AmAJugU,ri8P77arm'}      Methods  of  Estimating  Tannins.  419 
constant  stirring  becomes  very  tedious,  and  a  stream  of  air-bubbles 
forced  through  the  liquid  by  an  aspirator  may  be  substituted  with  great 
advantage. 
Neubauer  reckons  one  liter  of  decinormal  permanganate  as  equal  to 
4-157  grams  of  gallotannic  acid,  and  consequently  (according  to  Oser) 
to  6*235  of  oak  bark  tannin.  Further  research,  however,  is  needed 
before  percentages  can  be  calculated  with  certainty,  and  chemists,  in 
giving  results,  would  do  well  to  state  the  equivalent  in  permanganate, 
or  to  say  that  they  use  Neubauer's  or  Oser's  equivalent.  The  first  is 
applicable  to  sumach,  galls  and  myrobalans,  the  second  probably  to  oak 
bark,  valonia  and  chestnut  extract,  at  least  approximately.  It  is  a  sin- 
gular fact  that  gallic  acid  consumes  not  only  a  larger  proportion  of  per- 
manganate, weight  for  weight,  than  tannin,  but  even  a  larger  proportion 
than  the  tannin  from  which  it  is  derived,  as  I  proved  by  digesting  a 
solution  of  tannin  with  dilute  sulphuric  acid,  when  its  reducing  power 
was  notably  increased.  Hence,  commercial  tannin,  which  is  largely 
contaminated  with  gallic  acid,  consumes  more  permanganate  than  the 
above-mentioned  quantity. 
As  to  accuracy,  single  tests  should  never  differ  by  more  than  o*i  cc, 
or  say  i\  per  cent,  of  the  total  quantity,  but,  of  course,  in  so  rapid  a 
process,  no  one  would  rely  on  single  tests,  and  by  repeating  and  taking 
the  mean  any  required  accuracy  may  be  attained.  Separate  portions 
■of  liquor  precipitated  by  gelatin  give  identical  results,  at  least  within 
the  limits  named. 
I  should  perhaps  mention  that  Mr.  Estcourt  proposed  some  time 
since  to  precipitate  with  gelatin  in  conjunction  with  the  permanganate 
method  ("Chemical  News,"  xxix,  110),  but  as  he  heated  the  solution, 
and  tannate  of  gelatin  is  soluble  in  hot  gelatin  solution,  the  results  were 
not  satisfactory.  Still  he  undoubtedly  deserves  the  credit  of  the  idea, 
while  Lowenthal's  cold  gelatin  solution,  with  the  addition  of  salt  and 
acid,  completely  overcomes  the  difficulty. 
Several  other  oxidizing  methods  have  been  proposed.  Carpeni  pre- 
cipitates the  tannin  with  ammonio-acetate  of  zinc,  redissolves  and  esti- 
mates with  permanganate.  M.  Jean  oxidizes  with  iodine  in  solution 
of  sodic  carbonate,  and  M.  Pouchet  with  concentrated  permanganate 
in  a  caustic  potash  solution.  None  of  these  methods  seem  to  have 
any  advantage  over  Lowenthal's,  while  the  two  latter  are  in  my  expe- 
rience decidedly  inferior.    The  end-reactions  are  much  less  distinct, 
