18 
BULLETIN 1209, U. S. DEPARTMENT OF AGRICULTURE. 
both the upper and lower extremities of the bars in Figure 2 follow the 
line connecting the mean-yield points in a general way. There is an 
evident tendency, therefore, for the crosses having one parent in com- 
mon to read similarly. At the same time there are marked devia- 
tions from this tendency. The most striking exception is in the ease 
of the crosses involving the 13-1- strain. The reciprocal crosses be- 
tween the 10-3- and 13-1- strains yielded 74.06 and 75.14 bushels 
per acre, or 15 bushels more than the next best cross in which strain 
13-1- was used. The average of the 13-1- crosses is unduly high 
because of this, and in fact would be about 55 bushels if its crosses 
with 10-3- were not considered. This would be less than the cm->es 
of any of t he ol her strains except 7-1-. Undue influence in the oppo- 
site direction is evident in the crosses involving the 9-2- family. 
74 

-WMArLfy X SrCHA/fLES M//T£ _ 
72 
— ~ 
_ 
70 
— 
— 
68 
— ' 
NQ20/,/±,F 5 ,Fe ~ 
Ju 66 
— 
— 
£ 
— \ 
_ ,_ 
^ 64 
— 
_^ 
— 
cs 62 
5: 6o 
— 
_ 
!% 
- 
- 
- 
— 
— 
_^«v^ 
^ S3 
NJ °° 
$ 56 
ZO-3- 
= 
■s 
I 
— 
— 
- 
\ 
— 
5 
\ 
— 
\s* 
5 
-/- 
3-/- 
\ 
~ — 
52 
- 
_ 
^ 
/?-/- 
50 
43 
z © 
2-2- 
7- 
■4 
/3 
-/- /5-/~ 
46 
9-2- 7-/- 
Fi<;. 2.— Diagram showing the relative yields of Whatley X St. Charles White corn, of later generations 
ol No. 201, and of 41 crosses between self-fertilized hnes of No. 201 grouped to show the productive- 
ness by families. 
Here a single extremely poor cross with 7-1- yielding 8 bushels less 
than the next better cross, has lowered the mean yield nearly 2 
bushels per acre. 
On the whole, however, the tendency for certain strains to produce 
satisfactory crosses is promising. Certainly the results seem to offer 
a more logical basis for future selection and experiment than is pro- 
vided by the relative vigor of the selfed lines. This is demonstrated 
particularly well in the case o\' the 10-3- strain. Indeed, the most 
striking result in the experiment is the high yield of crosses having 
lo 3 asaparent. The mean yield of these seven crosses exceeds the 
productiveness of any single one of the 34 crosses not involving 10-3-. 
In spite of this, the 10-3- selfed lines are so thoroughly unpromising 
that they were marked for elimination during the season of 1921 
before the results of the crossing experiments were at hand. This 
emphasizes the importance of selection based in part at least on the 
performance of the crosses. 
