lt<0 



the gas bags attached, provided the composition of the 

 gas inside was siiitable. Also, some experiments in 

 which the gas supplied was controlled in glass gasomet c 

 instead of in rubber gas bags showed essentially the 

 same results as did the experiments with the bags. 



The oxygen and nitrogen used in the experiments 

 were commercial gases, purchased in compressed' form in 

 eg linders, from the Linde Air Products Company. ose 

 gases are made by the fractional distillation of liquid 

 air and contain practically no impurity except that 

 each gas may contain traces of the other, The traces of 

 nitrogen present in the oxygen are unimportant. Traces 

 of oxygen in the nitrogen are usually so small as to be 

 negligible in ordinary experiments. ".There extreme pre- 

 cision was required in the removal of all oxygen the 

 nitrogen was thoroughly washed with alkaline pyrogallol 

 before use. In a few cases, also, nitrogen was prepared 

 from air by removing the oxygen with this same reagent. 

 No attention was paid to the traces of argon and other 

 rare gases doubtless present in all of the gases used. 



E a soil used in the experiments was of 

 several types. For most cases a rather l^eayy loam of 

 quite ordinary character was used alone or mixed with 

 from one half to three times its volume of fine quartz 

 sand. Jach of three other loaas was used for a few 

 experiments, and in a few cases humus, stable manure, 

 ground bone or ground feldspar was added to one or more 



