1/10 
Research Bulletin No. 2 
;iud five for Dumber of stalks, as was assumed, over a milliou F 2 
plants, or in other words over one hundred acres, would be re- 
quired for an even chance of getting the desired combination of 
characters, and at least two or three times that number should 
be grown to make sure of results. Now, while perhaps breeding 
work on such a scale is not impossible, it is certainly out of the 
question under ordinary conditions. What method then can be 
used in such cases? Or are the proposed results impossible of 
accomplishment? 
Under the conditions assumed here, the method of back cross- 
ing with one parent type (Castle 1911), cannot be used, for, while 
insuring the ultimate recovery of say the many-stalked parent 
type, it would at the same time make impossible the recovery of 
the tall-stalked type of the other parent. The only method avail 
able in such cases is to grow as many F 2 plants as is practicable, 
select such plants as combine most nearly the desired combina- 
tion of height and number of stalks, and from selfed seed of 
these grow F 3 progenies, from which in turn further selections 
must be made, and so on for as many generations as are necessary. 
In this way, results can be secured with a much smaller total 
number of plants but of course at the expense of considerably 
more time. 
It may not be out of place here to show more definitely just 
how the element of time can be made to take the place of numbers 
in solving a breeding problem like that discussed above. For 
sake of simplicity, let us now assume that the parents differ in 
only four factors for height and four for number of stalks, eight 
in all, thus requiring something over 65,000 F 2 plants for an even 
chance of getting the desired combination of many stalks and 
tall stalks. F>y referring back to Table 1 (page 10), we can sec 
at a glance what behavior will result both in F 2 and in F 3 follow- 
ing a cross the parents of which differ in four factors A, />, (\ D. 
If the class headings are read as decimeters, they will represent: 
well actual heights of short and tall corn plants. Each single 
factor will then be assumed to add two decimeters to the initial 
height of 10 decimeters. If now we call the factors A', B f , C, D' 
and suppose each to add one stalk to an initial 1 -stalked con- 
dition, the same table will serve to illustrate the inheritance of 
number of stalks per plant where the parents are 1-stalked and 
9-stalked respectively. Now from the cross of a 1-stalked, 
26-decimeter type with a 9-stalked, 10-decimeter type, we see from 
Table 1 that, out of 256 F 2 individuals, one should have the 
formula AABBGCDB and also that one should have the formula 
A'A'B'B'C'C'D'D'. In other words, there should be one 26- 
decimeter plant and also one 9-stalked plant. But the 26-deci- 
