ASTEROIDEA OP NOBTH PACIFIC AND ADJACENT WATEES — FISHEB. 269 



So great and numerous are the variations in most of the species recorded hclow 

 that eacli is to be regarded more as a center of variation, deviations from the type 

 proccechng in many directions until thoy meet and often merge with aberrant mem- 

 bers of nearly related forms. This will very likely hold true in many other genera 

 when enough specimens from numerous localities are examined. Another wav of 

 stating the idea (wliich is not new) is tliis: So-called species of starfishes intergrade 

 in various ways (geographically, bathymetrically, by individual variation, and jjcr- 

 haps also by hybritlization) with neighboring antl sometimes also with accompanying 

 forms. Many species seem well marked only when wo have a few specimens. 

 Furthermore, wide differences in outward structure form no barriers to such inter- 

 gradation, wliich may bo accomplished from one species to another through a series 

 of intermediate species or races. 



With deliberate inconsistency I have called the well-marked types species, dis- 

 regarding certain indications of mtergradation between sanguinolenta and hviuscuh, 

 and between leviuscula and lu^pera; and I have called subspecies certain small species 

 which are obviously closely related to and were found to intergrade with sanguino- 

 leiita and leviuscula. Intergradation seems to take place between sanguinolcnta and 

 (schrichtii and leviuscula; between eschrichtii and leviuscula; between leviuscula, and 

 muUispina; between inultispina and eschrichtii (not so e\'ident); between leviuscula 

 aiul annectens; between annectens and aspera. Certain of these cases I feel sure are 

 due to hybritUsm Wien two or more species range together, "freaks" and aberrant 

 specimens immediately make their appearance. It is sometimes impossible to 

 classify' these; they are very bafHing. Since hybridism is possible and in some cases 

 probable, it has seemed better to keep such distinct types as leviuscula and aspera 

 separate from sanguinolenta, and to rank under each the races wliich are evidently 

 most nearly related to them. Thus we have Henriria leviuscula annectens and 

 Ilenricia leviuscula muUispina rather than Ilenricia sanguinolenta annectens and 

 Henricia sangidnolenta multispina, which would obtain if the custom of ornithologists 

 were followed. I am not con^^nced that the terrestrial subspecies has an exact 

 counterpart in the sea, nor the extreme and baffling variation of the Asteroidea 

 a counterpart in any class of vertebrates. 



If the supposed relationships were to he fuU;/ expressed in the name we should 

 be obliged to denominate the first six species somewhat as follows: (1) Ilenricia 

 sanguinolenta; (_>) Ilenricia sanguinolcnta eschrichtii; (.3) Ilenricia sanguinolenta 

 leviuscula or Ilenricia sanguinolenta eschrichtii leviuscula; (4) Ilenricia sanguinolenta 

 leviuscula muUispina or Ilenricia sanguinolenta eschrichtii Uviuscula muUiipina; (.5) 

 Henricia sanguinolenta leviuscula annectens; (6) Ilenriaa sanguinohnta hviuscula 

 annectens aspera. This will indicate why for practical purposes it is better to keep 

 the well-marked types separate, even if evidence of intergradation is present. The 

 subspecies, races, or varieties are in all ciuses simply deviations from the t]ipe. Ilen- 

 ricia sanguinolenta. annectens is strictly what should be written, but if the form in 

 question has obviously nothing to do with sanguinohnta an<l is as obviously related 

 to leviuscula, I believe the course followed to be the best, even if it is ultimately 

 ])roved beyond peradventure that sanguinohnta and leriuscula intergrade without 

 hvbridization. Similarly I think it better, if leinuscuhi, aspera, and sanguinohnta 

 constitute well marked jukI well differentiated t>-i)es to keep tiiem separate rather 



