Apr. 9,1917 
Water-Retaining Capacity of Soil 
33 
In filling the cylinders the soil was added very slowly with constant 
tamping, care being taken to insure the firmness of that already in before 
adding more. Blows as uniform as possible were delivered by means 
of a tamper, for which we used a 2-inch rubber stopper on the end of a 
^-inch gas pipe 3 feet long. As the cylinders were being filled, three 
samples of each soil were taken for moisture determinations. 
In such experiments as these the removal of exactly the desired depth 
of soil is an important operation, but in many cases a difficult one. 
Our practice with those of the metal cylinders in which there was no 
direct contact of the soil column under experiment with the subsoil mass 
was to place a cylinder on a table, and, by means of a can opener, to 
open it lengthwise, 6 to 12 inches at a time. The measured portions of 
the column of soil were then sliced off by means of a large spatula 
and placed in Mason jars, which were covered at once. With those of 
the cylinders open at the bottom and having the soil column in contact 
with the subsoil mass the successive layers were removed by a plate 
auger whose diameter was a little less than that of the cylinders. After 
the main portion of each section had been thus removed, the small resi¬ 
due next the cylinder wall was taken up with a spoonlike instrument. 
After all the samples had been secured from a cylinder, the contents of 
each jar were thoroughly mixed. The moisture determinations were 
made in an oven kept at no 0 C., through which passed a rapid current 
of dry air. Precautions were taken to insure the thorough drying of 
all the samples. 1 
As was to be expected, all the surface soils, after being tamped into the 
cylinders, much resembled their condition in a well-prepared seed bed. 
The structure of the subsoils, with the exception of A, very closely resem¬ 
bled that found in excavations in the field. In the cylinders soil A had 
in all cases a granular structure, very different from that observed in the 
field. This, however, did not appear to influence seriously the move¬ 
ment of moisture, as will be seen below in the comparison of field data 
with those obtained in the laboratory experiments. 
In most of the experiments duplicate cylinders were employed. The 
moisture conditions in the duplicates are so similar in all the soils, 
except the sand, Q, that we report only the average, no purpose being 
served by reporting the data from the two separately. With the sand 
the data on all the cylinders are reported. As illustrations of the degree 
of concordance with the soils other than the sand the data on two, C and 
I, are given in Table II. 
The results in these experiments and in those previously reported (3) 
would make it appear an unprofitable expenditure of time, at least in 
preliminary studies, to use duplicate cylinders. 
1 Further particulars of the method have already been given in a report on cylinder experiments (2, p. 29). 
