﻿W. 
  P. 
  White 
  — 
  Silicate 
  Specific 
  Heats. 
  25 
  

  

  The 
  constancy 
  of 
  C 
  v 
  at 
  high 
  temperatures 
  has 
  been 
  

   questioned, 
  and 
  for 
  platinum 
  and 
  some 
  other 
  metals 
  C„ 
  

   certainly 
  is 
  not 
  constant. 
  A 
  satisfactory 
  test 
  is 
  often 
  

   precluded 
  by 
  lack 
  of 
  the 
  mechanical 
  data, 
  but 
  in 
  silica 
  

   glass 
  we 
  have 
  a 
  substance 
  whose 
  expansion 
  and 
  com- 
  

   pressibility 
  28 
  are 
  known. 
  For 
  it 
  the 
  expansion 
  is 
  so 
  

   small 
  that 
  the 
  calculated 
  value 
  of 
  C 
  P 
  — 
  C 
  v 
  would 
  be 
  

   imperceptible 
  in 
  fig. 
  1. 
  The 
  Q 
  P 
  curve 
  is 
  the 
  same 
  as 
  the 
  

   C* 
  curve. 
  If 
  this 
  curve 
  is 
  compared 
  with 
  the 
  theoretical 
  

   curve 
  for 
  silica 
  glass 
  derived 
  by 
  Nernst 
  29 
  from 
  experi- 
  

   ments 
  between 
  — 
  247° 
  and 
  + 
  283°, 
  it 
  appears 
  that 
  above 
  

   600° 
  the 
  substance 
  is 
  above 
  the 
  theoretical 
  value. 
  Even 
  

   the 
  lower 
  values 
  of 
  cristobalite, 
  which 
  lie 
  on 
  the 
  theoret- 
  

   ical 
  curve 
  up 
  to 
  900 
  °, 
  rise 
  above 
  it 
  at 
  1100° 
  and 
  higher. 
  

   Above 
  600° 
  quartz 
  contracts 
  with 
  rising 
  temperature 
  for 
  

   400° 
  or 
  more, 
  and 
  if 
  the 
  thermal 
  effect 
  of 
  this 
  contraction 
  

   is 
  computed 
  and 
  applied 
  A 
  v 
  for 
  quartz 
  comes 
  to 
  about 
  

   that 
  of 
  silica 
  glass 
  within 
  the 
  uncertainty 
  of 
  all 
  the 
  data, 
  

   which 
  is, 
  however, 
  perhaps 
  1 
  per 
  cent. 
  By 
  analogy 
  it 
  

   seems 
  probable 
  that 
  the 
  low 
  value 
  of 
  cristobalite 
  com- 
  

   pared 
  to 
  silica 
  glass 
  may 
  be 
  partly 
  due 
  to 
  the 
  thermal 
  

   effect 
  of 
  a 
  contraction, 
  so 
  that 
  for 
  it 
  A 
  v 
  may 
  be 
  even 
  

   higher 
  than 
  the 
  values 
  given. 
  Hence 
  these 
  substances 
  

   agree 
  in 
  indicating 
  that 
  for 
  non-metallic 
  substances 
  A„ 
  

   at 
  high 
  temperatures 
  rises 
  above 
  the 
  present 
  accepted 
  

   theoretical 
  value. 
  Of 
  the 
  two, 
  cristobalite 
  is 
  more 
  sig- 
  

   nificant 
  than 
  the 
  glass, 
  for 
  the 
  theory 
  applies 
  strictly 
  only 
  

   to 
  solids, 
  and 
  the 
  glass, 
  being 
  amorphous, 
  has 
  some 
  of 
  the 
  

   properties 
  of 
  a 
  liquid. 
  

  

  For 
  albite, 
  C 
  p 
  is 
  probably 
  considerably 
  above 
  0„. 
  

   Roughly 
  approximate 
  calculations, 
  the 
  only 
  ones 
  now 
  

   available, 
  indicate 
  at 
  1000° 
  a 
  possible 
  difference, 
  but 
  not 
  

   one 
  of 
  great 
  magnitude, 
  between 
  C 
  v 
  and 
  the 
  theoretical 
  

   value, 
  which 
  should 
  be 
  higher 
  than 
  for 
  silica, 
  since 
  the 
  

   substance 
  contains 
  a 
  smaller 
  proportion 
  of 
  oxygen 
  atoms. 
  

   Hence 
  these 
  substances 
  confirm 
  as 
  well 
  as 
  can 
  be 
  ex- 
  

  

  28 
  Compressibility 
  3-2 
  x 
  10~ 
  6 
  from 
  unpublished 
  results 
  of 
  Leason 
  H. 
  Adams 
  

   and 
  Erskine 
  D. 
  Williamson 
  in 
  this 
  laboratory. 
  

  

  29 
  Op. 
  cit., 
  p. 
  430. 
  The 
  curve 
  here 
  is 
  slightly 
  adjusted 
  to 
  coincide 
  with 
  

   the 
  curve 
  of 
  the 
  observations 
  at 
  283°. 
  Nernst 
  had 
  used 
  at 
  that 
  temperature 
  

   Magnus' 
  result, 
  whose 
  error 
  has 
  already 
  been 
  treated. 
  From 
  700° 
  upward 
  

   the 
  Debye 
  formula 
  and 
  the 
  original 
  Einstein 
  formula 
  for 
  the 
  theoretical 
  

   atomic 
  heat 
  do 
  not 
  differ 
  perceptibly 
  from 
  the 
  Nernst-Lindemann, 
  used 
  by 
  

   Nernst, 
  so 
  there 
  is 
  no 
  uncertainty 
  regarding 
  the 
  form 
  in 
  which 
  the 
  theory 
  

   is 
  represented. 
  

  

  