128 
Journal of Agricultural Research 
Vol. VII, No. 3 
As a means of comparison the OH-ion concentration of solutions sat¬ 
urated with Ca(HC 0 3 ) 2 and CaC 0 3 were determined with an exactly 
similar technique. Ca(HCO s ) 2 gave a value for OH“ of 0.5 X io -7 and for 
CaC 0 3 0.3 Xio -9 . The figure for Ca(HC 0 3 ) 2 is almost identical with 
those obtained for the alkaline soils. This is in keeping with the theo¬ 
retical considerations advanced above. T. Saidel (29) determined the 
OH-ion concentration of several soil extracts after prolonged boiling. 
His values approximate that given by CaC 0 3 , which is distinctly higher 
than the normal OH concentration for soils of this class. 
At first attempts were made to estimate the H-ion concentration of 
filtered soil extracts. Serious difficulty was encountered in this pro¬ 
cedure, in that the nitrates present were slowly reduced by the hydrogen 
gas in contact with platinum black. Obviously this reduction of nitrates 
would result in the production of NH 3 and a residue of fixed alkali, 
with a corresponding increase in the concentration of OH ions. While 
the reduction of nitrates does occur to some extent in soil suspensions, 
yet in this case the results are not appreciably changed. This is accounted 
for by the fact that the acid soils contain a large excess of potential acid¬ 
ity, while the absolute amounts of alkalinity produced are exceedingly 
minute. Therefore, equilibrium would be immediately restored without 
sensibly affecting the true H-ion concentration. As evidence thereof, 
the voltmeter readings for acid soils became constant within a few 
minutes and remained constant for an indefinite time. 
In order to determine whether the reduction of nitrates gave rise to 
appreciable errors in the readings for alkaline soils, the OH - concentra¬ 
tion of a Ca(HC 0 3 ) 2 solution was measured in the presence of large 
amounts of NaN 0 3 . No disturbance of the equilibria was noted under 
the conditions of the experiment. It has been suggested that the use of 
palladium or iridium, instead of platinum, for coating the electrodes 
would practically prevent the reduction of nitrates, but the authors have 
no experimental data upon this point. 
While the reduction of nitrates prevents the use of the hydrogen elec¬ 
trode with extracts of acid soils, it is not believed that this factor would 
cause appreciable errors in the case of nutrient solutions containing 
nitrates. A nutrient solution would usually have a considerable poten¬ 
tial acidity or alkalinity—for example, the unhydrolyzed or undissociated 
fraction of alkaline phosphates—so that constant results could be 
obtained, just as with soil suspensions. Moreover, by use of the shaking 
method described by Michaelis (21) the reduction of NOJ may be reduced 
to a minimum, but even this slight production of alkali causes serious 
error in soil extracts, where the total quantity of acid present is exceed¬ 
ingly small. 
