516 
Journal of Agricultural Research 
Vol. XXXI, No. 6 
From these rows, composite soil samples were taken, and plates 
were poured on soil-extract agar from one series of dilutions. The 
counts showed the usual fluctuations from plus 14 per cent to minus 
19 per cent, but no relation to the percentage variation of the dry 
matter of the crop. If each component of the composite samples had 
been plated and the average calculated on such a basis, undoubtedly 
the variation of the counts would have been much less, but the work 
involved made such a procedure impracticable. 
SUMMARY 
The suspension of the bacteria in a soil sample may be effected by 
means of an electric vibrator of high frequency. 
Plate counts made from such a suspension and from a suspension 
made by hand shaking gave concordant results. 
Higher numbers obtained by the vibrator method would seem to 
indicate either incomplete sterilization of the vibrator or incomplete 
dispersion by hand shaking. 
The greater difficulty of manipulating the vibrator precludes its use 
as a routine procedure. 
Plates poured with soil-extract agar gave higher and more uniform 
results than those poured with other media. Egg-albumen agar gave 
lower counts, but was in general comparable to soil-extract agar. 
Mannite-salts agar proved to be too selective, and gave, accordingly, 
too low counts. Spreading colonies were equally rare on all these 
media. 
Asparaginate agar, synthetic agar, soil-extract gelatin, and other 
media were all inferior to soil-extract agar. 
Plate counts on soil-extract agars made from type soils ranging 
from loamy sand to muck showed that field soils of fair fertility, 
irrespective of their general character and location, are suitable for 
making soil-extract agars to be used in estimating the total number 
of soil microorganisms. 
Plate counts of soil samples taken horizontally from a level ap¬ 
proximately 4 inches below the surface are slightly higher and more 
uniform than those of samples taken vertically with an auger. 
Fluctuations of 20 per cent above or below the average total count 
may occur with a uniform soil, if one series of dilutions is made and 
soil-extract agar used as plating medium. A more selective medium, 
such as mannite-salts agar, gave much wider fluctuations. 
Duplicate series of dilutions of a composite soil sample showed 
similar fluctuations. Therefore, to be of any value, a total count 
should represent the average of three or more separate counts. 
LITERATURE CITED 
(1) Conn, H. J. 
1916. CULTURE MEDIA FOR USE IN THE PLATE METHOD OF COUNTING SOIL 
bacteria. Centbl. Bakt. (II) 44: 719-734. 
( 2 ) - 
1921. AN ABUNDANT BUT LITTLE KNOWN GROUP OF SOIL BACTERIA. (Au¬ 
thor’s abstract) Abs. Bact. 5: 8-9. 
(3) Cutler, D. W., Crump, L. M., and Sandon, H. 
1923. A QUANTITATIVE INVESTIGATION OF THE BACTERIAL AND PROTOZOAN 
POPULATION OF THE SOIL, WITH AN ACCOUNT OF THE PROTOZOAN 
fauna. Phil. Trans. Roy. Soc. London (B) 211: 317-350, illus. 
