fiOO Mit. K. II. wiirrr.iKji'.si: ON riii': [Apr.o, 



1 .am Jiwave. this ilitt'ereiice is universal aiiion^' Ti^lcosts, but, as 

 stated ahove, as yet I do nob assert it to be without exc'ei)tion. 

 'J'he matter is called attention to, in order to make it (}uite clear 

 that fin-rays of a caudal fin may be supported directly by un- 

 d()ubted radials. 



Concerning other features of the caudal tin of J'lolosus, the hust 

 dorsal ray is supported by an epural which is attached to the 

 jienultimate vertid)ra, and which shows the union of two elements, 

 a i-a<lial and a neural arch. Hypaxially is a lar^'e hypural firmly 

 united with the urostyle, and bearing ten dermotrichia. In the 

 specimen figured the penultimate centrum })ossesses two ventral 

 njipendages, a hypural supporting one ray, and a luemal arch. 

 CJuriously enough, the next anterior ray is without a support and 

 seems to anticipate a similar state of affairs ventrally as dorsally : 

 such, however, is not the case, for each ray preceding is a typical 

 anal fin-ray supported by a true radial. 



Hence the caudal fin of Plolosas is a very highly specialized 

 organ, totally different from the preceding type, Clarias. 



Syxodoxtis schall. (Siluridje.) (Plate XLVII. fig. 8.) 



The last six vertebrte form the basis of support for the deeply 

 cleft caudal fin in this fish. The urostyle is lost and the last 

 centrum is firndy fused to the last two l)road hypurals, which 

 occupy a terminal position. The hypaxial fin-rays are mainly 

 supported by these two hypurals, but in addition there are five 

 others, one to each of the precedinj; five vertebi-ie ; these two 

 hvi>urals afford excellent evidence of the coalescence of radials 

 and hipmal arches in their formation. E[)axially are four ei)ui'als, 

 one to each of the four vertebi'ic })receding the last two ; the.se 

 epural bones, like the hypurals, show the duplicate nature well ; 

 it will be seen that the neural ai'ch of the penultimate vertebra; 

 jiud its corresponding radial have not fused, but a little forward 

 movement and then the fu.sion with the neural arch, on the part 

 of the radial, woidd produce an exactly similar structure to the 

 preceding epurals. This cauilal liii provides one of the best 

 illustrations of Ihc fusion of radials with arches in the formation 

 of epurals and hypurals. Needless to say the fin is a highly 

 S|)ccializcil hoinorcrcal type. 



(Ii-neral reuiarls on tin' SiJnr'nIir, 



The .single family Siluridai, therefore, iiu-ludes great diversity 

 in the form of the (laudal fin structure, from a very generalized 

 to a highly .sjji'cialized type. 



A IM) DES. 



.■\\(!iiM,.\ VI i.cAitis. (Mur;enida\) 



llaviiig iil)tain('l a good series of stages. ! h.isc lnwMi able 

 to trace the development f)f the caudal tin in Ainindit from the 

 ij'jdoccphalaa stage to the adult. 



