﻿306 
  Dr. 
  C. 
  Chree's 
  Notes 
  on 
  Thermometry. 
  

  

  to 
  select 
  as 
  far 
  as 
  possible 
  times 
  of 
  steady 
  barometric 
  pressure 
  

   for 
  boiling-point 
  determinations. 
  

  

  Calibration 
  *. 
  

  

  § 
  25. 
  This 
  is 
  too 
  technical 
  and 
  mathematical 
  a 
  subject 
  to 
  

   be 
  profitably 
  discussed 
  here 
  in 
  detail. 
  The 
  methods 
  in 
  use 
  at 
  

   the 
  Bureau 
  International 
  are 
  described 
  in 
  several 
  of 
  their 
  

   publications 
  f 
  and 
  in 
  Guillaume's 
  Thermometries 
  pp. 
  43 
  et 
  seqq. 
  

   In 
  calibrating 
  an 
  ordinary 
  thermometer, 
  range 
  0° 
  to 
  100° 
  C, 
  use 
  

   is 
  made 
  of 
  threads 
  of 
  50, 
  40, 
  30,20,12, 
  10, 
  8,6,4,and 
  2 
  degrees. 
  

  

  The 
  publications 
  of 
  the 
  Reichsanstalt 
  recommend 
  that 
  the 
  

   shortest 
  mercury-threads 
  used 
  do 
  not 
  exceed 
  in 
  length 
  one 
  

   degree-division, 
  and 
  speak 
  of 
  an 
  error 
  as 
  large 
  as 
  o, 
  01 
  C. 
  % 
  

   arising 
  in 
  the 
  calibration 
  of 
  a 
  Tonnelot 
  thermometer 
  when 
  

   the 
  shortest 
  thread 
  equalled 
  2 
  Q, 
  5. 
  They 
  also 
  recommend 
  a 
  

   double 
  calibration 
  with 
  each 
  thread, 
  proceeding 
  from 
  one 
  

   end 
  of 
  the 
  thermometer 
  to 
  the 
  other 
  and 
  back 
  again. 
  This 
  

   is 
  to 
  serve 
  as 
  a 
  protection 
  against 
  the 
  error 
  that 
  might 
  arise 
  

   from 
  gradual 
  heating 
  of 
  the 
  mercury- 
  thread. 
  Alleged 
  "nicht 
  

   unerheblichen 
  " 
  errors 
  § 
  in 
  the 
  calibration-data 
  supplied 
  by 
  the 
  

   Bureau 
  International 
  for 
  one 
  of 
  the 
  Reich 
  sanstalt's 
  verre 
  dur 
  

   thermometers 
  are 
  ascribed 
  to 
  the 
  neglect 
  of 
  this 
  precaution. 
  

  

  The 
  object 
  of 
  calibration 
  at 
  both 
  these 
  institutions 
  is 
  to 
  

   supply 
  a 
  table 
  of 
  corrections 
  to 
  be 
  applied 
  habitually 
  to 
  every 
  

   reading 
  of 
  the 
  thermometer. 
  Degree-divisions 
  represent 
  

   equal 
  lengths, 
  not 
  equal 
  volumes, 
  on 
  the 
  French 
  and 
  German 
  

   standard 
  thermometers- 
  

  

  Kew 
  standard 
  thermometers, 
  on 
  the 
  other 
  hand, 
  are 
  cali- 
  

   brated 
  before 
  they 
  are 
  subdivided, 
  and 
  the 
  calibration 
  results 
  

   are 
  used 
  to 
  guide 
  one 
  in 
  dividing 
  the 
  stem 
  into 
  equal 
  volumes, 
  

   not 
  equal 
  lengths. 
  Most 
  authorities 
  take 
  for 
  granted 
  that 
  

   this 
  attempt 
  is 
  bound 
  to 
  fail 
  when 
  accuracy 
  of 
  the 
  order 
  

   0°*001 
  C. 
  is 
  in 
  question. 
  This 
  is 
  very 
  probably 
  a 
  true 
  con- 
  

   clusion, 
  but 
  it 
  is 
  not 
  always 
  safe 
  to 
  prophecy 
  failure 
  before 
  

   the 
  attempt 
  is 
  made. 
  The 
  fact 
  is 
  that 
  no 
  attempt 
  has 
  ever 
  

   been 
  made 
  to 
  construct 
  Kew 
  standards 
  suitable 
  for 
  this 
  degree 
  

   of 
  accuracy- 
  They 
  are 
  intended 
  to 
  be 
  read 
  easily 
  by 
  the 
  

   unaided 
  eye, 
  and 
  the 
  width 
  of 
  the 
  divisions 
  is 
  usually 
  a 
  very 
  

   considerable 
  multiple 
  of 
  the 
  y^oo 
  °f 
  a 
  degree 
  interval. 
  The 
  

   accuracy 
  ordinarily 
  aimed 
  at 
  in 
  the 
  construction 
  of 
  these 
  

   instruments 
  is 
  really 
  o, 
  05 
  F., 
  and 
  if 
  we 
  may 
  judge 
  from 
  

  

  * 
  For 
  & 
  description 
  of 
  the 
  principal 
  methods 
  see 
  a 
  report 
  by 
  Profs. 
  

   Balfour 
  Stewart, 
  Riicker, 
  and 
  Thorpe, 
  B. 
  A. 
  Report 
  for 
  1882, 
  p. 
  145. 
  

   f 
  Travaux 
  et 
  Memoires, 
  vol. 
  x. 
  1894 
  (Thermometres 
  Etalons, 
  pp. 
  10-16). 
  

   | 
  Wiss. 
  Abhandl. 
  vol. 
  i. 
  1894, 
  pp. 
  39, 
  40, 
  

   § 
  Ibid. 
  vol. 
  ii. 
  1895, 
  p. 
  6. 
  

  

  