24 BULLETIN TG. UNITED STATES NATIONAL MUSEUM. 



Diagnosis.— hike Porcellanaskr Wyville Thomson, but witli three (instead 

 of one) cribriform organs in each interradius. 



ilnrfriiiul plates thin; superomarginals not united along the median hne of 

 ray ami usually bearing a spme, these forming a series on each side of ray. Abac- 

 tiiial area covered with membrane containing very small, simple, perforated plates; 

 those on disk bearing usually a single spinelet each; in center of disk a sort of pro- 

 longation of the integument forms a peduncle, containing an extension of the body 

 cavity, but no intestine. Actinal interradial areas paved with thin plates, some- 

 times armed with spinelets. Adambulacral armature, one to three sharp spinelets. 

 Usually "segmental i)its and papillse" are present. No superambulacral ossicles; 

 tube feet conical, pointed. Ampulla} single. 



Remarks. — When this group was described as a subgenus in 1905, I gave three 

 characters, in order of importance, by which it could be distinguished from Por- 

 cellanaskr: namely, three cribriform organs, ])resence of segmental pits and papillae, 

 and one or two adambulacral spinelets. Since then I have examined much more 

 material, and have found that the last character is of no importance generically, 

 while the segmental pits and papilhe are numerically variable. The absolutely 

 constant character at all ages is the three cribriform organs. The segmental pits 

 and pajiilliB are valuable when taken in connection with the three cribriform organs, 

 as they show relationship. The second (aboral) spinelet of the adambulacral 

 plates of Porcellanaster is homologous with the segmental papilla of Eremicaster; 

 but the outer pait of the ray in Eremicaster lacks the segmental papilla, and is 

 therefore similar to Porcellanaster, except in those specimens in which three spmelets 

 are present in addition to the segmental ])apilla or its homologue. 



Ludwig has considered the segmental papilla^ and pits as of primary importance." 

 He placed his P. pacijicus which has three cribriform organs and no pits or papillae 

 in the restricted genus Porcellanaster. I have studied pacijicus and consider it 

 much more nearly related to ienehrarius than to P. cse.mleus. Owing to the slight 

 instability of the pits and jiapilla? in tenchrarius (includes waltharii Ludwig) it 

 seems better to rely upon the cribriform organs as mentioned above. Therefore 

 I have placed pacijicus in Eremicaster, although Ludwig relegates it to Porcellanaster. 



In raismg my subgenus Eremicaster to a genus Ludwig changed the published 

 type tenchrarius to crassus Sladen. This is contrary to a very elementary rule 

 in nomenclature. Eremicaster will stand oi- fall with ienehrarius as type. As a 

 matter of fact the two species are so close that it makes no material difference 

 which is type, but the habit of ignoring well-established rules of nomenclature 

 is a bad one. 



EREMICASTER TENEBRARIUS (Fisher). 

 PI. 1, figs. 1-4; pi. 2, fig. 4; pi. 53, figs. 4, 4a. 



Porcellanaster (Ercmkastcr) tmehrarius FisnEn, Bull. Rur. Fisheries for 1904, vol. 24, June 10, 



1905, p. 293.— LcDWUi, Zool. An?,., vol. 31, 1907, p. 317. 

 Porullanastcr uahharii Ludwig, Mem. Mus. Comp. Zool., vol. 32, July 17, 1905, p. 92, pi. 5, 



<• Zool. Anz., vol. 31, 1907, p. 317. 



