536 Mr. R. H. M. Bosanquet on Electromagnets. 



The most accurate law to be drawn from these results is 

 that, say, between IB = 2000 and % = 15,000 the magnetic 

 resistance does not vary more than about 2 or 3 per cent, for 

 such bars as I have been using. 



It must be remembered that % here refers to the equatorial 

 section of the magnet. Its average value throughout the 

 magnet would be less. 



My object has been to avoid refinements and deal with such 

 conditions as obtain in practical work. Thus I have not 

 touched the question of residual magnetism. I may note, 

 however, with reference to a recent remark by Prof. Ewing*, 

 that Rowland gave an account of the enormous values of resi- 

 dual magnetism in soft-iron rings in the paper before referred 

 to ; and the conditions of a wire or long bar approximate to 

 those of a ring as far as the magnetic resistance is concerned. 

 It is only when this is small that the residual magnetism is 

 largely developed. I have a bar 1 metre long and 1 centim. 

 in diameter, which retains a large charge just as Mr. Ewing 

 describes. 



Mr. Ewing remarks with surprise on the small magnetizing 

 forces needed to produce large effects with his wires. I must 

 point out that, if the bar be taken long enough, any difference 

 of potential, however great, can be represented as a small 

 magnetizing-force, since the force is potential / length. 

 Further, this conception of magnetizing-force tacitly assumes 

 that a long bar takes more potential to magnetize it than a 

 short one. The reverse is actually true within practical limits. 

 So that, by taking the bar long enough, the effect of a given 

 difference of potential may be continually increased, and the 

 corresponding magnetizing-force, as usually expressed, conti- 

 nually diminished up to a point beyond practical limits. 



Synopsis of Formulce. 

 p=L +-37rlO-' oo6 n: 



Then 



/=.650 + -440^. 

 r 



^ __ 4-7T x ampere-turns 



*~ ~10p ' 



Phil. Mag. November 1883, p. 382. 



