The Inheritance of Quantitative Characters in Maize 117 
many crosses between varieties having similar quantitative char- 
acters will be made with the avowed object of securing combina- 
tions of factors unlike the combinations responsible for the par- 
ent characters, but the idea is well worth keeping in mind. 
When the factors that have to do with the development of the 
principal crop plants have all been sorted out and tagged — as 
the} 7 will be some day — breeders can proceed in their undertak- 
ings with much more assurance than at present. 
By way of summary it can be said that a breeder dealing 
with quantitative characters should proceed just as he would if 
dealing with such qualitative characters as color where the 
parent types differ in respect to numerous characters. The prin- 
cipal differences between the two cases are that quantitative 
crosses are likely to differ in more factors than qualitative ones 
and that environmental influences cause more confusion. He 
should in short grow as many F 2 plants as his facilities allow. 
From these he should select numerous promising individuals, 
self pollinate them and grow their progenies on as large a scale 
as practicable. If any of these F 3 lots are of the desired type he 
will of course propagate them and discard the others, or if not 
he can at least discard some of them and continue his selection 
from the most promising ones. Or if no F 3 family shows by its 
range of variation any tendency to produce ultimately a type 
better than itself, resort should be had to intercrossing between 
the several more promising F 3 lots or between their F 4 prog- 
enies, with the hope (we might almost say assurance) that 
among these several types, tho they are similar in appearance, 
there exist sufficiently different sets of factors to insure the 
desired combination in crossing. When dealing with plants like 
corn that are much more vigorous when many characters are 
heterozygous, the breeder will find it necessary to practice cross 
breeding between different isolated types or to provide for 
natural crossing. He should not, therefore, be content with the 
production of a single strain of the desired type — say a single 
high-oil type — but should isolate two or more such types, both of 
which are high in oil content, but which differ by enough minor 
characters, quantitative or qualitative, to insure a vigorous de- 
velopment on crossing them or on allowing them to intercross 
naturally when grown in mixtures. 
These methods apply just as well to improvement by selection 
from tha complex hybrid mixtures, which, because of some out- 
standing features, have been assigned varietal names, as they do 
to the isolation of types from the similar, tho perhaps somewhat 
more complex, hybrid mixtures resulting from the crossing of 
two named varieties. The difference at most is only a matter 
