294  Finely  Divided  Metals  and  Ferric  Salts.  {^"^-/ubI'-^^"^' 
be  the  case  with  all  metals  M^'  where  M^'  -f  Hf)  =  M'^O  +  H2 
represents  an  endothermic  change.  Platinum  and  gold  do  not  reduce 
ferric  solutions.  ISTow,  with  the  exception  of  the  last  two  named,  the 
heats  of  formation  of  the  chlorides  of  all  the  foregoing  metals  are 
greater  than  55,540.  ^^'evertheless,  the  rapidity  of  reduction  bv  a 
metal  does  not  appear  to  be  a  function  of  the  energy  which  runs  down 
in  the  formation  of  its  chloride,  thus  [AP,  CP,  Aq]  =  475,650,  whilst 
[Zn,  CP,  Aq]  =  only  112,840;  yet  zinc  reduces  instantaneously, 
whilst  aluminium  reduces  the  most  slowly  of  all  the  metals  experi- 
mented with.  But  experiment  showed  it  to  be  undoubtedly  the  case 
that  those  metals  reduce  the  quickest  which  are  the  most  readily 
attacked  at  ordinary  temperatures  by  dilute  chlorine-water.  It  is  of 
interest  to  note  that  galena  in  a  finely-divided  state  also  reduces 
ferric  chloride  solution,  whereas  antimony  sulphide  has  not  this 
power. 
6.  From  the  whole  of  my  experiments,  I  conclude  that  zinc-dust  is 
practically  the  best  reducing  agent  for  the  purpose  in  hand.  True  it 
is  that  zinc-dust  may  sometimes  contain  a  little  iron,  and  that  titra- 
tion with  permanganate  cannot  be  conducted  in  an  acid  solution  con- 
taining zinc,  but  that  the  latter  must  be  first  removed.  But,  in  the 
first  place,  zinc  dust  contains  so  little  ii'on,  and  so  slight  a  solution  of 
zinc  takes  place  before  titration  by  my  method,  that  any  error  arising 
from  this  source  is  negligible.  However,  in  the  attempt  to  elaborate  a 
method  which  would  preclude  any  uncertainty  on  this  point,  I  pre- 
pared zinc-dust  free  from  zinc  oxide,  as  recommended  by  Sabatier,  by 
means  of  repeated  digestion  with  dilute  acid,  and  also  by  what  I  found 
to  be  a  more  rapid  method,  viz.,  by  digestion  with  solutions  of  ammo- 
nium chloride  and  ammonia,  in  both  cases  finally  drying  the  product 
on  porous  tiles  in  a  vacuum.  This  purified  zinc  was  shaken  up  with 
a  standard  ferric  chloride  solution  without  the  addition  of  any  acid  ; 
the  ferrous  solution  was  filtered  off,  acidified,  and  titrated  ;  but  the  iron 
was  not  fully  accounted  for  in  the  filtrate,  so  rapidly  does  w^ater  at- 
tack the  finely-divided  zinc  with  formation  of  hydrogen  and  zinc 
hydroxide;  the  latter  precipitating  solutions  of  iron  salts  in  con- 
tact with  it.  In  fact,  in  solutions  of  several  metallic  salts,  MgCl2, 
Al2(S04)3,  Co(N03,)2,  MnS04,  etc.,  finely-divided  zinc  very  soon  causes 
a  precipitate  either  of  the  hydroxides  or  of  basic  salts  of  the  metals 
present. 
In  the  second  place,  even  when  the  reduction  is  effected  by  metals 
