158  Oil  of  Bitter  Almonds.  {AmA$i;£8£Tm' 
Mallinckrodt  and  Stull  (Jour.  Am.'Chem.  Soc,  1904,  1029)  called 
attention  to  some  difficulty  they  had  in  testing  potassium  iodide  for 
iodate  by  reason  of  the  presence  of  nitrites  in  the  filter  paper 
which  they  had  used,  and  went  on  to  show  that  all  the  samples  of 
filter  paper  which  had  been  kept  in  their  laboratory  showed  more 
or  less  marked  traces  of  nitrites.  Procter  (your.  Soc.  Chem.  lnd.y 
1904,  9),  in  an  article  regarding  water  analysis,  also  calls  attention  to 
the  ready  absorption  of  ammoniacal  and  acid  vapors  from  the  air  of 
the  laboratory  and  the  necessity  of  washing  filter  papers  prior  to 
their  use,  which  were  desired  to  be  free  from  any  traces  of  these 
bodies.  We  have  found  the  same  thing  to  be  true  in  the  applica- 
tion of  this  silver  nitrate  test. 
The  Pharmacopoeia  directs  that  after  burning  the  oil  in  the 
presence  of  water  to  absorb  the  volatile  chlorine  compound  the 
resulting  solution  shall  be  filtered  free  from  carbon  and  the 
filtrate  tested  by  the  addition  of  silver  nitrate.  By  the  appli- 
cation of  this  test  exactly  as  directed  there  was  not  a  single  one  of 
the  twenty-two  samples  that  did  not  show  a  more  or  less  distinct 
test  for  chlorides  in  the  aqueous  solution.  The  water  used  was 
found  to  be  absolutely  free  from  the  minutest  taste  of  chlorine, 
and  in  looking  further  for  a  possible  source  of  error,  it  was  found 
that  not  a  filter  paper  in  the  laboratory  but  would  give  at  least  a 
slight  test  for  chlorine  when  extracted  with  water.  The  tests  on 
these  filter  papers  included  every  variety  in  stock,  from  the  heaviest 
sheet  filter  to  the  various  grades  of  Schleicher  &  Schull's  quantita- 
tive filters.  An  aqueous  extraction  of  the  filters  also  showed  the 
presence  of  ammonia  by  Nessler's  reagent,  so  that  this  source  of 
error  is  due  to  the  absorption  from  the  atmosphere  of  the  laboratory 
of  the  vapors  of  ammonia  and  hydrochloric  acid,  the  resulting  am- 
monium chloride  being  readily  extracted,  in  part  at  least,  by  the 
passage  of  even  a  small  quantity  of  water  through  the  filter.  Two 
methods  may  be  adopted  to  obviate  this  source  of  error:  First,  the 
filter  papers  used  may  be  washed  carefully  with  water  free  from 
chlorine  immediately  before  use;  or,  second,  the  test  for  chlorides 
may  be  made  without  filtering  off  the  carbon  produced  in  process 
of  the  test.  Tne  second  method  is  preferable,  as  it  renders  unneces- 
sary the  additional  time  and  trouble  to  wash  the  filter,  and  the 
finely  divided  carbon  floating  on  the  surface  of  the  liquid  does  not 
in  any  way  interfere  with  the  delicacy  of  the  test.     Thus  the  clos- 
