OP ARTS AND SCIENCES. 229 



for the filed carbons to shorten by any material amount. Before 

 each set of measurements the carbons after filing were screwed into 

 contact, the reading of the micrometer taken, and the carbons were 

 then separated by the desired amount. It was found, however, that 

 the expansion of the carbons by heat was a noticeable quantity, and 

 hence, except with very long arcs, the arc was allowed to form, and 

 then the carbons were quickly screwed into contact and again sepa- 

 rated by the desired amount. For arcs so long that this was not 

 practicable, the expansion was estimated and allowed for. The change 

 of reading from this cause appeared to vary from Jj; to ^ of a revolu- 

 tion of the micrometer screw; that is,' from ^^^ to ^Is o^ ^^ i"<^'^' 

 In order to avoid undue disturbance of the dynamo during the meas- 

 urements, the dynamo circuit was never broken, but by means of a 

 suitable key the current was thrown into a wire resistance approxi- 

 mately equivalent to the arc resistance before the arc was broken. 

 The current forming the arc was regulated by means of suitable resist- 

 ance coils interposed in the circuit. 



In most of the measurements Boulton carbons were used. Ten 

 successive measurements with reversals were usually made in each 

 series, keeping the current and electromotive force as nearly constant 

 as possible, and taking the mean value as the result of one set of 

 measurements for the length of arc used. The two following tables, 

 taken at random, are given in full to indicate the degree of variation 

 observed. One of the tables shows results with a silent arc, one with 

 a whistling arc. The column headed A contains the scale readings of 

 the ammeter ; that headed V, those of the voltmeter. The mean of 

 each of these is taken, and the current and electromotive force desired 

 are then found by multiplication of the mean by the proper instru- 

 mental constants, which, with the tables given, are 1.00 for the am- 

 meter and 2.12 for the voltmeter. 



