Isothermal  Distillation  of  Nitrogen  and  Oxygen.         641 
liquefied  gases;  for  in  this  case  one  cannot  work  with  such 
large  quantities  of  liquid  as  Zawidzki  used.  But  this  error 
and  many  others  may  be  eliminated  by  circulating  the 
vapour  through  and  through  the  liquid  until  no  further 
change  in  either  takes  place  :  and  then  collecting  and 
analysing  a  portion  of  the  vapour  which  has  in  this  way  been 
brought  into  equilibrium  with  the  liquid. 
With  such  an  arrangement  it  was  necessary  to  keep  the 
temperature  of  the  distillation-bulb  very  constant  for  a  con- 
siderable time,  and  an  accurate  means  of  measuring  the 
temperature  was  also  required.  The  distillation-bulb  was 
therefore  immersed  in  a  long  cylindrical  vacuum-vessel 
containing  about  600  c.c.  of  liquid  air,  and  the  pressure 
under  which  this  air  was  boiling  was  varied  by  means  of  a 
Fleuss  pump  so  as  to  keep  the  temperature  constant.  To 
measure  the  temperature,  a  bulb  containing  liquid  oxygen 
and  connected  to  a  manometer  was  immersed  in  the  liquid 
air  so  that  the  manometer  registered  the  vapour-pressure  of 
pure  oxygen  at  the  temperature  of  the  distillation-bulb. 
With  this  arrangement  the  Fleuss  pump  was  worked  at  such 
a  rate  that  the  manometer  showed  a  constant  pressure.  In 
the  different  experiments  this  pressure  amounted  to  either 
100,  200,  or  300  mm.,  and  the  variation  of  pressure  seldom 
exceeded  0*5  mm.  This  variation  corresponds  to  a  tem- 
perature difference  of  0°'03  C,  0a02  C,  and  0°-01  C.  in  the 
three  different  cases.  These  pressures,  according  to  Travers 
(Phil,  Trans.  200.  p.  105)  correspond  to  74°-7,  79°'07,  and 
82o,09  Absolute  measured  on  the  hydrogen  scale. 
The  apparatus  used  for  carrying  out  the  distillations  is 
shown  diagrammatically  in  fig.  1,  the  liquid  oxygen  bulb 
and  the  corresponding  manometer  being  omitted.  The 
rubber  cork  A,  which  fits  the  mouth  of  the  vacuum  vessel 
containing  the  liquid  air,  has  passing  through  it  the  liquid 
oxygen  bulb  and  several  tubes  of  which  three,  1,  2,  3,  are 
joined  to  the  distillation-bulb  P,  and  another,  not  shown  in 
the  figure,  leads  to  the  Fleuss  pump.  In  an  experiment,  B  is 
half  filled  with  the  mixture  to  be  distilled,  and  the  vapour, 
after  being  thoroughly  cooled  by  the  spiral  4,  is  blown  through 
the  liquid  in  B,  the  tube  3  then  carrying  the  vapour  on  into 
the  circuit.  The  tube  1  is  joined  to  the  bulb  B  quite  close  to 
the  bottom,  and  for  a  considerable  part  of  its  length  consists 
of  an  extremely  fine  capillary  drawn  out  of  an  ordinary  piece 
of  capillary  tubing.  This  tube  is  also  connected  to  a  gas- 
holder C,  so  that  when  the  reservoir  of  C  is  lowered  liquid 
is  sucked  up  the  fine  capillary  to  a  warm  part  above  the  liquid 
air,  and  there  boils  as  a  whole  and  is  collected  in  the  gasholder. 
