212 BULLETIN 100, UNITED STATES NATIONAL MUSEUM. 



The arm-disk is about 100 mm. wide, the perradial columns being 

 each 36 mm. wide. The subgenital ostia are much larger than in 

 C. tuberculata, but they were so mutilated that one can not state their 

 exact size, which appears, however, to be nearly that of the perradial 

 columns themselves. The specimen having been cut into pieces, we 

 can make no statement concerning the condition of the subgenital 

 porticus. 



The eight mouth-arms resemble those of C. tuberculata, but there 

 are windowlike openings in the lateral membranes, as in Lobonema 

 smithii. The total length of each arm is 81 mm., the upper arm being 

 one-fourth as long as the lower arm. At their widest part the arms 

 are about three-fourths as wide as they are long. The center of the 

 arm disk is thickly covered with slender filamentous appendages 

 which terminate in nematocyst-bearing, swollen, knoblike ends, as do 

 the appendages of the mouth-arms of C. tuberculata. In this Philip- 

 pine Island medusa the appendages of the outer parts of the mouth- 

 arms much less numerous and smaller than in C. tuberculata, but are 

 similar in general form to those of the Mediterranean medusa. The 

 longest are about 15 to 20 mm. long. 



The cruciform central stomach gives rise to about 140 radial-canals, 

 the eight rhopalar canals being about twice as wide as the others, in- 

 stead of being of the same caliber, as in C. tuberculata. All these 

 canals anastomose in a network under the zone of the circular muscles. 

 There is no distinct ring-canal. In formalin the specimen is dull 

 uniform yellowish brown. 



It differs from Cotylorhiza tuberculata in having no radial-muscles, 

 and in the circular muscles being interrupted in the eight principal 

 radii. The subgenital ostia and arm-disk are larger and the ap- 

 pendages of the mouth-arms smaller and fewer than in G. tuberculata. 

 Moreover, the peculiar perforations in its mouth-arm membranes at 

 once distinguish this species. 



These distinctions are indeed of such nature that if one felt so in- 

 clined a new genus could be established to receive this medusa. I 

 believe, however, that its relationships will be more clearly indicated 

 by placing it in the genus Cotylorhiza, within which it forms a well- 

 marked species. 



Genus CATOSTYLUS L. Agassiz, 1862. 



Catostylus (part) Agassiz, L., 1862, Contr. Nat. Hist. U. S., vol. 4, pp. 152, 

 153. 



Generic Characters. — Rhizostomata triptera, in which the mouth- 

 arms bear neither clubs, filaments, nor other appendages. Sixteen 

 radial-canals, 8 rhopalar, and 8 adradial. The rhopalar-canals ex- 

 tend to the bell-margin, but the adradial-canals end in the ring-canal. 

 On both its inner and outer sides the ring-canal gives off anas- 



