﻿54 
  W. 
  P. 
  White 
  — 
  Specific 
  Heat 
  Determination. 
  

  

  attained 
  between 
  charge 
  and 
  water. 
  The 
  logical 
  and 
  

   final 
  test 
  is 
  to 
  wait 
  so 
  long 
  that 
  longer 
  waiting 
  does 
  not 
  

   change 
  the 
  result, 
  that 
  is, 
  to 
  compute 
  for 
  the 
  same 
  exper- 
  

   iment 
  thermal 
  leakage 
  allowances 
  ("cooling 
  correc- 
  

   tions") 
  for 
  experimental 
  periods 
  of 
  increasing 
  lengths 
  

   until 
  a 
  constant 
  result 
  is 
  reached. 
  14 
  By 
  this 
  means 
  it 
  

   was 
  found 
  that 
  with 
  our 
  70-gram 
  charges 
  some 
  of 
  

   the 
  glasses 
  required 
  20 
  minutes 
  for 
  equilibrium. 
  Pre- 
  

   viously, 
  however, 
  in 
  the 
  attempt 
  to 
  get 
  a 
  simpler 
  and 
  

   more 
  certain 
  indication, 
  direct 
  measurements 
  had 
  been 
  

   made, 
  by 
  means 
  of 
  a 
  thermoelement, 
  of 
  the 
  temperature 
  

   at 
  the 
  middle 
  of 
  a 
  cooling 
  charge 
  of 
  glass 
  lumps 
  min- 
  

   gled 
  with 
  chemists' 
  glass 
  shot. 
  These 
  experiments 
  indi- 
  

   cated 
  that 
  in 
  our 
  work 
  9 
  minutes 
  would 
  be 
  more 
  than 
  

   enough 
  for 
  equilibrium. 
  It 
  was 
  through 
  reliance 
  on 
  this 
  

   conclusion 
  that 
  some 
  of 
  our 
  earlier 
  determinations 
  

   attained 
  only 
  inferior 
  precision, 
  and 
  had 
  to 
  be 
  classed 
  as 
  

   merely 
  "preliminary," 
  while 
  serious 
  errors 
  in 
  much 
  of 
  

   the 
  work 
  would 
  have 
  resulted 
  but 
  for 
  the 
  cross-check 
  

   with 
  different 
  cooling 
  corrections. 
  The 
  erf 
  or 
  in 
  the 
  con- 
  

   clusion, 
  once 
  found, 
  was 
  easily 
  accounted 
  for 
  by 
  compar- 
  

   ing 
  one 
  of 
  our 
  regular 
  charges 
  with 
  a 
  charge 
  of 
  glass 
  

   shot, 
  using 
  the 
  thermoelectric 
  method. 
  One 
  cooled 
  more 
  

   than 
  twice 
  as 
  fast 
  as 
  the 
  other. 
  The 
  source 
  of 
  error, 
  

   therefore, 
  was 
  the 
  assumption 
  that 
  one 
  charge 
  of 
  glass 
  

   could 
  be 
  taken 
  as 
  a 
  measure 
  of 
  others 
  in 
  the 
  matter 
  of 
  

   cooling. 
  This 
  assumption 
  was 
  not 
  unreasonable, 
  since 
  

   it 
  is 
  not 
  clear 
  even 
  now 
  why 
  there 
  should 
  be 
  any 
  notice- 
  

   able 
  difference. 
  15 
  Inferences 
  regarding 
  conduction 
  of 
  

   heat 
  seem 
  to 
  be 
  remarkably 
  risky; 
  their 
  uncertainty 
  in 
  

   the 
  case 
  of 
  the 
  attainment 
  of 
  equilibrium 
  in 
  the 
  furnace 
  

   has 
  already 
  been 
  mentioned, 
  and 
  this, 
  too, 
  was 
  responsi- 
  

   ble 
  for 
  the 
  putting 
  of 
  some 
  earlier 
  determinations 
  into 
  

   the 
  preliminary 
  class. 
  

  

  IV. 
  Probable 
  Efficiency 
  of 
  the 
  Dropping 
  Process. 
  

  

  It 
  has 
  already 
  been 
  shown 
  16 
  that 
  the 
  discrepancies, 
  or 
  

   accidental 
  errors, 
  in 
  the 
  final 
  results 
  not 
  infrequently 
  

  

  u 
  Cf. 
  C. 
  Bohn, 
  Noch 
  einige 
  Bemerkungeu 
  liber 
  die 
  Bestimrrumg 
  der 
  spezi- 
  

   (isdicii 
  Winnie 
  :nis 
  Mischversuchen, 
  Pogg. 
  Ann., 
  122, 
  296, 
  1864. 
  

  

  15 
  The 
  two 
  charges 
  were, 
  on 
  the 
  whole, 
  about 
  equally 
  fine-grained. 
  More- 
  

   over, 
  the 
  thermal 
  conductivity 
  of 
  water 
  is 
  not 
  very 
  different 
  from 
  that 
  of 
  

   glass. 
  

  

  " 
  ; 
  Preceding 
  paper, 
  pages 
  2 
  to 
  9. 
  

  

  