200 



A. F. GUSTAFSON 



to prevent evaporation and the bottles with soil weighed at once after wiping them perfectly 

 dry. After weighing, the soil was dried in the bottles at 105°C. for 8 hours and again weighed. 

 The loss is the hygroscopic moisture or, expressed as per cent on the basis of oven-dry soil, the 

 hygroscopic coefficient. 



The results of the determinations appear in table 7. 



As pointed out by Beaumont (3, page 492^93), the probable error of deter- 

 minations of hygroscopic coefficient is high. The accuracy of this statement 

 is well borne out by the figures in table 7, His results indicated that slight 

 variations in temperature made little difference in adsorption of vi^ater vapor 

 so the only effort in this work was to keep the temperature sufiiciently con- 

 stant to prevent condensation of free water on the particles, or in the inter- 



TABLE 7 

 Hygroscopic coefficients of air-dry soils used in experiment 4 



* Data taken from table 6. 



stices of the soil. This latter seems to have occurred during the early part of 

 this work as the hygroscopic coefficient was high and quite erratic. These 

 figures seem to bear out Beaumont's conclusion, also, that other factors than 

 total surface aflfect the adsorption of water. Soils 4 and 5 are very similar 

 in texture, consequently in total surface, and the hygroscopic moisture actually 

 in the soils, air-dry, is the same, yet when exposed to the same saturated 

 atmosphere, soil 4, with higher organic-matter and soluble-salt content, ad- 

 sorbs 0.482 per cent more actual moisture, 27 per cent excess over soil 5, the 

 one of lower salt- and organic-content. In this connection, let us compare 

 soils 5 and 6, in table 7. White silt loam has much more surface than Dunkirk 

 fine sandy loam.'' In the air-dry state they appear in their right order as 



^ No mechanical analysis available. 



