REPORT OK EXeCUTIVli COMMITTEE. 1 25 



value is probably nearer 1.433 than 1.434. A similar uncertainty- 

 exists relative to the Weston normal cell. 



To make the proposed determination it is necessary to design and 

 construct some form of electrodynamometer or ampere balance to 

 measure currents in terms of centimeters, grammes, and seconds. 

 The work referred to in 1899 was done with an imperfect instru- 

 ment, but the success attained was such as to warrant the construc- 

 tion of a better electrodynamometer with greater refinements of 

 detail, construction, and measurements. This has been done in 

 conjunction with one of Professor Carhart's colleagues, Prof. George 

 W. Patterson, without whose assistance, particularly in the mathe- 

 matical solution of the electromagnetic action of one coil on another 

 and the resultant torque, the work would have been almost fruitless. 

 They have constructed a large electrodynamometer composed of 

 one stationary and one movable coil. Both coils are wound on 

 cylinders made of plaster of Paris, accurately turned and covered 

 with a thin coating of shellac. The large coil has a winding of 593 

 turns of silk-covered copper wire, occupying a length of about 41 

 cm. , and the cylinder has a diameter of 47 cm. The relation be- 

 tween length and diameter was intended to be as nearly as possible 

 \2) to 2. The same relation holds for the inner suspended coil, 

 which has a diameter slightly over 10 cm. For the suspension both 

 phosphor-bronze and steel wires have been experimented with. The 

 principle of the instrument is the balancing of the torque, due to 

 the electromagnetic action between the two coils against the torque 

 of the suspending wire twisted through 360°. A twist of one com- 

 plete turn was chosen, because mirrors at the two ends of the wire 

 permit a complete turn to be measured with the greatest accuracy 

 by means of two telescopes and scales. 



The couple required to twist the suspending wire through one 

 turn is determined by separate observations on the period of tor- 

 sional vibration with a load whose moment of inertia can be com- 

 puted with great accuracy. The design of the instrument is such 

 that approximately one ampere is required to produce a balance. 



The current thus measured is carried through a standard ohm, 

 and the difference of potential between its terminals is then com- 

 pared with the electromotive force of the standard cell by means of 

 an accurately adjusted Wolff's potentiometer. 



About one hundred standard cells are available for the measure- 

 ment. The chief difficulty encountered up to the present is the 

 elastic fatigue of the suspending wire. In all the wires tested thus 



