Application of Biological Principles to Plant Breeding 135 



which are heteroz)'gous for starchiness the matter is cleared 

 up at once. If it is a question of which of a lot of beans is 

 homozygous for the gene-complex of large size, the com- 

 plexity of factors concerned may require a greater number 

 of progeny, but the test is valid in the end. 



There remains for mention a phenomenon of some inter- 

 est even though it has produced few varieties of plants of 

 commercial importance. This is the sudden appearance 

 of a branch with characteristics different from the mother 

 plant upon which it is borne that may be cut off and propa- 

 gated asexually. It is the so-called bud sport. It is of 

 practically no importance outside of the production of floral 

 novelties. Perhaps this is accounted for by the fact that 

 bud variations nearly always affect the same characters 

 that have previously been changed in the same way through 

 seed variations. Furthermore, the change is practicalh- 

 always the loss of a character which leaves little oppor- 

 tunity for the production of the real novelties through 

 progressive variations (Fig. 56). 



The production of the smooth-skinned peacli, the nec- 

 tarine, as a sport, from the ordinary peach tree, is the classi- 

 cal example of this type of variation. This is undoubtedl)- 

 simply the loss of the Mendelian factor for presence of the 

 dowTi upon the fruit, and might be expected to come about 

 through some abnormal cell division in much the same man- 

 ner that variations occur in the reproductive cells. They 

 are usually not inherited through the seeds. This is what 

 would be expected. It is said, however, that sometimes 

 such variations come perfectly true to seed. If this is so, 

 one must suppose that the varying plant cell was one which 

 could give rise to the reproductive cells. Since nothing 

 dehnite is known of this matter, however, speculation does 



