A^anuaryFi903m'}    Presence  of  Arsenic  in  Chemicals.  37 
strength.  The  solutions  were  consecutively  so  diluted  that  the 
writer  was  in  a  position  to  know  just  how  much  arsenic  was  added 
in  each  operation,  and  carried  to  such  an  attenuation  that  only  the 
minutest  coloration  was  developed  in  the  reactions. 
On  comparing  the  above  results  with  those  obtained  by  other 
observers,  it  can  be  readily  seen  that  the  Marsh-Berzelius  test  did 
not  appear  to  be  as  delicate  as  frequently  represented.  The  writer 
thinks,  however,  that  this  is  chiefly  due  to  the  fact  that,  with  few 
exceptions,  experimenters  have  given  the  degree  of  dilution  of  the 
solution  without  mentioning  the  quantity  employed.  The  point 
made  by  the  late  Dr.  Wormley1  in  this  connection  is  well  worth 
repeating.  He  says :  "  Thus,  it  has  been  stated  that  the  method 
(Marsh)  will  yield  satisfactory  deposits  when  the  solution  contains 
only  1 -2,000,000th  of  its  weight  of  arsenic.  This  is  true,  but  it 
requires  about  1,000  grains  of  such  a  solution  to  furnish  these 
results;  the  absolute  quantity  of  oxide  present  would,  therefore,  be 
about  i-2000th  of  a  grain." 
In  preparing  the  standard  tubes,  it  is  desirable  to  make  two  or 
more  so  as  to  be  certain  that  the  tubes  are  alike  and  uniformity 
exists  in  the  operation.  It  is  possible  that  the  standard  tubes  of 
one  worker  will  vary  slightly  from  those  of  another,  but  this  will 
not  materially  affect  results  if  details  are  observed.  The  tubes  must 
be  prepared  from  the  same  zinc  and  reagents  subsequently  to  be 
employed  for  the  work. 
How  long  the  tubes  will  be  reliable,  time  only  can  tell.  Those  in 
the  writer's  possession,  made  over  a  year  ago  and  carefully  sealed, 
show  by  comparison  with  new  tubes  not  to  have  changed  in  the 
least. 
After  making  a  careful  study  of  the  above  three  methods,  both  as 
to  reliability  and  ease  of  execution,  it  was  decided  to  employ  the 
Marsh-Berzelius  method  throughout.  This  process  is  frequently 
decided  against  as  being  too  difficult  of  operation.  Such  an  objec- 
tion is  probably  justified,  but  it  is  the  writer's  experience  that  when 
pure  reagents  are  at  hand,  and  the  apparatus  is  once  set  up,  the 
results  are  obtained  with  less  attention  and  trouble  than  by  any 
other  method. 
It  is  always  wise  to  make  two  mirrors — one  for  comparison  with 
1  1885,  "  Microchemistrv  of  Poisons,"  2d  Ed.,  p.  285. 
