BUCKINGHAM: STEM CORRECTION OF THERMOMETERS 169 



work the auxiliary stem must be made of the same glass as tin- 

 working stem so that the two may be thermally as well as geo- 

 metrically similar and similarly placed. 



In Mahlke's method the correction is not measured directly 

 as a length but is computed from equation (2). The mean tem- 

 perature, f°, of a length I of the working stem, including the emer- 

 gent part, is measured by a special "fadenthermometer" devised 

 by Mahlke for this purpose. The fadenthermometer has a long, 

 cylindrical, thick-walled bulb of similar dimensions and construc- 

 tion to a portion of the working stem, and is provided with a 

 still finer graduated capillary stem. When in use, it is placed 

 parallel and close to the working stem, with the upper end of its 

 bulb at the level of the meniscus in the working stem. The length 

 I in equation (2) is thus identical with that of the fadenthermom- 

 eter bulb. If the fadenthermometer has been standardized by 

 total immersion, its reading gives the mean temperature, /°, of 

 its bulb and therefore that of the adjacent length I, of the working 

 stem, for use in equation (2). The fadenthermometer may be 

 regarded as an auxiliary stem provided with a magnifying device 

 — the still finer capillary stem — for increasing the sensitiveness 

 and reducing the reading errors. 



The theory of the use of the fadenthermometer presents some 

 difficulties, and the results obtained by it are subject to several 

 small errors which are not easily estimated or determined and 

 which while not of importance in ordinary work, may become so 

 in work of the highest accuracy. In the more extended form of 

 this paper the theory is discussed at some length. In addition, 

 data on the values of a for several thermometer glasses are given 

 in convenient tabular form for practical use, and suggestions are 

 given for the selection of the most suitable fadenthermometer 

 for use in a given piece of work, when several are available. 



Careful examination of the relative merits of Guillaume's 

 and Mahlke's methods leads to the conclusion that the faden- 

 thermometer is not an improvement on the simple auxiliary stem 

 of uniform bore, and that its apparent advantages are illusory. 

 For work of moderate accuracy, either method is satisfactory, 

 though Guillaume's method is usually simpler in practice. For 

 work of the highest attainable accuracy, Guillaume's method is 

 to be preferred. 



