Irreciprocal Conduction. 135 



nearer to E p , the quantity of electricity will be greater if the 

 plate is the anode. 



The theory is thus highly ingenious, but its verification is 

 a matter of considerable difficulty, for we have to distinguish 

 the effects of polarization and chemical transition resistance 

 from the electrical transition resistance. As we have pointed 

 out, Christiani's own results are open to doubt, so that unless 

 some better method of experiment can be suggested the 

 hypothesis of Christiani must remain but an interesting 

 speculation. 



Section B. Descriptive of our Experimental Work. 



1. Arrangement of Apparatus. 



We have found the arrangement shown in fig. 3 very con- 

 venient for the larger part of our experimental work. 



M is a battery consisting usually of from 1 to 6" secondary 

 cells : 



h\ is a commutator for reversing the main current; 



k 2 is a four-way plug-key enabling the voltameter-circuit 

 to be broken or short-circuited, or cut out of the circuit 

 altogether; 



V is the electrolytic cell; 



G is a dead-beat Deprez and D'Arsonval galvanometer; 



& 3 is a commutator used to reverse the current through G; 



k 4 is a switch hinged at E. According as the contacts D and 

 A or those at C and B are made, then (1) the battery is in- 

 cluded or excluded from the voltameter-circuit, (2) the low- 

 resistance galvanometer-shunt S/> or the high resistance 

 galvanometer-shunt Ss is in use, and (3) the resistance-box 

 Rj, or B s is employed. The switch k± enables us, there- 

 fore, to study the changes of current-intensity when the 

 electrolytic cell is in circuit with the battery, and also 

 the changes of the polarization-current which ensue after 

 cutting the battery out of the circuit. 



2. Experiments with strong H 2 S0 4 and Pt Electrodes. 

 a. General description of the Phenomenon to be observed. — 

 In the opening paragraph of this paper it was stated that, 

 through a given voltameter, consisting of Pt electrodes in 

 strong H 2 S0 4 , currents of an intensity below a certain maxi- 

 mum could alone pass, and that if we tried to increase the 

 current strength above this maximum, it was found that the 

 deflection on a galvanometer in circuit diminished nearly to 

 zero, and that decomposition nearly ceased. Further inves- 

 tigation soon showed that the -ize of the electrodes was also 



