206 GENERAL DISCUSSION OF RESULTS. 



aqueous solvents lies in improving the conductivity method until it can be used for 

 this purpose. With this idea in mind, better forms of induction coils were con- 

 structed, and longer and more sensitive bridges emplo3 T ed. The rheostats were of 

 high grade and very carefully calibrated. More important, however, than any of 

 these improvements are those in the conductivity cells themselves. A form of cell 

 which had a constant less than three units, was devised. This was secured by using 

 platinum electrodes with large surfaces and placing these close together. The form, 

 as already stated, was concentric platinum cylinders. It was necessary to take 

 special precautions to purify the solvents, since work must be done in them in very 

 dilute solutions. Kreider extended his work also to mixed solvents mixtures of 

 methyl and ethyl alcohols with water. The salts studied in pure methyl and pure 

 ethyl alcohols are potassium iodide, ammonium bromide, potassium sulphocyanate, 

 lithium nitrate, sodium iodide, calcium nitrate, cobalt chloride, and copper chloride. 



Maxima in the molecular conductivities were found, the dilutions studied ranging, 

 in general, from v = 1,000 to v = 50,000. /j.^ was usually reached at a greater con- 

 centration at than at 25. This is what would be expected, since dissociation 

 diminishes with rise in temperature. In some cases, however, the maxima occur at 

 the same concentration at both temperatures. 



A relation was found between the values of // in methyl and in ethyl alcohols, 

 which is of interest. Take any given salt and divide the value of ju M for that salt 

 in methyl alcohol by the value of /loo for that salt in ethyl alcohol, and the result is 

 nearly a constant. 



/Zoo methyl alcohol 



Moo ethyl alcohol 



= constant. 



The value of this constant for methyl and ethyl alcohols is very nearly 2.37. 

 For cobalt chloride the constant came out 3.68, which is just about one and a half 

 times 2.37. Jones 1 had previously determined the dissociation of certain salts in 

 methyl and ethyl alcohols by the boiling-point method as improved by himself. If 

 we divide the dissociation in methyl alcohol by that in ethyl alcohol, we obtain an 

 average ratio which is very close to 2.37. 



Dissociation as measured by conductivity in this work is greater than as found by 

 the boiling-point method. This may be due, in part, to the polymerization of the 

 undissociated molecules in the solvent in question. With the improved conductivity 

 method used in this work it was possible in most cases to reach the true value of 

 fx^ for the dissolved substances; and, therefore, the conductivity method could be 

 used to measure dissociation in some non-aqueous solvents. 



The second investigation carried out by Dr. Kreider was an extension of the first 

 to the compounds sodium bromide, lithium bromide, and cobalt bromide. Better 

 methyl alcohol was obtained in this than in the former work , and dilutions as great 

 as 100,000 liters were studied. 



The relation earlier discovered between /jl^ in methyl alcohol and in ethyl alcohol 

 was confirmed by this second investigation. When a salt is completely dissociated 

 in each of two solvents, we have for the same concentration the same number of 



iZeit. phy8. Chem., 31, 114 (1899). 



