﻿244 
  Dr. 
  P. 
  Chapptiis 
  on 
  Gas 
  -Thermometry. 
  

  

  bearing 
  on 
  the 
  reduction 
  o£ 
  temperature 
  measurements, 
  

   in 
  which 
  reservoirs 
  of 
  Berlin 
  porcelain 
  have 
  been 
  employed. 
  

   I 
  drew 
  attention 
  in 
  a 
  former 
  note 
  * 
  to 
  the 
  fact 
  that 
  part 
  of 
  

   the 
  divergence 
  found 
  between 
  the 
  results 
  of 
  Messrs. 
  Callendar 
  

   and 
  Griffiths 
  and 
  of 
  Harker 
  and 
  myself 
  for 
  the 
  boiling-point 
  

   of 
  sulphur 
  may 
  be 
  attributed 
  to 
  the 
  uncertainty 
  in 
  the 
  values 
  

   assumed 
  for 
  the 
  expansion 
  of 
  porcelain. 
  I 
  propose 
  now 
  to 
  

   examine 
  in 
  what 
  way 
  our 
  results 
  would 
  be 
  modified 
  by 
  the 
  

   introduction 
  of 
  the 
  value 
  for 
  the 
  expansion 
  deduced 
  from 
  the 
  

   experiments 
  of 
  Messrs. 
  Holborn 
  and 
  Day. 
  I 
  would 
  like 
  at 
  

   the 
  same 
  time 
  to 
  make 
  clear 
  and 
  correct 
  one 
  or 
  two 
  errors 
  

   in 
  the 
  note 
  previously 
  mentioned, 
  to 
  which 
  Dr. 
  Chree 
  has 
  

   been 
  kind 
  enough 
  to 
  draw 
  my 
  attention. 
  

  

  1. 
  Messrs. 
  Holborn 
  and 
  Day 
  have 
  measured 
  the 
  expansion 
  

   of 
  a 
  rod 
  of 
  unglazed 
  porcelain 
  about 
  483 
  mm. 
  in 
  length 
  by 
  

   the 
  method 
  of 
  the 
  comparator. 
  Their 
  observations 
  are 
  re- 
  

   presented 
  in 
  a 
  fairly 
  satisfactory 
  manner 
  by 
  the 
  expression 
  

  

  L,-L 
  = 
  {2954£ 
  + 
  1-125* 
  2 
  }10- 
  9 
  

  

  between 
  the 
  limits 
  of 
  temperature 
  + 
  250° 
  and 
  625°. 
  This 
  

   function 
  diverges 
  considerably 
  from 
  the 
  observations 
  at 
  750° 
  

   and 
  at 
  875° 
  to 
  again 
  agree 
  better 
  at 
  1000°. 
  Below 
  250° 
  it 
  

   gives 
  values 
  which 
  are 
  too 
  high. 
  

  

  In 
  fact 
  the 
  expansion 
  of 
  Berlin 
  porcelain 
  cannot 
  be 
  

   represented 
  exactly 
  by 
  a 
  two-term 
  function 
  throughout 
  an 
  

   interval 
  of 
  temperature 
  exceeding 
  a 
  few 
  hundred 
  degrees. 
  

  

  It 
  follows 
  therefore 
  that 
  an 
  expression 
  for 
  the 
  expansion 
  

   deduced 
  empirically 
  cannot 
  be 
  applied 
  outside 
  the 
  limits 
  of 
  

   the 
  experiments 
  without 
  running 
  the 
  risk 
  of 
  committing 
  

   serious 
  errors 
  |. 
  

  

  The 
  following 
  table 
  contains 
  the 
  values 
  of 
  the 
  mean 
  co- 
  

   efficient 
  of 
  expansion 
  between 
  0° 
  and 
  T° 
  deduced 
  from 
  our 
  

   measurements, 
  those 
  of 
  Messrs. 
  Holborn 
  and 
  Wien 
  in 
  1892 
  J, 
  

   and 
  the 
  recent 
  ones 
  of 
  Messrs. 
  Holborn 
  and 
  Day. 
  

  

  * 
  Phil. 
  Mag. 
  [5] 
  vol. 
  1. 
  p. 
  436 
  (Oct. 
  1900). 
  

  

  t 
  The 
  case 
  is 
  not 
  the 
  same 
  with 
  platinum, 
  whose 
  expansion 
  measured 
  

   by 
  Messrs. 
  Holborn 
  and 
  Day 
  by 
  the 
  same 
  method 
  between 
  0° 
  and 
  1000° 
  

   has 
  been 
  found 
  to 
  be 
  (8889t+V274:t 
  2 
  ) 
  10-9. 
  M. 
  Benoit 
  had 
  previously 
  

   obtained 
  between 
  0° 
  and 
  100° 
  in 
  the 
  Fizeau 
  apparatus 
  ( 
  Trav. 
  et 
  Mem. 
  du 
  

   Bur. 
  Inter, 
  t. 
  vi. 
  p. 
  190) 
  the 
  almost 
  identical 
  expression 
  

  

  (8901*+l-21* 
  2 
  )Kr 
  9 
  . 
  

  

  Here 
  extrapolation 
  would 
  have 
  led 
  to 
  no 
  serious 
  inaccuracies. 
  

  

  X 
  Holborn 
  & 
  Wien, 
  Ann. 
  der 
  Phys. 
  und 
  Chem. 
  Bd. 
  xlvii. 
  p. 
  121 
  

   (1892). 
  

  

  