934 
Journal of Agricultural Research 
Vol. V, No. 20 
The bacterial growth from one agar slope was suspended in 12 c. c. of 
each of the above solutions, and 1 c. c. was diluted and plated quantita¬ 
tively on nitrogen-free ash agar. From each of the seven suspensions 
2 c. c. was added to each of five flasks of quartz sand, which was of the 
same quality and prepared exactly as in the preceding trial. 
These flasks were kept in a dark place at 22 0 to 25 0 C. Quantitative 
determinations, made at intervals, are based on plates from but a single 
sample of each set, consequently the opportunity for error is materially 
increased. It can not, therefore, be claimed that these figures (Table II) 
show accurate comparisons. However, it is quite evident that between 
March 26 and April 17, during which time the sand was so dry as to 
make the multiplication of organisms impossible, the rate of decrease 
in the numbers of organisms taken from broth, milk, and soil solution 
was noticeably less than that of organisms from the other solutions. 
This implies a certain protection gained from the presence of nitrogenous 
or albuminous constituents in the milk or broth. To what substance or 
substances in the soil solution such protection should be credited can 
not be stated definitely. The slime, mentioned by Van Suchtelen (10, 
p. 154), may be of influence in this connection. 
Table II .—Longevity of Pseudomonas fadicicola , dried in sand after suspension in 
different solutions 
Date. 
Salt solu¬ 
tion. 
Agar solu¬ 
tion. 
Gelatin 
solution. 
Gum-arabic 
solution. 
Broth. 
Milk. 
Soil solu¬ 
tion. 
March 18. . 
26. . 
April 6.... 
17.... 
1, 100, OOO 
—10, OOO 
“ 2 5 
-25 
1, 500, OOO 
— 10, OOO 
25 
-25 
1,440, 000 
10,125, 000 
50 
-25 
1, 613, 000 
— 10, 000 
-25 
-25 
1,024, OOO 
19, 967, 000 
220, 000 
-25 
1, 176, 800 
185, 000 
405, 000 
“25 
1,460,000 
40,000 
8, 600 
-25 
An additional experiment was conducted employing the same cultures 
used in the previous experiments. The procedure was the same, except 
that as a basis for quantitative determinations two samples were taken 
from each set instead of one. As the plates from several of the flasks 
showed no colonies whatever on May 3, even in the lowest dilutions, 
which represented 1/25 gm., it was thought advisable in making the next 
determinations, on May 13, to take i-gm. samples from these flasks and 
mix them directly with the melted medium in the Petri dish instead of 
plating 1 c. c. of a dilute suspension as previously done. It is quite evi¬ 
dent that the direct mixture of the sand with the plating medium tends 
to give higher counts than those secured by plating the washings of the 
sand, for in the latter case a large number of organisms undoubtedly 
remain attached to the sand particles instead of being washed off into 
the suspension. This difference in technic may account for the apparent 
increase in numbers in certain cases, as shown by the last plating. 
