260 PEor. p. J. BELL ON [Mar. 14, 



How is one less qualified than either of these two observers to 

 settle a question on which thej, with their ample knowledge, are 

 divided ? I, at any rate, pretend to no right or authority, but, 

 as I have to assign the specimens to a systematic place in a list of 

 Asteroidea which I am now preparing, it may be permitted to say 

 why they are to be found \dth the Archasteridfe. 



If we take the definitions offered us in the Introduction to the 

 Eeport on the ' Challenger ' collection of Starfishes, we find we 

 might place Odontaster almost indifferently with either Archasteridae 

 or Pentagonasteridae, though there is a slight balance in favour of 

 the former. If we take M. Perrier's latest revision of his scheme 

 of classification, there can be no doubt that the definition of the 

 order Paxillosse' {op. cit. p. K. 71) applies to Odontaster very much 

 better than does the definition of the Valvulatje ; and the Pentagon- 

 asteridae belong to the latter, the Archasteridae to the former of 

 these two orders. 



But at this point I must stop or I shall be drawn into a discus- 

 sion as to the classification of the subdivisions of Asteroids, which 

 would be far too lengthy for me just now. Only let me point the 

 moral that the present state of the classification of Asteroids is not 

 as satisfactory as one would hope to find it after the labours of two 

 industrious zoologists for five or more years, and that Odontaster 

 is sufficient to prove that no classification can be satisfactory which 

 allows of a Mide gap between the Archasteridae and the Pentagon- 

 asteridae. 



A list of the species of Odontaster with the synonyms may be 

 useful. 



OdOjSTASTEK. 



OdontaMer, YerrUl, Am. J. Sci. xx. (1880) p. 402. 

 Gmithaster, Sladen, ChaU. Eep. Ast. (1889) p. 285. 

 Asterodon, Perrier, Miss. Cap Horn, Echinodermes (1891), 

 p. K. 132. 



1. Odontaster belli. 



Pentaqonaster belli, Studer, Anhang zu Abb. k. Akad. "Wiss. 

 Berlin, 1884 (1885), p. 31. 



Prof. Perrier thinks this may be a synonym of Odontaster sinffit- 

 laris, but the ventral plates of that species are in no way paxilli- 

 form ; the adambulacral spine of 0. belli appears to be specially 

 modified, and in the distal fourth of the arm the intermediate 

 actinal plates are wanting, Por these reasons separation appears 

 to be, at present, the safe course. 



2. Odontaster dilatatus. 



Pentagonaster dilatatus, Perrier, Arch. Zool. exper. v. (1876) 

 p. 33. 



Gnathaster dilatatus, Sladen, Chall. Eep. Ast. (1889) p. 750. 



^ May I take this opportunity of remarking that "paxilli," and not " paxillje," 

 is the correct form of the name of these plates ? 



