26(i now cuors feicix 



Azotometer. {Chemisc/ies Centralhlatt, 1860, pp. 243 and 

 534.) 



By this method, which gives accurate results when ap- 

 phed to known quantities of ammonia-salts, Knop and 

 Wolff obtained the following results : 



Ammonia in dry soil. 



Verj' li<,^ht sandy soil from birch forest 0. 00077" jo 



Rich lime soil from beech forest 0.00087 



Sandy loam, forest soil 0.00012 



Forest soil 0.00080 



Meadow soil, red sandy loam .0.00027 



Average 0.00056 



The rich alluvial soils from tropical America are ten or 

 more times richer in ready-formed ammonia than those of 

 Saxony. These figures show then that the substance in 

 question is very variable as a constituent of the soil, and 

 that in the ordinary or poorer classes of unmanured soils 

 its percentage is scarcely greater than in the atmospheric 

 waters. 



The Quantity of Ammonia fluctuates, — Boussingault 

 has further demonstrated by analysis what we have insist- 

 ed upon already in this chapter, viz., that the quantity of 

 ammonia is liable to fluctuations. He estimated ammonia 

 in garden soil on the 4th of March, 1860, and then, moist- 

 ening two samples of the same soil with pure water, ex- 

 amined them at the termination of one and two months 

 respectively. He found, 



March 4th, 0.009" |„ of ammonia. 

 April " 0.014" " 

 May " 0.019 " " 



The simple standing of the moistened soil for two 

 months sufficed in this case to double the content of am- 

 monia. 



The quantitative fliuctuations of this constituent of the 

 soil has been studied further both by Boussingault and 

 by Knop and Wolff. The latter in seeking to answer the 



