276 BREEDING CROP PLANT* 



Stewart to conclude that there was more evidence in favor of 

 purity of the clone than in favor of the value of clonal selection 

 as a means of producing higher-yielding strains. Similar con- 

 clusions were reached from an experiment carried on by Tyson 

 brothers, in New York, with the York Imperial apple. Two 

 trees were selected which bore unusually similar fruits and these 

 were used for propagation. More than 8,000 trees were planted 

 in the new orchard. Examination of trees of this orchard when 

 they came into bearing showed them to be not superior to the 

 usual York Imperial apple (Dorsey, 1917). 



The cited cases show the present status of the problem of selec- 

 tion of bud sports as a means of improvement of fruit crops. 

 The studies with the citrus genus appear to justify the belief 

 that degenerate or inferior bud sports are of frequent occurrence. 

 This leads to a conclusion that only those limbs which produce 

 normally healthy fruit should be used for propagation purposes. 

 Even among the citrus fruits there is as yet no very conclusive 

 proof that the selection of cions from high-yielding trees will 

 accomplish more than to prevent possible "running out" of the 

 variety. The evidence from apples would seem to justify the 

 belief that bud sports are very infrequent. The breeder, then, 

 can well afford to make careful observations with the hope of 

 discovering bud sports. If apparently desirable sports are found, 

 these may then be used for propagation. 



In such crops as citrus fruits and with such plants as Coleus, 

 bud sports are of frequent occurrence. There is, then, some evi- 

 dence for the belief that sports occur more frequently in hetero- 

 zygous than in homozygous material. As Stout (1915) obtained 

 the same changes through asexual selection as by the use of self- 

 fertilized seed, it seems reasonable to suppose that some sort of 

 segregation and recombination occurs in somatic tissue. No 

 cytological evidence has been given to account for such a supposi- 

 tion. With heterozygous material the loss of a single dominant 

 factor would be immediately apparent in the soma. This is one 

 reason why bud sports occur more frequently in heterozygous 

 forms (East and Jones, 1919). Nabours (1919) has shown that 

 similar cross-overs occur in parthenogenetic reproduction in the 

 grouse locust as in those forms which are produced by the recom- 

 bination of gametes containing the haploid number of chromo- 

 somes. If the usual sort of cross-overs occurred in homozygous 

 material, there would be no change in the homologous parts of 



