4 CHARLES G. ROGERS AND ELSIE M. LEWIS. 



use of a highly sensitive galvanometer made by. Leeds and 

 Northrup. This instrument is of the D'Arsonval type and con- 

 structed especially for thermo-couple work. The apparatus 

 included thermostat, thermo-couples, switch, galvanometer and 

 scale. The thermostat used in all the work was provided with 

 fans and motor for keeping the water constantly stirred. The 

 thermo-couples were made of No. 32 double cotton covered 

 Advance wire made by the Driver-Harris company, and No. 36 

 double cotton covered copper wire. Both wires were carefully 

 shellacked before using. In making the couples we followed the 

 method of White. 1 The wires were uncovered for a short distance 

 at the ends, twisted together, dipped into melted resin and then 

 into melted solder. The ends of the wires were handled entirely 

 with grease-free forceps and the wires were protected from 

 physical strain throughout the course of the work. The couples 

 so made were inserted in small glass tubes sealed at one end. 

 The wires were twisted together through the whole length of the 

 tube and the open end of the tube was closed with wax. This 

 protected both wires from moisture. The advance wire between 

 the junctions was enclosed in a small rubber tube extending be- 

 tween the two glass tubes, and around the rubber tube was placed 

 a heavy wall of wool wadding to prevent changes of temperature 

 in this wire. The copper wires of the couples were connected to 

 heavier, well-insulated droplight cord used as leads to the switch 

 and to the galvanometer. Between the lead wire and one of the 

 copper wires of the couple there was inserted about one and a 

 half meters of No. 36 manganin wire to act as a resistance and 

 thus reduce the strength of current from the couple passing into 

 the galvanometer. All junctions were carefully soldered and 

 the manganin wire as well as the junctions were heavily wrapped 

 in wool and placed in a wooden box to provide against changes 

 of temperature. The switch was also protected in a similar way. 

 These precautions were found to be essential as the temperature 

 changes at these points when they were not so protected resulted 

 in wide variations in the readings of the galvanometer. The 

 junctions were kept, so far as possible, at a constant distance of 

 about one half inch from each other. This gave very constant 

 readings. 



1 White, Walter P., Jour. Am. Client. Soc., XXXVI., 1914, Nov., p. 2300. 



