6-iS  Mr.  Inglis  on  the  Isothermal  Distillation  of 
to  an  error  of  0"5  per  cent,  in  the  calculation  of  the  percentage 
of  the  vapour ;  so  that  the  apparently  large  differences  between 
the  calculated  and  observed  values  correspond  to  only  small 
errors  in  the  analysis  of  the  liquid.  The  formulae  given  can 
therefore  be  fairly  used  to  smooth  the  experimental  results. 
As  regards  the  total  pressure,  since  this  could  only  be 
smoothed  graphically,  the  results  were  plotted  on  squared 
paper  on  which  2  mm.  corresponded  to  1  per  cent,  in  the 
liquid,  the  pressures  being  plotted  full  size.  A  steady  curve 
was  then  drawn  through  the  experimental  points  and  the 
"smoothed"  values'of  the  total  pressure  taken  from  this  curve. 
Comparison  of  the  "  smoothed  "  and  "  observed "  values 
shows  a  fairly  good  agreement  considering  how  rapidly  the 
vapour-pressure  changes  with  the  temperature.  From  these 
"  smoothed  "■  values  of  the  total  pressure  and  the  composition 
of  the  vapour,  the  partial  pressures  of  the  nitrogen  and  the 
oxygen  may  be  calculated,  and  these  results  are  given  in  the 
last  columns  of  the  tables. 
When  now  these  partial  pressures  and  the  total  pressure 
are  plotted  in  the  usual  way  against  the  molecular  compo- 
sition of  the  liquid  (see  fig.  2  for  the  results  of  Table  II.),  it 
is  found  that  the  curves  obtained,  though  they  have  only 
slight  curvature,  are  certainly  not  straight  lines — a  result 
which  was  to  be  expected  from  the  fact  that  the  relation 
rv=krt  did  not  hold  good.  This  relation  is  the  mathematical 
expression  of  the  property  "  the  ratio  of  the  concentrations  in 
the  vapour  is  proportional  to  the  ratio  of  the  concentrations 
in  the  liquid/'  Hence- we  might  expect  to  find  the  concen- 
tration of  either  substance  in  the  liquid  proportional  to  its 
partial  pressure  as  vapour,  this  being  the  relation  known  as 
Henry's  law.  jSTow  the  vapour  is  usually  plotted  against  the 
molecular  percentage,  and  a  straight  line  for  the  vapour- 
pressure  w7ould  therefore  indicate  a  proportionality  between 
the  vapour-pressure  and  molecular  percentage.  But  since 
the  volume  of  a  gramme-molecule  of  a  liquid  is  a  quantity 
which  is  not  the  same  for  different  liquids,  the  molecular 
percentages  are  not  true  concentrations.  Hence  a  straight 
line  in  the  usual  method  of  plotting  wrould  indicate  a  deviation 
from  Henry's  law;  and  it  is  therefore  evident  that  the  correct 
method  is  to  plot  the  partial  pressures  against  the  true 
concentration. 
To  obtain  the  true  concentration  one  needs  to  know  the 
density  of  each  mixture  used.  But  a  close  approximation  to 
the  actual  value  of  the  density  may  be  obtained  by  calculating 
the  density  of  the  mixture  from  the  densities  of  the  pure  com- 
ponents,   assuming    that    no    contraction    takes  place.      Any 
