﻿Determination 
  of 
  Conductivity 
  of 
  Liquids 
  in 
  Thin 
  Layers. 
  253 
  

  

  magnetic 
  field 
  so 
  as 
  to 
  be 
  transversely 
  magnetized. 
  This 
  

   variation 
  is 
  greatest 
  in 
  nickel 
  and 
  in 
  cobalt, 
  and 
  least 
  in 
  iron 
  ; 
  

   the 
  variation 
  can 
  be 
  observed 
  in 
  nickel 
  at 
  much 
  lower 
  fields 
  

   than 
  in 
  cobalt, 
  in 
  some 
  cases 
  with 
  a 
  strength 
  of 
  1000 
  C.G.S. 
  

   units 
  : 
  in 
  cobalt 
  it 
  can 
  be 
  with 
  certainty 
  observed 
  from 
  2000 
  

   C.G.S. 
  units 
  onwards. 
  In 
  iron 
  the 
  variation 
  can 
  usually 
  not 
  

   be 
  observed 
  with 
  fields 
  weaker 
  than 
  6000 
  C.G.S. 
  units. 
  

  

  (2) 
  In 
  cobalt 
  films 
  a 
  direct 
  relation 
  exists 
  between 
  the 
  varia- 
  

   tion 
  of 
  the 
  resistance 
  and 
  the 
  Hall 
  effect 
  : 
  from 
  it 
  we 
  conclude 
  

   that 
  the 
  latter 
  is 
  proportional 
  to 
  the 
  magnetization 
  squared. 
  

  

  (3) 
  In 
  nickel 
  the 
  variation 
  of 
  resistance 
  differs 
  very 
  con- 
  

   siderably 
  according 
  to 
  the 
  specimen 
  used. 
  In 
  all 
  the 
  mirrors, 
  

   however, 
  it 
  shows 
  signs 
  of 
  attaining 
  a 
  maximum, 
  but 
  at 
  a 
  

   higher 
  field 
  than 
  that 
  required 
  for 
  the 
  Hall 
  effect 
  maximum. 
  

  

  Again, 
  the 
  variation 
  becomes 
  greater 
  in 
  some 
  plates 
  after 
  

   they 
  have 
  been 
  used, 
  so 
  that 
  a 
  plate 
  which 
  originally 
  had 
  a 
  

   resistance 
  variation 
  proportional 
  to 
  the 
  magnetization 
  squared 
  

   has, 
  after 
  first 
  being 
  used, 
  a 
  higher 
  variation 
  than 
  this 
  relation 
  

   would 
  lead 
  us 
  to 
  expect. 
  

  

  (4) 
  In 
  iron 
  the 
  variation 
  of 
  resistance 
  is 
  less 
  than 
  propor- 
  

   tionality 
  to 
  magnetization 
  squared 
  would 
  lead 
  us 
  to 
  expect. 
  

  

  I 
  have 
  to 
  express 
  my 
  thanks 
  to 
  Professors 
  Warburg 
  and 
  

   Rubens 
  for 
  their 
  great 
  kindness 
  and 
  assistance 
  to 
  me 
  while 
  

   engaged 
  in 
  carrying 
  out 
  the 
  above 
  experiments. 
  

  

  South 
  African 
  College, 
  Capetown, 
  

   December 
  1897. 
  

  

  XXVIII. 
  On 
  the 
  Determination 
  of 
  the 
  Conductivity 
  of 
  Liquids 
  

   in 
  Thin 
  Layers. 
  By 
  G. 
  B. 
  Bryan, 
  B.Sc, 
  1851 
  Exhibition 
  

   Science 
  Scholar, 
  University 
  College, 
  Nottingham, 
  St. 
  Johns 
  

   College, 
  Cambridge 
  *. 
  

  

  Introduction. 
  

  

  INTEREST 
  in 
  this 
  subject 
  has 
  been 
  aroused 
  by 
  the 
  work 
  

   of 
  Koller 
  f, 
  who 
  found 
  that, 
  with 
  certain 
  liquids, 
  the 
  con- 
  

   ductivity 
  decreased 
  as 
  the 
  thickness 
  of 
  the 
  layer 
  w 
  T 
  as 
  decreased; 
  

   and 
  it 
  was 
  suggested 
  by 
  Prof. 
  J. 
  J. 
  Thomson 
  that 
  it 
  might 
  

   be 
  worth 
  while 
  repeating 
  some 
  of 
  Roller's 
  experiments. 
  In 
  

   doing 
  this 
  the 
  apparatus 
  and 
  method 
  of 
  measuring 
  the 
  resis- 
  

   tance 
  differed 
  from 
  Koller 
  's, 
  and 
  therefore 
  his 
  method 
  of 
  

   experimenting 
  will 
  be 
  briefly 
  described. 
  

  

  To 
  obtain 
  a 
  thin 
  layer 
  of 
  a 
  liquid, 
  Koller 
  took 
  two 
  flat 
  

   circular 
  metal 
  plates, 
  and 
  placed 
  one 
  horizontally 
  on 
  the 
  

   bottom 
  of 
  a 
  glass 
  vessel 
  ; 
  the 
  other 
  was 
  laid 
  on 
  this 
  and 
  

  

  * 
  Communicated 
  by 
  Prof. 
  J. 
  J. 
  Thomson, 
  F.R.S. 
  

   t 
  Wien. 
  Ber. 
  98 
  ii. 
  a, 
  p. 
  201 
  (1889). 
  

  

  Phil. 
  Mag. 
  S. 
  5. 
  Vol. 
  45. 
  No. 
  274. 
  March 
  1898. 
  T 
  

  

  