﻿W. 
  P. 
  White 
  — 
  Specific 
  Heat 
  Determination. 
  53 
  

  

  evaporation 
  strongly 
  preponderated; 
  it 
  caused 
  in 
  one 
  

   case 
  a 
  fall 
  of 
  0-005° 
  per 
  minute 
  with 
  the 
  room 
  3° 
  warmer, 
  

   and 
  from 
  0-009° 
  to 
  0-014° 
  with 
  the 
  room 
  4° 
  colder. 
  The 
  

   lead-bath 
  furnace 
  at 
  500° 
  also 
  affected 
  the 
  calorimeter 
  

   noticeably, 
  raising 
  the 
  temperature 
  in 
  one 
  case 
  nearly 
  

   0-002° 
  per 
  minute 
  with 
  the 
  calorimeter 
  closed, 
  and 
  0-004 
  

   with 
  it 
  open. 
  Here 
  the 
  greater 
  absorption 
  of 
  radiation 
  

   by 
  the 
  water 
  overcame 
  the 
  evaporation. 
  It 
  follows 
  from 
  

   these 
  data 
  that 
  exposure 
  to 
  air 
  4° 
  cooler 
  for 
  from 
  7 
  to 
  10 
  

   seconds 
  would 
  cause 
  a 
  fall 
  of 
  nearly 
  0-002°, 
  and 
  from 
  

   0-0015° 
  to 
  0-0021° 
  were 
  actually 
  observed 
  in 
  5 
  blank 
  

   drops 
  with 
  cold 
  furnace. 
  Under 
  the 
  furnace 
  at 
  500°, 
  a 
  

   rise 
  of 
  0-0012° 
  or 
  more 
  was 
  observed 
  in 
  blank 
  drops, 
  

   which 
  is 
  a 
  little 
  more 
  than 
  would 
  be 
  expected. 
  The 
  error 
  

   is 
  only 
  the 
  variation 
  in 
  these 
  effects, 
  and 
  apparently 
  

   should 
  generally 
  be 
  under 
  0-001°, 
  which 
  is 
  only 
  0-3 
  per 
  

   mille 
  of 
  the 
  smallest 
  heats 
  with 
  the 
  furnaces, 
  those 
  at 
  

   300°. 
  At 
  100° 
  the 
  results, 
  as 
  will 
  presently 
  appear, 
  show 
  

   that 
  the 
  exposure 
  error 
  was 
  certainly 
  not 
  over 
  0-001°, 
  

   and 
  probably 
  not 
  over 
  0-0003°. 
  

  

  The 
  thermometric 
  sensitiveness 
  employed 
  in 
  most 
  of 
  

   the 
  work 
  was 
  that 
  of 
  24 
  copper-constantan 
  couples 
  read 
  

   to 
  0-1 
  microvolt 
  or 
  0-0001°, 
  three 
  times 
  as 
  good 
  as 
  in 
  the 
  

   work 
  of 
  1910. 
  This 
  was 
  several 
  times 
  as 
  good 
  as 
  the 
  

   average 
  final 
  precision 
  actually 
  attained 
  in 
  the 
  most 
  

   favorable 
  case. 
  Nevertheless 
  it 
  seems 
  to 
  have 
  been 
  well 
  

   worth 
  while 
  to 
  employ 
  such 
  sensitiveness. 
  It 
  was 
  of 
  the 
  

   very 
  greatest 
  advantage, 
  in 
  studying 
  the 
  efficiency 
  of 
  the 
  

   installation 
  and 
  investigating 
  the 
  puzzling 
  errors 
  encoun- 
  

   tered, 
  to 
  be 
  free 
  from 
  all 
  apprehension 
  of 
  appreciable 
  

   error 
  in 
  the 
  temperature 
  measurement, 
  while 
  the 
  obser- 
  

   vations 
  could 
  hardly 
  have 
  been 
  easier 
  or 
  more 
  certain, 
  

   even 
  with 
  considerably 
  less 
  sensitive 
  apparatus 
  of 
  other 
  

   types. 
  The 
  essentials 
  of 
  the 
  thermoelectric 
  method 
  used 
  

   have 
  already 
  been 
  described. 
  13 
  

  

  A 
  surprising 
  result 
  appeared 
  in 
  connection 
  with 
  the 
  

   time 
  needed 
  for 
  the 
  charge 
  to 
  come 
  to 
  equilibrium 
  in 
  the 
  

   calorimeter. 
  Ordinarily, 
  in 
  texts 
  or 
  articles 
  on 
  calorime- 
  

   try, 
  it 
  is 
  said 
  that 
  one 
  must 
  wait 
  till 
  the 
  cooling 
  is 
  

   "regular" 
  before 
  taking 
  the 
  final 
  reading 
  of 
  the 
  transfer 
  

   period. 
  This 
  rule 
  seems 
  inexact 
  and 
  may 
  prove 
  mislead- 
  

   ing, 
  since 
  the 
  variation 
  from 
  "regular" 
  cooling 
  usually 
  

   becomes 
  imperceptible 
  before 
  a 
  true 
  equilibrium 
  is 
  

  

  "Easy 
  Calorimetric 
  Methods 
  of 
  High 
  Precision, 
  loc. 
  cit., 
  p. 
  2327. 
  

  

  