1908.] IMPROVEMENT OF THE POTATO. 391 



4. METHODS OF BREEDING 

 GENERAL BASIS 



In potato breeding, there are four steps before the worker: 

 i. Selection of varieties for improvement. 2. Discovery of valuable 

 bud-variations. 3. Selections of mother plants, and their crossing. 

 4. Comparison and selection of the progeny. It is the province of 

 the breeder to discover the best methods for prosecuting this work, 

 the obstacles in the way, and the probabilities of success under 

 different conditions. The means of propagation of the potato is in 

 most of these steps a disadvantage. The comparatively simple 

 methods of the seed propagated annuals and biennials are seri- 

 ously complicated, and the advantage of grafting held forth by the 

 orchard fruits is lacking. 



In taking the first step, it is in some cases impossible to select 

 varieties with characters that are desirable, to combine with those 

 possessing other characters, because of the large number of varieties 

 which produce no viable pollen. A great number of varieties must 

 be grown, and the matter as to whether fertile pollen is produced 

 in quantities large enough to make hand pollination practicable 

 must be determined during their comparison as varieties. It has 

 been held by many horticulturists since the time of T. A. Knight 

 (60) that varieties which did not bloom readily could be stimulated 

 into fruiting by removing the soil from the underground stems, or 

 otherwise preventing the production of tubers. This does not seem 

 to be true at present with American varieties, for with ten different 

 varieties in Connecticut and with five in Illinois, we obtained by 

 this method no natural seed balls, and by microscopical examination 

 found no seeming increase in production of viable pollen. At least, 

 we might conclude that the reaction to this means is not certain 

 enough to be of great value to the breeder, and seed producing va- 

 rieties must be selected which come nearest to the ideal in the char- 

 acters needed. Some hybridists select parent varieties of similar 

 type with the idea of improving qualities already obtained, while 

 others cross widely divergent types, with the hope of obtaining a 

 smaller percentage of seedlings with much better characteristics. 

 Doubtless both methods are necessary, if it is certain in all cases 

 that there are eminently desirable characters in the types used. 



Having selected the varieties with which to work, it is of doubt- 

 ful value to use any of the old-time horticultural methods, such 

 as over supply of food materials, to induce variation. We may in 

 the future learn to use artificial means to produce mutations, but 

 the "plus" fluctuations induced by food supply are probably abso- 

 lutely valueless when we are about to resort to sexual reproduction 



