Pressures  of  Alcoholic  Solutions.  599 
marked  for  dilute  solutions*.  On  this  theory  a  curve, 
obtained  as  the  above  one,  should  cut  the  #-axis  at  that  point 
which  would  denote  the  strength  of  the  solution  whose 
attraction  for  the  solvent  would  just  be  balanced  by  the 
tendency  of  the  membrane  to  retain  it. 
A  further  remark  may  be  made  here  about  the  change  in 
form  of  the  curve  to  which  theory  would  point.  Since  the 
attraction  of  the  membrane  plays  a  more  important  part  in 
cutting  down  the  osmotic  pressure  of  the  weaker  solutions, 
the  curve  (plotted  as  above)  should  be  convex  towards  the 
y-axis.  The  above  curve  may  therefore  be  raised  in  its  lower 
portion  by  disturbing  influences.  Even  the  assumption  of 
considerable  dissociation  at  this  dilution  would  not  account 
for  the  slow  fall  of  the  curve. 
These  considerations  on  the  above  curve  showed  the 
necessity  for  a  more  careful  examination  of  the  osmotic 
pressures  of  such  solutions.     This  was  undertaken  as  follows  : 
"  Absolute  "  alcohol  was  redistilled  after  having  stood  over 
fresh  and  very  good  quicklime  for  three  weeks  and  then  over 
dehydrated  copper  sulphate  for  another  week.  It  was  dis- 
tilled on  a  water-bath  directly  from  the  latter.  The  first  and 
last  portions  of  the  liquid  were  neglected  ;  the  receiving- 
flask  was  only  open  to  the  air  through  a  calcium-chloride 
tube.  The  redistilled  alcohol  was  kept  in  a  flask  under  a 
desiccator  or  with  the  neck  of  the  flask  covered  tightly  with 
a  guttapercha  cap.  Every  precaution  was  taken  to  avoid 
absorption  of  moisture,  all  measurements,  &c.  being  made 
as  quickly  as  possible  where  it  was  necessary  to  expose  the 
alcohol  to  the  air.  The  specific  gravity  of  the  alcohol  after 
distilling  was  *7973  at  12°  C,  compared  with  water  at  the 
same  temperature.  That  of  the  alcohol  before  distillation 
was  '7980  at  the  same  temperature.  This  shows  that  the 
alcohol,  as  supplied  to  the  laboratory,  was  really  very  good. 
I  point  this  out  here,  because  in  the  last  of  the  next  series 
of  experiments  I  was  compelled  to  use  "  absolute  alcohol/' 
owing  to  having  run  short  unexpectedly  of  the  redistilled 
liquid. 
The  lithium  chloride  used  was  the  salt  supplied  to  the 
laboratory  as  "  pure  ;  "  the  salt  used  in  these  experiments 
was  kept  for  some  weeks  previously  in  a  weighing  tube  over 
phosphorus  pentoxide  in  a  small  desiccator.  Care  was  taken 
to  expose  the  salt  as  little  as  possible. 
Arrangements  were  made  for  carrying  out  the  experiments 
*  Prof.  Kahlenberg,  by  simultaneous  work  not  yet  published,  lias 
arrived  at  the  same  result  (Phil.  Mag.  Feb.  1905,  p.  2'2S). 
