V. TEMPERATURE DETERMINATIONS 145 



that 0.1°C. will be equal to 1 cm. of deflection. The accuracy of 

 measurement made by these methods is limited due to the fact that 

 the indicating instruments are not built to give linear displacement 

 throughout their range. Furthermore, the temperature and humid- 

 ity of the laboratory may be factors affecting the calibration of such 

 devices. 



The selection of the particular galvanometer system to be used 

 will depend upon the particular experiments. Insensitive, rugged 

 instruments of the pointer type can be used for large temperature dif- 

 ferences between hot and cold junctions. In ordering a galvanometer 

 for a specific purpose, it is important to specify the resistance of the 

 thermopile in order that the galvanometer may be chosen to give ap- 

 proximately the correct damping. Most of the modern moving-coil 

 instruments have magnetic shunts so that critical damping can be ob- 

 tained for a large range of external resistances — i.e., 10 to 300 ohms. 

 Careful thermal and electrical shielding are required for measure- 

 ments of verj'- small temperature changes. Introduction of the 

 breaker type amplifier (the Perkin-Elmer Corporation) has made it 

 possible to measure voltages conveniently as low as 5 X 10 ~^ v., 

 corresponding to 0.0001 °C. 



Potentiometric Method. Whereas the galvanometric method 

 gives a scale reading in direct proportion to the thermoelectric 

 force flowing through the instrument, the potentiometric method 

 uses a galvanometer or similar instrument as a null indicator for 

 balancing the thermoelectric potential against a known potential 

 in a second circuit, as shown in Figure 3. 



The temperature range, sensitivity, and accuracy of the potentio- 

 metric circuit are determined by the relative value of the various 

 resistances, the magnitude of the current supplied, and the sensitivity 

 of the galvanometer. To obtain a standardized, reproducible cur- 

 rent from the dry cell, the smtch Si is thrown into position 1 and the 

 cun-ent flowing through the galvanometer is adjusted until the voltage 

 drop across a standard cell resistor, AB, equals the voltage supplied 

 by the standard cell. Insertion and use of S2 and a resistance, R, in 

 the circuit is optional and serves to protect the standard cell. If 

 switch S2 is left open while coarse adjustment of current through the 

 standard cell circuit is being made and then closed to short-circuit 

 the resistance R during fine adjustment of the current, the standard 

 cell circuit will not be in danger of passing large amounts of current. 

 Switch Si is then thrown into position 2 and the voltage of the ther- 



