K. J. lirssKLi, AND A. Appleyakd 45 



a perceptible falling off of oxygen, and a still greater one in waterlogged 

 soils. 



2. Besides this free air there is another atmosphere dissolved in 

 the water and colloids of the soils. This consists mainly of COg and 

 nitrogen and has practically no oxygen. 



3. The fluctnations in composition of the free soil air are mainly 

 due to fluctuations in the rate of biochemical change in the soil, the 

 curves being similar to those showing the amount of nitrate and 

 the bacterial counts as far as they were taken. The rate of bio- 

 cliemical activity attains a maximum value in late spring and again 

 in autumn, and minimum values in summer and winter. In autumn 

 the bacteria increase first, then the CO2 rises, and finally the nitrate 

 increases. 



From November to May the curves closely follow those for the soil 

 temperature which thus appears to be the dominating factor; from 

 Ma}' to November they follow the rainfall and to a less extent the soil 

 moisture curves. The distinct difference between rainfall and soil 

 moisture indicates that rainfall does something more than add water 

 to the soil. It is shown that the dissolved oxygen brought in is probably 

 a factor of considerable importance in renewing the dissolved soil 

 atmosphere and facilitating biochemical change. 



4. Grass land usually contains more COj and less oxygen than 

 arable land but we cannot attribute the difference to the crop owing 

 to the large dift'erences in soil composition and conditions. It is 

 difficult to ascertain the precise effect of a crop, but as the soil differences 

 are eliminated so the dift'erences in composition of the soil air become 

 less and less. No evidence could be obtained that the growing crop 

 markedly increases the amount of COj in the soil air, and if it gives 

 rise to any great evolution of CO, in the soil it apparently exercises 

 a corresponding depressing effect in the activities of soil bacteria. 

 This result agrees with one obtained earlier in reference to the nitrates 

 in the soil. 



5. Such weather conditions as barometric pressure, wind velocity, 

 variations in temperature from the mean, small rainfall, etc. seem to 

 have but little ett'ect on the soil atmosphere. 



