﻿444 
  ON 
  THE 
  RELATIVE 
  HORIZONTAL 
  INTENSITIES 
  

  

  are 
  similar 
  in 
  their 
  general 
  proportions. 
  The 
  effects 
  of 
  figure 
  and 
  

   of 
  relative 
  dimensions 
  seem 
  to 
  be 
  without 
  influence 
  upon 
  the 
  result. 
  

  

  In 
  all 
  cases 
  pains 
  were 
  taken 
  to 
  allow 
  the 
  needles 
  time 
  to 
  arrive 
  at 
  

   the 
  temperature 
  of 
  the 
  inclosure, 
  but 
  observations 
  made 
  at 
  intervals 
  

   during 
  the 
  heating 
  or 
  cooling 
  seem 
  to 
  show 
  that 
  this 
  precaution 
  was 
  

   not 
  essential. 
  

  

  These 
  observations 
  conclusively 
  show 
  the 
  importance, 
  and 
  indeed 
  

   the 
  necessity, 
  of 
  determining 
  a 
  specific 
  correction 
  to 
  be 
  applied 
  to 
  each 
  

   needle 
  used 
  in 
  a 
  series 
  of 
  observations 
  for 
  intensity. 
  They 
  confirm 
  in 
  

   this 
  respect 
  conclusions 
  to 
  which 
  the 
  observations 
  of 
  captain 
  Sabine 
  

   for 
  obtaining 
  the 
  same 
  correction, 
  would 
  seem 
  to 
  lead. 
  The 
  variation 
  

   of 
  these 
  coefficients 
  from 
  each 
  other, 
  as 
  well 
  as 
  from 
  those 
  of 
  captain 
  

   Sabine, 
  and 
  from 
  that 
  quoted 
  as 
  having 
  been 
  determined 
  by 
  professor 
  

   Hansteen, 
  agrees 
  in 
  the 
  conclusion 
  to 
  which 
  they 
  lead. 
  It 
  is 
  the 
  more 
  

   necessary 
  to 
  call 
  special 
  attention 
  to 
  this 
  point, 
  because 
  the 
  coefficient 
  

   of 
  professor 
  Hansteen 
  has 
  been 
  applied 
  in 
  the 
  reduction 
  of 
  the 
  observa- 
  

   tions 
  (Royal 
  Soc. 
  Trans. 
  1828) 
  for 
  the 
  relative 
  intensities 
  at 
  Paris, 
  

   London 
  and 
  Edinburgh, 
  and 
  more 
  recently 
  in 
  a 
  very 
  extensive 
  series 
  

   of 
  observations 
  by 
  M. 
  Quetelet 
  of 
  Brussels, 
  whose 
  activity 
  in 
  this 
  

   branch 
  has 
  of 
  late 
  years 
  been 
  particularly 
  prominent. 
  

  

  A 
  further 
  inference 
  may 
  be 
  deduced 
  from 
  these 
  observations, 
  viz. 
  

   that 
  a 
  sensible 
  change 
  in 
  the 
  magnetic 
  state 
  of 
  a 
  bar, 
  will 
  be 
  attended 
  by 
  

   a 
  change 
  in 
  the 
  correction 
  to 
  be 
  applied 
  for 
  temperature. 
  So 
  that 
  a 
  cor- 
  

   rection 
  once 
  obtained 
  should 
  not 
  be 
  used 
  after 
  such 
  a 
  change 
  has 
  taken 
  

   place 
  in 
  any 
  considerable 
  degree. 
  In 
  an 
  extensive 
  series 
  of 
  observa- 
  

   tions, 
  it 
  would 
  therefore 
  be 
  necessary 
  to 
  investigate 
  this 
  correction 
  

   during 
  the 
  time 
  of 
  making 
  the 
  observations, 
  or 
  before 
  the 
  series 
  was 
  

   commenced 
  and 
  after 
  its 
  completion. 
  

  

  In 
  applying 
  the 
  correction 
  for 
  temperature, 
  it 
  is 
  convenient, 
  and 
  

   generally 
  admissible, 
  to 
  take 
  for 
  the 
  multiplier 
  of 
  the 
  coefficient 
  just 
  

   determined, 
  a 
  mean 
  time 
  of 
  vibration, 
  instead 
  of 
  the 
  actual 
  time 
  in 
  a 
  

   given 
  case. 
  

  

  That 
  is, 
  to 
  take 
  for 
  the 
  value 
  of 
  T, 
  

  

  T 
  = 
  T' 
  — 
  T" 
  . 
  m 
  (f 
  — 
  /), 
  

   where 
  T" 
  represents 
  the 
  mean 
  time 
  referred 
  to. 
  When 
  the 
  correction 
  

  

  