﻿688 
  

  

  Dr. 
  W. 
  F. 
  G. 
  Swann 
  on 
  a 
  Case 
  of 
  

  

  from 
  that 
  law, 
  and 
  not 
  to 
  some 
  imperfection 
  in 
  the 
  guard- 
  

   ring 
  arrangement, 
  for 
  on 
  such 
  a 
  view 
  there 
  would 
  be 
  no 
  

   reason 
  for 
  the 
  exact 
  similarity 
  of 
  the 
  two 
  curves. 
  Again, 
  

   the 
  effects 
  are 
  not 
  to 
  be 
  explained 
  by 
  any 
  inductive 
  action, 
  

   caused 
  by 
  a 
  lowering 
  of 
  the 
  potential 
  of 
  the 
  plate 
  C 
  due 
  to 
  

   a 
  passage 
  of 
  electricity 
  more 
  rapidly 
  from 
  its 
  edges 
  than 
  

   from 
  other 
  parts, 
  or 
  to 
  a 
  passage 
  of 
  electricity 
  from 
  the 
  rod 
  

   attached 
  to 
  the 
  plate 
  C, 
  for 
  such 
  effects 
  would 
  result 
  in 
  a 
  

   lowering 
  of 
  the 
  potential 
  of 
  both 
  plates. 
  

  

  Quite 
  apart 
  from 
  the 
  above 
  considerations, 
  it 
  may 
  be 
  

   shown 
  (though 
  it 
  is 
  hardly 
  necessary 
  here 
  to 
  give 
  the 
  

   analysis), 
  that 
  if 
  Ohm's 
  law' 
  were 
  obeyed, 
  any 
  alteration 
  of 
  

   the 
  potential 
  of 
  the 
  plates 
  or 
  of 
  any 
  part 
  of 
  the 
  apparatus, 
  

   for 
  that 
  matter, 
  which 
  could 
  possibly 
  arise 
  due 
  to 
  imper- 
  

   fections 
  of 
  the 
  guard-ring 
  arrangement, 
  would 
  necessarily 
  

   follow 
  an 
  exponential 
  law 
  with 
  the 
  time. 
  

  

  Now 
  on 
  referring 
  to 
  the 
  curves 
  in 
  figs. 
  2 
  and 
  3, 
  we 
  find 
  

   that 
  they 
  depart 
  from 
  the 
  exponential 
  form 
  to 
  a 
  degree 
  

   enormously 
  greater 
  than 
  the 
  possible 
  limit 
  o£ 
  error 
  in 
  the 
  

  

  Fig. 
  4. 
  

  

  jSO 
  

  

  ?. 
  3 
  A- 
  S 
  6 
  7 
  

  

  experiments, 
  for 
  if 
  these 
  curves 
  are 
  exponential 
  curves 
  they 
  

   should 
  lead 
  to 
  a 
  linear 
  relation 
  between 
  logy 
  and 
  t, 
  where 
  y 
  

   is 
  the 
  ordinate. 
  In 
  fig. 
  4 
  log 
  y 
  is 
  plotted 
  against 
  t, 
  y 
  being 
  

   taken 
  for 
  each 
  time 
  as 
  the 
  mean 
  of 
  the 
  corresponding 
  

  

  