268 BULLETIN 76, UNITED STATES NATIONAL MUSEUM. 



C-. Spinelels in definite pseudopaxillse or in fascicular groups. 

 d'. Adambulacral ppinelets less than twenty, 

 f'. General appearance re.sembling leviuscula with email pseudopaxillse, papular areas not 

 unusually large, often slightly larger than plates; spinelets not in spaced fascicular 

 groups, but in definite pseudopaxillse; no interniarginals. 

 /'. General form intermediate between aspera and leviuscula; papular areas usually small; 

 pseudopaxillse oblong with two or three rows of upward to twenty very short stubby 

 spinelets; prominent wide and short (compressed) marginals; no intermarginals; 

 adambulacral spinelets short, usually in two transverse series; only one spinelet in 



furrow all along ray annectens, p. 291. 



f^. General form intermediate between leriuscula and multispina, but nearer leviuscula; 

 abactinal pseudopaxillse small, often smaller than papular areas, and spinelets 

 numerous, delicate, ending in several minute points; actinal intermediate plates 

 ending near but not at tip of ray; two furrow spinelets near tip of ray; marginal plates 



not conspicuously compressed, with wide sutures between (f(/scn7a, p. 289. 



<?. General appearance not at all like leviuscula; papular areas fairly to very large; spinelets 



fairly to unusually long, in spaced fascicular groups; a long series of intermarginals; 



marginals not wider than long, armed with a cluster of comparatively long spinelets; 



adambulacral spinelets unusually long in a zigzag transverse series. 



/'. Papular areas not unusually large, about one to two times as wide as the intervening 



skeletal pieces; actinal interradial areas very email; secondary plates or calcareous 



grains in papular areas, when present, few longispina, p. 299. 



/-. Papular areas unusually large, the skeletal pieces slender and forming a wide mesh 

 bearing spaced fascicular groups of spinelets ; papular areas containing numerous small 

 calcareous grains, often with a single spinelet; large actinal interradial areas containing 



numerous plates and papulie; large intermarginal papular areas aleutica, p. 300. 



d-. Adambulacral spinelets more than twenty; abactinal plates rounded or crescentic (with 

 very numerous, delicate spines) and much less crowded than in leviuscula; adambulacral 



spinelets twenty-five to thirty spiculifera, p. 295. 



n^. Regularly two or more spinelets on the furrow face of the adambulacral plates all along ray; adambu- 

 lacral spinelets numerous; rays flexible. 

 6'. Abactinal and lateral pseudopaxillse small, evenly spaced; papulae single, arranged in fairly regu- 

 lar oblique transverse series on either side of the median radial line (at least when viewed 



internally) polyacantha, p. 302. 



b-. Plates small forming a sinuous irregular mesh work inclosing smaller, lower plates also more or less 

 joined in a close mesh between which emerge single papulae, which do not form transverse 

 oblique rows clarlci, p. 303. 



Remarks. — A little over one thousand specimens have been examined in prepar- 

 ing the following account of North Pacific species of Ilenricia. Probably no better 

 collection exists, at least for a region of similar extent. Although well acquainted 

 with the variability of starfishes, I have never before met \vith such an extreme 

 example as the species of this genus present. All attempts at classification are to be 

 regarded as provisional, but in the case of Henricia especial emphasis should be 

 directed to tliis fact. The more chverse individual and other variations become, the 

 more difficult it is to limit precisely the species of a genus, and the more must per- 

 sonal equation and fallible judgment figure in the final decision. For this reason it 

 is probable tliat another worker, udth the same collection would differ considerably 

 in his results. Tiie problem resolves itself into observation and interpretation of 

 not very obvious evidence — evidence in which many important gaps are present 

 and e^ndence which may be construed in several different ways. 



