AmjS?ri«5.arm-}  Benzin.  365 
The  existence  of  such  a  deviation  must  surely  have  been  sus- 
pected by  every  pharmacist  who  has  noticed  the  odor  of  the  benzin 
usually  handled  by  the  trade. 
In  order  to  determine  the  extent  of  this  variation,  the  writer 
examined  samples  of  benzin  which  were  purchased  at  eighteen 
retail  pharmacies  and  at  five  paint  stores,  all  of  which  were  located 
in  Philadelphia. 
All  of  the  samples  were  transparent  and  colorless. 
The  odors  were  noted  from  equal  volumes  of  the  samples  con- 
tained in  vessels  of  the  same  capacity.  When  examining  the  table 
of  results,  the  reader  should  understand  the  word  "  normal,"  when 
applied  as  the  description  of  odor,  to  mean  a  freedom  from  the 
odor  of  petroleum. 
The  samples  were  all  neutral  to  litmus  paper,  and  water  agitated 
with  them  remained  neutral  to  the  same  substance. 
While  ascertaining  this  last  fact,  peculiar  behaviors  of  some  of 
the  samples,  when  agitated  with  water,  were  noticed. 
When  shaken  in  graduated  cylinders  with  equal  volumes  of  water 
and  then  permitted  to  rest,  some  of  the  samples  demonstrated  their 
conformity  to  the  official  requirement  of  insolubility  in  water,  for 
the  two  layers  that  separated  were  equal  in  volume.  But  when 
other  samples  were  treated  with  water  in  this  manner  an  increase 
in  the  volume  of  the  lower,  or  water,  layer  was  noticed  ;  and  when 
still  other  samples  were  subjected  to  this  treatment  the  upper,  or 
benzin,  layer  was  found  to  have  been  increased. 
These  observations  were  first  made  upon  volumes  of  10  c.c.  each 
of  water  and  sample.  For  the  purpose  of  observing  these  phe- 
nomena from  larger  quantities,  a  line  of  experiments  was  instituted 
in  which  50  c.c.  of  each  liquid  were  employed.  These  experiments 
were  attended  with  interesting  results,  for  in  the  cases  of  some 
samples  a  strange  reversion  of  the  solubility  occurred.  When  the 
experiments  were  repeated  to  insure  certainty,  the  same  behaviors 
were  again  observed. 
The  specific  gravities  were  taken  at  the  temperatures  of  the 
samples  with  a  Westphal  balance.  As  the  densities  thus  indicated 
were  in  almost  all  instances  greater  than  -675,  it  was  considered 
unnecessary  to  determine  them  at  150  C. 
The  boiling  point  of  each  sample  was  found  by  distilling  a  conve- 
nient quantity  (22  c.c.  of  each  lot)  from  a  fractioning  bulb.  A  Centi- 
grade thermometer  was  inserted  almost  to  the  bottom  of  the  bulb. 
