752 GALILEO FERRARIS 



derstood what method of experiment I adopted in determining 

 the coefficient of efficiency of tl;ese instruments. That you shoald 

 bave tbus misunderstood ine is not altogetlier surprising for I 

 did not attempt to explain niy luethod of experiment, as the 

 purpose of my report was merely to inforni the company which 

 consulted me what the capability of their instruments really was. 

 I did not , as 1 gather that you assume , determine the mean 

 difference of potential between two points of the circuit and the 

 mean current in the circuit, and multiply these two together to 

 obtain the work done; for I bave for some years — in fact, 

 since I first studied the theory of alternating currents — known 

 that such a method would he erroneous. What I actually did 

 was by a single observation of tlie electrometer to determine the 

 mean value of the product at eacb instant of the current and 

 the difference of potential. This is done by a very ingenious 

 modification of the method of M. Joubert proposed by Professors 

 Ayron and Perry some years ago. The method is as follows. 

 Let A, B , C be three points of a conductor in which an alter- 

 nating current is passing ; let the resistance between A and B 

 be r, and let there be no self-induction or electrostatic capacity 

 between A and B : it is desired to determine the work done 

 by the current. Connect A and B to the two quadrants of the 

 electrometer, C to the needle, then if the quadrant electrometer 

 be in perfect adjustment the force tending to deflect the needle 

 will be proportional to 



. = (.-.)(c-^-±^). 



A,B and C being the potentials at the corresponding points. 



A-\- B 



Now C is the difference of iwtential between C 



2 



A — B 



and a point D midway between A and B . and is the 



r 



current passing in the conductor , hence the force tending to 

 deflect the needle at any instant is proportional io the work being 



