llf, BULLETIN 1059, V. S. DEPARTMENT OF AGRICULTURE. 



average weight was about 100 grams. Especially was it noted that 

 surface samples which were nearly air-dry when taken absorbed 

 vcrv little moisture in these short periods. 



Finally, from November 28, 1918, to March 29, 1919. a test was 

 made involving samples from nearly all of the regular soil-sampling 

 paints al the Fremont Experiment Station, as well as a miscellane- 

 ous Lot exemplifying various peculiar characters. Two bell jars 

 were employed, a small dish of sodium hydrate solution being placed 

 in each. At the end of the four-month period these solutions bad 

 nut absorbed vapor to quite the same extent in the two jars, and in 

 neither case was the total absorption equal to the losses from all of 

 the soils. However, for practical purposes the two container- were 



in equilibrium with each other, the osmotic pressures of the solu- 

 tions being 21.5 and 20.2 atmospheres, respectively. 



At the outset each sample of soil was given 10 per cent of moisture 

 above its air-dry weight, so that the moisture was available in liquid 

 form and the various soils were not radically unlike in their initial 

 conditions. While it is not certain that the time allowed was suffi- 

 cient to establish equilibrium, it is to be noted that the changes in 

 moisture content varied from losses of about 3 per cent to gains of 

 fractional percentages and, in one case, where there was much raw- 

 humus, a gain of 15 per cent. 



The results, as shown in small part in Table 6 and diagrams v «.». 

 and 10, are very elucidating. These diagrams are prepared some- 



