284 THE EFFECT OF FOOD 



Somewhat similar results to these were obtained by 

 Lawes and Gilbert for wheat, bean, clover, and other 

 crops, but it is deemed unnecessary to reproduce them 

 here. 



Most striking evidence as to the influence of nutri- 

 tion on variations has been obtained by De Tries.* 

 When carrying out his artificial selection experiments 

 on five-leaved clover, he found that in one series of ob- 

 servations seeds from some plants grown in a poor soil 

 yielded 39 per cent, of 3-leaved, and 48 per cent, of 5- 

 to 7-leaved clover. Those from plants of the same 

 stock which had been grown in a rich soil, however, 

 gave only 14 per cent, of 3-leaved, and 73 per cent, of 

 5- to 7-leaved clover. 



The effect of nutrition on Ranunculus bulbosus 

 (Crowfoot) was almost as striking. f Wild flowers col- 

 lected near Hilversum were found by De Vries to have 

 the following frequencies of distribution in the numbers 

 of their petals : 



Number of petals, 5 6 7 8 9 10 11 

 657 41 11 2 4 2 



In the autumn of 1887 De Vries planted some of 

 these plants in his culture garden, where they bloomed 

 the following year. Owing, presumably, to the better 

 nutrition, the proportion of flowers with more than five 

 petals was considerably increased: 



Number of petals, 5 6 7 8 9 10 



133 55 23 7 2 2 



The seed from the many-petalled flowers was collected 



*" Die Mutationstheorie," p. 448. 



fBer. d. deutsch. Bot. Ges., xii. p. 197, 1894. 



