510 
Journal of Agricultural Research 
Vol. XXXI, No. 6 
enough for drawing any definite conclusions, they are of interest 
when considered along with the others just given. 
Table VI. — Plate counts on different media, millions per gram of soil 
. Soil sample No. 
Aspar- 
aginate 
agar 
| 
Soil- 
extract 
gelatin 
Synthet¬ 
ic agar 
Mannite- 
nitrate 
agar 
Soil- 
extract 
agar 
1____ 
7.3 
7.0 
11.0 
10.0 
18.7 
2 ..__ 
3. 5 
3.3 
10.3 
8.1 
7.6 
10. 7 
14.4 
3___ 
4.0 
8.0 
4_ 
1.5 
5.0 
3.0 
9.0 
Averasre 
4.1 
5.1 
1 
8.6 
7.1 
13.2 
FLUCTUATIONS IN PLATE COUNTS OF SOIL ORGANISMS 
The fact that there is a seasonal rise and fall in the number of soil 
organisms has long been known. Daily or hourly fluctuations, how¬ 
ever, have been studied only recently. Cutler, Crump, and Sandon 
( 8 ) made 365 daily consecutive determinations of the bacteria and 
protozoa in a field soil which for a long time had received 14 tons of 
barnyard manure per acre annually. They observed great daily 
fluctuations in the numbers of bacteria, and noticed in many cases 
an inverse ratio between bacteria and protozoa. Soil moisture and 
temperature seemed to have no correlation with these fluctuations. 
Fortnightly averages showed the usual seasonal rise and fall and a 
marked parallelism in the numbers of bacteria and protozoa. 
More recently the work of Cutler et al. has been continued at 
Rothamsted, and similar fluctuations have been observed at intervals 
of two hours. (§) 
From the results obtained, it has been concluded that the usual 
method of making plate counts at intervals of one or several weeks 
does not give accurate information. Since, however, such wide 
fluctuations as recorded by the British investigators, as a rule, have 
not been recorded in bacteriological soil tests, and since, on the other 
hand, the fortnightly averages of those irregular counts agree closely 
with determinations made at longer intervals, it seemed advisable 
to make a series of comparative tests at frequent intervals for several 
weeks. 
THE EXPERIMENT PLOT AND METHODS OF SAMPLING 
The soil selected for sampling was a level piece of fertile lawn, 10 
feet square. The sod was removed and the earth spaded to a depth 
of about 6 inches. Small stakes were placed a foot apart around the 
edge, and strings were drawn across the plot in both directions, thus 
laying out 100 squares (fig. 1). In numbering them, the outside row 
around the entire plot, as well as the inside corner squares, were left 
as a border, from which no samples were taken. 
The test was started two weeks after spading. The vertical and 
the horizontal methods were used in taking samples. The former 
is the usual method of taking several borings for a mixed sample. 
In the horizontal method, about 3 inches of the surface soil were 
scraped off with a clean spade or trowel, and the sample taken from 
the underlying layer about 2 inches in thickness and extending over 
an area of about 10 square inches. By careful manipulation, both 
