61 



air from deep in the aoil show as much oxygen as they 

 do. It is possible that the samples have been con- 

 taminated by downward leakage of atmospheric air, for 

 instance, along the outside of the tube used for with- 

 drawing the sample, but, in any case, the data are not 

 sufficiently numerous to warrant any general conclusions, 

 even if they could be regarded as sufficiently accurate. 



In connection with the data of composition 

 of the soil air it is necessary to devote a word to the 

 total amount of air in the soil since this, jointly 

 with the oxygen content, will determine the total amount 

 of oxygen available to plant roots. In the simple theo- 

 retical case of an assemblage of spherical particles 

 all of the same size, the percentage of pore space, by 

 volume, may vary between 29. 95 and 47. 64 1 depending 

 on the "packing" or mutual arrangement of the spheres. 

 The percentage of pore space is independent of the size 

 of the spheres though the mean diameter of the indivi- 

 dual pore spaces decreases as the spheres are smaller. 

 In the actual soil the particles are not spherical 

 nor are they all of the same size, with the result that 

 the limits of pore space given above are widened great- 

 ly. The presence of particles of various sizes tends 

 to decrease the percentage of pore space. Thus the ad- 

 U. 



1. Schlichter,- f. S. Geol. Survey, Ann. Sep. 

 19: II: 306ff (1899). See also Soyka,- Forsch. Geb. 

 Agr. Physik j 1 (1685). 



