202 Dr. E. Bouty on the Magnetization of Steel by Currents. 



The value of c 2 A 2 is equal to the value fixed by the above- 

 mentioned experiments ; while c 1 A 1 is, as in the case of remag- 

 netization, equal to u v 



The corresponding distribution-formula, 



c 2 A 2 a 2 y#2 



-p 2 u 



e 2 +e 



£2? 



2 



■«!a 2 ^i 



e Pii 



-fou 



01* 



e 2 +e' 



^5 



2 



presents a singular peculiarity: it represents the superposition 

 of a long distribution and a shorter inverse distribution. The 

 resulting form of the distribution differs profoundly from that 

 of the ordinary distribution. 



Fig. 4. 



Fig. 5. 



Fig. 6. 



AP>AT. 



AP = AT. 



AP<AT. 



The above curves are intended to exhibit to the eye the 

 principal cases. With suitable values of the force F the dis- 

 tribution-curve cuts the axes of the abscissae, as is seen in fig. 6, 

 and comprises two regions with opposite signs. The bar that 

 carries this distribution presents two consequent points D, 

 symmetrical in relation to its centre 0. 



This consequence of the formulas of the moments is too 

 curious for me not to have sought to illustrate it by direct ex- 

 periments. For this purpose 1 regulated the intensity of the 

 demagnetizing current in such a manner that the action of a 

 bar previously saturated was nil upon a distant point — that is, 

 that the position of equilibrium of my galvanometer- needle was 

 the same, whether the bar was or was not inserted in the helix. 

 In one experiment the saturated bar, 40 centims. in length 

 and 1 centim. in diameter, produced a positive deviation of 

 102 divisions of the scale, and after the cessation of the current 

 the needle deviated further in the same direction 8 divisions. 

 The action during the passage of the current was under 0*25 

 of a division, which is exactly the limit of the errors of read- 



