and Law in Electro-optics. 171 



the screws were worked through small distances till the upper 

 conductor was in close and uniform contact with the slip. The 

 final adjustment -was obtained when the standard slip was 

 barely movable through the whole space between the conduc- 

 tors, sliding with tight friction, but without lifting or drag- 

 ging the upper conductor. The cell was now closed and 

 charged, the electro-optic measurements were made ; the cell 

 was again emptied and opened, the standard slip was again 

 applied, and the value of D was found to be absolutely un^- 

 altered. 



All the pieces being in position as they were left by the 

 preceding experiment, an unreserved slip was thrust in per- 

 manently between the brass floor of the cell and each of the 

 lateral supports; two of the standard slips, held togther as 

 one, were laid between the two conductors ; and the screws 

 were worked backwards and forwards through very small dis- 

 tances, till the double slip could slide without lifting or drag- 

 ging the upper conductor, but with tight friction as formerly. 

 The cell was then closed and charged, and the electro-optic 

 measurements were made for the new value 2 of D, as for- 

 merly for the value 1. It was also seen here, as in the 

 former case, that the faces of the conductors were parallel, 

 and that D had exactly the same value at the end of the expe- 

 riment as at the beginning. 



Distances 3 and 4 were managed in the same way. The 

 necessary time and pains were taken to make all the opera- 

 tions as correct as possible. No further explanation of the 

 method is required, the electrometer and compensator being 

 worked here in the same way exactly as before. The improve- 

 ments mentioned in (17) were retained here of course. 



20. Measurements. — As my time was now running short, I 

 thought it best to restrict the work to one very careful mea- 

 surement, or to two measurements at most, for each value of 

 D. The limits of error of observation in the measurement of 

 Q were drawn as close to each other as possible in every case; 

 and, as far as I could judge, the error lay always well within 

 the range ± 1 ; no observed value of Q could possibly be in 

 error by so much as ±2. 



(1) D=l. One measurement was made ; and the result was 

 as follows : — 



Q.D 2 ~63xl 2 



This experiment was made in very favourable circumstances ; 

 and the result was verified again and again. 



(2) D = 2. Two measurements were made and verified ; and 



