﻿Thermal 
  Expansion 
  of 
  Porcelain. 
  

  

  633 
  

  

  such 
  as 
  to 
  make 
  its 
  use 
  in 
  the 
  specified 
  cases 
  most 
  convenient 
  

   and 
  valuable. 
  Given, 
  however, 
  a 
  substance 
  capable 
  of 
  taking 
  

   a 
  good 
  polish, 
  and 
  whose 
  thermal 
  deformation 
  is 
  approxi- 
  

   mately 
  known, 
  the 
  compensator 
  is 
  both 
  unnecessary 
  and 
  

   inadvisable. 
  The 
  reflecting 
  surface 
  of 
  the 
  aluminium 
  disk 
  

   used 
  in 
  these 
  determinations 
  was 
  minutely 
  concave, 
  affording 
  

   bands 
  slightly 
  curved. 
  This 
  is 
  an 
  advantage, 
  as 
  it 
  enables 
  

   the 
  observer 
  to 
  know 
  at 
  once 
  whether 
  expansion 
  or 
  contraction 
  

   is 
  occurring, 
  according 
  as 
  the 
  bands 
  move 
  from 
  or 
  to 
  their 
  

   centre 
  of 
  curvature. 
  

  

  The 
  specimen 
  of 
  Bedford's 
  tube 
  used 
  by 
  Chappuis 
  was 
  

   15 
  mm. 
  long, 
  and 
  was 
  furnished 
  at 
  its 
  two 
  ends 
  with 
  plane 
  

   surfaces. 
  The 
  one 
  placed 
  uppermost 
  was 
  polished 
  as 
  far 
  as 
  

   possible, 
  and 
  the 
  fringes 
  of 
  sodium 
  light 
  obtained 
  with 
  this 
  

   annulus 
  were 
  used 
  for 
  his 
  determinations. 
  

  

  The 
  author 
  had 
  an 
  ample 
  amount 
  of 
  Bedford's 
  tube 
  to 
  

   furnish 
  three 
  specimens 
  for 
  investigation. 
  They 
  were 
  cut 
  in 
  

   lengths 
  somewhat 
  over 
  12 
  mm. 
  and 
  ground 
  with 
  truly 
  plane 
  

   ends 
  by 
  Messrs. 
  Troughton 
  and 
  Simms, 
  the 
  makers 
  of 
  the 
  

   dilatometer. 
  It 
  was 
  not 
  found 
  possible 
  to 
  get 
  a 
  satisfactory 
  

   polish 
  ; 
  and 
  this 
  has 
  been 
  shown 
  to 
  be 
  unnecessary. 
  Portions 
  

   of 
  the 
  two 
  ends 
  were 
  then 
  ground 
  out 
  somewhat, 
  at 
  three 
  

   places 
  in 
  each 
  case, 
  leaving 
  three 
  equidistant 
  small 
  patches 
  of 
  

   the 
  original 
  truly 
  plane 
  end, 
  so 
  that 
  the 
  specimens 
  might 
  

   stand 
  with 
  three-point 
  contact 
  on 
  the 
  truly 
  plane 
  polished 
  

   surface 
  of 
  the 
  platinum-iriclium 
  table 
  (that 
  side 
  of 
  the 
  table 
  

   being 
  employed 
  which 
  was 
  not 
  furnished 
  with 
  projecting 
  

   points). 
  Likewise 
  the 
  aluminium 
  disk 
  lay, 
  when 
  put 
  into 
  

   position, 
  with 
  three-point 
  contact 
  on 
  the 
  upper 
  porcelain 
  

   blunt 
  points. 
  The 
  diameter 
  of 
  the 
  pieces 
  of 
  porcelain 
  was 
  

   about 
  17 
  mm., 
  and 
  the 
  walls 
  were 
  about 
  3 
  mm. 
  thick 
  and 
  

   fairly 
  uniform, 
  the 
  hole 
  being 
  as 
  nearly 
  as 
  possible 
  central. 
  

  

  Fk. 
  1. 
  

  

  The 
  arrangement 
  of 
  the 
  tripod 
  and 
  its 
  contents 
  wil 
  

   from 
  the 
  accompanying 
  illustration 
  (fig. 
  1). 
  

  

  be 
  clear 
  

  

  