736  Prof.  J.  S.  Townsend  on  the  Field  of  Force 
It  was  found  that  perfectly  consistent  results  were  obtained 
when  the  connexions  from  the  electrode  and  the  exploring- 
wire  in  the  gas  to  the  terminals  of  the  voltmeter  were  made 
through  hi  oh  resistances  of  about  5  megohms.  A  dilute  salt 
solution  in  a  capillary  tube  3  or  -4  centimetres  in  length  is  a 
convenient  form  of  resistance  for  this  purpose  as  it  has  a 
small  capacity.  It  is  necessary  to  have  a  high  resistance  in 
each  line  leading  to  the  voltmeter,  and  it  is  of  advantage  to 
have  the  resistance  tubes  as  near  as  possible  to  the  electrodes 
and  the  exploring-wire.  With  the  high  resistances  thus 
placed  in  the  lines  conducting  to  the  electrostatic  voltmeter, 
the  following  readings  were  observed  : — 
Potential-difference  between  exploring-wire  and  negative  electrode  270  volts 
„  „  „  positive        „       102  volts 
The  sum  of  the  potentials  only  differs  by  2  volts  from  the 
ascertained  difference  of  potential  between  the  electrodes,  so 
that  the  disturbing  effects  prodnced  by  the  capacity  of  the 
voltmeter  are  eliminated  when  high  resistances  are  used  in 
tne  connexions. 
The  following  experiments  are  of  interest  in  this  connexion 
as  they  show  the  effect  of  capacity  in  a  simple  manner. 
A  high  resistance  was  joined  to  the  base,  another  to  the 
rod  Pl5  and  a  third  to  the  rod  P2,  and  the  connexions  to  the 
voltmeter  were  made  through  these  resistances,  so  that  the 
potential- differences  between  the  wire  and  each  of  the  electrodes 
were  not  disturbed  by  using  the  electrostatic  voltmeter  to 
make  the  necessary  observations.  The  effects  produced  by 
connecting  one  set  of  plates  of  a  condenser  to  the  wire  "W  and 
the  other  set  of  plates  to  either  A  or  B  may  then  be  easily 
examined.  For  this  purpose,  a  small  air-condenser  was  used 
which  consisted  of  an  inner  plate  P  12  centimetres  square  and 
two  outer  plates  Q  in  metallic  connexion  and  slightly  larger 
than  the  plate  P.  The  air-space  between  the  inner  plate  P 
and  the  plates  Q  was  approximately  1  centimetre,  so  that  the 
capacity  of  P  was  about  23  electrostatic  units.  The  plate  P 
was  fixed  in  a  vertical  position  to  the  rod  P:  making  contact 
with  the  wire  W  through  the  ebonite  plug  in  the  side  tube 
of  the  cover-glass,  and  the  plates  Q  were  set  up  on  a  paraffin 
block  so  as  to  be  equidistant  from  it  on  either  side.  There 
was  thus  complete  air  insulation  between  the  plates  P  and  Q  ; 
nevertheless  the  potential  of  the  wire  when  the  current  was 
flowing  depended  upon  whether  the  outer  plates  Q  were 
joined  to  the  positive  or  negative  electrode  in  the  gas. 
Using  the  same  current  as  in  the  previous   experiments 
the  following  results  were  obtained : — 
