﻿Physics. 
  215 
  

  

  as 
  a 
  "holograph 
  would 
  be. 
  The 
  form 
  of 
  this 
  curve 
  would 
  be 
  a 
  

   straight 
  line 
  only 
  broken 
  by 
  the 
  minute 
  accidential 
  deflection 
  of 
  

   the 
  galvanometer, 
  except 
  where 
  the 
  narrow 
  bands 
  of 
  radiation 
  at 
  

   the 
  above 
  wave-lengths 
  caused 
  narrow, 
  steep 
  sided 
  elevations. 
  

   If 
  we 
  suppose, 
  still 
  further, 
  that 
  either 
  before 
  or 
  after 
  this 
  curve 
  

   was 
  made 
  the 
  direct 
  sunlight 
  was 
  reflected 
  upon 
  the 
  slit 
  of 
  the 
  

   spectro-bolometer 
  for 
  several 
  minutes, 
  then 
  a 
  well-known 
  portion 
  

   of 
  the 
  solar 
  spectrum 
  energy 
  curve 
  would 
  appear 
  in 
  its 
  proper 
  

   relative 
  position 
  on 
  the 
  same 
  plate, 
  and 
  the 
  positions 
  of 
  all 
  the 
  

   sharp 
  elevations 
  corresponding 
  to 
  known 
  wave-lengths 
  could 
  

   readily 
  be 
  measured 
  on 
  the 
  comparator 
  with 
  reference 
  to 
  deter- 
  

   mined 
  solar 
  absorption 
  lines 
  of 
  the 
  short 
  holographs. 
  Thus 
  the 
  

   wave-lengths 
  at 
  as 
  many 
  points 
  as 
  desired 
  could 
  be 
  determined 
  

   without 
  any 
  circle 
  readings 
  whatever. 
  

  

  Practically, 
  this 
  process 
  (somewhat 
  altered 
  in 
  details) 
  was 
  

   gone 
  through 
  with 
  for 
  38 
  positions 
  between 
  wave-lengths 
  0.76/x 
  

   and 
  G.5/x; 
  and 
  not 
  only 
  once, 
  but 
  several 
  times, 
  with 
  all 
  the 
  care 
  

   for 
  accuracy 
  which 
  could 
  be 
  taken. 
  As 
  a 
  result, 
  it 
  may 
  be 
  said 
  

   that 
  the 
  wave-lengths 
  of 
  the 
  absorption 
  lines 
  in 
  the 
  infra-red 
  

   solar 
  spectrum 
  discovered 
  at 
  this 
  observatory 
  can 
  be 
  told 
  with 
  

   an 
  accuracy 
  of 
  about 
  3 
  parts 
  in 
  10,000, 
  while 
  previous 
  to 
  this 
  

   determination 
  1 
  part 
  in 
  100 
  would 
  have 
  been 
  all 
  that 
  could 
  be 
  

   claimed. 
  

  

  (2) 
  Distribution 
  of 
  energy 
  from 
  terrestrial 
  sources. 
  — 
  A 
  number 
  

   of 
  energy 
  curves 
  were 
  taken, 
  in 
  which 
  the 
  Kitson 
  lamp, 
  provided 
  

   with 
  mantles 
  of 
  various 
  kinds, 
  was 
  the 
  source. 
  Among 
  the 
  

   mantles 
  tried 
  were 
  the 
  ordinary 
  Welsbach 
  (which 
  consists 
  of 
  

   impure 
  thorium 
  oxide) 
  and 
  others 
  composed 
  of 
  pure 
  thorium 
  

   oxide, 
  iron 
  oxide, 
  uranium 
  oxide, 
  etc. 
  The 
  distribution 
  of 
  energy 
  

   between 
  different 
  wave-lengths 
  with 
  these 
  sources 
  so 
  different 
  in 
  

   illuminating 
  power, 
  is 
  much 
  less 
  diversified 
  than 
  would 
  be 
  sup- 
  

   posed, 
  and 
  goes 
  strongly 
  to 
  show 
  the 
  wastefulness 
  even 
  of 
  the 
  

   Welsbach 
  light 
  as 
  a 
  source 
  of 
  illumination. 
  For 
  the 
  invisible 
  infra- 
  

   red 
  in 
  all 
  cases 
  includes 
  by 
  far 
  the 
  major 
  portion 
  of 
  the 
  energy, 
  and 
  

   not 
  the 
  visible 
  spectrum, 
  as 
  is 
  the 
  case 
  with 
  the 
  sun 
  and 
  still 
  more 
  

   with 
  phosphorescent 
  substances. 
  However, 
  by 
  the 
  employment 
  

   of 
  a 
  second 
  spectroscope, 
  or 
  "sifting 
  train," 
  to 
  exclude 
  the 
  stray 
  

   infra-red 
  radiations, 
  we 
  were 
  able 
  to 
  determine 
  the 
  distribution 
  

   of 
  the 
  relatively 
  small 
  amount 
  of 
  energy 
  in 
  the 
  visible 
  spectra 
  of 
  

   the 
  various 
  sources, 
  and 
  to 
  show 
  how 
  tar 
  the 
  ordinary 
  Welsbach 
  

   mantle 
  outstripped 
  them 
  all 
  for 
  light, 
  especially 
  in 
  the 
  red, 
  

   orange, 
  and 
  yellow. 
  In 
  these 
  experiments 
  a 
  very 
  considerable 
  

   number 
  (at 
  least 
  fifty) 
  of 
  absorption 
  bands 
  were 
  discovered 
  at 
  

   wave-lengths 
  beyond 
  4/x, 
  which 
  were 
  most 
  probably 
  due 
  to 
  the 
  

   gases 
  given 
  off 
  by 
  the 
  lamp 
  in 
  burning, 
  and 
  perhaps 
  solely 
  to 
  

   carbon 
  dioxide. 
  

  

  (3) 
  Absorption 
  in 
  the 
  solar 
  spectrum. 
  — 
  All 
  the 
  holographic 
  

   records, 
  extending 
  back 
  to 
  1893, 
  were 
  carefully 
  examined 
  with 
  

   regard 
  to 
  the 
  changes 
  in 
  absorption 
  noted 
  in 
  last 
  year's 
  report, 
  

   and 
  such 
  changes 
  were 
  found 
  to 
  be 
  more 
  extensive 
  and 
  frequent 
  

  

  