162 Barker — Thermoelectromotive Forces of Potassium, 



eral cubic centimeters could be introduced with perfect safety 

 into a mass of melted paraffin, and, when melted under it, 

 showed in each case a brilliant metallic surface, which remained 

 untarnished indefinitely. Sodium when thus melted could be 

 simply poured from vessel to vessel without apparent risk ; 

 potassium was transferred when needful by means of a narrow 

 glass tube closed at one end by the finger, into which the metal 

 was drawn by suction. Paraffin oil was first tried, but aban- 

 doned in favor of solid paraffin, as the oil was found not with- 

 out actiou on sodium, especially when the latter was melted ; 

 moreover, the solid paraffin was more convenient and cleanly 

 to handle. 



A tube thus prepared will remain unimpaired indefinitely, 

 the precaution being taken to protect the free ends of the alkali 

 metal by a plug of paraffin a fraction of a centimeter long. 

 The surface of both metals in contact with the walls of the 

 tube is beautifully clean and lustrous, sodium being distinctly 

 crystalline in appearance, while potassium is not crystalline, or 

 obscurely so. 



The sodium and potassium were the commercially obtain- 

 able metals, presumably of fair purity. Cooling curves gave 

 the melting point of the potassium at 63*5°, and a rough spec- 

 troscopic examination showed no impurity other than the 

 expected trace of sodium. The mercury was the pure redis- 

 tilled metal supplied by a reputable firm. 



The purpose of the mercury tube was to afford a compari- 

 son with a metal whose thermoelectric behavior is less subject 

 to variation than is the case with commercial platinum, differ- 

 ent specimens of which have long been known to vary widely 

 in this respect. Mercury was chosen as readily obtainable in 

 the pure state and free from crystalline or other structure, fol- 

 lowing the example of K. Noll* and also as readily introduced 

 into a tube of the form described. 



When placed in position for measurement, the vertical parts 

 of a tube were inserted in wide test tubes containing mercury 

 so that the junctions were deeply immersed. These mercury 

 receptacles were surrounded respectively by water and by 

 melting ice contained in double walled copper vessels. In 

 each test tube a thermometer was also inserted, graduated to 

 - 2°, with the bulb as closely adjacent as possible to the actual 

 platinum-metal junction. The platinum wires were led to thin 

 glass mercury cups floating in a large mass of mercury, and 

 from these cups copper wires were used to connect with the 

 binding posts of the potentiometer. 



The thermometer at the cold junction was found to be indis- 

 pensable, the temperature being always a little above zero, and 

 * Ann. der Physik, 1894. 



