Changes in DieleGtric Constants and in Volumes. 65 



wood cylinder. Perhaps in view of the recent research by 

 Drude^ that was not a good plan, but these determinations 

 were finished before this last work of Drude came to my 

 notice. The nse of mercury for the inside coating of the 

 Leyden jar was a great convenience, for then the capacity could 

 be adjusted with great nicety. The vacuum tube for the 

 detector was on the Nernstf plan. 



The length of the induced wave in air w^as 71'5^'" and this 

 was the second wave from the exciter, not the one taking in 

 the brass tubes. TJie stationary bridge was always grounded. 



The acetone for standardizing came from Eimer and Amend. 

 ]^o mark. It was changed into the acid sodium sulphite 

 compound and this decomposed by chemically ^^ure normal 

 sodium carbonate. The liquid was dried over calcium chloride 

 for twenty-four hours and distilled from fresh calcium chloride. 

 Its boiling 'point was 56-8° to 57'8° cor., barometer = 763"^"^ 

 at 22°. 



The benzene for standardizing came from Kahlbaum. Marked 

 " Crystall." Tested for thiophene but none found. Not 

 otherwise examined or purified since an accuracy greater than 

 5 per cent, could not be obtained when the dielectric constants 

 were so low as that of benzene. 



The average room temperature was 22°, ranging from 19° 

 to 25°. 



Two sets of three double observations each of the wave 

 lengths were measured, the two sets being made at different 

 times, though with the same solutions. The mean of these 

 two sets was subtracted from the zero point, that is, from the 

 position of the slide when the condenser was replaced by a 

 metallic bridge, and the dielectric constant read off from the 

 calibration curve. The zero point was determined every day 

 and was the average of three observations. Two independent 

 series of observations were made for the calibration curve, 

 one at the commencement of the measurements of the dielec- 

 tric constant and one at the end of all the measurements. 

 These two series were plotted and form the calibration curve. 

 Each series consisted of two sets of three double observations, 

 the two sets being made at different times but with the same 

 solutions. So the curve is the result of 12 observations at 

 each |)oint of plotting. Notwithstanding much care, the curve 

 was not all that could be desired, for at times an observation or 

 a number of them would slip far away from neighboring ones. 



The solvents and solutes have already been described in 

 preceding papers. The solutions were made by placing the 

 solutions, saturated at a higher temperature, in running city 



^ Crude's Ann., ix, 293, 590 (1902). f Wied. Ann., Ix, 303 (1897). 



Am. Jour. Sci.— Fourth Series, Yol. XYI, No. 91.— July, 1903. 



