IMPROVEMENT OF PLANTS. 353 



II. Variation Furnishes the Starting-point. 



1. Variation of Seedlings. When the charac- 

 teristics of the desired variety have been defi- 

 nitely determined, then if one will diligently and 

 carefully search among his plants, he may find 



owing to variation individuals which possess 

 these characters in a more marked degree than 

 do the others. But if such individuals are not 

 found, then 



2. Variation maybe induced by (i) ENVIRON- 

 MENTAL CHANGES. 



Important among these is () a change in 

 food-supply. Darwin says: "Of all the causes 

 which induce variability, excess of food, whether 

 or not changed in nature, is probably the most 

 powerful." 



If heavy foliage and rank-growing plants rep- 

 resent the " ideal,'* they should be given a liberal 

 supply of nitrogeneous food (see " Effect of 

 Nitrogen," Chapter IV.) If dwarf size and 

 fruitfulness are the desired characters, then 

 foods containing potash and phosphorus should 

 be substituted. 



EXPERIMENT 26. (a) To show variation induced by 

 change of food supply. Secure one-half bushel of pure 

 white sand, and sterilize^ it by thoroughly baking it in 

 a hot oven. 



(<) The tomato, geranium, etc., are suitable plants for 

 this experiment. Select three small, similarly developed 

 plants grown from cuttings of the same stock (see page 

 220). 



