314 BULLETIN 82, UNITED STATES NATIONAL MUSEUM 



of these the centrodorsal, as viewed from the dorsal side, is above the level of the radial 

 pentagon, but in the second it is relatively much lower, so that its surface is flush 

 with that of the radials the proximal edges of which are convex, in correspondence 

 with the stellate outhne of the centrodorsal. In a similar case the centrodorsal 

 is pentagonal. 



The effect of the complete withdrawal of the centrodorsal into the radial pentagon 

 is to make it entirely invisible in a side view of the animal, while the dorsal surface of 

 the radial pentagon becomes very deeply hollowed for its reception instead of being 

 slightly convex, as is more usually the case. 



Carpenter noted that in this species the arms are most frequently 20 in number. 

 Among 50 individuals which he examined he found only 1 which did not have the 

 full complement of 10 IIBr series. In this one postradial series was entirely without 

 them, so that the number of arms is reduced to IS. Of the other specimens studied, 

 32 had 20 arms, 9 had 21 arms, 7 had 22 arms, and 1 had 23 arms. The IIIBr series, 

 when present, are alwaj^s 2 (1 + 2) like the IIBr series. 



Carpenter discussed the peculiar arrangement of the syzygies at the base of the 

 arms. The first brachial syzygy is always between brachials 1+2, even when IIBr 

 series are absent. On the 2 outer arms of each postradial series the next syzygy is 

 between brachials 3 + 4, so that, according to Carpenter, the normal sequence of 

 syzygies on these 2 outer arms is between brachials 1 + 2, 3 + 4, 12 + 13, and 17 + 18, 

 whereas on the 2 adjacent inner arms it is between brachials 1 + 2, 9 + 10, and 14 + 15. 



The amoimt of carination of the lower pinnules varies considerably. As a general 

 rule the pinnules of the first pair (Pi and Pa) have their basal segments somewhat 

 produced toward the dorsal side, and in the next 2 pairs of pinnules the second and 

 third segment have rather prominent keels, traces of which are sometimes visible as 

 far as the twelfth or fifteenth brachial. The terminal comb, which is very well 

 developed on the basal pinnules, becomes gradually smaller and disappears at about 

 the sixth or seventh brachial. 



The visceral mass is somewhat readily detached from the calj'x, and the Chal- 

 lenger occasionally dredged it in an isolated condition. It is not so completely 

 plated as that of Comatula Solaris is, for the ambulacra are unprotected and the 

 interradial areas are covered by larger and more nodular plates. But both species 

 may have the calcareous deposits considerably reduced in extent, though they are 

 rarely entirely absent. 



In Professor Doderlein's specimens from Thursday Island, Pi in individuals of 

 about equal size, with an arm length of from 105 to 80 mm., is about 21 mm. long, 

 with a comb occupying from one-third to one-fourth of the entire length. Most of 

 these specimens have 20-21 arms, but one has 27. The position of the syzygies 

 is quite variable. The syzygies on 2 rays of the individual with 27 arms are as 

 follows (syzygial pairs being counted as single ossicles) : 



