AfliuS,S1Ji-}     Valuation  of  Crude  Carbolic  Acid.  371 
be  made  serviceable  here,  with  proper  precaution,  to  show  the  pro- 
portions of  the  several  phenols  by  their  differences  in  bromine 
absorption.  To  determine  with  accuracy  by  this  method  the  pro- 
portion of  phenol  and  cresols  in  mixtures,  the  following  conditions 
must  be  fulfilled  :  The  mixture  must  be  either  free  from  all  sub- 
stances other  than  phenol  and  the  cresols,  or  they  must  be  present 
in  known  quantities  and  their  weight  deducted  from  the  weight  of 
the  sample  taken.  The  mixture  must  contain  no  other  substances 
that  bind  bromine,  or  it  must  contain  them  in  known  quantity,  in 
which  case  the  amount  of  bromine  they  will  take  up  must  be  known. 
The  contaminating  substances  must  not  in  any  other  way  interfere 
with  the  reaction. 
To  make  the  method  applicable  to  crude  carbolic  acid,  ready 
means  for  determining  the  non-phenolic  bodies  present  would  be 
required,  and  it  appeared  from  an  examination  of  the  literature  on 
the  subject,  that  simple  and  satisfactory  methods  for  this  purpose 
were  available,  but  experience  with  them  soon  showed  them  to  be 
wanting  in  accuracy. 
For  the  estimation  of  wTater,  it  is  usually  directed  to  shake  a 
known  volume  of  the  acid  with  a  saturated  solution  of  sodium 
chloride,  on  the  assumption  that  the  diminution  in  the  volume  of 
the  acid,  after  agitation  and  perfect  separation,  is  equivalent  to  the 
volume  of  water  present.  This  the  writer  found  far  from  being  the 
case.  A  sample  of  acid,  consisting  of  about  equal  parts  of  phenol 
and  the  cresols,  and  conforming  to  the  pharmacopceial  require- 
ments, was,  on  distillation,  found  to  contain  about  0-5  per  cent,  of 
water.  The  same  sample,  on  being  shaken  with  an  equal  volume 
of  the  salt  solution,  should  have  decreased  slightly  in  volume,  but 
instead,  it  was  found  to  increase  9-5  per  cent.,  with  a  simultaneous 
separation  of  sodium  chloride  crystals.  When  sufficient  water  was 
added  to  the  sample  to  make  the  mixture  contain  5  per  cent., 
and  tested  in  the  same  manner,  it  increased  4  per  cent,  in  volume. 
With  10  per  cent,  of  water  it  decreased  6  5  per  cent.,  showing 
two-thirds  of  the  water  present.  With  19  5  per  cent.,  a  saturated 
solution  of  water  in  the  acid  at  about  250  C,  15-5  per  cent,  was 
indicated.  These  results  show  the  test  to  be  entirely  worthless 
for  detecting  or  estimating  moderate  quantities  of  water,  at  least 
in  samples  containing  much  phenol.  But  the  error  is  lessened 
as  the  proportion  of  water  increases.    Acids,  consisting  mainly  of 
