﻿428 
  ON 
  THE 
  RELATIVE 
  HORIZONTAL 
  INTENSITIES 
  

  

  In 
  the 
  present 
  state 
  of 
  the 
  experimental 
  part 
  of 
  this 
  branch 
  of 
  

   science, 
  we 
  are 
  induced 
  to 
  present 
  these 
  observations, 
  not 
  only 
  as 
  show- 
  

   ing 
  the 
  horizontal 
  intensities 
  at 
  the 
  different 
  places 
  of 
  observation, 
  but 
  

   on 
  account 
  of 
  the 
  deductions 
  warranted 
  by 
  the 
  investigation 
  of 
  cor- 
  

   rections 
  for 
  temperature, 
  and 
  the 
  comparison 
  afforded 
  between 
  the 
  

   methods 
  of 
  oscillation 
  in 
  full 
  air, 
  and 
  in 
  a 
  rarefied 
  medium. 
  This 
  lat- 
  

   ter 
  mode 
  is 
  so 
  great 
  an 
  improvement, 
  that 
  its 
  claims 
  cannot 
  be 
  too 
  

   strongly 
  urged. 
  

  

  The 
  most 
  numerous 
  observations 
  of 
  the 
  first 
  series 
  above 
  referred 
  to, 
  

   were 
  made 
  at 
  Philadelphia 
  and 
  West 
  Point. 
  These 
  presented 
  discre- 
  

   pancies 
  which 
  we 
  were 
  entirely 
  at 
  a 
  loss 
  to 
  explain. 
  

  

  The 
  differences, 
  in 
  the 
  time 
  of 
  ten 
  vibrations, 
  as 
  observed 
  on 
  different 
  

   occasions 
  at 
  the 
  same 
  place, 
  were 
  of 
  an 
  amount 
  entirely 
  beyond 
  what 
  

   could 
  have 
  resulted 
  from 
  errors 
  of 
  observation, 
  which 
  the 
  method 
  of 
  

   observing 
  rendered 
  quite 
  small. 
  Pains 
  had 
  been 
  taken, 
  in 
  all 
  cases, 
  

   to 
  remove 
  magnetic 
  matter 
  which 
  might 
  have 
  affected 
  the 
  results 
  ; 
  and 
  

   in 
  other 
  respects 
  the 
  observations 
  were 
  carefully 
  made. 
  

  

  Some 
  part 
  of 
  the 
  discrepancies 
  might 
  be 
  charged 
  to 
  the 
  errors 
  in 
  

   the 
  time-pieces 
  used. 
  The 
  observations 
  at 
  West 
  Point 
  were 
  made 
  

   however, 
  with 
  a 
  chronometer 
  of 
  undoubted 
  reputation, 
  and 
  most 
  of 
  

   those 
  at 
  Philadelphia, 
  with 
  chronometers 
  of 
  good 
  standing; 
  and 
  the 
  

   discrepancies 
  could 
  not 
  thus 
  be 
  satisfactorily 
  accounted 
  for. 
  

  

  The 
  magnetic 
  needles 
  used 
  had 
  not 
  changed 
  their 
  magnetism 
  during 
  

   the 
  series 
  of 
  observations, 
  and 
  were 
  thus 
  shown 
  to 
  be 
  sufficiently 
  hard 
  

   to 
  retain 
  their 
  charge. 
  It 
  was 
  not 
  probable, 
  therefore, 
  that 
  they 
  would 
  

   readily 
  change, 
  from 
  day 
  to 
  day, 
  their 
  magnetic 
  state, 
  unless 
  from 
  

   changes 
  in 
  the 
  earth's 
  magnetism, 
  or 
  from 
  heat. 
  A 
  correction 
  for 
  

   temperature 
  was 
  ascertained 
  by 
  experiment, 
  and 
  the 
  results 
  show 
  that 
  

   this 
  correction 
  was 
  accurate. 
  Without 
  it 
  the 
  amount 
  of 
  the 
  differences 
  

   is 
  very 
  much 
  increased. 
  

  

  It 
  is 
  possible, 
  that 
  in 
  some 
  of 
  the 
  earlier 
  observations 
  in 
  the 
  first 
  

   series, 
  sufficient 
  care 
  was 
  not 
  taken 
  to 
  allow 
  the 
  needles 
  to 
  attain 
  the 
  

   temperature 
  of 
  the 
  surrounding 
  medium, 
  a 
  source 
  of 
  error 
  which 
  was 
  

   guarded 
  against 
  in 
  the 
  later 
  results. 
  

  

  The 
  discrepancies 
  seemed 
  to 
  us 
  to 
  show, 
  either 
  that 
  there 
  was 
  some 
  

   imperfection 
  in 
  the 
  apparatus, 
  other 
  than 
  those 
  already 
  enumerated, 
  or 
  

  

  