EVOLUTION 1019 



process, if a new position of organic stability has often been reached 

 brusquely, if the Proteus has frequently been leaping, marching, or 

 swinging, as well as vaguely feeling its way in creeping — in short, 

 if there be in Natura naturans also something of Natura saltatrix — 

 then what is required of the theory of Natural Selection will certainly 

 be less than Darwin supposed. Yet we do not find De Vries or Baur 

 jettisoning the theory; and we might give many other examples of 

 active investigators who remain natural selectionists also, not 

 slavishly, of course, but utilising its essential idea. 



(2) Of great importance is Johannsen's evidence that if the 

 descendants of an individual high-class bean-plant are kept apart 

 in a pure line, no amount of selection will get beyond the mean of 

 the line. There are indeed some tall plants and some short plants, 

 and other "fluctuations"; but there is nothing to choose between 

 the descendants of the tails and the descendants of the shorts. The 

 reason for this may be that the "fluctuations" in this case are non- 

 transmissible "modifications" or "acquired characters" due to 

 peculiarities in individual nurture. There is no use trying to select 

 from amidst non-heritable characters. 



But there are other reasons why we must not be swept away by 

 Johannsen's important evidence. It must be allowed, for instance, 

 that "pure lines" are not typical of wild stocks in natural conditions, 

 where cross-fertilisation is frequent; it is dangerous to argue from 

 brief pure-line experiments to the age-long processes of Nature; it 

 is premature to deny the possibility of heritable mutations occurring 

 in pure lines — ^very premature, since Baur has shown that they may 

 be common. 



(3) No doubt it is not always easy to see how a small new departure 

 could persist while still very incipient, too minute to be caught by 

 the meshes of the natural selection sieve. But it seems hardly 

 pla5dng the game to ignore the fact that Darwin, with his usual self- 

 criticism and anticipation of difficulties, gave careful consideration 

 to this very point. He laid stress on "the Correlation of Variations", 

 one being somehow linked to another; and he even suggested, as 

 Romanes and others have since done more fully, that the selection 

 of advances in a major variation might bring the minor variation 

 through the incipient stages until it also was gripped. It must also 

 be noted that a new departure may occur simultaneously in many 

 variants ; and if these are segregated, as by any of the many forms 

 of isolation, then the likelihood of the novelty being lost is greatly 

 diminished. Nowadays, of course, that difiiculty does not press if 

 the mutation in question "mendelises" ; for in that case, if it comes, 

 it comes to stay. And again, when we think of such difficult problems 

 as the evolution of the eye or the ear, it is a familiar reflection 

 that for ages the selection may have been in reference to a function 

 or significance different from that which is characteristic of the 



