Am'ipr\i; iP9?8arm }  Oil  of  Bitter  Almonds,  161 
per  cent.,  which  shows  that  the  small  amount  of  benzaldehyde  is 
not  caused  by  an  excessive  oxidation  to  benzoic  acid.  To  further 
try  the  process,  a  sample  of  hijrli-grade  commercial  binzildshyde 
was  obtained  and  assayed.  This  sample  had  a  specific  gravity  of 
1-053  at  1 50  C.  and  in  fractional  distillation  gave  the  following 
results  : 
Pressure,  739  m.m. 
Per  Cent.  Per  cent. 
1770     to  178°        4  1790  to  179/50  IO 
1780     "  i7S-5c,    26  i79'5°,  "  1800  5 
178-5°  "  179°,       45  Not  distilled.  10 
The  distillation  was  from  an  ordinary  distilling  flask  and  over  a 
free  flame  without  any  special  protection  of  the  upper  part  of  the 
distilling  flask.  Three  assays  of  this  benzaldehyde  gave  as  results 
79  8  per  cent,  8o-i2  per  cent.,  and  80.07  per  cent.,  which  deter- 
minations are  very  close  indeed.  The  amount  of  benzoic  acid 
present  in  this  benzaldehyde  was  very  small,  amounting  to  nothing 
more  than  a  distinct  trace.  Here,  again,  we  have  results  which  are 
5  per  cent,  below  the  U.S. P.  standard,  showing  that  the  assay  proc- 
ess used  is  certainly  at  fault  in  some  particular. 
The  results  above  given,  both  on  the  oil  and  the  benzaldehyde,  as 
obtained  by  the  writer,  were  duplicated  within  the  limits  of  their 
variations  by  another  analyst,  so  that  the  trouble  is  evidently  not 
due  to  a  misreading  of  the  end-point. 
Still,  a  third  sample,  which  was  marked  "  Oil  of  Bitter  Almonds, 
Artificial,"  assayed  78-4  per  cent. 
As  we  have  stated,  the  chief  difficulty  with  the  process  is  the 
uncertainty  attendant  upon  the  determination  of  the  end. point, and 
while  Sadtler  recommends  rosolic  acid  as  giving  better  results,  I 
rather  prefer  phenolphthalein.  Unless  extreme  care  be  taken,  par- 
ticularly if  one  is  unacquainted  with  this  process,  the  end-point  will 
be  considered  reached  before  such  is  actually  the  case,  for  the  final 
pink  tint  of  the  phenolphthalein  is  exceedingly  faint.  For  this  rea- 
son it  is  absolutely  essential  that  the  titration  should  be  carried  out 
in  full  daylight,  a  cloudy  day  even  being  detrimental  to  the  best  re. 
suits.  It  is  evident  that  results  which  are  fairly  concordant  may  be 
obtained,  but  it  seems  equally  evident  that  concordance  of  results 
is  not  indicative  of  accuracy.  It  is  of  interest  to  note  here  the  state- 
ments in  the  first  article  of  Sadtler's  dealing  in  any  way  with  the 
assay  of  benzaldehyde  by  this  process,  wherein  he  says  that  it  gave 
