10 "TERRA NOVA" EXPEDITION. 



way distinctive are the " perfectly -segmented " body and the " nine-jointed " palps.* 

 The first character is shared by two species of Colossendeis, C. articulata and C. doflelni 

 of Loman (1908, p. 22, and 1911, p. 4). As regards the second character, N. germanica 

 agrees in this, but, apparently, in no other respect, with the species described below as 

 C. wilsoni. If the genus were to be retained it would be hard to decide whether it 

 should include C. articulata and C. dofleini on the one hand or C. ivlUoni on the other ; 

 it could not embrace all three. 



A laro-e number of species of Colossendeis have lieen described, most of them from 

 a very small number of specimens. They are distinguished mainly by proportional 

 differences of measurement, and there is reason to beUeve that some of fehem would not 

 survive a critical revision based on adequate collections. In the absence of such a 

 revision it is necessary, before venturing to describe any additional species, to attempt 

 to reduce to some sort of order those already known. Bouvier has made a beginning 

 by dividing the species into two groups according to the relative lengths of the distal 

 segments of the legs. In the first or " longitarsal " group the carpus, propodus, and 

 claw together measure at least three-quarters of the length of the second tibia ; in the 

 " brevitarsal " group the proportion is always very much less. Proceeding on these 

 lines, the following key may be offered for the "longitarsal" group, which includes all 

 the Antarctic species. 



Key to the " Longitarsal " species of Colossendeis. 



a. Sixth segment of palp more than three times as long as thick. Proboscis 

 distinctly longer than trunk. 



a'. Trunk segmented G- artkulata, Loman 



U. Trunk not segmented. 



a". Lateral processes in contact. 



a'". Seventh segment of palp longer than eighth. Eyes absent . G. prohoscidca (Sabine) 

 &'". Seventh segment of palp shorter than eighth. Eyes present . G. scotti, sp. n. 

 h". Lateral processes separated. 



ft'". Seventh segment of palp equal to eighth . . . . G. australis, Hodgson 



i G. media, Hoek f 

 \G. hrevipes, Hoek 

 &'". Seventh segment of palp distinctly shorter than eighth. 

 a"". Eyes absent. 



a'"". Proboscis dilated distally G. orcadensis, Hodgson 



„, ^ ,...,,,..-, T, {G. anqusta, G. O. Sarst 



h . Proboscis dilated m middle . . . . . • ) /i v • tt v 



* I learn from Mr. Hodgson that he does not accept the view of Bouvier (1913, p. 37), according to 

 which only nine segments are counted in the palp of normal species of Golossendeis. Bouvier, no doubt 

 rightly, excludes from the enumeration the " saillie basilaire " of the palps, which is generally counted as 

 a segment, although it is precisely similar to the process (never reckoned as a segment) that lies alongside 

 it and carries the oviger. Loman is said by Bouvier to have been the first to call attention to this point, 

 but I cannot discover the passage in which he did so. 



f G. media and G. hrevipes were described by Hoek as doubtfully distinct from C. gracilis, which 

 again is identified by Miibius with G. angusta. Meinert's observation (1899, p. 59) as to the difierences in 

 form of the palpal segments in immature and fully adult specimens of the last-named species deserves to 

 be noted as having possibly a wider application. 



