LUTHER BURBANK 



experiment are greatly facilitated by the fact that 

 the flowers are self-fertilized. With these plants, 

 as with the small grains, this becomes an important 

 aid in fixing a type, and in maintaining a pure 

 race once it has been developed. 



For the most part, my experiments with the 

 clovers have been made through selection, and 

 without resort to hybridization. But in excep- 

 tional cases I have cross pollenated these plants, 

 to test the possibilities of work in this line. I 

 found that the process involves no great difficul- 

 ties, notwithstanding the small size of the flowers. 



In practice I found it better to remove all but 

 two or three flowers in a dlover head. 



The remaining ones have the petals and 

 stamens removed with a small pair of forceps, 

 after which the application of pollen from another 

 clover head presents no special difficulties; care 

 being taken, of course, to see that the pistil is at 

 the right stage of development. 



Developing New Characteristic? of Stem 

 AND Leaf 



In the course of these experiments I have grown 

 in the neighborhood of two hundred species of 

 clover. Many of these are native species, some of 

 which invaded my grounds unasked. Others have 

 been received from far away regions, in particular 

 from Peru, Bolivia, and Chile, 



[84] 



