40 COMPARATIVE MORPHOLOGY OF VERTEBRATES. 



salts on their outer ends, thus forming a small plate of dentine (p. 24) 

 with a central spine into which the papilla extends. The overlying 

 epidermal cells form an enamel organ, the lower surface of which 

 secretes an even harder layer of enamel 1 upon the dentine base, this 

 being thickest on the tip of the spine. The mesenchyme in the papilla 

 is the so-called pulp. With continued growth the spine projects through 

 the epidermis, giving the skin of the shark its characteristic rough 

 (shagreen) condition. This is the placoid type of scale. 



FISHES. In the adult elasmobranchs the scales may be large and remote from 

 each other (skates) or small and closely set. In the torpedo scales are lacking, 

 while in the chimaeroids they occur only on the claspers, on the frontal horn, and 

 as extreme forms, in a great spine in front of the dorsal fin. 



FIG. 32. Ventral armor of Stegocephals (after Credner-Zittel). A, Branchiosaurus; B, 

 detail of same; C, detail of Archegosaurus; D, of Petr abates. 



A few ganoids lack scales (Polyodori), while the sturgeon have minute granules 

 and five rows of large plates along the sides. Amia has scales of the cycloid type, 

 soon to be described. With these exceptions the ganoids have ganoid scales, which 

 are rhomboid in outline and joined to each other like parquetry. They consist of 

 two layers, the lower apparently homologous with the dentine of sharks, except that 

 it is formed in, not on, the corium. The outer layer of ganoin is formed by the 

 corium and consequently cannot be enamel as once was thought. 



A few teleosts are scaleless (some eels), but elsewhere scales are formed in 

 pockets in the corium (fig. 181). At first they lie side by side, but with growth they 

 overlap like shingles. There is only one layer of bone mixed with a large amount 

 of ossein.- In cycloid scales the element is circular and is marked with concentric 

 and radiating lines. The ctenoid scales differ in having the posterior edge of 



1 There is some question whether this layer is really enamel; the usual statement as to 

 its nature is followed here. 



