\ 
152        ON  WEIGHING  VERY  SMALL  PORTIONS  OF  MATTER. 
succeed  in  obtaining  the  weights  of  extremely  small  parts  of 
matter.  For  that  purpose  I  tried  the  elasticities  both  of  torsion 
and  flexure,  and  found  the  latter  only  to  answer  the  purpose. 
The  following  is  a  description  of  the  construction  of  my  ap- 
paratus with  which  I  have  succeeded  in  estimating  portions  of 
matter  equal  in  weight  to  the  thousandth  part  of  a  milligram.* 
Heating  a  rod  of  soft  glass  in  one  spot  to  bright  redness,  I  drew 
it  out  quickly,  and  thereby  obtained  a  filament  uniformly  cylin- 
drical, of  about  the  diameter  of  fine  human  hair.  Taking  from 
the  middle  of  this  fine  glass  thread  a  piece  of  such  a  length 
i  about  three  inches)  that  its  weight  would  barely  reduce  it  from 
the  horizontal,  one  end  of  it  was  fastened  by  means  of  good  seal- 
ing wax  to  the  edge  of  a  mahogany  block,  and  the  other  end 
slightly  hooked  by  approaching  quickly  a  small  spirit  flame.  In 
order  to  obtain  a  pan  in  which  to  place  the  substance  whose 
weight  I  would  estimate,  I  cut  with  the  common  microscopic 
section-cutter  some  disks  of  elder  pith  from  -001  to  »002  inch  in 
thickness  ;  and  drawing  out  a  still  finer  filament,  the  end  was 
likewise  hooked,  and  the  other  extremity  being  passed  through 
a  pith  disk,  a  small  knob  of  glass  was  made  on  this  end  by  the 
spirit  flame,  just  of  sufficient  size  to  prevent  this  disk  slipping 
off  the  suspending  rod.  The  filament  with  attached  disk  was 
now  hooked  on  the  end  of  the  rod  fixed  to  the  block,  and  was 
then  ready  for  graduation. 
Not  being  able  at  the  time  to  procure  silver  wire  of  sufficient 
fineness,  I  substituted  some  very  fine  and  long  hair,  taken  from 
the  head  of  a  child ;  and  having  brought  the  centre  of  gravity 
and  centre  of  motion  of  a  very  sensitive  analytical  balance 
almost  to  coincide,  I  obtained  a  piece  of  the  middle  of  a  hair 
weighing  exactly  one-half  milligram.  This  being  divided  into 
five  equal  parts  (each  about  one  inch  long)  gave  us  tenths  of  a 
milligram.  One  of  these  tenths  being  placed  on  the  pithpan, 
the  glass  filament  was  deflected  a  certain  quantity,  which  was 
marked  on  an  arc  formed  of  bristol  board,  and  so  as  to  be  almost 
touched  by  the  deflected  rod  in  its  revolution  about  the  edge  of 
the  block.  Another  tenth  was  added  and  another  division  ob- 
tained :  and  so  on,  until  all  five  divisions  were  marked,  the 
*  [l-65000th  of  a  grain,  Troy.] 
