﻿Distribution 
  of 
  Energy 
  in 
  the 
  Spectra 
  of 
  Gases. 
  825 
  

  

  Daring 
  the 
  observations 
  represented 
  by 
  Curve 
  10 
  some 
  water 
  

   was 
  noticed 
  in 
  the 
  discharge-tube. 
  It 
  appears 
  that 
  the 
  

   influence 
  of 
  the 
  discharge 
  on 
  the 
  brass 
  walls 
  o£ 
  the 
  hollow 
  

   electrodes 
  is 
  to 
  make 
  them 
  become 
  porous 
  in 
  time. 
  The 
  

   effect 
  of 
  the 
  water 
  seems 
  only 
  to 
  have 
  been 
  to 
  increase 
  some- 
  

   what 
  the 
  intensity 
  of 
  the 
  continuous 
  spectrum 
  underlying 
  

   the 
  peaked 
  curve. 
  Comparing 
  this 
  curve 
  (10) 
  with 
  (11) 
  the 
  

   interesting 
  fact 
  comes 
  out 
  that 
  there 
  is 
  a 
  very 
  marked 
  shift 
  

   of 
  the 
  energy 
  towards 
  the 
  longer 
  wave-lengths. 
  Not 
  only 
  

   is 
  the 
  ratio 
  of 
  H^ 
  to 
  H 
  a 
  much 
  smaller 
  at 
  the 
  higher 
  pressure, 
  

   but 
  the 
  still 
  shorter 
  wave-lengths 
  have 
  practically 
  disap- 
  

   peared. 
  Curve 
  10 
  seems 
  to 
  suggest 
  that 
  the 
  peaks 
  could 
  

   be 
  joined 
  by 
  a 
  smooth 
  curve, 
  probably 
  an 
  exponential 
  one. 
  

   But 
  the 
  values 
  are 
  scarcely 
  accurate 
  enough 
  to 
  fall 
  in 
  with 
  

   such 
  a 
  stringent 
  test, 
  It 
  is 
  true 
  that 
  the 
  two 
  chief 
  peaks 
  in 
  

   the 
  curve 
  represent 
  the 
  means 
  of 
  many 
  readings, 
  but 
  even 
  

   then 
  they 
  are 
  liable 
  to 
  errors 
  of 
  a 
  few 
  per 
  cent. 
  

  

  The 
  curves 
  for 
  air 
  (12 
  A 
  and 
  B) 
  show 
  clearly 
  that 
  the 
  gas 
  

   was 
  contaminated 
  with 
  hydrogen, 
  doubtless 
  occluded 
  and 
  re- 
  

   emitted 
  by 
  the 
  electrodes. 
  The 
  rest 
  of 
  the 
  spectrum 
  is 
  made 
  

   up 
  of 
  lines 
  too 
  close 
  to 
  be 
  resolved 
  by 
  the 
  instrument, 
  and 
  

   exhibits 
  a 
  very 
  great 
  increase 
  in 
  intensity 
  in 
  the 
  red 
  region 
  

   when 
  the 
  pressure 
  is 
  raised. 
  This 
  increase 
  may 
  not 
  be 
  so 
  

   abnormal 
  as 
  at 
  first 
  sight 
  appears, 
  for 
  at 
  the 
  higher 
  pressure 
  

   the 
  proportional 
  amount 
  of 
  hydrogen 
  present 
  must 
  have 
  been 
  

   much 
  smaller, 
  and 
  if 
  the 
  blue 
  region 
  of 
  the 
  spectrum 
  be 
  con- 
  

   sidered, 
  the 
  proportional 
  rise 
  in 
  the 
  red 
  is 
  not 
  much 
  greater 
  

   than 
  in 
  the 
  other 
  part. 
  There 
  is, 
  of 
  course, 
  the 
  bare 
  possi- 
  

   bility 
  that 
  some 
  of 
  the 
  energy 
  in 
  the 
  red 
  end 
  of 
  the 
  spectrum 
  

   may 
  be 
  due 
  to 
  the 
  heating 
  of 
  the 
  capillary 
  portion 
  of 
  the 
  tube 
  

   to 
  redness, 
  but 
  several 
  distinct 
  pieces 
  of 
  evidence 
  can 
  be 
  

   found 
  to 
  negative 
  this 
  supposition. 
  In 
  the 
  first 
  place, 
  there 
  

   is 
  no 
  reason 
  wdiy 
  the 
  same 
  large 
  maximum 
  should 
  not 
  have 
  

   been 
  observed 
  in 
  the 
  spectra 
  of 
  hydrogen. 
  Secondly, 
  accom- 
  

   panying 
  such 
  a 
  large 
  increase 
  in 
  magnitude 
  of 
  the 
  maximum 
  

   there 
  must 
  be 
  in 
  the 
  case 
  of 
  a 
  pure 
  temperature 
  radiation 
  a 
  

   very 
  marked 
  shift 
  in 
  the 
  position 
  of 
  this 
  maximum. 
  

  

  Against 
  this 
  we 
  may 
  set 
  the 
  fact 
  that 
  the 
  opacity 
  of 
  the 
  

   glass 
  apparatus 
  may 
  have 
  been 
  responsible 
  for 
  the 
  downward- 
  

   sloping 
  portions 
  on 
  the 
  right 
  of 
  both 
  curves, 
  so 
  that 
  the 
  

   maxima 
  of 
  both 
  would 
  have 
  been 
  hidden. 
  

  

  Theoretical. 
  

  

  The 
  subject 
  of 
  the 
  radiation 
  from 
  gases 
  in 
  discharge-tubes 
  

   is 
  a 
  difficult 
  one 
  to 
  approach 
  theoretically; 
  because, 
  in 
  the 
  

  

  