STARFISHES OF THE PHILIPPINE SEAS. 253 



intestinal coeciun resembles that of R. symholicus^ but each interra- 

 dial division is entire and not divided into 2 distal lobes. There are 

 no rudimentary superambulacral ossicles. 



Type.— C^t. No. 30540, U.S.N.M. 



'Type-locality. — Station 5281. between Lubang and Luzon (lat. 13° 

 52' 45" N.; long. 120° 25' E.), 201 fathoms, dark gray sand. 



Distribution. — Known only from the vicinity of the type-locality. 



Specimens e.i'amined. — Four, the type, and 3 from [•station 5282, 

 type-locality, 248 fathoms, dark gray sand. 



Re7)i-a7''ks.—l do not think this species is so closely related to 

 R. hipunctuR as is R. nannxis. A small example of R. mimicus with 

 11^32 mm., and therefore comparable to the tj^pe of bipunctus is 

 just as different as the larger type, except that the large radial 

 plates have upward of 12 central granules and the adambulacral 

 plates have 12 compressed, truncate furrow spines instead of the 

 maximum of 14 or 15 of the type. 



The general appearance of R. mimicus is much like a Ceramaster 

 or a Mediaster., with long rays and numerous superomarginals in 

 contact. At the same time it greatly resembles a true NyTnphaster, 

 from which it is separated by the anatomical features characterizing 

 Rosaster. 



ROSASTER SYMBOLICUS (Sladen). 



Plate 91, fig. 6. 



yyinpliasivr ■\iJii:bolicii.s Slaokx, lysi!, p. i::iT, pi. ."0. figs. 1 mv.\ '2; pi. 53, 



figs. 7 and 8. 

 Xercidaster symbolicus Vekrill, 1899, p. 187. 

 Rosaster symbolicus Fisher, 1913a, p. 630. 



Notes on Philippine specimens. — The 2 specimens differ somewhat 

 from Sladen's figures, but agree more closely with his description, 

 the latter not being in strict accord with the plates. 



The specimen from station 5520 has 9 or more superomarginals 

 in contact at the end of the ray. Sladen's figure shows them sepa- 

 rated throughout the ray. But he says, on page 299 : " Toward the 

 end of the ray the paxillar plates lose their hexagonal outline and 

 become square, and finally become so small that their serial succes- 

 sion is interrupted by the contact of the 2 corresponding supero- 

 marginal plates from each side of the ray. The radial series then 

 rapidly aborts and the extremity of the ray is occupied entirely by 

 the superomarginal plates." On the following page he says: 

 " Nymphaster symbolicus is readily distinguished from Nymphaster 

 protentus and its allies * * * by the presence of a medioradial 

 series of plates which separate the two opposite series of supero- 

 marginal plates throughout the ray." He contradicts himself in the 

 same way in his key, his statement being obviously erroneous there, 

 because in N. bipunctus the distal superomarginals are regularly in 



