HEEBDITY 



amination of 717 flowers in the field made by 

 De Vries in 1887 showed the rather frequent 

 occurrence of 6 and 7 petaled flowers also, the 

 average number of petals in the entire collec- 

 tion being 5.13. De Vries set himself the task 

 to see if the proportion of many petaled flowers 

 could be increased or the number of petals to 

 a flower be further increased. In both these 

 respects he succeeded surprisingly well. As a 

 result of five successive selections the average 

 number of petals was raised from 5.6 to 8.6, 

 the upper limit of variation from 8 to 31, and 

 the mode (or commonest condition) from 5 to 9. 

 Singularly enough De Vries concludes, in ac- 

 cordance with general ideas which he had 

 adopted, that selection had in this case done 

 practically all that it could accomplish, that 

 further selection, while it might advance the 

 average somewhat farther, would have no per- 

 manent effect in modifying the type. This be- 

 lief seems to have rested on considerations 

 such as these. De Vries had found, as had 

 others, that variations which are heritable have 

 their origin in the germ-cells only. He recog- 

 nized that the tendency to produce double 



108 



