774 DISCUSSION ON WHEAT: 
same species but even between individuals of the same variety. The 
majority of plants constituting improved varieties of cereals, in this 
country, reveal on close examination striking morphological and substantive 
differences. 
Morphological differences are most readily noticed. Tach plant in a 
plot cf thousands may have its own individuality and still bear un- 
mistakable evidence of belonging to the same variety. Substantive 
differences, on the other hand, are more subtle. In individual plants these 
qualities cannot be measured, weighed, or accurately determined. Know- 
ledge of correlations may be of service in helping to arrive at a decision, 
but the ability of a plant to transmit its characters must be the final test. 
The closest study of the physical characters of a number of plants, with 
a view to securing uniformity in general conformation, will come far short 
of enabling the breeder to determine the projected efficiency of the indi- 
viduals under consideration. Such information can be obtained only by 
testing separately the progeny of each plant under conditions as nearly 
uniform as possible. 
For this purpose the centgener system was devised—a system which 
renders possible the testing of a large number of mother plants whereby 
a record of the performance of each plant is obtained. The plant is the 
unit of selection. The centgener is a trial plot of one hundred grains 
planted from the product of a single mother. 
A prime requisite in planting a foundation bed from which mothers are 
to be selected is to have all the plants grown under conditions as nearly the 
same as possible. To secure this equality of opportunity, so far as external 
conditions can be controlled, carefully chosen grains taken from hand- 
selected heads of the most promising varieties are planted in foundation 
beds four inches apart each way. The plants resulting from this seeding 
are pulled and studied separately in the field at harvest time and are 
subsequently subjected to a more severe test in the laboratory. So rigorous 
is this selection, based on the physical characters of the plants, that not 
more than one in five hundred is considered eligible for trial in the cent- 
geners. The same fundamental tests are applied to the centgeners as were 
applied the previous year to the individual plants in the foundation beds. 
This reduces the number eligible for registration to one in one thousand, and 
frequently this proportion cannot be had. Full notes are taken on the 
centgeners in the field, and the next season multiplying plots are sown 
broadcast from those centgeners which have the best performance record. 
In the fifth year the product of the multiplying plots enters into competition 
with the original variety. If, in this contest, it proves superior to the 
original stock it will be multiplied on the general farm as rapidly as 
possible and sold to the farmers at a reasonable price. 
In our work so far, numerous variations, some desirable and many 
undesirable, have been isolated from the 550,000 plants studied during the 
past three years. The progeny of many of these mothers has differed widely 
from the parent plant and consequently from each other; the progeny 
of each mother from the best plants has always been remarkably uniform 
within itself, and, with few exceptions, which we have not as yet proven 
to our satisfaction, has come true in the multiplying plots. In well- 
established varieties of oats of known breeding and purity, some strains 
have been isolated which ripened two weeks before others, and equally 
striking differences have been observed and perpetuated. This holds true 
under practically every heading under which oats are judged in the field 
or in the laboratory. Even among the most productive individuals the 
range in yield is surprising, the area covered in the third generation from 
one kernel of Joanette oats in one case being 127°78 square feet and in the 
other 0°89 square foot, one 144 times the other. The latter, in addition 
to being a very low-yielding strain, was also very poor in quality. 
