Relation of Xonavailable Water to Hygroscopic Coefficient 49 
weighed quantity of the air-dry soil, of which the moisture con- 
tent had been determined, on a smooth cement floor and adding 
the calculated amount of water in small portions while the mass 
was being shoveled over. The whole was mixed thoroly. first by 
shoveling, then by passing it twice thru a swinging sieve of one- 
fourth-inch mesh, and finally by again shoveling the mass 
thoroly. The moist soil thus prepared was immediately placed 
in a large covered can in which it was kept until transferred to 
the cylinders. The moisture percentages reported in the tables 
are those actually found in the soil after it had been mixed and 
placed in the large cans; thus, for example, where it was desired 
to obtain a series of moist subsoils containing 6.0. 10.0. 14.0. and 
18.0 per cent of water, there was obtained instead a series with 
5.8, 9.8, 14.1, and 18.2 per cent respectively. 
Solid instead of one-hole rubber stoppers were placed in the 
drainage nipples of these 12 cylinders before the filling was be- 
gun. A small amount of the subsoil was added at a time and 
tamped before adding more. When a cylinder was filled to 
within twelve inches of the top all of the second 6-inch bulk sam- 
ple, and also all of the first 6-inch bulk sample, except the final 
inch, was added with constant tamping. The two portions of 
the surface foot were added in an air-dry condition. Then using 
the method described on page 48 there was added the amount of 
water required to raise the total moisture content of the surface 
foot to 20 per cent. When the water had soaked away the cylin- 
ders were placed in the pit in the greenhouse, the seed planted, 
and the final inch of air-dry soil added. 
SEPARATION OF ROOTS FROM THE SUBSOIL. 
As the subsoil was very dry in most of the cylinders, it was 
easy to screen out the roots. In the case of each 3-inch section, 
after a pint sample, to be used for the moisture determination, 
had been placed in a jar, the rest was passed thru an eighth-inch 
screen. The portion of the section placed in the jar was not 
thus screened, but as it was being transferred to the jar a care- 
ful watch for roots was maintained. Where the soil was too 
moist to pass readily thru the screen, as in the lower part of 
cylinder VIII, it was allowed to stand in a jar of water for a 
few hours and then washed thru the screen, using a jet of 
water. All the roots thus secured from the four 3-inch sections 
of each foot were placed in a jar of 2 per cent formaldehyde and 
left until all the cylinders had been opened. Finally the roots 
were washed on a screen in running water, carefully separated 
from other organic residues, placed on disks of black filter paper, 
and photographed. Then they were dried and weighed. No at- 
tempt was made to separate the roots from the surface foot of 
