a Dielectric Normal to the Lines of Force. 345 



potassium and red iodide of mercury (n D = 1*69 approximately). 

 The transmitted light was of a deep orange colour. In order 

 to obtain good definition, water from the water-mains was 

 allowed to run through the jacket for a couple of days, and 

 the tube was taken out and rolled at intervals. 



Under the most favourable conditions of sunlight and 

 homogeneity of the liquid, four observations were obtained 

 during the course of a month, in which the sensibility of the 

 apparatus was such that -01° change in the plane of polariza- 

 tion could with certainty be detected. In other observations 

 "02° conld usually be observed. 



In making the observations, the heliostat was adjusted 

 until the image of the sun appeared in the centre of the field 

 of the polarizer, and the analyser was then adjusted until each 

 half of the field appeared of the same intensity. The magnet 

 was then excited and, if any change of intensity occurred, 

 the arm A was shifted by the screw M until the two halves 

 became of the same intensity again. The current was then 

 reversed and the same adjustments made until no effect could 

 be detected when the current was reversed. In this case the 

 mirror E was evidently made parallel to the lines of force. 



By rotating the polarizing combination, the same observa- 

 tions could be made with the plane of polarization making 

 different angles with the lines of force. Observations were 

 made over a range of 90°, but the chief ones were those at 

 45° with the lines of force ; of these four were obtained with 

 a sensibility of , 01°. The current used was generally 9 amperes, 

 but in one case it reached 40 amperes, but this probably did 

 not increase the field by more than 50 per cent. 



In no case could an effect be observed after the neutral 

 position was obtained, indicating that there is no double 

 refraction within the limits of the observations. 



During each experiment a determination of the sensibility 

 was made. This was effected in two different ways. The 

 first was to determine the rotation of the analyser which 

 would just produce a perceptible change in the two halves of 

 the field. The second was to shift the arm A when the 

 current was on and read the micrometer screw M. The pitch 

 of this screw was 1/32 of an inch. 1/40 of a turn of M 

 could certainly be detected during several of the most 

 favourable observations. These two independent methods 

 serve as a check on each other. An iris dinphragm placed 

 in front of the telescope T allowed the aperture to be varied 

 to give the best results. The following schedule illustrates 

 the varying conditions, the sensibility being determined by 

 one or both of the two methods : — 



