30 • DISCOVERY REPORTS 



in several species, however, these setae are short and few in number, e.g. A'^. articulare, 

 N. capense, N. co^npachim. In group I, the setae, as a rule, are either few and fine or 

 wanting; in a few exceptional cases (e.g. N. adareanum, Gordon, 1932, p. 99, fig. i a) 

 a few strong setae are present. 



With regard to the chela there is so much variation in each group that no generaliza- 

 tion ^ can be made. The palm may be shorter than wide (A^. proceroides and A^. neumayri, 

 n.sp. (Figs. 9 and 29 d), or as much as 3-4 times as long as wide (A^. biarticidatum, 

 N. longicoxa, N. procerum, Fig. 32 a and Gordon, 1932, figs. 7 and 9 a). The number 

 of spinules on each finger of the chela is very variable (6-1 10, see Table III); in a few 

 instances there are many more spinules on one finger than on the other, e.g. A'^. longicoxa 

 (70 and 90-1 10). 



Palp. Segments 2-5 are long and subequal in a few exceptional cases (A^. charcoti, 

 Nymphon, sp.? (St. 181), A^. lanare and A^. hrachyrhynchiim — the last is not closely 

 related to the other species). In all other forms, segments 2-5 are of difl^erent lengths 

 and segment 2 is usually longest although the third may not be much shorter. As a 



general rule the ratio ^-^ — , — ^ ^ — 5 ijgg between 0-7 c; and 1-3 ; in five species 



length of segment 2 / j j r 



of group II it is less than o-6 (usually less than 0-5, see Table III). 



Number of spines on oviger. With one exception (A^. adoreamim, Hodgson, 1907, 

 pi. iii, fig. 3 b) each of the four terminal segments of the oviger bears a number of 

 denticulate spines. Loman (1923, p. 11) has shown that the total number of these 

 spines remains within certain definite limits for each species. This total may be less 

 than ten or as high as seventy in exceptional cases (see Table III); also it may be the 

 same in several unrelated species, e.g. 30-38 in A^. tenuipes, N. hamatum and A", brachy- 

 rhynchmn of group I and A'^. anstrale, N. compoctiim and A^. orcadense of group II. 



Third leg. In all the species of group II and most of those of group I a (in which 

 segment 5 of the male oviger is long, slender and distinctly curved) and in A^. charcoti 

 the second coxa is approximately equal in length^ to the sum of the first and third. 

 In all the other species of group I the second coxa, especially in the male, is much longer. 

 The femoral gland openings in the male vary greatly in number; in several instances 

 they have not been detected. In the female, the femur is at least five times as long as 

 wide in all species of group I, A^. pfefferi and A'^. adareanum excepted. In group II it 

 is usually less than four times as long as wide, but is nearly five times in each of the 

 atypical forms. 



The second is longer than the first tibia, as a rule, in group I, and the reverse holds 

 for most of the species in group II (Table III). 



1 Loman (1923, p. 10) says that in " Nymplioii" the "hand" is not conspicuously broadened distally, 

 whereas in " Chaetonymphun" it is much broadened distally, but I am afraid I do not quite understand this 

 distinction. The palm in Figs. 32 and 29 d is not much broadened distally, yet these species are typical 

 " Chaetonymphom" . The only Antarctic species in which the palm is broadened distally is TV. rmdtidens, n.sp. 



(Fig- 35 d). 



^ The dorsal measurements are always given for the segments of the leg so that the third coxa is much 

 shorter than if measured ventrally. 



