840 
Journal of Agricultural Research 
Vol. XXIV, No. 10 
The majority of variates in a population of Gradus types have a mean 
stipule ratio between 1.50 and 1.79 and none have been observed with a 
stipule ratio above 1.99. For Gradus rogue no plants have been found 
with a stipule ratio of less than 2.00 and by far the greatest number 
occur between the limits 2.10 and 2.49. The mean stipule range of 
Mummy extends over portions of the ranges of both Gradus and Gradus 
rogue, but the majority of individuals occur in classes 1.95 to 2.15, in¬ 
clusive. Consequently, if linkage existed between the factor for rogue 
stipules and the a factor for white flowers, one would expect proportion¬ 
ately more of the aa variates to lie close to the mean of the rogue parent, 
2.339, and a greater concentration of the variates with the A factor 
around the mean of Mummy, 2.088, than would be found close to the 
mean of Gradus rogue. The modes of the frequency distributions of 
both the colored (AA and Aa) and white (aa) segregates are actually at 
2.45, or very near the mean of the rogue parent, however. It is evident, 
therefore, that no linkage exists between the factor for rogue stipules 
and the a factor. 
The difference in means of the two populations is caused in a large 
measure by the linkage of the AY and ay factors, so that among the non¬ 
rogue segregates the AA and Aa variates tend to have a stipule ratio 
approaching that of Mummy and the mean of the aa segregates approaches 
that of the Gradus parent. The effect of linkage is to lower the mean 
of the aa plants and to increase the mean of the AA and Aa segregates. 
For this reason more aa segregates than AA and Aa segregates are found 
in classes 1.55 to 1.95, inclusive. Linkage of the two factors also accounts 
for the tendency to bimodality that both curves exhibit. (Fig. 3.) 
GRADUS ROGU^ X MUMMY, AND RKCIPROCAU IN TH^ Fg GENERATION; 
CLASSIFICATION OF Fg FAMILIES 
On the hypothesis that Gradus rogues are produced from Gradus by 
mutation of a single factor, x to X, which is allelomorphic to x' in Mummy, 
the expectation in regard to the Fg rogues of Gradus rogue X Mummy is 
that approximately 66 per cent will show segregation in the Fg generation 
into rogues and nonrogues, while the remaining 33 per cent should 
breed true. The Fg broads and intermediates (nonrogues) should not 
contain rogues among their progeny except as the product of primary 
mutation. 
No record was kept of the Fg plants saved for seed as to their general 
appearance, i. e., whether they were rogues, intermediates, or broads. 
Instead, stipule measurements were made and the stipule ratio was used 
for a criterion as to which of the three categories, rogue, intermediate, or 
broad, that the plant belonged. Altogether 184 plants were chosen 
from a number of Fg frequency classes (Tables XVI to XVIII). As 
expected, the parent plants of families containing no rogues came for the 
most part from classes 1.65 to 2.25. The parent plants of progenies 
breeding true to the rogue character or segregating into rogues and non¬ 
rogues were largely from classes 2.35 to 2.95. This and a classification of 
seregating Fg families of Gradus rogue X Mummy and reciprocal are 
shown in Table XVI. 
