406 BULLETIN 82, UNITED STATES NATIONAL MUSEUM 



arm. The second, the Comanthus type, has a greater number of teeth, which are 

 small and low and therefore well separated. They become smaller proximally, and 

 become obsolete toward the tip of the pinnule, leaving the latter smooth. The 

 combs occur on a very few proximal pinnules — never further out on the arm than 

 Pe — and are in a continuous series, never on every second or third pinnule as in the 

 Comaster type. The third type, which to a certain extent is an intermediate one, is 

 the subgenus Comanthus type, which has rather high and large teeth right out to the 

 tip of the pinnule. As in the Comanthus type, the combs occur in an unbroken series. 



Both the first 2 types are usually very definitely associated with certain genera 

 within the subfamily Comasterinae. The Comaster type is found in the genus 

 Comaster. The Comanthus type is, according to Gisl^n, characteristic of the genera 

 Comantheria, Comanthina{f) , and the subgenus Cenolia in Comanthus. 



But Gisl^n pointed out that in the subgenus Comanthus the use of this character, 

 like so many others, becomes impossible. 



To the subgenus Comanthus are assigned 2 forms, distinguished from each other 

 by such an artificial character, in Gisl^n's opinion, as the number of arms. He 

 believed the facts to be that Comanthus timorensis represents one group of forms with 

 a generally greater number of longer and stouter arms, and C. parvicirra another group 

 with a smaller number of more slender arms. But he said that both the so-called 

 species vary so considerably that they intrude upon each other's spheres. 



The specimens obtained by Bock show that within the subgenus Comanthus 

 there occur combs of both the Comaster and the Comanthus types, and Mortensen's 

 Japanese specimens show that combs of the subgenus Comanthus type occur also. 



In reporting upon the Bock collection Gisl4n divided Comanthus {Comanthus) 

 parvicirra into 2 subspecies, comasteripinna and comanthipinna, based upon the 

 occurrence of the Comaster or the Comanthus type of comb. He said that he was 

 neither able nor willing to discuss the innumerable synonyms of this species, and 

 remarked that it is not possible to distribute the synonyms among the two new 

 subspecies because most of the authors have given no information about the occur- 

 rence and appearance of the combs. 



In his report upon the crinoids of Mjoberg's expedition he mentioned Carpenter's 

 statement regarding the occurrence of the Comaster type of comb in 3 of the forms 

 referred by the present author to the synonymy of parvicirra, namely, Actinometra 

 elongata, Act. simplex, and Act. quadrata. All of Mjoberg's specimens, however, had 

 4 elements in the IIBr series and combs extending far out on the arms as in Comaster; 

 so, trusting to the generic diagnoses of the present author, he referred the specimens, 

 properly referable to Comanthus (Comanthus) parvicirra, to the genus Comaster, 

 assigning the examples with stouter cirri to Comaster multijida and those with rudi- 

 mentary cirri to C. typica. Since then he has been able to show that the occurrence 

 of combs far out on the arms is not pecuhar to the genus Comaster, but is characteristic 

 of certain forms in the subgenus Comanthus in the genus Comanthus. 



Gisl^n remarked that it is not strange that both of the species in the subgenus 

 Comanthus have caused great trouble to all authors by their extreme variability in 

 practically all diagnostic characters elsewhere successfully employed. Not only are 

 the cirri and cirrus segments, division series and brachials, very variable in appearance 



