24 BULLETIN 869, U. S. DEPARTMENT OE AGBICHLTUBE. 
tions may be associated with the same variability which manifests 
itself in seasonal fluctuations. They may be due to the differences 
in the progressive density from the base to the tip of the rachis. which 
is more marked in some than in other varieties. Other explanations 
might be suggested, but hi the absence of definite proof it seems 
unwise to attempt a more detailed analysis of the results. 
CONCLUSIONS. 
Despite the handicaps of the investigations, a number of points are 
established. 
■ 1 Internode length in the barley rachis is a very stable character, 
which is much less affected by environmental conditions than many 
size characters. 
(2) Segregation occurs hi the F 2 generation of crosses, and forms 
homozygous for density appear in this generation, their purity being 
demonstrated in the F 3 generation. 
(3) In some crosses new lines with densities differing much from 
those of then parents can not be secured, while in others lines with 
very different densities may be isolated. 
(4) The inheritance of internode lengths may be interpreted on the 
factor hypothesis. Some of the crosses studied appeared to differ 
by a single main factor of density, while hi others two or three main 
factors are necessary to explahi the genetic results. Minor factors 
were evident whose number or nature was not established and through 
whose action the means of homozygous forms of intermediate 
densities hi some crosses may become more or less continuous between 
the means of the parents. 
