498 SCIENCE PROGRESS. 



in the majority of the Palaeoniscidae and Platysomidee the 

 mouth is terminal, and although the tail is heterocercal it 

 approximates in most cases towards the homocercal condi- 

 tion in the fact that the ventral lobe is so large as to be 

 equal or nearly equal to the dorsal lobe which contains the- 

 upturned extremity of the vertebral column. 



In the Dapediidse, Pycnodontidse, Caturidse and Lep- 

 tolepididse, which are represented principally in Mesozoic 

 strata, we have obviously the forerunners of the homocercal 

 Teleosteans, and in them the homocercal condition of the tail 

 practically complete. 



If as I have suggested the benefit of the heterocercal 

 tail lies in the effect it has of throwing the hinder end of 

 the body upwards when it is moved, still the lateral motion 

 of the tail could not give a stimulus to the hypertrophy of 

 the ventral fin-rays. The origin of the modification must 

 be sought in some particular stimulation of these rays. 

 Such a stimulation might well be produced by the direct 

 movement of the rays themselves. There is in fishes a 

 movement of the rays by the muscles attached to them, in 

 addition to the movement of the caudal region by the lateral 

 trunk muscles. My theory therefore is that the develop- 

 ment of the heterocercal tail from the diphycercal was due 

 to the concentration of active exertion in the ventral fin- 

 rays, so that in course of generations these were hypertro- 

 phied and the others atrophied. The object of this 

 exertion is assumed to be the raising of the tail and the 

 depression of the head in fishes which fed upon the bottom. 

 The lobe having once been hypertrophied the further 

 modification in fishes which moved actively, even when 

 they ceased to feed on the bottom, and swam chiefly in a 

 horizontal direction, would take the form of a further modifi- 

 cation in the shape and position of the ventral lobe until the 

 homocercal condition was attained. 



It is a remarkable fact that although the Pleuronectidae 

 and Gadidse have generally been considered to be so 

 similar that they have been placed in the same order 

 Anacanthini, the structure and development of the tail 

 described above occur in the flat fishes but are entirely 



