§52 
Determination  of  Morphine.        /  Am  jour.  Pharm. 
;  r  1       Dec,  1918. 
Of  those  methods  which  direct  that  morphine  be  extracted  by 
lime,  that  of  Douzard  {Pharm.  J.,  1903  [IV],  17,  909-910)  is  a 
popular  one.  It  suffers  from  a  fundamental  error  in  reckoning  that 
r  Cc.  of  3^-H2S04  =  0.0283  gramme  of  anhydrous  morphine.  Ref- 
erence to  experiments  1  and  2  in  Part  I  of  this  communication  will 
show  that  the  figure  derived  directly  from  theory  is  correct. 
An  objection  against  all  methods  in  which  lime  is  used  is  that 
while  aliquot  parts  of  the  extract  are  taken  for  the  determination, 
these  are  reckoned  on  a  basis  other  than  that  of  their  relation  to  the 
water  known  to  be  present:  instead  of  the  mixture  of  opium,  lime, 
and  water  being  made  to  a  measured  volume,  it  is  a  measured  volume 
of  water  which  is  added.  An  irregularity  in  the  amount  of  water 
previously  present,  or  formed  in  the  reaction,  will  thus  produce  a 
corresponding  error. 
In  all  these  methods,  the  morphine  is  precipitated  in  the  presence 
of  alcohol,  and  sometimes  the  proportion  of  alcohol  is  large.  The 
magnitude  of  the  errors  to  which  this  gives  rise  is  fairly  well  known, 
and  is  again  displayed  in  Experiments  8,  9  and  10,  Part  I  of  this 
communication.  Many  standard  methods  introduce  fixed  corrections 
to  meet  this  loss.  Douzard  (Joe.  cit.)  adds  0.05  Gm.  of  anhydrous 
morphine  to  the  amount  found  by  titration  in  a  reputed  4  Gm.  of 
opium,  i.  e.,  1.25  per  cent.  Stevens  {Pharm.  Arch.,  V,  41-45)  adds 
1. 12  per  cent,  to  the  percentage  of  crystallized  morphine  found  by 
his  titration.  The  British  Pharmacopoeia  does  not  specify  a  correc- 
tion, but  admits  a  possible  error  of  —  1  per  cent. 
In  the  latter  case,  even  if  the  solvent  effect  of  alcohol  and  ether 
be  disregarded,  the  loss  through  water  alone  (of  which  104  Cc.  is 
present)  cannot  be  less  than  0.65  per  cent,  as  shown  in  Part  1,  Ex- 
periment 12. 
"  Corrections "  of  this  order  of  magnitude,  which  represent 
about  10  per  cent,  of  the  morphine  present  in  average  opium,  must 
give  rise  to  grave  doubts  of  the  accuracy  obtained.  Furthermore, 
it  makes  it  impossible  to  use  such  methods  when  the  percentage  of 
morphine  is  small,  and  analyses  of  opium  are  on  record  showing 
as  little  as  2  per  cent,  of  morphine.  The  writer,  himself  has  ex- 
amined a  sample  of  smoking  opium  which  gave  no  morphine  when 
submitted  to  the  United  States  Pharmacopoeia  method  of  estimation, 
so  small  was  the  actual  proportion  and  so  great  the  errors  of  the  pro- 
cedure.   Yet  this  sample  was  obtained  from  an  illicit  dealer  whose 
