﻿Icient 
  of 
  Diffusion. 
  871 
  

  

  The 
  densities 
  were 
  obtained 
  in 
  the 
  first 
  place 
  at 
  the 
  tem- 
  

   perature 
  o£ 
  the 
  laboratory; 
  but 
  the 
  concentrations 
  have 
  been 
  

   calculated 
  at 
  0° 
  C, 
  the 
  temperature 
  of 
  the 
  experiments, 
  by 
  

   means 
  of 
  a 
  determination 
  of! 
  the 
  absolute 
  coefficient 
  o£ 
  expan- 
  

   sion 
  o£ 
  the 
  several 
  solutions 
  over 
  this 
  range 
  o£ 
  temperature, 
  

   using 
  for 
  this 
  purpose 
  apparatus 
  identical 
  in 
  principle 
  with 
  

   that 
  employed 
  by 
  Dulong 
  and 
  Petit 
  for 
  a 
  similar 
  purpose. 
  

  

  6. 
  Determination 
  of 
  8. 
  

  

  It 
  has 
  been 
  long 
  known 
  that 
  when 
  salts 
  enter 
  into 
  solution, 
  

   a 
  change 
  usually 
  takes 
  place 
  in 
  the 
  volume 
  occupied 
  by 
  the 
  

   solvent. 
  

  

  F. 
  Kohlrausch 
  and 
  Hallwachs 
  (Wied. 
  Ann. 
  liii. 
  1894, 
  p. 
  1) 
  

   and 
  Thomsen 
  have 
  studied 
  these 
  phenomena 
  and 
  have 
  estimated 
  

   the 
  change 
  in 
  volume 
  produced 
  in 
  several 
  cases 
  by 
  the 
  deter- 
  

   mination 
  of 
  the 
  density 
  of 
  the 
  solution 
  formed, 
  when 
  a 
  known 
  

   mass 
  of 
  salt 
  is 
  dissolved 
  in 
  a 
  known 
  mass 
  of 
  water. 
  W. 
  F. 
  

   Magie 
  (Phys. 
  Rev. 
  xviii. 
  1904, 
  pp. 
  449-452) 
  and 
  Macgregor 
  

   have 
  endeavoured 
  to 
  express 
  in 
  a 
  formula 
  the 
  results 
  obtained 
  

   by 
  previous 
  experimenters. 
  

  

  We 
  have 
  seen 
  that 
  in 
  the 
  present 
  investigation 
  the 
  phe- 
  

   nomena 
  referred 
  to 
  play 
  an 
  important 
  part, 
  and 
  have 
  deduced 
  

   an 
  expression 
  for 
  k 
  involving 
  the 
  knowledge 
  of 
  the 
  change 
  

   in 
  volume 
  which 
  accompanies 
  a 
  slight 
  change 
  in 
  concentration. 
  

  

  Now 
  8 
  is 
  not 
  a 
  constant, 
  but 
  varies 
  with 
  the 
  strength 
  of 
  

   the 
  solution, 
  and 
  according 
  to 
  Magie 
  depends 
  on 
  the 
  relation- 
  

   ship 
  existing 
  between 
  the 
  volumes 
  occupied 
  by 
  the 
  dissociated 
  

   and 
  the 
  non-dissociated 
  molecules 
  of 
  salt. 
  

  

  Its 
  value 
  can 
  be 
  easily 
  deduced 
  from 
  the 
  published 
  tables 
  

   giving 
  the 
  density 
  of 
  salt 
  solutions 
  o£ 
  known 
  concentration. 
  

  

  7. 
  Results 
  for 
  KC1 
  and 
  KNO 
  ? 
  . 
  

  

  The 
  accompanying 
  graphs 
  (PI. 
  XXV.), 
  which 
  are 
  numbered 
  

   in 
  accordance 
  with 
  the 
  order 
  of 
  the 
  experiments, 
  show 
  some 
  of 
  

   the 
  curves 
  obtained 
  by 
  plotting 
  the 
  weight 
  o£ 
  the 
  suspended 
  

   apparatus 
  against 
  the 
  time 
  as 
  abscissae. 
  

  

  The 
  curves 
  suggest 
  that 
  the 
  method 
  is 
  at 
  any 
  rate 
  capable 
  

   of 
  giving 
  fairly 
  consistent 
  results, 
  and 
  they 
  indicate 
  a 
  dimi- 
  

   nution 
  in 
  the 
  coefficient 
  of 
  diffusion 
  with 
  a 
  decrease 
  in 
  the 
  

   concentration 
  in 
  the 
  cases 
  of 
  both 
  NaCl 
  and 
  KC1, 
  a 
  result 
  

   which 
  has 
  been 
  found 
  by 
  several 
  other 
  observers. 
  

  

  The 
  actual 
  figures 
  obtained 
  in 
  this 
  work 
  are 
  omitted 
  for 
  

   the 
  reason 
  mentioned 
  below. 
  

  

  Before 
  proceeding 
  further, 
  it 
  was 
  decided 
  to 
  test 
  whether 
  

   the 
  outer 
  vessel 
  containing 
  the 
  salt 
  solution 
  was 
  large 
  enough. 
  

  

  