Classification of Teleostean Fishes. 297 



bizarre de caract^res que pr^sente ce genre me parait une 

 confirmation eclatante du rapprochement que j'ai fait des 

 Pleuronectes et des Chetodontes, et si I'on fait abstraction des 

 rayons epineux de ces derniers, ou plutot si I'on consid^re le 

 peu d'importance que m^rite ce caract^re dans une faniiile 

 qui compte des genres conformds comme les Platax et les 

 Psettus, on ne m^connaitra pas leur intirae affinite. Qx'on 

 ne m'objecte pas la conformation bizarre et irreguli^re de la 

 t^te des Pleuronectes ; car oil qu'on les range, ils se distin- 

 gueront toujours par la de tons les poissons connus," 



Somewhat the same suggestion as that of Agassiz was 

 again incidentally made by E. W. L. Holt in 1894 * : — 

 " Messrs. Cunningham and MacMunn find a difficulty in 

 accepting reversion or atavism as an explanation of the 

 ambicolorate condition, in that the hypothetical vertically 

 swimming ancestor of the flat-fish must have had an un- 

 pigmented white or silvery ventral surface, as other sym- 

 metrical fishes have, whereas completely ambicolorate flat-fish 

 are uniformly pigmented all over. The difficulty certainly 

 arises if we assume that the ancestor really was paler on the 

 ventral region than elsewhere ; but is it not equally reasonable 

 to assume a stage of evolution in which the fish resembled 

 such forms as Platax or JJascyllas, to take instances from 

 families widely separated from each other by systematists? 

 Both forms have high compressed bodies, and in some species, 

 at any rate, of both genera the ventral region is as deeply 

 pigmented as the dorsal. Even in the John Dory (^Zeus 

 falser), in which the ventral abdominal region is flattened, it 

 is nevertheless rather darkly pigmented, and to me it certainly 

 seems more probable that the Pleuronectida^ of the present 

 day began to take on their asymmetrical characters as com- 

 pressed and uniformly coloured forms than in the condition 

 of ordinary round fish." 



Merely for the sake of completeness would I allude to the 

 suggestion made by Gill in lb87 f: — " I am half inclined to 

 think that the Heterosomatous fishes may have branched off 

 from the original stock, or progenitors of the Ta3niosomous 

 fishes [Trachypterida3]. I shall investigate the subject when 

 I can get the requisite material." Dr. (iill has not published 

 the reasons which made him incline towards such a con- 

 clusion, and the position in which lie has left the Pleuronectids 

 in his latest classification %, with the Anacantliini between 

 them and the Trachypterids^ seems to show that the idea of 



* P. Z. S. 1894, p. 438. 



t Amer. Natur. xxi. p. 86. 



X Mem. Nat. Acad. Washington, vi. 1893, p. 137. 



Ann. & Mag. N. Hist. Ser. 7. Vol. x. 22 



