REMARKS  ON  DIALYSIS. 
313 
Liquid  Diffusion  is  the  tendency  which  particles  of  matter  in 
solution  have  to  move  from  one  part  of  a  liquid  menstruum  to 
another,  even  in  opposition  to  the  force  of  gravity,  under  par- 
ticular circumstances. 
The  first  arrangement  suggested  by  Mr.  Graham  consists  in 
placing  a  small  jar,  containing  a  solution  of  the  substance  to  be 
diffused,  in  a  large  jar  of  distilled  water,  so  carefully  that  the 
fluids  be  not  mixed  mechanically,  when,  on  standing,  the  dissolv- 
ed body  will  be  found  to  have  diffused  itself  into  the  distilled 
water.  Whilst  muriatic  acid,  common  salt,  cane  sugar,  and 
many  other  bodies  possess  this  property  in  a  strong  degree, 
others,  such  as  albumen,  gum,  etc.,  are  found  to  enjoy  it  but 
slightly.  Now  Mr.  Graham  discovered  that,  as  a  general  rule, 
crystallized  bodies  diffuse  much  more  readily  than  amorphous 
substances,  and  hence  he  has  called  the  former  Crystalloids,  and 
the  latter  Colloids,  from  eollin,  the  best  sort  of  gelatin.  (See 
,Mr.  Graham's  paper.) 
Of  the  bodies  heretofore  tested,  hydrochloric  acid  is  the  most 
diffusive  ;  and  assuming  this  body  as  unity,  the  times  of  equal 
diffusion  for  several  substances  are  as  follows  :  hydrochloric 
acid  1  ;  chloride  of  sodium  2-33  ;  sugar  7  ;  sulphate  of  magne- 
sia 7  ;  albumen  49  ;  caramel  98. 
Mr.  Graham  considers  crystallizable  bodies  to  be  in  a  statical 
condition,  whilst  the  colloids  are  in  a  dynamical  state;  that  is 
to  say,  the  latter  have  a  greater  susceptibility  to  the  influence 
of  external  agents,  and  from  this  cause  are  adapted  to  the  trans- 
forming influences  of  the  processes  of  organic  life,  like  albumen, 
gelatin  and  fibrin  of  the  animal  body.  These  bodies  generally 
form  a  feeble  union  with  water,  and  the  solid  hydrates  of  some 
colloids  possess  a  remarkable  property  in  reference  to  the  dif- 
fusion of  other  bodies.  For  instance,  jelly  of  gelatin  will  per. 
mit  crystalline  matter  in  an  aqueous  solution  in  contact  with  it 
to  diffuse  or  pass  through  or  into  it,  whilst  it  will  refuse  a  pas- 
sage partially  or  wholly  to  a  body  of  the  colloid  class.  Accord- 
ing to  Prof.  Redwood,  if  ten  per  cent  of  chloride  of  sodium  be 
added  to  one  part  of  such  a  jelly  while  rendered  fluid  by  heat, 
and  this  having  been  put  into  the  bottom  of  a  diffusion  jar,  the 
latter  be  carefully  filled  up  with  seven  parts  of  pure  liquified  jelly, 
after  the  first  has  solidified,  and  quickly  cooled,  the  chloride  of 
