Haigh — Estimation of Hygroscopic Moisture in Soils. 533 



methods agree fairly closely, the amount obtained in tlie case of the carbide 

 method is always slightly lower than that obtained by heating in the oven. 

 The difference increases in a rough proportion to the hygroscopic moisture 

 present. In these soils, which were of a similar type, with the exception of 

 7 and 8, the increased hygroscopic moisture present was due almost entirely 

 to an increase in organic material or humus. In the case of 7, whieh is a 

 pure peat, the difference is as much as 1'5 per cent., while in 8, a sand with 

 no organic matter, the results agree. It would appear from this that when a 

 soil contains raucli organic material, the loss of weight on heating to 100° 0. 

 represents more tlian the liygroscopic moisture present in tlie soil, part of 

 the loss being made iip of other volatile constituents which are driven off on 

 lieating. Tliis being so, it should be possible to drive off this volatile 

 material by heating to a moderate temperature, then allowing the soil to 

 cool and absorb moisture from the atmosphere, and the hygroscopic moisture 

 then estimated by both methods should agree very closely. Accordingly the 

 following experiment was carried out. A sample of 6, above, was taken, and 

 the hygroscopic moisture determined by both methods. The result gave 

 •20 per cent, more in the oven than with the carbide. Another sample of the 

 same soil was heated in an air-oven to 180° C. for several hours. It was then 

 exposed to the air for several days, and the moisture again estimated by both 

 methods, and the results were found to agree to within '04 per cent. 



In carrying out these experiments it was found that in the case of any 

 particular sample of soil the results obtained by the carbide method were 

 more consistent and agreed more closely between themselves than did the 

 results obtained from the same sample in the oven. 



The question arises as to how far the carbide affects the combined water 

 in the hydrated silicates in the soil. H. A. Danne' has shown that, in 

 applying the method to the estimation of hygroscopic moisture in organic 

 substances, a marked interval exists between the evolution of acetylene from 

 hygroscopic and that from combined water. No such interval has been 

 observed in the case of soils. The gas is evolved steadily and rapidly. 

 Masson, when determining the water of crystallization of salts, fouud that 

 only the simple salts, such as sodium carbonate and sulphate, were completely 

 desiccated by the carbide at ordinary temperatures, wliile others react, 

 some completely, some partially, only when heated. Another class is shown 

 to be stable to carbide even when heated to 170° C. From this we may 

 conclude that the effect on hydrated silicates, if any, which in themselves 



1 Op. c'lt. 

 SCIENT. PKOC. R.D.S., VOL. XIV., NO. XL. 4 Q 



