﻿8 
  W. 
  P. 
  White 
  — 
  Silicate 
  Specific 
  Heats. 
  

  

  100, 
  or 
  50 
  per 
  cent, 
  more 
  than 
  this 
  necessary 
  number 
  were 
  

   made 
  for 
  various 
  extra 
  checks 
  and 
  comparisons. 
  

  

  Except 
  in 
  one 
  case 
  there 
  was 
  not 
  even 
  a 
  perceptible 
  

   effect 
  from 
  interchanging 
  calorimeters, 
  thermoelements, 
  

   etc. 
  This 
  of 
  course, 
  tends 
  to 
  indicate 
  the 
  absence 
  of 
  con- 
  

   stant 
  sources 
  of 
  concealed 
  error. 
  This 
  indication 
  is 
  

   corroborated 
  by 
  other 
  things 
  mentioned 
  already, 
  that 
  is 
  : 
  

   by 
  the 
  excess 
  precision 
  obtained 
  in 
  the 
  calibrations 
  of 
  

   thermoelements 
  and 
  calorimeters 
  ; 
  by 
  the 
  great 
  constancy 
  

   of 
  the 
  working 
  thermoelement 
  in 
  the 
  platinum 
  furnace; 
  

   and 
  by 
  the 
  temperature 
  uniformity 
  attained 
  in 
  that 
  fur- 
  

   nace, 
  since 
  almost 
  the 
  only 
  possible 
  sources 
  of 
  constant 
  

   error 
  lie 
  in 
  the 
  temperature 
  and 
  heat 
  determinations. 
  

   The 
  exceptional 
  case 
  of 
  perceptible 
  constant 
  error 
  is 
  

   shown 
  in 
  Table 
  II, 
  which 
  gives 
  all 
  the 
  comparisons 
  made 
  

   between 
  the 
  two 
  furnaces 
  at 
  500°, 
  and 
  indicates 
  that 
  the 
  

   platinum 
  furnace 
  was 
  possibly 
  1 
  per 
  mille 
  low 
  at 
  this 
  

   temperature. 
  This 
  difference 
  is 
  no 
  greater 
  than 
  the 
  

   value 
  often 
  reached 
  by 
  the 
  visible 
  (or 
  accidental) 
  error. 
  

  

  Table 
  II. 
  

  

  Determinations 
  of 
  the 
  Interval 
  Specific 
  Heat 
  from 
  0° 
  to 
  500° 
  of 
  Quartz 
  and 
  

   of 
  glasses 
  formed 
  from 
  Magnesium 
  Silicate, 
  Microcline 
  and 
  Calcium 
  

   Silicate. 
  Comparison 
  of 
  the 
  two 
  furnaces 
  at 
  500°. 
  

   Furnace 
  Quartz 
  MgSi0 
  3 
  gl 
  Micro, 
  gl 
  CaSi0 
  3 
  gl 
  

  

  Pt-Wound 
  -23781 
  -24729 
  -23222 
  -22034* 
  

  

  764 
  723 
  219 
  019 
  

  

  Metal 
  bath 
  -23799 
  -24754 
  -23200 
  -22095* 
  

  

  791 
  746 
  212 
  077* 
  

  

  * 
  Preliminary 
  determination. 
  

  

  Tests 
  of 
  Slowly 
  Varying 
  Causes 
  of 
  Error. 
  

  

  The 
  insidious, 
  slowly 
  or 
  occasionally 
  varying 
  causes 
  of 
  

   error 
  can 
  only 
  be 
  tested 
  by 
  numerous 
  intercomparisons. 
  

   Many 
  of 
  the 
  extra 
  determinations 
  were 
  devoted 
  to 
  repeti- 
  

   tions 
  of 
  the 
  same 
  result 
  at 
  different 
  times. 
  The 
  most 
  

   important 
  of 
  these 
  are 
  given 
  in 
  Table 
  III 
  and 
  give 
  no 
  

   indication 
  of 
  appreciable 
  long 
  period 
  variations. 
  

  

  A 
  more 
  comprehensive 
  test 
  of 
  insidious 
  errors 
  is 
  given 
  

   by 
  comparing 
  the 
  results 
  obtained 
  with 
  similar 
  silicates. 
  

   This 
  can 
  be 
  done 
  with 
  the 
  aid 
  of 
  Table 
  X 
  at 
  the 
  end 
  of 
  this 
  

   paper, 
  whose 
  data 
  cover 
  the 
  work 
  from 
  beginning 
  to 
  end. 
  

   Such 
  discrepancies 
  as 
  occur 
  there 
  are 
  clearly 
  sporadic; 
  

   and 
  out 
  of 
  45 
  comparable 
  results 
  above 
  100° 
  all 
  but 
  9 
  

   depart 
  from 
  the 
  probable 
  smooth 
  curve 
  less 
  than 
  0-3 
  

   per 
  mille, 
  and 
  only 
  3 
  as 
  much 
  as 
  1 
  per 
  mille. 
  There 
  is, 
  

  

  