GRAIN-SORGHUM EXPERIMENTS IN THE PANHANDLE OF TEXAS. 61 
between the two is in height. The original head row of No. 30 in 
1907 was heterozygous for stature, but of six head rows grown from 
it in 1908, two were uniformly dwarf and probably inherently so. 
~ In 1909, 10 head rows were grown from seed of row No. 340-1 of 
1908. All the rows were uniformly dwarf and early. To test the purity 
of this progeny in regard to dwarf stature, 10 heads were selected 
from the tallest of the dwarf plants and 10 from the lowest. The. 
average difference in the height of the two sets of parents was about 
6 or 8inches. In 1910 the seed from each lot of heads was used to 
sow 10 head rows. ‘There was absolutely no difference in the progeny 
of the two lots called, respectively, Nos. 340-1—Ad and 340-1-At. 
Fic. 10.—A plat of Dawn (Dwarf) kafir, C. I. No.340, at the Amarillo Cereal Field Station, August 17,1915, 
yield, 68.3 bushels per acre. 
For six generations since, or nine m all, this race has bred true for 
dwarf stature and earliness. A plat of Dawn kafir grown at 
Amarillo, Tex., m 1915, is shown in figure 10. 
The number of plats grown in each year in the 9-year period 
has varied from 1 to 3, the total number bemg 17. Selection has 
been practiced within this race in the hope of obtaining still lower 
stature or earlier maturity, but without special success. The results 
of the experiments conducted are shown in Tables XX and XXII, 
and the annual and average yields of this selection are compared 
with those of other kafirs in Table XXX. 
Besides being more dwarf than the ordmary Blackhull kafir, the 
Dawn kafir is also a little earlier, as will be seen in Table XXII. In 
1908, when all varieties developed normally, the difference was 
