﻿High 
  Temperatures. 
  167 
  

  

  The 
  glass 
  bulb 
  proved 
  to 
  be 
  exceedingly 
  well 
  adapted 
  to 
  the 
  

   purpose 
  up 
  to 
  500°, 
  showing 
  no 
  appreciable 
  change 
  in 
  the 
  zero 
  

   point 
  after 
  being 
  repeatedly 
  heated, 
  while 
  with 
  porcelain, 
  even 
  

   when. 
  nitrogen 
  was 
  afterward 
  substituted 
  for 
  the 
  hydrogen, 
  the 
  

   zero 
  point 
  did 
  not 
  remain 
  constant, 
  but 
  rose 
  slightly 
  after 
  each 
  

   heating, 
  denoting 
  an 
  increase 
  in 
  the 
  gas 
  pressure. 
  Further- 
  

   more^ 
  the 
  expansion-coefficient 
  of 
  the 
  gas 
  between 
  0° 
  and 
  100° 
  

   often 
  showed 
  changes 
  larger 
  than 
  could 
  be 
  accounted 
  for 
  by 
  

   the 
  errors 
  in 
  observation. 
  

  

  Such 
  changes, 
  even 
  when 
  they 
  remain 
  small 
  and 
  the 
  tem- 
  

   peratures 
  measured 
  with 
  the 
  bulbs 
  agree 
  fairly 
  among 
  them- 
  

   selves, 
  are 
  persistent 
  and 
  disturbing 
  factors 
  in 
  the 
  measure- 
  

   ment 
  of 
  high 
  temperatures. 
  

  

  Another 
  misfortune 
  arises 
  from 
  the 
  fact 
  that 
  porcelain 
  is 
  not 
  

   a 
  definite 
  chemical 
  compound, 
  and 
  bulbs 
  from 
  different 
  sources 
  

   may 
  behave 
  quite 
  differently 
  ; 
  even 
  bulbs 
  ordered 
  at 
  different 
  

   times 
  from 
  the 
  same 
  source 
  have 
  the 
  same 
  unfortunate 
  

   tendency. 
  

  

  These, 
  with 
  the 
  above 
  mentioned 
  fact 
  that 
  bulbs 
  glazed 
  on 
  

   the 
  inside 
  do 
  not 
  admit 
  of 
  use 
  above 
  1100° 
  at 
  all, 
  and 
  the 
  

   unglazed 
  ones 
  only 
  with 
  difficulty 
  and 
  under 
  conditions 
  of 
  

   greatly 
  diminished 
  sensitiveness, 
  led 
  us 
  to 
  direct 
  our 
  attention 
  

   once 
  more 
  to 
  metallic 
  bulbs 
  and 
  to 
  investigate 
  the 
  behavior 
  of 
  

   a 
  platin-iridium 
  cylinder 
  under 
  the 
  above 
  conditions. 
  

  

  From 
  the 
  results 
  thus 
  far 
  obtained 
  we 
  do 
  not 
  hesitate 
  to 
  say 
  

   that 
  in 
  our 
  opinion 
  such 
  a 
  bulb 
  is 
  certainly 
  to 
  be 
  preferred 
  to 
  

   porcelain 
  for 
  the 
  higher 
  temperatures. 
  

  

  This 
  bulb 
  filled 
  with 
  nitrogen 
  and 
  heated 
  to 
  1300° 
  showed 
  a 
  

   constancy 
  of 
  the 
  zero 
  point 
  and 
  coefficient 
  of 
  gas 
  expansion 
  

   such 
  as 
  we 
  have 
  only 
  been 
  able 
  to 
  obtain 
  from 
  glass 
  bulbs 
  at 
  

   comparatively 
  low 
  temperatures. 
  This 
  result 
  is, 
  however, 
  in 
  no 
  

   small 
  degree 
  dependent 
  upon 
  arrangements 
  for 
  heating 
  by 
  

   electricity 
  and 
  the 
  consequent 
  freedom 
  from 
  all 
  sorts 
  of 
  com- 
  

   bustion 
  products, 
  which 
  if 
  present, 
  as 
  we 
  observed 
  repeatedly, 
  

   pass 
  rapidly 
  through 
  the 
  bulb 
  wall, 
  even 
  against 
  an 
  excess 
  

   pressure 
  within. 
  

  

  For 
  comparison 
  of 
  the 
  gas-thermometer 
  readings 
  among 
  

   themselves, 
  they 
  were 
  referred 
  to 
  a 
  platinum, 
  platin-rhodium 
  

   thermo-element 
  observed 
  simultaneously. 
  

  

  For 
  temperatures 
  up 
  to 
  750° 
  bulb 
  and 
  thermo-element 
  were 
  

   dipped 
  into 
  a 
  bath 
  of 
  nitre 
  and 
  in 
  the 
  vicinity 
  of 
  900° 
  a 
  zinc 
  

   vapor 
  bath 
  was 
  tried. 
  As 
  no 
  bath 
  is 
  yet 
  available 
  for 
  tempera- 
  

   tures 
  above 
  that 
  point 
  which 
  does 
  not 
  seriously 
  injure 
  the 
  walls 
  

   of 
  the 
  bulb, 
  the 
  remaining 
  temperatures 
  were 
  attained 
  electri- 
  

   cally 
  by 
  heating 
  a 
  coil 
  of 
  nickel 
  wire 
  wound 
  upon 
  a 
  thin 
  tube 
  

   of 
  porcelain 
  or 
  clay. 
  The 
  last 
  proved 
  the 
  most 
  convenient 
  of 
  

   all 
  iDoth 
  for 
  producing 
  and 
  maintaining 
  constant 
  any 
  tempera- 
  

   ture 
  within" 
  the 
  range 
  of 
  the 
  present 
  series. 
  

  

  