78 BULLETIN 7G. UNITED STATES NATIONAL MUSEUM. 



all ulon-,' the upper edge of the ray, and by the fact that tlie bristling inferomarginal 

 spines form about two longitudinal rows along at least two-thirds length of ray, 

 and one row along the outer third. 



Ku'liler " has described the genus Phidiaster (Nicobar Islands, 888-930 fath- 

 oms), wliich is distinguished from PsUusier by the great development of the supero- 

 margimds. These plates, as well as the inferomarginals, carry each two large robust 

 spines, forming thus four longitudinal series along the ray, one dorsal, one ventral, 

 and two lateral. .Vnother character is the presence on the inner angle of each 

 mouth plate of two enlarged teeth. 



In some resi)Octs Psilaster pedinatus is intermediate between Psilasfer and 

 Phidiaster. It has the superomarginal plates suliiciently large to satisfy the quali- 

 fications of Phidiaster, but has only one superomarginal series of spines, and the 

 lower of the two infcromarginal series extends only part way along the ray. The 

 teeth are not conspicuously enlarged. The characters of Phidiaster appear to be 

 somewhat trivial for generic value. The superomarginal spines vary in the genus 

 Perse phonaster, being absent in the species formerly ranged under Psilastcropsis. 

 Koehler has, himself, called attention to the fact that the presence or absence of 

 superomarginal spines can not be regarded as of generic importance in Persephone 

 aster. Certainly the same remark ap[)lies ecjually well to the nearly related Psilaster 

 and Astropedcn. The superomarginals of Phidiaster, judging by plate 9, fig. 3 

 (Ka'lJer, 1909), are not larger than in Psilaster pectinxitus, and some other species, 

 as, for example, Psilaster florae (Verrill). The superomarginals of Astropecten vary 

 greatly in relative size, and I have seen enough species of Persephonaster to know 

 that they do also in that genus. The remaining character, that of the teeth, is not 

 shared by typical Psilaster, although there are two (instead of four) more or less 

 enlarged median teeth at each mouth angle. These, in Psilaster, are usually bent 

 up and not directed across the actinostome as in Phidiaster. From an analysis 

 of these characters and owing to the intermediate position of Psilaster pcctinatus 

 I am inclined to unite Phidiaster with Psilaster. 



Genus THRISSACANTHIAS Fisher. 



Thrissacanthias Fisher, Ann. and Mag. Nat. Biet., eer. 8, vol. 5, Feb., 1910, p. 171. Type, Perse- 

 phonaster ipenicillatus Fisher. 



Diagnosis. — Near Persephonaster Alcock, but differing in the arrangement of 

 gonads which are not confined to the interradii but extend for a short distance 

 along the ray as a number of separate tufts depending from the genital rachis; 

 adambulacial plates with one or two enlarged actinal spines; disk of medium size, 

 depressed, rays long; marginals opposite, with conspicuous spines; abactinal sur- 

 face %vith true paxilla; arranged in regular oblique transverse rows; no enlarged 

 radial series; papula- all over abactinal surface except at tip of ray; actinal inter- 

 mediate plates extending far along ray; interradial areas rather small; spiniform 

 fascicular pediccUariiv on marginals; anus present; madreporic body large, not 

 hidden. 



lieep fSea Aslcroidea of the Investigator, 1909, p. 28, pi. 9, figs. 1, 2, 3. 



