56 T. (iilSLEN, 



fication. In the Comasterids they are lai-ge and well developed, they 

 appear smallest in species with powerful cirri as in Cmnanlhns pingim. 

 Everyone who has worked somewhat with species of this genera, 

 knows that for instance a Comaster or Variia j>arvicirra has a strong 

 power of attaching itself by means of the dorsal hooks of the distal 

 pinnules. Consequently the cirri will become superfluous or perhaps 

 even be a hindrance if the animal wishes to creep about, which by the 

 observations of H. L. Clark is proved to be the manner of locomotion 

 of some Comasterids. Because of this they become reduced and in 

 many genera they have totally obliterated (as in species of CapiUaster, 

 Comahda, Gomaster, Co'inantheria, Comauthiwt and Yanin). 



As appeal's from the above investigation the subgenus Vania 

 is a very critical genus, forming to a certain extent a middle link be- 

 tween Comaniheria-Gomanihus and Gomaster. The reason why I have 

 not united all these genera is that I perceive the possibility of Vania 

 representing a rallying group of hybrids and hybrid-splittings with com- 

 binations of characteristics from different foi'm-eircles. I should have 

 liked to put Vania as a separate genus between Gomaster and Goman- 

 theria-Gomantlms. In this case I would have distinguished it fiom the 

 the first-mentioned genus by the III Br — s chiefly consisting of -I com- 

 ponents, from the two latter ones by comb-provided piinudes often 

 occurring far out on the arms. Nevertheless the characterizing in 

 such a case is not satisfactory, if one does not transfer all the Vanias 

 of comanfhipinna-iyi^e to the genus Gomanthus (Genolia). Perhaps there- 

 fore, it is for the present quite as good to keep the classification 

 made by A. H. Clark and put Vania as a sub-genus under Gomanthus. 

 Still it is then to be observed that, according to his scheme of exa- 

 mination, V. parvicirra a comasteripinna ought to be referred to Go- 

 mantlnis if one pays attention to the division-series, to (^'omaster if one 

 takes into consideration the occurrence of the combs. Compare also 

 in this connection the above quoted statement of A. H. Clark. Ano- 

 ther solution, and this a rather good one, might be possible. This 

 x^ould 1)0 to subdivide Gomasterinre into 2 genera: Gomaster and Goman- 

 thus., the fii-st one characterized by comb-provided pinnules occuring 

 distally on every second or third pinnule, the latter one with comb- 

 jirovidod |)innnlos in an unbi-oken series, ^^'ith such a di\ision the 



