﻿872 
  Mr. 
  B.W. 
  Clack 
  on 
  the 
  

  

  As 
  the 
  salt 
  diffuses 
  up 
  the 
  tube 
  D 
  the 
  water 
  outside 
  becomes 
  

   slightly 
  denser, 
  but 
  the 
  change 
  in 
  concentration 
  is 
  so 
  small 
  

   that 
  it 
  does 
  not 
  appreciably 
  affect 
  the 
  quantity 
  of 
  salt 
  diffused 
  

   up 
  the 
  tube; 
  but 
  the 
  suspended 
  apparatus 
  has 
  such 
  a 
  large 
  

   volume, 
  that 
  a 
  very 
  slight 
  variation 
  in 
  the 
  density 
  of 
  the 
  

   water 
  outside 
  has 
  a 
  considerable 
  etfed 
  on 
  the 
  hydrostatic 
  

   force 
  upwards. 
  When 
  the 
  earlier 
  experiments 
  were 
  per- 
  

   formed, 
  the 
  importance 
  of 
  this 
  action 
  was 
  not 
  fully 
  

   appreciated. 
  

  

  It 
  is 
  impossible 
  to 
  calculate 
  the 
  change 
  in 
  the 
  buoyancy 
  

   of 
  the 
  water, 
  but 
  there 
  is 
  no 
  doubt 
  that 
  the 
  heavier 
  liquid 
  

   will 
  tend 
  to 
  sink 
  to 
  the 
  bottom, 
  and 
  that 
  then 
  diffusion 
  will 
  

   tend 
  to 
  make 
  the 
  density 
  uniform. 
  Any 
  variation 
  in 
  the 
  

   density 
  o[ 
  the 
  liquid 
  below 
  the 
  suspended 
  apparatus 
  will 
  

   produce' 
  no 
  effect 
  on 
  its 
  weight, 
  and 
  so 
  it 
  is 
  an 
  advantage 
  to 
  

   have 
  as 
  deep 
  a 
  vessel 
  as 
  possible 
  ; 
  to 
  have 
  the 
  suspended 
  

   apparatus 
  as 
  high 
  as 
  possible 
  in 
  the 
  water; 
  and 
  to 
  have 
  the 
  

   volume 
  of 
  the 
  water 
  as 
  great 
  a- 
  possible. 
  

  

  Experiments 
  similar 
  to 
  the 
  preceding 
  were 
  carried 
  out 
  

   with 
  a 
  deeper 
  vessel, 
  C 
  C, 
  shown 
  in 
  fig. 
  2, 
  having 
  also 
  a 
  volume 
  

   about 
  twice 
  that 
  of 
  the 
  original 
  one, 
  and 
  they 
  proved 
  that 
  

   the 
  first 
  vessel 
  was 
  undoubtedly 
  too 
  small. 
  For 
  this 
  reason 
  

   a 
  number 
  of 
  results 
  obtained 
  with 
  the 
  earlier 
  vessel 
  must 
  be 
  

   discarded. 
  The 
  results 
  given 
  below 
  have 
  all 
  been 
  obtained 
  

   with 
  the 
  Larger 
  vessel. 
  

  

  In 
  the 
  theory 
  ii 
  lias 
  been 
  assumed 
  that 
  the 
  density 
  of 
  the 
  

   solution 
  at 
  the 
  bottom 
  of 
  the 
  tube 
  is 
  constant 
  over 
  a 
  hori- 
  

   zontal 
  plane, 
  and 
  that 
  this 
  density 
  is 
  the 
  same 
  as 
  that 
  of 
  the 
  

   mass 
  of 
  the 
  solution 
  in 
  the 
  neighbourhood. 
  It 
  has 
  likewise 
  

   been 
  assumed 
  that 
  the 
  density 
  at 
  the 
  top 
  of 
  the 
  tube 
  in 
  a 
  

   horizontal 
  plane 
  is 
  constant, 
  and 
  practically 
  the 
  same 
  as 
  that 
  

   of 
  pure 
  water. 
  These 
  assumptions 
  would 
  only 
  be 
  strictly 
  

   true 
  if 
  gravity 
  were 
  infinitely 
  powerful, 
  and 
  in 
  practice 
  one 
  

   would 
  expect 
  the 
  density 
  at 
  the 
  bottom 
  of 
  the 
  tube 
  to 
  be 
  

   rather 
  less 
  than 
  that 
  assumed, 
  and 
  at 
  the 
  top 
  rather 
  greater. 
  

   These 
  deviations 
  from 
  the 
  assumptions 
  will 
  give 
  a 
  calculated 
  

   value 
  of 
  k 
  smaller 
  than 
  the 
  true 
  one, 
  but 
  even 
  in 
  the 
  worse 
  

   possible 
  case 
  studied, 
  it 
  can 
  be 
  shown 
  by 
  an 
  approximate 
  

   calculation 
  that 
  the 
  error 
  introduced 
  is 
  certainly 
  much 
  less 
  

   than 
  1 
  per 
  cent. 
  

  

  The 
  narrower 
  the 
  tube 
  the 
  nearer 
  will 
  the 
  assumptions 
  be 
  

   to 
  the 
  truth, 
  but 
  a 
  narrow 
  tube 
  permits 
  only 
  a 
  small 
  amount 
  

   of 
  diffusion 
  to 
  take 
  place. 
  

  

  In 
  order 
  to 
  test 
  wmether 
  it 
  was 
  necessary 
  to 
  have 
  a 
  tube 
  

   of 
  extremely 
  narrow 
  bore 
  three 
  diffusion-tubes 
  were 
  prepared, 
  

  

  