THE LIMITS TO VAEIATION 49 



It has already been shown that Darwin entirely 

 recognized the limits which individual variations, 

 or, as they are called by de Vries, ' fluctuations,' ^ 

 may set to the progress achieved by artificial 

 selection, and that he admitted the necessity 

 of waiting for a fresh 'start in the same line'. 

 In this respect he agreed with modern writers on 

 mutation ; but differed from them in believing 

 that the fresh start would ultimately be made. 

 He also differed, as has been already abundantly 

 shown, in the magnitude assigned to the varia- 

 tions forming the steps of the onward march of 

 evolution. His observation and study of nature 

 led him to the conviction that large variations, 

 although abundant, were rarely selected, but 

 that evolution proceeded gradually and by small 



' It is to be feared that confusion will result from Dr. A. E. 

 Shipley's treatment of this subject in his address to the Zoological 

 Section of the British Association at Winnipeg as reported in the 

 Times of Aug. 28, 1909. The account of Dr. Shipley's address- 

 by now probably widely read - contains the following statement : — 

 ' Mutations were variations arising in the germ-cells and due to 

 causes of which we were wholly ignorant ; fluctuations were varia- 

 tions arising in the body or " soma " owing to the action of external 

 conditions. The former were undoubtedly inherited, the latter very 

 probably not.' The term ' Fluctuation ' or ' Fluctuating Variability ' 

 has been applied by de Vries to what Darwin called ' individual 

 variability ', — 'determining the differences which are always to be 

 seen between parents and their children, or between the childi-en 

 themselves ' (Species and Varieties, H. de Vries, 1906, 190). To 

 speak of these differences as ' very probably not ' inherited, is to 

 follow neither Darwin, nor Weismann, nor de Vries, but simply 

 to cause gratuitous confusion by questioning an accepted con- 

 clusion based upon universal experience. The reported statement 

 as to the nature of fluctuations would, if it were correct, prove that 

 the hereditary transmission of acquired characters takes place on 

 the vastest imaginable scale. But, although no one disputes that 

 fluctuations are hereditary, very few indeed will agree that they 

 are due ' io the action of external conditions ', or in other words 

 ' acquired characters '. See Appendix D, p. 258. 



