﻿International 
  Unit 
  of 
  Candle 
  Power. 
  267 
  

  

  due 
  to 
  atmospheric 
  changes. 
  Hence, 
  in 
  the 
  table 
  given 
  later 
  

   on 
  in 
  the 
  paper 
  a 
  correcting 
  factor 
  has 
  had 
  to 
  be 
  assumed 
  in 
  

   cases 
  where 
  the 
  Carcel 
  lamp 
  is 
  corrected 
  for 
  a 
  difference 
  of 
  

   humidity. 
  

  

  Accuracy 
  of 
  Comparisons. 
  — 
  It 
  is 
  well 
  to 
  explain 
  in 
  giving 
  

   the 
  results 
  of 
  experiments 
  that 
  different 
  limits 
  of 
  accuracy 
  

   must 
  be 
  attribut(.'d 
  to 
  photometric 
  measurements 
  of 
  different 
  

   types 
  of 
  standards. 
  

  

  It 
  is 
  usual 
  in 
  giving 
  photometric 
  results 
  to 
  write 
  down 
  the 
  

   fourth 
  figure, 
  but 
  even 
  in 
  the 
  most 
  favourahle 
  circumstances 
  

   this 
  must 
  be 
  written 
  small 
  and 
  the 
  value 
  considered 
  liable 
  to 
  

   an 
  error 
  of 
  + 
  or 
  — 
  0*1 
  per 
  cent. 
  In 
  the 
  case 
  of 
  the 
  best 
  

   comparisons 
  of 
  electric 
  sub-standards 
  this 
  inaccuracy 
  should 
  

   not 
  be 
  occasioned 
  by 
  imperfection 
  in 
  the 
  bench 
  or 
  photo- 
  

   meter 
  head 
  nor 
  yet 
  to 
  the 
  electrical 
  measurements, 
  but 
  must 
  

   be 
  attributed, 
  in 
  the 
  author's 
  opinion, 
  chiefly 
  to 
  want 
  of 
  

   constancy 
  in 
  the 
  individual 
  who 
  is 
  making 
  the 
  photometric 
  

   observations. 
  

  

  In 
  some 
  of 
  the 
  comparisons 
  which 
  are 
  tabulated 
  the 
  

   electrical 
  measurements 
  are 
  probably 
  not 
  so 
  accurate 
  as 
  in 
  

   others. 
  The 
  fuller 
  appreciation, 
  however, 
  of 
  the 
  exact 
  values 
  

   of 
  the 
  national 
  and 
  international 
  electrical 
  units 
  which 
  has 
  

   recently 
  resulted 
  from 
  the 
  labours 
  of 
  the 
  International 
  Con- 
  

   ference, 
  makes 
  it 
  possible 
  now 
  to 
  attain 
  an 
  accuracy 
  which 
  

   leaves 
  nothing 
  to 
  be 
  desired 
  from 
  this 
  point 
  of 
  view. 
  

  

  As 
  matters 
  stand 
  now, 
  undoubtedly 
  the 
  photometric 
  com- 
  

   parisons 
  in 
  wdiich 
  the 
  highest 
  precision 
  is 
  attainable 
  are 
  

   those 
  between 
  properly 
  seasoned 
  electric 
  glow-lamps 
  of 
  the 
  

   same 
  coloured 
  light. 
  With 
  a 
  potentiometer 
  which 
  is 
  accurate 
  

   to 
  one 
  part 
  in 
  10,000 
  and 
  a 
  substitution 
  method 
  of 
  photo- 
  

   metric 
  comparison 
  * 
  on 
  a 
  bench 
  which 
  can 
  be 
  read 
  to 
  

   0*5 
  mm., 
  an 
  accuracy 
  is 
  attainable 
  with 
  a 
  set 
  of 
  good 
  sub- 
  

   standards 
  in 
  which 
  the 
  fourth 
  figure 
  is 
  almost 
  definite. 
  

  

  When, 
  on 
  the 
  other 
  hand, 
  comparisons 
  are 
  made 
  against 
  

   or 
  between 
  flame 
  standards, 
  the 
  probable 
  inaccuracy 
  is 
  

   greater. 
  

  

  How 
  much 
  the 
  inaccuracy 
  is 
  must 
  depend 
  largely 
  on 
  the 
  

   flame 
  adjustments 
  and 
  the 
  consistent 
  behaviour 
  of 
  the 
  

   standard 
  in 
  question. 
  It 
  also 
  depends 
  upon 
  the 
  accuracy 
  

   of 
  measurement 
  of 
  atmospheric 
  conditions 
  and 
  the 
  precise 
  

   knowledge 
  which 
  we 
  have 
  of 
  their 
  effect 
  on 
  the 
  light 
  of 
  the 
  

   standard 
  lamps. 
  

  

  It 
  follows 
  from 
  this 
  that 
  a 
  relatively 
  large 
  number 
  of 
  

   observations 
  must 
  be 
  made, 
  when 
  using 
  a 
  flame 
  standard, 
  

  

  * 
  See 
  '* 
  Photometry 
  of 
  Electric 
  Lamps," 
  by 
  Br. 
  J. 
  A. 
  Fleming, 
  M. 
  A., 
  

   F.R.S., 
  JouiTi, 
  Inst. 
  Elect. 
  Eng. 
  vol. 
  xxxii. 
  p/144. 
  

  

  