Nitrogen  and  Oxygen  and  of  Argon  and  Oxygen.        643 
and  thus  sends  it  round  and  round  the  circuit.  The  mercury 
after  it  falls  down  the  circulator  is  collected  at  the  exit  Gr, 
the  height  of  which  above  or  below  F  is  adjusted  so  as  to 
keep  the  surface  of  the  mercury  just  below  the  side  tube 
leading  from  F. 
The  sample  of  vapour  is  collected  in  the  bulbs  L  through 
which  the  vapour  is  circulated ;  by  closing  the  taps  HK 
this  sample  of  vapour  is  shut  off  from  communication  with 
the  liquid  in  P,  and  can  be  collected  by  opening  the  tap  M 
which  leads  to  a  Topler  pump.  The  pressure  under  which 
the  liquid  is  boiling  can  be  measured  in  two  different  ways. 
It  can  either  be  calculated  from  the  difference  in  height  of 
the  mercury  surfaces  in  F  and  Gr  and  the  height  of  the 
barometer  ;  or  it  may  be  measured  directly  by  means  of  the 
closed  manometer  NP.  The  shorter  open  limb  N  of  this 
manometer  can  be  shut  off  from  the  closed  limb  P  by  means 
of  the  tap  R.  By  means  of  the  taps  QR  and  the  reservoir  S 
the  amount  of  mercury  in  the  manometer  can  be  so  regulated 
that  when  R  is  open  the  mercury  surface  in  N  is  close  below 
the  side  tube  coming  from  the  tap  H.  In  this  way  dead 
space  is  as  far  as  possible  avoided.  This  manometer  could 
be  used  for  pressures  up  to  770  mm.,  and  the  manometer 
GF  was  used  only  for  pressures  greater  than  this.  Readings 
of  the  manometer  NP  were  made  during  the  circulation,  so 
that  one  could  see  when  the  reading  became  constant ;  but 
the  final  readings  were  taken  by  means  of  a  telescope  after 
the  circulator  had  been  stopped  and  after  the  tap  T  connecting 
the  tubes  2,  3  had  been  opened.  It  was  found  convenient 
to  close  the  tap  R  a  moment  after  T  had  been  opened  and 
before  the  readings  were  taken. 
The  gases  were  stored  over  water  in  ordinary  glass  gas- 
holders— the  argon  alone  being  kept  over  mercury  ;  and 
after  passing  through  soda-lime  and  phosphorus-pentoxide 
tubes,  were  admitted  to  the  apparatus  by  means  of  the 
tap  U.  As  regards  the  gases  themselves,  the  oxygen  was  in 
all  cases  obtained  by  the  decomposition  of  potassium  per- 
manganate. The  nitrogen  was  prepared  by  heating  a  mixture 
of  solutions  of  ammonium  sulphate  and  potassium  nitrite, 
and  was  fractionated  by  means  of  liquid  air  before  being- 
passed  into  the  gasholders.  A  small  quantity  of  oxygen  was 
then  added  and  the  gas  left  standing  over  water.  In  this 
way,  any  traces  of  nitric  oxide  were  turned  into  nitrous  or 
nitric  acids  which  dissolved  in  the  water,  and  the  small 
quantity  of  oxygen  (0'5  per  cent.)  remaining  did  not  matter, 
as  mixtures  of  nitrogen  and  oxygen  were  to  be  used.     The 
