306 It. Sissingh on Kerr's 



It follows from these that 



(j)™ for 4- magnetization = — 4*9, 



= +5-6, 

 ., = — o*y, 



>T P » + 



Up » 



+ 3-9. 



The mirror was then demagnetized again, and the obser- 

 vations gave : — For polarizer perpendicular to plane of 

 incidence, 385 "8 ; analyser parallel to plane of incidence. 

 283*5. These numbers agree exactly with those obtained 

 above. 



The mean values of the rotations for positive and negative 

 magnetization are : — 



6>\ 0':\ 

 7 ip 1 m 



• lp v lu 



Magn. 

 + 



+5-2 

 5-3 



4-7 



+5-2 



Magn. Magn. Magn. 

 - 1 + 



Magn. 

 + 



j 

 Magn. 



Magn. 



+ 



Magn. 



-4-2 4-3-6 

 53 | 3-7 

 4-9 3-9 

 5-6 4-5 



-5-0 4-3-9 



-4-0 



3-5 

 37 



4-5 



-39 



-39 

 4-5 

 4-6 

 3-4 



4-3-95 



3 '75 



34 



44 



+3-9 



-49 

 4-6 

 5-3 

 54 



-5-0 



+ [>•(' 

 4-6 

 51 

 50 



4-5-1 



As the scale-divisions were 1^'= 4-1*30, l a '=— 1*27, we 

 obtain, on reducing these minimum rotations to minutes of 

 angle: — 



<p m . 



' ia 



| * 



ha' 



Magn. 



Magn. 



Magn. 

 + 



Magn . 



! Magn. 

 + 



Magn. 



Magn. 

 + 



Magn. 



4-4-0 



-3-8 



-3-1 i +3-1 1 -3-2 



-30 



44-0 



-4-0 



On looking over the whole series of observations it was seen 

 that the deviations from the mean value amounted in each set 

 to about 4'; so that the final mean values will have a slightly 

 greater accuracy than this. Within the limits of error the 



