AmMS•J9hot.rm•}        The  Detection  of  Methyl  Alcohol.  107 
tion  a  matter  of  importance  both  for  the  pharmaceutical  chemist 
and  the  food  analyst. 
If  we  turn  to  standard  books  like  "  Allen's  Commercial  Organic 
Analysis  "  for  the  tests  for  methyl  alcohol  in  the  presence  of  ethyl 
alcohol  we  will  find  that  the  methods  there  found  are  far  from  being 
satisfactory  in  results  or  capable  of  entire  dependence  as  to  accuracy. 
We  find  first  several  methods,  such  as  those  of  Reynolds  and  of 
Cazeneuve,  based  upon  the  common  presence  of  acetone  as  an  im- 
purity in  methyl  alcohol,  so  that  its  detection  by  inference  leads  the 
observer  to  draw  conclusions  as  to  the  presence  of  methyl  alcohol. 
This  is  a  very  unreliable  and  unsafe  way  of  detecting  the  presence 
of  methyl  alcohol.  Some  of  the  deodorized  and  purified  methyl 
alcohols  on  the  market  at  present,  like  "  Columbian  Spirits,"  at 
times  are  nearly  free  from  acetone,  so  that  such  methods  can  be  dis- 
missed from  further  consideration. 
The  method  of  Riche  and  Bardy,  dependent  on  the  formation  of 
methyl  iodide  and  from  this  methyl-aniline  violet,  is  fairly  reliable, 
but  much  too  elaborate  and  involved  to  serve  as  a  ready  test.  In 
this  case  the  mixture  of  alcohols  is  distilled  with  iodine  and  red 
phosphorus  when  methyl  and  ethyl  iodides  are  formed.  These  col- 
lect under  the  aqueous  layer  and  are  separated  and  transferred  to  a 
flask  containing  aniline  with  which  they  react  readily.  After  an 
hour's  time,  the  product  is  boiled  with  water  and  soda  solution 
added  when  the  bases  rise  to  the  top  as  an  oily  layer.  This  is  drawn 
off  and  oxidized  by  the  aid  of  cupric  nitrate.  The  product  of  this 
oxidation,  which  takes  some  hours,  is  exhausted  with  warm  alcohol 
and  filtered  when,  if  pure  ethyl  alcohol  had  been  taken  as  the 
original  sample,  a  red  liquid  is  obtained,  while  if  methyl  alcohol  had 
been  present  a  violet  shade  is  obtained.  Still  more  conclusive  results 
are  obtained  if  the  colored  solution  so  obtained  is  used  to  dye  a 
piece  of  white  merino  wool  which  takes  up  the  violet  color,  but  is 
not  dyed  in  the  absence  of  the  methyl-aniline  compound.  This  test 
is  obviously  too  difficult  of  execution  and  too  detailed  to  be  avail- 
able for  ordinary  pharmacopceial  testing. 
Still  another  test  is  given  in  Allen,  viz.,  that  of  Miller.  This  is 
based  on  the  fact  that  when  oxidized  with  potassium  dichromate 
and  sulphuric  acid,  methyl  alcohol  produces  formic  acid,  capable  of 
reducing  silver  nitrate  solution.  However,  pure  ethyl  alcohol  when 
oxidized  yields  a  trace  of  formic  acid,  or  other  reducing  substance, 
