STAEFISHES OF THE PHILIPPINE SEAS. 455 



an angular series, the first to sixth or seventh being on the ac- 

 tinostomial margin, the seventh or eighth on the angle of the plate 

 at mouth of ambulacra! furrow, and sometimes directed across 

 mouth of latter, while the remaining 5 or 6 form a fascicle or 

 group on the surface of the plate adjacent to the very short furrow 

 margin and to the first adambulacral comb. All the spines are 

 joined by membrane, the innermost being nearly as long as median 

 suture; the others decrease in length to the spine at the angle of 

 plate, which is one-half the length of the median suture, while the 

 longest of the remaining 5 or 6 spines is slightly longer, the others 

 nearly as long. 



Madreporic body slightly convex, 6 mm. in diameter, and situated 

 slightly adcentral to the middle of r. The actinostome is wide 

 (38 mm.) and the tube feet large, with large sucking disks. Gonads 

 large, in a compact mass at base of ray. Apparently there are 2 

 ducts, one opening near ambitus and another about one-third r from 

 margin, on the interradial line. The eggs are large, 1 to 2 mm. 

 in diameter. A membranous interbrachial septum is present. It 

 extends about one-third r from margin toward center of disk, and 

 not so far as the inward extension of the gonads. 



Type.— C?it. No. 32650, U.S.N.M. 



Type-locality.— Station 5605, Gulf of Tomini, Celebes (lat. 0° 21' 

 32" N.; long. 121° 34' 10" E.), 647 fathoms, bottom not recorded. 



Remarks. — In the original description, I suggested that this species 

 might really belong in Myxaster, if the type of that genus should 

 be found to possess the same peculiarity in respect to the disposi- 

 tion of the adambulacral spines. On account of the structure of the 

 adambulacral plates, I placed medusa in Asthenactis, in which each 

 adambulacral fan is divided between 2 plates, the outer spines of 

 each series standing on an oblique aboral extension or apophysis of 

 the next adorally situated plate. No mention of this highly char- 

 acteristic arrangement of plates and spines was made by Perrier in 

 his description of Myxaster sol. In June, 1914, I examined the 

 type of this species at the Museum d'Historie Naturelle of Paris. 

 In Myxaster sol the adambulacral plates are constructed essentially 

 as in M. medusa., 3 spines, as a rule standing on the apophysis, while 

 5 or 6 continue the series on the body of the next aboral plate. 



It becomes evident, therefore, that the structure of the adambu- 

 lacral plates and the peculiar arrangement of spines are the same in 

 both Asthenactis and Myxaster, and that the presence of an actino- 

 lateral membrane is the one trenchant character which separates 

 Asthenactis from Myxaster. 



Myxaster medusa differs from M. sol in having a much thicker 

 skin, which completely hides the underlying plates; the mouth plates 

 are broader, with a broader actinostomial margin and a decided angle 



