﻿152 
  

  

  Prof. 
  Kuenen 
  and 
  Mr. 
  Robson 
  

  

  and 
  the 
  temperature 
  change 
  of 
  coils 
  and 
  wire, 
  we 
  found 
  that 
  

   at 
  the 
  steam-point 
  differences 
  of 
  as 
  much 
  as 
  0°*06 
  were 
  

   possible. 
  

  

  (a) 
  Resistance 
  Measurements. 
  — 
  The 
  coils 
  and 
  bridge-wire 
  

   were 
  of 
  german-silver, 
  the 
  smallest 
  box-coil 
  being 
  0*1 
  ohm, 
  

   and 
  the 
  wire 
  having 
  a 
  resistance 
  of 
  0*22 
  ohm 
  per 
  metre. 
  

   The 
  temperature-coefficient 
  of 
  the 
  box-coils 
  was 
  about 
  '00040, 
  

   and 
  all 
  readings 
  were 
  reduced 
  to 
  a 
  standard 
  temperature 
  of 
  

   15° 
  C. 
  The 
  coils 
  were 
  immersed 
  in 
  paraffin-oil, 
  and 
  their 
  

   temperature 
  read 
  on 
  a 
  mercury 
  thermometer 
  divided 
  to 
  o, 
  05. 
  

   It 
  is 
  to 
  be 
  noticed 
  that 
  german-silver 
  is 
  on 
  account 
  of 
  its 
  high 
  

   temperature-coefficient 
  very 
  unsuitable 
  for 
  measurements 
  of 
  

   this 
  kind. 
  With 
  our 
  apparatus 
  an 
  error 
  of 
  o, 
  2 
  in 
  the 
  box 
  

   temperature 
  — 
  an 
  error 
  which 
  is 
  quite 
  possible 
  when 
  the 
  room 
  

   temperature 
  is 
  changing 
  rapidly 
  — 
  corresponded 
  at 
  the 
  steam- 
  

   point 
  to 
  a 
  resistance 
  error 
  of 
  *00032 
  ohm, 
  i. 
  e. 
  the 
  platinum 
  

   temperature 
  would 
  be 
  wrong 
  by 
  o, 
  032. 
  At 
  the 
  ice-point 
  

   the 
  same 
  error 
  in 
  box 
  temperature 
  would 
  cause 
  an 
  error 
  in 
  

   the 
  platinum 
  temperature 
  of 
  o, 
  023. 
  The 
  expensive 
  platinum- 
  

   silver 
  alloy 
  would 
  only 
  reduce 
  these 
  errors 
  to 
  three-fifths 
  of 
  

   their 
  values 
  ; 
  and 
  we 
  would 
  recommend, 
  therefore, 
  that 
  the 
  

   coils 
  of 
  boxes 
  meant 
  for 
  use 
  with 
  the 
  platinum 
  thermometer 
  

   be 
  made 
  of 
  some 
  alloy, 
  such 
  as 
  manganin, 
  whose 
  temperature- 
  

   coefficient 
  is 
  only 
  '00002, 
  so 
  that 
  the 
  above 
  errors 
  would 
  be 
  

   reduced 
  to 
  one-twentieth 
  of 
  their 
  values. 
  

  

  (b) 
  Stem 
  Immersion. 
  — 
  The 
  platinum 
  thermometer 
  is 
  said 
  

   to 
  be 
  free 
  from 
  errors 
  due 
  to 
  incomplete 
  immersion 
  of 
  the 
  

   stem, 
  being 
  on 
  this 
  account 
  superior 
  to 
  the 
  mercury 
  ther- 
  

   mometer 
  *. 
  Our 
  experience 
  with 
  regard 
  to 
  this 
  is 
  entirely 
  in 
  

   accord 
  with 
  that 
  of 
  Chreef. 
  Defining, 
  as 
  he 
  does, 
  the 
  

   " 
  bulb 
  " 
  to 
  be 
  that 
  part 
  of 
  the 
  thermometer 
  extending 
  to 
  the 
  

   first 
  mica 
  disk 
  — 
  in 
  our 
  thermometer 
  this 
  is 
  1 
  cm. 
  above 
  the 
  

   top 
  of 
  the 
  spiral 
  — 
  we 
  find 
  the 
  following 
  errors 
  due 
  to 
  

   insufficient 
  immersion 
  : 
  — 
  

  

  Immersion. 
  

  

  Ice-point. 
  

  

  Steam-point. 
  

  

  Bulb 
  -f- 
  Hems. 
  

  

  +o°-n 
  

  

  -0°-26 
  

  

  ., 
  + 
  3£ 
  „ 
  

  

  •05 
  

  

  

  » 
  + 
  7 
  „ 
  

  

  •03 
  

  

  — 
  

  

  » 
  + 
  12 
  „ 
  

  

  •00 
  

  

  00 
  

  

  * 
  Griffiths, 
  ' 
  Nature,' 
  vol. 
  liii. 
  p. 
  45. 
  

   t 
  Proceedings 
  Roy. 
  tSoc. 
  4.'3o. 
  p. 
  28. 
  

  

  