OPHIUROIDEA 211 



than half the disk radius ; proximal dorsal plates rounded, 

 quadrangular ; mouth shield rounded, triangular, with 

 convex outer edge . . . A. grandisquama Lym. 



Lower arm spine not curved, sabre shaped ; upper and lower 

 spine somewhat longer than the others ; radial shields large, 

 more than half the disk radius ; proximal dorsal plates 

 triangular ; mouth shields triangular, somewhat heart shaped, 

 the outer edge being more or less excavated ... 9 



9. Arm spines 5 ^ .... A. josephince 'Ljungm. 



Arm spines 4^ ..... . ^. ms Lyman 



10. Arm spines simple, 4 or 3, ventral side of disk partly scaled^ 



A. digna Koehler 



A. abyssorum Norm. 



Arm spines serrate or, partly, hatchet shaped ; ventral side of 



disk naked . . . . . . . .11 



11. Arm spines 7, with distal edge serrate, proximal edge smooth, 



spine often ending in a fine hook ; outer mouth papilla small, 



conical . . . . . 3. A. denticulata Koehler 



Armi spines partly hatchet shaped . . . . .12 



12. Dorsal side of disk wholly or partly naked ; outer mouth 



papilla small, scale-like . . . . . .13 



Dorsal side of disk wholly scale -covered ; outer mouth papilla 

 conical ......... 14 



13. Dorsal side of disk wholly naked, excepting the narrow, con- 



tiguous radial shields . 6. A. securigera (Diib. and Koren) 



Dorsal side of disk naked in the interradial spaces, scale -covered 



in the middle . . . . 5. A. horealis (G. O. Sars) 



14. Ventral plates with concave outer edges and produced outer 



corners ; broadly contiguous ; disk scales of dorsal side 

 coarse, primary plates not distinct 4. A. filiformis O. F. M. 

 Ventral plates with rounded outer edge, outer corners not pro- 

 duced ; narrowly contiguous ; scales of dorsal side of disk 

 fine, primary plates fairly distinct . A. Sarsi Ljungm. 



^ It seems very difficult to distinguish the species iris and josephince ; 

 the difference in the number of arm spines being scarcely constant enough 

 for separating the two forms. In any case, it would seem very probable 

 that the Atlantic specimens referred to A. iris are really A. josephince ; 

 whether the Pacific form may deserve being recognised as a separate species, 

 different from A. josephince, is a question that requires further study ; but 

 this does not concern us here. 



^ Amphiura abyssorum is insufficiently known (has never been figured), 

 and it is therefore impossible to indicate how it may be distinguished from 

 A. dignn ; the character of the ventral side of the disk is unknown (cf. 

 Proc. R. Soc. London, vol. 25, 187G, p. 215). 



