BLACKWELL. — DISPERSION IN ELECTRIC DOUBLE REFRACTION. 659 



fifty volts per coil in use, and the resistance was measured with only- 

 six volts, it was an open question whether the increased leakage at the 

 higher voltage between such an enormous number of adjacent turns 

 would be negligible. To settle this question, a d'Arsonval galvanometer 

 was calibrated with a known resistance box and standard cell to read 

 current. The current through each coil of a set of nine was then 

 measured with six volts per coil and with one hundred and fifty. 

 Within the limit of error of the observations, one tenth of one per cent, 

 the current proved proportional to the voltage. 



Static Voltmeter. — With this resistance and potentiometer satisfactory 

 measurements of potential up to fifteen thousand volts were obtainable, 

 but another difficulty presented itself. The electromotive force of the 

 battery with the potentiometer connected falls perceptibly in a few sec- 

 onds, even with the small discharge rate of fifteen milliamperes. An 

 auxiliary static voltmeter calibrated immediately before and after using 

 is necessary. A very simple device has proved exceedingly sensitive. 

 Between two parallel plates of brass about fifteen centimeters square a 

 silvered glass disc is hung, four or five centimeters in diameter, firmly 

 cemented to a small, straight, steel clock spring, thus probably securing 

 nearly perfect elasticity in the moving parts within the limits of defor- 

 mation. The whole is enclosed in wood and glass. The spring and one 

 plate are connected to one terminal and the other plate to the other 

 terminal of the high voltage line, and deflections can be read by a tele- 

 scope and micrometer eyepiece to one tenth of one per cent. Moreover, 

 successive calibrations agree to this degree of accuracy. 



Purification of CS 2 . 



Though small quantities of impurities often do not preceptibly influ- 

 ence the behavior of a substance, it is more satisfactory to work with a 

 pure substance, or definite mixture. Pure carbon bisulphide can be 

 ordered, but is said to deteriorate in a short time by the formation of 

 more complex sulphides. 19 Kerr, Quincke, and Schmidt all lay stress 

 on the need of freeing the carbon bisulphide from solid particles in order 

 to prevent sparks. It is possible also that traces of dissolved water 

 lower the obtainable difference of potential. Accordingly the following 

 method of purifying and drying CS 2 was used. 



(1) To commercial carbon bisulphide add quicklime, stand a day, 

 digest four hours (with return condenser), distil. 



19 Bertliellot, A. Ch., (7), 14, 155; 19, 150. 



