white: thermoelectric methods 319 



PHYSICS. — New modified thermoelectric methods in calorimetry. 

 Walter P. White. Geophysical Laboratory. Communi- 

 cated by A. L. Day. 



In the common calorimetric method, the " ^lethod of IMixtiires," 

 the work of the last five years has shown that the error due to 

 heat loss from the caloruneter into the air, once thought to be 

 an unconquerable foe to accuracj', is, in fact, ordinarily quite 

 negligible. With a proper installation, the attainment of very 

 high precision, to 0.1 per mille or better, requires merely sufficient 

 precision in the temperature measurement. Such precision has 

 been attained for ordinary' systems by using electric thermome- 

 ters. It remains to make the operation of these thermometers 

 as simple and as free as possible from experimental precautions 

 and corrections, to facilitate the attainment of unusual precision 

 where this is desirable, and to take advantage of the facilities 

 which the electrical system offers for making various measure- 

 ments such as could not be attempted with the older mercurj- 

 thermometers. 



For a variety of measurements the thermoelectric thermometer 

 system is especially well adapted, as has already been shown 

 elsewhere. The present paper deals with the securing of certainty 

 and high precision by means of the thermoelement. The essential 

 feature of the methods is to use, around the "cold junction" of 

 the thennoelement, in place of an ice bath, a body at nearly the 

 same temperature as the calorimeter. The smallness of the 

 electrical quantity to be measm-ed then gives to the temperature 

 measurement extraordinary^ precision combined with extraordi- 

 nary ease. These methods, accordingly, while specially advan- 

 tageous for precision in determining very small heat quantities, 

 are in nearly all cases among the best in convenience and cer- 

 tainty. 



Determinations of great delicacy have already been made on 

 this principle, especially in Nernst's laboratory. In these the 

 other body was a twin calorimeter. If the two calorimeters are 

 ahke, the external temperature and its changes affect both alike, 

 and hence have but a slight effect upon the thermoelement read- 



