202 A. E. Verrill — Revision Genera and Species of Starfishes. 



of these spines arises from near the apex of each dentary plate. 

 Both conditions sometimes occur, abnormally, on the same specimen. 



An odd interradial marginal plate, above and below, on each side. 



Abactinal surface covered with more or less paxilliform plates, 

 parapaxillae or protopaxillae, with intervening large papular pores 

 on the radial areas. The abactinal plates may bear clusters of more 

 or less elongated spines, or a group of small granules. They usually 

 form obliquely transverse lines on the rays, not always regular. 



Actinal plates angular, covered either with sj^ines or granules. 

 Small simple pedicellarife sometimes occur on the actinal or abactinal 

 plates. They may have two, thi'ee, or four papilliform blades. 



Adambulacral plates usually bear elongated spinules arranged in 

 three or four pairs of small transverse rows, generally only two or 

 three of the furrow-series are on each plate ; sometimes only one. 

 Dentary plates usually have elongated, acute marginal and apical 

 spines. They are sometimes closely united along the median suture; 

 in other cases {Odo7itaster),,i\xej are separated by a space covered 

 only by membrane. 



The marginal plates are covered either with spinules or granules; 

 sometimes the upper ones are granulated and the lower spinulose, 

 like the corresponding disk-plates ; they usually have deep fas- 

 ciolated sutures. 



Perrier, 1894, instituted a sub-family under the name Gnathaster- 

 ince to include the present group, together with some other forms 

 [Hoplaster] in which no recurved jaw-spines occur. He based the 

 group more particularly on the odd interradial marginal plate. But 

 the latter character seems to me to be of less importance, for I have 

 found it abnormally present in various species of Tosia and allied 

 genera. (See under Pyrenaster dentatus, p. 167, above). Moreover 

 in JBenthojyecten, which normally has an odd interradial plate, I 

 have found it replaced by two plates, on some of the margins (see p. 

 218 below and pi. xxx, figs. 7, la). 



Therefore, I have taken the presence of the recurved dentary 

 spines as the special feature of the group.* 



As the name Gnathaster is nearly a synonym of Odontaster, and 

 may, therefore, be dropped from the system by future authors, it 

 seems desirable to change the name of the group to Odontasteridce. 



Perrier has divided the group into three genera : Gnathaster, 

 Asterodon, and Goniodon. 



* A similar recurved tooth is found in certain species of Pterasteridce. 



