156  Caramclization  in  Rivas's  Test.       { Am;^S£m1£rm- 
odor  and  the  color  are  not  inseparably  connected,  and  it  is  likely 
that  the  odor  is  due  to  another  and  colorless  substance  that  may 
form  at  the  same  time  caramel  is  formed,  but  may  also  form  at 
other  times  as  well. 
The  explanation  suggested  by  some  of  these  experiments  is  that 
in  the  browning  of  the  glucose  solutions  wTe  have  to  deal  with  the 
condensation  of  glucose  molecules.  Such  condensation  is  to  be 
expected  because  glucose  is  an  aldehyde.  In  the  pale  yellow  color 
that  appears  before  the  brown  shades,  and  that  may  be  preserved 
for  some  time  by  cooling  the  solution  quickly,  we  may  have  to  deal 
with  an  alcoholate  or  with  mixtures  of  several  alcoholates,  formed 
by  the  action  of  the  alkali  on  the  alcohol  groups  of  the  glucose. 
The  experiments  with  the  polariscope  suggest  that  racemization  of 
the  glucose  molecule  may  also  occur  at  the  time  the  alkali  acts. 
Accordingly  I  examined  the  literature  to  see  if  I  could  find  any 
ground  against  these  suggestions  or  any  evidence  in  favor  of  cara- 
melization.    I  shall  give  data  from  several  sources. 
When  a  solution  of  glucose  in  absolute  alcohol  is  treated  with 
sodium  ethylate,  a  compound  is  precipitated  which  when  properly 
dried  is  a  white  to  yellow-white  powder  of  the  composition 
CeH^OgNa.  When  heated  to  a  little  over  70 0  C,  this  substance 
begins  to  brown  and  to  decompose,  even  in  an  atmosphere  of  hydro- 
gen, with  the  final  formation  of  a  brown,  flaky,  amorphous  mass.14 
"  If  an  alcoholic  caustic-alkali  solution  is  added  to  an  alcoholic 
solution  of  glucose,  an  amorphous  precipitate  of  insoluble  alkali 
compound  is  formed.  On  warming  this  compound  it  decomposes 
easily  with  the  formation  of  a  yellowish  or  brownish  color,  which 
is  the  basis  of  Moore's  Test."  15 
When  glucose  is  heated  with  water  and  potassium  hydroxide  in 
a  loosely  stoppered  flask  on  a  water-bath  at  temperatures  from 
350  to  400  C,  the  solution  broAvns  after  a  short  time.  This  color 
increases  in  intensity  for  the  first  few  days,  but  after  some  time  the 
liquid  becomes  colorless.  Lactic  acid  may  be  isolated  from  this 
colorless  liquid  with  a  yield  of  about  41  per  cent.  Although  the 
rate  of  this  reaction  varies  greatly  with  the  dilution  and  the  relative 
concentration  of  the  alkali,  lactic  acid  is  always  formed  and  may  be 
separated  from  the  colorless  solutions  that  finally  result.  Sodium 
14  Honig  and  Rosenfeld,  Ber.  CJiein.  Gesell.,  Berlin,  10,  p.  871,  1877. 
15  Hammersten-Mandel,  p.  80. 
