II6 Hayes. 
environment plays such a large part in their development. Fluctuating 
variability is always high, yet many careful investigations have proved 
that such fluctuations are not inherited. Such fluctuations greatly 
interfere with the analysis of the inheritance of size characters, as it 
is impossible except by a breeding test to determine whether a certain 
deviation from the mean is of a heritable nature or is simply due to 
evironmental conditions. Fluctuating variability as used here refers 
to the somatic variations which are non-inherited, and not to germinal 
variations which are inherited. 
The only way to determine the inheritance of quantitative 
characters is to compare the variability of the F, and F, generations 
and the parents, growing all under as nearly similar environmental 
conditions as possible. The variability of the F, generation should 
be no greater than that of the parents, and the F, generation should 
give a greatly increased variability. When sufficient numbers are 
studied, the F, generation should show a range of variation equal to 
the combined range of the parents and Fy. Certain Fy, forms should 
breed true in F3, giving no greater variability than the parent types; 
other F, forms should give decreased variability in F3 when compared 
with the F, generations; and others should show as great variability 
as the F, generation itself. 
In most of the above investigations no F, generations have been 
reported, and where reported they comprise only a few forms. In an 
earlier paper, data were given for the parental types, also for the 
F, and F, generations of tobacco crosses, and the inheritance of such 
characters as leaf number, size of leaf and height of plants was inter- 
preted as due to the interaction of several independently inherited 
factors, the heterozygous condition of each factor being half the homo- 
zygous. The purpose of this paper is to give further data on the 
inheritance of leaf number in tobacco, of which a number of Fs; 
generations have been grown, 
The Methods Used. 
The parents of the crosses reported here, with the exception of 
the Broadleaf, were strains which had been inbred for a number of 
years and which were therefore of uniform type. Unguarded fertilized 
seed of the Breadleaf plants used for crossing has proved very uni- 
form, so we are justified in concluding that this cross was also made 
between pure parents. 
