The Edingburgh New Philosophical Journal. 255 



ments ; these consist of masses of enamel, often of considerable dimen- 

 sions, at other times reduced to small points, such as the rings of the 

 rays and the different shagreened surfaces of rays and sharks. 2d 

 Order : The Ganoides. — The character common to all these is the an- 

 gular form of the scales, which are composed of two substances, namely, 

 corneous or osseous leaflets placed one above another, and covered 

 with a thick layer of enamel. M. Agassiz includes many families 

 both recent and fossil in this order, such as the Sclerodermes, Gymno- 

 dontes, Lophobranches, Goniodontes, Silures, and Sturgeons. 3d Order : 

 The Ctenoides. — The scales are formed of plates pectinated on their 

 posterior edge ; the pectinations of these numerous plates superimposed 

 on each other, so that the margin of the under one projects beyond 

 that above it, rendering these scales rough to the touch. This structure 

 is particularly remarkable among the Chenodontes and Pleuronectes. 

 Here likewise are arranged the Percoides, Polyacanthes, Scienoides, 

 Sparoides, Scorpiono'ides, and the Aulostomes. We shall afterwards 

 have an opportunity of learning that these combs are very far from 

 being simple plates, so arranged that the under ones always project 

 beyond the upper : we shall see that they are true teeth, all the deve- 

 lopment of which admit of being traced. This structure must needs 

 have escaped M. Agassiz's observation, in consequence of the insuffi- 

 ciency of the means he employed. For the same reason, this distin- 

 guished observer has not discovered the form of the teeth in many 

 of the families he has placed in the following order. 4th Order : The 

 CycLoiDEs. — The families belonging to this order have scales formed 

 of simple plates, with a smooth border; a structure which does not 

 prevent their surface being frequently ornamented with various designs, 

 imprinted at once on all the scales on their outer part, which is not 

 covered. Here we must place the Labroides, Muges, and Atherines, 

 Scomberoides, Gadoides, Gobioides, Murenoides, Lucloides, Salmones, 

 Clupeae, and Cyprini. We shall subsequently see that this order 

 contains families whose scales present the most strongly marked 

 differences. Assuredly we cannot assimilate the scales of the Cyprini 

 to those of the Gobioides, nor arrange the Muges with the Atherines ; 

 the Muges present distinct teeth, although it may be true that the 

 scales are not rough to the touch. In the family of the Cyprini there 

 is room for sub-divisions. In this case also, a simple magnifying 

 glass produced quite an insufficient enlargement of the objects for such 

 investigations. We shall revert to this interesting point, which has 

 acquired great importance since M. Agassiz's inquiries regarding the 

 form of scales. 



" In finishing the history of researches into the structure of scales, 



