738 
Purified  Ether. 
{Am.  Jour.  Pharm. 
November,  19 19. 
until  it  has  gained  room  temperature,  or,  preferably,  for  an  hour. 
Under  these  conditions  the  test  has  given  very  satisfactory  results. 
The  reason  for  adopting  this  procedure  was  the  experience  that, 
with  a  freshly-made  and  sensitive  Schiff,  a  bluish  violet  coloration 
was  always  obtained  that  was  very  confusing,  no  doubt  due  to 
acetaldehyde,  and  which  would  vary  from  time  to  time.  But  it  was 
noticed  that  by  heating  as  suggested  this  color  would  suddenly  dis- 
appear at  about  45 °,  and  then  the  color  due  to  the  formaldehyde 
asserts  itself.  With  0.001  Cc.  of  methyl  alcohol  it  requires  about  a 
quarter  of  an  hour,  but  with  0.003  Cc.  the  color  asserts  itself  almost 
as  soon  as  the  last  trace  of  the  fallacious  coloration  has  disappeared. 
As  thus  set  out  the  test  makes  comparison  of  an  "  apparent "  0.02, 
0.04,  and  0.06  per  cent,  of  methyl  alcohol  in  the  5  Cc.  of  ether. 
One  cannot  urge  too  strongly  the  desirability  of  running  the  con- 
trols every  time  a  test  is  made.  There  are  too  many  variables  in- 
fluencing the  test  for  one  to  rely  solely  on  judgment  of  coloration. 
Alkaline  Silver. — A  little  2  per  cent,  silver  nitrate  is  added  to  an 
equal  volume  of  caustic  soda  solution  (20  per  cent.)  and  cleared  by 
the  addition  of  just  sufficient  ammonia.  Very  pure  ethers  when 
shaken  with  this — say  about  5  Cc.  with  3  Cc.  of  reagent — give  very 
little  reduction  indeed ;  on  setting  aside  for  five  to  ten  minutes  some 
samples  are  all  but  negative  in  reaction. 
Formaldehyde  may  be  detected  by  the  usual  gallic  acid  ring  test. 
Acetone. — This  may  be  detected  qualitatively  by  applying 
Rothera's  modification  of  the  nitroprusside  reaction  as  described  by 
Finnemore  (P.  /.,  25/7/14,  pp.  139  and  160).  Scott-Wilson's 
method3  consists  in  precipitating  the  acetone  or  ketone  as  the  mer- 
cury compound  HgCsO  (HgCN)4  by  reacting  with  an  alkaline 
silver-mercury  cyanide  solution  and  determining  the  mercury  in  the 
precipitate.  It  is  an  exceptionally  delicate  reaction,  and  one  must 
always  be  careful  to  have  an  excess  of  reagent.  In  making  qualita- 
tive tests  nothing  is  easier  than  to  achieve  a  negative  indication 
through  taking  too  much  acetone.  Alcohol  does  not  appear  to  have 
any  influence  in  small  quantity,  but  aldehyde  reacts  after  the  manner 
of  acetone,  and  ammonia,  if  present  would  also  interfere.  The 
reagent  is  prepared  by  dissolving  5  Gm.  of  mercuric  cyanide  and  90 
Gm.  of  caustic  soda  in  600  Cc.  of  water,  then  1.5  Gm.  of  silver 
nitrate  in  200  Cc.  of  water,  and  gradually  pouring  this  solution  into 
the  former  under  constant  stirring.   The  reagent  is  then  set  aside  to 
3Joum.  Physiol,  1911,  Vol.  42,  444. 
