54 
Journal oj Agricultural Research 
Vol. XXX, No. 1 
LEMMA COLOR AND SPIKELET DISARTICULATION 
The summarized data on correla¬ 
tion between spikelet disarticulation 
and lemma color, as presented in 
Table XIV, indicate that some 
correlation exists between these 
characters. The coefficient of associa¬ 
tion was determined in this case by 
combining the black and brown color 
classes as dark and the red, yellow, 
and white as light. Two disarticula¬ 
tion classes were combined as before, 
that is, abscission and semiabscission 
were added together in contrast to 
the fracture class. The data thus 
arranged appear as follows: 
Lemma color 
1 
Disarticulation 
Abscis¬ 
sion and 
semiab¬ 
scission 
Fracture 
Dark..... 
3 ,258 
1,685 
2, 792 
4,347 
Light.... 
Q=0.501 ±0.005 
The data indicate that most of the 
kernels, withbut respect to color, dis¬ 
articulated by fracture. As a rule in 
those strains in which a high percent¬ 
age of kernels of one color disarticulated 
by abscission, high percentages of the 
kernels of all colors disarticulated in 
the same manner. 
There is association between lemma 
color and spikelet ' disarticulation or 
basal form. Dark kernels most often 
have the prominent basal cavity re¬ 
sulting from disarticulation by ab¬ 
scission, while yellow or w T hite kernels 
seldom are found in strains disarticulat¬ 
ing in this way. 
The data indicate that generally the 
yellow and white kernels disarticulate 
by fracture. This agrees w T ith the ob¬ 
servations of Love and Craig {71) and 
others. The darker-colored kernels are 
found associated more often with the 
spikelet disarticulation by abscission 
while light-colored kernels are more 
often associated with spikelet disarticu¬ 
lation by fracture. The association 
between dark color and abscission 
apparently is not as strong as that 
between light color and fracture. 
LEMMA COLOR AND AWNS 
The data on the correlation between 
lemma color and awns are presented in 
Table XV. As in the cases previously 
described, the classes for lemma color 
and awns were grouped so as to make 
only two classes of each, as follows: 
Awns 
Lemma color 
Present 
Absent 
Dark.... 
4,104 
1,946 
Light__ 
2,837 
■1 
3,195 
Q= 0.407 ±0.005 
The results obtained in this study 
indicate that the twisted awn must be 
different in genetic constitution from 
the other classes of awns. In the study 
of correlation between awns and lemma, 
color it is found that practically the 
same percentages of twisted awns ap¬ 
peared in each color class with the 
exception of white. 
If the data for the long and short 
nontwisted awns are combined as non- 
twisted, the following figures are ob¬ 
tained : 
Lemma color 
Character of awn 
Twisted 
I 
Non- i 
twisted j 
[ 
Absent 
Percent 
Per cent J 
Per cent 
Black... 
\ 7.9 
64.6 1 
27. 5 
Brown__ 
9.4 
58.3 i 
32.3 
Red... 
8.0 
44.8 
47.2 
Yellow__ 
10.8 
21.4 
67.8 
White... 
! 0 
6.2 i 
93. 8 
The figures indicate a consistent and 
significant decrease in the nontwisted 
awn class in passing from the darker 
to the lighter colors. 
The data for each of the strains are 
in general agreement with the conclu¬ 
sions drawn from the summary data, 
although in some strains the relation 
between lemma color and awns is less 
definite than in others. 
The presence or absence of awns 
may be rather uniformly either low or 
high for all colors in a given strain. 
The variations or irregularities which 
occur in these strains probably may be 
accounted for by the fact that only a 
few individuals were studied. 
In general, it may be stated that 
some correlation exists between color 
of lemma and the type of awn in the 
Burt oat used in these experiments. 
The lighter-colored kernels as a rule 
have fewer nontwisted awns than the 
darker kernels. 
The three awn classes—twisted, non¬ 
twisted, and absent—are clearly differ¬ 
ent in breeding behavior as indicated 
in the studies of independent inherit¬ 
ance and as shown by their associations 
with other spikelet characters. 
