﻿the 
  Absolute 
  Scale 
  of 
  Temperature. 
  375 
  

  

  Then 
  from 
  equation 
  (v.) 
  we 
  find 
  the 
  following 
  values 
  of 
  T 
  :— 
  

  

  

  V0. 
  

  

  i 
  

   T 
  

  

  

  273-035 
  

  

  272-48 
  

  

  272-13 
  

  

  268-47 
  

  

  268-47 
  

  

  272-81 
  

   273-27 
  

   273-20 
  

   274-83 
  

   273-48 
  

  

  Air 
  

  

  Nitrogen 
  

  

  Carbon 
  dioxide 
  

  

  

  With 
  regard 
  to 
  these 
  results 
  we 
  may 
  remark 
  that 
  in 
  each 
  

   case 
  the 
  data 
  leave 
  something 
  to 
  be 
  desired, 
  Hydrogen, 
  no 
  

   doubt, 
  is 
  the 
  best 
  thermometric 
  substance, 
  since 
  its 
  divergence 
  

   from 
  a 
  perfect 
  gas 
  is 
  so 
  small 
  ; 
  and 
  the 
  numbers 
  suggest 
  that 
  

   perhaps 
  even 
  taking 
  e 
  = 
  o, 
  039 
  is 
  an 
  overestimate 
  of 
  the 
  

   divergence. 
  Moreover, 
  as 
  we 
  shall 
  see 
  below, 
  there 
  appears 
  

   to 
  be 
  an 
  anomaly 
  in 
  its 
  behaviour 
  which 
  makes 
  the 
  course 
  of 
  

   the 
  hydrogen 
  scale 
  between 
  0° 
  and 
  100° 
  run 
  differently 
  from 
  

   those 
  of 
  the 
  other 
  gases. 
  For 
  both 
  these 
  reasons, 
  more 
  

   information 
  on 
  the 
  thermal 
  effect 
  due 
  to 
  free 
  expansion, 
  and 
  

   its 
  variation 
  with 
  temperature, 
  is 
  especially 
  needed 
  ; 
  and 
  it 
  

   is 
  essential 
  that 
  the 
  experiments 
  should 
  be 
  made 
  with 
  pure 
  

   hydrogen. 
  

  

  The 
  value 
  for 
  T 
  derived 
  from 
  the 
  air-thermometer 
  is 
  

   probably 
  entitled 
  to 
  about 
  the 
  same 
  weight 
  as 
  that 
  from 
  

   hydrogen. 
  It 
  is 
  the 
  only 
  gas 
  for 
  which 
  e 
  appears 
  to 
  be 
  satis- 
  

   factorily 
  determined 
  ; 
  on 
  the 
  other 
  hand, 
  it 
  is 
  probable 
  that 
  

   ft 
  is 
  a 
  little 
  higher 
  than 
  is 
  assumed 
  above 
  ; 
  for, 
  according 
  to 
  

   Jolly, 
  whose 
  measurements 
  seem 
  to 
  be 
  more 
  reliable 
  than 
  

   Regnault's, 
  air 
  has 
  a 
  coefficient 
  of 
  pressure 
  higher 
  than 
  that 
  

   of 
  nitrogen 
  by 
  2 
  x 
  10~ 
  6 
  ; 
  this 
  is 
  all 
  the 
  more 
  likely 
  because 
  

   it 
  brings 
  air 
  nearly 
  into 
  the 
  position 
  it 
  would 
  occupy 
  if 
  its 
  

   constituent 
  gases 
  exercised 
  no 
  influence 
  on 
  one 
  another. 
  Now 
  

   if 
  this 
  is 
  true, 
  and 
  Chappuis's 
  value 
  for 
  nitrogen 
  is 
  correct, 
  the 
  

   pressure-coefficient 
  for 
  air 
  would 
  become 
  0*0036730 
  and 
  the 
  

   value 
  of 
  T 
  derived 
  from 
  it 
  273°-05. 
  

  

  The 
  determination 
  from 
  nitrogen 
  is 
  not 
  of 
  much 
  value 
  on 
  

   account 
  of 
  the 
  absence 
  of 
  information 
  on 
  the 
  value 
  of 
  e. 
  Of 
  

   the 
  two 
  values 
  from 
  carbon 
  dioxide, 
  the 
  first 
  is, 
  of 
  course, 
  

   quite 
  inadmissible 
  ; 
  if 
  the 
  theory 
  given 
  in 
  this 
  paper 
  is 
  cor- 
  

   rect, 
  we 
  must 
  conclude 
  that 
  the 
  measurements 
  of 
  the 
  cooling 
  

   effect 
  contain 
  some 
  error, 
  and 
  that 
  the 
  variation 
  of 
  the 
  

   coefficient 
  of 
  pressure 
  is 
  more 
  to 
  be 
  depended 
  on 
  in 
  this 
  

   instance. 
  It 
  may 
  be 
  remarked 
  too, 
  that 
  if 
  we 
  take 
  Regnault's 
  

  

  