66 Speyers — Heat of a Change in Conneotion loith 



water and letting tliem stand over night. That gave plenty 

 of time for proper separation and the temperature of the run- 

 ning city water w^as very constant, ranging from 22*0° to 22*8° 

 cor. An ordinary thermostat could not liave been used below 

 30° on account of the hot summer weather, and then would have 

 come the trouble with crystallization when the solutions were 

 cooled to the room temperature to go into the condenser. 



The dielectric constants for the solid solntes were obtained 

 by melting and letting them solidify in the condenser. When 

 the series of measurements with the solid solutes was finished, 

 the condenser was filled with pure acetone and the dielectric 

 constant measured again. No change in value could be detected, 

 showing no change in the position of the condenser plates. 



The conductivity of the solutions was found to be so low 

 that there was no danger of interference from that source. 

 The highest conductivity was that of acetamid in water = 

 O'OOIS coulombs per sec. through 1^=^ at 1 volt p. d.^ while the 

 lowest was < 0*062 coulombs per sec. through l^"^ at 1 volt p. d. 

 and this was a solution in toluene. 



The dielectric constants are given in the large table 

 together with their estimated uncertainties. Also the func- 

 tion 1 — 9/(K-f-2y with its "mean uncertainty." This "mean 

 uncertainty" is found by computing the function when the 

 estimated uncertainty is added to K and when it is subtracted 

 from K, which two values will be quite unequally distant from 

 1 — 9(K+2)^ wdien K is small, and taking the mean of these 

 differences from the function when the experimentally observed 

 value of K is used. When K is greater than 16, the value of 

 the function is about the same whether the uncertainty^ in K is 

 added to K or subtracted from it. 



The densities of the solutes, solvents, and solutions are 

 needed to find their A^olumes. The densities of the last two 

 have already been determined,^ but not the densities of all 

 the solutes. Landolt and Bornstein quote some data from 

 Schroeder and others, but the trouble which Schroeder had 

 with air bubbles made his results a little questionable and a 

 redetermination of them seemed desirable, while at the same 

 time the necessary new data could be obtained. 



The densities were found by w^eighing in saturated solutions 

 of kerosene or of amyl alcohol, according to the nature of the 

 solid. Kerosene was excellent ; only naphthalene, acenaph- 

 thene, and phenanthrene were too soluble in it ; for these, 

 amyl alcohol was used. The air bubbles with kerosene were 

 insignificant, but with amyl alcohol they were sometimes quite 

 annoying and could not be altogether removed, but those left 

 are not believed to have influenced the second decimal place at 

 * This Journal, xiv, 293 (1902). 



