﻿W. 
  P. 
  White 
  — 
  Silicate 
  Specific 
  Heats. 
  

  

  therefore, 
  almost 
  no 
  irregularity 
  except 
  the 
  small 
  visible 
  

   errors 
  already 
  shown 
  to 
  characterize 
  the 
  original 
  obser- 
  

   vations. 
  An 
  even 
  more 
  important 
  thing 
  is 
  the 
  agreement 
  

   between 
  the 
  mean 
  of 
  anorthite 
  and 
  albite 
  on 
  the 
  one 
  hand 
  

   and 
  andesine, 
  whose 
  composition 
  is 
  the 
  mean 
  of 
  the 
  two, 
  

   on 
  the 
  other, 
  and 
  the 
  similar 
  agreement 
  between 
  pseudo- 
  

   wollastonite, 
  magnesium 
  silicate 
  pyroxene, 
  and 
  diopside, 
  

   for 
  these 
  go 
  far 
  to 
  show 
  the 
  absence 
  of 
  errors 
  systematic 
  

   with 
  the 
  different 
  substances. 
  

  

  Table 
  III. 
  

  

  Results 
  of 
  repetition 
  of 
  the 
  same 
  determination 
  at 
  different 
  times. 
  (At 
  

  

  100° 
  mean 
  values 
  only 
  are 
  given. 
  The 
  individual 
  results 
  here 
  differed 
  

  

  more 
  from 
  each 
  other 
  than 
  the 
  means 
  at 
  different 
  times.) 
  

  

  The 
  values 
  are 
  specific 
  heats 
  for 
  even 
  intervals. 
  

  

  0°-100° 
  Quartz 
  MgSi0 
  3 
  Microcline 
  

  

  1911 
  -18680 
  

  

  1911 
  -18680 
  -18715 
  

  

  1914 
  • 
  -18682 
  -20336 
  

  

  1916 
  -18688 
  -20323 
  -18707 
  

  

  

  Quartz 
  

  

  Quartz 
  

  

  Quartz 
  

  

  CaSi0 
  3 
  

  

  

  0°-500° 
  

  

  0°-300° 
  

  

  0°-900° 
  

  

  0°_300° 
  

  

  1912 
  

  

  •23803 
  

   804 
  

  

  •21689 
  

   693 
  

  

  

  

  1913 
  

  

  - 
  

  

  

  •25956 
  

   950 
  

   954 
  

   950 
  

  

  

  1914 
  

  

  

  

  •26971 
  

   963 
  

  

  

  1915 
  

  

  

  

  

  •20449 
  

  

  448 
  

  

  1916 
  

  

  •23799 
  

  

  •21685 
  

  

  

  •20452 
  

  

  

  791 
  

  

  674 
  

  

  

  439 
  

  

  Concealed 
  Errors 
  as 
  Indicated 
  by 
  the 
  Visible 
  Errors. 
  

   It 
  remains 
  to 
  inquire 
  if 
  the 
  visible 
  errors 
  are 
  entirely 
  

   accidental. 
  Errors 
  can 
  often 
  be 
  made 
  to 
  be 
  accidental, 
  

   and 
  that 
  is 
  mainly 
  true 
  of 
  the 
  visible 
  error 
  in 
  the 
  present 
  

   case, 
  for 
  by 
  subtracting 
  the 
  heat 
  given 
  by 
  the 
  empty 
  con- 
  

   tainer 
  from 
  that 
  given 
  by 
  the 
  container 
  with 
  charge 
  all 
  

   constant 
  error 
  due 
  to 
  the 
  container 
  is 
  necessarily 
  elim- 
  

   inated. 
  On 
  the 
  other 
  hand, 
  there 
  is 
  no 
  important 
  source 
  

   of 
  accidental 
  error 
  except 
  those 
  affecting 
  the 
  container, 
  as 
  

   is 
  shown 
  by 
  the 
  fact 
  that 
  below 
  1000° 
  the 
  visible 
  errors 
  

   average 
  about 
  as 
  large 
  (in 
  calories) 
  with 
  the 
  container 
  

   empty 
  as 
  when 
  the 
  charge 
  is 
  added, 
  10 
  although 
  the 
  total 
  

  

  10 
  See 
  Table 
  I 
  in 
  the 
  paper 
  following 
  this. 
  

  

  