﻿Dispersion, 
  and 
  Surface- 
  Colour 
  of 
  Selenium. 
  617 
  

  

  inch. 
  This 
  grating 
  was 
  sent 
  to 
  me 
  by 
  Mr. 
  Thorp 
  of 
  Man- 
  

   chester. 
  It 
  was 
  made 
  by 
  mounting 
  one 
  of 
  his 
  celluloid 
  

   replicas 
  of 
  a 
  Rowland 
  grating 
  on 
  a 
  concave 
  Ions, 
  the 
  celluloid 
  

   film 
  being 
  subsequently 
  silvered. 
  It 
  was 
  only 
  intended 
  as 
  a 
  

   device 
  for 
  illustrating 
  the 
  action 
  of 
  the 
  concave 
  grating 
  *, 
  

   but 
  I 
  found 
  that 
  if 
  its 
  aperture 
  were 
  reduced 
  somewhat 
  it 
  

   gave 
  surprisingly 
  sharp 
  lines. 
  

  

  A 
  film 
  of 
  selenium 
  was 
  deposited 
  on 
  a 
  thin 
  plate 
  of 
  quartz, 
  

   and 
  mounted 
  in 
  front 
  of 
  the 
  slit 
  of 
  the 
  grating 
  outfit, 
  the 
  

   lower 
  half 
  of 
  the 
  slit 
  being 
  covered 
  with 
  clear 
  quartz, 
  the 
  

   upper 
  half 
  with 
  quartz 
  and 
  selenium. 
  

  

  The 
  slit 
  was 
  illuminated 
  by 
  a 
  spark-discharge 
  between 
  zinc 
  

   terminals, 
  and 
  the 
  spectrum 
  photographed. 
  The 
  absorbing- 
  

   action 
  of 
  the 
  selenium 
  manifested 
  itself 
  as 
  a 
  more 
  or 
  less 
  

   complete 
  obscuration 
  of 
  the 
  upper 
  half 
  of 
  the 
  spectrum. 
  By 
  

   giving 
  the 
  selenium-covered 
  portion 
  of 
  the 
  slit 
  a 
  very 
  long- 
  

   exposure, 
  and 
  the 
  clear 
  portion 
  a 
  short 
  one, 
  any 
  given 
  part 
  

   of 
  the 
  spectrum 
  could 
  be 
  balanced, 
  the 
  ratio 
  of 
  the 
  times 
  

  

  giving 
  us 
  an 
  approximate 
  value 
  of 
  ^7. 
  

  

  We 
  exposed 
  a 
  number 
  of 
  plates, 
  and 
  fully 
  established 
  the 
  fact 
  

   that 
  the 
  turning-point 
  of 
  the 
  extinction 
  curve,if 
  it 
  exists 
  at 
  all, 
  

   lies 
  beyond 
  wave-length 
  "00022. 
  In 
  the 
  case 
  of 
  one 
  plate 
  the 
  

   clear 
  portion 
  of 
  the 
  slit 
  was 
  exposed 
  for 
  30 
  seconds, 
  the 
  

   selenium-covered 
  portion 
  for 
  62 
  minutes. 
  The 
  resulting 
  photo- 
  

   graph 
  showed 
  lines 
  down 
  to 
  wave-length 
  '00022 
  for 
  the 
  clear 
  

   part 
  of 
  the 
  slit, 
  while 
  the 
  selenium-covered 
  part 
  only 
  recorded 
  

   itself 
  down 
  to 
  *00028, 
  the 
  lines 
  being, 
  however, 
  very 
  much 
  

   fainter 
  than 
  the 
  corresponding 
  ones 
  in 
  the 
  other 
  half 
  of 
  the 
  

   spectrum. 
  The 
  lines 
  were 
  of 
  about 
  the 
  same 
  intensity 
  at 
  

   wave-length 
  '00033, 
  for 
  which 
  we 
  get, 
  from 
  the 
  ratio 
  of 
  the 
  

   times 
  of 
  exposure, 
  the 
  value 
  1/124 
  for 
  I/I'. 
  This 
  gives 
  us 
  

   the 
  value 
  1'255 
  for 
  k. 
  The 
  thickness 
  of 
  the 
  film 
  was 
  the 
  

   same 
  as 
  in 
  the 
  previous 
  experiments. 
  {Still 
  longer 
  exposures 
  

   would 
  have 
  to 
  be 
  given 
  to 
  balance 
  the 
  two 
  halves 
  of 
  the 
  

   spectrum 
  for 
  shorter 
  waves 
  than 
  these, 
  from 
  which 
  it 
  is 
  

   evident 
  that 
  the 
  extinction 
  curve 
  is 
  still 
  rising 
  rapidly 
  at 
  

   wave-length 
  0*(!0022, 
  where 
  the 
  coefficient 
  has 
  as 
  high 
  a 
  

   value 
  as 
  in 
  the 
  case 
  of 
  the 
  metals. 
  

  

  The 
  extinction 
  curve 
  as 
  figured 
  is 
  not 
  to 
  be 
  regarded 
  as 
  

  

  * 
  I 
  have 
  mounted 
  this 
  grating 
  on 
  a 
  light 
  wooden 
  frame 
  designed 
  

   according 
  to 
  the 
  well-known 
  Rowland 
  device, 
  which 
  can 
  be 
  placed 
  on 
  

   the 
  lecture-table 
  and 
  used 
  for 
  class 
  illustration. 
  The 
  slit 
  is 
  illuminated 
  

   with 
  sun 
  or 
  arc 
  light 
  and 
  the 
  spectra 
  received 
  on 
  a 
  strip 
  of 
  ground 
  glass. 
  

   It 
  is 
  useful 
  for 
  showing 
  how 
  the 
  spectra 
  and 
  the 
  camera 
  move 
  relatively 
  

   to 
  each 
  other. 
  

  

  