New YoRK AGRICULTURAL EXPERIMENT STATION. 43 
conducted without suitable check plats by which to measure the 
effect. Accordingly we have tested this method of inoculation in a 
considerable number of cases, with an equal number of check plats 
for comparison. The results thus obtained are useful both as an 
indication of what can be expected from the use of soil and also 
as a basis for measuring the relative value of the various other 
methods which have been suggested for producing the inoculation. 
During 1905 we shipped soil to 54 farmers for the purpose of 
making these tests. The soil was obtained from one of our fields 
which had been in alfalfa for a long period and had produced heavy 
crops. Nodules were not easily found upon the alfalfa roots since 
the plants were old and were feeding at a considerable depth. 
The use of the soil from this field the preceding season at the 
rate of 100 lbs. per acre had produced good inoculation. 
Owing to the late cutting of the alfalfa in the fall of 1904 there 
was little aftergrowth and the unusually severe winter of 1904-5 
destroyed practically all of the plants. The results from the use 
of the soil show that the bacteria associated with the alfalfa were 
not destroyed. 
The soil was prepared by removing a little of the surface and 
passing the remaining soil, down to a depth of about a foot, 
through a coarse sieve, in order to remove the small stones and 
put the soil in condition to pass through the fertilizer attachment 
of a seeder. The sifted soil was shipped to the farmers in bags 
or barrels. Each was directed to apply it to a portion of the field 
just previous to sowing the seed, leaving another portion of the 
field untreated as a check. 
The results obtained from thirty-two of these experiments where 
the experiment was properly conducted and reported are given in 
Table II. 
