1921] ■ The Genus Raspailia 59 



actually happened, since, it would seem, we must recognize as a real 

 1. e. natural, process the independent varying of characters. This 

 consideration brings with it the conviction that classification in its de- 

 tails cannot be a completely satisfactory guide to relationship. Minor 

 uncertainties will occur everywhere. 



The only species of Raspailia recorded for our waters (North Amer- 

 ica and West Indies), are the following: 



R. haniata O. Schmidt (1870, p. 62), West Indies. The diagnosis 

 is very incomplete and Schmidt puts a query as to the genus. But 

 the species, which is of typical habitus, is probably to be referred here. 

 (R. tenuis Ridley and Dendy, 1887, p. 188, occurs off the Brazilian 

 coast in shallow water to twenty fathoms) . 



R. acanthifera (George and Wilson), North Carolina coast (Fort 

 Macon Beach). Somewhat reluctantly I conclude that this species 

 originally referred to Axinella shoud be transferred to Raspailia, be- 

 cause of the presence of acanthostyles and the character of the der- 

 mal brushes, which consist of a bunch of ordinary styles together with 

 one to a few long ones projecting far beyond the others. The species 

 is to be regarded (cf. George and Wilson, 1919, p. 160), as a connecting 

 link between Raspailia and Axinella. 



When one speaks, as in this paper, of the independent variability 

 of structures, it is not meant to imply that any part of a body is reallj^ 

 an independent variable. It would rather seem from all that is known 

 concerning racial differences, that where a race has varied from the 

 ancestral type in one feature which is conspicuous from our present 

 view-point, it has probably come to differ from the type in a great 

 many other features, although these may become known to us only 

 gradually as we become intimately familiar with the race in question. 

 It is possible and convenient to refer, as I and others have pointed out, 

 all this kind of variation in natural races to the gene theory, since in 

 contemporary speculation a gene is a germinal unit which, like a 

 molecule in a radical, may influence the organism (compound) in 

 many different directions. 



A good case of close correlation between specific characters in the 

 matter of variability has been pointed out in recent years by Hent- 

 schel (1913 a, b), who has shown that in the silicious sponge Mycale 

 (Esperella) a large number of characters ("character values" as Hent- 

 schel calls them, representing the concrete realizations of a schematic 

 character, itself an abstraction drawn from a survey of the several 



