10 BULLETIN 1103, U. S. DEPARTMENT OF AGRICULTURE. 
determinations made by this method usually agreed well, variations 
between different determinations averaging about 5 per cent of the 
values obtained. 
In determining the adsorption of ammonia, 5 to 10 gram samples 
of soil or 2 to 5 gram samples of colloid which had previously been 
dried for 18 hours at 110° C. were used. The sample was placed in 
a U-tube which was evacuated for 15 minutes while immersed in 
boiling water. After replacing the boiling water about the U-tube 
by an ice-pack, ammonia gas was let into the U-tube through a dry- 
ing train for two to four hours. The pressure of ammonia gas 
during this time was maintained at 1 centimeter above that of the 
atmosphere by a mercury manometer. When the mercury level in 
the manometer remained without perceptible change for 30 minutes 
with the ammonia cut off, the adsorption was regarded as complete. 
Two bottles containing a saturated solution of boric acid were then 
attached to the U-tube. the ice-pack was removed, the pressure in 
the apparatus was kept reduced by about 10 centimeters of mercury 
in the manometer by means of a filter pump, air was let in through 
a soda-lime tower, and boiling water was again placed around the 
U-tube. In this way all the ammonia was swept out of the appara- 
tus within 30 to 40 minutes. The ammonia in the boric acid solu- 
tion was titrated, using methyl orange as an indicator. The quantity 
of ammonia thus obtained was corrected for the blank of unadsorbed 
ammonia in the apparatus. This method gave the amount of am- 
monia adsorbed at 0° C. which was driven off at 100° C. Variations 
in duplicate determinations by this method were usually not greater 
than 3 per cent of the values obtained. 
The adsorption of water vapor was determined as follows : About 2 
grams of colloid or 2 to 4 grams of soil, previously air dried and passed 
through a 100-mesh sieve, were weighed into shallow weighing bot- 
tles. The uncovered weighing bottles were placed in a desiccator 
containing a large amount of 3.3 per cent sulphuric acid, and the 
desiccator was then evacuated to 50 millimeters or less of mercury. 
The desiccator was kept in a thermostat maintained at 30° C. for 
five days; at the end of which time the adsorption of water va- 
por was practically complete. The samples were then weighed 
moist and weighed again after drying for 18 hours at 110° C. The 
difference between the two weights represented the quantity of 
water adsorbed by the sample. A more detailed description of the 
procedure is given in a previous paper (Robinson. 26). 
The adsorption of the soil and of the corresponding colloid were 
determined in the same run. Under these conditions duplicate de- 
terminations usually agreed within 2 per cent of the values obtained. 
Duplicate determinations made in different runs might show con- 
siderabh 7 larger variations. 
EXPERIMENTAL RESULTS. 
THE ADSORPTIVE CAPACITIES OF SOILS AND EXTRACTED COLLOIDAL MATERIAL. 
The adsorptive capacities of 32 soils and of samples of the colloidal 
matter extracted from these soils are given in Table 1. In each case 
the adsorptive capacity is expressed as the weight of dye, water, or 
ammonia adsorbed by 1 gram of the material. 
