124 



Part III. — Twenty-fourth Annual Report 



that among specimens of luscus the anus was situated exactly below the 

 beginning of the first dorsal in three cases; in two cases it was in front of 

 that point to the extent of 1 per cent, of the length of the fish, and in 

 seven cases it was behind the beginning of the first dorsal to the extent 

 of 2 per cent, of the fishes' length. 



In luscus, then, the anus is situated below the beginning of the first 

 dorsal fin. , 



In minutus the anus occupies a very different position in relation to the 

 first dorsal fin. Of 34 examples of minutu*, in 31 the anus was below 

 the second half of the first dorsal, in one case it was below the middle 

 point of the fin, and in two cases it was just in front of the middle point 

 to the extent of 1 per cent, of the length of the fish. In minutus, there- 

 fore, the anus is under the second half of the first dorsal fin, i.e., from 

 the middle of the fin to the hind limit of its base. The amount by 

 which the anus was in front of the middle point of the fin would not 

 be noticeable to the eye, and for all practical purposes it would be 

 regarded as coinciding with the middle point. 



In esmarkii the anus is below the second half of the first dorsal fin ; 

 this was found to be the case in all the fishes in which these characters 

 had been recorded — viz., 56 in number. In one of these the anus was 

 exactly below the end point of the fin-base. In esmarkii, then, the anus 

 is situated below the second half of the first dorsal fin. 



The comparative depth of the members of the two species, luscus and 

 minutus, has been insisted on in a specific description. It is not, however, 

 very easy to reduce this relation to a specific character. In some speci- 

 mens it is well marked; but while luscus is always a deep fish, minutus is 

 variable in this respect. Sometimes in a collection of minutus it is 

 possible to divide them up into two lots differing markedly in the depth 

 of body, and in the deeper lot simulating in depth luscus, It is necessary 

 to know the range of variation in this character as determined by the 

 measurements of various examples. If, therefore, we examine the 

 average dorso-ventral height of each species, as Vt out in Table XIII, we 

 find the average height at the anus is, for luscus, 26 per cent ; for minutus, 

 21 per cent.; and for esmarkii, 18 per cent, of the length of the fish. In 

 luscus, however, the depth ranges from 28 per cent, to 22 per cent., while 

 in minutus the limits were 24 per cent, and 18 per cent. The relation 

 between the two species might be summarised thus : — Where the dorso- 

 ventral height reaches one-quarter or more of the length of the fish, we 

 have to deal with luscus ; where the same character is as small as 

 one-fifth or less of the length, the fish is probably minutus or esmarkii; 

 but where the relation lies between one-quarter and one-fifth, either of 

 the two first species may be represented. It is evident, then, that a 

 character such as this is only of small importance for specific diagnosis. 



The Lengths of the Bases of the Unpaired Fins. 



The inter-relations of the sizes of the different fin-bases may be here 

 considered. In the following Table the average size of the fin-base, 

 expressed as a percentage of the length of the fish, and its range of varia- 

 tion is given for each of the three species. 



Species. 



1 I). 



2 D. 



3 D. 



! 



Max. 



Av. 



Vari- 

 ants. 



Min. 



Max. 



Av. 



Fori- 



ants. Mm " 



Max. 



Av. 



Vari- 

 ants. 



Min. 



G. luscus 

 G. minutus 

 G. esmarkii 



13 

 13 

 15 



11-6 

 11 g 



10 



34 

 56 



10 

 Id 

 9 



27 



26 

 25 



85 



'it 



10 24 



33 22 

 48 ; 20 



15 

 17 



18 



12 7 



15 2 



16 5 



10 

 34 

 47 



11 

 13 

 12 



