166  ON  METALLIC  BISMUTH. 
strongest  objections  to  the  Pharmacopoeia  method.  It  is  cer- 
tainly true  that  fusion  with  nitre  is  useless  for  the  removal  of 
very  small  quantities  of  copper.  Down  to  what  proportion  it 
is  possible  to  reduce  the  copper  by  this  means  I  am  not  pre- 
pared to  say,  and  I  do  not  know  that  any  experiments  have 
been  published  on  the  point.  I  cannot  admit,  however,  that  the 
nitre  fails  to  remove  any  portion  of  this  impurity,  as  some 
have  implied ;  for  the  following  experiment  goes  to  show  the 
contrary.  Messrs.  Johnson  and  Matthey  were  kind  enough  to 
prepare  for  me  a  piece  of  bismuth  containing  2*9  per  cent,  of 
copper.  I  fused  this  for  ten  minutes  with  one-fifth  its  weight  of 
nitre,  and  then  analysed  the  product.  I  found  it  to  contain 
only  1*51  per  cent.  Consequently,  nearly  one-half  the  copper 
had  in  this  case  been  removed.  Nevertheless,  I  cannot  deny  that 
fusion  with  nitre  fails  to  remove  the  last  portions  of  copper, 
and  is  therefore  useless  as  far  as  small  percentages  of  this 
impurity  are  concerned. 
Admitting  this,  it  becomes  important  to  know  the  exact 
amount  of  copper  commonly  present  in  the  metallic  bismuth  of 
commerce.  To  ascertain  this  point,  consequently,  I  have  taken 
three  commercial  samples  of  metal,  and  have  made  quantitative 
determinations  of  the  amount  of  copper  in  each.  The  analysis 
was  performed  as  follows : — One  hundred  grains  of  the  metal 
were  dissolved  in  dilute  nitric  acid,  and  the  solution  evaporated 
until  a  pellicle  formed.  About  an  ounce  of  a  saturated  solution 
of  sal  ammoniac  was  then  added,  the  mixture  slightly  warmed, 
and  diluted  to  the  bulk  of  thirty  or  forty  ounces  with  cold  water. 
All  the  bismuth  was  thus  completely  precipitated  as  insoluble 
oxychloride,  leaving  the  copper,  etc.,  in  solution.  After  some 
hours'  repose  the  liquor  was  filtered,  and  the  precipitate  washed. 
The  filtrate  was  evaporated  to  about  two  ounces,  and  a  slight 
excess  of  ammonia  added.  After  filtration,  the  liquor  was  acidi- 
fied and  precipitated  by  sulphuretted  hydrogen.  The  sulphide 
was  collected,  washed  with  dilute  sulphide  of  sodium,  and  dis- 
solved in  aqua  regia.  The  solution  was  evaporated  to  dryness. 
The  copper  in  the  residue  was  then  estimated,  by  precipitation 
with  zinc  in  a  platinum  dish  and  weighing  as  metal,  after  the 
manner  recommended  by  Fresenius. 
