TRANSACTIONS OF THE SECTIONS. lOl 



nor, in all cases, successful. The cause of the failure of the general processes, where 

 a horse-shoe magnet was employed, seemed to be the production of consecutive 

 poles. The action of a powerful magnet applied to a hard thin bar seemed too local; 

 so that in the passage of its two nearly contiguous poles, a kind of magnetic wave is 

 raised, highest under the magnet, which leaves behind it, probably, like the passage 

 of a ship or boat, a series of other waves of diminishing altitude. 



To remedy this supposed defect in the ordinary processes, I placed two pairs of 

 thin horse-shoe shaped magnets upon each other, each pair arranged in the form, 

 nearly, of the figure of 00, with converse poles in contact. The arrangement was such, 

 that whilst the two bars of each series or stratum, as laid on the table, had their 

 converse, or mutually attracting poles in contact, the two series had correspondent 

 polarities laid on each other. The an- 

 nexed figure represents the arrange- ^_^^ 

 ment and the position of the operating u\) 

 magnet nearly at the commencement of J^^J 

 the process. The compound or operating /^^^n\ 

 magnet is placed on the upper surface |[[j f^v!\ ) 

 of the upper pair of bars at the curve w\\ \ 11)11 1 

 with its N. pole towards the S. and the Wiw/// 



S. towards the N. of the bars to be mag- ^-t:rT^:;jt~-.„^_^ --Wci}th^:>^ 



netized. It is then slid gradually for- /r^^^^^^^^^^^^^;:— =^§?\\ 



ward, S. pole in advance, towards the |^^\\ ---^^^3^^5?~-~~--^ ^J^-^) ^*"*' 



end designed for the N. pole of the bar >^^;i^;- ^$^^ ^^=a^^~£^:^»:;*-jj::^ 

 beneath, and continued across the June- ^*'"** ""^ 



tion of the bars, in the course of the 

 dotted line, keeping the axis of the 



two poles in the central line of the bar, until the magnet comes round to tiie point at 

 which the process commenced; it is then slid off in the direction of the small arrow- 

 shaped mark. The upper pair of bars is then removed and the lower pair turned 

 over: the upper pair being also turned over is replaced on the top, and the process 

 of manipulation, changing also the poles of the operating magnet, is repeated. Two 

 complete circuits being thus made on the two surfaces of the upper pair of bars, 

 developes in the highest possible degree (if the operating magnet be sufficiently ener- 

 getic) the magnetic power of the lower pair of bars, whilst the upper pair is found to 

 be comparatively weak. Before separating either pair of bars, those above must be 

 removed ; and then, if the highest capacity be wished to be determined, a separate 

 conductor should be laid across the two poles of each magnet to sustain the power 

 when thej' are separated.. 



The effect of this process may be advantageously illustrated by giving the powers 

 of the bars of a small five-bar magnet, weighing altogether 2'86 lbs. as magnetized in 

 a single series in the form above figured, and when magnetized (taking the powers of 

 the lower pair) by the process now described. These bars, it should be noted, were 

 of best shear-steel, annealed in oil (after being made quite hard) at a temperature of 

 about 490°; so that the process in the single series was much more effective than in the 

 case of harder bars of similar thickness. The powers of the bars were determined by a 

 spring balance, and those registered are the powers after the removal, at least once, 

 of the iron conductor, so that these may be considered as the permanent powers. In 

 the employment of the double series process of magnetizing, it will be observed, that 

 there must be two bars left, after the magnetizing of three others, whose highest 

 magnetic energy could not be developed, if no additional or subsidiary bars were used. 

 But in this case I employed two bars of a corresponding kind belonging to another 

 magnet for completing the process. If no additional bars are in possession of the 

 magnetizer, then similar bars of iron can be substituted, or, without a spare bar of 

 any kind, one of the two bars employed as the upper series in the previous manipu- 

 lations can be magnetized by placing the other upon it, and an iron conductor across 

 the poles of each bar whilst the manipulations are in progress. There will then 

 remain only a single bar to be magnetized by another process. The following com- 

 parative experiments show the advantage gained by the new double-series process 

 over that of the single series, or of any other method previously in use, for all the 

 modes heretofore described were tried. 



