1910.] CAUDAL FIX OF THE TKLEOSTOMI. GOij 



General remark-'^ on ilie Heteronii. 



All fishes belonginof to tliis sub-order, wliethei- fossil or living, 

 appear to be either eel-like or deep-sea forms ; from this it is to be 

 expected that the tail-fin when present is similar tlnoiif'hout, 

 i. e. gephyrocei'cal. 



Catosteomi. 



( Jasterosteus AruLEATUS. (Hemibranchii.) 



The caudal fin of this fish, was well descri})ed b}' Huxlev in 

 1859. The vertebral column ends in a long urostjde which is 

 fused along the dorsal edge of the last hypural bone. There is a 

 small neural arch to the last centrum, which Huxley figures as 

 being continuous with it, though there is a line of junction 

 separating arch and urostyle. The penultimate vertebra possesses 

 an epural and hypurals ; between the epural and the urostyle is 

 a dorsal caudal ladial (the " epural " of Huxley). 



Oentriscus scolopax. (Hemibranchii.) 



The urostyle here is suppressed and the " cone " of the last 

 vertebra shows only the slightest suggestion of upturning. Two 

 hypurals are attached and equally well fused to the last centrum. 

 An epural and hypural are attached to the last vertebra but one, 

 and both show well that they are composed of a neural and a hasmal 

 arch i-espectively, together' with a radial. A single dorsal caudal 

 radial is situated between the epural and the last hypural. 



Thus the caudal fin of Centt'iscics is highly specialized in 

 contiast with the nearly related form Gasterosteus. 



Syngnathus acus. (Lophobi\anchii.) (Plate XLVIII. fig. 15.) 

 The caudal fin of the Pipe-fish presents a perfectly symmetrical 

 appearance, but morphologically it is quite uns3'mmetrical since all 

 the eight dermotrichia are supported by hypaxial elements con- 

 sisting of two expanded hypural bones. Although the young- 

 stages show a distinct urostyle, this structure is absent in the 

 adult. There are no complications of any kind and the whole 

 caudal structure is exceedingly simple. The spinal cord can be 

 traced to the distal end of the hypural bone. On account of the 

 presence of a urostyle in the larva and tlie hj-paxial natuie of the 

 fin, the tail must be classed as homocercal ; moreover, it is a 

 highly specialized form of homocercy, 



SiPHOXOSTOMUS ROXDELETii. (Lophobranchii,) 



Precisely the same structure is found in this species as in the 

 preceding type, hence it is unnecessary to i-epeat the description. 

 The similarity is no doubt associated with the similarity in habits 

 of the two fishes. 



Proc. ZooL. Soc— 1910, No. XL. 40 



