Polarization of Heat by Magnetism. 483 



and that they act simultaneously upon a system already very 

 distant and almost completely astatic. It might be feared that 

 the electro-magnet, without action on the needle at zero, acted 

 on the needle already displaced by the action of the calorific 

 radiation. This would be possible in fact, if, in its first po- 

 sition, the needle had the same direction as the line which 

 joins its centre to the electro-magnet, and if, when it deviates, 

 it made a notable angle with that direction. In our experi- 

 ments, precisely the inverse condition was realized ; so that 

 the component of the magnetic action diminished more and 

 more during the movement of the needle, and became per- 

 ceptibly null when it attained its greatest deviation. If there- 

 fore it had no action in the first case, such ought for a stronger 

 reason to be the case in the second. 



By means of a commutator the electric current could be 

 made to pass, now in one direction, now in another, through 

 the wires of the electric magnet. We shall designate the two 

 currents by the abridged expressions Current A, Current B. 



The following are the deviations observed : — 



Experiments of September 22. 

 (A Muncke's battery of 50 elements with large surfaces, but 

 already worn, was employed.) • 



First Series. 



Deviations. 



Current A 21-0 



Without current . . . 19*0 



Current A 21 '4 



Without current . . r 18*6 



Second Series. 



(Acid was added.) 



Without current . . . 20*5 



Without current . . . 20*6 



Current B 18-6 



Without current . . . 20-9 



Current A 23-6 



Current B 18'8 



Current A 22-0 



Current B 18 



Without current . . . 19*9 



Third Series. 



Current B 1 7*4 



Currents IT'l 



Current A 19-5 



Without current . . . 18-3 

 2 I 2 



