408 BULLETIN 82, xmiTED STATES NATIONAL MUSEUM 



So he considered that it is for the present quite as good to retain the classification 

 proposed by the present author. 



He suggested as a rather good sohition the division of the subfamily Comas- 

 terinae into 2 genera, Comaster and Comanthus, the first characterized by having 

 combed pinnules occurring distally on every second or third segment, the second 

 characterized by having the comb bearing pinnules in an imbroken series. With 

 such a division the two types would be easy to keep apart if it were a question of 

 closely related forms, for such forms are to be found no matter what division is 

 made. He pointed out that in the same individual the combs seldom or nevei 

 vary in the manner of their occurrence on different arms, and so to this extent they 

 offer a better distinguishing feature than the division series. The subdivision 

 proposed is still, however, impossible to carry out because of the incompleteness of 

 the specific descriptions in regard to the appearance and the occurrence of the 

 combs. 



He said that to the first genus there might be referred Comaster, the subgenus 

 Comanthus (part), and eventually Comanthina and Comantheria (part). To Co- 

 manthus in the new sense would be referred Comantheria, Comanthina (part), the 

 subgenus Cenolia, and the subgenus Comanthus (part). 



Although it is the largest of the subfamilies of the Comasteridae and the most 

 abundantly represented in species and in individuals wherever it occurs, this sub- 

 family is the most restricted in its range both geographically and bathymetrically. 

 It is not found in the Atlantic, where species of CapUlasterinae occur on both the 

 eastern and western sides and where a species of Comactiniinae is very abundant 

 along the western shores, and it does not descend below 548 meters. Its range in 

 Polynesia, however, on the basis of the available information, would seem to be 

 much more extensive than that of the other two subfamilies. But this undoubtedly 

 is due to insufficient knowledge. 



History. — The subfamily Comasterinae was established in 1909 to include the 

 genera Comaster and Comanthus. The genera Comantheria and Comanthina were 

 subsequently established — originally as subgenera — to include groups of species 

 which previously had been referred to Comanthus. 



The Comasterinae include all the species referred by Dr. P. H. Carpenter to the 

 Typica group (all of which fall in the genus Comaster), all of those placed by him in 

 the Valida group (which are merely individual variants of Comanthus timorensis or 

 of C. parvicirra, also placed, under the names parvicirra, quadrata, and littoralis, in 

 the Parvicirra group), and all of the species referred to the Parvicirra group (which 

 are distributed among the genera Comaster, Comantheria, Comanthina, and Coman- 

 thus). 



KEY TO THE GENERA O? THE SUBFAMILY COMASTERINAE 



o'. Elements of the IBr series and first 2 ossicles following each axillary united by syzygy; IIBr 

 series 4 (1 + 2, 3 + 4), or partly 4 (1 + 2, 3 + 4) and partly 2 (1 + 2); IIIBr and succeeding 

 division series 2 (1 + 2), only very exceptionally of 4 elements; in large species the IIBr and 

 following axillaries tj'pically bear a division series on one face and an undivided arm on the 

 other, so that the IIBr (more rarely IIIBr) axillary is the base of a stout trunk composed of 

 division series giving off arms on alternate sides; this feature serves to identify the young of 

 large species in which the 2 ossicles following each axillary are stiU united by synarthry 



