28 Prof. E. Edlund on the Path of Electrical Induction- and 



may conceive it as follows : — The induction- current in question 

 produces a disintegration of the polar surfaces ; and this gives 

 rise to a disjunction-current which is in the opposite direction to 

 the former disjunction-current, and therefore diminishes the de- 

 flection of the magnetic needle. Now, from experiments 22 to 

 24, we know that the greatest disjunction-current is obtained 

 when the valve has such a position that the discharge which 

 the disintegration causes goes from the disk to the point. Hence 

 it follows that the greatest diminution in the deflection must oc- 

 cur when the induction-current in question goes from the disk 

 to the point, or, what is the same thing, when the discharge goes 

 from the point to the disk. Subsequent experiments show, how- 

 ever, that the fact is quite the reverse, and therefore that the 

 assumption, that that induction-current which is opposite in di- 

 rection to the discharge is the one which traverses the spark most 

 readily, cannot be correct. In these experiments the brass disk 

 previously mentioned was fastened to one rod of the glass cylin- 

 der, and the platinum wire surrounded by a glass tube was 

 firmly screwed to the other rod. The pressure of air in the 

 glass cylinder was 1 atmosphere. 



Experiment 27. — The valve was first of all so arranged that the 

 positive discharge of the machine went from the wire to the disk. 

 There was thus obtained : — 



Without indue- Coil between Without 

 tion-coil. e and Jc. coil. 



Mean deflections . . 37'1 6*2 37*6 



The valve was then reversed^ so that the discharge went from 

 the disk to the wire : — 



Mean deflections . . 47*4 8*5 46'2 



In the former case, then, the induction produces a diminution 



/37'l4-37'6 \ 



in the deflection of 31*2= ( - 6*2 ); andin the latter 



case, of 38*3 divisions. Hence the diminution in the deflection 

 in the first case was not greater, but less than in the latter. 



Experiment 28. — This experiment was in all respects like the 

 preceding, except that the induction-coil was inserted between e 

 and g. The discharge of the machine went first from the wire to 

 the disk : — 



Without Coil between Without 

 coil. e and g. coil. 



Mean deflections . . 38-4 5*9 387 



The valve reversed, so that the discharge went from the disk 

 to the wire : — 



Mean deflections . • 45'0 5*1 48*6 



