INHERITANCE IN THE BARLEY SPIKE. 17 
FAMILY HANNA (460) X REID TRIUMPH (404). 
The parental forms, Hanna and Reid Triumph, are of distinctly 
different densities, and there is no overlapping of frequency distri- 
butions during the three years in which they have been grown. In 
Table II (sec. D) the mean of the Hanna parent ranges from 
4.12 ±0.02 mm. in 1916 to 4.56 ±0.01 mm. in 1918. The Reid 
Triumph variety has much less seasonal variation, the mean in 1917 
being 2.73 ±0.01 mm. and in 1916, 2.64 ±0.01 mm. It is of interest 
to note that the Reid Triumph has about the same average mean as 
the Svanhals 2-rowed form, while the Hanna is considerably more 
lax than the Manchuria form which was crossed with the Svanhals 
variety. 
The F 2 generation of the cross between Hanna and Reid Triumph 
proved more variable than the parents and frequently gave distri- 
bution from below the mode of the Reid Triumph to considerably 
above the mode of the Hanna parent. Twenty F 2 plants were grown 
in F 3 , some giving as variable a population as obtained in F 2 , while 
other F 3 lines were no more variable than the parental forms. 
Fourteen of these F 3 lines which gave the clearest indication of 
being homozygous were further tested in the F 4 generation. The 
method was similar to that previously used, 4 to 10 plants of a line 
being grown and the combined result being the basis of conclusions 
as to purity. Of the 14 lines tested in F 4 , 8 gave evidence in the com- 
bined F 3 and F 4 data to show that they are homozygous for density. 
Those which are of questionable purity will be briefly considered. 
Selection 406-3 gave a mean of about the same density as the Reid 
Triumph parent, but the coefficient of variability is somewhat higher 
than in the pure parental lines. Selection 406-4 proved to be 
heterozygous. One of the head selections, 406-4-3, produced a type 
which seems pure for density. The mean of this line is 3.72 ±0.03 
mm. Selection 406-9 seems to be heterozygous. Probably 406-9-1 
is homozygous, the average mean being about the same as that of 
the Hanna parent. Selection 406-10 also is more variable than the 
pure parental variety. The frequency distribution indicates that 
fewer density factors are involved than in the F 2 generation. Selec- 
tions 406-16 and 406-18 appear to be heterozygous. In later gen- 
erations two selections of 406-18 seem to be homozygous. Thus 
406-18-5 is probably breeding true with a mean density of 3.40 ± 0.02 
mm,, while 406-18-9 gives evidence of being homozygous for a 
mean of 2.66±0.02mm. 
Those which seem nearly homozygous by an examination of their 
frequency ranges and coefficients of variability as obtained in F 3 
and F 4 generations are as follows: 406-1, mean 2.81 ±0.01 mm.; 
406-5, mean 4.43 ±0.01 mm.; 406-7, mean 2.43 ±0.01 mm.; 406-8, 
mean 4.32±0.01 mm.; 406-11, mean 4.32±0.02 mm.; 406-12, mean, 
