FLUCTUATIONS AND MUTATIONS 99 



the more cautiously. An authority on domesticated animals, 

 Prof. Keller of Zurich, finds but little evidence of it in the history 

 of the well-known stocks. 



It seems to us that in emphasising the importance of mutations 

 De Vries has swung to the extreme of greatly depreciating the 

 importance of fluctuations. Until we know more about animal 

 mutations, it does not seem to us legitimate to deny that fluc- 

 tuations may form, as Darwin believed, an important part of 

 the raw material on which selection operates. 



We cannot but regard with suspicion the distinction between 

 large fluctuations and small mutations. It seems to us a verbal 

 distinction. 



Finally, it must be remembered that, as De Vries frankly 

 points out, we are ignorant in regard to the conditions in which 

 mutations occur. The Mutation Theory does not as yet give 

 us a theory of mutations. 



" Pure Lines." The position held by De Vries has been strength- 

 ened by the work of Johannsen and Jennings on " pure lines." If we 

 succeed in starting a " pure line " " the progeny of a single self- 

 f ertilised homozygous plant" say an innately exceptional bean-plant 

 with very large seeds, we shall find slight individual differences in the 

 size of the beans from generation to generation ; if we take the biggest 

 and the smallest of these and start afresh, we find that their progeny 

 are neither larger nor smaller than the average. The original bigness 

 was a -fixed mutation ; the other differences were probably mere 

 modifications and non-transmissible. If we take a considerable 

 number of the largest beans and the smallest beans from a field and 

 sow them, we are likely to get in the progeny of the former a larger 

 average size than in the progeny of the latter, for we are almost sure 

 to have started with a number of beans which are innately (not modi- 

 ficationally) large-sized and small-sized. What Johannsen did for 

 the bean and some other plants, Jennings has done for the slipper- 

 animalcule, Paramascium. He isolated eight pure lines differing in 

 average size, and found that he made no progress by selecting the 

 largest in an established large pure line, the exceptional largeness 

 being probably the accidental result of peculiar nurture. Selection 

 from a mixed population, however, resulted, as in the case of the 

 beans, in a distinctly altered average size. 



