86 abstracts: physics 



from stray portions of power or lighting currents, and the frequently 

 more sensitive resistance-measuring system is of course in greater danger 

 still. All such trouble is absolutely prevented by an equipotential 

 shield, which is merely a connected system of metal plates, wires, etc., 

 which interposes itself at every point of solid contact between the 

 measuring system and external bodies. This shield need not be, and 

 preferably should not be, "earthed." 



Slight modifications of this shield are also useful in electric furnaces, 

 in the measurements upon power circuits, and within the potentiometer 

 circuit itself. 



These arrangements are easy to install, most of them require no sub- 

 sequent attention, and all are easily tested. W. P. W. 



PHYSICS. — T her mo-elements of 'precision, especially for calorimetry. 

 Walter P. White. J. Am. Chem. Soc, 36: 2292-2313. 1914. 



Inhomogeneity, once a serious foe to precision in thermo-elements, 

 and still often supposed to be such, can without difficulty be rendered 

 practically negligible in copper-constantan thermo-elements used for 

 any precision up to 50 parts per million. Such thermo-elements, ac- 

 cordingly, may, except for imperfect insulation, easily preventable, be 

 free from all appreciable errors other than those (such as incomplete 

 depth of immersion) which are possible with all thermometers. To at- 

 tain this freedom from error the wire used must be tested, and the es- 

 sential though easily satisfied requirements peculiar to a thermo-elec- 

 tric system must be observed. These requirements this paper attempts 

 to consider m detail, and it also describes simple but important details 

 regarding the operations of construction, insulation, inclosure, calibra- 

 tion, etc., of the thermo-elements. 



Constantan wire for thermo-elements has been so far improved that 

 continuous lengths are frequently obtainable which vary (in electro- 

 motive force against copper) less than 0.0002, making sensitive thermo- 

 elements with errors usually less than 20 per million. 



The testing of wire enough for a thermo-element of maximum sensi- 

 tiveness takes but an hour or two, with simple apparatus. 



On account of the ease with which thermo-elements can be constructed, 

 the more sensitive combination of several couples is generally prefer- 

 able to a single couple, even for cruder measurements. W. P. W. 



