362 
Estimation  of  Nitroglycerin. 
<  Am.  Jour.  Pharm. 
\     August,  1.911. 
it.  It  suggests  the  Kjeldahl  method  of  estimating  nitrogen, — which 
the  writer  has  tried  on  nitroglycerin  with  good  results. 
For  pharmaceutical  purposes  the  need  applies  particularly  to  an 
estimation  of  minute  quantities,  as  in  i/ioo  grain  tablets,  etc. 
The  writer  first  attempted  tO'  use  "  nitron  "  for  this  purpose  but 
the  insolubility  of  nitroglycerin  in  water  makes  this  impracticable. 
The  use  of  phenol  disulphonic  acid,  as  in  water  analysis  for  the 
estimation  of  nitrates,  then  suggested  itself.  On  trial  this  worked 
well,  and  appeared  to  give  excellent  results.  It  then  remained  to 
ascertain  whether  such  results  were  correct. 
Since  this  reagent  must  be  applied  to  a  dry  residue,  the  first 
question  relates  to  the  volatility  of  nitroglycerin,  some  writers  hav- 
ing asserted  that  it  is  slightly  volatile. 
A  (supposedly)  10%  alcoholic  solution  was  evaporated  under 
three  conditions,  two  samples  of  10  c.c.  each  being  used  for  each 
test. 
No.  I  the  alcohol  was  driven  off  on  a  steam  bath,  the  residue 
being  removed  from  the  bath  before  the  last  traces  of  alcohol  had 
disappeared,  then  dried  in  a  vacuum  desiccator. 
No.  2  was  subjected  to  a  blast  of  warm  air  until  the  alcohol  had 
been  dissipated,  then  dried  in  a  vacuum  desiccator. 
No.  3  was  placed  in  a  vacuum  desiccator  and  the  vacuum  main- 
tained during  60  hours. 
The  residues  weighed : 
No.  I  A  =  1.030  (b)  =  1.029  Gm. 
No.  2  A=  1.0135  (b)  =  T.0115  Gm. 
No.  3    A=  1.027        (b)  =  1.032  Gm. 
On  subjecting  some  of  these  residues  to  a  moderate  heat,  as  on 
top  of  a  steam-bath,  above  and  removed  from  a  steam-bath,  on  warm 
sand,  etc.,  they  all  lost  weight  with  varying  rapidity,  while  residues 
which  were  maintained  in  a  vacuum  in  the  desiccator  and  weighed 
daily,  lost  only  1.5  and  1.6'  milligrams  in  6  days.  It  was  further 
learned  that  when  in  evaporating  the  alcohol  by  aid  of  a  very  mod- 
erate heat — (40°  to  50°  C.)  if  the  residue  was  left  in  the  heat  after 
the  alcohol  had  disappeared,  the  results  were  lower  and  were  uneven. 
It  appears  therefore  that  nitroglycerin  is  not  volatili/:ed  even  in 
a  vacunm  in  ordinary  temperature  but  that  dry  nitroglycerin  is 
slowly  deco})iposed  by  a  very  moderate  temperature.  This  sample  of 
solution  had  a  density  of  0.8650  at  25°  C.  and  tested  by  the  Kjeldahl- 
