134 
A second solution with a different hydroxyl-ion concentration gives 
in the same way: 
Y 
C 
E' = a. + Won = Re + 0,0595 log — 
c 
Therefore: 
E'.— EH = z'on — zon —0,0595 log be 
c 
from which, substituting pon = — loge and p'on = — log c!: 
ph 0005 (p'on — POH): 
Thus, by measuring the electromotive force of the calomel elec- 
trode, combined, first with a solution of known por, e.g. à SORENSEN’S 
phosphate solution, and then with an 8°/, phosphate solution, this 
equation for £’— ME gives at once the value p'on. 
Determinations of this kind with two SÖRENSEN’s solutions gave 
nearly the right value. 
As is well known however, an OH electrode does not give at 
all as constant and accurate results as a H electrode. The potential 
of an OH electrode has always been found about 150 millivolts too 
low, which fact is commonly ascribed to the formation of some 
suboxyde of platinum. 
This constant depression is, however, eliminated in our formula. 
Moreover it proved possible to arrange the experiments in such a 
way, that the influence of the inconstancy was repressed considerably. 
Some 10 ec. of the liquid to be examined were introduced into 
a smal] tube of about 1 cm. diameter. By dipping the electrode tube 
into this small quantity a quick and thorough saturation with oxygen 
could be obtained. 
The oxygen, free from hydrogen, was washed in a bottle with 
3°/, KI solution to keep back possible traces of ozone. It was 
brought to the temperature of 27° by keeping the washbottle and 
rubber tubing as well as the other apparatus (the same as used 
for the py determination) in the thermostat. 
After oxygen has been through for some 10 minutes, the electrode 
tube was lifted out of the liquid under examination and connected 
with the calomel electrode as usual by means of the KCl-cup; the 
electromotive force was read within about one minute. 
This process was carried through alternately with a SÖRENSEN’s 
solution (usually 5,6 cc. alkaline + 4,4 ¢.c. acid phosphate), then 
with the 8°/, phosphate, and then with another SérEensen’s solution 
(9 ce. alkaline + 1 ce. acid phosphate). Immediately after this the 
same series of three observations was once or twice repeated. Every 
estimation of the unknown pon was thus preceded and followed at 
