Irreciprocal Conduction. 137 



volts and upwards Lad been employed. Measurements of the 

 resistance were also taken by Kohlrauseh's method, this being- 

 done by breaking the circuit of fig. 3 between a and b, and 

 including the arrangement in one of the arms of a Wheat- 

 stone's bridge. It was found very difficult to get anything 

 but a very approximate value of the resistance during the 

 insulating condition, for the passage of the alternate currents 

 apparently tended to destroy the insulation, and caused con- 

 tinual vibration of the needle of the dead-beat galvanometer. 

 This effect was much less when the current was sent in one 

 direction through the primary coil of the induction-coil than 

 in the other. The values of the resistance obtained, however, 

 fully confirmed those obtained by other methods ; so that we 

 may conclude that the insulating condition is due to a re- 

 sistance at the anode. In what follows it will be convenient 

 to regard this as due to a film of badlv-conducting matter. 



c. The Film-forming Current-density. — If we increase the 

 size of the anode a greater current is necessary to produce the 

 insulating condition, hence the current-density at the anode 

 is an important factor in determining the formation of the 

 film (see Table I.). Calling C the minimum current which 

 will form the film and S the whole surface of the anode im- 

 mersed, then C/S = A we shall define as the film-forming 

 current-density. The actual value of A depends upon several 

 conditions, such as will be best understood after an examina- 

 tion of Tables I. and II. 



Examination of the following and similar observations show 

 that A is not perfectly definite in value, but is greatly in- 

 fluenced by the previous treatment of the anode. For example, 

 if the anode has been heated to redness and washed before 

 putting in the current, the film forms with a less value of A 

 than if the anode had been previously used for electro- 

 lysis (see experiments 2 and 3, Table I., and 5, 6, 7, and 

 21, Table II.). Generally speaking, previous currents in 

 either direction cause A to increase, but after the current has 

 been on for some time in the fi direction, a reversal for a 

 short period seems to assist the formation of the film. Ex- 

 periment 6, Table I., shows very markedly the effect of un- 

 clean electrodes. It is noteworthy that, when the current- 

 density in the /3 direction is at first of a value A' greater than 

 A, the rapidity and completeness of the film-formation depends 

 upon A'— A. 



