﻿Coefficient 
  of 
  Diffusion. 
  869 
  

  

  The 
  time 
  recorded 
  by 
  the 
  clock 
  and 
  tlie 
  corresponding- 
  

   mass 
  on 
  the 
  pan 
  of 
  the 
  balance 
  are 
  recorded 
  in 
  tables, 
  and 
  

   from 
  these 
  the 
  graphs 
  following 
  have 
  been 
  drawn. 
  

  

  A 
  slightly 
  smaller 
  mass 
  is 
  then 
  placed 
  on 
  the 
  pan. 
  The 
  

   lath 
  is 
  replaced 
  on 
  the 
  trigger, 
  the 
  clock 
  reset 
  and 
  restarted, 
  

   and 
  the 
  key 
  K 
  switched 
  on, 
  and 
  the 
  whole 
  process 
  repeated. 
  

   This 
  method, 
  involving 
  the 
  determination 
  of 
  the 
  time 
  required 
  

   for 
  a 
  given 
  decrease 
  in 
  weight 
  to 
  take 
  place, 
  has 
  an 
  advantage 
  

   over 
  the 
  converse 
  process, 
  in 
  that 
  it 
  causes 
  far 
  less 
  disturbance 
  

   in 
  the 
  solution, 
  and 
  hence 
  it 
  has 
  been 
  adopted 
  in 
  the 
  present 
  

   investigation. 
  

  

  Another 
  precaution 
  which 
  should 
  be 
  mentioned 
  is 
  that, 
  the 
  

   arrangement 
  employed 
  being 
  extremely 
  sensitive 
  to 
  changes 
  

   in 
  density, 
  great 
  care 
  must 
  be 
  taken 
  to 
  prevent 
  any 
  but 
  

   minute 
  changes 
  in 
  temperature 
  occurring 
  in 
  the 
  solution. 
  

  

  The 
  large 
  amount 
  of 
  damping 
  introduced, 
  and 
  the 
  effect 
  of 
  

   the 
  surface-tension 
  of 
  the 
  liquid, 
  prevented 
  any 
  accurate 
  work 
  

   being 
  done 
  until 
  the 
  effect 
  of 
  a 
  film 
  of 
  oil 
  on 
  the 
  surface 
  of 
  

   the 
  solution 
  was 
  tested. 
  This 
  greatly 
  reduces 
  the 
  surface- 
  

   tension, 
  and 
  increases 
  the 
  accuracy 
  of 
  weighing 
  to 
  a 
  re- 
  

   markable 
  extent. 
  Of 
  a 
  large 
  number 
  of 
  oils 
  tested, 
  Fleuss 
  

   pump-oil 
  was 
  found 
  to 
  be 
  superior 
  to 
  any 
  of 
  the 
  others, 
  in 
  

   that 
  it 
  always 
  remained 
  quite 
  fluid. 
  

  

  As 
  showing 
  the 
  effect 
  of 
  this 
  oil-film 
  a 
  few 
  figures 
  might 
  

   be 
  quoted. 
  

  

  The 
  apparatus 
  was 
  filled 
  with 
  water 
  and 
  suspended 
  in 
  water 
  

   at 
  a 
  constant 
  temperature. 
  If 
  set 
  in 
  oscillation, 
  in 
  no 
  case 
  

   would 
  the 
  balance-beam 
  make 
  more 
  than 
  two 
  swings, 
  and 
  

   the 
  resting 
  place 
  varied 
  in 
  different 
  attempts 
  between 
  10 
  

   and 
  15 
  on 
  the 
  scale 
  for 
  the 
  same 
  mass 
  in 
  the 
  pan. 
  Two 
  

   drops 
  of 
  Fleuss 
  pump-oil 
  allowed 
  to 
  fall 
  on 
  the 
  liquid 
  near 
  

   the 
  suspending 
  wire 
  made 
  a 
  remarkable 
  difference. 
  Five 
  

   turning-points 
  could 
  now 
  be 
  easily 
  obtained, 
  and 
  assuming 
  

   the 
  ratio 
  between 
  any 
  two 
  consecutive 
  deflexions 
  to 
  be 
  con- 
  

   stant, 
  great 
  consistency 
  was 
  found 
  between 
  the 
  calculated 
  

   positions 
  of 
  the 
  resting 
  place 
  of 
  the 
  balance. 
  For 
  example, 
  

  

  Turning 
  Points. 
  

   20-8 
  «.„ 
  

  

  14-4 
  ;.$ 
  

  

  11-5 
  ' 
  8 
  

  

  Resting 
  Points, 
  

   10-0 
  

   10-2 
  

   10-1 
  

   10-2 
  

   10-3 
  

   10-0 
  

  

  19-1 
  4 
  . 
  7 
  

   13-8 
  li 
  

   11-0 
  bl 
  

  

  Now 
  in 
  the 
  diffusion 
  experiments 
  the 
  beam 
  does 
  not 
  oscil- 
  

  

  