— 16 — 



will fall clear of that of the two others , this can only be determined by examina- 

 tion of a greater number of specimens than I have had at my disposal. 



With regard to the length at which the full number of rays appears, it may 

 be mentioned that larvae of A. vulgaris about 38 mm long had apparently already 

 their half number of rays, while a specimen 42 mm in length showed 16 rays. 



5. Number of caudal rays. 



The number of caudal rays is the last of the characters to be here dealt with 

 at length. The caudal rays, it is scarcely necessary to state, are those emanating 

 from the last two hypurals. Their number has been found to vary from 8 (a single 

 specimen with 7 should perhaps be regarded as abnormal) to 12. Table VI A 

 shows the total number of caudal rays, whereas in Table VI B, distinction is made 

 between the rays of the last hypural [d] and those of the penultimate (w). 



It will thus be seen, that the number of caudal rays is very nearly identical 

 with that of the branchiostegal rays ^(cf. Table IV), the range of variation also 

 being the same. There is, however, this essential difference, that the frequency of 

 the separate variants exhibits an entirely different character. If we compare, for 

 instance, the variation for the whole of the A. vulgaris material (cf. Table IV A, 

 1 — 3 and Table VI A, 1 — 3) we find that about 50 7o of the branchiostegal variants 

 fall to other figures than that of most frequent occurrence, whereas in the case of 

 the caudal rays, only about 10 7o of the variants lie outside the corresponding 

 figure here. Thus, in spite of the fact that the range of variation is in both cases 

 the same, the number of caudal rays actually varies far less than that of the 

 branchiostegal rays, and is therefore less adapted to the purpose of specific or 

 racial distinction; in other words, it is of less value in classification than the latter. 

 Table VI A further shows, that in spite of the very considerable number of speci- 

 mens, no charactei'istic difference between A. vulgaris and A. rostrata has been 

 i-evealed ; still less, as might be expected, between the different samples of European 

 eels. In all samples the figure 10 is absolutely predominant over all others, and 

 we do not here, as with the branchiostegal and pectoral rays, find that even the 

 investigation of smaller samples quickly furnishes a characteristic view of each 

 species, with a constant order of precedence as regards frequency of the separate 

 variants. It would seem that the average number of rays is lower in the case of 

 A. rostrata than for A. vulgaris; the difference is, however, so slight as only to be 

 apparent in the second decimal figure. A noticeable peculiarity in the table is the 

 fact that a greater difference may be found to exist between two samples af A. vul- 

 garis than between the one of these and the A. rostrata sample. From all this it 

 is evident that the number of caudal rays is a feature of little systematic value, 

 and that in any case, a far greater quantity of material would be required in order 

 to demonstrate with certainty a slight difference between the species. Nor does the 

 distinction made in Table VI B between the rays of the last and those of the 

 penultimate hypural seem to indicate any characteristic difference between the 

 species; in addition to which, the material thus drawn up cannot be treated and 

 judged by the usual statistical method. 



