10(5 MR F. E. SMITH ON THE CONSTRUCTION OF 



A Determination of the Temperature Coefficient of the Resistance of 



1. Mercury in Jena 16'" ylass. 



2. Mercury in Verre dur glass. 



3. A constant volume of mercury. 



Two very excellent series of observations on the Temperature Coefficient of 

 Mercury have been made in the past by GUILLAUME,* and by KREICHGAUER and 

 jAGER.t On p. 112 it is shown that no very great difference exists between the two 

 results. Nevertheless, the arrangements already dealt with for the resistance 

 measurements of the standards were so convenient for such a determination that a 

 series of observations (the temperature ranging from C. to 22 C.) were made. 



Owing to its high temperature coefficient, the resistance of a mercury column 

 increases approximately 1 per cent, for an increment of temperature of 10 C., so that 

 in conducting the measurements some difficulty arises if it be desired to make all the 

 comparisons with one standard coil. The Kelvin bridge is, however, particularly 

 suited for such measurements. In the case previously dealt with, the ratio arms 

 R and S (see fig. 8) consist of 1000-ohm coils. The shunt necessary to alter the 

 value of QR/S by 2 per cent, is, therefore, not small, and need only be approximately 

 known. (This assumes that the shunted coil is either R or S.) Also the variation 

 of resistance of the copper leads, owing to varying temperature, is known, since the 

 resistance of the leads is measured immediately after that of the standard. At any 

 temperature, therefore, the same degree of accuracy as before can be attained, even 

 though the extreme measurements vary by 3 or 4 per cent. 



The maintenance of the temperature of the standard so as to remain constant to 

 0'01 C. for an hour or more, is an apparent difficulty only. Formerly, the tube was 

 immersed in ice, and an interval of 5 hours was considered necessary before a 

 measurement could Ije recorded. The conditions now, however, are changed. Liquid 

 contact is secured, and this, assisted by violent agitation of the liquid, promotes 

 uniformity of temperature. The interval of constancy therefore essential in order 

 that the temperature of the mercury may be regarded as identical with that of the 

 bath is much smaller than before. If the temperature can be regulated so as not to 

 vary by more than 0'005 C. in an hour, and also, if this temperature can be 

 accurately read, then the conditions may be regarded as ideal. Such ideal conditions 

 were practically realised, as will afterwards be shown. 



The most accurate oKservations of a number at different temperatures will 

 certainly be those at the ice-point, since not only is the temperature steady, and 

 absolutely known, but, in addition, the observations can be made with more leisure. 

 The mercury columns chosen were therefore such that at the high temperatures 



* M. C. E. GWIXAUME, 'Bureau International des Folds et Mesures,' 1889-1890. 

 t ' WIEDEMANX'S Annalen.' vol. 47, 1892. 



