14 THE BIOLOGY OF FLOWERING PLANTS 



attempted a new classification based on his freezing-point 

 method. The amount of moisture in the soil freezing out 

 at any given temperature, may be estimated by the expansion 

 of the soil (the pores of which are previously filled with 

 a non-freezing organic fluid, ligroin) determined in a special 

 instrument, the dilatometer. Bouyoucos finds that a fraction 

 of the moisture freezes out from o° C. to — 1'5° C, a further 

 fraction between —4° C. and —78° C, and a third fraction 

 not even at the latter low temperature. The fraction which 

 freezes at — 1'5° C. is termed free zvater, the remainder 

 imfree water. The free water is to be identified with what 

 we have termed capillary moisture ; of the unfree water, 

 that fraction which freezes between —4° C. and —78° C. 

 is termed capillary absorbed, and corresponds to the water 

 absorbed by the colloids. The water which does not freeze 

 is water of hydration and solid solution. Bouyoucos holds 

 that the amount of unfree water is a constant for a given 

 soil independent of the total moisture present ; but Keen 

 (1919, 1922) has obtained evidence from the depression of 

 the freezing-point at difi"erent moisture contents that this 

 is not the case. Bouycucos's method has the advantage 

 of enabling us to separate the water exactly into the 

 different fractions ; the divisions do not correspond exactly 

 to those of the older classification. 



We know, however, principally from the work of 

 Keen (1914, 1922) on the mode of evaporation of water 

 from soil, that such fractions are not in any case really 

 sharply separated ; the one merges into the other. If we 

 think of a colloid like glue in a dry condition absorbing water, 

 we see that after it has become saturated the surface is no 

 longer sharply delimited ; the imbibition water of the soil 

 colloids which coat the sand grains, therefore, passes without 

 a break into the thicker film of moisture which may be 

 retained by capillarity. On drying out there can be no 

 sudden change from film water to imbibed water. There is, 

 however, a range of water content in which the forces 

 retaining water in the soil increase very rapidly though 

 continuously. The amount of capillaiy water is relatively 



