The Theory of Evolution 225 



Darwin was obliged in the later editions of the "Origin of 

 Species" to assume that a beneficial variation, in order to get a 

 foothold, must appear in a relatively large number of individuals. 

 His critics were not slow in pointing out that if we suppose a 

 variation to be so common that it occurred in many individuals 

 at the same time, it is probably the result of some definite process 

 taking place, and the change into a new species might be brought 

 about in time without selection of any kind being necessary. 

 Hence one assumption will do instead of two, and the explanation 

 is correspondingly simplified. If we still try to save Darwin's 

 theory by assuming that we need only suppose that half of the 

 fluctuations in a given direction, i.e. those on one side of the 

 mean, survive and furnish the basis for the next generation, and 

 that this process repeats itself in every generation, we meet with 

 a fact, apparently well established, that shows how futile it is to 

 attempt to save the theory in this way. It has been shown by 

 actual experiment that all that such a process accomplishes is 

 to raise the average in the direction of selection, and that how- 

 ever long the process is continued, and however severe the selec- 

 tion, nothing new, no new species, can be created in this way. 

 As soon as the selection ceases, the form quickly sHps back to its 

 original condition. 



Let us see if the mutation theory can surmount this difficulty, 

 for if mutations appear in the small numbers shown by de Vries's 

 experiment, it may seem that they, too, might quickly become 

 swamped by back-crossing with the parent type. 



In the first place the new type is at the start different from the 

 parent, and may be capable of living in a different locahty, i.e. 

 under somewhat different external conditions. Hence the chance 

 of becoming swamped by back-crossing is lessened. If more 

 than one individual occurs in the proper locahty, the chance for 

 propagation is given. Since the mutations in the evening prim- 

 rose are produced year after year, this condition may at some 

 time be reahzed. In hermaphroditic forms, moreover, the oppor- 

 tunity of self-fertihzation exists, and this increases the chance of 

 the mutation establishing itself. 



Q 



