﻿G. 
  C. 
  Comstock 
  — 
  The 
  Secular 
  Variation 
  of 
  Latitudes. 
  481 
  

  

  These 
  methods 
  as 
  commonly 
  applied 
  require 
  an 
  accurate 
  

   knowledge 
  of 
  the 
  declinations 
  and 
  proper 
  motions 
  of 
  the 
  stars 
  

   but 
  it 
  is 
  perfectly 
  feasible 
  to 
  eliminate 
  both 
  declinations 
  and 
  

   proper 
  motions 
  and 
  leave 
  the 
  resulting 
  variation 
  of 
  latitude 
  

   almost 
  if 
  not 
  quite 
  free 
  from 
  systematic 
  error. 
  To 
  illustrate, 
  

   suppose 
  two 
  stations 
  to 
  be 
  selected 
  as 
  nearly 
  as 
  possible 
  on 
  the 
  

   same 
  parallel 
  of 
  latitude, 
  one 
  in 
  longitude 
  70° 
  west 
  of 
  Green- 
  

   wich 
  and 
  the 
  other 
  110° 
  east 
  and 
  let 
  the 
  latitudes 
  of 
  the 
  sta- 
  

   tions 
  be 
  simultaneously 
  determined 
  by 
  zenith 
  telescope 
  obser- 
  

   vations 
  of 
  the 
  same 
  pairs 
  of 
  stars. 
  The 
  difference 
  of 
  the 
  

   latitudes 
  of 
  the 
  stations 
  thus 
  determined 
  is 
  entirely 
  independ- 
  

   ent 
  of 
  the 
  star 
  places, 
  and 
  I 
  know 
  of 
  no 
  source 
  of 
  systematic 
  

   error 
  by 
  which 
  this 
  difference 
  can 
  be 
  affected 
  except 
  possible 
  

   personal 
  peculiarities 
  of 
  the 
  observers 
  which 
  can 
  be 
  eliminated 
  

   by 
  an 
  interchange 
  of 
  observers 
  if 
  this 
  should 
  be 
  thought 
  de- 
  

   sirable. 
  The 
  periodic 
  variation 
  of 
  the 
  latitude 
  would 
  be 
  

   eliminated 
  from 
  the 
  mean 
  of 
  observations 
  made 
  at 
  epochs 
  six 
  

   months 
  apart. 
  An 
  annual 
  motion 
  of 
  the 
  pole 
  of 
  O'^OIS 
  will 
  

   alter 
  the 
  difference 
  of 
  latitude 
  of 
  these 
  stations 
  by 
  twice 
  this 
  

   amount 
  per 
  year 
  giving 
  a 
  change 
  in 
  the 
  difference 
  of 
  latitude 
  

   amounting 
  to 
  \" 
  in 
  eleven 
  years, 
  a 
  quantity 
  which 
  cannot 
  pos- 
  

   sibly 
  escape 
  careful 
  observations 
  with 
  the 
  zenith 
  telescope 
  or 
  

   prime 
  vertical 
  transit. 
  If 
  similar 
  observations 
  be 
  conducted 
  

   near 
  the 
  meridian 
  20° 
  east 
  of 
  Greenwich 
  they 
  will 
  furnish 
  the 
  

   best 
  attainable 
  data 
  for 
  determining 
  the 
  direction 
  of 
  motion 
  of 
  

   the 
  pole. 
  The 
  execution 
  of 
  this 
  program, 
  which 
  can 
  be 
  

   effected 
  within 
  a 
  dozen 
  years, 
  will 
  add 
  more 
  to 
  our 
  knowledge 
  

   of 
  the 
  variation, 
  or 
  possible 
  permanence, 
  of 
  terrestrial 
  latitudes 
  

   than 
  can 
  be 
  furnished 
  by 
  all 
  the 
  astronomical 
  observations 
  

   that 
  have 
  hitherto 
  been 
  made. 
  By 
  a 
  proper 
  selection 
  of 
  sta- 
  

   tions 
  it 
  will 
  even 
  be 
  possible 
  within 
  a 
  year 
  or 
  two 
  to 
  test 
  the 
  

   results 
  above 
  obtained. 
  The 
  following 
  pairs 
  of 
  stations 
  ap- 
  

   proximately 
  satisfy 
  the 
  conditions 
  above 
  indicated 
  and 
  in 
  addi- 
  

   tion 
  possess 
  the 
  great 
  advantage 
  that 
  at 
  each 
  one 
  of 
  them 
  a 
  

   good 
  value 
  of 
  the 
  latitude 
  was 
  determined 
  by 
  the 
  Talcott 
  

   method 
  prior 
  to 
  1875 
  : 
  

  

  Vladivostok, 
  Lat. 
  43° 
  6'*6 
  Peking, 
  Lat. 
  39° 
  54'-3 
  

  

  Madison, 
  Wis. 
  43 
  4'6 
  Columbus. 
  O. 
  39 
  57*7 
  

  

  Nagasaki, 
  Lat. 
  32° 
  43' 
  -4 
  

  

  Macon, 
  Ga. 
  32 
  50 
  '4 
  

  

  San 
  Diego, 
  Cal. 
  32 
  43-1 
  

  

  If 
  these 
  stations 
  can 
  now 
  be 
  reoccupied 
  the 
  simultaneous 
  

   determination 
  of 
  latitudes 
  at 
  the 
  two 
  stations 
  composing 
  a 
  

   group 
  will 
  furnish 
  the 
  beginning 
  of 
  the 
  program 
  above 
  indi- 
  

   cated 
  while 
  the 
  latitudes 
  thus 
  derived 
  will 
  be 
  immediately 
  

  

  