﻿374 
  On 
  the 
  Constitution 
  of 
  the 
  Mercury 
  Green 
  Line. 
  

  

  induction-coil. 
  The 
  tubes 
  were 
  exhausted 
  bj 
  a 
  Fleuss 
  two- 
  

   cylinder 
  pump, 
  and 
  were 
  in 
  open 
  connexion 
  with 
  a 
  P 
  2 
  5 
  

   drying-tube. 
  On 
  heating 
  the 
  tube 
  slightly 
  the 
  effect 
  was 
  

   soon 
  obtained. 
  The 
  discharge 
  in 
  the 
  tube 
  became 
  very 
  

   brilliant 
  and 
  very 
  broad 
  and 
  continuous, 
  suggesting 
  a 
  

   breakdown 
  of 
  the 
  insulating 
  properties 
  or! 
  the 
  vacuum, 
  

   while 
  the 
  appearance 
  in 
  the 
  echelon 
  exactly 
  corresponded 
  to 
  

   Janicki's 
  description 
  of 
  "five 
  equidistant 
  bands.'" 
  With 
  

   the 
  arc, 
  as 
  described 
  above, 
  we 
  were 
  unable 
  to 
  obtain 
  this 
  

   effect. 
  The 
  cases 
  differ 
  in 
  three 
  respects: 
  (1) 
  temperature, 
  

   (2) 
  perfection 
  of 
  vacuum, 
  (3) 
  nature 
  of 
  discharge. 
  

  

  (1) 
  We 
  have 
  measured 
  the 
  temperature 
  at 
  which 
  the 
  

   effect 
  was 
  obtained, 
  and 
  find 
  it 
  varies 
  from 
  200° 
  C. 
  to 
  260° 
  C. 
  

   or 
  more, 
  depending 
  apparently 
  upon 
  the 
  degree 
  of 
  vacuum. 
  

   The 
  temperatures 
  attained 
  by 
  the 
  arc 
  varied 
  as 
  stated 
  above 
  

   from 
  130° 
  C. 
  to 
  over 
  400° 
  C. 
  Hence 
  the 
  effect 
  is 
  not 
  due 
  

   solely 
  to 
  temperature. 
  

  

  (2) 
  We 
  find 
  that 
  the 
  effect 
  is 
  more 
  easily 
  obtained 
  with 
  

   an 
  imperfectly 
  exhausted 
  tube, 
  though 
  we 
  could 
  always 
  

   obtain 
  it 
  with 
  the 
  highest 
  vacua 
  that 
  our 
  pump 
  could 
  give. 
  

  

  (3) 
  The 
  nature 
  of 
  the 
  discharge 
  seems 
  to 
  have 
  some 
  slight 
  

   •effect. 
  When 
  the 
  hammer 
  of 
  the 
  coil 
  was 
  working 
  unsteadily, 
  

   the 
  appearance 
  would 
  flicker 
  in 
  and 
  out 
  even 
  at 
  temperatures 
  

   below 
  200° 
  C, 
  while 
  when 
  the 
  proper 
  conditions 
  had 
  been 
  

   attained 
  it 
  worked 
  smoothly. 
  

  

  We 
  have 
  succeeded 
  in 
  obtaining 
  photographs 
  of 
  the 
  effect, 
  

   and 
  from 
  these 
  the 
  measurements 
  have 
  been 
  taken. 
  O 
  wing- 
  

   to 
  the 
  great 
  width 
  of 
  the 
  bright 
  lines 
  it 
  was 
  not 
  easy 
  to 
  

   judge 
  their 
  centres, 
  and 
  we 
  do 
  not 
  claim 
  an 
  accuracy 
  of 
  more 
  

  

  than 
  '005 
  A.U 
  . 
  The 
  whole 
  appearance 
  so 
  strongly 
  suggested 
  

   an 
  inversion 
  due 
  to 
  absorption, 
  that 
  we 
  have 
  also 
  measured 
  

   the 
  distances 
  between 
  the 
  dark 
  lines 
  to 
  compare 
  with 
  the 
  

   normal 
  pattern 
  given 
  above. 
  

  

  Bright 
  lines 
  ... 
  -'206 
  -'103 
  -'059 
  -.000 
  '089 
  '177 
  

  

  Dark 
  lines 
  ... 
  -'234 
  -'124 
  --051 
  "000 
  -075 
  -163 
  

  

  Normal 
  pattern 
  - 
  *235 
  - 
  -187 
  -102 
  - 
  -067 
  '000 
  -083 
  128 
  -174 
  -212 
  

  

  The 
  line 
  at 
  — 
  *059 
  is 
  very 
  faint 
  and 
  only 
  obtained 
  on 
  some 
  

   of 
  the 
  photographs. 
  The 
  one 
  at 
  "177 
  is 
  very 
  broad 
  and 
  may 
  

   be 
  two 
  lines 
  or 
  rather 
  bands 
  with 
  a 
  dark 
  line 
  between. 
  

   The 
  correspondence 
  of 
  patterns 
  is 
  better 
  shown 
  by 
  a 
  diagram 
  

   in 
  which 
  the 
  relative 
  brightness 
  of 
  the 
  lines 
  is 
  shown. 
  

  

  By 
  omitting 
  the 
  faint 
  band 
  at 
  — 
  *059 
  Janicki's 
  system 
  

   of 
  five 
  "equidistant"" 
  bands 
  is 
  obtained. 
  We 
  consider 
  that 
  

  

  