﻿in 
  the 
  Magnetic 
  Field. 
  331 
  

  

  of 
  which 
  the 
  two 
  border- 
  lines 
  are 
  somewhat 
  nebulous 
  on 
  their 
  

   outside 
  edges. 
  The 
  other 
  line 
  Dj 
  is 
  of 
  the 
  double 
  doublet 
  

   type 
  (fig. 
  4), 
  and 
  shows 
  as 
  two 
  pairs 
  of 
  sharp 
  lines. 
  It 
  is 
  thus 
  

   seen 
  that 
  the 
  sodium 
  lines 
  and 
  the 
  blue 
  cadmium 
  line 
  4800 
  do 
  

   not 
  show 
  as 
  normal 
  triplets 
  ; 
  and 
  it 
  is 
  probably 
  for 
  this 
  reason 
  

   that 
  different 
  observers 
  have 
  differed 
  in 
  their 
  accounts 
  as 
  to 
  

   the 
  effects 
  observed 
  in 
  the 
  case 
  of 
  these 
  lines*. 
  Difficulties 
  

   no 
  doubt 
  encumber 
  the 
  observation 
  of 
  the 
  true 
  effect 
  in 
  the 
  

   case 
  of 
  the 
  sodium 
  lines 
  by 
  reason 
  of 
  the 
  ease 
  with 
  which 
  they 
  

   become 
  reversed. 
  If 
  the 
  sodium 
  salt 
  is 
  present 
  in 
  any 
  consi- 
  

   derable 
  quantity, 
  there 
  is 
  constant 
  sputtering 
  and 
  reversal 
  

   of 
  the 
  lines 
  which 
  completely 
  masks 
  the 
  true 
  effect. 
  

  

  Explanation 
  of 
  the 
  Various 
  Appearances. 
  

  

  All 
  the 
  modifications 
  of 
  the 
  normal 
  triplet 
  described 
  above 
  

   may 
  be 
  sufficiently 
  accounted 
  for 
  by 
  the 
  supposition 
  that 
  

   reversal, 
  or 
  more 
  or 
  less 
  complete 
  absorption, 
  takes 
  place 
  in 
  

   the 
  case 
  of 
  one 
  or 
  more 
  of 
  the 
  constituents 
  of 
  the 
  triplet. 
  

   Thus, 
  if 
  the 
  middle 
  line 
  of 
  the 
  triplet 
  (1) 
  becomes 
  reversed, 
  

   t. 
  e* 
  has 
  its 
  middle 
  absorbed 
  5 
  we 
  are 
  furnished 
  with 
  the 
  weak 
  

   middled 
  quartet 
  (2) 
  ; 
  ? 
  and 
  if 
  the 
  middle 
  line 
  becomes 
  com- 
  

   pletely 
  absorbed 
  in 
  the 
  layers 
  of 
  colder 
  vapour 
  surrounding 
  

   the 
  spark, 
  then 
  we 
  have 
  the 
  doublet 
  (3) 
  . 
  While 
  if, 
  in 
  

   addition, 
  the 
  side 
  lines 
  of 
  (3) 
  become 
  reversed, 
  we 
  have 
  the 
  

   double 
  doublet 
  (4) 
  j 
  and, 
  finally, 
  if 
  each 
  line 
  of 
  the 
  triplet 
  CI) 
  

   becomes 
  reversed, 
  we 
  have 
  the 
  sextet 
  (5). 
  

  

  Although 
  the 
  hypothesis 
  of 
  reversal 
  or 
  absorption 
  accounts 
  

   easily 
  for 
  the 
  number 
  of 
  the 
  constituents 
  in 
  the 
  modified 
  lines, 
  

   yet 
  it 
  must 
  be 
  remarked 
  that 
  the 
  appearance 
  of 
  these 
  modifi- 
  

   cations 
  is 
  not 
  that 
  which 
  is 
  usually 
  associated 
  with 
  reversal. 
  

   The 
  constituents 
  are 
  not 
  sharp 
  on 
  one 
  side 
  and 
  nebulous 
  on 
  

   the 
  other, 
  but 
  are 
  sharp 
  and 
  well 
  defined 
  on 
  both 
  sides 
  ; 
  nor 
  

   do 
  the 
  original 
  lines 
  when 
  unaffected 
  by 
  the 
  magnetic 
  field, 
  

   that 
  is 
  when 
  photographed 
  in 
  the 
  free 
  field, 
  show 
  any 
  sign 
  of 
  

   reversal. 
  For 
  this 
  reason, 
  in 
  order 
  to 
  test 
  this 
  matter, 
  I 
  have 
  

   spent 
  much 
  time 
  in 
  taking 
  photographs 
  under 
  various 
  con- 
  

   ditions 
  in 
  regard 
  to 
  length 
  of 
  exposure 
  and 
  vapour-density 
  in 
  

   spark, 
  but 
  so 
  far 
  without 
  any 
  conclusive 
  result. 
  On 
  the 
  

   whole, 
  however, 
  the 
  weight 
  of 
  evidence 
  lends 
  itself 
  to 
  the 
  

   supposition 
  that 
  all 
  these 
  variations 
  of 
  the 
  normal 
  triplet 
  type 
  

   are 
  produced 
  by 
  absorption 
  in 
  the 
  vapour 
  of 
  the 
  spark. 
  

  

  An 
  interesting 
  point 
  to 
  be 
  remembered 
  in 
  this 
  connexion 
  

   is 
  that 
  the 
  spectral 
  lines 
  may 
  be 
  more 
  or 
  less 
  complex 
  in 
  their 
  

  

  * 
  See 
  Fievez, 
  Bull, 
  de 
  VAcad. 
  Roy. 
  de 
  Belgique, 
  3rd 
  series, 
  vol. 
  ix. 
  

   p. 
  381 
  (1885) 
  ; 
  Lodge 
  and 
  Davies, 
  « 
  Nature/ 
  vol. 
  lvi. 
  p.'237 
  (1897) 
  ; 
  Zeernanl 
  

   Phil. 
  Mag. 
  vol. 
  xliv. 
  p. 
  55 
  (1897), 
  &c. 
  

  

  