﻿the 
  Conductivity 
  of 
  Liquids 
  in 
  Thin 
  Layers* 
  271 
  

  

  Conclusion. 
  

  

  It 
  will 
  be 
  seen 
  that 
  the 
  expectation 
  of 
  reducing 
  the 
  

   apparent 
  differences 
  in 
  conductivity 
  of 
  the 
  different 
  thick- 
  

   nesses 
  (Tables 
  VIII. 
  and 
  XL) 
  has 
  been 
  realized 
  to 
  a 
  great 
  

   extent 
  ; 
  in 
  most 
  cases 
  reduced 
  within 
  the 
  limits 
  of 
  error 
  of 
  

   adjustment. 
  This 
  is 
  partly 
  due 
  to 
  the 
  improvement 
  in 
  the 
  

   surfaces 
  of 
  the 
  plates, 
  and 
  partly 
  to 
  better 
  cleaning. 
  The 
  

   plates 
  were 
  washed 
  in 
  alcohol, 
  after 
  dipping 
  in 
  dilute 
  sul- 
  

   phuric 
  acid, 
  until 
  the 
  liquid 
  to 
  be 
  tested 
  would 
  thoroughly 
  

   wet 
  them. 
  

  

  The 
  readings, 
  in 
  the 
  case 
  of 
  the 
  very 
  thin 
  layers 
  of 
  dilute 
  

   solutions, 
  depended 
  very 
  much 
  on 
  the 
  state 
  of 
  cleanliness 
  of 
  

   the 
  plates. 
  Thus 
  for 
  water 
  (Table 
  XII.) 
  the 
  Sp. 
  R. 
  of 
  *1 
  turn 
  

   is 
  27 
  per 
  cent, 
  greater 
  than 
  that 
  for 
  10 
  turns 
  (of 
  this 
  difference 
  

   about 
  12 
  per 
  cent, 
  can 
  be 
  accounted 
  for 
  by 
  error 
  of 
  adjust- 
  

   ment) 
  ; 
  whilst 
  for 
  very 
  dilute 
  KC1 
  (Table 
  XVI.) 
  when 
  the 
  

   plates 
  were 
  more 
  carefully 
  cleaned, 
  the 
  corresponding 
  differ- 
  

   ence 
  is 
  only 
  15 
  per 
  cent., 
  although 
  the 
  resistance 
  is 
  five 
  times 
  

   less. 
  

  

  With 
  alcohol 
  (Table 
  XVII.) 
  the 
  readings 
  from 
  1 
  to 
  12 
  turns 
  

   agree 
  very 
  closely, 
  and 
  below 
  1 
  they 
  are 
  within 
  the 
  limit 
  of 
  

   error 
  of 
  adjustment. 
  

  

  With 
  aniline 
  (Table 
  XVIII.) 
  all 
  the 
  values 
  of 
  the 
  Sp. 
  R. 
  

   agree 
  fairly 
  closely, 
  the 
  greatest 
  difference 
  from 
  the 
  mean 
  

   being 
  less 
  than 
  2 
  per 
  cent. 
  

  

  The 
  range 
  of 
  the 
  experiments 
  may 
  be 
  considered 
  to 
  be 
  from 
  

   *5 
  turn 
  ("423 
  mm.) 
  to 
  20 
  turns 
  (16*92 
  mm.) 
  ; 
  below 
  *5 
  th^ 
  

   values 
  of 
  the 
  Sp. 
  R. 
  show 
  a 
  difference 
  from 
  the 
  rest 
  which 
  

   could 
  not 
  be 
  got 
  rid 
  of; 
  above 
  20 
  turns 
  several 
  experiments 
  

   were 
  made 
  with 
  a 
  tube 
  1(3 
  cm. 
  long 
  and 
  about 
  4 
  cm. 
  in 
  

   diameter. 
  The 
  values 
  of 
  the 
  Sp. 
  R. 
  obtained 
  were 
  in 
  very 
  

   close 
  agreement 
  with 
  those 
  at 
  10 
  or 
  20 
  turns 
  for 
  the 
  same 
  

   liquid. 
  

  

  The 
  above 
  results 
  show 
  that 
  for 
  dilute 
  solutions, 
  alcohol, 
  and 
  

   aniline, 
  the 
  conductivitv 
  i> 
  the 
  same 
  for 
  all 
  thicknesses 
  within 
  

   the 
  limits 
  of 
  the 
  experiments. 
  It 
  has 
  been 
  found 
  that 
  the 
  great 
  

   differences 
  obtained 
  by 
  Roller, 
  for 
  water 
  and 
  alcohol, 
  can 
  be 
  

   eliminated 
  by 
  taking 
  care 
  that 
  the 
  plates 
  are 
  true 
  and 
  clean, 
  and 
  

   by 
  using 
  an 
  alternate 
  current 
  instead 
  of 
  a 
  continuous 
  one. 
  

   Koller 
  allowed 
  for 
  polarization 
  by 
  estimating 
  it 
  immediately 
  

   after 
  stopping 
  the 
  current, 
  but 
  this 
  is 
  not 
  satisfactory, 
  sinc^ 
  

   the 
  polarization 
  falls 
  off 
  verv 
  rapidly. 
  He 
  estimated 
  that 
  the 
  

   greatest 
  polarization 
  could 
  not 
  have 
  been 
  greater 
  than 
  

   •02 
  volt., 
  which, 
  though 
  great 
  compared 
  with 
  the 
  whole 
  

   E.M.F. 
  (1 
  Daniell) 
  would 
  not 
  account 
  for 
  the 
  difference 
  of 
  

  

  U2 
  

  

  