﻿188 
  Holhorn 
  and 
  Day 
  — 
  Gas 
  Thermometer 
  at 
  

  

  the 
  elements 
  through 
  the 
  tube 
  carrying 
  the 
  coil, 
  the 
  three 
  bits 
  

   of 
  fire 
  clay 
  which 
  served 
  to 
  hold 
  the 
  bulb 
  in 
  place 
  in 
  the 
  oven, 
  

   and 
  the 
  metallic 
  bulb 
  itself. 
  The 
  error 
  due 
  to 
  this 
  leakage 
  

   within 
  the 
  range 
  of 
  temj3erature 
  observed 
  (to 
  1150°) 
  was 
  

   readily 
  eliminated 
  by 
  reversing 
  the 
  current 
  in 
  the 
  oven.* 
  

  

  For 
  temperatures 
  above 
  this 
  point 
  a 
  larger 
  oven 
  is 
  being 
  pre- 
  

   pared. 
  The 
  oven 
  as 
  used 
  up 
  to 
  the 
  present 
  is 
  shown 
  in 
  cross 
  

   section 
  in 
  fig. 
  1. 
  

  

  TemjperatuTe 
  variations 
  along 
  the 
  hiolb. 
  — 
  The 
  difficulty 
  

   caused 
  by 
  the 
  variations 
  of 
  temperature 
  within 
  the 
  oven 
  we 
  

   succeeded 
  in 
  overcoming 
  for 
  the 
  most 
  part 
  by 
  varying 
  the 
  

   winding 
  of 
  the 
  coil 
  somewhat. 
  

  

  Instead 
  of 
  a 
  uniform 
  space 
  between 
  the 
  separate 
  turns 
  of 
  

   wire 
  approximately 
  equal 
  to 
  the 
  diameter 
  of 
  the 
  wire 
  (2"™) 
  as 
  

   we 
  had 
  previously 
  used 
  it, 
  it 
  was 
  only 
  so 
  wound 
  at 
  the 
  ends, 
  

   the 
  turns 
  in 
  the 
  middle 
  for 
  some 
  10^"^ 
  being 
  twice 
  as 
  far 
  apart 
  

   {4z'^'^) 
  and 
  coming 
  gradually 
  closer 
  together 
  toward 
  the 
  ends, 
  i. 
  e. 
  

   roughly, 
  a 
  logarithmically 
  wound 
  coil. 
  

  

  Before 
  the 
  oven 
  thus 
  equipped 
  was 
  used 
  for 
  thermometric 
  

   measurements 
  the 
  differences 
  of 
  temperature 
  at 
  various 
  points 
  

   within 
  were 
  determined 
  as 
  follows 
  : 
  Two 
  thermo-elements 
  

   were 
  melted 
  together 
  at 
  their 
  junctions 
  and 
  drawn 
  from 
  point 
  

   to 
  point 
  through 
  the 
  empty 
  tube 
  at 
  various 
  temjperatures. 
  

   Especially 
  thin 
  elements 
  were 
  chosen 
  for 
  this 
  purpose, 
  0-25°^"' 
  

   diameter 
  instead 
  of 
  the 
  usual 
  0*6'''"', 
  in 
  order 
  to 
  reduce 
  the 
  

   heat 
  conducted 
  away 
  from 
  the 
  junction 
  by 
  the 
  wires 
  themselves 
  

   to 
  a 
  minimum. 
  

  

  Table 
  YIII 
  contains 
  the 
  results 
  of 
  this 
  trial 
  for 
  the 
  old 
  and 
  

   the 
  new 
  oven 
  coils, 
  the 
  dimensions 
  of 
  the 
  tube 
  being 
  approxi- 
  

   mately 
  the 
  same 
  for 
  both. 
  In 
  the 
  first 
  case 
  (uniform 
  winding) 
  

   the 
  conditions 
  improve 
  with 
  the 
  higher 
  temperatures, 
  which 
  

   seems 
  to 
  indicate 
  that 
  the 
  tendency 
  to 
  a 
  condition 
  of 
  equili- 
  

   brium 
  through 
  radiation 
  is 
  stronger 
  than 
  the 
  loss 
  at 
  the 
  ends 
  

   from 
  conductivity. 
  The 
  presence 
  of 
  the 
  platin-iridium 
  bulb 
  in 
  

   the 
  tube 
  diminishes 
  the 
  irregularities 
  some 
  3° 
  at 
  the 
  lower 
  

   temperatures, 
  at 
  1000° 
  scarcely 
  any. 
  The 
  placing 
  of 
  fire-clay 
  

   partitions 
  in 
  the 
  tube 
  close 
  to 
  the 
  ends 
  of 
  the 
  bulb 
  also 
  in- 
  

   fluences 
  the 
  results 
  between 
  500° 
  and 
  900° 
  ; 
  between 
  these 
  

   limits 
  the 
  temperature 
  as 
  measured 
  by 
  .the 
  gas 
  thermometer 
  is 
  

   raised 
  some 
  3*5° 
  without 
  changing 
  the 
  temperature 
  relation 
  

   between 
  the 
  middle 
  and 
  end 
  of 
  the 
  bulb 
  ; 
  that 
  is, 
  the 
  parti- 
  

   tions 
  affect 
  only 
  the 
  radiation 
  from 
  the 
  end 
  surfaces. 
  

  

  * 
  A 
  simple 
  reversal 
  of 
  the 
  current 
  is 
  insufficient 
  for 
  correcting 
  a 
  large 
  error 
  

   from 
  this 
  cause, 
  as 
  the 
  resistance 
  of 
  porcelain 
  increases 
  with 
  the 
  time 
  of 
  passage 
  

   of 
  the 
  current, 
  and 
  one 
  must, 
  therefore, 
  wait 
  until 
  the 
  resistance 
  is 
  approximately 
  

   constant. 
  In 
  the 
  observations 
  here 
  presented 
  the 
  reversal 
  of 
  the 
  current 
  influ- 
  

   enced 
  the 
  thermo-elements 
  about 
  5° 
  at 
  1000^' 
  and 
  1*5° 
  at 
  1100°, 
  and 
  some 
  live 
  

   minutes 
  were 
  required 
  to 
  produce 
  constant 
  conditions. 
  

  

  