282 BULLETIN 100, UNITED STATES NATIONAL MUSEUM. 



until after the twelfth plate (fourth or fifth in 'ineseres) and the 

 inferomarginals as viewed from below do not become longer than 

 wide until after the fifth or sixth. In ludwigi the subambulacral 

 granules are fewer in 2 regular series, and there are no adambulacral 

 pedicellariae. N. helli Koehler has the radial plates conspicuously 

 wider than long, while in N. ternalis (Koehler, but not Perrier) the 

 marginal i^latcs have different proportions, the adambulacral fur- 

 row spines are fewer (8) and pedicelUxriae much more numerous. 



I do not think the species which Koehler identifies as Nymphaster 

 ternalis is really that form, apart from the improbability of a West 

 Indian species being found in the Indian region. Judging by Per- 

 rier's figure, the marginal plates of ternalis are more tumid and 

 the abactinal radial plates larger. Perrier states that the apophysis 

 of the adambulacral plates of ternalis starts with the twenty-fourth 

 plate. In Koehler's species it starts on the third or fourth to 

 fifteenth. In most of the species described in this report the apophy- 

 sis starts near the base of the furrow on the first five adambulacral 

 plates and is a rather conservative character. It is unlikely that 

 so great a variation would be present in a single species. I have 

 examined what purports to be N. ternalis from the West Indies, 

 but as it agrees with neither the figure nor description of Per- 

 rier I think it likely that in the United States National Museum 

 collection other species than true ternalis masquerade under that 

 name. This species has the apophysis starting on the third or 

 fourth plate, and its superomarginals, instead of being swollen on 

 the ray and slightly sunken in the midradial line, slope upward 

 toward the median radial line. 



NYMPHASTER HABROTATUS Fisher. 



Plate 66, fig. 3; plate 67, fig. 3; plate 68, fig. 4; plate 69, fig. 4; plate 92, 



figs. 8, 8a. 



"Nymphaster hahrotatus Fisher, 1913a, p. 639. 



Diagnosis. — Rays long and slender as in N. ai'throcnemis^ but not 

 tumid, the dorsal surface of ray evenly arched proximally and oral 

 spines 12 to 15; contour of ray as seen from below even, not con- 

 stricted at intervals ; breadth of ray at inner end of the first pair of 

 superomarginals which meet medially equal to first 3 to 3.5 supero- 

 marginals measured on ambitus; superomarginals, as seen from 

 above, wider than long up to the sixth or seventh; the next 2 or 3 

 squarish ; granules round and well spaced ; a few marginal 2-jawed 

 pediceHariae; median radial plates slightly wider than long, all 

 slightly elevated; primary apical plates conspicuous; adambulacral 

 plates with proximally 7 or 8 and farther along ray upward of 14 

 furrow spines; apophysis beginning on second or third plate, but 



