498 
Journal of Agricultural Research 
Vol. XXVII, No. 7 
THE THIRD GENERATION OF THE CROSSES 
In each of the four F 2 progenies grown in 1922, flowers on a number 
of individuals were bagged to insure strict self-fertilization. The selfed 
individuals were selected for average grade of petal spot as follows: (a) 
Highest, (6), most nearly intermediate, (c) lowest. Twenty-four F 3 
progenies, two representing each condition of petal spot in each F 2 
progeny, were grown in 1923. Eight parental progenies were grown in 
the same plat, each of the original parents of the hybrids having been 
represented by two progenies. 
The frequency distributions for the several parental and F 3 progenies 
are given in Table VI, the unit having been the average of several flowers 
graded on each individual plant. The grading was begun on July 14 
and ended on August 18. 
It is clear from the frequency distributions in Table VI that in every 
case the behavior of the F 3 progeny accorded with the position of its 
parent in the F 2 frequency distribution. The eight F 2 parents having 
average grades higher than 7.5 proved to be pure dominants, the eight 
F 2 parents having average grades lower than 0.5 proved to be pure 
recessives, and the eight F 2 parents having average grades between 5 and 
6.5 proved to be heterozygous. The percentages of spotless plants and 
departures from the expected 25 per cent in the segregating F 3 progenies 
are stated in Table VII. 
Tabi,E VII .—Percentages of spotless individuals and departures from the expectation in 
the segregating F z progenies grown in 1923 
j i 
F3 progeny. 
! Number of 
plants. 
! 
Percentage 
spotless.« 
Departure 
from the 
expectation 
(25 per cent). 
(1-3-12-14X12) 20-27. 
39 
2 °- 5 
4 - 5 ^ 4-33 
(1-3-12-14X12) 20-35. 
33 
21.2 
3.8^4-82 
(1-3-12-14X12) 28-14. 
48 
25.0 
o± 4 - 2 i 
(1-3-12-14X12) 28-53. 
43 
39-5 
14.5i5.03 
(1—0—A—T y o.r— 7 . 
44 
45 
22.7 
17.8 
2 . 3 ± 4.25 
7-2^3.85 
VO * T’ 1 AiJ/ /... 
( 3 - 2 - 4 -lXl 3 ) 21-29.! 
( 3 - 2 - 4 - 1 X 13 ) 24-33 . 
45 
3 1 • 1 j 
6.1^4.65 
U-2-4-1X13) 24-35 ..| 
42 
33-3 
8.3 ±4 .90 
8 F 3 progenies as one array. 
339 
26.5 
i-5±i.6i 
a The probable error of the percentage is omitted, being the same as that of the departure from the 
expectation. 
Reference to Table VII shows that some of the F 3 progenies gave rather 
wide departures from the expected 25 per cent of spotless individuals. 
In no case, however, is the departure mathematically significant. If all 
of the segregating F 3 progenies are considered as one array, the departure 
from 25 per cent spotless is smaller than its probable error. As was the 
case in the F 2 progenies taken as one array (Table V), it is the spotless 
class which is slightly in excess of the expectation. 
X X X X X X X X X X X X 
G \ I \ i \ \ 4 
h « o w « « 
i I I I > f 
v 1* v 
« *> 
* \ 
ty 1 rM 
v V V v- 
i < V 7 7 ! 
1-5 l’** 
rv* r s 
K e> 
U 
