ASSAY  OP  ALKALOIDS. 
13 
whole  mixture.  If,  after  the  reaction  on  the  watch-glass  be- 
comes indistinct,  the  glass  beaker  is  shaken  for  one  minute,  the 
mixture  will  settle  almost  clear,  and  upon  now  adding  drop  for 
drop  more  of  the  reagent,  it  sinks  to  the  bottom  and  at  once 
forms  a  dense  cloud  over  the  stratum  of  precipitate,  if  any 
quinia  be  left  in  solution.  Whenever  the  cloud  no  longer  makes 
its  appearance,  the  numbers  consumed,  with  the  usual  deduction 
for  surplus,  will  represent  6  equivalents  of  mercury.  In  the 
meantime  the  color  of  the  precipitate  will  have  changed  to  white. 
My  view  of  the  cause  for  this  retardation  is,  that  2  eq.  of 
quinia  require  6  eq.  of  mercury  for  complete  precipitation  ;  that 
when  a  considerable  excess  of  KI  has  accumulated,  it  will  tend 
to  re-dissolve  a  portion  of  Hgl,  and  destroy  the  yellow  color; 
and  that  this  is  the  cause  why  a  fresh  solution  of  quinia  or  any 
other  alkaloid  will  then  produce  a  precipitate  ;  while  in  the 
dilute  mixture  the  action  of  the  now  very  much  diluted  alkaloid 
on  the  mercury  is  overpowered  by  the  solubility  of  the  latter  in 
the  great  excess  of  KI ;  but  that  when  a  fresh  quantity  of  mer- 
cury is  added,  held  in  solution  by  a  small  excess  of  KI  only, 
this  will  in  part  rescue  the  re-dissolved  mercury  and  then  act 
on  the  remaining  alkaloid,  similar  to  a  solution  of  nitrate  of 
silver  that  will  precipitate  all  the  diver  from  a  solution  in  cya- 
nide of  potassium,  but  is  itself  involved  in  the  precipitation. 
In  fact,  the  filtrate  shows  the  presence  of  mercury  if  none 
whatever  is  added  in  excess,  but  not  a  trace  of  quinia.  I  think 
nothing  remains  but  to  conclude  that  the  mercury  in  solu- 
tion will  have  to  be  deducted  from  the  formula  of  the  alkaloid- 
precipitate,  which  originally  was  2Chi+3HI+6HgI,  and  became 
+ 
2Chi+3HI-f5HgI,  or  was  changed  in  some  similar  manner. 
How  closely  my  reasoning  coincides  with  what  actually  occurs 
is  shown  by  Mr.  Groves  himself.  The  formula  for  the  quinia 
compound,  which  he  published  in  1858  and  confirms  in  this 
paper,  is  Chi  HI-{-2HgL    As  will  be  seen,  the  result  of  his 
analysis  does  not  agree  with  the  formula ;  but  to  admit  that  the 
formula  was  incorrect  would  upset  his  theory  of  the  precipitate. 
So  we  have  : 
