388      Hypophosphorous  Acid  and  Hypophosphites.  {AnAjgust^88a9rm' 
The  insoluble  salt,  Fea  (H2  P02)6  was  the  cause  of  much  work,  the 
result  of  which  establishes  the  superiority  of  the  bromine  method  and 
the  unreliability  of  the  HgCl2  method  in  determining  ferric  hypo- 
phosphite.  The  objection  to  the  latter  method  is  that  the  insolubility 
of  this  salt  demands  it  to  be  first  brought  in  solution,  by  use  of  HC1, 
before  the  addition  of  HgCI2 ;  in  dissolving,  however,  a  portion  of  the 
ferric  salt  (more  or  less,  depending  on  conditions)  is  reduced  to  ferrous, 
with  simultaneous  oxidation  of  the  hypophosphite,  which  does  not 
reduce  mercuric  chloride ;  hence,  the  results  are  low  ;  if  the  hypophos- 
phite be  not  dissolved  the  particles  coming  in  contact  with  the  sepa- 
rated mercurous  chloride  reduce  this  to  metallic  mercury  and  loss 
again  results.  If  heat  be  applied  to  assist  the  solution  of  the  iron  salt, 
the  results  are  still  lower.  One  sample  by  Br  method  gave  98*80  per 
cent.  Fe2  (H2  P02)6 ;  with  HgCl2  method  using  5  cc.  HC1  and  5  cc.  water 
to  dissolve  the  hypophosphite  and  then  adding  the  HgCl2  and  10  cc. 
water  before  heating,  working  as  rapidly  as  possible,  only  94*82  per 
cent,  was  obtained;  working  more  leisurely  in  dissolving  the  salt 
84*73  per  cent,  resulted.  Another  sample  with  Br  gave  90*32  per 
cent. ;  with  HgCl2,  as  above,  working  rapidly,  89*28  per  cent ;  apply- 
ing a  little  heat  in  dissolving,  86*41  per  cent. 
The  difficulties  encountered  in  the  Br  method  were  too  high  results 
due  partly  to  the  bromine  but,  chiefly,  to  the  ready  decomposition  of 
ferric  phosphate  in  the  titration  with  NaOH.  No  indicator  was 
found  with  this  salt  which  would  indicate  sharply  the  change  of  acid 
phosphate  into  normal  phosphate  :  Phenolphthalein  was  useless  because 
it  only  produces  the  red  color  in  alkaline  solution,  even  if  a  slight  ex- 
cess of  NaOH  was  added  this  would  instantly  react  with  the  ferric  phos- 
phate and  a  neutral  solution  result ;  by  the  time  an  alkaline  reaction 
was  obtained  the  greater  part  of  the  ferric  phosphate  had  been  decom- 
posed ;  Methyl-orange  and  Tropaeolin  00  do  not  appear  to  be  sensi- 
tive towards  acid  phosphates  even  after  addition  of  calcium  chloride ; 
Congo-red  and  Litmus,  sensitive  to  acid  phosphates  after  addition  of 
calcium  chloride  (the  former  only  after  warming),  can  be  made  to  give 
fair  results  but  the  end-reaction  is  rather  indistinct  owing  to  forma- 
tion of  lakes  with  the  iron  phosphate.  This  objectionable  feature  was 
overcome  by  completely  decomposing  the  ferric  phosphate  by  addition 
of  normal  NaOH  to  the  boiling  hot  solution  of  ferric  phosphate 
(after  addition  of  the  CaCl2)  until  phenolphthalein  indicated  an  alka- 
line reaction.    This  procedure,  however,  necessitated  another  determi- 
