1874.] on a Magnetized Copper Wire. 313 



cases the first effect of putting on the magnetism was a marked increase 

 of resistance. For instance, we have 



Date. First off. On first effect. Second off. 



Dec. 16 +36 on (2) +3 



17 +34on(l) +17 



18 ...... ........ + 54on(l) ........ +24 



19 +33on(l) .. .. -18 



It was soon seen that this first effect had some reference to the time 

 elapsing since the last experiments were made. For instance, in the 

 above Table, we see for December 18th a marked increase of resistance 

 when the magnet was first put on; but, on the afternoon of that day, the 

 experiments were repeated, and there was no apparent increase of 

 resistance in ibis first effect. Next, with regard to the average effect : on 

 Dec. 16th, 17th, and 18th, this average effect of magnetism was a decrease 

 of resistance; but on Dec. 19th there was an apparent increase of resist- 

 ance when the magnetism was on. We cannot say that nothing had 

 been done to the arrangement between the 18th and 19th of December 

 that might account for this change ; but whatever was done must have 

 escaped our recollection. Undoubtedly a good many experiments were 

 made during the time between the 15th and 19th of December, and the 

 direction of the magnetism was frequently changed. This curious 

 anomaly, occurring unexpectedly, induced us to limit our future experi- 

 ments to a definite set each day. 



4. The experiments were resumed on January 7th, the arrange- 

 ment having remained untouched during the holidays. From this date 

 until January 10th inclusive, the key was taken out before beginning 

 experiments in the morning : there was no peculiar first effect ; while, on 

 the other hand, an average effect denoting a decrease of resistance came 

 out very prominently. On January 12th and 13th the key was only 

 taken out before magnetizing, and on these occasions the first effect, 

 denoting increased resistance, was sufficiently marked. 



Our method of procedure was varied in the above manner up to 

 January 27th ; and it was invariably found that, whenever the key was 

 taken out before commencing experiments, there was no first effect; but 

 when it was kept in until before magnetizing, this first effect was suffi- 

 ciently marked. These experiments concur in proving that the first 

 effect has" some reference to the previous treatment of the wire ; but they 

 do not prove that it is at the same time connected with the putting on 

 of the magnetism. To determine this point we made a set of experi- 

 ments on January 22nd, 26th, and 27th. When the current had become 

 constant the key was taken out, but the magnetism was not put on ; and 

 on these occasions there was no first effect of the current upon itself in the 

 direction of increased resistance, but rather in the opposite direction. It 

 thus appears that the first effect which increases the resistance has not 



