SOME MKKCURY STANDARDS OF RESISTANCE, ETC. 



97 



made, nor is it likely that any theoretical computation (for the particular connecting 

 vessels riiipluyril) \\milil ! si. ;n-iMir;itr ;is an c\ | n-riiiiiMit ;il (lftfnniii;it ii .n. 



Throughout the experiments, the distance hetweeu the plane ends of the standard 

 and the potential points was maintained constant. It was equal to 1'5 centims. 

 The constancy of this distance was of some importance, since the resistance of a 

 column of mercury 1 centim. long and of the cross-section of the connecting vessels 

 is 0'00002G ohm (cross-section of connecting vessel = 4 sq. centims.). Now, 

 although the value f = 0*82 is from a theoretical standpoint a little too great for the 

 conditions imposed, it is sufficiently near to the true value to indicate that the 

 difference of the resistance measurements, when erection Methods I. and II. are 

 employed, is of the order 1 per cent, for unit tubes. If, then, the resistance 

 measurements in the two cases are accurate to O'OOl per cent, only, the value of /can 

 only be calculated with an accuracy of 1 per cent. To attain greater accuracy, the 

 following method was adopted : 



One of the tubes (H, Jena 16'"), originally intended for a mercury standard, 

 became of little value owing to the development of a flaw near one end. Its cross- 

 section was, however, accurately known, and the calibration curve had been plotted. 

 At the point where the flaw had developed, the tube was broken, and the new end 

 ground, so as to avoid irregularities. It was then filled with all the care pertaining 

 to a standard, and its resistance measured by the Kelvin double bridge and the 

 potentiometer, the mode of erection being necessarily Method II. The value so 

 obtained was 0'95109 a ohm. The length of the tube was 98'9 centims. and the 

 mean cross-section 0'9792 sq. millim. 



After removal from the ice bath and subsequent cleaning, the tube was cut into 

 five lengths, these being respectively 10, 25, 22, 29, and 13 centims. The corre- 

 sponding points of fracture were marked on the calibration curve and the cross- 

 sections at these points calculated therefrom. The divided tube was then re-erected, 

 four glass vessels serving to connect the five portions, and a distance of 3 centims. 

 maintained between any two. Thus the total " end corrections " were now con- 

 siderable, and, since the junctions were all similar and the cross-sections well known, 

 a fairly accurate evaluation of f was possible. The resistance at 0'0 C. was now 

 0'95450 3 ohm, showing a difference of 0-00341! ohm from the first measurement. 



The mean cross-section of the whole tube was, as above stated, equal to 0'9792 sq. 

 millim. 



The cross-sections and radii of the new ends introduced were : 



VOL. CCIV. A. 



