106 
Research Bulletin No. 2 
variety had an internode length almost exactly half way between 
that of the other parent variety and of the F. generation. One 
of the F 2 families of this cross had a range of variation extending 
from considerably below the minus extreme of the short-internode 
parent to much above the plus extreme of the long-internode 
parent and even considerably above the plus extreme of F x . This 
extreme range in F 2 is thought to be due to new combinations of 
the internode-length factors of the two parents. 
In two crosses, one parent of which produced numerous tillers 
and the other few tillers, F t was intermediate in number of stalks 
per plant. In both crosses F 2 was more variable than F, and in 
one cross the F 2 range was from one to eight stalks, while five 
was the largest number observed in the parent variety that 
tillered most freely. Among the F 3 families, a few were practi- 
cally one-stalked types and a few others had a somewhat larger 
mean number of stalks than the tillering parent and a range of 
variation so great as to suggest the possibility of isolating by 
selection a type with a still larger number of stalks. 
On account of the fact that in one cross the tall parent was 
a few-stalked type and the short parent a many-stalked type, the 
two parent varieties did not differ greatly in total length of stalks 
per plant. Owning to the intermediate number of stalks in ¥ 1 and 
to the tallness of F 1 stalks in this cross, the total length of stalks 
in F x was much greater than that of either parent — in fact almost 
equal to the combined stalk lengths of the parents. The mean 
total length of stalks of F 2 was much less than of F 1 but the 
variability was considerably greater. In another cross, where 
the parents differed much more in total length of stalk, no 
records of the F x generation are available, but the mean total 
stalk length in F 2 (which was doubtless less than in Fj was 
somewhat greater than that of either parent. Owing apparently 
to a lack of correlation — either physiological or genetic — between 
number of stalks and height of plant, some of the F 3 lots of this 
cross had a mean total length of stalk twice as great as that of 
the long-stalked parent and one F 3 family had a mean length 
greater than the plus extreme of that parent. This great stalk- 
length is thought to be due in the main to a combination of the 
factors influencing total stalk length, some of which come from 
one parent and some from the other. 
In earliness F 3 plants were intermediate between their 
parents. The F 2 generation more than filled in the gap between 
the parents in all cases where exact records were made and in 
one case had a range from below the mean of the early parent to 
above the mean of the late parent. Very distinct types were 
obtained in F 3 , some of which were practically as early in flower 
