﻿Energy 
  of 
  the 
  Ions 
  emitted 
  by 
  Hot 
  Bodies. 
  897 
  

  

  This 
  was 
  shown 
  by 
  the 
  fact 
  that 
  the 
  ratio 
  of 
  the 
  current 
  

   through 
  the 
  slit 
  to 
  the 
  total 
  current 
  was 
  never 
  quite 
  sym- 
  

   metrical 
  about 
  the 
  centre 
  of 
  the 
  pattern. 
  This 
  lack 
  of 
  

   symmetry 
  was 
  of 
  little 
  importance 
  in 
  the 
  experiments 
  on 
  

   e/m 
  where 
  the 
  deflexion 
  of 
  the 
  whole 
  pattern 
  in 
  a 
  magnetic 
  

   field 
  was 
  measured, 
  but 
  it 
  is 
  of 
  vital 
  importance 
  in 
  the 
  present 
  

   experiments. 
  If 
  the 
  strip 
  is 
  bent 
  the 
  applied 
  electric 
  field 
  

   alone 
  will 
  cause 
  a 
  spreading 
  of 
  the 
  ions 
  which 
  may 
  easily 
  be 
  

   greater 
  than 
  that 
  due 
  to 
  their 
  original 
  kinetic 
  energy. 
  The 
  

   author 
  now 
  believes 
  that 
  he 
  has 
  obtained 
  a 
  better 
  method 
  of 
  

   cutting 
  the 
  strips 
  which 
  will 
  avoid 
  these 
  difficulties 
  in 
  future 
  

   experiments. 
  

  

  There 
  were 
  other 
  errors 
  arising 
  from 
  the 
  finite 
  widths 
  of 
  

   the 
  strip 
  and 
  the 
  slit, 
  from 
  the 
  finite 
  length 
  of 
  the 
  strip, 
  and 
  

   from 
  the 
  scattering 
  of 
  the 
  ions 
  by 
  the 
  small 
  quantity 
  of 
  gas 
  

   present 
  in 
  the 
  apparatus. 
  It 
  is 
  difficult 
  to 
  see 
  how 
  this 
  could 
  

   be 
  reduced 
  much 
  below 
  the 
  value 
  (about 
  3 
  x 
  10 
  ~ 
  4 
  mm.) 
  it 
  

   had 
  in 
  most 
  of 
  the 
  experiments, 
  with 
  apparatus 
  of 
  this 
  degree 
  

   of 
  complexity. 
  This 
  difficulty 
  could 
  be 
  allowed 
  for 
  to 
  some 
  

   extent 
  by 
  making 
  observations 
  at 
  different 
  pressures. 
  A 
  

   rough 
  attempt 
  to 
  correct 
  some 
  of 
  the 
  observations 
  for 
  the 
  

   effect 
  of 
  the 
  gas 
  was 
  made 
  in 
  this 
  way. 
  

  

  Originally 
  it 
  was 
  intended 
  to 
  determine 
  the 
  temperature 
  

   of 
  the 
  strip 
  from 
  its 
  resistance 
  by 
  the 
  usual 
  methods, 
  but 
  

   this 
  part 
  of 
  the 
  plan 
  was 
  not 
  successful 
  as 
  the 
  strip 
  broke 
  

   before 
  the 
  necessary 
  observations 
  could 
  be 
  completed. 
  The 
  

   temperature 
  was 
  therefore 
  estimated 
  from 
  the 
  magnitude 
  of 
  

   the 
  ionization 
  currents 
  and 
  the 
  dimensions 
  of 
  the 
  strip. 
  It 
  

   is 
  subject 
  to 
  a 
  possible 
  error 
  of 
  about 
  10 
  per 
  cent. 
  (100° 
  abs.), 
  

   but 
  the 
  uncertainties 
  arising 
  out 
  of 
  the 
  structure 
  of 
  the 
  

   apparatus 
  do 
  not 
  warrant 
  our 
  proceeding 
  to 
  a 
  higher 
  order 
  

   of 
  accuracy 
  than 
  this 
  in 
  the 
  determination 
  of 
  the 
  temperature. 
  

  

  The 
  experimental 
  points 
  from 
  which 
  curves 
  like 
  that 
  shown 
  

   in 
  fig. 
  2 
  were 
  constructed 
  represent 
  the 
  values 
  of 
  the 
  electro- 
  

   meter 
  reading 
  for 
  the 
  current 
  through 
  the 
  slit, 
  divided 
  by 
  

   that 
  for 
  the 
  slit 
  and 
  the 
  plates 
  together. 
  These 
  measure- 
  

   ments 
  were 
  taken 
  precisely 
  as 
  described 
  in 
  the 
  author's 
  

   previous 
  paper 
  *. 
  To 
  obtain 
  the 
  value 
  of 
  the 
  quantity 
  denoted 
  

   by 
  i 
  in 
  the 
  preceding 
  equations 
  it 
  is 
  necessary 
  to 
  divide 
  each 
  

   ordinate 
  by 
  the 
  ratio 
  of 
  the 
  sum 
  of 
  the 
  capacities 
  of 
  the 
  

   electrometer 
  and 
  the 
  condenser 
  to 
  the 
  capacity 
  of 
  the 
  electro- 
  

   meter 
  alone. 
  This 
  ratio 
  was 
  not 
  determined 
  accurately 
  but 
  

   it 
  was 
  known 
  to 
  be 
  not 
  far 
  from 
  12. 
  The 
  relative 
  values 
  of 
  

   the 
  different 
  ordinates 
  do 
  not, 
  of 
  course, 
  depend 
  on 
  the 
  

  

  * 
  Phil. 
  Mag. 
  [6] 
  vol. 
  xvi. 
  p. 
  740 
  (Nov. 
  1908). 
  

   Phil. 
  Mag. 
  B. 
  6. 
  Vol. 
  16. 
  No. 
  96. 
  Dec. 
  1908. 
  3 
  N 
  

  

  