EXPERIMENT STATION BULLETINS. 673 



It must be stated at once tliat a thorough experimentation and a criti- 

 cal theoretical examination of both these factors have failed to indicate 

 that they are the causal agents of the results noted. Some of the evi- 

 dences both experimental and theoretical which bear out very strongly 

 this conclusion may be mentioned as follows: (1) The data in Table 5 

 show that the depression increases in a geometric progression as the 

 percentage of water decreases in an arithmetic progression, consistently 

 and uniformally from the highest to the lowest water content. Now if 

 the pressure of the films influenced the depression the rate of increase 

 should not be the same, but entirely different, since the pressure of the 

 films does not begin to be manifested until the moisture content is con- 

 siderably reduced. Probably there should be a rather abrupt change in 

 the results at the point where the water is reduced almost entirely to 

 the film water. 



(2) If the pressure of the films influences the freezing point lower- 

 ing of soils at the low moisture content, it should afi'ect it also at the 

 high moisture content because after the free water is solidified the re- 

 maining water exists in films just as in the low moisture content and 

 the effect would more or less constant in both cases. 



(3) There are abundant evidences Avhich go to indicate that 

 when the soil mass is sufficiently supercooled and the water be- 

 gins to freeze the physically adsorbed water freezes like free 

 water, if it freezes at all, and the forces which hold the water 

 do not infiuence the lowering of the freezing point. If pressure or the 

 allied forces influence the depression at all, it is in causing some of the 

 water in the soil to become unavailable or inactive, and thus not be in a 

 free condition to act as a solvent and take part in dissolving the salts 

 of the soil. This question, however, will be dealt with more in detail 

 subsequently. 



(4) The solid particles themselves do not influence the freezing-point 

 depression as proved from the following evidences: (a) The lowering of 

 the freezing point of very dilute soil suspensions is practically insignifi- 

 cant, (b) The freezing-point lowering of colloids is also so small as to 

 be negligible. Indeed one of the characteristic properties of colloids, 

 which distinguishes them from the crysalloids, is that they produce 

 practically no depression of the freezing point, (c) The depression of 

 a solution such as n/10 NaNO.^ determined in pure quartz sand is almost 

 identical Avith that obtained in the solution itself. Thus the freezing- 

 point depression of n/10 NaNOg solution alone was ..370°, that of the 

 solution Avith quartz sand .368''C or a difl'erence of only .002°C. 



(5) The washed quartz which a priori would not be expected to con- 

 tain a solution of high concentration gave a depression of only .070° C. at 

 1.5% moisture, while clay loam which a priori would be expected to 

 possess a solution of very high concentration gave a depression of 1.075 °C 

 at 39.28% of moisture. 



There are still other evidences which could be brought forward to sup- 

 port the view but the foregoing may be considered sufficient. 



The conclusion seems to be inevitable, therefore, that the unusual and 

 significant lowerings of the freezing-point noted are caused by and repre- 

 sent actual concentration of solution, and therefore contradict the views 

 held that the soil solution is very dilute, that its magnitude of concentra- 

 tion tends to remain constant, and to be the same for all soils. 



