1026 
Journal of Agricultural Research 
Vol. XXVIII, No. 10 
Table Y. — Correlation between various characters of 52 F 2 segregates of Amby X 
Kubanka and of 152 F 2 segregates of Marquis X Alaska 
Characters correlated 
Amby X 
Kubanka 
Marquis X 
Alaska 
Type of culm and— 
Type of keel__ 
Correlations 
0.86±0.03 
Correlations 
0. 24±0.05 
Shape of spikelet_ 
.90± .03 
. 23± . 05 
. 27± . 05 
. 48± . 05 
. 39± .05 
Awn length__ _ _. 
. 80± .04 
Grain texture_ _ 
.68± .07 
Type of keel and— 
Shape of spikelet_ _ _ ... . .. 
. 84± .03 
Awn length_ 
. 74± . 04 
. 30± .05 
Compactness of head_ 
. 24db . 09 
. 26± .05 
Grain texture_ ............. ... 
. 80± .03 
. 17± . 05 
Grain shape_ _ 
. 90dr .02 
. 12± . 05 
Shape of spikelet and— 
Grain shape_ _ . _ _ _ __ 
. 89± .02 
. 07 ± .05 
Grain texture__*.. 
• 69± .05 
. 28± . 05 
Length of awns and— 
Grain shape_ 
. 36± . 08 
. 23± .05 
Grain texture__ 
. 49± . 07 
— . 13± . 05 
Grains per spikelet and— 
Type of culm__ 
. 40± . 08 
. 11± . 05 
Type of keel____ _ . 
. 53± .07 
. 32± . 05 
. 03± . 05 
Homozygous factors_ __ 
. 42± .08 
Do _ ________ 
—. 02± . 09 
. 03± .05 
Heterozygous factors______ 
-. 19± . 09 
-. 08± . 05 
Correlations between various characters of a comparatively sterile cross, 
Amby X Kubanka, and a comparatively fertile cross, Marquis X Alaska, are 
shown in Table V. The various characters, type of culm, type of keel, shape of 
spikelet, and in most cases awn length and grain texture, are all characters which 
differ greatly in the emmer and vulgare groups and these characters in each 
group are more or less characteristic and are not found in varieties within the 
other group. In the cross Amby X Kubanka the correlation between the various 
characters is rather high in all cases, indicating considerable association of the 
parental characters. In the more fertile cross, Marquis X Alaska, the correla¬ 
tions are in most cases relatively low or are of little or no significance. In the 
correlation between type of culm and grain texture the correlation is relatively 
high. Although Marquis has a harder grain than Alaska when grown in eastern 
Washington, the durum and emmer segregates usually had flinty kernels harder 
than those of Marquis. This is due to the fact that Alaska has hard grain if 
the factors for “yellowberry ” are absent. These correlations show clearly that 
in the more sterile hybrids the parental characters are more or less closely asso¬ 
ciated, while there is less association as fertility increases. It is possible, of 
course, that ultimately the more fertile hybrid would result in segregates resem¬ 
bling one parent or the other, and that the elimination of intermediate types 
occurs more slowly in the more fertile combinations. This is indicated by the 
degrees of sterility found in various types of segregates, as shown in Table IV. 
The abnormal types which do not resemble any of the economic species are the 
most sterile while segregates resembling the parent species are the most fertile. 
One might expect the segregates resembling the parent to be more fertile than 
the intermediate types. This is found to be the case in the cross Amby X Ku¬ 
banka, as indicated by the correlation between grains per spikelet and type of 
culm. The culm type and keel type resembling these parental forms are asso¬ 
ciated with comparatively fertile individuals, while the intermediate forms are 
associated with relatively sterile individuals in F 2 . In the cross Marquis X 
Alaska there was, however, little correlation between fertility and type of segre¬ 
gate. An attempt was also made to classify the various individuals according 
to the number of homozygous and heterozygous factors for the various characters 
involved. There is some indication in the first cross that with the increase in 
the number of homozygous factors the number of grains per spikelet is increased, 
