Magneto-optic Phenomenon. 



303 



The value of a minute of angle is 1*56 scale-division for the 

 polarizer and 1*66 for the analyser*. 



After the first demagnetization the Mcol prisms were left 

 in their last position. On repeating the process of demag- 

 netization, for the second experiment, it was seen that the 

 dark band had not changed its position. This, together with 

 the agreement of the mean values of the scale-readings, proves 

 that the demagnetization was fully accomplished by the method 

 adopted. The results of a few other sets of experiments were 

 as follows : — 



Demagnetizing 

 Current to 



Pri ncipal 



Positions. 



Minimum Positions. 



Polarizer. 



Analyser. 



Polarizer. 



Analyser. 



+ 002 amp 



144-75 



304-4 







-002 „ 



144-8 



304-3 







+2-3 , 









142-5 



308-0 



-23 „ 







147-3 



300-5 



4-002 , 



145-0 



304-5 







These results verify the above conclusion. By a demag- 

 netization stopping at 2*3 amperes the minimum rotations of 

 the polarizer on account of permanent magnetism were —2*3 

 and + 2'45, of the analyser +3*6, —3*9; or, expressed in 

 minutes, -l'*5 and V% +2'"2 and -2''3. 



These observations were made with an angle of incidence 

 of 51° 22', because at this angle the minimum rotations can 

 be observed with greater accuracy (cf. § 4) , while the influence 

 of permanent magnetism is not too small, its effect being a 

 maximum for an angle of incidence of between 60° and 70° 

 (see §16). 



12. A second proof of the full demagnetization of the 

 mirror is furnished by the agreement between the principal 

 positions of the Mcol prisms obtained with the demagnetized 

 mirror and with a silver mirror respectively. The two obser- 

 vations must be made immediately following each other. The 

 table of the spectrometer carrying the silver mirror comes in 

 the place of the bridge above the spectrometer and carrying 

 the iron mirror. However great be the care taken in this 

 operation to keep the mirror in the proper position, the ad- 

 justment and centering of the mirror have always to be tested, 

 and if necessary a new adjustment must be made. In place 



* In the remainder of the paper these will be denoted by l w '=l-56, 



la' = l-66. ^ * ' 



