^"sepr'issr"'}        Polyhydric  Alcohols  and  Borax.  449 
819)  published  the  results  of  experiments  on  the  action  of  glycerol 
upon  borax,  from  which  it  appeared  that  free  boric  acid  was  liberated 
and  a  more  basic  borate  formed.  Experiments  made  with  a  view  of 
isolating  the  free  boric  acid,  either  directly  or  indirectly,  gave  negative 
results.  These  same  observers  (Journ.  Chem.  Soc,  xxxiii,  438,  Am. 
Jour.  Phar.y  1883,  p.  483)  proposed  to  utilize  the  fact  that  a  mixture 
of  glycerol  and  borax  yields  a  green  color  to  the  flame  as  a  test  for 
glycerol,  for  they  found  that  the  reaction  was  not  shared  by  the  majority 
of  polyhydric  alcohols,  erythrol  and  glycol  alone  resembling  glycerol  in 
this  respect.  In  1881,  Donath  and  Mayrhofer  (Zeits.  fllr  anal.  Ohem, 
XX.,  379)  found  that  the  acid  solution  resulting  from  the  action  of  gly- 
cerol upon  borax  became  alkaline  when  heated,  and  proposed  this 
reaction  as  a  test  for  glycerol.  In  1882  (PAarm.  Journ.,  [3],  257, 
Am.  Jour.  Fhor.,  1882,  p.  537),  I  showed  the  limits  of  this  test 
and  proposed  a  new  method  of  applying  it,  at  the  same  time  point- 
ing out  that  it  could  not  be  relied  upon  as  a  distinctive  test  for 
glycerol  unless  special  precautions  were  taken  to  exclude  other  polyhy- 
dric alcohols,  for  I  had  found  that  the  same  reaction  was  given  by 
mannitol,  erythrol,  guaiacol,  pyrogallol,  saligenol,  dextrose,  levulose, 
lactose  and  mycose,  and  in  fact  might  be  considered  to  be  character- 
istic of  all  polyhydric  alcohols.  Since  that  time  my  work  has  been  ex- 
tended with  the  view  of  discovering  the  rationale  of  the  general 
reaction. 
Action  of  Glycerol  upon  Borax. — When  anhydrous  borax  and  an- 
liydrous  glycerol,  the  latter  in  excess,  are  heated  together  the  mixture 
becomes  acid  ^nd  imparts  a  vivid  green  tinge  to  the  flame.  Anhy- 
drous borax  and  anhydrous  glycerol,  the  latter  slightly  in  excess,  were 
heated  to  a  temperature  of  120°  C.  As  the  temperature  neared  100° 
C.  water  began  to  be  abundantly  given  off,  but  it  was  necessary  to 
raise  the  temperature  to  120°  C.  in  order  to  eflect  entire  expulsion  of 
the  water.  Very  many  experiments  were  made  with  varying  quantities 
of  the  two  substances  to  estimate  the  amount  of  water  given  ofl*  and  so 
to  determine  the  formula  of  the  compound  produced,  but  no  reliable 
results  were  obtained,  owing  to  the  inevitable  loss  of  glycerol  at  this 
temperature.  The  mass  obtained  in  the  above  manner,  and  which 
gave  the  boron  flame  reaction,  was  reduced  to  powder  and  extracted 
with  ether,  which  removed  a  glassy,  uncrystallizable,  extremely  deli- 
quescent body.  It  was  freely  soluble  in  alcohol  and  was  left  as  a 
gummy  film  after  the  spontaneous  evaporation  of  the  alcohol.  Upon 
29 
