936 
Journal of Agricultural Research 
Vol. V, No. 20 
It is evident from the data above tabulated that a larger number of 
organisms survive a limited period of desiccation in clay loam than in 
quartz sand. This may be partly explained by the difference in grain 
size and hygroscopic moisture of the two. A given weight of coarse 
quartz sand consisting of large particles has a surface much less than 
that of the same quantity of finely divided garden soil, and it therefore 
retains a much smaller amount of moisture in the hygroscopic form. If 
the grain size were the only distinction between sand and clay-loam soil, 
it might properly be concluded that the longevity of organisms in such 
materials is directly proportional to the percentage of hygroscopic water 
retained. Such a conclusion is not permissible, however, for the clay- 
loam soil differs from the, sand not only in texture but in content of 
organic constituents. The amount of such material in any sand is small, 
and in this case, where the sand was treated with acid, it may be regarded 
as having been absent. The experiments already described indicate that 
the soil solution contains substances which offer to the bacteria some 
protection against desiccation. The soil solution used in our experiments 
was extracted from just such a soil as was used in the experiment now 
under discussion. 
A further experiment was conducted to compare the changes in num¬ 
bers and kinds of organisms when soil solution is dried in different types 
of soils. Soil solution extracted from a rich garden loam was used for 
this experiment. The soils, obtained from the Soil Physics Department 
of the Michigan Agricultural College, were of five different types: Muck, 
sand, sandy loam, clay, and clay loam. 
Fifty-gm. portions of these soils in the air-dry condition were placed 
in ioo c. c. Erlenmeyer flasks plugged with cotton and were then sterilized 
in the autoclave for 45 minutes under 15 pounds pressure. For greater 
exactness the total quantity of soil solution was agitated and then divided 
into five 250 c. c. portions; from each of these 1 c. c. was plated on ordi¬ 
nary agar in dilutions of 1 to 10,000, 1 to 100,000 and 1 to 1,000,000. 
Ten flasks of each type of soil were then inoculated with the soil solution, 
all the solution used for any one type of soil being taken from a single 
flask. Although it was desired to have the inoculum approximately 
equal in all cases, a quantity of liquid which barely moistened the muck 
and clay loam was found to more than saturate the coarser soils. So, 
to make the physical conditions more nearly alike, 15 c. c. of the solution 
was used for each flask of clay, clay loam, and muck, but only 10 c. c. 
for the flasks of sand and sandy loam. 
The inoculated flasks were kept on a shelf in the laboratory at a tem¬ 
perature of 20 0 to 25 0 C., exposed to very dim diffused light, and subject 
to the influence of normal variatiQns in the humidity of the room atmos¬ 
phere. At intervals of about four weeks quantitative determinations 
were made, samples being taken from two flasks of each soil. After the 
first plating, samples were taken from one flask opened at the previous 
