PRODUCTIVENESS OF SELF-FERTILIZED CORN 
owing probably in part to greater variability and in part to the 
fewer lines in these earlier generations. The 12-1- strain is the only 
one having parallel lines in three successive generations. The 12—1—1— 
lines were more productive than those of 12-1-2- in each generation . 
Four comparisons between parallel pairs of lines are possible in the 
fifth and sixth generations, in each of which the more productive line 
of the pair in the sixth generation also was the more productive in 
the fifth. The average yields of the more productive and of the less 
productive lines of these pairs, together with the average yields of their 
antecedent lines in the third and fourth generations, are shown as 
averages C and D in Table 3 and are shown graphically in Figure 1. 
The difference in the fourth generation is due entirely to the effect 
of the difference between 12-1-1-S and 12-1-2-S, so that little impor- 
tance attaches to the 
spread in that gener- 
ation. For compari- 
son with these data 
the theoretical curve 
for decreased yield in 
six successive gener- 
ations of inbreeding 
without selection also 
is shown in Figure 1. 
This curve is plotted 
to decrease from 50 
bushels per acre, an 
approximation of the 
yield of the open-fertil- 
ized noninbred check 
seed in the present 
experiment, to 17.4 
bushels per acre in 
the third self -fertilized 
generation, the aver- 
SO 
% 
X 
\ 
N 20 
SO 
/o 
o 
0.£ 
\ 
\ 
\ 
\ 
\ 
\ 
\ 
w 
v-7-/i£0& 
\ 
\ 
\ 
\ 
\ 
\ 
\ 
\ 
3£rr£se l/a/£s 
N 
T 
u 
^—»~-*~~ 
"~~~~~* s **-. 
v,^ 
<~ 
^zr 
POOA 
?£/e l/ 
ty£S 
/ Z S *■ J" 
Fig. 1.— Diagram showing the average yields of four better yielding 
self-fertilized lines of corn (average C of Table 3) and of four poorer 
yielding lines (average D of Table 3) after three to six generations 
of self-fertilization and the theoretical decrease in yield in unselected 
self-fertilized lines of corn 
age yield of the four 
selfed lines in that 
generation. This as- 
sumes that the aver- 
age yield of these four 
lines in the third gener- 
ation had not been affected by selection, which may or may not be 
true, but will serve for purposes of approximation. 
It is evident that two lines have been isolated in each of the four 
strains after three or four generations of self-fertilization, one of 
which is more and the other less productive. Moreover, the average 
of the four less productive lines decreased about 20 per cent between 
the fifth and sixth generations, whereas that of the four more produc- 
tive lines remained essentially constant. 
The averages of the strain means in the different generations are 
shown as average A of Table 3. There is a slight, reasonably con- 
sistent decrease from the third to the sixth generation. The difference 
between the yields in these generations is 1.8 bushels per acre. The 
decrease in the theoretical curve of inbreeding (shown in fig. 1) be- 
tween the third and sixth generations is from 17.4 to 13.3 bushels, or 
