172 DISCOVERY REPORTS 



Actinal intermediate plates about 8, carrying 1-4 delicate thorny spinelets, and not 

 extending beyond third or fourth adambulacral plate, except for i or 2 rudimentary 

 plates near middle of some of the rays. 



Adambulacral plates very short but broad with prominent muscular intervals. 

 Furrow comb consisting of.proximally 4 or 3 (exceptionally 5) delicate, slender, long, 

 basally webbed spines ; and a transverse actinal comb of usually 4 or 3 subambulacrals of 

 which characteristically the outer 2 (or i) are much thornier and a little shorter than the 

 inner 2 (the former i-6 mm., the latter i-8 mm.. Fig. F, ic, id). The spine numbers 

 run as follows along one side of ray, the Roman numerals of each couplet being sub- 

 ambulacral, 4-IV; 4-IV; 3-V; 3-IV; 3-III; 3-111; 3-III. On another ray it runs 

 5-IV; 3-V; 4-IV; 4-IV; 4-II; 3-III; 3-IV; 3~IV; 3-III; 3-III; 3-IV; 3-IV; 

 3-III. On outer part of ray the 2 furrow spinelets gradually align themselves with the 

 subambulacrals so that there is a single transverse fan of 5 or 6 spinelets to a plate. 



Marginal mouth spines 7 or 8, similar to adambulacrals, basally webbed ; on the con- 

 vex outer part of mouth-plates, 3-5 slender subambulacrals, the inner slightly the longer. 



Tube-feet with sucking disk about as broad as column. 



Type locality. St. 1563, Marion Island, 46° 48-4' S, 37° 49-2' E, 101-106 m., i 

 specimen. 



Remarks. The similarity of this species to L. fiircilUger^ Fisher, Alaska to Galapagos 

 Islands, 192-1100 fathoms, has been mentioned in the diagnosis. In that form the 

 abactinal 2- and 3-pronged glassy spinelets, i-i-2 mm. long, radiate in all directions 

 forming a subspherical group. The spinelets of the marginal paxillae are not 

 markedly unequal in length nor is the pedicel of the inferomarginals compressed at the 

 end. A series of spaced actinal platelets extends along the ray. In L. suluensis- Fisher the 

 pedicels of abactinal and marginal paxillae are conspicuously shorter than the spinelets, 

 which are not trifurcate; the subambulacral spines are much longer than the furrow 

 spines, and are longer than marginal paxillae with their spines. L. mariouis is probably 

 most nearly related to L. quadrispimis Clark from the Cape region. South Africa. Clark 

 compares quadrispimis to L. ftircilliger but it is more like L. stellans Sladen. There is a 

 well-developed series of actinal paxillae extending nearly to tip of ray and 8-12 similar 

 but smaller paxillae in the remainder of the interradial area. Clark states that quadri- 

 spinus has much more numerous actinal paxillae than has stellans. But stellans also has a 

 series of well-developed actinal paxillae extending to end of ray, while marionis has none. 

 I think quadrispimis and stellans are closely related. In the matter of proportions and 

 general habit my specimens of stellans agree with Clark's rather poor figures ; and in 

 addition to the series of actinal paxillae adjacent to the adambulacral plates there are 

 4-10 small paxillae in each interradial angle — a number which overlaps the 8-12 of 

 quadrispimis. In stellans the furrow spines are generally 4, 3, 2; but also, as in quadri- 

 spimis, there may be 5 or 6^ at base of ray. In quadrispimis the reported number of 



1 Fisher, 191 1, p. 334, pi. 79, figs, i, 2; pi. 114, figs, i-i^; pi. 116, fig. 5.— 1930, p. 198. 

 - Fisher, 1919, p. 449, pi. 123, fig. 2; pi. 124, fig. 3; pi. 132, figs. 8, 8a. 

 ^ Mortensen, 1933, p. 272. 



