252 DISCOVERY REPORTS 



and proximal processes: in E. lucens the terminal is longer than the proximal, in E. vallentini the 

 proximal is longer than the terminal, and in E.frigida it is proportionately longer still. If the species 

 form a true series one would expect that if any one of them had a structure, possessed by all three, 

 more strongly developed than the other two it would be an end member. E. frigida has the distal 

 tooth between the membranous lobes of the proximal process more strongly developed than in 

 E. vallentini and E. lucens ; I have not found a secondary process on the median lobe in E. lucens as 

 in the other two. 



E. vallentini has, however, other characters sharply distinguishing it from E. lucens and E.frigida, 

 which make it appear unlike a species standing midway between them. It is larger than either of 

 them. It has a dorsal process on the posterior margin of the third abdominal segment which caused 

 Hansen to include it in his Group c. The front end of its carapace and the rostrum are differently 

 shaped to those oi E. lucens and E.frigida which are similar to one another (Fig. 31). E. lucens has 

 a moderate or small, E.frigida a very small triangular lobe from the end of the first segment of the 

 antennular peduncle ; E. vallentini has a very characteristic lobe, large and rounded (Fig. 32). It has 

 a high rounded dorsal keel on the third segment of the peduncle, whereas E. lucens and E. frtgida 

 have lower keels with straight upper margins (Fig. 33). 



a '/be 



Fig. 33. Third segments of the left antennular peduncles of the species of the southern group showing the 

 shapes of the dorsal keels, etc., from the outside, a, E. lucens; b, E. vallentini; c, E. frigida; d, E. superba 

 (after Hansen) ; e, E. crystallorophias. Not on the same scale. 



I am inclined to attach more importance to the structure of the male copulatory organs which 

 places E. vallentini between E. lucens and E.frigida, as it occurs in space, than to size and the other 

 structures — the abdominal spine, the sharp rostrum and the antennular lappet — which make it 

 appear more distant from E. lucens and E. frigida than they are from one another. 



Of these three species E. frigida alone occurs in the Antarctic like E. superba and E. crystallo- 

 rophias, its range overlapping that of the former but not that of the latter. It would be difficult to 

 show that either of them is more closely related to it than the other. On the whole E. crystallorophias 

 shows the fewest differences. The terminal process of its copulatory organ is divided distally into 

 two parts, and the inner edge of the outer is serrated like the hinder edge of the outer, to which 

 I think it corresponds, in E.frigida. The proximal process is more like that of E.frigida in general 

 shape and in the possession of a strong secondary process than is that of E. superba. The antennular 

 peduncle of E. crystallorophias, with no lobes on the first and second segments, is less unlike that of 

 E. frigida than E. superba's is. But there are big differences too, and no good reason that I can see 

 for regarding E. crystallorophias as intermediate between E. frigida and E. superba. 



To sum up, I think the structural evidence may be sufficient to represent the coloniza- 

 tion of southern waters by this well-marked group or branch of the genus Euphausia in 

 the following diagrammatic way : ^ ^^^^^^^^ 



'1 



E. vallentini 



I 

 E. frigida 



/ \ 

 / E. crystallorophias 



E. superba 



