﻿170 
  Ilolborn 
  and 
  Day 
  — 
  Gas 
  Thermometer 
  at 
  

  

  glass; 
  the 
  observations 
  made 
  with 
  it 
  also 
  gave 
  good 
  results 
  but 
  

   the 
  arrangement 
  was 
  inconvenient 
  and 
  awkward. 
  At 
  a 
  dis- 
  

   tance 
  of 
  1()1-61'"'" 
  below 
  the 
  nickel 
  point, 
  the 
  tube 
  contained 
  a 
  

   second 
  point 
  of 
  colored 
  glass 
  projecting 
  out 
  from 
  the 
  wall 
  of 
  

   the 
  tube 
  itself, 
  for 
  purposes 
  of 
  volumetric 
  measurement. 
  

   The 
  volume 
  included 
  between 
  the 
  points 
  was 
  50-723^'='°. 
  

  

  The 
  longer 
  tube 
  (B) 
  was 
  2"" 
  long 
  and 
  its 
  inside 
  diameter 
  

   12™"'. 
  Its 
  upper 
  end 
  is 
  connected 
  by 
  means 
  of 
  a 
  ground 
  glass 
  

   joint 
  and 
  a 
  metallic 
  return 
  tube 
  with 
  a 
  wash 
  bottle 
  containing 
  

   phosphoric 
  pentoxide, 
  thereby 
  assuring 
  dry 
  air 
  above 
  the 
  mer- 
  

   cury 
  column. 
  

  

  To 
  change 
  the 
  gas 
  in 
  the 
  bulb 
  a 
  water 
  aspirator 
  could 
  be 
  

   attached 
  to 
  the 
  wash 
  bottle 
  to 
  secure 
  the 
  diminished 
  pressure 
  

   necessary 
  for 
  drawing 
  the 
  mercury 
  from 
  the 
  shorter 
  tube 
  (A) 
  

   when 
  the 
  pressure 
  in 
  the 
  bulb 
  was 
  low. 
  When 
  the 
  mercury 
  is 
  

   thus 
  drawn 
  below 
  the 
  three-way 
  cock, 
  J, 
  this 
  may 
  be 
  turned 
  and 
  

   connection 
  made 
  through 
  a 
  second 
  similar 
  cock, 
  K, 
  either 
  with 
  

   the 
  mercury 
  pump 
  or 
  the 
  gasometer 
  and 
  drying 
  tubes. 
  All 
  

   connections 
  were 
  made 
  with 
  glass 
  or 
  sealing 
  wax 
  and 
  tested 
  

   from 
  time 
  to 
  time 
  to 
  see 
  that 
  they 
  were 
  and 
  remained 
  abso- 
  

   lutely 
  air-tight. 
  Rubber 
  connections 
  were 
  conscientiously 
  

   avoided. 
  

  

  The 
  temperature 
  of 
  the 
  mercury 
  in 
  the 
  manometer 
  was 
  

   determined 
  by 
  the 
  thermometer 
  fitted 
  into 
  the 
  cap 
  above 
  

   described 
  and 
  three 
  other 
  similar 
  ones, 
  whose 
  bulbs 
  dipped 
  into 
  

   short 
  tubes 
  containing 
  mercury 
  placed 
  directly 
  beside 
  the 
  long 
  

   manometer 
  tube 
  at 
  proper 
  intervals 
  and 
  giving 
  the 
  tempera- 
  

   ture 
  at 
  the 
  corresponding 
  height 
  in 
  the 
  long 
  tube 
  with 
  suffi- 
  

   cient 
  accuracy. 
  

  

  The 
  oven 
  being 
  always 
  in 
  a 
  separate 
  room, 
  the 
  changes 
  of 
  

   temperature 
  about 
  the 
  manometer 
  were 
  very 
  small 
  and 
  the 
  

   time 
  between 
  observations 
  necessary 
  for 
  regulating 
  the 
  oven 
  

   temperature 
  to 
  the 
  required 
  degree 
  of 
  constancy, 
  always 
  suffi- 
  

   cient 
  to 
  enable 
  the 
  thermometers 
  to 
  indicate 
  any 
  small 
  changes 
  

   which 
  may 
  have 
  existed. 
  

  

  When 
  mercury 
  was 
  let 
  into 
  the 
  apparatus 
  from 
  the 
  reservoir 
  

   (G) 
  at 
  the 
  top, 
  more 
  or 
  less 
  air 
  bubbles 
  always 
  came 
  with 
  it, 
  

   which 
  collected 
  in 
  the 
  main 
  reservoir 
  (R), 
  having 
  no 
  means 
  of 
  

   escape. 
  To 
  avoid 
  this 
  in 
  the 
  new 
  instrument, 
  a 
  short 
  tube, 
  D, 
  

   with 
  a 
  glass 
  cock, 
  F, 
  was 
  introduced 
  into 
  R 
  at 
  its 
  highest 
  point, 
  

   while 
  the 
  manometer 
  tubes 
  extended 
  much 
  deeper, 
  thereby 
  

   allowing 
  the 
  air 
  to 
  be 
  drawn 
  oil 
  at 
  any 
  time 
  without 
  endanger- 
  

   ing 
  the 
  gas 
  in 
  the 
  bulb 
  — 
  a 
  matter 
  of 
  the 
  highest 
  importance. 
  

  

  The 
  manometer 
  scale 
  was 
  ruled 
  in 
  millimeters 
  for 
  a 
  length 
  

   of 
  1-87"', 
  and 
  carried 
  upon 
  a 
  slider 
  the 
  knife-edge 
  and 
  a 
  micro- 
  

   meter 
  screw 
  and 
  vernier 
  reading 
  intervals 
  of 
  0-02'"'" 
  directly. 
  

   The 
  largest 
  scale 
  error 
  amounted 
  to 
  04'"'". 
  

  

  