482 ANNUAL. REPORT OF THE Off. Doc. 



of the sire is unknown. In the case of the other cereals or of pota- 

 toes, the size of the kernel or tuber is no necessary measure of the 

 productiveness of the parent. A small kernel from a fine well-bred 

 individual is better than a large kernel from a poor indifferently 

 bred individual. Other things equal, a small tuber from a large 

 hill of potatoes is better than a large tuber from a small hill. In 

 case the large and small seeds come from equally good heads of 

 wheat, which will probably be the case under average conditions, 

 the large seeds may perhaps give the best results, especially as under 

 field conditions the larger size may be of advantage in enabling the 

 plant to get a more vigorous start. Hays believes it to be estab- 

 lished that the best heads of wheat as well as the best plants should 

 be selected. To succeed in plant breeding, the seed must be selected 

 from individuals which possess the characteristics it is desired to 

 perpetuate, while the characteristics of the seed are of minor im- 

 portance, provided the vitality has not been impaired. 



Steps in the Improvement of Plants : There are three steps in the 

 improvement of plants or animals, viz., (A) inducing variation; (B) 

 selection of forms having desired characteristics; (C) testing the 

 power of specific forms to reproduce themselves. 



Variation: Variation is the basis of selection. Plants must vary 

 or they could not be selected. There are two general methods of 

 producing variations, viz., (1) environment, such as soil, climate, 

 space, cultivation, etc.; (2) crossing. 



The causes of variability cannot be discussed here but the follow- 

 ing facts should guide the breeder of plants: 



(1) Horticulturists do take advantage of a superabundance of food 

 in causing modification or multiplication of parts, such as the de- 

 velopment of petals from samens. After this habit becomes fixed 

 it will be transmitted in some measure, even in poor soil. 



(2) Nevertheless, the most important value of cultivation in the 

 case of most plants is to allow the plant breeder or cultivator to 

 study individual forms. It enables him to select the desirable forms 

 and reject the undesirable ones. By milking the cow and testing her 

 milk we are able to select the best milkers. By trotting horses we 

 are enabled to breed those best able to trot. Whatever influence 

 milking or trotting may have, the fact remains that it makes possi- 

 ble intelligent selection. 



(3) The variations selected should be those induced under the en- 

 vironment in which we expect to continue to grow the crop. If we 

 expect to grow three stalks of corn to the hill in general field culture 

 it is desirable to select the ears for planting from corn grown in a 

 similar manner rather than from ears where but one stalk is grown 

 in a hill. In the latter case the size of the ear will not be a criterion 

 of the size of the ear where three stalks are grown in a hill. Where it 

 is not possible to make selection under field conditions, care should 

 be taken to select from among plants under like environment and 

 subsequently subject to field conditions. 



"In selecting sugar beets," says Vilmorin, "those roots are sought 

 for that are straight, long and free from lateral branches. This is 

 right, for those that are branched are more difficult and hence more 

 expensive to gather. Now, certain growers of beet seed in the 

 north of France once formed the idea — thinking, no doubt, in this 



