﻿642 
  On 
  the 
  Thermal 
  Expansion 
  of 
  Porcelain. 
  

  

  attained, 
  is 
  the 
  maximum 
  which 
  can 
  with 
  safety 
  be 
  employed 
  

   in 
  an 
  interference-dilatometer. 
  

  

  The 
  conclusion 
  now 
  arrived 
  at 
  with 
  respect 
  to 
  Bayeux 
  

   porcelain 
  is 
  very 
  similar 
  to 
  that 
  derived 
  from 
  a 
  research 
  by 
  

   Holborn 
  and 
  Day 
  {Ann. 
  der 
  Phys. 
  una 
  1 
  Chem. 
  vol 
  ii. 
  p. 
  505, 
  

   1900), 
  on 
  the 
  expansion 
  of 
  Berlin 
  porcelain 
  between 
  0° 
  and 
  

   1000°. 
  A 
  rod 
  of 
  unglazed 
  porcelain 
  483 
  mm. 
  long 
  was 
  

   employed, 
  and 
  the 
  method 
  of 
  the 
  comparator 
  used 
  in 
  making 
  

   the 
  determinations 
  of 
  its 
  expansion. 
  It 
  was 
  found 
  that 
  

   between 
  250° 
  and 
  625° 
  the 
  results 
  are 
  fairly 
  reproduced 
  by 
  

   the 
  expression 
  

  

  l 
  t 
  = 
  l 
  [l 
  + 
  (2954)5 
  4- 
  1'125* 
  2 
  )10- 
  9 
  ]. 
  

  

  But 
  below 
  250° 
  the 
  values 
  afforded 
  by 
  the 
  formula 
  are 
  too 
  

   high. 
  Again, 
  between 
  700° 
  and 
  900° 
  the 
  determinations 
  are 
  

   not 
  accurately 
  reproduced 
  by 
  the 
  formula 
  ; 
  but 
  agreement 
  

   becomes 
  better 
  again 
  at 
  1000°. 
  

  

  Prof. 
  Callendar, 
  in 
  discussing 
  the 
  relative 
  results 
  of 
  

   Chappuis 
  and 
  Bedford 
  (Proceedings 
  of 
  the 
  Physical 
  Society, 
  

   Nov. 
  22nd, 
  1901), 
  stated 
  that 
  " 
  Bedford's 
  results 
  agreed 
  very 
  

   fairly, 
  allowing 
  for 
  difference 
  of 
  material, 
  with 
  Holborn 
  and 
  

   Day's 
  from 
  200° 
  to 
  600°, 
  and 
  that 
  both 
  differed 
  from 
  those 
  of 
  

   Chappuis 
  between 
  C 
  and 
  80° 
  when 
  extrapolated 
  in 
  a 
  precisely 
  

   similar 
  manner. 
  It 
  was 
  quite 
  possible, 
  as 
  he 
  (Prof. 
  Callendar) 
  

   had 
  previously 
  suggested, 
  that 
  the 
  expansion 
  of 
  porcelain 
  

   between 
  0° 
  and 
  100° 
  was 
  anomalous. 
  It 
  appeared 
  certain 
  

   that 
  some 
  anomaly 
  in 
  the 
  expansion 
  at 
  800° 
  was 
  indicated 
  in 
  

   the 
  experiments 
  of 
  Bedford, 
  and 
  also 
  in 
  those 
  of 
  Holborn 
  

   and 
  Day." 
  

  

  The 
  supposition 
  of 
  Prof. 
  Callendar, 
  as 
  to 
  an 
  anomalous 
  

   expansion 
  between 
  0° 
  and 
  100° 
  in 
  the 
  case 
  of 
  porcelain, 
  

   appears 
  to 
  be 
  w 
  T 
  ell 
  founded, 
  the 
  anomaly 
  consisting 
  in 
  a 
  very 
  

   much 
  larger 
  increment 
  for 
  this 
  interval 
  than 
  for 
  the 
  higher 
  

   interval 
  200° 
  to 
  600°. 
  The 
  author 
  fully 
  agrees 
  with 
  Dr. 
  

   Chappuis 
  that 
  the 
  thermal 
  expansion 
  of 
  the 
  porcelain 
  employed 
  

   for 
  air-thermometers, 
  whether 
  of 
  the 
  Bayeux 
  or 
  Berlin 
  

   varieties, 
  cannot 
  be 
  represented 
  by 
  a 
  two-term 
  function 
  

   throughout 
  an 
  interval 
  of 
  temperature 
  exceeding 
  a 
  very 
  few 
  

   hundred 
  degrees. 
  On 
  no 
  account 
  can 
  such 
  an 
  expression 
  be 
  

   safely 
  employed 
  outside 
  the 
  interval 
  for 
  which 
  it 
  was 
  actually 
  

   determined, 
  extrapolation 
  leading 
  to 
  grave 
  errors 
  in 
  the 
  case 
  

   of 
  porcelain. 
  

  

  