•348 



Part X IIL — Eighteenth Annual Report 



primitive notochord proceeds to the extremity of the caudal region as in 

 the true diphycereal tail — it is impossible to say until the developmental 

 stages are examined. If the former were the case, however, we should 

 expect to find some evidence of the urostyle amongst the posterior 

 vertebrae, but there is no tiace of it. Hence we should be inclined to 

 believe that the true form of the diphycereal tail reappears again 

 through convergence. The gradual transformation which is undergone 

 by the urostyle is evidence for this view. Only one stage is here repre- 

 sented, that in Soles (Fig. 15, u.), but an analogous phenomenon is found 

 amongst the eels. Ryder (46), in his treatise on this subject, gives a 

 drawing (Plate IV., Fig. 4) of the caudal region of the Common Eel, 

 which might well represent a further stage in the specialisation of the 

 flat-fishes to that shown here in Soles. The end of the notochord is 

 only slightly bent up, and is almost in line with the other vertebras. 

 The diphycereal form is obtained further in Murcena, which may be 

 compared with Ammopleurops and Ci/noglossus amongst the flat-fishes. 

 It is possible, therefore, that we should call the last degree of specialisa- 

 tion in these fishes the true diphycereal form, and not " pseudodiphy- 

 cercal " as was formerly suggested. 



The development of the caudal region in the flat-fishes has been 

 described by Agassiz (-4), M'Intosh and Prince (38), and other writers, 

 but for somewhat earlier stages than those required for the present 

 purpose. It seems clear, however, that Cunningham's view that the 

 hypural bone represents the ventral spines of a number of vertebrae 

 fused together (9, p. 39) is quite erroneous, and that the conclusions of 

 Huxley (28) and Ryder (46) for other forms hold good also for the flat- 

 fishes. In Fig. 16 is shown a drawing of the posterior caudal region of 

 a post-larval Sole 1 3mm. in length. The hypural elements which later 

 develop into the fan-shaped bone (Fig. 15, hy. 1 ) are not yet fused 

 together, but appear as four separate and distinct plates. Their form 

 and position show that they rather correspond to interspinous elements 

 than to haemal spines. Inasmuch as they fuse later, however, to the 

 base of the urostyle, and support distally the caudal fin-rays, we may 

 accept Huxley's conclusion that they are really composed of both 

 haemal and interhaemal elements. 



From the foregoing description of the caudal region it will be con- 

 cluded that the diverse forms and changes in structure which are found 

 amongst the flat-fishes may be of use as showing various stages in 

 specialisation, and thus of showing affinities between species in certain 

 groups, but do not aid in the general work of classification, as was 

 imagined by Sauvage. Only with the Soleidce do they become of such 

 use, because in the other groups the form and structure of the caudal 

 region are essentially the same, and even in the Soleidce so little is yet 

 known of the intermediate stages between the true Soles and such as 

 Cynoglossus that no classification based on this character could be 

 proposed. 



Characters of Most Importance for Classification. 



We may once again pass under review briefly the characters which 

 have just been described, in order to consider their respective merits for 

 the purposes of classification. Only those which gave promise of being 

 of importance have been mentioned here, and the choice of characters 

 is therefore not a large one. In a work such as this, further, which 

 endeavours to group the various genera according to their natural 

 affinities, it is self-evident that not one character alone but several 

 must be choseii. So long as the number of species known was rela- 

 tively small, and classifiers wished to have some external character or 



