39 
disappearing  from  the  upper  part  of  the  liquid  and  becoming 
increasingly  dense  below. 
The  amount  of  quinine  in  the  supernatant  liquid  so  obtained 
was  then  estimated,  and  from  this  the  total  amount  of  quinine  still 
remaining  in  solution  determined.  The  method  of  estimation 
employed  was  that  of  Giemsa  and  Schaumann,*  slightl}'  modified  to 
suit  the  conditions  under  which  our  investigation  was  carried  out. 
To  the  supernatant  liquid,  which  was  contained  in  a  stoppered 
glass  bottle  holding  300  c.cm.,  about  2  g.  of  solid  potassium  hydrate 
were  added.  When  the  potassium  hydrate  had  dissolved,  the 
liquid  had  111  most  cases  become  milky  in  aspect,  though  no  precipi¬ 
tate  was  recognisable.  70  c.cm.  of  ether  were  now  added,  the 
milkiness  at  once  disappearing.  The  mixture  of  ether  and  water)' 
solution  was  then  shaken  for  half  an  hour.  If  the  ether  did  not 
separate  out,  a  small  amount  of  alcohol  was  added,  when  ready 
separation  took  place.  The  ether  was  then  transferred  to  a  flask, 
by  means  of  a  capillary  siphon,  the  amount  of  the  upper  layer  of 
ether  left  behind  being  one-tenth  to  one-twentieth.  Ether  was  now 
added  again  in  three  successive  amounts  of  50  c.cm.,  and  siphoned 
off  as  before,  in  each  case  after  half  an  hour’s  shaking.  It  may  here 
be  mentioned  that  three  extractions  with  ether  as  recommended  by 
Giemsa  and  Schaumann  were  not  always  sufficient  to  remove  the 
last  traces  of  quinine.  The  ether  was  then  distilled  off  and  the 
residue  dried  in  a  water  oven  at  98°  C.  When  quite  dry  the  residue 
was  taken  up  with  chloroform.  The  chloroform  extract  was  put  up 
in  a  weighing  bottle,  allowed  to  evaporate  in  a  water  oven  at  70°  C., 
then  kept  at  I20°G.  for  one  hour  and  weighed.  The  residue  thus 
obtained  w’as  glassy  in  appearance,  and  colourless  or  of  a  faint 
brownish  lint.  It  was  found  necessary  to  re-distil  the  ether  and 
chloroform,  and  to  test  for  the  presence  of  any  non-volatile  residue 
before  use.  In  the  hands  of  Giemsa  and  Schaumannt  three 
estimations  of  0'05  g.,  O'l  g.  and  0’02  g.  of  quinine  respectively  gave 
an  experimental  error  of  -  0  4  per  cent.,  -t-  ro  per  cent,  and  —  i'2  per 
cent.  In  our  estimations  two  control  amounts  of  o’lioog.  are 
*  Pharmakalogische  und  chemisch-physilcalische  .Studien  iibi'r  Cliiniti.  .‘\rch.  I. 
.SchifTs-  11.  Tropenhygieiit,  1907,  Bd.  XI,  S.  7. 
I  l.oc.  cil.  p.  18. 
