410 PROCEEDINGS OF THE AMERICAN ACADEMY. 



some simple fraction thereof, since n and m are always positive whole 

 numbers by definition. 



We can obtain the values of A for the experiment whose results are 

 given in Figure 2 by finding the slope of the straight line which best 

 satisfies the individual points. This method permits a possible error of 

 perhaps one per cent. The value so obtained is 0.159, if E is expressed 

 in volts. Similarly, from another experiment at the same temperature 

 of 22° we obtain a value for A of 0.157. The maximum value consist- 

 ent with ecpiation (4) is 0.058, when m and n are both equal to unity. 



Between 35.5° and 36.5° four experiments were maile. The values 

 of A obtained from these were 0.166, 0.157, 0.162, 0.166, with an aver- 

 age 0.163. The difference of over 3 per cent between this value and the 

 average value, 0.158, found at 22°, is probably considerably greater than 

 the error in determining these values, and shows that A increases with 

 the temperature; but whether proportionally to the temperature, as equa- 

 tion (4) demands, cannot be determined accurately from our experiments. 



Since we find that our experiments are inconsistent with equation (4), 

 we must conclude that at least one of the assumptions upon which that 

 equation was based is false. The assumption which seems most likely to 

 be erroneous is that the process determining polarization is a chemical 

 reaction in a homogeneous phase. Even without the evidence of our 

 experiments, is not this assumption improbable?* Is it not more prob- 

 able that the processes, whatever they are, take place in that surface film 

 which marks the transition from metal to electrolyte, and that processes 

 of diffusion to, from, or through this film play an important rdle ? 



Two important diffusion processes suggest themselves at once as pos- 

 sible causes of the polarization, — the diffusion of gaseous hydrogen from, 

 and the diffusion of hydrogen-ion to, the electrode. But these can both 

 be excluded by the aid of certain general criteria. 



If any polarization is due to the accumulation of the final products of 

 electrolysis, there will be, as a rule, especially if this product is a gas, 

 some potential at which its concentration becomes so great that a new 

 phase will suddenly form and the current will begin to rise more rapidly. 

 There will be, in other words, a break in continuity in the polarization 

 curve. As a matter of fact, we observed no discontinuity, in the region 

 of our experiments, that is up to over 0.7 volt, although at this potential, 

 if the polarization were due to the accumulation of gaseous hydrogen, the 

 hydrogen would be at a pressure of over 10 24 atmospheres, — an incon- 



* See Nernst, Zeit. phys. Chem., 47, 52 (1904). 



