﻿the 
  Velocity 
  of 
  Light. 
  337 
  

  

  mirror, 
  which 
  reflects 
  the 
  beam 
  through 
  the 
  lens 
  /to 
  the 
  

   distant 
  mirror 
  upon 
  the 
  surface 
  of 
  which 
  the 
  second 
  image 
  of 
  

   the 
  slit 
  is 
  formed. 
  The 
  light 
  then 
  retraces 
  its 
  course 
  and 
  

   returns 
  to 
  its 
  source 
  at 
  a, 
  part 
  being 
  reflected 
  to 
  a 
  1 
  for 
  con- 
  

   venience 
  of 
  observation 
  by 
  the 
  eyepiece. 
  

  

  The 
  limit 
  of 
  closeness 
  of 
  the 
  grating-space 
  is 
  determined 
  

   by 
  the 
  aperture 
  of 
  the 
  revolving 
  mirror 
  viewed 
  from 
  c. 
  

  

  If 
  the 
  " 
  radius 
  " 
  be 
  is 
  3 
  metres 
  and 
  the 
  revolving 
  mirror 
  is 
  

   6 
  cm. 
  wide, 
  this 
  angle 
  will 
  be 
  CM) 
  2 
  ; 
  and 
  the 
  breadth 
  of 
  the 
  

   ■diffraction 
  image 
  at 
  e 
  will 
  be 
  o£ 
  the 
  order 
  \/*02, 
  or, 
  say, 
  

   1/40 
  mm. 
  The 
  grating-space 
  should 
  therefore 
  be 
  at 
  

   least 
  1/20 
  mm., 
  and 
  probably 
  better 
  0*1 
  mm. 
  If 
  the 
  number 
  

   of 
  revolutions 
  is 
  250, 
  and 
  the 
  distance 
  to 
  the 
  fixed 
  mirror 
  be 
  

   3 
  kilometres, 
  the 
  displacement 
  of 
  the 
  first 
  image 
  over 
  the 
  

   grating 
  surface 
  will 
  be 
  18 
  cm., 
  corresponding 
  to 
  1800 
  eclipses. 
  

   There 
  need 
  be 
  no 
  difficulty 
  in 
  counting 
  the 
  order 
  of 
  the 
  

   eclipse 
  observed, 
  if 
  the 
  speed 
  is 
  gradually 
  increased 
  to 
  its 
  

   final 
  value. 
  The 
  fractions 
  could 
  probably 
  be 
  observed 
  cor- 
  

   rectly 
  to 
  something 
  like 
  2 
  per 
  cent., 
  so 
  that 
  this 
  element 
  of 
  

   the 
  computation 
  for 
  V 
  could 
  be 
  measured 
  to 
  something 
  like 
  

   one 
  part 
  in 
  a 
  hundred 
  thousand. 
  

  

  This 
  same 
  or 
  even 
  a 
  higher 
  order 
  of 
  accuracy 
  may 
  be 
  

   obtained 
  in 
  the 
  measurement 
  of 
  all 
  the 
  other 
  elements. 
  

   Previous 
  experiments 
  have 
  shown 
  that 
  the 
  speed 
  of 
  the 
  

   mirror 
  may 
  be 
  obtained, 
  by 
  means 
  of 
  a 
  rated 
  tuning-fork, 
  to 
  

   within 
  one 
  in 
  one 
  hundred 
  thousand. 
  

  

  The 
  measurement 
  of 
  the 
  distance 
  may 
  be 
  made 
  directly 
  on 
  

   a 
  base 
  line 
  especially 
  prepared 
  for 
  such 
  work 
  to 
  within 
  one 
  

   in 
  two 
  hundred 
  thousand. 
  

  

  Finally, 
  the 
  grating 
  may 
  be 
  calibrated 
  to 
  an 
  order 
  of 
  

   accuracy 
  depending 
  on 
  the 
  angle 
  subtended 
  by 
  the 
  diffraction- 
  

   fringes, 
  i. 
  e., 
  1/40 
  x 
  1/3000, 
  or 
  less 
  than 
  one 
  in 
  100,000. 
  

  

  It 
  seems 
  not 
  unreasonable 
  to 
  hope 
  that 
  with 
  proper 
  care 
  

   and 
  patience 
  the 
  value 
  of 
  this 
  great 
  fundamental 
  constant 
  of 
  

   Nature 
  may 
  be 
  found 
  to 
  within 
  five 
  kilometres 
  or 
  less. 
  

  

  It 
  will 
  be 
  noted 
  that 
  in 
  the 
  form 
  of 
  experiment 
  here 
  pro- 
  

   posed, 
  the 
  more 
  serious 
  of 
  the 
  difficulties 
  pointed 
  out 
  by 
  

   M. 
  Oornu 
  no 
  longer 
  exist, 
  or 
  are 
  very 
  much 
  diminished 
  ; 
  

   while 
  the 
  possible 
  accuracy 
  is 
  greatly 
  increased. 
  It 
  may 
  be 
  

   hoped, 
  therefore, 
  that 
  the 
  result 
  of 
  this 
  combination 
  of 
  the 
  

   methods 
  of 
  Foucault 
  and 
  Fizeau 
  will 
  be 
  to 
  reconcile 
  the 
  

   differences 
  which 
  thus 
  far 
  seem 
  to 
  exist 
  between 
  the 
  results 
  

   of 
  the 
  work 
  of 
  their 
  respective 
  followers. 
  

  

  Phil. 
  Mag. 
  S. 
  6. 
  Vol. 
  3. 
  No. 
  15. 
  March 
  1902. 
  Z 
  

  

  