314 Prof. B. Stewart and Dr. A. Schuster [May 7, 



only reference to the previous treatment of the wire, but depends also 

 upon the magnetism being put on. 



This result is confirmed by experiments made previous to Dec. 12th, 

 in which the key was not taken out at all. For instance, we have on 



Dec. 9th, 



First off. On first effect. Second off. 



+54 ". . + 45 



"We have hitherto only spoken of the first effect obtained after 

 January 7th ; we now come to the average effect. From January 7th 

 to January 27th inclusive, the magnetism was always put on in the 

 direction (1), and the average effect invariably denoted a decrease of 

 resistance when the magnetism was on. 



5. On January 28th the magnetism was reversed ; the effect during this 

 day was very irregular. On January 29th, 30th, 31st, and February 2nd 

 the key was left in until before magnetization. The first effect was now 

 extremely large ; but it was suspected that, during these experiments, the 

 contact of the key was not very good. 



On January 29th the average effect denoted a decrease of resistance, 

 but on January 30th, 31st, February 2nd, 4th, 6th, the average effect 

 denoted an increase of resistance. 



6. From February 6th until February llth the wires were left broken ; 

 on February llth there was a very slight first effect in the direction of 

 increased resistance, and a slight average effect in the direction of de- 

 creased resistance. On February 12th a mercury interrupter was used 

 instead of a metal key, both the wires being broken by it, and its use 

 was continued until February 18th. The interrupter was left in over 

 night and the current was only broken before magnetization, but no 

 first effect was observed. 



From February 19th to February 26th one wire only was broken by 

 the fluid interrupter; nevertheless there was no first effect. 



On February 12th, when the fluid interrupter was first employed, there 

 was a very small average effect in the direction of increased resistance ; 

 but, in all the experiments afterwards, this average effect was in the direc- 

 tion of decreased resistance. The magnetism had been in the direction 

 (2) from January 28th ; but, during the experiment of February 25th, it 

 was reversed and retained in this condition through the experiment of 

 February 26th, without appearing to affect the results. 



7. From these experiments we may perhaps conclude as follows : 

 In the first place, there is a first effect in the direction of increased 



resistance which appears to have reference to three things namely, the 

 previous state of the wire, the solidity of the circuit, and its mag- 

 netization. 



In the second place, we have an average effect, of which the normal state 

 appears to denote a decreased resistance while the magnetism is on, 

 without reference to the direction of the magnetism. 



