HUBEECHT ON 'FLUCTUATIONS' 269 



forth different races within the limits of these species, but 

 whenever selection ceases the race is turned back to the 

 parent form. The maximum deviation in these races is 

 generally obtained after three or four generations of con- 

 tinuous selection ; it takes about as many generations to 

 bring back the parent form.' (p. 219.) 



(7) ' The fact that artificial selection of fluctuating varieties, 

 as well as hybridizing, etc., has already led to such indis- 

 putable improvements in the different races of animals and 

 plants may, however, etc' (p. 223.) 



Finally in an article only published about a 

 year ago in the Contemporary Review^ Professor 

 Hubrecht says : — 



'Wherever our agriculturist succeeds by the most 

 careful artificial selection in producing (e. g.) a beetroot of 

 which the percentage of sugar has been raised, say, to 15 per 

 cent, out of roots which originally stood at 7 to 8 per cent., 

 he knows that the fluctuating variation of the beetroot has 

 permitted him to attain this end ; but he knows, at the 

 same time, that what he has obtained is not a new species 

 of beetroot, richer in sugar, but a product of nature which 

 the moment it is left to itself and freed from the bonds of 

 artificial selection goes back to an inferior sugar-producing 

 root again.' (p. 633.) 



I will now prove, although more briefly, that 

 other writers have understood de Vries cor- 

 rectly. The sectional heading employed by 

 Professor C. B. Davenport — ' Mutation vs. Sum- 

 mation OF Fluctuations ' ^ — is sufficient to show 

 this ; for summation would be impossible without 

 hereditary transmission. We do not, however, 



' For Nov., 1908, ' Darwinism veraus Wallaceism.' 

 2 Fifty Years of Danvmism, New York (1909j, 173. 



