8 BULLETIN 1122, U. S. DEPARTMENT OF AGRICULTUEE. 
persing some of the colloidal material, reliable data on the non- 
colloidal absorption in a soil cannot well be obtained from an investi- 
gation of soil fractions. It was necessary, therefore, to attack the 
problem in another manner. 
If the noncoUoidal part of the soil is made up of various mineral 
particles above the arbitrary limit of 1 micron in diameter, a deter- 
mination of the absorptive capacities of soil minerals in the proper 
state of subdivision should afford data for calculating the absorption 
of the noncoUoidal soil constituents. Data obtained in this way 
should be more reliable than that yielded by the preceding method 
of investigation, as mineral powders can be prepared which are 
essentially free from colloidal material. 
Some 21 minerals are of common occurrence in soils. Although 
only a few of these are present in a soil in any considerable quantity, 
it was thought advisable to determine the absorptive capacity of 
each of them. Samples of the 21 minerals were kindly provided by 
the National Museum. They were characteristic mineral specimens, 
and were apparently free from decomposition products, although 
not absolutely pure, a few containing traces of calcite. 
The work of previous investigators indicates that the absorptive 
capacity of any one mineral powder is somewhat proportional to the 
size of the particles,^ and that some minerals have very low absorp- 
tive capacities. It was, therefore, advisable to test samples in a fine 
state of subdivision in order to obtain an accurate comparison of the 
absorptive capacities of different minerals. The absorptive capaci- 
ties of larger particles could then be calculated closely enough for our 
purpose from the values of the fine particles, on the assumption that 
the absorptive capacity of any one kind of mineral powder is to a 
large degree inversely proportional to the size of the particles. 
Accordingly, a sample of each mineral was prepared which contained 
only particles ranging in size from 1 to 50 microns in diameter. This 
group of particles is just above the size we arbitrarily set for the col- 
loids and corresponds to the ''fine fraction'' separated from the soils 
in the study just described. 
PREPARATION OF MINERAL POWDERS. 
After the minerals had been crushed to small pieces in an arbor press 
and in a steel mortar, they were ground either in an agate mortar or 
in a steel ball mill, depending on the hardness of the mineral. The 
harder minerals were more effectively handled in the ball mill. The 
samples were frequently passed through a 130-mesh sieve in order to 
obtain the desired fineness with a minimum amount of material finer 
than 1 micron. The sieved powders were then further graded by 
subsidence in absolute alcohol in order to remove particles that were 
larger than 50 microns and smaller than 1 micron in diameter. Abso- 
lute alcohol was used for the subsidence rather than water, in order to 
prevent the formation of colloidal matter by hydrolysis of the mineral. 
The mineral powders were carefuUy examined under the micro- 
scope and counts made of the approximate number of particles present 
in the three following groups: 0.025 to 0.050 millimeter, 0.002 to 
0.025 millimeter, and under 0.002 millimeter. There were but 
7 The data of Ammon (/) and o JVon Dobeneck (8) show that while the absorption ofany one mineral 
increases with the degree of subdivision, the in^ease in absorption is somewhat less than the increase 
in surface exposed. 
