on the Phanomena of Magnetic Reaction. 195 



other. On extending this experiment to compound magnets, 

 I was at first surprised to find that the attraction of the or- 

 dinary keeper exceeded that of the interrupted one by con- 

 siderably more than double. These keepers were applied to 

 two compound magnets, the one having three plates, and the 

 other five, of about the same size. The interrupted keeper was 

 attracted nearly the same by each, but the common one was 

 held much more firmly by the magnet having the greater 

 number of plates. This brought to my recollection a fact 

 which liad often appeared very surprising, and for which I 

 could find no explanation. After taking to pieces a compound 

 horseshoe magnet consisting of eight blades, and of consider- 

 able sustaining power, 1 found that the quantity of magnetism 

 in each blade was very trifling, and apparendy for too little to 

 account for the attractive power of the entire magnet. These 

 magnets were each remagnetized to saturation, but the same re- 

 duction in their power took place on replacing them with their 

 similar poles in contact. They were made of very soft steel, 

 and after having their extremities hardened they lost much 

 less by the contact of their similar ends: this improvement was 

 not so perceptible on the application of the keeper to the whole 

 magnet as to the individual parts. The explanation of these 

 facts appears to be as follows : When magnets are placed with 

 their similar poles in contact, they lose a considerable quantity 

 of their magnetic influence. The absolute amount of loss will 

 depend upon the temper of the steel; the harder it is the less 

 it will lose : and also upon the number of magnets ; the more 

 there are the less will each magnet retain. Now, if the inter- 

 rupted keeper be applied to a compound magnet, it will be 

 attracted only with a force equal to the quantity of residual 

 magnetism, which being very small the attraction will be com- 

 paratively feeble, though probably a correct measure of the 

 absolute strength of the magnet. When the common keeper is 

 applied, it will be obvious, from the double source of induction 

 by which it is influenced, that its state of magnetism will exceed 

 that of the weakened magnet, and by reacting upon it increase 

 its power. The magnet having thus acquired an addition of 

 magnetism will now react more forcibly upon the keeper, and 

 the keeper in its turn upon the magnet, and so on, until the 

 resistance from the contact of the similar poles, and the na- 

 ture of the steel, cease to allow of any further induction from 

 this source. The softness of the steel and the number of magnets 

 being circumstances which favour the reaction of the keeper 

 upon the magnet, are facts corroborative of the probability of 

 this explanation, and show that the force with which the com- 

 mon keeper is attracted by a magnet is a fallacious indication 



2C2 



