﻿Minute 
  Metal 
  Particles 
  for 
  Light-waves. 
  399 
  

  

  next 
  found 
  that 
  cooling 
  the 
  films 
  by 
  the 
  application 
  of 
  ice 
  to 
  

   the 
  exterior 
  of 
  the 
  bulb 
  produced 
  most 
  extraordinary 
  changes 
  

   of 
  colour. 
  Pale 
  green 
  films, 
  almost 
  transparent, 
  on 
  being 
  

   cooled 
  ten 
  or 
  fifteen 
  degrees, 
  changed 
  to 
  a 
  violet 
  as 
  deep 
  as 
  

   that 
  shown 
  by 
  dense 
  cobalt-glass. 
  Pink 
  films 
  were 
  changed 
  

   to 
  deep 
  blue-green, 
  while 
  films 
  originally 
  deep 
  blue 
  became 
  

   transparent. 
  These 
  effects 
  were 
  most 
  puzzling 
  at 
  first, 
  but 
  I 
  

   finally 
  succeeded 
  in 
  finding 
  out 
  the 
  cause, 
  as 
  I 
  shall 
  show 
  

   later 
  on. 
  At 
  the 
  beginning 
  of 
  the 
  work 
  it 
  was 
  almost 
  

   impossible 
  to 
  predict 
  what 
  sort 
  of 
  a 
  film 
  would 
  be 
  obtained 
  

   in 
  any 
  experiment. 
  Sometimes, 
  on 
  first 
  heating 
  the 
  sodium, 
  

   a 
  film 
  of 
  a 
  deep 
  claret 
  colour 
  would 
  be 
  obtained; 
  once 
  formed, 
  

   it 
  could 
  be 
  driven 
  from 
  one 
  part 
  of 
  the 
  bulb 
  to 
  another 
  with- 
  

   out 
  any 
  great 
  change 
  of 
  colour. 
  In 
  another 
  experiment, 
  

   made 
  under 
  exactly 
  the 
  same 
  conditions, 
  so 
  far 
  as 
  I 
  could 
  see, 
  

   it 
  was 
  impossible 
  to 
  obtain 
  anything 
  but 
  a 
  deep 
  blue. 
  I 
  

   prepared 
  and 
  experimented 
  with 
  fully 
  fifty 
  bulbs, 
  without 
  

   being 
  able 
  to 
  arrive 
  at 
  any 
  definite 
  conclusion 
  in 
  regard 
  to 
  

   the 
  conditions 
  necessary 
  for 
  producing 
  a 
  definite 
  result. 
  It 
  

   was 
  finally 
  found 
  however, 
  by 
  blowing 
  connected 
  bulbs, 
  that 
  

   if 
  the 
  sodium 
  was 
  thoroughly 
  boiled 
  in 
  one 
  of 
  them, 
  and 
  

   while 
  still 
  molten 
  was 
  shaken 
  into 
  an 
  adjoining 
  bulb, 
  the 
  

   surface 
  of 
  which 
  was 
  of 
  course 
  free 
  from 
  any 
  deposit, 
  on 
  

   heating 
  the 
  globule 
  in 
  the 
  fresh 
  bulb, 
  only 
  blue 
  films 
  could 
  

   be 
  produced. 
  This 
  made 
  it 
  appear 
  as 
  if 
  the 
  claret-coloured 
  

   film 
  were 
  produced 
  by 
  something 
  volatile 
  in 
  the 
  sodium 
  

   which 
  was 
  driven 
  off 
  by 
  the 
  first 
  heating. 
  It 
  was 
  necessary, 
  

   if 
  possible, 
  to 
  find 
  out 
  what 
  this 
  was, 
  for 
  these 
  red 
  films 
  

   were 
  found 
  to 
  be 
  more 
  interesting 
  than 
  the 
  blue, 
  and 
  it 
  

   was 
  desirable 
  to 
  be 
  able 
  to 
  produce 
  them 
  at 
  will. 
  As 
  a 
  result 
  

   of 
  many 
  experiments, 
  I 
  had 
  formed 
  a 
  vague 
  notion 
  that 
  bits 
  

   of 
  sodium 
  cut 
  from 
  certain 
  lumps 
  in 
  the 
  bottle 
  produced 
  the 
  

   red 
  films 
  on 
  the 
  first 
  heating, 
  while 
  portions 
  taken 
  from 
  other 
  

   lumps 
  did 
  not. 
  It 
  seemed 
  not 
  unlikely 
  that 
  the 
  sodium 
  might 
  

   contain 
  potassium 
  as 
  an 
  impurity, 
  which 
  might 
  be 
  the 
  agent 
  

   concerned. 
  I 
  accordingly 
  prepared 
  a 
  set 
  of 
  bulbs 
  containing 
  

   potassium, 
  and 
  immediately 
  found 
  that 
  magnificent 
  reds 
  and 
  

   purples 
  could 
  be 
  obtained 
  at 
  will; 
  and 
  I 
  felt 
  firmly 
  convinced 
  

   that 
  the 
  anomalies 
  shown 
  by 
  sodium 
  were 
  due 
  to 
  potassium 
  

   impurities. 
  Similar 
  coloured 
  films 
  could 
  be 
  obtained 
  with 
  

   lithium, 
  though, 
  owing 
  to 
  its 
  higher 
  boiling-point, 
  bulbs 
  of 
  

   bohemian 
  glass 
  are 
  required. 
  Letting 
  this 
  serve 
  as 
  an 
  

   introduction, 
  it 
  will 
  be 
  well 
  to 
  take 
  up 
  the 
  experiments 
  

   somewhat 
  in 
  detail. 
  

  

  As 
  the 
  optical 
  behaviour 
  of 
  the 
  films 
  is 
  extremely 
  interesting, 
  

   regardless 
  of 
  the 
  ultimate 
  cause 
  of 
  the 
  colour, 
  I 
  shall 
  describe 
  

  

  