﻿182 
  Mr. 
  E. 
  P. 
  Harrison 
  on 
  the 
  Temperature 
  Variation 
  of 
  

  

  and 
  poured 
  over 
  the 
  cork 
  and 
  top 
  of 
  the 
  tube, 
  which 
  was 
  then 
  

   warmed 
  till 
  the 
  glue 
  ran 
  into 
  every 
  hole. 
  Finally 
  a 
  coating 
  

   of 
  beeswax 
  was 
  spread 
  over 
  the 
  glue 
  — 
  a 
  precaution 
  which 
  

   was 
  found 
  to 
  be 
  very 
  necessary 
  to 
  the 
  preservation 
  of 
  the 
  

   vacuum, 
  though 
  as 
  beeswax 
  melts 
  at 
  a 
  low 
  temperature, 
  

   the 
  top 
  of 
  the 
  hot-junction 
  tube 
  had 
  to 
  be 
  kept 
  well 
  protected 
  

   from 
  the 
  heat 
  of 
  the 
  burner. 
  The 
  glass 
  tube 
  which 
  passed 
  

   out 
  of 
  the 
  cork 
  was 
  then 
  sealed 
  directly 
  on 
  to 
  a 
  Fleuss 
  pump. 
  

   Two 
  P 
  2 
  5 
  tubes 
  were 
  interposed 
  between 
  the 
  porcelain 
  tube 
  

   and 
  the 
  pump, 
  and 
  a 
  mercury 
  gauge 
  was 
  set 
  up 
  with 
  a 
  

   barometer 
  by 
  its 
  side 
  for 
  comparison. 
  The 
  wires 
  above 
  

   the 
  hot 
  junction 
  were 
  in 
  the 
  first 
  instance 
  insulated 
  by 
  being 
  

   separately 
  wrapped 
  in 
  several 
  layers 
  of 
  asbestos-paper 
  as 
  far 
  

   up 
  as 
  the 
  mouth 
  of 
  the 
  tube. 
  Subsequently, 
  the 
  wires 
  inside 
  

   the 
  tube 
  were 
  insulated 
  by 
  little 
  circular 
  mica 
  disks 
  instead 
  

   of 
  asbestos, 
  because 
  the 
  latter 
  was 
  found 
  to 
  evolve 
  gas 
  on 
  

   heating 
  which 
  contaminated 
  the 
  couples. 
  With 
  mica 
  insula- 
  

   tion 
  no, 
  difficulty 
  was 
  found 
  in 
  keeping 
  the 
  vacuum 
  as 
  good 
  

   as 
  that 
  above 
  the 
  barometer 
  for 
  several 
  days. 
  

  

  Regulation 
  of 
  the 
  Temperature. 
  

  

  It 
  was 
  thought 
  possible 
  that 
  hysteresis 
  occurred 
  at 
  the 
  

   singular 
  point 
  of 
  the 
  nickel 
  curve. 
  To 
  investigate 
  whether 
  or 
  

   not 
  this 
  was 
  the 
  case, 
  it 
  was 
  necessary 
  to 
  take 
  observations 
  

   with 
  a 
  very 
  slowly-rising 
  or 
  slowly- 
  falling 
  temperature. 
  

   A 
  special 
  apparatus 
  was 
  constructed 
  for 
  regulating 
  the 
  gas- 
  

   pressure, 
  and 
  for 
  slowly 
  turning 
  on 
  or 
  off 
  the 
  gas-supply 
  to 
  

   the 
  small 
  Fletcher 
  burner.. 
  This 
  apparatus 
  worked 
  very 
  

   satisfactorily, 
  although 
  only 
  a 
  small 
  hysteresis 
  effect 
  was 
  

   detected, 
  which 
  is 
  shown 
  by 
  the 
  double 
  line 
  at 
  the 
  summit 
  of 
  

   difference 
  -curve 
  D 
  (M) 
  (fig. 
  4) 
  ; 
  the 
  upper 
  branch 
  was 
  

   obtained 
  with 
  rising, 
  the 
  lower 
  one 
  with 
  falling 
  temperature. 
  

   The 
  difference 
  could 
  not 
  be 
  explained 
  by 
  temperature-lag 
  in 
  

   either 
  the 
  thermometer 
  or 
  the 
  couple. 
  

  

  Method 
  of 
  taking 
  the 
  Observations. 
  

  

  The 
  temperature 
  of 
  the 
  hot 
  junction 
  being 
  steady, 
  or 
  

   cnanging 
  not 
  more 
  than 
  1 
  degree 
  in 
  five 
  minutes, 
  a 
  stop-watch 
  

   was 
  started 
  when 
  the 
  pen 
  of 
  the 
  Recorder 
  was 
  crossing 
  a 
  hori- 
  

   zontal 
  "time-line." 
  The 
  scale 
  of 
  the 
  potentiometer 
  was 
  

   adjusted 
  by 
  reference 
  to 
  the 
  cadmium-cell, 
  and 
  the 
  latter 
  was 
  

   then 
  cut 
  out 
  and 
  the 
  Cu-Ni 
  couple 
  switched 
  into 
  the 
  galvano- 
  

   meter-circuit. 
  Time 
  on 
  the 
  stop-watch 
  at 
  which 
  balance 
  

   occurred, 
  and 
  the 
  bridge-reading 
  at 
  balance 
  being 
  noted, 
  as 
  

   soon 
  after 
  each 
  observation 
  as 
  possible 
  the 
  temperature 
  of 
  

   the 
  cold 
  junction 
  was 
  read 
  to 
  1/10 
  degree. 
  The 
  same 
  

  

  