ELECTROMOTIVE FORCES OF THE ELECTRIC ARC. 319 



the carbons together. The testing current was adjusted to its working value, in 

 most cases O'l ampere, as read on T, and the resistance p t in series with M was 

 adjusted until V B gave a deflection of 100 scale divisions on M. The reason that this 

 adjustment was made without the direct current flowing was that it formed a check 

 on the satisfactory working of the measuring circuit, since any apparent change in 

 the sensibility of M when the direct current was re-established would have indicated 

 an error somewhere. 



With solid carbons the positive or upper electrode was adjusted to project G centims. 

 from its holder and the lower or negative 4 centims. With cored carbons these 

 lengths were 7 and 4'5 centims. respectively. The object of adjusting these lengths 

 was to make the mean resistance and self-induction of the loop formed by the frame 

 of the arc lamp and carbons as nearly as possible the same in every experiment. 



The arc was now re-started, and the length and direct current having been adjusted, 

 the carbons were fed together as they burnt away, so as to keep the direct current 

 constant during the whole of the time (about half an hour) that V A , V R , and V were 

 being determined. This kept the P.D. arc constant, as long as the P.I), of the 

 battery remained constant ; any slight drop in this P.I), was compensated for by 

 adjusting p lf any considerable drop necessitated recommencing the experiment. The 

 arc length with solid carbons also remained constant ; but with cored carbons it 

 constantly varied about a mean value according to the amount of material from the 

 core present in the arc. In all cases, readings were oidy taken when the length was 

 observed to be correct as well as the direct P.D. and current. 



As soon as about 5- millims. had burnt off the end of the positive carbon, the 

 deflections of M corresponding to V A , V R , and V were observed in turn, the zero of M 

 being taken after each reading, until in most cases five consecutive sets were obtained 

 which were reasonably consistent with one another, the li.M.S. value of the testing 

 current as read by T and its frequency being kept constant. It was easy to obtain 

 individual deflections corresponding with V K which differed from the mean by less 

 than 0'3 per cent. The deflections corresponding with V A and V were not so definite, 

 V A being within 1 per cent, and V within O'G per cent, of the mean, except in a few 

 exceptionally unsteady arcs, such as long-cored arcs and small-current arcs. 



The values of the direct P.D. arc lamp and direct P.D. R were noted, and the 

 drop of volts in the frame of the lamp and carbons was found by pressing the carbons 

 together, the direct current being so adjusted that when the carbons were in good 

 contact its value was that used for the experiment. This observation was repeated 

 until consistent results were obtained with the carbons hot as in use. By deducting 

 this value from P.D. arc lamp, P.D. arc was obtained. 



The relative calibration of the thermo-galvanometer M was then determined by 

 means of direct currents. This completed the observations required for a single 

 experiment. 



Mean deflections corresponding with V A , V B , and V, were calculated and corrected 



