Light at certain Metal- Liquid Surfaces. 243 



Compensator left or right means that the first fringe to 

 the left or right of the central one was under the cross-wires. 

 On two occasions in the course of the investigation it became 

 necessary to take the compensator to pieces, once to clean it 

 and the second time to replace the cross-wires. On each 

 occasion it was re-calibrated, and both times with the same 

 result within the probable error of the determinations. The 

 constant of the compensator, i. e. the factor by which its 

 readings must be multiplied in order to reduce them to 

 degrees, deduced from the above readings is 40*11. The 

 position of the central fringe (zero point of the compensator) 

 was found to vary slightly with temperature, but otherwise 

 was constant. The maximum variation observed was 0*035 

 turns. The scale on the compensator is so placed that its 

 direct readings yield the supplement of the angle of phase- 

 difference rather than the angle itself. 



The observations of the phase difference and the angle of 

 restored polarization were made in the following manner in 

 every case :— (1) With both the telescope and collimator 

 arms level and the surface-cleaning apparatus lowered out 

 of the way, the zero point of the compensator was determined 

 as the mean of at least twelve settings taken coming up 

 from both directions. (2) With the two arms set to give 

 an angle of incidence = 60°, the mirror was brought into 

 position, adjusted and cleaned in the manner which has been 

 described. (3) With the polarizer set to give light vibra- 

 ting in a plane inclined at an angle of 45° to the plane of 

 incidence, the displaced position of the central fringe was 

 observed (again as the mean of at least twelve settings from 

 each direction), and the position of the analyser making the 

 fringes blackest was determined. For this last at & least 

 twenty-four settings were made in each case. (4) With 

 the polarizer set to give light vibrating in the other 45° 

 plane, the compensator and analyser positions were again 

 determined and with the same number of readings for each. 

 (5) Finally, with the arms brought back into line, the zero 

 point was again determined as in (1), so as to be able to take 

 account of ^ any temperature change in the compensator. 

 The reflecting surface was renewed, of course, as often as 

 might be necessary in the course of a set of observations. 



The phase difference is obtained from the difference of 

 the compensator readings of (3) and (4) and those of (1) 

 and (5). The double angle of restored polarization is the 

 difference of the positions of the analyser in (3) and (4). 



There were made in all nine complete sets for A and 2^ 

 as described above : four in air, three in water, and one each 



R2 



