STARFISHES OF THE PHILIPPINE SEAS. 481 



close resemblance to MaTumaster Perrier, the only differential charac- 

 ters that I can find being stated in the key to genera. It may be 

 necessary eventually to merge the 2 genera. Perrier (1894, p. 128) 

 states that there are forcipiform (croise) pedicellariae on the disk; 

 I have found numerous forficiform pedicellariae. The former are 

 not supposed to occur in this family. 



In both Gneinidaster and Maminaster superambulacral plates are 

 present and the first is modified into a conspicuous buttress or stay 

 extending from the upper end of the combined first 2 ambulacral 

 platas to the body wall at the interradius, as in Myxod&nna and 

 Bythiolophus. 



There are in the collection no specimens strictly comparable with 

 the very small type of wyvillii, although 1 example, as already 

 stated, is only slightly larger. This specimen seems to be referable 

 to Sladen's species, although it must be admitted that the specific 

 determination of very young specimens is difficult, owing to the 

 fact that important characters, such as granules, spines, and pedicel- 

 lariae may not appear until comparatively late. The young of dif- 

 ferent species may, therefore, be very closely similar. Owing, how- 

 ever, to the similarity between the 2 forms, and the geographical 

 proximity of the Challenger station, I have thought it better to list 

 the Albatross specimens as wyvilUi. They are closely related to both 

 Cnemidaster zea, and G. squameus. For instance, there is consid- 

 erable variability, one specimen, otherwise perhaps referable to G. zea 

 has central spines on the distal carinal and marginal plates, and 

 spaced granules on the disk plates, while another from the same 

 station (5602) has no granules on the swollen disk plates. The fol- 

 lowing description will aid in determining specimens : 



Description. — Taken from specimens from Gulf of Tomini, Celebes. 

 R=109 mm., r=12 mm., 11=9 r (station 5607). Disk flat-topped and 

 elevated above base of rays, the plates essentially as in the young, but 

 definitely scalloped or indented for papulae. In the grooves are a 

 few fair-sized pedicellariae (1.5 mm. long) about 3 times as long as 

 broad at base, and a few scattered immersed granuliform spines. One 

 specimen from station 5602 has scattered small granules under the 

 skin of the disk and proximal abactinal plates of the ray, while an- 

 other from station 5606 has fewer granules on the disk only. Four 

 other adult specimens, one from 5602, lack the granules. Eays evenly 

 arched, without a marked carinal ridge. There is a carinal series of 

 slightly convex plates proximally wider than long, distally longer 

 than wide. On either side of this is a close double row of papular 

 pores emerging on either side of a nearly hidden series of small ad- 

 radial plates. Then follows 3 (or 4) series of overlapping, smooth, 

 skin-covered plates, followed by 3 series covered with spaced, lan- 

 ceolate, flat, squamiform spinelets, which in the interbrachial areas 



