E. J. RussEL]. AND A. Applbyard 



Table I. Comparisori of composition of soil air taken from a depth of 

 6 and 18 in. in the soil. 30 January 1914. Percentage by volume. 



Unlike atmospheric air, however, the soil air is not constant in com- 

 position but changes somewhat from day to day and even on the same 

 day at different spots in the field ; nevertheless the values fall within 

 fairly narrow limits. 



There are two Icinds of variation in composition ; the local daily 

 ones just referred to, and the greater variations produced by season, 

 treatment, etc. : the latter may be so great as to mask altogether the 

 local fluctuations. In our experiments the greatest factor of all was 

 the effect of season. Whatever the history of the soil its atmosphere 

 in spring and to a less extent in autumn was characterised by high 

 amounts of carbon dioxide indicating rapid biochemical changes at 

 these seasons of the year, while in summer and winter the amounts 

 were much lower. The effect is complex and includes at least two 

 others each of which was found to be very potent: the temperature 

 during the period December to June, and moistness of the soil during 

 part of the summer months. (Figs. 7 and 8.) 



In addition there is the possibility that a certain amount of partial 

 sterilisation has taken place during the winter and during the dry 

 summer, leading to considerable bacterial activity immediately con- 

 ditions become favourable once more. 



This seasonal effect dominates all the others and impresses on all 

 the curves the same general type seen in Figs. 1-6^. Other factors, 

 such as manuring, cropping, etc., simply raise or lower the whole curve 

 according as they give rise to more or less carbon dioxide; in par- 

 ticular the effect of the crop proved to be considerably less than was 

 anticipated. 



Within these major variations there fall the smaller fluctuations 



See Table VI for data. 



1—2 



