﻿RATE 
  OF 
  PROPAGATION. 
  77 
  

  

  Before 
  leaving 
  this 
  subject 
  it 
  may 
  be 
  well 
  to 
  devote 
  a 
  few 
  words 
  

   to 
  the 
  consideration 
  of 
  the 
  degree 
  of 
  accuracy 
  of 
  the 
  results 
  obtained. 
  

   This 
  1 
  regard 
  as 
  much 
  greater 
  than 
  might 
  reasonably 
  have 
  been 
  ex- 
  

   pected, 
  or 
  than 
  I 
  myself 
  expected, 
  from 
  the 
  data 
  available. 
  It 
  is 
  

   true 
  that 
  the 
  hodograph 
  is 
  a 
  one-sided 
  one, 
  as 
  no 
  observations 
  of 
  

   any 
  value 
  were 
  obtained 
  except 
  to 
  the 
  westwards, 
  and 
  to 
  this 
  extent 
  

   its 
  value 
  is 
  discounted, 
  as 
  the 
  more 
  distant 
  observations 
  could 
  not 
  

   be 
  checked 
  by 
  corresponding 
  ones 
  lying 
  in 
  a 
  different 
  direction 
  from 
  

   the 
  centre. 
  On 
  the 
  other 
  hand, 
  the 
  observations 
  in 
  that 
  one 
  direc- 
  

   tion 
  show 
  a 
  surprisingly 
  close 
  agreement 
  among 
  themselves, 
  an 
  

   agreement 
  which 
  is 
  in 
  no 
  way 
  the 
  result 
  of 
  the 
  selection 
  of 
  observa- 
  

   tions 
  on 
  account 
  of 
  their 
  concordance. 
  All 
  the 
  observations, 
  which 
  

   are 
  given 
  in 
  this 
  chapter, 
  were 
  selected 
  in 
  the 
  first 
  instance 
  on 
  the 
  

   grounds 
  stated 
  above 
  (pages 
  63 
  to 
  67), 
  because, 
  from 
  internal 
  evi- 
  

   dence 
  of 
  care 
  or 
  other 
  reasons, 
  they 
  appeared 
  intrinsically 
  more 
  trust- 
  

   worthy 
  than 
  those 
  which 
  did 
  not, 
  on 
  the 
  face 
  of 
  them, 
  appear 
  to 
  

   possess 
  that 
  value 
  which 
  would 
  alone 
  make 
  it 
  worth 
  while 
  consider- 
  

   ing 
  them. 
  The 
  selection 
  was 
  made 
  before 
  the 
  distances 
  were 
  cal- 
  

   culated, 
  and 
  was 
  moreover 
  originally 
  made 
  at 
  a 
  time 
  when 
  it 
  was 
  still 
  

   believed 
  that 
  the 
  earthquake 
  originated 
  in 
  a 
  very 
  different 
  

   situation 
  to 
  what 
  afterwards 
  turned 
  out 
  to 
  be 
  the 
  case, 
  a 
  difference 
  

   which 
  would 
  have 
  introduced 
  a 
  very 
  appreciable 
  modification 
  in 
  the 
  

   relative 
  distances 
  of 
  all 
  the 
  places 
  lying 
  within 
  300 
  or 
  even 
  more 
  

   miles 
  of 
  the 
  centre 
  finally 
  accepted. 
  

  

  No 
  subsequent 
  additions 
  were 
  made 
  to 
  the 
  selected 
  list, 
  though 
  

   there 
  are 
  a 
  large 
  number 
  of 
  observations 
  which 
  show 
  a 
  close 
  agree- 
  

   ment 
  with 
  the 
  times 
  deduced 
  from 
  the 
  hodograph, 
  nor 
  were 
  any 
  sub- 
  

   sequently 
  rejected, 
  except 
  those 
  mentioned 
  in 
  this 
  chapter. 
  But 
  

   briefly, 
  it 
  may 
  be 
  said 
  every 
  care 
  was 
  taken 
  to 
  prevent 
  the 
  results 
  

   being 
  affected 
  by 
  any 
  bias, 
  conscious 
  or 
  unconscious, 
  and 
  they 
  may, 
  I 
  

   consider, 
  be 
  accepted 
  as 
  accurate 
  at 
  any 
  rate, 
  within 
  an 
  error 
  of 
  less 
  

   than 
  five 
  per 
  cent. 
  

  

  ( 
  77 
  ) 
  

  

  