1907.] THE SCALES OF FISH. 769 



The remainder of the Actinopterygii, except the Teleostei, all 

 ■appear to have the Lepidosteoid type of scale, with the peculiar 

 tubules described above (p. 758). 



The Protospondyli. — In the family Semionotidse, I have 

 examined Semionotus kajypfi, Lepidotus mantelli, and Dapedius 

 sp. ; in the Pycnodontidfe, Microdon sp. ; in the Eugnathidse, 

 JiJicgnathus formosus (PI. XLYI. fig. 20) ; in the Pachycormid^, 

 Pachycormus heterurus ; and Ophiopsis in the Macrosemiidas. 

 These genera display a remarkably uniform type of lepidosteoid 

 scale. The scales of the Amiidfe (Amia and the extinct 2Iegalurus) 

 are so thin that little can be made out in sections beyond the fact 

 that they are formed of laminae and contain bone-cells, except in 

 the topmost layer. However, a section of a cranial dermal bone 

 of Amice shows the lepidosteoid tubules developed in the most 

 ^perfect manner (PI. XLVI. fig. 18). This observation is of some 

 practical importance, as it may help us to determine the affinity 

 of those fish in which the scales are degenerate or absent, but of 

 which cranial plates can often be obtained*. 



The uEtheospondi/li. — Both the Aspidorhynchidae and the 

 Lepidosteidfe show the same characteristic structure ; the genera 

 examined being Lepidosteus (PI. XLVI. fig. 16, and text-fig. 199), 

 and Aspidorhynchus (fig. 19), already described by Williamson, 



It is interesting to find that the scale of Pholidophorus also 

 belongs to the Lepidosteoid type. 



Summary. — From what has been said in the foregoing pages, 

 it will be gathered that the scales of the Dipnoi and Teleostomi 

 present very few distinct types of structure. That in fact only 

 three distinct types exist : the cosmoid, the palseoniscoid, and the 

 lepidosteoid. -Certain other varieties are found, such as the 

 Rhizodont and Dipnoan, which are probably to be derived from 

 the cosmoid. The position of the Ocelacanth scale is at pi-esent 

 difficult to determine : it may be a primitive form in which the 

 denticles have not yet fused to a cosmine layer, as Williamson 

 supposed ; or it may be simply a degenerate cosmoid scale to the 

 surface of which denticles have become attached. 



Further, it appears that the structure of the scales is very 

 uniform within the families, and that closely allied families 

 usually have very similar scales. The cosmoid scale occurs in the 

 extinct Ostcolepidoti (Orossopterygii) and Dipnoi ; but in no 

 other group of fish. Similarly the ganoid scale occurs in the 

 Teleostomi and never elsewhere. The palasoniscoid type is 

 restricted to the Palseoniscidse and their immediate allies ; while 

 the lepidosteoid type is universal among the Protospondyli, the 

 ^theospondyli, and the Pholidophoridae, and not found in any 

 other group as far as is known. 



These different kinds of scales, then, are of great systematic 

 value, and the position of a fish in any of the large divisions can 



* Stewart lias described similar bone in Fistularla (Cat. Coll. Surgeons). 



