﻿494 
  

  

  Mr. 
  W. 
  C. 
  Roberts 
  on 
  

  

  The 
  most 
  important 
  of 
  the 
  corrections 
  applied 
  to 
  these 
  results 
  was 
  that 
  

   for 
  change 
  of 
  volume 
  of 
  the 
  iron 
  vessel 
  which 
  attended 
  the 
  introduction 
  

   of 
  the 
  molten 
  metal. 
  

  

  Different 
  qualities 
  of 
  wrought 
  iron 
  vary 
  considerably 
  as 
  to 
  dilatation 
  

   by 
  heat. 
  This 
  fact, 
  together 
  with 
  the 
  known 
  increase 
  in 
  the 
  expansion 
  

   at 
  high 
  temperatures, 
  rendered 
  it 
  necessary 
  to 
  determine 
  the 
  mean 
  coeffi- 
  

   cient 
  between 
  0° 
  C. 
  and 
  the 
  melting-point 
  of 
  silver. 
  For 
  this 
  purpose 
  a 
  

   modification 
  of 
  Earn 
  sd 
  en's 
  method 
  was 
  adopted, 
  the 
  iron 
  being 
  placed 
  in 
  

   a 
  graphite 
  trough 
  and 
  surrounded 
  by 
  molten 
  silver. 
  The 
  micrometer- 
  

   reading 
  was 
  taken 
  when 
  the 
  length 
  of 
  the 
  iron 
  remained 
  for 
  a 
  short 
  

   period 
  invariable, 
  as 
  this 
  was 
  the 
  true 
  solidifying-point 
  of 
  the 
  silver, 
  

   the 
  loss 
  of 
  the 
  latent 
  heat 
  of 
  liquefaction 
  rendering 
  the 
  temperature 
  

   constant. 
  

  

  A 
  great 
  number 
  of 
  experiments 
  were 
  made 
  ; 
  and 
  although 
  they 
  were 
  

   attended 
  with 
  much 
  difficulty, 
  I 
  believe 
  the 
  following 
  results 
  to 
  be 
  trust- 
  

   worthy 
  (the 
  numbers 
  represent 
  the 
  mean 
  coefficient 
  of 
  linear 
  expan- 
  

   sion 
  "per 
  1° 
  0. 
  of 
  the 
  Low-Moor 
  iron 
  employed, 
  up 
  to 
  the 
  temperature 
  

   of 
  melting 
  silver) 
  : 
  — 
  

  

  •00001242, 
  

   ♦00001254, 
  

   •00001215, 
  

   •00001219, 
  

   •00001271, 
  

   Mean.... 
  -00001240, 
  

  

  which 
  gives 
  a 
  mean 
  coefficient 
  of 
  cubical 
  expansion 
  

  

  •00003720. 
  

  

  This 
  result 
  is 
  considerably 
  higher 
  than 
  that 
  of 
  Einmann, 
  who 
  gives 
  

   0*00002808 
  as 
  the 
  mean 
  coefficient 
  of 
  cubical 
  expansion 
  of 
  wrought 
  iron 
  

   between 
  15° 
  C. 
  and 
  " 
  a 
  white 
  or 
  welding 
  heat." 
  

  

  The 
  results, 
  of 
  the 
  experiments 
  made 
  with 
  a 
  view 
  to 
  ascertain 
  the 
  

   densities 
  of 
  pure 
  silver 
  and 
  of 
  Levol's 
  homogeneous 
  alloy 
  when 
  in 
  

   a 
  molten 
  state 
  are 
  given 
  in 
  the 
  accompanying 
  Table 
  (p. 
  495). 
  This 
  

   alloy 
  was 
  chosen 
  for 
  the 
  experiment, 
  as 
  its 
  density 
  when 
  solid 
  very 
  

   nearly 
  agrees 
  with 
  that 
  calculated 
  from 
  the 
  densities 
  of 
  the 
  constituent 
  

   metals. 
  

  

  The 
  cubic 
  dilatation 
  was, 
  in 
  the 
  case 
  of 
  pure 
  silver, 
  in 
  the 
  ratio 
  of 
  

   9*4612 
  : 
  10*57. 
  Deducing 
  from 
  this 
  the 
  mean 
  coefficient 
  through, 
  say, 
  

   1050° 
  0., 
  we 
  obtain 
  

  

  ■ 
  0*00011164 
  per 
  1° 
  C. 
  

   The 
  coefficient 
  of 
  linear 
  expansion 
  was 
  therefore 
  

  

  0*00003721. 
  

  

  