﻿68 
  Messrs. 
  M. 
  Greenwood, 
  C. 
  Hodson, 
  and 
  A. 
  E. 
  Tebb. 
  

  

  The 
  ends 
  of 
  the 
  shell 
  are 
  approached 
  in 
  an 
  uncomfortable 
  stooping 
  attitude, 
  

   there 
  are 
  many 
  dies 
  in 
  the 
  stamp 
  box 
  but 
  the 
  right 
  one 
  is 
  rapidly 
  selected. 
  

   The 
  marks 
  to 
  be 
  stamped 
  are 
  chalked 
  on 
  the 
  shell 
  and 
  great 
  accuracy 
  is 
  

   needed. 
  The 
  rate 
  is 
  seven 
  blows 
  in 
  10 
  seconds. 
  

  

  Labouring 
  {Sweeping 
  and 
  Barrow). 
  — 
  Steel 
  filings 
  are 
  removed 
  from 
  trays 
  under 
  

   the 
  lathes 
  with 
  large 
  shovels 
  and 
  placed 
  in 
  iron 
  barrows, 
  the 
  loaded 
  barrows 
  

   are 
  then 
  wheeled 
  away 
  and 
  the 
  contents 
  dumped 
  outside 
  the 
  shop. 
  We 
  have 
  

   also 
  included 
  ordinary 
  rubbish 
  sweepings, 
  which 
  seemed 
  to 
  be 
  equally 
  

   laborious 
  work. 
  This 
  is 
  the 
  only 
  type 
  of 
  work 
  seen 
  to 
  produce 
  sensible 
  

   perspiration 
  and, 
  as 
  will 
  appear, 
  the 
  experimental 
  results 
  confirm 
  the 
  opinion 
  

   that 
  a 
  lightly 
  regarded 
  and 
  relatively 
  ill-paid 
  task 
  is 
  physiologically 
  the 
  most 
  

   costly 
  of 
  all. 
  

  

  Walking 
  and 
  Carrying. 
  — 
  The 
  few 
  observations 
  under 
  this 
  heading 
  refer 
  

   to 
  persons 
  carrying 
  such 
  objects 
  as 
  boxes 
  of 
  bullets, 
  a 
  tea 
  tray, 
  gas 
  

   containers, 
  etc. 
  In 
  no 
  case 
  was 
  the 
  load 
  heavy 
  ; 
  perhaps 
  of 
  the 
  order 
  of 
  an 
  

   ordinary 
  coal 
  scuttle 
  or 
  a 
  marketing 
  basket. 
  

  

  Section 
  3. 
  

  

  At 
  this 
  point 
  it 
  will 
  be 
  advisable 
  to 
  touch 
  upon 
  the 
  material 
  accuracy 
  of 
  

   our 
  experiments, 
  merely 
  as 
  experiments, 
  quite 
  apart 
  from 
  the 
  interpretation 
  

   of 
  experiments 
  deemed 
  to 
  be 
  in 
  themselves 
  exact. 
  

  

  When 
  the 
  whole 
  volume 
  of 
  expired 
  air 
  is 
  analysed, 
  as 
  for 
  instance 
  in 
  the 
  

   research 
  of 
  Benedict 
  and 
  Cathcart 
  upon 
  the 
  metabolism 
  of 
  cyclists, 
  the 
  only 
  

   error 
  involved 
  is 
  due 
  to 
  inexact 
  graduation 
  (fineness 
  of 
  the 
  weighing 
  technic) 
  

   of 
  the 
  apparatus 
  or 
  defective 
  absorption 
  by 
  the 
  chemical 
  reagents. 
  In 
  the 
  

   sampling 
  method 
  errors 
  of 
  this 
  type 
  persist, 
  and 
  there 
  is 
  in 
  addition 
  a 
  series 
  

   of 
  errors 
  introduced 
  due 
  to 
  (a) 
  incorrect 
  measurement 
  of 
  the 
  volume 
  of 
  

   expired 
  air 
  — 
  whether 
  dependent 
  upon 
  faults 
  of 
  the 
  gas 
  meter 
  or 
  failure 
  to 
  

   express 
  the 
  bag, 
  or 
  again 
  to 
  inaccurate 
  measurement 
  of 
  temperature 
  and 
  

   barometric 
  pressure 
  ; 
  and 
  (b) 
  failures 
  of 
  technic 
  in 
  transferring 
  samples 
  to 
  the 
  

   gas 
  analyser 
  whereby, 
  in 
  particular, 
  contamination 
  with 
  room 
  air 
  may 
  be 
  

   produced. 
  With 
  respect 
  to 
  errors 
  of 
  class 
  (a), 
  in 
  our 
  earlier 
  experiments 
  the 
  

   measurement 
  of 
  temperature 
  was 
  imperfect 
  because 
  the 
  thermometer 
  was 
  

   not 
  inserted 
  into 
  the 
  meter's 
  inlet, 
  and 
  it 
  was 
  found, 
  on 
  subsequent 
  investiga- 
  

   tion, 
  that 
  the 
  temperature 
  of 
  the 
  expired 
  air 
  at 
  the 
  moment 
  of 
  its 
  volumetric 
  

   measurement 
  was 
  understated 
  to 
  the 
  extent 
  of 
  about 
  2° 
  C. 
  ; 
  this 
  error 
  was 
  

   subsequently 
  corrected. 
  So 
  far 
  as 
  the 
  volume 
  of 
  expired 
  air 
  is 
  concerned, 
  

   careful 
  experiment 
  showed 
  that 
  with 
  the 
  light 
  bag 
  used 
  in 
  most 
  of 
  the 
  

   experiments 
  any 
  error 
  due 
  to 
  incomplete 
  expression 
  of 
  the 
  contents 
  

   through 
  the 
  meter 
  was 
  very 
  small. 
  The 
  combined 
  effect 
  of 
  error 
  classified 
  

  

  