210 MESSRS. W. R BOUSF1ELD AND W. ERIC BOUSFIELD 



3 The Electrical Arrange**-** are shown diagrammatically in fig. ^ 

 M ishe r Ury thermometer-resistance heater. M, is the 2-ohm resistance , of 



The spirals of the resistance M> were immersed in a current of 



W-. - ** ** p A ^l. ampere 



, the seL coils of a watt balance, W 2 the shunt coils of this balance in 

 the watt balance resistance. Where used quantitatively B. indicate. 

 sun resistance from V u to V 2 , amounting to about 500 ohms and Medicates 

 he total series resistance from V to V 2 , for observations in which the watt balance 

 was used so that the resistance W, and the leads and part of the mercury leads had 

 to be deducted to give the net resistance B of the heater. The resistance R was 

 carefully calibrated for different temperatures and was checked from time to time. 

 The main current passed from the point V through the series coils of 

 balance W, and the heater to the point V 



At V, is placed a switch by means of which the current could be shunted from the 

 heater M, through an equivalent idle resistance X, This shunt was arranged so as 

 to be able to turn on the current in the heater at the required instant without 

 disturbing the readings of the ampere and watt balances. 



The ends of the resistance M t were connected up through a key and galvanometer 

 to a battery of nine standard cadmium cells S. By means of an adjustable resistance X 2 

 in the main circuit the current was regulated by hand so as to keep the galvanometer 

 always at zero. The resistance of M! being nearly constant and always accurately 

 known, the total current d passing in the circuit was thus kept nearly constant and 

 always accurately known. A series of thirty secondary cells D across the poles of 

 the dynamo served to steady the dynamo current. The current C passing through 

 the heater is given by the expression C = C,Ro/(Bo+M a ). Whilst the watt balance 

 was in circuit, the temperature of R was noted from time to time and E was thus 

 accurately known. 



It was finally found that the watt balance was not sufficiently accurate, and its 

 use was discontinued. The shunt circuit V W 2 IloV 2 was then disconnected and the 

 coils W, were cut out, leaving everything else the same. The whole current Ci then 

 passed through the mercury heater. In the later experiments without the watt 

 balance M 2 where used quantitatively indicates the resistance of the mercury 

 thermometer-resistance only, including the long tubular mercury leads, part of which 

 had to be deducted from M 3 to give the net resistance R of the heater. The ampere 

 balance was always very useful, as its indications were found to keep accurate to 

 1 in 5000, even in a long run, and though the actual regulation of the current by the 

 resistance X 2 was determined by the standard cells and galvanometer, any fluctuation 

 of the current showed itself at once on the ampere balance and gave warning of the 

 necessity for regulation. 



4. The Primary Standards. Time. The timing of all observations was made by 

 a Dent's ship's chronometer. This was compared at frequent intervals, and found to 



