21 G JONES AND SCHUMB. 



The lowest curve shows the results of Sucheni on the potential of 

 thallium amalgam electrodes at 0° (measured against the decinormal 

 calomel electrode). The sharp break in the curve occurs very close 

 to the point on the horizontal axis indicated by the composition of the 

 compoimd, Tl2Hg5. The evidence cited strongly supports the follow- 

 ing conclusions: (a) mercury and thallium form a solid compound 

 TUHgs (not TlHgo) ; (b) tliis compound can dissolve excess of either 

 mercury or thallium within definite limits; (c) solid thallium forms a 

 solid solution with mercury up to a limit roughly indicated to be 18% 

 of mercury; (d) the eutectic mixture which is stable at 0°C. cannot 

 contain pure solid thallium as one phase; (e) a two-phase solid amal- 

 gam would not be expected to have the same potential as pure metallic 

 thallium. 



In our investigation we determined to avoid the use of amalgam 

 electrodes, if metallic electrodes could be obtained which would be 

 constant, reproducible, and reversible. Although the metallic elec- 

 trodes first prepared did not meet these tests satisfactorily, the diffi- 

 culty was traced to atmospheric oxidation. After apparatus was 

 developed which permitted the rigid exclusion of atmospheric oxygen 

 from all contact with electrodes or solution the metallic electrodes 

 pi'oved to be satisfactory. INIany of them were measured at intervals 

 for several days and did not vary more than 0.2 millivolt. Each of 

 the measurements recorded below refer to an independent experiment. 

 Between each experiment the cells were completely dismantled and 

 set up again with fresh thallium electrodes, new solutions, and new 

 calomel electrodes. As will be seen, variations of more than 0.2 milli- 

 volt are rare. Most of the cells contained two thallium electrodes 

 side by side in the same solution. This arrangement made it possible 

 to pass a small current between the two electrodes, thus polarizing 

 one of them cathodicalh- and the other anodicallv, therebv causing a 

 difference of potential of several millivolts between them. This 

 polarization, however, disappeared completely on standing or on 

 rinsing the electrodes with a fresh solution, thus showing the electrodes 

 to be reversible. In our first series of experiments no amalgam 

 electrodes were used but two metallic thallium electrodes in contact 

 with a saturated solution of thallous chloride were measured against 

 a tenth-normal calomel electrode. These cells gave a value for the 

 normal potential 2.6 millivolts higher than the result of Lewis and 

 von Ende. We, therefore, determined to make a direct comparison 

 of a metallic electrode and an amalgam electrode prepared according 

 to the method of Lewis and von Ende. As will be seen in the table 



