ASTEROIDEA OF NORTH PACIFIC AND ADJACENT WATERS FISHER. 143 



marginal usually with a very proiuinent spine. Actinul iuterriuiial ureas with very 

 few intermediate (ventral) plates. Armature of adambulaeral plates consisting of a 

 semicircular furrow series of small uniform spines, radiating fan-like, and one or 

 more larger actinal spines. The furrow series are typically separateil from each 

 other by a considerable interval. Pectinate pedicellaritr may be present on any or all 

 of the following regions: Abactinal plates, inferomarginal plates, actinal inter- 

 radial areas. Superambulacral plates alwaj-s absent; tube feet with sucking disks. 

 Intestinal ccEcum saccular, four-parted. 



Remarks. — Ludwig m his recent paper on the Xotomyota recognizes Parar- 

 chaster (type, P. pedidfer) for those species which have several spinelets on each 

 abactinal plate, reserving for Benthopecten the forms which have one spine or spinelet 

 to a plate. His genus Pararchastcr corresponds to the second section [B] of Sladen's 

 key (OhaUenger Astewidea., p. 5), less arrnatus. 



Such a division of Benthopecten mto two genera, or even subgenera, would bo 

 useful, but in practice is difhcult, in the absence of other correlative characters. At 

 the start we find that Benthopecten in the restricted sense lias three species, incertus, 

 cognatus, and simplex with certain plates bearmg two or three spinelets. 



Among the following species claviger would undoubtedly fall in Parar chaster. 

 Mutahilis has generally single spmelets on the rays, but the large sjiines of the disk 

 are freriuently surroimded by a circle of upward to ten or eleven short spinelets, 

 while other plates have one to three or four short stubby spinelets, in no wise differ- 

 ent from those present on tj'pical " Parar chaster." But mutahilis, while presenting 

 several points of diiference, greatly resembles Benthopecten spinosus, the type of the 

 genus. One would not consider placing the two species in separate genera. Ben- 

 thopecten acanthonotus is difficult to classify, if two groups are recognized. It would 

 seem to belong with Benthopecten restricted, but in the type-specunen I find 

 numerous plates with two and three spinelets. Another specimen, doubtfully 

 referred to this species, and which, if not acanthonotus, is very closely relatetl, has 

 five or six spinelets surrounding the enlarged spines of the disk. The type also 

 has two or three small sj)inelets next to some of the enlarged spmes. The type of 

 Pararchaster is not so well differentiated from Benthopecten as some other species 

 included in the genus by Lud\vig. 



I have maintained Benthopecten in the extended sense in wliich Verrill and I have 

 already used it, and coextensive with Pararchaster Sladen, Perrier, Ka^hler, and Lud- 

 wig previous to 1910. I do this because I find great practical difficulties in using the 

 two generic names. I believe that it will be |)ossible to spht up Benthopecten, but I 

 doubt if Pararchaster can be used as a name, because its type is too closely allied to 

 B. spinosus. 



Some hint of a possible division was afforded by the distribution of papulfe in 

 B. acanthonotus. Here the papuhe extend far along the ray or are confined to the 

 basal portion, det)ending apparently upon the age of the individual. The character 

 is, however, too variable to bo of use as a generic feature in this group, and I think 

 the same is true of the number of spines on the abactinal plates, unless accompanied 

 by some more conservative structural feature as j'et undetermined. 



