﻿332 
  Mr. 
  T. 
  Preston 
  on 
  Radiation 
  Phenomena 
  

  

  own 
  constitution, 
  and 
  an 
  important 
  investigation 
  bearing 
  on 
  

   this 
  subject 
  has 
  already 
  been 
  made 
  by 
  Prof. 
  A. 
  A. 
  Michelson*, 
  

   who 
  has 
  shown 
  that 
  the 
  D-lines 
  of 
  sodium 
  and 
  the 
  green 
  and 
  

   blue 
  lines 
  of 
  cadmium 
  (with 
  others) 
  are 
  really 
  complex 
  in 
  

   constitution. 
  Now 
  these 
  lines 
  are 
  amonost 
  those 
  which 
  show 
  

   deviations 
  from 
  the 
  normal 
  triplet 
  type, 
  and 
  it 
  is 
  consequently 
  

   of 
  importance 
  to 
  determine 
  if 
  this 
  complexity 
  of 
  structure 
  

   gives 
  rise 
  to 
  the 
  deviations 
  from 
  the 
  normal 
  triplet, 
  or 
  if, 
  on 
  

   the 
  other 
  hand, 
  the 
  deviations 
  as 
  w 
  T 
  ell 
  as 
  the 
  complexity 
  of 
  

   structure 
  observed 
  by 
  Michelson 
  are 
  due 
  to 
  a 
  common 
  cause, 
  

   namely 
  intermittent 
  reversal. 
  For 
  example, 
  if 
  a 
  bright 
  line 
  is 
  

   not 
  simple 
  but 
  really 
  a 
  close 
  doublet, 
  then 
  in 
  the 
  magnetic 
  field 
  

   each 
  member 
  of 
  this 
  doublet 
  will 
  give 
  rise 
  to 
  a 
  corresponding 
  

   triplet. 
  These 
  triplets, 
  however, 
  will 
  overlap 
  and 
  appear 
  as 
  

   one 
  triplet 
  in 
  the 
  field 
  of 
  view. 
  Yet 
  it 
  is 
  possible 
  to 
  conceive 
  

   a 
  state 
  of 
  affairs 
  such 
  that, 
  by 
  the 
  action 
  of 
  the 
  magnetic 
  field, 
  

   one 
  or 
  more 
  of 
  the 
  members 
  of 
  one 
  triplet 
  shall 
  be 
  opposite 
  in 
  

   phase 
  to 
  the 
  corresponding 
  member 
  of 
  the 
  other 
  triplet. 
  

   Under 
  such 
  circumstances 
  the 
  overlapping 
  lines 
  will 
  " 
  inter- 
  

   fere 
  " 
  and 
  produce 
  doublets. 
  Thus 
  if 
  the 
  central 
  lines 
  alone 
  

   are 
  opposite 
  in 
  phase 
  they 
  will 
  obliterate 
  each 
  other 
  where 
  

   they 
  overlap, 
  and 
  the 
  weak-middled 
  quartet 
  will 
  be 
  produced 
  

   as 
  in 
  the 
  case 
  of 
  the 
  blue 
  cadmium 
  line, 
  and 
  so 
  on. 
  

  

  It 
  is 
  to 
  be 
  remembered, 
  however, 
  that 
  the 
  conditions 
  in 
  the 
  

   magnetic 
  field 
  are 
  not 
  the 
  same 
  as 
  in 
  the 
  free 
  field, 
  for 
  in 
  the 
  

   former 
  the 
  spark 
  is 
  violently 
  blown 
  about, 
  and 
  this 
  may 
  lead 
  

   to 
  a 
  sharpness 
  on 
  both 
  sides 
  of 
  the 
  reversed 
  lines 
  not 
  obtain- 
  

   able 
  in 
  the 
  free 
  field, 
  and 
  thus 
  account 
  for 
  the 
  fact 
  that 
  the 
  

   modifications 
  described 
  above 
  do 
  not 
  look 
  like 
  ordinary 
  

   reversals. 
  

  

  Viewed 
  from 
  the 
  theoretical 
  standpoint, 
  however, 
  we 
  have 
  

   no 
  reason 
  to 
  demand 
  that 
  all, 
  or 
  indeed 
  necessarily 
  any, 
  of 
  

   the 
  spectral 
  lines 
  should 
  be 
  resolved 
  into 
  sharp 
  triplets 
  when 
  

   the 
  source 
  of 
  light 
  is 
  viewed 
  across 
  the 
  lines 
  of 
  magnetic 
  force. 
  

   For, 
  in 
  order 
  that 
  a 
  spectral 
  line 
  should 
  exhibit 
  itself 
  as 
  a 
  

   characteristic 
  triplet 
  under 
  the 
  influence 
  of 
  the 
  magnetic 
  field, 
  

   it 
  is 
  necessary 
  that 
  the 
  freedom 
  of 
  vibration 
  should 
  be 
  equal 
  

   in 
  all 
  directions, 
  and 
  in 
  this 
  case 
  the 
  intensity 
  of 
  each 
  rect- 
  

   angular 
  component 
  will 
  be 
  the 
  same. 
  Hence 
  the 
  middle 
  line 
  

   of 
  the 
  triplet 
  will 
  contain 
  as 
  much 
  light 
  as 
  the 
  two 
  side 
  lines 
  

   taken 
  together. 
  If, 
  however, 
  the 
  vibrations 
  are 
  not 
  equally 
  

   free 
  in 
  all 
  directions 
  the 
  foregoing 
  result 
  will 
  not 
  hold, 
  and 
  it 
  

   becomes 
  possible 
  to 
  have 
  a 
  triplet 
  with 
  a 
  weak 
  middle 
  line 
  or 
  

  

  * 
  A. 
  A. 
  Michelson, 
  'Brit. 
  Assoc. 
  Report,' 
  1892, 
  and 
  Phil. 
  Mag. 
  

   vol. 
  xxxiv. 
  p. 
  280 
  (18^2). 
  I 
  have 
  to 
  thank 
  Prof. 
  G. 
  F. 
  FitzGerald 
  for 
  

   attracting- 
  my 
  attention 
  to 
  this 
  point. 
  

  

  