ASTEROIDEA OF NORTH PACIFIC AND ADJACENT WATERS FISHER. 205 



intermediate granules; the first actinal a(laml)ulacral granules or spinelets some- 

 tiinea slightly spaced from furrow series as in Mediaster. Pediceiiarifc bivalved 

 or with two or three spatulate jaws; absent in type. No trace of rudimentary 

 superambulacral plates, such as occur in Mediaster; no internal abactinal radiating 

 ossicles connecting the plates. 



Remarks. — Tliis genus is better characterized than any other nearly related 

 group, and while the species included in it arc quite diverse in some particulars, 

 they hang together fairly well in respect to the structure of the abactinal, mar- 

 ginal, and actinal intermediate i)lates. The genus Tosia, as limited by Verrill in 

 1899 and adopted by me in 1906, is too comprehensive. Plinthaster, a section 

 of Tosia under Vcrrill's classification, is very different from Ceramaster, as I found 

 after having examined Plinthaster pcrrieri, P. cnmptus, P. niiidus, and P. dentatiis. 

 As explained above (p. 105), Pyrenasier does not seem to mc to be separable from 

 PlintTiaster. Plinthaster has flat, naked abactinal plates, while in Ceramaster the 

 plates arc granulated and there is a distinct tabulate stnicture. The Toina {Cera- 

 master) micropelta described by mo in 1906 is not a Ceramaster, as it has small, round, 

 flat, granulated plates, with intermediate smaller plates on the central part of the 

 disk and proximal radial regions. It belongs to Peltaster or a closely related genus. 



Of the list of species given by Verrill (1S99) as belonging to Ceramaster I would 

 not include, without personal examination, mammiHatus, hxsitans, and pulvinus; 

 placenta is in many respects an aberrant member of this group, though I have not 

 seen specimens. The tabulate form of the plates of the papular areas seems to 

 be minimized. In formulating our conception of a genus we hark back to the 

 type, in this case granuhris. Typical species, as granulans and patagonicus, stand 

 about midway between such extremes as japonicus and placenta. C. leptoceramus 

 with its secondary abactinal plates m adults is decidedly aberrant, though appar- 

 ently near to japonicus. It might be made the type of a new subgenus. C. arcticus, 

 with the excei)tionally extensive papular areas, short adambulacral plates, and 

 few furrow spinelets, is still another variant from the typical form of stnicture. 



Perhaps the nearly related genera wliich can be grouped about Ceramaster 

 are Nereidaster, Mediaster, Tosia (as here used, a very restricted austral genus), 

 and Peliaster, with Plinthaster, Eugoniaster, anil Pentagonaster in an outer circle. 

 Mediaster arcuatus (Sladen), which has usually been classed as Pentagonaster, but 

 which posses.ses the internal independent radiating ossicles of Mediaster sequalis 

 shows how near the two groups approach in external appearance. From the 

 original figures and description Professor Verrill was led to place C. japonicus and 

 patago7iicus in Mediaster, but they are more nearly allied to C. granulans and 

 must therefore be classified in Ceramaster. C. placenta .seems to tend toward Tosia. 

 Tliis small grouj), as alread}' explained, is homogeneous enough to form an independ- 

 ent genus. Plinthaster and Eugoniaster are its nearest relatives, next to Penta- 

 gonaster (restricted). Peltaster has small flat (not tabulate) granular abactinal 

 plates, with few to many secondarv ones among them. If the aberrant P. micro- 

 pelta (Fisher) is admitted to the genus the form of the pedicellaria^ an<l the ailam- 

 bulacral armature will have sis great a range as in the most diverse species here 

 classified under Ceramaster. 



