FINS OF GAlfOIDS AND TELEOSTS. 543 



only from tlie anterior surface of its supporting radial element, 

 but also from opposite sides of the intermuscular septum in 

 which the lamina of the next anterior element is developed, and 

 thence run obliquely backwards to their insertion into the base 

 of the spine or fin-ray as the ease may be *. 



It may be mentioned that Mivart [6] seems to have entirely 

 misunderstood the nature of the fin-supports of Polypferus. In 

 his description of tlie dorsal fin he says : — •' This fin is supported 

 by radials which give off on one side small secondary rays pro- 

 ceeding dorsad and postaxiad " (l. c. p. 458 ; also pi. Ixxix. fig. 6). 

 It is clear from the use of the term " radials," as well as from 

 the accompanying figure, that Mivart is here describing the 

 spines and soft rays of the series of finlets, and has entirely 

 overlooked the true " radials," which are situated beneath and 

 support the finlets. It is probable that this usually accurate 

 morphologist only had access to an imperfectly prepared skeleton. 

 Anal fin. — The anal fin of Polypterus consists of six radial 

 elements (fig. 6), of which the first (r.e}) is a simple bony rod, 

 slightly thickened and tipped with cartilage at its ventral end. 

 The remainder are bisegraental, each consisting of a ventral 

 segment (v.st.^ similar to the simple segment of the first, and a 

 slender styliform dorsal segment (d.st.). The segments of all 

 the elements are well ossified, with the exception of the dorsal 

 segment of the last one, which is cartilaginous. The distal 

 extremities of the ventral segments are in close contact with 

 one another so as to form a continuous, even if somewhat ir- 

 regular, peripheral margin. The first five radial elements are 

 situated in front of the first complete haemal arch (h.s.), to the 

 spine of which the sixth is attached by ligament. 



Thirteen soft in ulti articulate and slightly branched fin-rays 

 are supported by the six radial elements, the ventral segments 

 of the latter being embraced for a third of their length by the 

 cleft rays. Each element, however, obviously contributes to the 

 support of at least two fin-rays. 



In older and. larger specimens than that described above, 

 the ventral divisions of the difi'erent radial elements are not 

 merely larger and relatively more expanded towards their distal 

 extremities, but the three anterior ones, which are longer than 



* Eyder (2- PI. v, fig. 2) gives a figure, "from Agassiz's 'Poissons Fossiles,' 

 modified after KoUiker," in which these structures are described as " non-ray- 

 bearing interspinous epural elements." 



