EXPERIMENTS CONCERNING CORROSION 207 



diffusing, and outside of which, as at C, C,' C," the concentration 

 of dissolved oxygen is constant. Inside this sphere the amount 

 of oxygen will gradually fall towards the surface of the iron, 

 at which place it will be lowest. The same argument applies 

 whatever shape the iron may possess, but the configuration of 

 the shell will, of course, vary accordingly. Such a shell is 

 known as the " Corrosion Zone." 



Now what will happen if we bring a second ball of iron into 

 the same tank of water? If the distance between the two 

 spheres is greater than twice the radius of the corrosion zone, 

 the metals will not affect each other, and they will each corrode 

 at their maximum rate. But if, as in fig. 8, the corrosion zones 

 intersect, the amount of oxygen that can diffuse towards each 

 metal ball is reduced, and corrosion is proportionately retarded. 

 If three such balls are brought together in line, clearly the two 



Fig. 8. 



outer ones stand the best chance of corroding, for oxygen can 

 diffuse towards the middle one in two directions only, namely 

 from above and below. 



This illustrates the importance of using tanks of sufficiently 

 large capacity, and of having the metals a sufficient distance 

 apart when an attempt is made to determine the relative rates 

 of corrosion of a series of samples. 



The same argument applies to the employment of series of 

 small containing vessels in cases where only one piece of metal 

 is suspended in each. Air can only penetrate to the sides and 

 bottom of the vessel from the surface ; hence, if the vessel is not 

 larger than the corrosion zone (as in fig. 9), the air at the surface 

 will pass into the corrosion zone and be absorbed by the metal, 

 and there is none left to replenish that at A and B, which is 

 likewise diffusing into the corrosion zone. In a short time, 

 therefore, we shall have equilibrium after the manner of fig. 10, 

 and the rate of oxidation of the metal now becomes a function 

 of the surface area of the liquid. It is difficult to arrange a 



