THKUMAL CONDUCTIVITIES OP SOMK KLICCTKICAL IXSUL.VI 



187 



The upper end of each spiral was soldered to a No. 20 bare copper wire enclosed in 

 a thin glass tube in contact with the heating tubes. Each copjter wire had a total 

 length of 1)2 centims. 



To give strength to the arrangement, the tubes passed through two short plugs of 

 plaster of Paris, 10 and 22 centims. respectively from the lower ends of the heating 

 tubes. 



The material to be tested was fused in a brass tube D, 8*4 centime long, 2 047 

 centims. external, 1*950 centims. internal diameter, with a base and lid 

 of the same thickness. 



While the material was in the molten state, the testing apparatus 

 AiAjBtBg was slowly lowered into the tube till the platinoid wires of 

 the heating coils occupied the middle two-thirds of the length of the 

 brass tube. The heating tubes were kept in symmetrical positions 

 laterally by two thin rods of wood of length equal to the internal 

 diameter of the brass tube, which were attached to the tubes and 

 projected equally from the two on each side. Through each wooden rod 

 there passed at right angles a thin brass wire of length equal to the 

 inside diameter of the brass tube. 



The apparatus could thus be removed and replaced in the brass tube 

 in the same symmetrical position. 



The brass tube and its contents were placed near the bottom of a 

 vertical straight Dewar tube E (fig. 2), 4'5 centims. internal diameter 

 and 24 centims. internal length, and were held in place in the centre 

 by three wires which extended horizontally from the brass tube to the 

 surface of the Dewar tube. 



Fig. 2. 



The top of the Dewar tube was closed by a plug of cotton-wool, F (fig. 2), through 

 which the conducting wires and their surrounding glass tubes passed to the cells and 

 galvanometer. 



The electric current for the heaters was supplied by three storage cells, and ]iassed 

 through regulating resistances (fig. 3) which could be adjusted by means of merrmy 

 cup connections to '01 ohm, and a Weston ammeter which had been carefully 

 compared with a Kelvin current Iwilance 

 standardized by copper deposition. The 

 potential difference at the ends of the 

 copper leads attached to the heating coils 



Fig. 3. 



was measured l>v a Keiser and Schmidt moving-coil voltmeter, standardized by the 

 potentiometer method. A correction for the resistance of these leads was determined 

 and applied to the readings thus obtained. 



The bridge circuit for determining the resistance of one and the difference of 

 resistance of the two platinum wire spirals was supplied with an indejxjudent current 

 from a small Leclanchd cell (fig. 4) with 53 ohms resistance in series with it. 



