PIERCE. — THE COOPER HEWITT MERCURY INTERRUPTER. 403 



covered by a paper screen, leaving only the electrodes and the lower 

 part of the bulb to be photographed. In this case (O = .0043 micro- 

 farads, V= 15,000 volts) about sixty complete discharges occur during 

 the half-cycle of the transformer, which is t -^q second. 



By lowering further the capacity of the condensers and raising the 

 potential of the transformer, the number of discharges may be greatly 

 increased and the period of rest at the reversal of the cycle can be made 

 small, so that the attempt of Simon and Reich to make the mercury 

 interrupter operate on a direct current can be approximately realized 

 with the transformer as source of current. 



In order to obtain an understanding of the succession of charges and 

 discharges through the interrupter let us examine the photograph of 

 Figure 8, Plate III, which is a negative. In this case the capacity was 

 .0130 microfarads, the maximum potential of the transformer cycle 12,000 

 volts. The numbers of complete charges and discharges during succes- 

 sive half-cycles of the transformer are seen to be 12, 11, 12, 13, 13, 9, 

 11, 12, 11, 12, 12, 16, 13, 13, 12, 11, 13, 11, 14, 12, 12. By a separate 

 experiment it has been shown that with the particular interrupter the 

 condensers begin to discharge when their potential is 7070 volts, whatever 

 the capacity of the condensers. 



If the condensers always discharge at 7070 volts throughout each series, 

 and if we neglect the reactance of the discharges on the potential of the 

 transformer, the diagram of Figure VIII would represent approximately 

 the manner in which the discharges occur. The sine-curve of Figure 

 VIII represents the potential of the open-circuited secondary of the trans- 

 former with condenser in series. This curve is plotted with voltage 

 as ordinates and epoch as abscissas. When about 7070 volts is reached, 

 for this particular interrupter, the condenser discharges with a series of 

 oscillations up and down the line A M A'. The condenser again charges 

 along the practically straight line M B, and discharges again along the 

 line B N B', and so on. The rapidity with which the condenser charges, 

 after any given series of oscillations, is approximately proportional to the 

 potential of the transformer during the charge, so that the areas M A G N, 

 N G II O, . . . should be equal ; thus a division of the area M A G II K S 

 into equal n smaller areas M A G N, N G II O, . . . ought to give the 

 distribution of the discharges M, N, O, . . . S. 



The construction of Figure VIII is slightly erroneous, for the discharge 

 is never complete, but with the present interrupter always leaves the con- 

 denser charged to about IGOO volts. This residual voltage chances 

 sometimes to be positive and sometimes negative, which is a possible 



