288 ERYLUS SOLLASII. 



form B, most of the triaene-clades arc simple; and the microrliabds thinner and 

 shorter, and the aspidasters absolutely shorter and relatively broader than in 

 any of the others. In race I, form C, most of the triaene-clades are simple; and 

 the rhabd megascleres thicker and the aspidasters more irregular than in any 

 of the others. In race I, form D, most of the triaene-clades are branched ; and 

 the triaene-rhabdomes relatively thicker than in any of the others. In race II 

 most of the triaene-clades are simple; and the acanthtylastei-s larger, the aspid- 

 asters absolutely longer and relatively considerably narrower and thinner and 

 their rays provided with a larger number of lateral spines than in any of the 

 others. In race III most of the triaene-clades are branched; and the rhabd 

 megascleres longer, the microrliabds very considerably shorter, the triaene- 

 rhabdomes much longer, the triaene-dadonaes much broader, and the asj^id- 

 asters relatively tliicker than in any of the others. 



That the megascleres of race I, form A, are smaller in size and more simple 

 in character than those of the other forms and that there are other differences 

 of this kind, appears to be due to differences in the age (size) of the specimens. 

 Some peculiarities, as for instance the irregularity of the aspidasters of race I, 

 form C, may be pathological. Some are, no douljt, to be accounted for by 

 differences in the external forces which acted on the different specimens. All 

 these can be considered as mere somatic non-germinal characters, destitute 

 of systematic significance. There remain however some, the nature of which 

 is more doubtful and which might well be germinal. These peculiarities are 

 the exceptional narrowness and thinness of the aspidasters and the richness of 

 their rays in spines in race II, and the exceptional shortness of the micro- 

 rliabds and the exceptionally large size of the triaene-cladomes in race III. 

 If these peculiarities are considered germinal three systematic groups must be 

 distinguished, one for the forms A, B, C, and D of race I, one for race II, and 

 one for race III. 



There can, I think, be no doubt that these three groups must be united in 

 one and the same species; it is another question, however, whether or not 

 varieties should be established for them. After a careful consideration I have 

 decided that these differences are probably germinal and systematically impor- 

 tant, but sufficient only for racial distinction, and I distinguish three races, 

 designated I, II, and III, in this species accordingly. 



