460 
Journal of Agricultural Research 
Vol. XXIII, No. 6 
per cent for the unselected check. Other tests of selections from Turkey 
and Red Russian add proof to the conclusion that the bunt-free plants in 
varieties planted under conditions favoring maximum infection are not 
resistant mutants but rather escape infection by accident. 
BREEDING FOR BUNT RESISTANCE 
methods 
In testing different wheats for comparative resistance to bunt, condi¬ 
tions favoring maximum infection were, as nearly as possible, maintained. 
The seed was inoculated with fresh viable spores just before planting. 
This was accomplished by stirring into each packet of seeds from 1 to 5 
per cent of its weight in bunt spores, so that each seed was literally 
blackened by spores clinging to it, besides a surplus of loose spores among 
the seeds. The inoculum was obtained from the smutted heads of many 
different varieties of wheat in the field, in order to get a representative 
sample of the organism in its native environment. The heads were 
ground up and the spores sifted out and kept in a cool, dry place during 
the planting season. The packets of inoculated seeds were planted by 
hand in rows 18 inches apart. The seeds were spaced 4 to 6 inches apart 
in the row to avoid confusion in separating the plants at harvest time. 
Each experiment, whether a variety test or an F 3 family, was planted at 
approximately the same date to avoid error due to changing seasonal 
conditions. As many as 10 men were required at times in order to plant 
a given experiment in a single day. 
The field was kept free of weeds and volunteer grain by rotation and 
cultivation, com and field peas preceding the cereal nursery in a 3-year 
rotation in every case. The rainless summers of eastern Washington 
make clean cultivation under these conditions comparatively easy. 
A stake bearing a printed label with date, name of experiment, name of 
variety and pedigree number was placed at the end of each row. The 
records at harvest time were obtained as follows: To get a quantitative 
measure of resistance, the plants of each row were pulled and separated 
into three piles (bunt-free, all bunt, and part bunt); and the number of 
plants in each was recorded in a field note book. The partly bunted 
plants were then divided into heads of wheat and heads of bunt, and the 
numbers were entered in their proper columns. With these five figures 
it is easy to reduce the amount of bunt in a given row to terms of a single 
number for direct comparison with the amount produced by any other 
row. The computation used gives the total bunt in terms of percentage 
of the whole row, according to the formula ab + c = d, in which a is the 
percentage of bunted heads on the partly bunted plants, b is the percent¬ 
age of partly bunted plants in the row, c is the percentage of entirely 
bunted plants in the row, and d is the total percentage of the row 
that is bunted. The computations were made with a calculating 
machine and checked with a slide rule. The results thus obtained are 
not materially different from those obtainable by a straight head count 
of all the heads in the row. This method takes much less time and gives 
information that a straight head count would not give—that is, the 
quantity of partly bunted plants and the amount of wheat produced on 
them, which in itself is an important measure of resistance, and is useful 
in checking errors in counting or computations. 
The work was all carried on out of doors, in the field, the main care 
being to have uniform conditions for all the rows of a given test, so that 
