JAPANESE ASTEROIDEA. 221 



Superomarginal spines three in P. triacanthus, and two in P. 

 hrevisjmius. Only in the interbrachial arcs and the terminal por- 

 tions of the arms are these spines less in P. triacanthus. 



The fnrrow series of the adambnlacral spines are four or five 

 for P. hrevispinus, and 5-7 for P. triacanthus. They are long and 

 slender in the latter species but short, robust, and subpapilliform 

 in the former. The spines on the actinal plates of P. iriacaiithus 

 are also tolerably long and slender but small and stumpy in P. 

 hrevispinus. 



The armature of the mouth-plates are similar in both ; but 

 the spines appears to be somewhat less numerous in P. hrevispinus, 

 there being about 16 (six pairs + 4) or 18 on each plate as against 

 18-23 for P. triacanthus. 



The terminal plate bears three spines in P. triacanthus, but 

 apparently none in P. hrevispinus. This may however be due to 

 abrasion. 



Perseplionaster hrevispinus (Sladen). 



This is the xiäropeden hrevispinus of Sl.\den. It is not represented 

 in any of the collections studied by me, but there is no doubt in my mind 

 that it must be referred to PersepJionaster, because the character of its 

 adamhulacral armatm^e shows it to he closely related to triacanthus. The 

 only single character of decisive value that distinguishes this genus from 

 Asfropeden is the presence of the anus. Sladen was not able to make out the 

 anus, but this is very small in some s^Decies and is very apt to be over- 

 looked, unless special pains be taken to prove its presence ; and in dealing 

 with a large mass of material such as Sladen had before him one is but 

 too hable to rely more on external characters that are comparatively easy 

 of obsei-vation. It did not however escape the experienced eyes of Sladen 

 that the present species presents some characters that are not usually foimd 



