JOURNAL 



OF THE 



WASHINGTON ACADEMY OF SCIENCES 



Vol. io September 19, 1920 No. 15 



PHYSICS. — Methods of increasing the precision of thermostats. 

 Walter P. White, Geophysical Laboratory, Carnegie 

 Institution of Washington. (Communicated by Robert B. 

 Sosman.)^ 



In the July number of the Physical Review I published the 

 abstract of a paper on the performance of mercury-contact thermo- 

 stat regulators. Further work has brought out additional facts 

 which are of considerable importance where a practical applica- 

 tion is desired. It therefore, seems best to present at once 

 briefly the subject as modified by these facts, without waiting 

 for the final publication. 



The only phenomenon peculiar to the thermostat is the back- 

 lash of the mercury contact; the mercury, descending, clings 

 to the wire and then drops away, so that the temperature of the 

 make is higher than that of the break. The thing of direct im- 

 portance is the corresponding bulb temperature difference, A^b, 

 the temperature change required to move the mercury meniscus 

 from the break to the make position. Since the effect upon 

 this AL'b of a change in the backlash can be completely offset 

 by a corresponding change in the length of the bulb, a knowledge 

 of the magnitude of the backlash is not important in a brief 

 discussion, though desirable for efficient designing. The back- 

 lash can probably be kept as small as iO;u, and the largest prob- 

 able value is only 6 times as great as that. 



The lag also has a profound influence on the constancy of the 

 thermostat temperature. Sligh- has given a formula which, 



1 Received July 17, 1920. 



^ T. S. Sligh, Jr. Some characteristics of the Gouy thermoregulator . Journ. 

 Amer. Chem. Soc. 42: 66. 1920. 



429 



