OPHIURANS OF THE PHILIPPINE AND ADJACENT WATEPtS. 97 



mal angle bounded by two straight sides which appear only toward 

 the third or fourth plate; the two sides are divergent and very 

 strongly excavated by the tentacle pores; the distal border is very 

 broad and convex. These plates are at first in contact, later becom- 

 ing separated by a narrow space. 



The side arm plates, which project only slightly, bear four spines, 

 which are subequal and a little longer than the segment ; their length 

 increases somewhat from the first ventral to the last dorsal (figs. 1 

 and 4). These spines are flattened, pointed, and appressed against 

 the lateral surfaces of the arms; their surface is slightly rugose. 

 They are hollow, and their wall is very thin. 



The first arm segments bear three tentacle scales, two very large, 

 oval and lanceolate, inserted on the side arm plate, and a third, 

 smaller, which arises from the side arm plate and extends perpen- 

 dicularly to the axis of the arm. This third scale disappears beyond 

 the disk, and the two large scales alone persist. 



The color of the specimen in alcohol is white. 



Affinities and distinctive features. I established the genus Opliio- 

 medea for an ophiuran discovered by the expeditions of the Travail- 

 leur and the Talisman in the North Atlantic at a depth of 2,000 

 meters (1,100 fathoms), in which the outer mouth papilla, borne on 

 the adoral plate, is considerably elongated and spinif orm ; the tenta- 

 cle pores show, in addition to the principal scale inserted on the side 

 arm plate, a smaller scale inserted on the under arm plate, and the 

 disk is covered with imbricated scales, each armed with a spine. The 

 type of the genus Ophiomedea is 0. duplicata, which is only repre- 

 sented by a unique specimen in which the diameter of the disk does 

 not exceed 4 millimeters. 



Although the species which I have just described differ in some 

 important features from 0. duplicata, I believe that nevertheless it 

 may be assigned to the same genus. The differences are, the presence 

 of two spiniform mouth papillae inserted on the adoral plate, instead 

 of only one, the existence of three tentacle scales instead of two, and 

 the presence of granules instead of spines covering the dorsal surface 

 of the disk. But none of these characters is sufficient for generic 

 differentiation ; the presence of a tentacle scale inserted on the under 

 arm plate and the occurrence of one or of two spiniform external 

 papillae are evidently very characteristic features. It is certainly 

 regrettable that the type species of the genus Ophiomedea should be 

 represented by a single specimen of very small size, rendering com- 

 parisons very difficult, but I do not believe I am in error in assign- 

 ing our species to this genus. 



Furthermore, I am in doubt whether the genus Ophiophrura, estab- 

 lished by H. L. Clark in 1911 for a Japanese species, also represented 

 55269 22 Bui. 100 7 



