1902.] on Musical and Talking Electric Arcs. 69 



shunt circuit (exclusive of the condenser, which should of course be 

 very high). 



Besides resistance, any other dissipation of energy such as 

 hysteresis or eddy -currents, may, if sufficient, cause the arc to fail to 

 produce its note, and for this reason the inductive coils used are pre- 

 ferably without iron cores. 



The pitch of the note depends on the frequency of the alternating 

 currents in the shunt circuit. If the resistance of the circuit is small, 

 as it should be for the best results, the frequency of these currents 

 may be calculated from Lord Kelvin's formula mentioned above 



1 



Frequency = 2 tt VKF: 



where L is the self-induction and F the capacity in the oscillatory 

 circuit shunting the arc. 



In order to show that this formula is really obeyed, a constant self- 

 induction was taken and a series of condensers calculated to produce 

 the eight notes of an octavo which were connected to a keyboard ; 

 to produce a second octave J of the self-induction was used in con- 

 junction with the same condensers, with this arrangement popular airs 

 were played on the arc. 



In view of the great sensitiveness of the ear to small differences 

 from true musical intervals, the fact that the notes produced by this 

 set of calculated condensers was even approximately in tune, indicate s 

 that Lord Kelvin's law was closely obeyed. 



Chords could not be played on the above-mentioned key-board, 

 as owing to the same inductive coil being used for each note, the 

 depressing of two or more keys simply put the condensers in parallel 

 and produced a lower note. Several experiments were made to see 

 if it were possible to make one and the same arc give out two notes 

 at the same time by shunting it with two distinct oscillatory circuits. 

 When the periodic times of the two oscillatory circuits had a simple 

 ratio such as the octave, the arc would emit both notes together and 

 form a chord, but the condition seemed to be unstable, the chord 

 often changing to a sound which was neither one note nor the other. 



To obtain loud results, several arcs in series should be used all 

 shunted by the one oscillatory circuit. A very convenient arrange- 

 ment for working on 200-volt mains is to use four arcs in two parallel 

 circuits, each circuit consisting of two arcs and a steadying resistance 

 in series ; by this means the four arcs, though supplied with power 

 in two parallel circuits from the mains, will be metallically connected 

 in series, and may be thus all four shunted by one oscillatory circuit. 



A method of still further increasing the sound, first shown me by 

 Professor Slaby, is to shunt the main condenser by a condenser con- 

 sisting of alternate loose sheets of paper and tin-foil= In this case a 

 loud note due to the alternating current is produced by the loose 

 paper condenser as well as the note by the arcs. 



