662 BULLETIN 82, UNITED STATES NATIONAL MUSEUM 



In 1 of the specimens collected by the Challenger at Zainboanga, 1 of the IBri is 

 absent, so that the IBr axillary rests directly on the radial. Of the 2 IIBr series 

 borne by this axillary, 1 similarly consists of a single axillary ossicle, and the other is 2. 

 The other 8 IIBr series are all 4 (3 + 4); all of the IIIBr series and the single IVBr 

 series, which is on the outermost side of a ray, are 4 (3 + 4). There are 32+ arms. 

 The centrodorsal is pentagonal with irregular sides and apparently bears II inter- 

 radial cirri. 



In a specimen with 16 arms from the Danish expedition to the Kei Islands 

 station 20, one of the IBr axillaries is represented by 2 ossicles situated side by side, 

 with their distal borders making an obtuse angle with each other. The dividing line 

 runs from just to the left of the angle proximally and to the right to a point one- 

 third of the distance from the right to the left side of the distal border of the IBri. 

 Where this line reaches the distal border of the IBri there is a prominent triangular 

 area, and just to the left another in contact with it, these represcntmg rudimentary 

 segments. Each of these elements into which the IBr axillary is divided is followed 

 by a syzygial pair, of which the epizygal is axillary. 



Remarks. — The true status of the assemblage of more or less widely varying 

 types here considered as the species Comanthus parvicirra is as yet undetermined. 



The line of delimitation between parvicirra and timorensis as here drawn is not 

 sharply defined, and, as may be seen from the notes on the specimens of both of 

 these, many intermediates occur. Yet the line seems really to exist, for the vast 

 majority of specimens fall readilj^ on one side or the other. 



It is worthy of special note that while the small form, parvicirra, is excessively 

 variable, the large, timorensis, is relatively stable, being only slightly variable in all 

 the several features wherein the other is remarkably inconstant. 



The form here called timorensis is confined within the area from Ceylon to 

 tropical Australia, Tonga, Fiji, and the Philippmes, while parvicirra occurs far beyond 

 these limits. Yet the specimens of parv^icirra from the limits of its range seem to be 

 identical with those taken withm the area occupied by timorensis . 



As the young of comasterids always have slender arms, no matter how stout 

 these may become in later life, and as all the species with very numerous arms some- 

 times are found with an arm number far below the average (as, for instance, Comatella 

 nigra, p. 95), it may be that some of the individuals here assigned to parvicirra are 

 in reality slender individuals of timorensis with a reduced number of arms. 



In both timorensis and parvicirra. the cirri are extremely variable, but usually few 

 and more or less undeveloped, sometimes confined to the interradial angles of the 

 centrodorsal, and occasionally wholly absent. In timorensis the number of cirrus 

 segments is somewhat greater than in panncirra, a difference which appears to be pro- 

 portionate to the difference in size of the individuals. In timorensis they are much 

 more commonly greatly reduced or absent altogether than is the case in parvicirra, 

 and also they are much more uniform in structure. 



In parvicirra individuals are sometimes found with the cirri more numerous, 

 better developed, and stouter than usual, and through such individuals there is a 

 more or less complete intergradation with sainoana. 



The eversion of the distal edges of the brachials and the relative development of 

 spines on the edges and the dorsal surface of the brachials and pinnule segments varies 



