98 DISCOVERY REPORTS 



c^. Abactinal plates with a distinct tabulum crowned with short to fairly long spinelets; 

 marginal plates small, to well-developed, more or less tabulate, spinulose; actinal 

 area densely spinulose. Odont aster Verrill. 



c^. Abactinal plates without a distinct tabulum, but slightly convex. They are crowned 

 with low granules, sometimes involved in a thin integument, which, however, in 

 no way hides them. Marginal plates prominent to small, granulose, not tabulate; 

 actinal areas covered with granules which, in some species, are gradually transformed 

 to spinelets as oral angle is approached. Acodontaster Verrill. 



From the above synopsis it will be seen that five genera are recognized, whereas 

 Koehler (1920, p. 190) admitted eleven. Heuresaster Bell was overlooked, but it is 

 obviously the same as Koehler 's Tridontaster. 



I regret the necessity for departing so widely from the results obtained by Koehler ; 

 but it is inevitable since in certain cases he adduced to distinguish his genera characters 

 of only specific, or even less than specific, significance. 



It would be useful to separate Odotitaster from Gnathaster, but it cannot be done. For 

 one thing the marginal plates of Gnathaster vary widely, if we include such diverse 

 species as Odontaster validiis, where they are paxilliform, and O. penicillatus forma 

 grayi, where they are block-like and covered by elongate granules or short spinelets. 

 But Gnathaster must include the widely variable penicillatus, which bridges the gap to a 

 restricted Odontaster. 



Unfortunately Koehler made O. penicillatus iorraa grayi (i.e. O. grayi Bell) the type of 

 Peridontaster. Grayi intergrades perfectly with Odontaster penicillatus (O. pilulatus 

 Sladen) which in turn is closely related to O. meridionalis (Smith), the type of Gnathaster. 

 Such differences as Odontaster grayi shows to O. meridionalis are therefore not generic. 

 In another direction, O. grayi (apart from its pentagonal shape) obviously resembles 

 O. hispidus, the type of Odontaster \ Odofiiaster, Gnathaster and Peridontaster cannot 

 be separated on the basis of characters now known. 



I believe that the small type of Epidontaster is the young of Odontaster meridionalis 

 having accessory teeth — not an uncommon variation in that species. Only one specimen 

 is cited. 



The character of the plates and their armature is the same in Acodontaster, Pseudont- 

 aster, Metadontaster, and Tridonaster [Heuresaster Bell). Pseudo?itaster simply has 

 somewhat smaller marginal plates than has Acodontaster. Pseudontaster moderatus 

 Koehler and Ps. stellatus Koehler, in my opinion, are variants of the circumpolar 

 Acodontaster elongatus (Sladen), type of Acodontaster. For some curious reason 

 Koehler never recognized elongatus (see Odontaster cr emeus Ludwig, 1903 ; Koehler, 

 1912; Acodontaster cremeus Koehler, 1920). 



In view of the obviously close relationship of Tridontaster laseroni to Acodontaster 

 conspicuus and the occasional presence of accessory teeth in the latter, I cannot take 

 seriously the use of accessory teeth alone as the criterion of a new genus. 



Metadontaster contains a single species M. zvaitei, which is so close to Acodontaster 

 conspicuus that I am uncertain whether to call it a species or a form of conspicuus. 



There follows a list of the Odontasteridae of the southern hemisphere as given in 



