SALTS  OF  ALUMINA  IN  THE  ANALYSIS  OF  PLANTS.  153 
length  of  the  divisions  being  about  one-fourth  of  an  inch  ;  they 
were  very  readily  subdivided  into  ten  equal  parts  which  gave  me 
immediately  j^ths  of  a  milligram.  The  weight  of  any  quantity 
of  matter  less  than  one-half  milligram  may  be  now  estimated 
to  ^th  of  a  milligram  by  placing  it  on  the  pan  and  observing 
the  deflection. 
For  the  thousandths,  still  more  care  and  patience  is  required, 
the  filament  being  much  finer  and  somewhat  shorter,  and  the 
pith  disk  smaller  and  as  thin  as  possible.  In  order  to  obtain  the 
primary  graduations  of  hundredths,  one  of  the  above  pieces  of 
hair  equal  to  ith  milligram  is  divided  into  ten  equal  parts,  which 
gives  us  weights  of  ^th  milligram.  The  deflections  caused  by 
these  weights,  divided  into  ten  equal  parts,  give  ^  of  a  milli- 
gram. 
As  the  least  breath  of  air  interferes  with  the  graduations  and 
weighing,  the  whole  instrument  is  protected  by  a  glass  case,  the 
end  of  the  case  next  the  graduated  arc  being  on  a  hinge. 
In  elastic  rods  of  square  section,  the  deflection  is  proportional 
to  the  weight ;  in  those  of  circular  section  this  law  is  slightly  de- 
parted from  ;  but  by  the  above  method  of  ascertaining  directly 
the  value  of  each  division,  the  error  is  avoided. 
Those  who  may  have  the  necessity  to  construct  this  apparatus 
should  arm  themselves  beforehand  with  scrupulous  care  and  un- 
bounded patience. 
From  the  great  simplicity  of  the  above  arrangement,  it  seems 
very  strange  that  some  person  did  not  long  ago  invent  it ;  but, 
to  my  knowledge,  it  has  never  been  attempted. — American  Jour- 
nal of  Science  and  Arts,  January,  1858. 
ON  THE  EMPLOYMENT  OF  THE  SALTS  OF  ALUMINA  IN  THE 
ANALYSIS  OF  PLANTS. 
Rochleder  has  pointed  out  the  superiority  of  alumina  over 
hydrate  of  oxide  of  lead  for  the  separation  of  the  proximate  con- 
stituents of  plants.  The  author  in  the  first  place  remarks  that 
organic  substances  may  be  divided  into  two  classes  with  refer- 
ence to  their  behaviour  toward  alumina.  Many  coloring  matters, 
as  well  as  other  substances,  are  precipitated  by  alumina  from 
their  solutions,  while  others  on  the  contrary  are  not  affected. 
