﻿416 
  Mr. 
  L. 
  Regard 
  : 
  Researches 
  upon 
  

  

  be 
  determined 
  very 
  exactly 
  for 
  pressures 
  from 
  to 
  over 
  

   100 
  atm. 
  ; 
  and 
  the 
  exactness 
  will 
  in 
  general 
  increase 
  when 
  

   the 
  leak 
  diminishes, 
  as 
  a 
  smaller 
  osmotic 
  leak 
  generally 
  

  

  corresponds 
  to 
  a 
  greater 
  value 
  of 
  --. 
  

  

  In 
  general, 
  however, 
  we 
  shall 
  not 
  get 
  a 
  better 
  value 
  by 
  

   taking 
  the 
  mean 
  of 
  several 
  observations 
  corresponding 
  to 
  

   different 
  leaks, 
  Only 
  in 
  the 
  case 
  where 
  we 
  had 
  a 
  series 
  of 
  

   values 
  whose 
  differences 
  were 
  less 
  than 
  the 
  error 
  in 
  the 
  

   determination 
  of 
  tt 
  ' 
  would 
  it 
  be 
  natural 
  to 
  take 
  the 
  mean 
  

   value. 
  

  

  The 
  Temperature 
  Effect. 
  

  

  21. 
  The 
  influence 
  of 
  temperature 
  upon 
  the 
  reversion 
  

   pressure 
  has 
  been 
  examined 
  in 
  three 
  experiments, 
  I., 
  II., 
  III., 
  

   which 
  are 
  described 
  earlier. 
  The 
  pressure 
  7r 
  ' 
  will 
  generally 
  

   depend 
  upon 
  the 
  qualities 
  of 
  the 
  membrane. 
  If 
  then 
  the 
  

   membrane 
  and 
  consequently 
  7r 
  ' 
  undergo 
  a 
  change 
  which 
  is 
  

   not 
  a 
  temperature 
  effect, 
  the 
  experiment 
  must 
  be 
  arranged 
  

   so 
  as 
  to 
  give 
  a 
  determination 
  of 
  this 
  change. 
  This 
  can 
  be 
  

   obtained 
  in 
  the 
  following 
  manner. 
  

  

  The 
  temperature 
  is 
  to 
  be 
  changed 
  in 
  one 
  and 
  the 
  same 
  

   experiment 
  without 
  taking 
  down 
  the 
  cell 
  and 
  changing 
  the 
  

   solution. 
  The 
  time 
  between 
  the 
  measurements 
  by 
  the 
  two 
  

   temperatures 
  ought 
  to 
  be 
  as 
  short 
  as 
  possible. 
  Further, 
  the 
  

   reversion 
  pressure 
  must 
  be 
  examined 
  after 
  going 
  back 
  to 
  the 
  

   original 
  temperature. 
  This 
  mode 
  of 
  proceeding 
  has 
  only 
  

   been 
  entirely 
  followed 
  in 
  the 
  last 
  two 
  experiments. 
  

  

  In 
  this 
  manner 
  we 
  can 
  reduce 
  the 
  temperature 
  coefficient 
  

   to 
  a 
  constant 
  membrane, 
  or 
  we 
  get 
  the 
  temperature 
  effect 
  

   ttq 
  as 
  it 
  actually 
  exists 
  at 
  a 
  certain 
  moment. 
  

  

  It 
  will 
  be 
  a 
  special 
  question 
  to 
  decide 
  how 
  far 
  the 
  deter- 
  

   mined 
  temperature 
  coefficient 
  will 
  give 
  any 
  value 
  for 
  the 
  

   effect 
  of 
  temperature 
  upon 
  Osmotic 
  Pressure. 
  

  

  By 
  the 
  temperature 
  coefficient 
  of 
  tt 
  ' 
  we 
  shall 
  understand 
  

  

  the 
  quantity 
  ( 
  -^p 
  J 
  , 
  meaning 
  the 
  change 
  in 
  ir 
  ' 
  per 
  degree 
  

  

  when 
  the 
  concentration 
  is 
  kept 
  constant. 
  

  

  When 
  working 
  between 
  two 
  temperatures 
  we 
  can 
  only 
  

   find 
  the 
  average 
  temperature 
  coefficient 
  a 
  '. 
  Assume 
  that 
  

   by 
  changing 
  the 
  temperature 
  from 
  to 
  t 
  the 
  pressure 
  ttq 
  

   increases 
  by 
  a 
  quantity 
  A7r 
  ', 
  then 
  

  

  ,_1A<> 
  ' 
  C 
  R 
  

   t 
  ttq 
  a\j 
  7Tq 
  

  

  