﻿of 
  Pressure 
  which 
  accompany 
  Point 
  Discharge. 
  41 
  

  

  clearly 
  tend 
  towards 
  the 
  particular 
  percentage 
  (between 
  

   01 
  and 
  0*2) 
  which 
  brings 
  the 
  peak 
  to 
  the 
  right 
  place 
  to 
  

   correspond 
  with 
  the 
  oxygen 
  peak. 
  

  

  In 
  Curves 
  VI. 
  (circles) 
  are 
  given 
  the 
  results 
  of 
  a 
  second 
  

   nitrogen 
  series, 
  taken 
  directly 
  after 
  the 
  first. 
  For 
  this 
  the 
  

   apparatus 
  was 
  alternately 
  pumped 
  out 
  and 
  filled 
  with 
  " 
  pure 
  " 
  

   hydrogen 
  until 
  the 
  nitrogen 
  left 
  from 
  the 
  first 
  series 
  was 
  

   reduced 
  several 
  thousand 
  times, 
  and 
  the 
  oxygen 
  in 
  it 
  probably 
  

   more 
  than 
  a 
  million 
  times. 
  The 
  negative 
  curve 
  now 
  shows 
  

   signs 
  of 
  the 
  fall 
  in 
  the 
  height 
  of 
  the 
  peak 
  which 
  we 
  mentioned 
  

   above, 
  and 
  it 
  is 
  impossible 
  to 
  say 
  where 
  the 
  peak 
  ends. 
  We 
  

   therefore 
  cannot 
  use 
  the 
  position 
  of 
  the 
  peak 
  to 
  find 
  the 
  per- 
  

   centage 
  of 
  oxygen 
  present 
  as 
  before. 
  But 
  in 
  Table 
  4 
  is 
  

   given 
  a 
  set 
  of 
  percentages 
  worked 
  out 
  similarly 
  to 
  those 
  in 
  

   Table 
  3. 
  

  

  Table 
  4. 
  

  

  Percentage 
  of 
  

   Ni 
  trogen. 
  

  

  Pei'centage 
  of 
  Oxygen 
  

   in 
  the 
  Nitrogen. 
  

  

  0-006 
  

  

  ? 
  

  

  •01G 
  

  

  ? 
  

  

  •065 
  

  

  •is 
  

  

  •125 
  

  

  •12 
  

  

  •62 
  

  

  •12 
  

  

  105 
  

  

  ■14 
  

  

  Here 
  the 
  first 
  two 
  nitrogen 
  percentages, 
  though 
  larger 
  

   than 
  those 
  in 
  Table 
  3, 
  gave 
  values 
  of 
  p 
  which 
  were 
  too 
  small 
  

   to 
  rely 
  on, 
  and 
  the 
  0*18 
  of 
  the 
  third 
  is 
  also 
  rather 
  uncertain 
  

   for 
  the 
  same 
  reason. 
  There 
  can, 
  however,- 
  be 
  little 
  question 
  

   that 
  we 
  are 
  here 
  dealing 
  with 
  quantities 
  of 
  impurity 
  which 
  

   bear 
  a 
  constant 
  proportion 
  to 
  the 
  nitrogen 
  present 
  with 
  them, 
  

   and 
  which 
  were, 
  therefore, 
  doubtless 
  to 
  be 
  found 
  in 
  that 
  

   nitrogen 
  before 
  it 
  entered 
  the 
  discharge 
  vessel. 
  

  

  The 
  initial 
  charge 
  of 
  oxygen, 
  indicated 
  by 
  Table 
  3, 
  is 
  no 
  

   longer 
  present, 
  but 
  the 
  fact 
  that 
  the 
  numbers 
  in 
  Table 
  4 
  

   correspond 
  to 
  the 
  limiting 
  value 
  of 
  those 
  in 
  Table 
  3, 
  and 
  that 
  

   this 
  limiting 
  value 
  brings 
  the 
  first 
  peak 
  to 
  the 
  point 
  which 
  

   corresponds 
  with 
  oxygen, 
  renders 
  it 
  extremely 
  probable 
  that 
  

   in 
  Table 
  4 
  we 
  are 
  also 
  dealing 
  with 
  oxygen, 
  introduced 
  in 
  

   the 
  nitrogen. 
  

  

  