The Hydrogen-ion Concentration, etc., of Sea-Water. 49 



The apparatus was filled with the palladium solution, a platinum wire 

 was inserted for an anode, and a current of 4 volts passed until the gold 

 became black. When necessary to repalladize, the apparatus was left 

 full of nitric acid over night, which completely removed the palladium; 

 it was then rinsed and the process was repeated. Palladium black 

 does not seem to have the effective life of platinum black, but it was 

 necessary to use it, as it can be removed without destroying the glass 

 parts or the use of aqua regia. There is no necessity of electrolyzing 

 H 2 S0 4 in a hydrogen electrode, since any free chlorine is removed with 

 the anode when it is taken out unless some aqua regia has been formed 

 in it. If the palladium black is deposited too rapidly it will not stick. 



We do not understand why a freshly palladized electrode behaves in 

 a more reliable manner than one which has soaked in distilled water 

 for many days. The electrode is thoroughly charged with hydrogen 

 when palladized, but much of this is probably lost during the rinsing. 

 Nitric and, to a smaller extent, other mineral acids attack palladium 

 black and oxidizing agents probably affect it; hence we never cleaned 

 palladized electrodes with cleaning fluid. It seems probable that the 

 alcohol and indicator were slowly reduced by hydrogen at the surface 

 of the palladium black, although we did not detect any decrease hi the 

 indicator concentration. It is probable that most of the alcohol had 

 been removed from the solution by the current of gas before our final 

 readings were taken. Such suppositions as to the effect of alcohol 

 were based on variations of the potential of the electrode, and it is well 

 known that such occur whenever free oxygen or oxyhemoglobin, that 

 are reduced at the electrode, are present. The phenomenon is probably 

 general, the ease with which the substance is reduced probably deter- 

 mining merely the rate at which reduction occurs. Even with a small 

 amount of free oxygen in the solution, we found it possible to obtain 

 the correct result by tilting the apparatus so that the platinized or 

 palladized disk communicated with the main mass of fluid by only a 

 thin film. 



The oxygen is apparently used up in the immediate vicinity of the 

 electrode in a relatively short tune and the proximity to a large store 

 of hydrogen hastens the resaturation of the electrode. But because 

 of our ignorance of some of the factors involved, we did not wish to rely 

 on any such assumptions in this set of experiments, and consequently 

 we made about half of the determinations soon after immersion and 

 half soon after removal of the electrode by tilting, and took the mean 

 value. In some cases it was necessary to apply corrections for pressure 

 of C0 2 and H 2 when they were not exactly the same in the gas and 

 water phases, but these differences never exceeded 1 per cent of the 

 partial pressure of the gas. We were very much impressed with the 

 slowness with which equilibrium is established between the solution 

 and the gas above it. 



