﻿$±G 
  Prof. 
  R. 
  W. 
  Wood 
  on 
  the 
  

  

  the 
  return 
  to 
  the 
  atoms 
  of 
  electrons 
  expelled 
  by 
  the 
  action 
  

   of 
  the 
  light-waves, 
  for 
  there 
  is 
  no 
  trace 
  of 
  polarization 
  of 
  

   the 
  light. 
  The 
  fluorescence 
  of 
  iodine, 
  sodium, 
  and 
  potassium 
  

   vapours 
  is 
  strongly 
  polarized, 
  however, 
  as 
  I 
  have 
  shown 
  in 
  

   previous 
  papers, 
  and 
  the 
  polarization 
  is 
  for 
  the 
  light 
  of 
  the 
  

   complete 
  resonance 
  spectrum 
  (i. 
  e. 
  not 
  confined 
  to 
  the 
  R.R. 
  

   line). 
  This 
  makes 
  it 
  seem 
  probable 
  that 
  the 
  fluorescence 
  

   results 
  directly 
  from 
  disturbances 
  set 
  up 
  in 
  the 
  atom, 
  and 
  

   not 
  from 
  the 
  falling 
  back 
  of 
  electrons. 
  On 
  this 
  hypothesis 
  

   we 
  should 
  expect 
  phosphorescence 
  to 
  be 
  shown 
  only 
  by 
  

   mercury 
  vapour, 
  for 
  it 
  is 
  inconceivable 
  that 
  vibrations 
  set 
  up 
  

   in 
  the 
  electron 
  system 
  of 
  an 
  atom 
  could 
  persist 
  long 
  enough 
  

   to 
  be 
  detected. 
  If 
  we 
  have, 
  however, 
  something 
  analogous 
  

   to 
  dissociation 
  and 
  recombination, 
  it 
  is 
  clear 
  that 
  phosphor- 
  

   escence 
  may 
  be 
  apparent 
  if 
  only 
  the 
  latter 
  process 
  is 
  

   sufficiently 
  delayed. 
  

  

  LXXI. 
  The 
  Polarization 
  of 
  the 
  Light 
  of 
  Resonance 
  Spectra. 
  

   By 
  R. 
  W. 
  Wood, 
  Professor 
  of 
  Experimental 
  Physics, 
  

   Johns 
  Hopkins 
  University 
  *. 
  

  

  [Plate 
  XVII. 
  j 
  

  

  ONE 
  of 
  the 
  most 
  remarkable 
  and 
  important 
  phenomena 
  

   observed 
  in 
  connexion 
  with 
  resonance 
  spectra 
  is 
  the 
  

   strong 
  polarization 
  of 
  the 
  light 
  even 
  when 
  the 
  exciting 
  light 
  

   is 
  unpolarized. 
  In 
  an 
  earlier 
  paper 
  (Phil. 
  Mag. 
  Oct. 
  1911) 
  

   I 
  drew 
  attention 
  to 
  this 
  fact, 
  and 
  published 
  a 
  rough 
  drawing 
  

   of 
  the 
  Savart 
  fringes 
  as 
  they 
  appeared 
  cutting 
  across 
  the 
  

   spectrum-lines 
  at 
  right 
  angles. 
  During 
  the 
  past 
  year 
  I 
  have 
  

   repeated 
  the 
  observations 
  with 
  improved 
  apparatus, 
  and 
  am 
  

   now 
  able 
  to 
  publish 
  photographs 
  which 
  record 
  the 
  pheno- 
  

   menon. 
  En 
  the 
  more 
  recent 
  work 
  I 
  have 
  made 
  use 
  of 
  a 
  

   Presnel 
  quartz 
  double 
  prism 
  of 
  right- 
  and 
  left-handed 
  rotatory 
  

   power 
  (such 
  as 
  is 
  used 
  in 
  the 
  study 
  of 
  the 
  magnetic 
  rotation 
  

   of 
  absorbing 
  vapours). 
  This 
  was 
  found 
  to 
  give 
  fringes 
  

   somewhat 
  more 
  distinct 
  than 
  the 
  Savart 
  plate 
  which 
  was 
  at 
  

   my 
  disposal. 
  

  

  The 
  light 
  from 
  a 
  large 
  quartz 
  mercury 
  arc 
  running 
  on 
  a 
  

   220 
  volt 
  circuit, 
  with 
  the 
  minimum 
  resistance 
  allowable, 
  

   was 
  focnssed 
  along 
  the 
  axis 
  of 
  one 
  of 
  the 
  iodine 
  vacuum 
  

   tubes 
  described 
  in 
  the 
  previous 
  paper. 
  The 
  light 
  was 
  first 
  

   filtered 
  through 
  a 
  strong 
  solution 
  of 
  neodymium 
  which 
  

   removed 
  every 
  trace 
  of 
  the 
  yellow 
  Hg 
  lines, 
  so 
  that 
  the 
  

  

  * 
  Communicated 
  by 
  the 
  Author. 
  

  

  