A^uS^898.rm'}     Caustic  Soda  Sold  as  Caustic  Potash.  397 
The  best  crude  caustic  potash  available  in  the  New  York  market 
at  present  is  not  greatly  inferior  in  quality  and  strength  to  the 
standard  set  by  the  Pharmacopoeia  for  the  purified  article,  as  the 
subjoined  analysis  of  a  sample  recently  examined  will  show.  It  can 
be  obtained  in  drums  containing  10  pounds,  or  in  any  larger  quantity 
desired.  It  is  in  the  form  of  a  yellowish-white,  coarse,  granular 
powder,  and  gives  a  nearly  clear  and  colorless  solution.  The  com- 
position is  as  follows : 
Per  cent. 
Potassium  hydrate   86*4 
Sodium  hydrate   i*8 
Potassium  carbonate   4*85 
Potassium  chloride   14 
Potassium  silicate   o'  1 
Water   5-55 
A  total  absence  of  sulphate,  as  noted  in  this  sample,  is  unusual  in 
crude  potash,  and  the  chloride  is  seldom  less  than  2  per  cent. 
Alumina  and  lime  are  practically  absent. 
The  methods  used  in  the  examination  of  this  sample  and  those 
discussed  at  the  beginning  of  this  paper,  are  essentially  those  recom- 
mended by  Fresenius.  Potassium  was  estimated  by  precipitation 
with  platinum  chloride,  and  in  some  cases  the  results  were  confirmed 
by  weighing  the  mixed  chlorides  of  potassium  and  sodium,  and  esti- 
mating the  chlorine  in  the  mixture.  Further  corroboration  of  the 
absence  of  more  than  small  quantities  of  potassium  in  some  of  the 
samples  was  obtained  by  the  failure  of  precipitation  with  an  access 
of  tartartic  acid  or  with  sodium  cobaltic  nitrite. 
In  this  connection  an  experiment  was  made  to  determine  the 
sensitiveness  of  the  sodium  cobaltic  nitrite  test.  Pure  potassium 
bicarbonate  was  supersaturated  with  acetic  acid  and  diluted  to  defi- 
nite volumes.  On  adding  the  test  solution  to  a  solution  containing 
the  equivalent  of  0  3  per  cent,  of  potassium  hydrate,  a  turbidity 
began  to  appear  in  1^  minutes.  A  0-2  per  cent,  solution  became 
faintly  turbid  in  4  minutes.  Greater  dilutions  failed  to  react  even 
after  standing  over  night.  To  detect  for  instance,  2  per  cent,  of 
potash  in  a  sample  of  soda,  the  test  would  have  to  be  applied  to  a 
solution  containing  at  least  10  per  cent,  of  the  sample.  Very  rough 
quantitative  estimations  of  potash  may  be  made  by  diluting  a  solu- 
tion of  the  sample  to  the  point  at  which  a  reaction  is  barely  visible 
within  4  minutes  and  assuming  this  dilution  to  contain  0  2  per  cent. 
