46 VERRILL 



be restricted to this type as a subgenus, if retained at all. The 

 group including glacialis had also been previously named Marth~ 

 asterias by Jullien (Bull. Zool. Soc. France). 



No species of typical Coscinasterias is known to me from the 

 northwestern coast, but a species (C. acutispina Stimp., sp.), similar 

 to C. tenuispina, and often having some diplacanthid adambulacral 

 plates, occurs at Ousima Island, Japan. It is autotomous. 



C. TABLE OF EXTRALIMITAL SPECIES OF COSCINASTERIAS 

 AND CLOSELY ALLIED FORMS. 



A. More than one distinct row of dorsal spines (usually three rows). 



B. One row of peractinal spines and plates well developed. Rays variable, 



six to twelve. Mostly fissiparous. 



C. Rays mostly seven to twelve. Adambulacral spines strictly uniserial 



(monacanthid). Lower marginals bear two spines each. Usually 

 two or more madreporic plates are present. 

 C. muricata (VER.) 1867. (J = C. calamaria.) Tt*^ 

 C. calamaria (LAM.). Australia; Indian Ocean. 

 C. jehennesii (PER.). Madagascar. (? = calamaria, teste Sladen.) 

 C. echinata (GRAY). Valparaiso. Rays eight 

 C. gemmifera (PER.). Chile;' Fiji Is. (? = echinata.) 

 C. (f)aster (GRAY). Rays twelve to thirteen. Five rows of dorsal 



spines (Gray). 



CC. Some of the adambulacral plates diplacanthid. Rays variable. 

 C. acutispina ( STIMP.). Ousima Island. 

 C. tenuispina LAM. (Large specimens.) 



BB. No interactinal spines ; peractinal plates rudimentary or lacking. 

 AA. Only one row of large dorsal plates and spines. Adambulacral plates 



monacanthid. 



b. Rays mostly five to nine. Fissiparous. Subgenus Stolasterias (Sla. restr.). 

 C. (S.) tenuispina (LAM.) when young. Eastern Atlantic; Brazil; 



West Indies. 



Variety atlantica (VER.). Bermuda; Cuba, 

 bb. Rays constantly five or six. Type not fissiparous. 



Marthasterias (JuL.) when young. Type, M. glacialis (MULLER). 



D. OBSERVATIONS ON VARIOUS GENERA. 



Genus Heterasterias Verrill, nov. 

 Type, Stolasterias volsellata Sla., Philippines. 



The rudimentary state of the dorsal skeleton and the lack of 

 peractinal plates in the named species may entitle it to be con- 

 sidered the type of a new generic division, as suggested by Sladen. 

 This view is strengthened by reason of the presence of large ungui- 

 culate dermal pedicellariae. In this character and in general appear- 



