484  American  Pharmaceutical  Association.     { Ase!ptemb6v^s?* 
defects  are:  (1)  the  boiling  with  milk  of  lime,  (  2)  prolonged  heating  necessary 
to  dry  the  drug  after  that  treatment,  (3)  weighing  the  caffein  instead  of  esti- 
mating it  with  Wagner's  reagent  volumetrically,  (4)  the  means  employed  to 
separate  the  so-called  "kolanin,"  (5)  estimating  caffein  kolatannate,  "  kola- 
nin,"  by  weight  as  such  instead  of  by  its  caffein  content.  (1)  Boiling  with 
milk  of  lime  or  other  aqueous  alkalies  tends  to  decompose  caffein  itself  and  to 
liberate  caffein  from  its  tannate,  and  thus  prevent  an  exact  determination  of  the 
proportion  of  "free  "  to  combined  alkaloids  that  originally  existed  in  the  drug. 
The  starch  of  kola,  amounting  to  nearly  40  per  cent.,  causes  the  drug  to 
become  of  a  jelly-like  consistency  when  boiled  with  aqueous  liquids,  and  when 
dried  the  mass  is  very  difficult  to  powder,  and  even  when  powdered  the  condition 
of  it  is  well  calculated  to  resist  the  penetrating  action  of  the  solvent,  and  thus 
cause  incomplete  extraction.  (2)  The  gelatinous  condition  of  the  drug  after 
boiling  with  lime  water  renders  it  hard  to  dry  completely,  and  as  very  pro- 
longed heating  is  necessary,  there  is  danger  of  loss  of  caffein  by  sublimation. 
(3)  We  do  not  consider  that  the  alkaloids  of  kola  are  sufficiently  pure 
when  removed  in  this  manner  to  be  weighed  as  such.  Gomberg's  volu- 
metric method  gives  more  accurate  results.  (4)  If  any  caffein  kolatannate 
escapes  decomposition  during  the  first  part  of  the  assay,  which  will  occa- 
sionally happen,  it  is  removed  by  the  extraction  with  alcohol.  But  as 
it  is  somewhat  soluble  in  water  and  more  soluble  in  aqueous  solutions  contain- 
ing tannin,  the  directions  to  collect  and  weigh  the  portion  left  undissolved  by 
water  will  generally  be  found  superfluous,  for  the  small  amount  of  it  present 
will  generally  pass  into  and  remain  in  solution.  Dieterich  had  this  experience 
and  met  with  nothing  but  disappointment  in  trying  to  estimate  caffein  kolatan- 
nate by  this  method.  (5)  Caffein  kolatannate  has  been  shown  to  be  a  body  of 
somewhat  variable  composition,  with  a  caffein  content  ranging  from  19  to  25 
per  cent.  Inasmuch  as  the  value  of  kola  as  a  stimulant  depends  primarily  on 
its  percentage  of  caffein,  and  as  the  weight  of  its  caffein  compound  indicates 
only  approximately  the  amount  of  its  combined  caffein,  there  would  not  seem 
to  be  any  reason  for  attaching  much  importance  to  the  weight  of  this  compound 
if  a  very  precise  valuation  of  the  drug  is  desired.  We  hold  that  a  direct  esti- 
mation of  the  caffein  of  this  compound  is  preferable,  and,  at  least,  as  expedi- 
tious. 
Method  of  Carles. — Ten  grammes  of  kola,  1  gramme  of  calcium  hydroxide, 
and  20  grammes  of  80  per  cent,  alcohol,  are  mixed  together  and  dried  on  the 
water-bath  until  the  weight  is  reduced  to  14  grammes.  The  mixture  is  then 
powdered  and  transferred  to  a  100  c.c.  flask  containing  35  c.c.  of  a  mixture  of 
100  parts  of  chloroform  and  20  parts  of  alcohol,  and  heated  for  one  hour  on  a 
water-bath.  After  filtration,  the  residue  is  extracted  next  with  20  c.c.  of  the 
same  solvent,  and  finally  with  10  c.c.  The  united  extracts  aie  evaporated  to 
dryness  and  the  residue  taken  up  with  10  c.c.  of  boiling  water,  containing  four 
or  five  drops  of  1  per  cent,  sulphuric  acid,  then  with  6  c.c,  and  finally  with  5  c.c. 
The  solutions  are  united,  filtered,  evaporated  to  constant  weight  and  weighed 
as  caffein.  For  "kolanin,"  which  the  author  recognizes  as  caffein  kolatan- 
nate, another  sample  of  the  drug  is  taken  and  extracted  with  water  to  remove 
the  caffein  and  other  soluble  constituents.  The  drug  is  next  extracted  with 
70  per  cent,  alcohol,  the  ex'ract  evaporated  to  dryness,  transferred  to  a  filter, 
washed  with  cold  w&ter,  ihen  dried  by  a  gentle  heat  and  weighed.    If  it  is 
