﻿178 
  Mr. 
  E. 
  P. 
  Harrison 
  on 
  the 
  Temperature 
  Variation 
  of 
  

  

  couple 
  was 
  either 
  platinum 
  or 
  an 
  alloy 
  of 
  platinum 
  and 
  

   iridium 
  ; 
  he 
  used 
  mercury 
  thermometers 
  up 
  to 
  300°, 
  but 
  did 
  

   not 
  correct 
  for 
  stem- 
  exposure, 
  and 
  the 
  experimental 
  difficulties 
  

   due 
  to 
  chemical 
  action 
  and 
  the 
  measurement 
  of 
  temperature 
  

   made 
  observations 
  above 
  500° 
  quite 
  unreliable. 
  His 
  results 
  

   within 
  a 
  moderate 
  range 
  of 
  temperature 
  (obtained 
  with 
  the 
  

   hot 
  junction 
  in 
  an 
  oil-bath) 
  showed 
  that 
  the 
  E.M.F. 
  tempera- 
  

   ture-curves 
  were 
  made 
  up 
  of 
  a 
  series 
  of 
  "excellent 
  parabolas." 
  

   At 
  higher 
  temperatures 
  a 
  the 
  parabola 
  was 
  slightly 
  steeper 
  

   on 
  the 
  hotter 
  than 
  on 
  the 
  colder 
  side.'"' 
  He 
  also 
  found 
  that 
  

   between 
  200° 
  and 
  the 
  melting-point, 
  two 
  or 
  more 
  distinct 
  

   neutral 
  points 
  existed. 
  

  

  Professors 
  Fleming 
  and 
  Dewar's 
  experiments 
  on 
  the 
  same 
  

   subject"* 
  extend 
  from 
  the 
  temperature 
  of 
  boiling 
  oxygen 
  

   to 
  100° 
  0. 
  Their 
  paper 
  gives 
  the 
  experimental 
  methods 
  used 
  

   and 
  the 
  numerical 
  results, 
  but 
  a 
  discussion 
  of 
  the 
  latter 
  

   is 
  reserved 
  by 
  the 
  authors 
  for 
  a 
  further 
  communication. 
  

   The 
  temperature 
  of 
  the 
  " 
  hot 
  junction 
  " 
  in 
  their 
  experiments 
  

   was 
  measured 
  by 
  a 
  platinum 
  thermometer, 
  and 
  their 
  results 
  

   were 
  expressed 
  on 
  the 
  platinum 
  scale. 
  

  

  Holborn 
  and 
  Day 
  f 
  compared 
  a 
  platinum-platinum-rhodium 
  

   couple 
  with 
  the 
  air-thermometer, 
  and 
  J 
  tested 
  a 
  number 
  of 
  

   similar 
  couples, 
  but 
  were 
  unable 
  to 
  obtain 
  consistent 
  results 
  

   with 
  oxidizable 
  metals 
  at 
  high 
  temperatures. 
  They 
  adopt 
  

   Tait's 
  method 
  of 
  representing 
  their 
  results 
  by 
  portions 
  of 
  

   parabolas 
  for 
  limited 
  ranges 
  of 
  temperature. 
  

  

  Stansfield 
  §, 
  from 
  observations 
  on 
  couples 
  formed 
  of 
  plati- 
  

   num 
  and 
  platinum-iridium 
  or 
  platinum-rhodium 
  alloys, 
  found 
  

  

  that 
  for 
  these 
  couples 
  the 
  Peltier 
  effect 
  | 
  T 
  -=- 
  ) 
  approximated 
  

  

  closely 
  to 
  a 
  linear 
  function 
  of 
  the 
  temperature, 
  but 
  that 
  the 
  

  

  thermoelectric 
  power 
  ( 
  -^- 
  ) 
  could 
  not 
  be 
  so 
  represented. 
  

  

  Description 
  of 
  Apparatus. 
  

  

  A. 
  Measurement 
  of 
  E.M.F. 
  

  

  An 
  ordinary 
  potentiometer 
  method 
  was 
  used 
  and 
  the 
  

   general 
  arrangement 
  of 
  apparatus 
  is 
  shown 
  in 
  fig. 
  1. 
  A 
  two- 
  

   metre 
  bridge-wire 
  BP 
  was 
  connected 
  in 
  series 
  with 
  a 
  resist- 
  

   ance-box 
  R, 
  and 
  a 
  1 
  ohm 
  rheostat 
  O. 
  The 
  P.D. 
  due 
  to 
  the 
  

  

  * 
  Phil. 
  Mag. 
  1895, 
  vol. 
  xl. 
  pp. 
  95-119. 
  

   f 
  Wiedemann's 
  Annalen, 
  August 
  1899. 
  

   % 
  Berlin 
  Academy, 
  July 
  1900. 
  

   § 
  Phil. 
  Mag. 
  1898, 
  vol. 
  xlviii. 
  

  

  