MAGNET STEELS AND PERMANENT MAGNETS 393 



is because, compared with straight magnets, their effective lengths 

 are greater than their actual lengths. The difference between effective 

 and actual lengths is negligible, however, except for magnets whose 

 poles are brought quite close together, either actually or by means 

 of pole pieces of soft iron. Watt-hour meter damping magnets are 

 good examples of magnets whose poles are brought thus closely to- 

 gether without the aid of pole pieces. 



On the other hand, the general conclusions drawn from the curve of 

 Fig. 6 are applicable to magnets of any shape if L is assumed to stand 

 for the effective length of the magnet. Fig. 7 also is perfectly general 

 on the same basis. By effective length is meant the length of a straight 

 bar magnet of equal cross-section and magnetic properties, having the 

 same remanance as the given magnet. 



Correlation Between Straight and Bent Magnets 



Subsequent to the establishment of the relationships shown in Fig. 

 6, a limited amount of data were secured which have a bearing upon 

 the effect on Brem of departure from the form of a straight bar magnet. 



To get these data, one H by l^i in. rod each of cobalt steel and 

 tungsten steel were formed into rings of 2-in. inside diameter, with 

 the ends touching. Companion test bars were cut adjacent to each 

 end of the rods used in forming the rings. All three pieces of each 

 kind of steel received as nearly identical heat treatments as possible, 

 even to heating and cooling the straight test bars at the time of forming 

 the rings. After hardening, demagnetization curves were obtained on 

 each of the two test bars of tungsten steel and of cobalt steel. Then 

 data for Brem vs. length curves were obtained from both the test bars 

 and rings by the process of cutting off the ends and remeasuring Brem 

 already described. 



The demagnetization curves and the Brem vs. length curves, for the 

 two straight cobalt steel bars are practically identical. Inasmuch as 

 these bars were cut adjacent to the ends of the bar from which the ring 

 was formed, and heat treated with it, it is safe to assume that the 

 material in all three samples is alike magnetically, and that the demag- 

 netization curve for the ring is the same as for the straight bars. The 

 same conclusions hold for the tungsten steel, as shown in Fig. 9. 



It follows then that it is valid to compare the Brem vs. length curves 

 for the straight bars with those for the rings, and conclude that all 

 differences in the curves are due to differences in shape. It is in- 

 teresting to note that the ring for cobalt steel yields the same curve 

 for Brem VS. length as the straight bar, but that, in the case of the 

 tungsten steel, the proximity of the ends of the ring has an appreciable 



