24 The Afiitosp/i(ir o/ tin Soil 



to come into play. In the first j)!acc the rain does not immediately 

 distribute itself throughout the soil but produces a more or less saturated 

 layer which seals the surface and prevents the escape of COg from the 

 soil air. Further, rain appears to be nearly saturated with dissolved 

 oxygen. We have already seen that the dissolved atmosphere in the 

 soil tends to lose oxygen more rapidly than to gain it and in consequence 

 is largely anaerobic. A large fall of rain bringing with it oxygen in 

 solution affords the possibility of partially renewing the dissolved 

 atmosphere and giving the organisms a new lease of acti\'ity. In 

 time, however, the oxygen is used up and the activity falls off even 

 though the moisture remains constant. This effect is probably most 

 marked when the soil is dry and the new dissolved atmosphere can 

 most completely replace the old one. We could find no determinations 

 of the amount of dissolved oxygen in rain water but a number of analyses 

 of stream waters have been made by the Sewage Commission, and they 

 show that on an average about ten parts per million by weight of 

 dissolved oxygen is present. If we suppose that rain contains approxi- 

 mately the same amount then 1 inch of rain brings down 2^ lbs. of 

 oxygen per acre; this if converted into COg would add 0-8 to the normal 

 0-2 per cent, by volume and make the total up to I per cent. In 

 addition the rain itself brings down a certain amount of CO2, but not 

 much, and considerably less than the amount of oxygen. 



Relation between soil air and atmospheric air. The experiments 

 described in this section show that COo is produced at maximiim rates 

 in spring and in autumn and at minimum rates in summer and winter. 

 As it is constantly escaping from the soil into the atmosphere we should 

 naturally expect to find that the CO., in the atmospheric air also reaches 

 maximum amounts in spring and autunni, minimum amounts in 

 summer and winter. 



Systematic determinations of the amount of CO.2 in atmospheric air 

 are not numerous, but those made prior to 1899 were collected by Letts 

 and Blake in their paper already quoted^. A statistical examination 

 of the data shows that, as far as they can be relied upon, they indicate 

 an increase in atmospheric COg during the period I\Iarch-]May, a falling 

 off during the period May to August, and a rise during the period 

 October to January. Thus a very close agreement is obtained with our 

 soil results. 



' Proc. Hoy. Sor. Dublin. 1000. 9. K)7-270 and especially pi). 205 et seq. 



