HELIOMETRINAE i35 



hoffeniamis differing from P. kerguelensis in having: (i) longer cirrus segments, (ii) long 

 and slender axillaries and second brachials, (iii) no lateral notch between the radial and 

 the costal, and (iv) well-developed side-plates along the pinnule ambulacra. 



Vaney (1910, pp. 158-62, figs, i, 2) described the new species P. joubini, from one 

 specimen from the west of Graham Land. It is described as being distinguished from 

 P. kerguelensis by having clusters of spines on the lower brachials, by the position of its 

 syzygies, and by the possession of a naked dorsal pole. Like P. kerguelensis it has 

 rhombic axillaries and lacks an ambulacral skeleton ; its cirri are tolerably like those of 

 P. vanhoffeniamis. Vaney considered it to be nearer the former than the latter. 



Clark (191 5) regards both Minckert's and Vaney 's species as invalid, believing the 

 specimens upon which they were based to have been immature P. kerguelensis. 

 Mortensen (1918, p. 19) is not satisfied that Clark is correct, more particularly in re- 

 garding P. vanhoffeniamis as identical with P. kerguelensis ; nor is he convinced that the 

 presence of an ambulacral skeleton is a sign of immaturity, a doubt shared by Grieg 

 who is otherwise inclined to accept Clark's opinion (Grieg, 1929 a, p. 4; 1929 b, p. 4). 



My examination of the present collection of 163 specimens from low and high lati- 

 tudes in the Falkland sector of the Antarctic zone, and of thirty-eight specimens from 

 the Ross Sea, and of one or two taken by the ' Challenger ' near Kerguelen and Heard 

 Island, has convinced me that there is only one, very variable, species. 



Description. In the great majority of specimens there are 10 rays and 20 arms but 

 individuals with 12, 16, 17, 18, 19, 22 and 23 ^ arms occur. The arms of large specimens 

 are commonly 180-200, and may be up to 250, mm. long. 



The centrodorsal is usually a large high rounded cone, but it is variable. It may be 

 more sharply conical; in smaller specimens it is often a moderately low cone. It is 

 closely covered with cirri which may, exceptionally, number 200 or more. The cirrus 

 sockets are arranged in slightly irregular alternating rows, so that those of alternate 

 rows make slightly irregular columns. The dorsal pole is usually of medium size, smooth 

 and rounded. In large specimens it is sometimes sunken and rough. It may, in either 

 large or small specimens, be a sharply triangular or a rough truncated pillar-like pro- 

 jection. It may be very small or absent : though this is most often so in small specimens 

 it may be so in large. 



Cirri : up to CC or more. There is always a sharp contrast between the apical cirri and 

 those of the more ventral part of the centrodorsal : the latter are long, of up to 47 or, 

 exceptionally, 65 segments ; the former very much shorter and of fewer segments. The 

 contrast is greatest in large specimens ; the longest peripheral cirri may be three or more 

 times as long as the shortest apical cirri. They may be up to 130 mm. long. 



The segments of the cirri of different specimens vary greatly in length. I have seen 

 none in which those of one cirrus are so nearly equal to one another as in Carpenter's 

 figure. The range of variation is illustrated by the extremes shown in Fig. i ; ^i and a^ 

 are a long and a short cirrus of a specimen with cirri of short segments, and a^ and a^ are 



1 Mortensen, 1918, p. 19. 



