164 
METALLIC  OXIDES  IN  COMMERCIAL  AMMONIA. 
port  from  the  above  noticed  reaction  of  ammonia.  In  its  bear- 
ings upon  analytical  chemistry  the  question  also  formed  one  of 
interest,  so  that  it  appeared  worth  while  to  obtain  a  precise 
answer  from  experiment.  With  this  end  in  view,  various  samples 
of  the  liquor  ammonias  of  commerce  were  quantitively  treated 
with  the  annexed  results  : — 
Experiment  1.— The  sample  used  had  a  brownish  hue  indica- 
tive of  empyreumatic  matter,  but  was  perfectly  transparent  and 
not  clarified  in  the  least  by  filtration.  1000  grs.  were  diluted 
five-fold  with  distilled  water,  and  a  stream  of  purified  sulphide 
of  hydrogen  gas  passed  through.  The  presence  of  sulphide  of 
hydrogen  in  the  ammonia  was  instantly  marked  by  the  pro- 
duction of  a  dense  green  color.  Upon  continuing  the  stream 
of  gas  a  precipitate  formed,  which  fell  completely  after  the  lapse 
of  several  hours.  The  precipitate  was  separated  by  filtration, 
washed,  dissolved  in  nitric  acid,  and  the  amount  of  iron  deter- 
mined by  metrical  use  of  ferrocyanide  of  potassium.  The  analy- 
sis showed  2-734  grs.  of  sesquioxide  of  iron,  equivalent  to 
0-2734  grs.  per  100  grs.  of  the  liquor  ammonias. 
Experiment  2. — 1000  grs.  of  liquid  ammonia  quite  trans- 
parent and  translucent  also  were  operated  on  as  detailed  under 
experiment  1.  The  result  was  1-333  grs.  of  sesquioxide  of  iron, 
equal  to  0-1333  grs.  per  cent. 
Experiment  3. — Experiment  1st  was  repeated  upon  1000  grs. 
of  a  solution  of  ammonia  remarkably  limpid  and  free  from  color, 
when  the  amount  of  iron  found  equalled  1-62  grs.  of  sesquioxide 
of  iron,  or  0-162  per  cent. 
Experiment  4  The  solution  had  a  perceptible  tinge  of  brown, 
but  not  removed  by  filtration  through  close  Swedish  paper. 
1000  grs.  afforded  2-361  grs.  of  sesquioxide  of  iron,  or  0-2361 
grs.  per  cent. 
Experiment  5.— -This  sample  had  a  singular  greenish  hue,when 
viewed  through  a  considerable  depth  of  liquid,  readily  suggesting 
the  existence  of  copper  in  the  ammonia.  To  test  this,  3000 
grs.  of  the  solution  were  acidified  by  pure  nitric  acid  and  sulphide 
of  hydrogen  passed  through  to  excess,  when  a  scanty  precipitate 
appeared.  The  amount  of  copper  was  estimated  by  the  com. 
parison  method  relating  to  metre  analysis,  viz.,  by  comparing 
its  solution  in  purified  ammonia  with  its  equivalent  shade  in  a 
solution  of  known  strength.  By  this  plan  1*104  gr.  of  protoxide 
