﻿292 
  Messrs. 
  Spiers, 
  Twyman, 
  and 
  Waters 
  on 
  Variations 
  in 
  

  

  Considering 
  that 
  this 
  adjustment 
  was 
  done 
  purely 
  by 
  the 
  

   method 
  of 
  trial 
  and 
  error, 
  the 
  curves 
  obtained 
  may 
  be 
  regarded 
  

   as 
  highly 
  satisfactory. 
  They 
  approximate 
  so 
  closely 
  to 
  straight 
  

   lines 
  that 
  it 
  was 
  thought 
  unnecessary 
  to 
  reproduce 
  them. 
  In 
  

   some 
  of 
  the 
  cycles 
  we 
  cooled 
  first 
  and 
  and 
  heated 
  afterwards, 
  

   thus 
  reversing 
  the 
  cycle. 
  At 
  the 
  conclusion 
  of 
  each 
  experi- 
  

   ment 
  the 
  E.M.F. 
  of 
  the 
  standard 
  was 
  again 
  tested 
  absolutely. 
  

  

  The 
  accuracy 
  of 
  the 
  method 
  was 
  all 
  that 
  was 
  desired. 
  

   The 
  thermostat 
  would 
  keep 
  the 
  temperature 
  of 
  the 
  standard 
  

   constant 
  to 
  o, 
  01 
  C, 
  and 
  the 
  temperature 
  of 
  the 
  cell-bath 
  was 
  

   uniform 
  to 
  0°*02 
  at 
  least. 
  The 
  difference 
  in 
  the 
  E.M.Fs. 
  

   could 
  be 
  read 
  on 
  the 
  stretched 
  wire 
  to 
  one 
  hundred-thousandth 
  

   of 
  a 
  volt. 
  The 
  E.M.F. 
  of 
  the 
  standard 
  never 
  varied 
  more 
  

   than 
  one 
  ten-thousandth 
  of 
  a 
  volt 
  during 
  an 
  experiment, 
  and 
  

   in 
  general 
  it 
  probably 
  varied 
  much 
  less. 
  

  

  Three 
  different 
  rates 
  of 
  change 
  of 
  temperature 
  were 
  tried 
  : 
  

   1° 
  C. 
  in 
  seven 
  minutes; 
  1° 
  in 
  fifteen 
  minutes; 
  and 
  1° 
  in 
  

   thirty 
  minutes. 
  A 
  cycle 
  generally 
  lasted 
  about 
  five 
  or 
  six 
  

   hours. 
  The 
  curves 
  are 
  plotted 
  for 
  E.M.F. 
  and 
  temperature, 
  

   and, 
  to 
  avoid 
  confusion, 
  the 
  curves 
  for 
  the 
  different 
  cells 
  

   have 
  been 
  separated 
  by 
  displacing 
  the 
  origin 
  of 
  E 
  M.F. 
  

   differences. 
  

  

  It 
  will 
  be 
  seen 
  that 
  the 
  curve 
  connecting 
  E.M.F. 
  and 
  

   temperature 
  is 
  not 
  the 
  same 
  for 
  both 
  the 
  rising 
  and 
  the 
  falling 
  

   temperatures 
  in 
  a 
  cycle. 
  As 
  explained 
  above, 
  this 
  indicates 
  

   the 
  existence 
  of 
  lag. 
  The 
  lag 
  of 
  E.M.F. 
  at 
  a 
  rate 
  of 
  change 
  

   of 
  temperature 
  of 
  1° 
  in 
  seven 
  minutes 
  is 
  about 
  four 
  ten- 
  

   thousandths 
  of 
  a 
  volt 
  ; 
  at 
  a 
  rate 
  of 
  change 
  of 
  1° 
  in 
  fifteen 
  

   minutes, 
  it 
  is 
  about 
  two-and-a-half 
  ten-thousandths 
  ; 
  at 
  a 
  rate 
  

   of 
  1° 
  in 
  thirty 
  minutes, 
  it 
  is 
  about 
  one 
  twenty-thousandth 
  of 
  a 
  

   volt. 
  By 
  this 
  lag 
  is 
  meant 
  half 
  the 
  difference 
  between 
  the 
  

   E.M.Fs. 
  given 
  by 
  the 
  rising 
  and 
  falling 
  curves 
  for 
  any 
  par- 
  

   ticular 
  temperature. 
  If 
  these 
  lags 
  are 
  compared 
  with 
  those 
  

   obtained 
  for 
  the 
  B.T.-type 
  of 
  cell, 
  in 
  the 
  paper 
  above 
  referred 
  

   to, 
  we 
  see 
  that 
  the 
  lag 
  in 
  the 
  case 
  of 
  the 
  H-form 
  of 
  cell 
  is 
  less 
  

   than 
  one 
  quarter 
  of 
  that 
  in 
  the 
  ordinary 
  form. 
  (See 
  fig. 
  7. 
  

   N.B. 
  — 
  The 
  zero-lines 
  on 
  this 
  diagram 
  are 
  purely 
  arbitrary 
  ; 
  

   the 
  curves 
  are 
  merely 
  for 
  the 
  purpose 
  of 
  comparison.) 
  

  

  The 
  temperature 
  at 
  the 
  close 
  of 
  a 
  cycle 
  was 
  sometimes 
  not 
  

   exactly 
  the 
  same 
  as 
  at 
  the 
  commencement, 
  hence 
  the 
  tempera- 
  

   ture-time 
  curves 
  are 
  not 
  closed. 
  It 
  is 
  highly 
  probable 
  that 
  if 
  

   the 
  temperature 
  had 
  been 
  the 
  same, 
  the 
  E.M.F. 
  w 
  T 
  ould 
  have 
  

   returned 
  to 
  its 
  original 
  value. 
  On 
  keeping 
  the 
  temperature 
  

   constant 
  at 
  the 
  end 
  of 
  the 
  cycle, 
  the 
  lag 
  passed 
  off 
  rapidly 
  ; 
  

   at 
  a 
  rate 
  of 
  variation 
  of 
  temperature 
  of 
  1° 
  in 
  fifteen 
  minutes 
  

   it 
  disappeared 
  in 
  about 
  ten 
  minutes. 
  

  

  