﻿526 
  Mr. 
  F. 
  Soddy 
  on 
  the 
  Production 
  of 
  

  

  smallest 
  possible 
  volume, 
  is 
  muck 
  greater 
  than 
  was 
  formerly 
  

   deduced 
  by 
  calculation. 
  Thus 
  while 
  the 
  estimates 
  o£ 
  the 
  

   amounts 
  of 
  helium 
  to 
  be 
  expected 
  have 
  grown 
  less, 
  the 
  estimates 
  

   of 
  the 
  capabilities 
  of 
  the 
  spectroscope 
  have 
  also 
  diminished 
  ; 
  

   so 
  that 
  the 
  task 
  is 
  now 
  known 
  to 
  be 
  more 
  difficult, 
  for 
  a 
  double 
  

   reason, 
  than 
  was 
  at 
  first 
  supposed. 
  

  

  In 
  the 
  present 
  series 
  of 
  tests 
  the 
  processes 
  of 
  sealing 
  off, 
  

   absorption, 
  and 
  subsequent 
  tilling 
  of 
  the 
  apparatus 
  with 
  

   mercury, 
  were 
  carried 
  out 
  exactly 
  as 
  in 
  actual 
  experiments. 
  

   Tests 
  were 
  conducted 
  with 
  three 
  different 
  types 
  of 
  spectrum- 
  

   tubes, 
  all 
  furnished 
  with 
  electrodes 
  of 
  simple 
  platinum 
  wire. 
  

   Those 
  used 
  in 
  the 
  early 
  work, 
  but 
  now 
  abandoned, 
  had 
  a 
  volume 
  

   of 
  about 
  0'3 
  c.c, 
  and 
  were 
  made 
  of 
  soda-glass. 
  The 
  bore 
  of 
  

   the 
  capillary 
  was 
  rather 
  too 
  wide 
  for 
  the 
  best 
  results. 
  The 
  

   kind 
  now 
  used 
  exclusively 
  consist 
  of 
  a 
  very 
  small 
  tube 
  indeed, 
  

   with 
  hardly 
  any 
  unnecessary 
  space 
  around 
  the 
  electrodes, 
  

   made 
  of 
  lead-glass 
  with 
  a 
  total 
  volume 
  of 
  only 
  about 
  0*04 
  c.c. 
  

   A 
  third 
  type 
  were 
  exactly 
  similar 
  except 
  that 
  they 
  were 
  made 
  

   of 
  soda-glass. 
  They 
  all 
  became 
  exhausted 
  under 
  the 
  discharge 
  

   with 
  great 
  rapidity, 
  but 
  for 
  the 
  present 
  work 
  the 
  advantage 
  

   of 
  smaller 
  volume, 
  made 
  possible 
  by 
  the 
  use 
  of 
  platinum 
  

   electrodes, 
  outweighed 
  the 
  advantage 
  of 
  longer 
  life 
  to 
  be 
  

   derived 
  from 
  using 
  aluminium 
  electrodes. 
  

  

  Several 
  mixtures 
  of 
  pure 
  helium 
  with 
  pure 
  oxygen, 
  con- 
  

   taining 
  only 
  a 
  few 
  parts 
  per 
  million 
  of 
  helium, 
  were 
  made 
  up 
  

   by 
  the 
  principle 
  of 
  successive 
  dilutions, 
  with 
  the 
  greatest 
  care, 
  

   using 
  the 
  spectrum-tube 
  filler 
  devised 
  by 
  Sir 
  William 
  Ramsay. 
  

   With 
  the 
  same 
  device, 
  a 
  known 
  fraction 
  of 
  a 
  cubic 
  centim. 
  of 
  

   the 
  mixture 
  was 
  admitted 
  to 
  the 
  exhausted 
  calcium 
  furnace. 
  

   The 
  furnace 
  was 
  sealed 
  off, 
  and 
  a 
  correction 
  applied 
  for 
  the 
  

   part 
  of 
  the 
  gas 
  not 
  used. 
  The 
  absorption 
  of 
  the 
  ox3^gen 
  and 
  

   rilling 
  of 
  the 
  apparatus 
  with 
  mercury 
  were 
  then 
  carried 
  out 
  

   in 
  the 
  usual 
  manner. 
  

  

  The 
  general 
  result 
  has 
  been 
  to 
  show 
  that 
  with 
  the 
  lead-glass 
  

   tube 
  described, 
  a 
  quantity 
  of 
  helium 
  below 
  a 
  " 
  tenth-gram," 
  

   that 
  is 
  10~ 
  10 
  gram, 
  cannot 
  be 
  detected 
  ; 
  while 
  in 
  a 
  perfect 
  

   experiment 
  two 
  tenth-grams 
  can 
  be 
  detected 
  with 
  certainty. 
  

   For 
  the 
  tubes 
  of 
  soda-glass 
  otherwise 
  similar, 
  the 
  quantity 
  

   detectable 
  is 
  about 
  twice 
  this. 
  In 
  the 
  first 
  place 
  the 
  brightness 
  

   of 
  the 
  neighbouring 
  sodium 
  lines 
  interferes 
  with 
  the 
  test; 
  

   and, 
  secondly, 
  the 
  glass 
  conducts 
  much 
  more 
  readily 
  than 
  

   lead-glass 
  ; 
  so 
  that 
  the 
  discharge 
  under 
  some 
  circumstances 
  

   passes, 
  not 
  through 
  the 
  glass 
  at 
  all, 
  the 
  spectrum 
  of 
  which 
  is 
  

   not 
  observed, 
  but 
  almost 
  wholly 
  along 
  the 
  glass 
  surface, 
  

   causing 
  it 
  to 
  fluoresce 
  brilliantly 
  and 
  show 
  the 
  sodium 
  lines. 
  

   Test 
  C 
  VI. 
  of 
  the 
  table 
  was 
  a 
  very 
  instructive 
  one 
  in 
  this 
  

  

  