PLANT BREEDING 191 
books. Careful study of the gardens or fields of some suc- 
cessful breeder is also necessary. The five rules which follow 
are here given only to illustrate some of the precautions 
necessary to insure success. 
1. For breeding experiments choose plants like wheat, corn, 
and apples, which naturally show many varieties or elementary 
species (fig. 163). 
2. Breed for one 
character at atime; 
that is, do not try 
to get such a prod- 
uct as astrawberry 2 j 5 
whicheballsarpass Zio 18, os, esmaay onthe ri to 
most others in size, the Early Williams apple 
sweetness, bearing Modified from photograph by Burbank 
qualities, and abil- 
ity to stand shipping long distances — all the desirable quali- 
ties in one berry. 
3. Do not try to get opposite qualities in the same breed. 
For instance, corn cannot be bred for high percentage of 
starch and of protein at the same time. 
4. Choose plants for seed by inspection as they grow in 
the field. A melon or a cucumber plant which bears many 
rather large fruits is likely to be a better parent than a plant 
which bore only one very large fruit. The total product of the 
individual should be ascertained by field study, and the vigor 
and general condition of the plant should be carefully noted. 
5. Keep up the type by constant selection of the best indi- 
viduals for seed, even after breeding has resulted in securing 
plants that come up to the desired standard. This is abso- 
lutely necessary in perpetuating, for instance, the variations 
which make the difference between the sweetest and the 
poorest sugar beets; but it is also useful in keeping such 
elementary species as wheat varieties up to the highest stand- 
ard, because the seed of a desirable variety might be taken 
from imperfect or diseased ears and yield a poor crop. 
