ON THE ORIGIN OF SPECIES. 487 



rience, do not exist for Scott. Each species remains unchanged as 

 long as its period of existence lasts. All its characters vary more or 

 less according to the law of Quetelet, but the type, to which all varia- 

 tions return, remains the same through centuries. 



A species changes only when it produces others. Or rather, it does 

 not change, but continues to exist next to the species newly formed. It 

 may be compared to a tree, which, though it produces branches, does 

 not cease growing in length. Only when among its descendants there 

 are types better fitted for the battle of life, a species may locally suc- 

 cumb. But it would require a long time before the new species had 

 entirely taken the place of the old. 



It is clear that one must distinguish by some simple term varia- 

 tion by jumps from variation obeying the law of Quetelet. It is not 

 practical to use the terms, sport, discontinuous variation or spontane- 

 ous variation, since they tend to produce the impression of something 

 incomprehensible. Scott did not use these terms. He speaks of 'mu- 

 tations. ' A mutation occurs when one species is formed from another. 

 As it is, 'mutation' is the expression in general use before the days of 

 Darwin, and at first used by Darwin himself. Since it has apparently 

 fallen into disuse, except in paleontology, where it is met with in vari- 

 ous authors, always conveying the same meaning, it seems best to 

 continue to use this term. Hence as long as species produce others 

 they are called mutable, and this part of the doctrine of variability is 

 known as mutability. 



Once it has been conceded that species originate from others by 

 mutation, one can go on to investigate what deductions must be made 

 in regard to this process from the facts with which we are acquainted. 

 And as long as an empirical investigation was impossible it was of the 

 utmost importance to be able, even in this manner, to form an opin- 

 ion about it. 



First of all, we can come to the conclusion that mutations must 

 be the smallest changes which can produce a difference between two 

 species or rather between two constant types. Ordinarily the estima- 

 tion of the differences existing between two related species is too 

 great. Differences as between a horse and a donkey are of course not 

 the result of a single mutation; there must have been a series of 

 transition forms, at present extinct. Nobody will expect to see so 

 great a change occur at once. Even much slighter differences, for 

 instance those existing between our native violets, are still too large; 

 here also there must have existed transition forms. And indeed a 

 comparison with the floras of other countries actually does show a 

 number of forms which bridge over these differences. 



