136 Records nf the Imluni Museum. [Vol.. XV, 



form and the inner skeleton, though modification of the latter may not 

 always be very strongly marked. The ventral contribution to the 

 caudal fin is larger than the dorsal, and in most cases markedly so, as 

 illustrated in the Ganoids Aciyenser, Polyodon and Amia. Associated 

 with this, as would be expected, the skeletal supports of the ventral 

 fin-rays are more strongly developed than those on the dorsal side. 



Though often less marked in Elasmobranchs than in the Ganoids, 

 from the base of the fin there is an upward bend of the axis which is 

 continued to the extremity, and when centra are present, they remain 

 distinct tliroughout this upwardly directed part. We thus see that 

 the main characteristics of the heterocercal tail are (1) an enlarged 

 ventral lobe compared with the dorsal ; (2) a bending upward of the 

 axis at its end ; (3) the retention of individual centra, when present, 

 to the end of the axis. 



Turning now to the homocercal type, there can be no question as to 

 its having succeeded directly the heterocercal form. Reference to the 

 skeletal structure of the caudal fin of Amia will readily shew that this 

 particular fin requires but slight modification to convert it into a homo- 

 cercal form, for homocercy is characterized by (1) external symmetry ; 

 (2) strongly asymmetrical internal skeletal structure by which the 

 majority of the fin-rays are always supported by ventral elements ; 

 and (3) the presence in the larval or adult stage of a urostyle which re- 

 presents a much shortened axis. 



The caudal fin of Amia has always been recognized as deserving of 

 some special distinctive designation, and it has usually been referred to 

 as hemi-heterocercal ; however since its distinctive feature is its close 

 approach to the homocercal form, I have elsewhere^ proposed that a 

 better term would be hemi-homocercal. 



It is well known that the earliest fishes of the Devonian period pos- 

 sessed heterocercal tail fins and that diphycercal forms appeared later. 

 Thus the evidences of embryology have been said to be at variance with 

 those of palaeontology ; it is true that the evidences mentioned do not 

 agree, but it would seem that there is no ground for denying that the 

 embryological data are indicative of the sequence of forms in the evo- 

 lution of the caudal fin. Professor Dollo^ in dealing with the Dipnoi 

 recognized the difTerence between the embryological and palaeonto- 

 logical evidences and suggested a way out of the difficulty. He argues 

 that since the caudal fins of the earliest fossil fishes found are heterocercal, 

 and later forms diphycercal, these later forms are secondarily diphy- 

 cercal, i.e., gephyrocercal. 



Now gephyrocercy implies the complete loss of the original caudal 

 fin elements, a great reduction of the axis having brought this about ; 

 and also that the new tail fin is in reality the result of a " bridging over " 

 of the gap thus produced, by the dorsal and anal fins, each having an 

 equal share in the formation of the new fin. Fierasfer is the best illus- 

 tration of a gephyrocercal fin ; the larva is known to possess a long 

 filamentous termination to the chordal axis, which is entirely lost in the 



1 Proc Zool. Soc. London, October 1910, where most of the caudal fins mentioned in 

 this paper are figured. 



2 The Phylogeny of the Diimoi. 



