18 BULLETIN 869, U. S. DEPARTMENT OE AGRICULTURE. 
2.84 ±0.02 mm.; 406-19, mean 3.29 ±0.01 mm.; 406-22, mean 
4.37 ±0.01 mm. 
Aside from these, individual heads grown in F 4 which appear to 
give homozygous progeny as a result of the single season's test are 
as follows: 406-4-3, mean 3.72 ±0.03 mm. ; 406-9-1, mean 4.30 ±0.04 
mm.; 406-18-5, mean 3.40±0.02 mm.; 406-18-9, mean 2.66±0.02 
mm. 
The means for these four F 4 families are somewhat unreliable 
because of the small number of individuals grown. All coefficients 
of variability, however, are very small. 
These results show that homozygous intermediates may be pro- 
duced, as well as homozygous types, which give about the same aver- 
age density as the parental forms. No analysis of average differ- 
ences as small as 0.2 to 0.3 mm. has been attempted. The fact that 
environmental or other seasonal characters may modify the expres- 
sion of a character nullifies such close analysis. 
FAMILY HAKNA (460) X ZEOCRITON (1039). 
The Hanna used in the cross with Zeocriton is the same pure line 
that was used in the cross with Reid Triumph. Zeocriton is a very 
dense 2-rowed form. This cross is between the most dense and the 
most lax form used in this study. 
The F 2 generation shown in Table II (sec. E) ranged from above 
the modal class of Hanna to the modal class of Zeocriton, even though 
only 141 individuals were studied. It has a correspondingly high 
coefficient of variability. 
An examination of the coefficients obtained in later generations 
show that some are as large as those obtained in the F 2 line. Others 
are intermediate, being significantly larger than any obtained in the 
pure forms, while still others are as small as those obtained for the 
pure parental lines. This would in dicate that the mode of inheritance 
was more complex than in the cross between Pyramidatum X Jet 
previously mentioned. 
Selection 448-9, which was almost as variable in the F 3 as in the 
F 2 generation, was selected for further experiment, the progeny of 
30 plants being measured in the F 4 generation. Data from 7 of the 
30 progeny lines are presented, as the remaining 23 all appeared to 
be segregating. Results of density studies in F 4 lines 448-9-7, 
448-9-14, 448-9-16, and 448-9-29 are given, as these indicate the 
segregation obtained in the unpresented lines. No F 4 line of greater 
coefficient of variability than 448-9-7 was obtained, and none with a 
wider frequency range than 448-9-16. Three lines appear to be 
homozygous, as determined by the frequency distribution and coeffi- 
cient of variabilitv. These are shown in Table III. 
