[3] THE EVOLUTION OF THE FINS OF FISHES. 983 



will not only describe this larval condition, but also be in harmony with 

 the terms descriptive of the adult tail, will be, lophocercal and its deriv- 

 ative, lopliocercy, which will imply that the membranous tail is formed 

 merely of a fold of skin or epiblast, continuous with that on the axial 

 part of the tail, and that the inner surfaces of this fold are more or less 

 nearly in contact. 



The next term will apply to the next stage of development as indi- 

 cated by the most undiffereutiated median fin-system as seen in Cera- 

 todns, FrotopterKs, &c. Diphycercal and diphycercy will therefore imply 

 a condition in which the end of the axial column bears not only hypural, 

 but also epural intermediary pieces which support rays, as defined by 

 Huxley. Diphyceny may coexist with heterocercy^ and does not com 

 pletely vanish until all of the rays of the caudal are supported by hy- 

 pural spines only, or when an ideal hypocereal condition is attained. 



The next stage of caudal development is attained in the heter' cereal 

 condition, when the hinder end of the vert bral axis is flexed upwards, 

 and the words Jiettrocercal and heterocercy throujihout the ]>ret^ent i)aper 

 will invariably mean that such an upbeudiug lias hap])ened, even if it 

 involves the modification of but a single terminal vertebra. 



The words homotercal and homocercy will merely express the epaxial 

 and hypaxial symmetry presented by the fan-shaped caudal of Teleosts 

 and other fishes. It is the final stage of the evolution of that fiu. 



Gephyrocercy and gephyrocercal are terms which will apply to the type 

 of caudal structure appearing normally in only a few forms, such as 

 Mola and Fierasfer. The end of the urosome in these forms, together 

 with the posterior end of the chordal axis, is aborted. In consequence 

 of this, the hinder epaxial and hypaxial tissues concerned in the forma- 

 tion of rays and their supports are approximated or swung round over 

 the rayless interval existing over the stump of the axis, and by such a 

 secondary process of growth the caudal fin-rays are formed, together 

 with their interspinous supports. The interval between the vertical fins 

 biidged in this way leads to the formation of gephyrocercal tail. 



In the discussion of the neural and haemal arches of the caudal ver- 

 tebrae of heterocercal forms, it is also expedient to use some convenient 

 terms which will indicate without circumlocution what is meant in speak- 

 ing of the structure of these greatly modified elements. They will be 

 si)oken of as spineless if they are without dorsal or ventral spines, or as 

 dorsally spineless, or as ventrally two or three spined, as in the case of 

 the concrescent hypural bones, or as dipla- or triplacanthous. 



The caudal part of the axial column may be perfectly )nonospondylic, 

 that is, each of its vertebrae may have its dorsal and ventral spine, or 

 every alternate vertebra may have these aborted, and thus become re- 

 duced to a centrum only, or become diplospondyUc in the terminology of 

 Von Jhering, or every alternate centrum may be said to be sinneless. 



The meanings of the terms epural and hypural, as applied to indicate 

 the i>osition of the caudal apophysial elements in reference to the cen- 



