of some Teleostean Fishes. 357 



numerous in the adult than in the young and to increase by 

 development of rays posteriorly ; in large adult Tarpon 

 (fig. 2 b), 1500 to 1^00 mm. in length, I count from two to 

 four inserted on the uppermost hypaxial ray, but in young 

 examples (300 to 400 mm.) one or two. The young differ 

 from the adult fish also in that the articulations of the 

 principal caudal rays are far less numerous; but the anterior 

 epaxial rays do not increase in size at anything like the same 

 rate as the hypaxial rays, and are relatively much larger in 

 the young, so that the longer of them are articulated in the 

 young, but, from their more proximal position, spinous in the 

 adult. In adult Megalops cypriuoldes (300-350 mm.) the 

 posterior epaxial rays vary in number from six to nine. These 

 posterior epaxial rays or "fulcral spines" are wanting in Elops 

 and in all other homocercal fishes, but the anterior epaxial rays 

 persist in most and are usually articulated, so that we appear 

 here to have a case of the reconversion of spines into articu- 

 lated rays. 



In the Elopidce the principal caudal rays (i. e. all the 

 hypaxial rays except the lower graduated ones) are 10 in 

 number, and 17 of these are branched. This seems to be the 

 case in a large number of families, e. g. the Leptolepidse, 

 Chanidse, Albulida?, Salmonidse, Clupeidse (except Coilia), 

 Esocida?, Aulopid;e, Cynrinidse, Berycidse, Lampridye. In 

 the Serranidse, Sparidse, Carangidse, &c, the number appears 

 to be constantly 17 (15 branched), and in various more 

 specialized groups is still further reduced. Without laying 

 too much stress on this character, I may say that I have 

 found it of considerable value in determining the affinities of 

 some types of doubtful systematic position. 



In Elops, but not in Megalops, there is an oblong bony 

 scale above and below, partly covering the first upper and 

 lower rays; this was present in the Leptolepida;, and persists 

 in AHula, Aulopus, and some Clupeids. 



I have already mentioned Mr. Whitehouse's paper on the 

 caudal fin of fishes, and I have shown that the element which 

 he terms " urostyle " in Clupea is formed by one or more 

 displaced posterior neural arches or " uroneurals " ; the homo- 

 cercal fin should not then be defined by the presence of a 

 urostyle formed by the fusion of upturned vertebrae, but by the 

 modification of posterior neural arches into uroneurals which 

 functionally replace and so lead to the suppression of the centra 

 of the upturned vertebra;. There are some otiier points which 

 call for co nment ; thus I think the term "epaxial basalia" 

 preferable to and more correct than " dorsal caudal radials," 

 and I cannot agree that Fierasfer is " gephyiocercal." 



