68 
causing osmosis, is undoubtedly regulatory of all the conditions 
within the cell. It is, indeed, through the study of soils that the 
present conception of the nature of osmosis and of the importance of 
the colloids of the plant in this process is arrived at. The work of 
Bouyoucos (109) followed by Hoagland (127) has been especially 
instructive on this point. Bouyoucos and McCool (106) first at- 
tempted to determine directly the concentration of soluble salts in 
the soil by measuring the freezing-point depressions of soils in their 
natural conditions of moisture and with definite amounts of water 
added to them. At the outset, no doubt, he supposed that, allowing 
for increased solubility of the soil substances with increased moisture, 
the freezing-point depressions would increase proportionately as the 
concentration of the solutes increased by reduction of the whole 
moisture of the soil. This he found not to be the case, for, while the 
effect of increasing concentration was shown in greater freezing-point 
depressions as the moisture was reduced, a point was rather suddenly 
reached at which no freezing at all occurred. From this Bouyoucos 
concluded that there must be at all times in the soil a certain amount 
of " unf ree " water, probably not in a liquid state but so adsorbed or 
chemically combined that at no time does it constitute a part of the 
soil solution. This water was conceived to be that which is held 
within the colloidal or clay masses of the soil, but it is fairly evident 
that, since pure sand may contain a small amount of such water, it 
may be in part water in extremely thin films, or more probably in 
the form or independent molecules, on the surfaces of the crystalline 
particles. 
In his later work Bouyoucos (109) succeeded in measuring the 
volume of this unfree or combined water indirectly, by noting the 
volume expansion of the whole mass at the instant when freezing 
occurred. By this means he was able to determine just what propor- 
tion of the whole amount of water entered into the freezing process. 
Adding in every case 5 cubic centimeters of water to 25 grams of air- 
dried soil, he found that the proportion of this 5 cubic centimeters 
which did not freeze was only 2 per cent (0.10 cubic centimeters) in 
the case of a quartz sand, but 60 per cent (3 cubic centimeters) in the 
case of a clay, and 74 per cent for heavy silt loam. 
The exact amount of unfree water for a given soil was found to be 
dependent on several experimental conditions, which led to the belief 
that under certain conditions it could be transformed into free water. 
In other words, there is no fixed point at which the water begins to 
be unfree, but under given conditions each soil exhibits certain definite 
characteristics. Thus there was in all soils, except the quartz sand, 
less water which failed to freeze when 10 cubic centimeters had been 
added to a standard sample than when only 5 cubic centimeters were 
used. There was also less when the supercooling was carried below the 
