in  a  Discharge  between  Parallel  Plates.  735 
^  a  volt  from  80  to  265  volts,  and  an  aluminium-leaf  electro- 
scope graduated  to  read  potentials  above  150  volts  were  used 
to  determine  the  potential-differences  between  the  wire  and 
the  electrodes.  When  either  the  voltmeter  or  the  electro- 
scope were  connected  to  the  wire  and  charged  with  a  small 
replenisber,  they  maintained  their  charges  without  any  appre- 
ciable diminution  for  several  minutes,  showing  that  the  in- 
sulation was  satisfactory.  The  cases  of  the  voltmeter  and 
electroscope  were  also  insulated. 
4.  An  unexpected  difficulty  arose  when  preliminary  expe- 
riments were  being  made,  which  was  subsequently  traced  to 
an  effect  arising  from  the  capacity  of  the  insulated  system  of 
the  voltmeter  or  the  electroscope  in  connexion  with  the  wire 
in  the  gas.  The  following  experiment  illustrates  the  general 
nature  of  the  effects  which  are  produced.  The  plates  were 
set  at  8  millimetres  apart  and  the  wire  was  placed  midway 
between  the  plates,  the  gas  being  hydrogen  at  6*7  mms. 
pressure.  The  terminals  of  a  battery  of  452  volts  were 
connected  through  a  galvanometer  and  resistances  amounting 
to  371,000  ohms  to  the  electrodes,  which  gave  a  constant 
current  of  2'1  X  10~4  ampere  through  the  gas.  The  potential- 
fall  along  the  external  resistances  being  78  volts,  the  differ- 
ence of  potential  between  the  electrodes  in  the  gas  was 
374  volts.  This  was  verified  by  another  electrostatic  volt- 
meter which  gave  a  constant  deflexion  of  374  volts  when  it 
was  connected  to  the  electrodes. 
When  the  sensitive  electrostatic  voltmeter  was  used  to  find 
the  potential-differences  between  the  wire  and  the  electrodes, 
the  following  results  were  obtained  : — 
volts 
Potential-fall  between  wire  and  negative  electrode   .     .     178 
„         „  „  positive         „       less  than  80 
The  sum  of  the  potentials  should  be  374  volts  instead  of 
being  less  than  258  volts,  which  shows  that  the  potential  of 
the  wire  must  be  disturbed  by  connecting  it  to  the  voltmeter. 
It  was  found  that  the  discrepancy  could  not  be  due  to  faults 
in  the  insulation  or  to  errors  in  the  voltmeter,  as  the  readings 
of  the  latter  were  found  to  be  very  accurate  over  the  whole 
range  of  the  scale  when  tested  with  voltages  that  had  been 
measured  with  a  standard  milliampere  balance. 
In  the  above  experiments  the  connexions  to  the  voltmeter 
were  made  with  ordinary  copper  wires  having  small  resistance. 
A  number  of  experiments  were  made  to  find  the  cause  of 
the  disturbance  of  the  potential  of  the  exploring  wire  and  to 
eliminate  the  effect  if  possible. 
