380  CRYSTALLIZATION  OF  SUPERSATURATED  SOLUTIONS,  ETC. 
I.  The  Crystallization  of  the  Supersaturated  Solution  of 
Sulphate  of  Soda  is  Determined  by  the  Fall  of  a  Solid  Body. — 
We  know  that  supersaturated  solutions  exposed  to  the  air  in  a 
laboratory  always  end  by  forming  into  a  mass.  By  varying  the 
experiments,  I  have  ascertained  the  conditions  necessary  for  the 
crystallization.  It  is  always  produced  in  matrasses  or  flasks,  if 
the  dust  in  the  air  will  by  falling  vertically  encounter  the  surface 
of  the  liquid.  I  found  that  in  the  same  place  the  crystallization 
is  much  more  rapid  where  the  surface  directly  in  contact  with 
the  air  is  largest,  or  where  the  air  has  easiest  access ;  it  is  very 
slow  where  the  orifice  is  narrow,  in  which  case  it  always  begins 
at  one  point,  whence  it  extends  in  all  directions  in  the  form  of 
radiated  needles  ;  and  if,  for  instance,  this  takes  place  in  a  bal- 
loon with  a  straight  neck,  this  point  is  invariably  in  the  hori- 
zontal projection  of  the  orifice. 
By  inclining  the  neck  of  the  balloon  or  matrass  containing 
the  liquid  while  still  hot,  or  by  leaving  in  the  usual  position  a 
retort  filled  with  the  liquid,  no  crystallization  takes  place.  It 
would,  then,  seem  necessary  that  the  corpuscles  of  the  air  should 
be  able  to  reach  the  surface  of  the  liquid  in  their  vertical  fall. 
Several  experiments,  especially  Lcewel's,  would  have  led  to 
the  opinion  that  crystallization  is  due  to  the  presence  of  a  solid 
body  in  the  air,  had  not  other,  and  apparently  contradictory, 
experiments  been  made.  Thus  air  which  has  traversed  cotton 
or  starch  no  longer  determines  crystallization.  I  have  found 
that  all  causes  which  arrest  the  movement  of  the  solid  bodies  in 
suspension  in  the  air  produce  the  same  effect.  To  ascertain  this 
I  merely  repeated  with  supersaturated  solutions  the  experiments 
made  by  M.  Pasteur  with  putrescible  liquids ;  among  others,  the 
experiment  with  balloons  with  sinuous  necks. 
Finally,  hardly  perceptible  dust  deposited  on  any  body  put  in 
contact  with  a  supersaturated  solution  always  determines  its 
solidification.    It  is  the  same  with  a  thick  layer  of  dust. 
II.  The  Solid  Body  Determining  the  Crystallization  is  Soluble 
in  Water. — In  a  second  series  of  experiments  I  found :  1.  That 
the  bodies  determining  the  crystallization  of  the  supersaturated 
liquid  lose  their  property  when  washed  in  ordinary  water  and 
dried,  when  cooled  in  flasks  having  a  layer  of  sulphuric  acid  at 
