461 THE AMEBIC AN NATURALIST [Vol. XLIV 



second generation plants was planted from certain par- 

 ent plants, and the records were kept in such a way that 

 the characters of the offspring could be compared with 

 those of the parents. As stated above, the .number of 

 individuals was of necessity small. yet since the study 

 was made under two different soil environments it 

 seems quite representative of what one is to expect in 

 ordinary selection work. 



The difference between this study and regular selec- 

 tion work is that no selection of parents was made, while 

 in ordinary selection work the best individuals arc saved 



These data are presented to answer the questions — 

 1 'What basis have we to expect a cumulative effect of 

 selection? Do we rind the resulting offspring following 

 the law of regression ? ' ' 



