﻿186 
  Mr. 
  E. 
  P. 
  Harrison 
  on 
  the 
  Temperature 
  Variation 
  of 
  

  

  and 
  copper-iron 
  curves 
  corresponding 
  to 
  definite 
  temperatures. 
  

   It 
  has 
  no 
  neutral 
  point 
  above 
  0°, 
  though 
  at 
  the 
  temperature 
  

   of 
  liquid 
  air 
  there 
  is 
  an 
  indication 
  of 
  one. 
  The 
  fact 
  that 
  no 
  

   neutral 
  point 
  occurs 
  in 
  the 
  iron-nickel 
  " 
  line 
  " 
  is 
  also 
  evident 
  

   from 
  the 
  fact 
  that 
  copper-nickel 
  and 
  copper-iron 
  thermo- 
  

   electric-power 
  curves, 
  when 
  plotted 
  on 
  the 
  same 
  scale, 
  do 
  not 
  

   intersect. 
  

  

  The 
  E.M.F. 
  curve 
  is 
  a 
  nearly 
  linear 
  function 
  of 
  the 
  tem- 
  

   perature 
  up 
  to 
  about 
  900°, 
  when 
  a 
  decrease 
  in 
  the 
  E.M.F. 
  is 
  

   noticed 
  (fig. 
  5). 
  Above 
  700° 
  the 
  curve 
  was 
  obtained 
  by 
  

   direct 
  observation 
  of 
  the 
  E.M.F. 
  of 
  a 
  fresh 
  nickel-iron 
  

   couple, 
  the 
  copper 
  being 
  omitted 
  in 
  order 
  to 
  avoid 
  the 
  

   uncertainty 
  due 
  to 
  its 
  evaporation, 
  which 
  became 
  quite 
  

   appreciable 
  at 
  900° 
  C. 
  The 
  junction 
  of 
  iron 
  and 
  nickel 
  was 
  

   made 
  by 
  fusing 
  the 
  wires 
  in 
  the 
  electric 
  arc. 
  

  

  The 
  dotted 
  parts 
  of 
  the 
  nickel 
  and 
  iron 
  " 
  lines 
  " 
  (fig. 
  3) 
  

   were 
  obtained 
  when 
  the 
  couples 
  were 
  known 
  to 
  be 
  affected 
  

   by 
  oxidation, 
  i. 
  e. 
  when 
  the 
  vacuum 
  was 
  known 
  to 
  be 
  faulty. 
  

   The 
  shift 
  only 
  occurred 
  at 
  the 
  higher 
  temperatures. 
  The 
  

   E.M.F., 
  when 
  determined 
  on 
  the 
  same 
  " 
  oxidized 
  " 
  couple 
  

   for 
  lower 
  temperatures 
  (such 
  as 
  200°), 
  coincided 
  with 
  the 
  

   original 
  values. 
  A 
  ny 
  error 
  in 
  measuring 
  temperature 
  or 
  

   E.M.F. 
  would 
  be 
  constant 
  for 
  the 
  copper-nickel 
  and 
  copper- 
  

   iron 
  "lines." 
  The 
  observed 
  facts, 
  however, 
  show 
  that 
  the 
  

   effect 
  is 
  not 
  the 
  same 
  in 
  the 
  two 
  cases. 
  The 
  vertical 
  dis- 
  

   placement 
  of 
  the 
  nickel 
  curve 
  is 
  about 
  4-25 
  microvolts, 
  while 
  

   at 
  the 
  same 
  temperature 
  that 
  of 
  the 
  iron 
  curve 
  is 
  about 
  

   120 
  microvolts, 
  the 
  horizontal 
  shift 
  in 
  the 
  first 
  case 
  being 
  

   14°, 
  and 
  in 
  the 
  second 
  case 
  18°, 
  in 
  the 
  same 
  direction. 
  

  

  The 
  effect 
  for 
  Cu-Fe 
  is 
  as 
  if 
  an 
  applied 
  E.M.F. 
  were 
  

   acting 
  so 
  as 
  to 
  produce 
  a 
  current 
  in 
  the 
  same 
  direction 
  as 
  

   that 
  due 
  to 
  the 
  junction. 
  Thus 
  the 
  fact 
  that 
  the 
  effect 
  is 
  an 
  

   increase 
  of 
  ordinate 
  above 
  500°, 
  and 
  a 
  decrease 
  of 
  ordinate 
  

   below 
  500° 
  (in 
  the 
  case 
  of 
  Fe-Cu) 
  is 
  accounted 
  for. 
  For 
  

   Cu-Ni 
  there 
  is 
  always 
  a 
  decrease, 
  such 
  as 
  might 
  be 
  due 
  to 
  an 
  

   E.M.F. 
  giving 
  a 
  current 
  in 
  the 
  opposite 
  direction 
  to 
  that 
  due 
  

   to 
  the 
  junction. 
  Moreover, 
  the 
  two 
  dotted 
  bits 
  are 
  not 
  abso- 
  

   lutely 
  parallel 
  to 
  the 
  original 
  curves, 
  making 
  it 
  still 
  more 
  

   probable 
  that 
  the 
  explanation 
  of 
  the 
  shift 
  is 
  to 
  be 
  found 
  

   in 
  the 
  fact 
  that 
  chemical 
  action 
  took 
  place 
  and 
  altered 
  the 
  

   E.M.F.'s. 
  

  

  When 
  the 
  couples 
  were 
  heated 
  in 
  hydrogen, 
  so 
  as 
  to 
  

   preclude 
  the 
  possibility 
  of 
  oxidation, 
  those 
  points 
  taken 
  below 
  

   895° 
  C. 
  lay 
  on 
  the 
  main 
  curve, 
  which 
  bears 
  out 
  the 
  idea 
  that 
  

   the 
  shift 
  in 
  question 
  is 
  due 
  to 
  oxidation. 
  

  

  Above 
  895° 
  other 
  effects 
  were 
  observed 
  when 
  the 
  couples 
  

  

  