b) lateral line double, ventral and niediolateral: 



L. esmarkii Coll. 



L. ezidipletirostictus Jensen. 



/.. pallidus Coll. 



L. platyrhiiuts Jensen. 



c) lateral line single, niediolateral: 



L. rossi Malmgr. 

 /,. liltkcnu Coll. 

 L. veticulatus Reinh. 

 L. soniiiudus Reinh. 

 Z. agiiDStiiS Jensen. 



With regard to the second group, it must be mentioned that the niediolateral branch of the 

 lateral line is frequently indistinct in two of the species, nanieh- L. esmarkii and L. pallidas, so that 

 it is often only after a very careful study of a large number of specimens that one can rightly deter- 

 mine their position — this holds especially for L. pallidus, which stands as a sort of transition form 

 between groups a and b, so far as the lateral line is concerned ■). 



For the rest, the groups a and b seem in other respects also, to stand near to one another and 

 to form together a separate subdivision contrasted with group c. Thus, in groups a and b the tail is 

 relatively long, whilst the head and trunk together (or the distance between the snout and the anus) 

 most often amount only to 36,5 — 45 % of the total length (sometimes reaching 47% in males of L.frigidus) ; 

 in group c on the other hand the tail is relatively short, whilst the head and trunk together amount 

 to 43 52 '^/o of the total length. Groups a and b may therefore be described as long- 

 tailed, group c as short- tailed. 



In close connection herewith is the number of rays in the unpaired fins. This is throughout 

 laro-er in the long-tailed species than in the short-tailed, and ver>- naturally so, since the anal fin 

 entirely and the dorsal fin for the most part, belong to the tail. In groups a and b the number 

 of rays in the dorsal fin is 94 — 118, in the anal fin Si — 102^), in group c the number 

 is 90—97 for the dorsal fin, 70—78 for the anal'^). 



It will appear from the foregoing that the groups of Liitken based on the lateral line only, 

 are not of equal value, but that the groups having the ventral and ventral-mediolateral lines form 

 too-ether one subdivision over against the group with the niediolateral lateral line. For practical 



■) whilst speaking of the lateral line, it should be mentioned that one finds in some of the species, and in all three 

 jjroups, a shorter or longer series of pores placed relatively remote from one another on each side of the back an indication 

 of a dorsal lateral Una. 



2) Both here and in the special portion of die work, the upper rays of the tail fin are reckoned witli the dorsal fin, 

 the lower rays with the anal fin. since the unpaired fins pass without break right round the tip of the tail. — I think it not 

 unnecessary to remark that all my statements of the number of fin-rays are based on my own obser\'ations , which do not 

 always agree with those given in the literature. 



i) Probablv the number of the vertebra; will also be greater in the long-tailed than in tlie short-tailed species, but 

 the material in my hands is too little to allow any certain conclusions to be drawn in this regard; in four species of 

 groujjs a and b I have counted 98— 118 vertebriC (L. vahlii 98 — 116, L. frigidus 103—107, L. eudipleurosticlus 106 and L. 

 esmarkii 115—118), in two species of group c {L. relicu/atus and L. semiiiudusj 93—96 vertebrae. 



