62 ART. 2. — H. MATSUMOTO: 



i.e. genital burscx\ also in these genera, is a decisive proof against 

 Lyman's interpretation. 



According to Lyman, AstrogompJius vaUatiis Lyman has a very 

 shnple oral skeleton, wliich " somewhat recalls the shape in 

 OpMoscoIex, though the peristomial plates are entire and trans- 

 versely oval." Ludwig's 'first ventral arm plate' is present at the 

 upper outer corner of the oral slit. The radial shields are long, 

 bar-like, narrow, composed of several overlapping secondary plates 

 soldered together. Again, according fo Ta^ian, in Astrocnida iskUs 

 the digestive cavity has its roof firmly attached to the body wall, 

 but the floor is entirely and the sides partially free. The perihœmal 

 canal is entirely closed, and judging from Lyman's statements, the 

 genital bursa? appear to communicate witli one another so as to 

 form a single compound cavity, regarded by Lyman as the " peri- 

 toneal cavity." I do not know any other genus of the present 

 family with entirely closed periluxîmal canal. 



My own observations have been made upon Astrochele 

 li/mani and Asleroporpa Jiadracdniha, as representatives of the 

 simple-armed genera of the Gorgonocephcdince. In the first named 

 species, the radial shields are long, narrow, bar-like, almost reach- 

 ing the centre of the disk, composed of several overlapping 

 secondary plates soldered together. 'The genital plates are stout, 

 lying close to the sides of the basal vertebra^. The genital scales 

 are very small, flat," iamella-like, articulated to the genital plates 

 at a short distance from the outer ends of the latter. The basal 

 vertebra) witliin the disk are not covered over by the muscles 

 lying between them and the genital plates. The peristomial plates 

 are entire, very large, roughly pentagonal, with the longest side 

 inwards and the most obtuse angle outwards, much wider than 

 long, closely soldered witli the oral frames, wliich are ratlier 



