CRYSTALLIZATION  OF  SUPERSATURATED  SOLUTIONS,  ETC.  381 
the  bottom.  In  the  same  way,  dust  slowly  dried  after  being 
washed  becomes  perfectly  inactive. 
2.  I  passed  into  a  saturated  solution,  which  had  remained 
intact,  more  than  1500  litres  of  air  washed  in  distilled  water, 
and,  to  use  the  least  possible  quantity  of  water,  I  passed  the  air 
through  a  series  of  inclined  glass  tubes,  in  each  of  which  a  drop 
of  water  uplifted  by  the  bubble  of  air  rose  the  length  of  the  tube, 
diminishing  in  volume,  and  forming  again  in  the  lower  part  of 
the  tube,  thus  serving  an  indefinite  number  of  times. 
3.  A  layer  of  oil  poured  on  a  solution  of  sulphate  of  soda  pre- 
serves it  from  contact  with  the  air  and.  prevents  its  crystallizing. 
In  place  of  oil  I  employ  a  layer  of  water,  when  no  crystallization 
is  produced  by  plunging  a  rod  into  the  solution ;  at  least,  if  this 
is  done  very  rapidly,  and  there  is  no  thick  coating  of  deposited 
matter  on  the  rod,  in  which  case  should  a  morsel  fall  into  the 
solution  and  not  be  completely  dissolved,  the  solution  will  be 
crystallized  before  the  rod  has  touched  the  supersaturated 
solution. 
III.  The  Solid  Body  Determining  the  Crystallization  loses  its 
Property  under  the  Influence  of  Heat. — I  verified  this  fact,  which 
has  long  been  established. 
IY.  The  Air  ivhich  Determines  Crystallization  contains  Sul- 
phate of  Soda. — I  collected  the  few  drops  of  water  which  had 
washed  the  air  in  the  inclined  tubes  in  experiment  section  II., 
and  which  contained  the  soluble  matters  of  more  than  1500  litres 
of  air ;  they  gave,  with  chloride  of  barium,  a  precipitate  of  sul- 
phate of  baryta ;  one  drop  of  liquid  gave,  with  the  spectroscope, 
the  soda  ray  with  remarkable  intensity. 
Dust  deposited  even  out  of  the  laboratory,  treated  by  distilled 
water,  gave  a  lixivium  presenting  the  same  reactions,  and  con- 
tained considerable  quantities  of  soda  and  of  a  soluble  sulphate. 
In  short,  the  crystallization  of  the  supersaturated  solution  of 
sulphate  of  soda  is  determined  by  the  contact  of  a  solid  body 
soluble  in  water,  alterable  under  the  influence  of  heat,  giving 
with  chloride  of  barium  a  precipitate  of  sulphate  of  baryta,  and 
containing  soda ;  these  are  exactly  the  characteristics  of  ordinary 
sulphate  of  soda,  which  would  then  determine  the  crystallization 
of  its  supersaturated  solutions. 
