for those with the peculiar 

 structure of Fteraspis ; and the Section Osteostraci [oarkov, bone) for those with the bony 

 structure of Cephalaspis Lyellii. There are other characters which separate these two 

 Sections ; as, for example, the disposition of the orbits mentioned above. 



It is interesting to observe that each of the Sections presents a genus in which the 

 shield is composed of a single simple piece : in the Heterostraci this shield presents 

 no trace of orbits, but is a simple discoid body [Scaj^hasjjis) : in the Osteostraci the 

 simple shield has large diverging appendages (cornua), and mainly consists of an orbital 



1 I have ascertained the presence of bone-lacuna3 in Auchenaspis and Didymaspis, genera allied in 

 external character to the type Cephalaspis Lijellii. 



