26o 



Journal of Agricultural Research 



Vol. XVI, No. lo 



extract of the soil, which showed it to contain o.ooi per cent of lime, 

 (CaO), 0.003 per cent of magnesia (MgO), and o.ooi per cent of potash 

 (K2O). After two months contact with the glass bottles in the thermo- 

 stat, an analysis of the solution gave o.oi per cent of lime, 0.002 per cent 

 of magnesia, and a trace of potash. 



Table I. 



-Solubility of the lime in epidote, the magnesia in chrysolite, and the potash in 

 muscovite, in contact with a slightly acid soil extract 



[Minerals ground to pass a screen of loo meshes to the inch] 



o Probably too low. 



The results tend to show that, under the conditions of the experiment, 

 more lime and potash was extracted from the silicates epidote and musco- 

 vite, ground to pass a sieve of 100 meshes to the inch, by two months' 

 contact with the soil extract kept at 25° C. in a thermostat, than of mag- 

 nesia from its silicate in the form of chrysolite similarly ground and sub- 

 jected to the same conditions. 



The removal of such proportionally large amounts of lime and potash 

 from silicates by an acid soil extract would seem to indicate that in time a 

 soil's fertility index, with respect to lime and potash, would under proper 

 conditions of acidity be quite appreciably lowered. That the potash 

 results are truly representative is strengthened by the fact that in perco- 

 lation experiments tried out by the author a short time ago with dilute 

 acid solutions, such as phosphoric acid and sulphuric acid, in every in- 

 stance considerable amounts of potash were removed from soil silicates, as 

 was the case also when a peaty soil from Maine intermixed with white 

 mica was shaken up with distilled water and allowed to stand for 24 

 hours before filtering, the filtrate showing potash to be present in a water 

 soluble condition to the extent of 600 p. p. m. 



Because of the small amounts of the samples used in Nos. i to 4 in rela- 

 tion to the quantity of solution, it would seem to follow as a natural con- 

 sequence that the results for those samples would be high and should 



