1884.] Chemical Corrosion to Voltaic Current. 339 



calculating from the quantity of deposited silver, together with the 

 valencies and atomic weights of the positive metal and of silver, the 

 amount of the positive metal required to produce the current. Each 

 of these methods however is imperfect ; the second is usually much the 

 least so. In several cases, by employing the former one, the amount 

 of deposited metal was found to be nearly equivalent to the amount 

 of corrosion accompanying external current, after making allow- 

 ance for a small deficiency of deposited silver caused by a portion of 

 the current traversing the free potassic cyanide in the depositing 

 liquid. 



That the external current affects in some cases the amount of "local 

 action " was shown by the following experiment : — A sheet of platinum 

 and one of ordinary zinc, not connected together, were partly 

 immersed in dilute sulphuric acid, until a moderate stream of gas 

 arose from the zinc. On now connecting the two plates together by 

 means of a wire, the stream from the zinc was conspicuously dimi- 

 nished. The contact of a negative metal with the positive one 

 usually increased the total loss by corrosion, whilst it commonly 

 decreased that caused by " local action." 



The results contained in the table further show that the propor- 

 tion of corrosion attending external current (as measured by the 

 amount of silver deposited) to that by " local action," depends upon 

 several conditions : — 1st, upon the kind of metal ; thus it was usually 

 greater with cadmium (average 75*63 per cent.) than with copper 

 (30*33 per cent.). 2nd, upon the kind of liquid ; it was greater with 

 solution of potassic cyanide (average 63*27 per cent.) than with dilute 

 nitric acid (31*14 per cent.). And 3rd, upon the degree of concentra- 

 tion of the liquid ; a greater strength of solution frequently increased, 

 and as frequently decreased, the proportion of corrosion accompanying 

 external current to that by " local action." A variation of kind of 

 metal also was attended by a greater change in the proportion of such 

 corrosions than a variation of kind of liquid. This latter circumstance 

 is in harmony with the view that the potential molecular activity of 

 metals is usually greater than that of electrolytes (see " Relations of 

 Heat to Voltaic and Thermo-Electric Action," " Proc. Roy. Soc," 

 vol. 36, p. 50). 



The effect of amalgamation of the plates upon the proportion of 

 those actions to each other was usually distinct. In ten cases it 

 increased, and in five decreased, the proportion of external electric 

 current to " local action" currents. With tin in dilute nitric acid, it 

 greatly increased the proportion of external current to total corro- 

 sion, whilst with lead in a solution of potassic hydrate it had the 

 opposite effect. It appeared to increase that proportion the most in 

 cases where much hydrogen was evolved, as with zinc in dilute sul- 

 phuric or hydrochloric acid. In seven cases it increased and in nine 



