﻿Distribution 
  of 
  Energy 
  in 
  the 
  Spectra 
  of 
  Gases. 
  823 
  

  

  increase 
  of 
  the 
  variable 
  with 
  increasing 
  pressure 
  points 
  to 
  

   some 
  regularity 
  and 
  leaves 
  little 
  doubt 
  as 
  to 
  the 
  existence 
  of 
  

   the 
  phenomenon. 
  

  

  It 
  is 
  not 
  to 
  be 
  expected 
  that 
  the 
  ratio 
  of 
  H 
  a 
  to 
  the 
  whole 
  

   energy 
  would 
  vary 
  in 
  exactly 
  the 
  same 
  way 
  as 
  the 
  ratio 
  of 
  

   H 
  a 
  to 
  Hp, 
  but 
  H 
  rt 
  being 
  by 
  far 
  the 
  most 
  prominent 
  feature 
  

   of 
  the 
  spectrum 
  we 
  should 
  expect 
  to 
  find 
  the 
  variation 
  in 
  the 
  

   same 
  sense 
  in 
  both 
  cases. 
  

  

  The 
  same 
  phenomenon 
  — 
  i. 
  e. 
  shift 
  of 
  energy 
  towards 
  longer 
  

   wave-lengths 
  with 
  increased 
  excitation 
  — 
  has 
  been 
  observed 
  

   by 
  Vaillant 
  (Comptes 
  Rendus, 
  1906, 
  p. 
  81) 
  in 
  the 
  case 
  of 
  the 
  

   mercury 
  arc. 
  In 
  some 
  respects 
  the 
  discharge 
  from 
  a 
  large 
  

   condenser 
  at 
  comparatively 
  high 
  pressures 
  resembles 
  the 
  

   mercury 
  arc. 
  

  

  0. 
  The 
  Continuous 
  Distribution 
  of 
  Energy 
  in 
  the 
  Spectrum. 
  

  

  This 
  was 
  investigated 
  by 
  taking 
  a 
  long 
  series 
  of 
  readings 
  

   at 
  appropriate 
  intervals 
  of 
  time. 
  Curves 
  9, 
  10, 
  11, 
  12 
  (A 
  

   and 
  B) 
  (PL 
  XIV.) 
  contain 
  observations 
  on 
  the 
  spectrum 
  of 
  

   hydrogen 
  at 
  three 
  different 
  pressures, 
  and 
  on 
  that 
  of 
  air. 
  

  

  Table 
  IV., 
  representing 
  Curve 
  10, 
  is 
  given 
  as 
  an 
  example. 
  

   Curve 
  9 
  was 
  really 
  only 
  obtained 
  as 
  a 
  test, 
  and 
  as 
  only 
  one 
  

   reading 
  was 
  taken 
  for 
  each 
  point 
  the 
  values 
  are 
  but 
  very 
  

   approximate. 
  The 
  general 
  form 
  of 
  the 
  curve 
  is 
  apparent, 
  

   but 
  even 
  an 
  inspection 
  of 
  the 
  spectrum 
  with 
  an 
  eyepiece 
  

   shows 
  that 
  the 
  peak 
  representing 
  H^ 
  is 
  too 
  sharp. 
  A 
  

   perfectly 
  monochromatic 
  line 
  gives 
  two 
  intersecting 
  straight 
  

   lines, 
  such 
  as 
  can 
  be 
  seen 
  to 
  a 
  fair 
  degree 
  of 
  perfection 
  in 
  

   the 
  case 
  of 
  H 
  a 
  , 
  even 
  at 
  higher 
  pressures. 
  

  

  In 
  order 
  to 
  obtain 
  enough 
  energy 
  a 
  rather 
  wide 
  slit 
  has 
  to 
  

   be 
  employed, 
  giving 
  a 
  considerable 
  breadth 
  to 
  the 
  trace 
  of 
  a 
  

   monochromatic 
  line 
  on 
  the 
  curve. 
  The 
  readings 
  for 
  Curves 
  

   10 
  and 
  11 
  were 
  made 
  more 
  systematically 
  than 
  those 
  for 
  

   Curve 
  9. 
  They 
  were 
  always 
  taken 
  in 
  pairs 
  belonging 
  to 
  the 
  

   same 
  spectroscope 
  setting, 
  thus 
  affording 
  some 
  sort 
  of 
  check. 
  

   Further, 
  a 
  single 
  progression 
  from 
  one 
  end 
  of 
  the 
  spectrum 
  

   to 
  the 
  other 
  was 
  not 
  allowed 
  to 
  suffice, 
  and 
  a 
  return 
  was 
  

   made 
  to 
  some 
  of 
  the 
  more 
  important 
  points, 
  thus 
  avoi 
  ding- 
  

   any 
  errors 
  due 
  to 
  the 
  possible 
  slow 
  change 
  in 
  the 
  conditions 
  

   of 
  the 
  discharge. 
  The 
  reading 
  marked 
  " 
  (a) 
  " 
  in 
  Table 
  IV. 
  

   is 
  a 
  test 
  for 
  disturbing 
  effects 
  with 
  shutter 
  closed. 
  Both 
  of 
  

   these 
  curves 
  show 
  a 
  continuous 
  spectrum 
  of 
  some 
  intensity, 
  

   and 
  it 
  appears 
  that 
  a 
  smooth 
  curve 
  could 
  be 
  drawn 
  along 
  the 
  

   groundwork 
  of 
  the 
  spectrum, 
  thus 
  suggesting 
  strongly 
  the 
  

   continuous 
  spectrum 
  given 
  by 
  an 
  ordinary 
  heated 
  body. 
  

  

  