Dec. 1893. EVOLUTION OF THE SCALES OF FISHES. 449 



decidedly suggestive of the truth of Professor Ryder's theory of the 

 effect of mechanical strains. 



Among undoubted fishes it is not easy to discover any facts par- 

 ticularly apposite to the question under consideration in the groups 

 of Dipnoi and Fringe-finned "Ganoids"; but it seems possible to 

 turn with more satisfaction to the higher " Ganoids " of the sub-class 

 Teleostomi. Here it is quite evident, in a general way, that as the 

 fishes become higher in organisation — become specialised — the scales 

 are more deeply overlapping, thinner, and almost or regularly cycloidal. 

 We can now even quote instances in which the anterior part of the 

 trunk is covered by typical quadrangular ganoid scales, united by a 

 peg-and-socket articulation ; while the base of the tail of the same 

 fish is enveloped by deeply-overlapping thin rounded scales, which 

 would be typically " cycloid" were it not for the presence of a few 

 remnants of superficial enamel. 



Such an instance is a deep-bodied fish related to the European 

 Liassic genus Dnpedhis, obtained from the Hawkesbury Formation 

 of New South Wales and shortly to be described under the name of 

 Aetheolepis. Typical examples of its scales are shown in the accom- 

 panying figures. Those of the anterior series, as shown by the 



Scales oi Aetheolepis from Hawkesbury Formation, New South Wales. 



drawing to the left, are thick and rhombic with well-developed peg- 

 and-socket articulation ; while on the caudal region all the scales 

 are very thin and much overlapping, and they gradually degenerate 

 backwards in the manner here indicated. The specimens are not 

 sufficiently well-preserved to admit of the making of microscopic 

 sections ; but it is probable that the hinderniost scales are destitute 

 of bony tissue, and the traces of enamel are merely a few isolated 

 tubercles. Aetheolepis, in fact, so far as its scales are concerned, 

 belongs in its anterior half to the " order Ganoidei " of Agassiz, while 

 posteriorly it almost exactly enters the " order Cycloidei " of the 

 same author. 



There is a tendency to the same kind of scale-arrangement also 

 in a fish from the English Lias, apparently of the genus Endactis ; 

 and from both instances it may probably be inferred that the cycloidal 

 scales have resulted from the modification of the thick quadrangular 



2 G 



