[     640     ] 
LYIII.    The  Isothermal  Distillation  of  Nitrogen  and  Oxygen 
and  of  Argon  and  Oxygen,     By  J.  K.  H.  Inglis  *. 
MANY  investigations  have  been  made  in  order  to  find 
the  relation  connecting  the  composition  of  the  vapour 
with  the  composition  of  the  liquid,  when  a  mixture  of  two 
liquids  is  distilled  isothermally  (see  Young's  '  Fractional 
Distillation/  Macmillan  &  Co.,  where  a  full  summary  of  the 
literature  on  this  subject  in  ay  be  found)  ;  and  it  has  been 
shown  that  in  some  cases  the  relation  takes  the  simple  form 
rv=k  .  ri,  where  rv  is  the  ratio  of  the  two  substances  in  the 
vapour,  ti  the  corresponding  ratio  in  the  liquid,  and  k  a 
constant.  In  most  cases,  however,  k  is  not  an  absolute 
constant  but  varies  slightly  with  the  molecular  composition 
of  the  liquid ;  and  we  thus  get  mixtures  (a)  which  have  a 
maximum  vapour-pressure,  (b)  which  have  a  minimum 
pressure,  and  (c)  which  although  having  neither  maximum  or 
minimum  vapour-pressure  do  not  satisfy  the  above  relation. 
These  different  cases  have  been  very  fully  investigated  by 
Zawidzki  and  others  at  ordinary  temperatures  ;  but  the  only 
paper  dealing  with  distillation  at  low  temperatures  is  one  by 
Mr.  E.  C.  C.  Baly  (Phil.  Mag.  vol.  xlix.  p.  517,  1900),  who 
carried  out  a  series  of  distillations  of  liquid  air  under  a 
pressure  of  one  atmosphere.  In  isobaric  distillations,  how- 
ever, the  conditions  are  not  so  simple  as  they  are  when  the 
temperature  is  constant  ;  so  at  Mr.  Baly's  suggestion  I 
decided  to  complete  his  work  by  making  a  series  of  isothermal 
distillations  at  low  temperatures,  and  I  received  considerable 
help  from  him  in  the  early  experiments  when  we  hoped  to 
make  the  research  a  joint  one. 
Many  forms  of  apparatus  have  been  devised  for  carrying 
out  distillations  isothermally,  but  none  of  them  were  suitable 
for  work  at  low  temperatures.  Moreover,  most  of  the  forms 
used  are  open  to  the  objection  that  care  is  not  taken  to 
ensure  that  the  vapour  is  in  complete  equilibrium  with  the 
liquid ;  and  in  addition  considerable  error  may  be  introduced 
by  back  condensation,  &c.  The  apparatus  used  by  Zawidzki 
(Zeit.f.  phys.  Chemie,  vol.  xxxv.  p.  129)  avoids  most  of  these 
sources  of  error ;  but  even  in  this  case,  since  the  sample  of 
distillate  is  obtained  by  distilling  over  a  small  quantity  of 
the  liquid,  the  composition  of  the  liquid  may  change  during 
this  operation,  and  so  introduce  error.  Such  an  error  would 
be    of    considerably   greater    importance    in    dealing    with 
*  Communicated  by  the  Physical  Society  :  read  January  26,  1906. 
