154  Caramelization  in  Rii'as's  Test.  {^'a^iIw™' 
slightest  dextrorotation.  The  conclusions  are  that  the  glucose  is 
seriously  involved  in  the  action  of  the  alkali,  and  that  the  brown 
color  is  the  color  of  a  secondary  product. 
To  insure  the  dissimilarity  of  the  brown  color  in  the  glucose 
solutions  to  the  brown  color  of  caramel,  solutions  of  caramel  made 
by  heating  glucose  at  2000  C.  for  70  minutes  and  by  adding  "  cara- 
mel brown  "  used  for  coloring  whiskies  to  water,  were  used  in 
the  next  experiment.  Each  solution  was  divided  into  two  parts, 
one  of  which  was  acidified.  The  color  of  the  acidified  solution  was 
then  compared  with  that  of  the  original  solution  corresponding  to 
it.  The  acid  produced  a  reduction  in  the  intensity  of  the  colors 
that  was  noticeable  when  a  careful  comparison  was  made,  but  the 
change  was  negligible  when  compared  with  the  entire  destruction 
by  acids  of  the  equally  intense  color  of  the  browned  glucose  solution. 
One  of  the  well  known  tests  for  caramel  was  then  selected  and 
applied  to  the  brown  solution  obtained  when  glucose  is  warmed  with 
alkali.  I  selected  Marsh's  test  as  recommended  by  Crampton  and 
Tolman  11  and  used  in  their  extensive  researches  on  the  aging  of 
whiskies.  The  reagent  is  an  emulsion  made  by  adding  3  c.c.  of 
water  and  3  c.c.  syrupy  phosphoric  acid  to  100  c.c.  amyl  alcohol. 
When  a  whiskey  colored  with  caramel  is  shaken  with  twice  its 
volume  of  this  reagent  and  time  is  allowed  for  two  distinct  layers 
to  form  in  the  liquid,  it  is  found  that  the  lower  (aqueous)  layer  is 
always  colored  brown,  whereas  in  the  absence  of  caramel  the  lower 
layer  is  colorless.  When,  for  the  purpose  of  testing  my  preparation 
of  this  reagent,  some  of  it  was  applied  to  a  "  straight  whiskey  "  that 
I  happened  to  have  in  my  laboratory,  the  lower  layer  was  colorless. 
When  the  browned  glucose  solution  was  tested  in  the  same  way,  the 
lower  layer  was  brown.  Therefore,  the  test  for  caramel  was  posi- 
tive. It  is  evident,  however,  that  the  results  on  the  browned  glucose 
solution  are  not  comparable  with  those  on  the  whiskey,  for  the 
glucose  was  alkaline  and  the  whiskey,  of  course,  acid.  Accordingly 
the  browned  glucose  solution  was  acidified  and  then  tested  as  before, 
with  the  result  that  the  lower  layer  was  now  colorless  and  the  same 
solution  that  before  gave  a  positive  test  now  gave  a  negative  test, 
for  caramel.  Similarly,  some  of  the  whiskey  was  rendered  alkaline 
and  then  tested.  The  whiskey  that  had  before  given  a  negative  test 
now  gave  a  strong-  positive  test,  showing  a  brown  color  in  the 
11  Journ.  Amer.  Chem.  Soc,  30,  p.  100,  1908. 
