﻿of 
  the 
  Solar 
  Prominences. 
  . 
  461 
  

  

  a 
  number 
  of 
  photographs 
  o 
  . 
  its 
  spectrum 
  were 
  made 
  with 
  

   both 
  radial 
  and 
  tangential 
  slit. 
  

  

  In 
  reducing 
  the 
  wave-lengths 
  of 
  these 
  lines 
  it 
  might 
  be 
  con- 
  

   sidered 
  easy 
  to 
  obtain 
  values 
  for 
  a 
  given 
  line 
  agreeing 
  closely 
  

   in 
  the 
  hundredths 
  place 
  of 
  tenth-meters, 
  but 
  two 
  causes 
  have 
  

   combined 
  to 
  lessen 
  the 
  accuracy 
  of 
  determinations. 
  The 
  H 
  

   and 
  K 
  reversals 
  almost 
  invariably 
  show 
  some 
  indications 
  of 
  mo- 
  

   tion 
  of 
  the 
  prominences 
  in 
  the 
  line 
  of 
  sight, 
  and 
  the 
  conse- 
  

   quent 
  distortion 
  renders 
  somewhat 
  difficult 
  the 
  proper 
  setting 
  

   of 
  the 
  spider 
  line 
  of 
  the 
  measuring 
  machine. 
  Again 
  the 
  plate- 
  

   holder 
  used 
  was 
  made 
  for 
  another 
  purpose, 
  which 
  required 
  

   that 
  the 
  plane 
  of 
  the 
  plate 
  should 
  be 
  at 
  right 
  angles 
  to 
  the 
  axis 
  

   of 
  the 
  observing 
  telescope. 
  As 
  the 
  object-glass 
  of 
  the 
  tele- 
  

   scope 
  is 
  corrected 
  for 
  the 
  visual 
  region, 
  it 
  is 
  evident 
  that 
  near 
  

   K 
  there 
  must 
  be 
  a 
  slight 
  change 
  in 
  focus 
  from 
  one 
  side 
  of 
  the 
  

   plate 
  to 
  the 
  other, 
  and 
  a 
  small 
  error 
  is 
  thus 
  introduced. 
  It 
  will 
  

   be 
  seen, 
  however, 
  that 
  the 
  measures 
  are 
  sufficiently 
  accurate 
  to 
  

   allow 
  very 
  little 
  doubt 
  as 
  to 
  the 
  identity 
  of 
  most 
  of 
  the 
  lines. 
  

   The 
  fact 
  that 
  the 
  solar 
  spectrum, 
  due 
  to 
  the 
  diffuse 
  light 
  

   of 
  the 
  atmosphere, 
  is 
  photographed 
  simultaneously 
  with 
  the 
  

   prominence 
  spectrum, 
  is 
  of 
  great 
  advantage 
  in 
  determining 
  

   the 
  position 
  of 
  the 
  prominence 
  lines, 
  though 
  it 
  has 
  a 
  corres- 
  

   ponding 
  disadvantage 
  in 
  concealing 
  very 
  faint 
  lines, 
  which 
  

   would 
  otherwise 
  be 
  brought 
  out. 
  The 
  wave-lengths 
  of 
  certain 
  

   standard 
  lines 
  in 
  the 
  solar 
  spectrum 
  have 
  been 
  taken 
  from 
  the 
  

   list 
  published 
  by 
  Professor 
  Rowland,* 
  and 
  it 
  has 
  thus 
  been 
  

   easy 
  to 
  find 
  the 
  wave-lengths 
  of 
  the 
  prominence 
  lines 
  by 
  

   simple 
  interpolation. 
  The 
  value 
  of 
  the 
  micrometer 
  screw 
  has 
  

   been 
  determined 
  for 
  several 
  regions 
  on 
  every 
  plate 
  by 
  measur- 
  

   ing 
  the 
  positions 
  of 
  properly 
  distributed 
  standard 
  lines, 
  the 
  

   number 
  of 
  separate 
  settings 
  of 
  the 
  spider 
  line 
  in 
  each 
  case 
  

   ranging 
  from 
  five 
  to 
  fifteen, 
  depending 
  upon 
  the 
  character 
  of 
  

   the 
  line 
  measured. 
  In 
  the 
  following 
  table 
  the 
  first 
  column 
  

   contains 
  the 
  wave-lengths 
  of 
  the 
  ultra-violet 
  prominence 
  lines 
  ; 
  

   the 
  second 
  the 
  positions 
  assigned 
  by 
  Ames 
  to 
  the 
  lines 
  in 
  the 
  

   hydrogen 
  stellar 
  series 
  ; 
  and 
  the 
  third, 
  the 
  wave-lengths 
  of 
  the 
  

   calcium 
  lines 
  at 
  H 
  and 
  K, 
  which 
  Professor 
  Rowland 
  has 
  been 
  

   kind 
  enough 
  to 
  furnish 
  in 
  advance 
  of 
  publication. 
  I 
  am 
  in- 
  

   formed 
  that 
  they 
  are 
  provisional 
  only, 
  but 
  may 
  be 
  relied 
  on 
  to 
  

   within 
  1 
  or 
  2 
  in 
  the 
  last 
  place 
  of 
  decimals. 
  In 
  the 
  case 
  of 
  

   hydrogen, 
  Ames 
  considers 
  that 
  the 
  error 
  in 
  any 
  wave-length 
  

   cannot 
  amount 
  to 
  more 
  than 
  0'05 
  of 
  a 
  unit,f 
  and 
  my 
  own 
  

   values 
  for 
  the 
  prominence 
  lines 
  must 
  possess 
  at 
  least 
  an 
  equal 
  

   degree 
  of 
  accuracy, 
  though 
  I 
  am 
  inclined 
  to 
  believe 
  some 
  of 
  

   them 
  even 
  more 
  reliable. 
  In 
  the 
  fourth 
  and 
  fifth 
  columns 
  I 
  

  

  * 
  This 
  Journal, 
  p. 
  182, 
  188Y. 
  fPhil. 
  Mag., 
  July, 
  1890, 
  p. 
  49. 
  

  

  