﻿Coefficient 
  of 
  Diffusion. 
  875 
  

  

  condensation 
  o£ 
  water 
  from 
  the 
  air 
  on 
  the 
  salt 
  solution 
  in 
  

   Method 
  A, 
  so 
  that 
  part 
  o£ 
  the 
  increase 
  in 
  weight 
  of 
  the 
  

   suspended 
  apparatus 
  may 
  be 
  due 
  to 
  the 
  diminution 
  of 
  

   buoyancy 
  of 
  the 
  solution, 
  on 
  account 
  of 
  its 
  dilution 
  with 
  

   this 
  condensed 
  water. 
  

  

  The 
  error 
  produced 
  by 
  this 
  cause 
  would 
  probably 
  not 
  be 
  

   great, 
  but 
  no 
  quantitative 
  experiments 
  were 
  actually 
  carried 
  

   out 
  on 
  the 
  subject, 
  as 
  the 
  experience 
  gained 
  in 
  the 
  research 
  

   convinces 
  us 
  that 
  the 
  first 
  method 
  described 
  in 
  this 
  paper 
  is 
  

   distinctly 
  superior 
  to 
  Method 
  A. 
  

  

  For 
  this 
  reason 
  no 
  further 
  experiments 
  are 
  being 
  carried 
  

   out 
  by 
  the 
  earlier 
  Method 
  A. 
  

  

  Some 
  of 
  the 
  points 
  in 
  which 
  the 
  method 
  mentioned 
  in 
  § 
  4 
  

   shows 
  superiority 
  over 
  Method 
  A 
  may 
  be 
  summarized. 
  First, 
  

   in 
  this 
  better 
  method, 
  the 
  condensation 
  of 
  water 
  mentioned 
  

   above, 
  instead 
  of 
  being 
  harmful, 
  is 
  even 
  somewhat 
  of 
  an 
  

   advantage. 
  

  

  Again, 
  this 
  method 
  necessitates 
  the 
  preparation 
  of 
  much 
  

   smaller 
  amounts 
  of 
  solution 
  ; 
  then 
  it 
  requires 
  much 
  less 
  

   manipulation 
  in 
  commencing 
  an 
  experiment, 
  as 
  the 
  apparatus, 
  

   filled 
  with 
  the 
  solution 
  and 
  having 
  the 
  open 
  tube 
  temporarily 
  

   closed 
  by 
  a 
  thin 
  glass 
  plate, 
  may 
  be 
  simply 
  lowered 
  into 
  the 
  

   cooled 
  distilled 
  water. 
  

  

  Instead 
  of 
  this 
  simple 
  process, 
  in 
  the 
  Method 
  A 
  in 
  com- 
  

   mencing 
  an 
  experiment, 
  the 
  solution 
  under 
  investigation 
  was 
  

   slowly 
  run 
  in 
  through 
  a 
  glass 
  tube 
  passing 
  to 
  the 
  bottom 
  of 
  

   the 
  vessel 
  C 
  (fig. 
  2), 
  while 
  the 
  distilled 
  water, 
  which 
  is 
  thus 
  

   displaced, 
  was 
  removed 
  from 
  the 
  top 
  by 
  means 
  of 
  a 
  syphon. 
  

   Great 
  care 
  was 
  taken 
  to 
  ensure 
  the 
  complete 
  removal 
  of 
  the 
  

   water, 
  so 
  that 
  the 
  vessel 
  C 
  was 
  filled 
  with 
  a 
  solution 
  of 
  uniform 
  

   density. 
  

  

  Thus 
  the 
  whole 
  of 
  the 
  suspended 
  system 
  remains 
  filled 
  

   with 
  distilled 
  water, 
  but 
  is 
  immersed 
  in 
  a 
  salt 
  solution 
  of 
  

   uniform 
  density, 
  so 
  that 
  the 
  diffusion-tube, 
  initially 
  filled 
  

   with 
  pure 
  water, 
  is 
  brought 
  into 
  contact 
  at 
  its 
  lower 
  end 
  

   with 
  a 
  salt 
  solution 
  whose 
  concentration, 
  on 
  account 
  of 
  

   the 
  large 
  volume 
  present, 
  may 
  be 
  considered 
  to 
  remain 
  

   constant. 
  

  

  Again, 
  the 
  better 
  method 
  enables 
  an 
  experiment 
  to 
  be 
  

   repeated 
  with 
  precisely 
  the 
  same 
  solution, 
  so 
  that 
  the 
  working 
  

   details 
  are 
  much 
  more 
  under 
  control. 
  Moreover, 
  it 
  has 
  

   another 
  advantage 
  in 
  that 
  greater 
  accuracy 
  is 
  introduced 
  in 
  

   the 
  determination 
  of 
  the 
  quantity 
  8. 
  

  

  It 
  is 
  only 
  in 
  very 
  dilute 
  solutions 
  that 
  the 
  value 
  of 
  8 
  is 
  

   doubtful, 
  and 
  as 
  only 
  a 
  very 
  slight 
  decrease 
  occurs 
  in 
  the 
  

   concentration 
  of 
  the 
  solution 
  inside 
  the 
  diffusion 
  apparatus 
  

  

  