110 FISHES. 



also consisting of thin horny plates, but having their pos- 

 terior margins fringed with spines, or cut into comb-like pro- 

 jections. 3. Ganoid scales, composed of an inferior layer 

 consisting of bone, covered by a superficial layer of hard 

 polished enamel (the so-called "ganoine"). These scales 

 (fig. 486) are usually much larger and thicker than the 

 ordinary scales, and though they are often articulated to one 

 another by special processes, they only rarely overlap. 4. 

 Placoid scales, consisting of detached bony grains, tubercles, 

 or plates, of which the latter are not uncommonly armed 

 with spines. 



It is very important for the geologist to recognise the 

 characters of these diiferent scales, as he may have to decide 

 upon the characters of a fossil fish merely from detached 

 scales. Such decisions, however, are always more or less 

 hazardous, since the scales of the different orders of the liv- 

 ing fishes are not invariably of the same kind in all the forms 

 of the order. Thus, ganoid scales are not peculiar to the 

 order of the Ganoid fishes, but occur also in some of the 

 Bony Fishes {Tekostei). The scales, also, form at best but 

 one character, and they can hardly be said to constitute the 

 most important cliaracter of any fish. A classification, there- 

 fore, which is based primarily upon the nature of the scales, 

 necessarily is more or less " artificial," and is liable to bring 

 into juxtaposition forms which have no real affinity to one 

 another. For these reasons, most zoologists do not accept 

 the classification of the Fishes into the four orders of the 

 Cycloidci, Cteiwidd, Ganoidci, and Placoidci, since this classi- 

 fication, though sanctioned by such an eminent authority as 

 Professor Agassiz, is founded solely upon the nature of the 

 integumentary covering. The palaeontologist, however, whose 

 materials often consist of nothing more than detached scales, 

 has been not rarely driven, by the necessity of the case, 

 to provisionally classify his specimens in accordance with 

 the nature of these appendages. 



As regards their true osseous system or endoskeleton, 

 Fishes vary very widely. In the Lancelet there can hardly 

 be said to be any skeleton, the spinal cord being simply 

 supported by the gelatinous notoeliord, which remains through- 



