﻿W. 
  P. 
  White 
  — 
  Silicate 
  Specific 
  Heats. 
  5 
  

  

  trical 
  calibration 
  of 
  the 
  whole 
  calorimetric 
  installation, 
  6 
  

   giving 
  in 
  joules 
  the 
  heat 
  corresponding 
  to 
  a 
  given 
  interval 
  

   measured 
  with 
  the 
  thermoelement, 
  so 
  that 
  the 
  exact 
  value 
  

   of 
  that 
  interval 
  in 
  degrees 
  is 
  of 
  no 
  importance 
  for 
  the 
  

   present 
  purpose. 
  In 
  expressing 
  the 
  result, 
  4-183 
  joules 
  

   were 
  called 
  a 
  calory, 
  and 
  this 
  value 
  undoubtedly 
  differs 
  

   from 
  the 
  15-degree 
  calory 
  by 
  less 
  than 
  the 
  probable 
  error 
  

   of 
  one 
  of 
  the 
  specific 
  heat 
  determinations, 
  but 
  the 
  1911 
  

   "international" 
  joule 
  is 
  the 
  real 
  unit 
  used. 
  

  

  The 
  final 
  accuracy 
  of 
  this 
  calibration 
  was 
  almost 
  cer- 
  

   tainly 
  several 
  times 
  as 
  good 
  as 
  1 
  per 
  mille, 
  but 
  changes 
  

   subsequently 
  made 
  in 
  the 
  calorimeter 
  rendered 
  its 
  equiva- 
  

   lent 
  doubtful, 
  possibly 
  by 
  as 
  much 
  as 
  1 
  per 
  mille. 
  The 
  

   value 
  of 
  the 
  equivalent 
  was 
  therefore 
  checked 
  at 
  the 
  close 
  

   of 
  the 
  work 
  by 
  dropping 
  water 
  at 
  100° 
  sealed 
  in 
  a 
  silver 
  

   bulb, 
  into 
  the 
  calorimeter. 
  Two 
  determinations, 
  agree- 
  

   ing 
  (as 
  it 
  happened) 
  to 
  0-1 
  per 
  mille, 
  gave 
  0-999 
  for 
  the 
  

   specific 
  heat 
  of 
  water 
  from 
  30° 
  to 
  100°. 
  The 
  best 
  avail- 
  

   able 
  values 
  are 
  -998 
  and 
  1-0016, 
  so 
  that 
  the 
  calorimeter 
  

   equivalent 
  seems 
  still 
  to 
  be 
  known 
  as 
  well 
  as 
  the 
  mean 
  

   calory. 
  The 
  not 
  improbable 
  systematic 
  error 
  of 
  1 
  per 
  

   mille 
  in 
  all 
  the 
  results 
  is 
  of 
  no 
  great 
  importance, 
  since 
  

   it 
  is 
  relative 
  values 
  that 
  are 
  important. 
  

  

  Uniformity 
  of 
  Furnace 
  Temperature. 
  

   At 
  high 
  temperatures 
  the 
  relation 
  of 
  the 
  mean 
  tempera- 
  

   ture 
  of 
  the 
  charge 
  to 
  the 
  temperature 
  of 
  the 
  thermometer 
  

   used 
  to 
  read 
  it 
  is 
  harder 
  to 
  know 
  satisfactorily 
  than 
  is 
  

   the 
  calibration 
  of 
  the 
  thermometer. 
  The 
  furnace 
  used 
  

   above 
  600°, 
  like 
  most 
  of 
  its 
  kind, 
  was 
  really 
  a 
  vertical 
  air 
  

   bath, 
  heated 
  along 
  the 
  sides 
  and 
  cooled 
  at 
  the 
  ends. 
  

   Within 
  such 
  furnaces 
  strong 
  temperature 
  gradients 
  are 
  

   to 
  be 
  expected. 
  To 
  diminish 
  these 
  in 
  the 
  charge 
  itself 
  

   (a) 
  the 
  furnace 
  chamber 
  was 
  made 
  relatively 
  slender, 
  

   20 
  cm. 
  high 
  by 
  4-8 
  cm. 
  in 
  diameter; 
  (b) 
  the 
  furnace 
  wind- 
  

   ing 
  was 
  somewhat 
  concentrated 
  at 
  the 
  ends 
  ; 
  (c) 
  platinum- 
  

   faced 
  partitions 
  were 
  used 
  above 
  and 
  below 
  the 
  

   container; 
  (d) 
  the 
  thermoelement 
  ran 
  into 
  the 
  middle 
  of 
  

   the 
  charge 
  (a 
  scheme 
  used 
  by 
  Regnault) 
  ; 
  (e) 
  this 
  portion 
  

   of 
  the 
  thermoelement, 
  to 
  reduce 
  thermal 
  conductivity, 
  was 
  

   made 
  of 
  specially 
  fine 
  wire 
  (0-2 
  mm.), 
  and 
  was 
  protected 
  

   only 
  by 
  the 
  capillary 
  insulators, 
  which 
  were 
  ground 
  flat 
  

   on 
  an 
  emery 
  wheel, 
  so 
  that 
  they 
  would 
  enter 
  a 
  platinum 
  

  

  6 
  Walter 
  P. 
  White, 
  A 
  Test 
  of 
  Calorimetric 
  Accuracy, 
  Phys. 
  Rev., 
  31, 
  

   686, 
  1910. 
  

  

  