Skip to main content

Full text of "Lectures on illuminating engineering delivered at the Johns Hopkins University, October and November, 1910"

See other formats


This  is  a  digital  copy  of  a  book  that  was  preserved  for  generations  on  library  shelves  before  it  was  carefully  scanned  by  Google  as  part  of  a  project 

to  make  the  world's  books  discoverable  online. 

It  has  survived  long  enough  for  the  copyright  to  expire  and  the  book  to  enter  the  public  domain.  A  public  domain  book  is  one  that  was  never  subject 

to  copyright  or  whose  legal  copyright  term  has  expired.  Whether  a  book  is  in  the  public  domain  may  vary  country  to  country.  Public  domain  books 

are  our  gateways  to  the  past,  representing  a  wealth  of  history,  culture  and  knowledge  that's  often  difficult  to  discover. 

Marks,  notations  and  other  maiginalia  present  in  the  original  volume  will  appear  in  this  file  -  a  reminder  of  this  book's  long  journey  from  the 

publisher  to  a  library  and  finally  to  you. 

Usage  guidelines 

Google  is  proud  to  partner  with  libraries  to  digitize  public  domain  materials  and  make  them  widely  accessible.  Public  domain  books  belong  to  the 
public  and  we  are  merely  their  custodians.  Nevertheless,  this  work  is  expensive,  so  in  order  to  keep  providing  tliis  resource,  we  liave  taken  steps  to 
prevent  abuse  by  commercial  parties,  including  placing  technical  restrictions  on  automated  querying. 
We  also  ask  that  you: 

+  Make  non-commercial  use  of  the  files  We  designed  Google  Book  Search  for  use  by  individuals,  and  we  request  that  you  use  these  files  for 
personal,  non-commercial  purposes. 

+  Refrain  fivm  automated  querying  Do  not  send  automated  queries  of  any  sort  to  Google's  system:  If  you  are  conducting  research  on  machine 
translation,  optical  character  recognition  or  other  areas  where  access  to  a  large  amount  of  text  is  helpful,  please  contact  us.  We  encourage  the 
use  of  public  domain  materials  for  these  purposes  and  may  be  able  to  help. 

+  Maintain  attributionTht  GoogXt  "watermark"  you  see  on  each  file  is  essential  for  in  forming  people  about  this  project  and  helping  them  find 
additional  materials  through  Google  Book  Search.  Please  do  not  remove  it. 

+  Keep  it  legal  Whatever  your  use,  remember  that  you  are  responsible  for  ensuring  that  what  you  are  doing  is  legal.  Do  not  assume  that  just 
because  we  believe  a  book  is  in  the  public  domain  for  users  in  the  United  States,  that  the  work  is  also  in  the  public  domain  for  users  in  other 
countries.  Whether  a  book  is  still  in  copyright  varies  from  country  to  country,  and  we  can't  offer  guidance  on  whether  any  specific  use  of 
any  specific  book  is  allowed.  Please  do  not  assume  that  a  book's  appearance  in  Google  Book  Search  means  it  can  be  used  in  any  manner 
anywhere  in  the  world.  Copyright  infringement  liabili^  can  be  quite  severe. 

About  Google  Book  Search 

Google's  mission  is  to  organize  the  world's  information  and  to  make  it  universally  accessible  and  useful.   Google  Book  Search  helps  readers 
discover  the  world's  books  while  helping  authors  and  publishers  reach  new  audiences.  You  can  search  through  the  full  text  of  this  book  on  the  web 

at|http: //books  .google  .com/I 




University  of  California. 















October  and  November,   1910 




BALtlMOM,  Md. 


>y  Google 

CoPTRiOHt,  1911,  ir 

Ztt  Bttb  i^aftiinoii  tpttM* 

>y  Google 


This  Course  of  Lectures  on  lUuminatiog  EugiDeering  was  giren 
at  the  Johns  Hopkins  University,  Baltimore,  between  the  dates 
October  26  and  November  8,  1910,  nnder  the  joint  auspices  of  the 
University  and  the  Illuminating  Engineering  Society.  The  origin 
and  objects  of  the  lectures  are  clearly  stated  in  the  preliminary 
announcement  of  the  course,  from  which  the  following  quotation 
is  made: 

"  The  Illuminating  Engineering  Society  recognizing  the  fact 
that  there  is  an  increasing  demand  for  trained  illuminating  engi- 
neers, and  that  the  present  facilities  available  for  the  specialized 
instruction  required  are  inadequate,  determined,  through  an  act 
of  the  Council  of  the  Society,  to  encourage  the  establishment  of  a 
course  of  lectures  on  the  subject  of  illuminating  engineering.  This 
course  should  have  three  objects:  (1)  to  indicate  the  proper  coordi- 
nation of  those  arts  and  sciences  which  constitute  illuminating  engi- 
neering; (2)  to  furnish  a  condensed  outline  of  study  suitable  for 
elaboration  into  an  undergraduate  course  for  introduction  into  the 
curricula  of  undergraduate  technical  schools;  and  (3)  to  give 
practising  engineers  an  opportunity  to  obtain  a  conception  of  the 
science  of  illuminating  engineering  as  a  whole. 

"  Inasmuch  as  such  a  course  is  most  appropriately  given  at  a 
university  where  graduate  instruction  is  emphasized,  and  as  the 
Johns  Hopkins  University  has  regularly  offered  courses  by  non- 
resident lecturers  as  part  of  ita  system  of  instruction,  and  is  now 
preparing  to  extend  its  graduate  work  into  applied  sciences  and 
engineering,  an  arrangement  has  been  effected  by  which  the  lectures 
will  be  given  at  this  University  under  the  joint  auspices  of  the 
University  and  the  Illuminating  Engineering  Society.  The  sub- 
jects and  scope  of  the  lectures  have  been  proposed  by  the  Society 
and  approved  by  the  University.  The  lecturers  have  been  invited 
by  the  University  upon  the  advice  of  the  Society." 

The  lectures  were  attended  by  840  men  from  various  parts  of 
the  United  States,  many  of  them  representatives  of  technical 
schools,  gas  and  electric  central  stations,  and  manufacturing  com- 



vi  Pbefacb 

panies.  A  large  nnmbeT  of  the  attendants  at  the  lectures  ako 
followed  the  course  of  laboratory  work  which  had  been  arranged. 
The  general  inteieet  in  the  course  encourages  the  hope  that  these 
published  Tolumes  may  serve  to  advance  our  knowledge  of  this  new 
and  important  branch  of  engineering. 

>y  Google 

Volume  I 


L  Tbm  Physical  Basib  or  tkb  PsoDU<nios  of  Lisht.    Three 

lectureg    1 

JoevH  B.  AuES,  Ph.  D.,  Profeeaor  of  FhrslCB  ftnd  Direc- 
tor  of   the   Pbyslcal   LHboratory,   Tb«   Jolins   Hopkins 
II.   Thb  Phtsicai.  Chabactkbistios  or  LcMiNous  SoiTBCEs.    Two 

lecture* 3C 

EmVABD  P.  Htdb,  Ph.  D.,  President,  Illuminating  Bn- 
glneerlng  Sodetr;  Director  of  Pliyaical  Laboratorr, 
National  Electric  Lamp  Assoclatlnn. 

HI.  Thk  Chcuistbt  or  LTTHntOQa  SoiiBciCB.    One  lecture 93 

WiLus  R.  Whithkt,  Pb.  D.,  Director  of  Reaeardi  Labo- 
rabny.  General  Electric  Co.;  Past  President,  American 
Chemical  Sodetr- 

IV.  Blectsic  Illdkihaktb.    Two  lecture* 10& 

Chauxb    P,    STxqracBiz,    Ph.  D.,   Consulting   Bnglneer, 
General  Electric  Co.;   Professor  at  Electrical  Engineer- 
Ing,  Union  University. 
V.   il)    Oas   aho  On.  iLLumiTAirra,    (3)    iNOAKimciifT   Gas 

MAioTLEa.    Tv>o  lectures IGT 

(1)  AixzANi^  C.  HumRBETs,  M.  B^  Hon.  Sc.  D.,  Presi- 
dent of  BtaTens  Institute  of  Technology;  Past  Preaident 
American  Gas  Institute. 

(S)  H.  C.  WnrrAEBB,  B.  S.,  H.  8.,  Professor  of  Industrial 
Cbemlatry,  Colombia  Unlversltr. 
VI.  Thi  Gm^uTioN  Am  DisTBinrTiON  or  BLxcTBicrrT  with 

Special  Retebenci  v)  liiaHTiNa.    Two  lecture* 331 

John  B.  Whiteheas,  Ph.  D.,  Professor  of  Applied  Elec- 
tricity, The  Johns  Hoiking  UnWenjIty. 
TIL  The  UAiruPACTDEE  Ain>  Dia^msunoH  or   AKnnciAi.  Gas 

with  Si>BciAL  Betebehce  to  LiGHTiRa.    Txoo  iectureB 277 

(1)    Ml   E.   G.   Cowd^t,  Tice-President,   Peoples   Gas 
Ught  and  Coke  Company,  Chicago,  Til. 
<2)  Ma.  WALTn  R.  Addickb,  Vice-President,  Consolidated 
Gas  Co.,  New  York, 

Till.   PnoTOMnsio  Uitits  and  8rAin>ABDB.    One  lecture 887 

BnwAiD  B.  Rosa,  Ph.  D.,  Physicist,  National  Bureau  of 

>y  Google 

ii  General  Contents 

IX.  Tbb  HsASTiBCMEnT  or  Light.    Two  lectures 411 

Ci^ATTOR  H.  Shabp,  Pb.  D.,  Test  OfBcer,  Blectrical  Test- 
Ins  LftborataiT,  New  York  City;  Past  PreBident,  Illnml- 
natlng  Engineer Ing  Society. 

X.  The  AECK1TECT08AL  Abpecis  or  Illuminatino  BnoiNBBB- 

iBQ.     One  lecture  607 

Walteb  Cook,  A.  M.,  Vice-President,  American  Institute 
of  Arcbltects;  Past  President,  Society  of  Beaux  Arts 

XI.  Thb  Fhtsioumical  Asracra  or  IixirMiNATino  BNoinEEBiNO. 

Two  lecture* 525 

P.  W.  Cobb,  B.  S.,  M.  D.,  Physiologist.  Physical  Labora- 
tory. National  Electric  Lamp  Association. 

XII-  The  Pstchologioal  Asfeots  or  Iixumihatiko  Esoinkeriso. 

One  lecture  575 

RoBEBT  M.  Tebkes,  Ph.  D.,  AselBtant  Professor  of  Psy- 
chology, Harvard  University. 

XIII.   The  Pbinciplgs  and  Dbbioh  of  Intebiob  Iixuuination.    BUe 

lectwet    605 

(1)  W.  B.  Babbowb,  Jb.,  Assistant  Professor  Electrical 
Engineering,  Armour  Institute  of  Technology,  Chicago. 

(2)  L.  B.  Habkb,  B.  S.,  M.  M.  E.,  Consulting  Engineer, 
New  York  Clt^;  Past  President,  Illuminating  Engineer- 
ing Society. 

(S)  Mb.  Nobmait  Macbeth,  Illnmtnatlng  Engineer,  The 
Welsbach  Co. 

XIV.  The  Pbircipixs  and  Desior  of  DrTraios  Illumination. 

Three  lectures   796 

(1)  Lome  Bell,  Ph.D.,  Consulting  Engineer,  Boston, 
Mass.;  Past  President,  lUumtnatlng  Engineering  Society. 

(2)  B.  N.  Wbiohtisoton,  A.  B.,  Boston  Consolidated' 
Qas  Co. 

XV.  Shades,  Reflectobs  and  DiTFUsiNa  Media.    One  lecture. . .     885 
Van    Rensselaxb    LanbikoHi    B.  &.,    General    Manager 
Holophane  Co. 

XVI.  LioHTiNe  FixTUKBs.    One  lecture 931 

Mb.  Edwabd  F.  Caldwell.  Senior  Member  of  Firm  and 
Designer,  Edward  F.  Caldwell  A  Co.,  New  York. 

>y  Google 

Geneeal  Contents  ii 

XVII.  The  CoHiiEBciAi.   Abprctb  of  EIlbctbic   LieHTina.     One 

lecture    946 

John  W.  Lieb,  Ja.,  M.  B.,  Third  Vice-President,  New 
York  Sdlson  Co. ;  Past  President,  American  Institute  of 
Electrical  Engineers. 

XVIII.  The  Comuebciai.  Aspect  of  Oab  Business  wrrs  Special 

Retxbence  to  Oas  Lightino.    0ns  lecture 1009 

Walton  Cubk,  M.  E.,  President  of  The  Franklin  Instl- 
tnte,  Philadelphia;  Third  Vice-President,  United  Oas 
Improvement  Co..  Philadelphia, 


Lists  of  experiments  In  connection  with  the  Lecture  Course,  to- 
together  with  the  necessary  bibliographies 1011 

Charus  O.  Bohd,  Manager  of  Photometric  Laboratory,  United 

Gas  Improvement  Co.,  Philadelphia. 
Hebbebt    E.    Itbs,    Ph.    D.,    Physicist,    Physical    Laboratory, 

National  Electric  Idjnp  Association. 
Pbeston  S.  MiiXAB,  Electrical  Testing  Laboratory,  New  York. 
A.  H.  PruND,  Ph.  D.,  Associate  In  Pbysles,  The  Johns  Hopkins 

Index 1047 

>y  Google 





By  Joseph  S.  Ames 



PhyaUsal  Quantitiei  and  MeaturemMOa 
Objects  and  general  principles  of  pbystcs. 
Uethoda  of  assigning  numbers  to  iriiyilcal  quanUtte>. 

a.  Uesaurement  In  tenns  of  nnlti. 

b.  Indirect  means,  e.  g.,  temperature. 
Simple  Ideas. 

a.  Intnltlve:  space  and  time. 

b.  Experimental:  e.  g.,  force  (Illustrated  by  properties  of  matter), 
Units  of  length,  of  time,  of  force;  C.  O.  S.;  English. 

Derived  mechanical  quantities,  and  their  units;  e,  g.,  density,  proMnr& 

Heasnrement  of  length,  volume,  time,  force,  pressure. 

Errors  of  Instruments. 

Discussion  of  observations. 

Definition  of  electrical  quantities,  and  their  units. 

Heasurement  of  electrical  guantltles  by  portable  InstrumeDta. 


Energy  and  Thermal  PAenomena 

Definition  of  work  and  energy;  mechanical  Illustrations. 

Our  temp^^tnre  sense.    Thermal  phenomena. 
Thermal  effects. 

Methods  of  producing  these  effects. 
Explanation  In  terms  of  energy. 

Meaning  of  "  Conservation  of  Energy." 
Illustrations:  battery,  dynamo,  etc 

DiBcnsalon  of  temperature  and  Its  "  measurement." 

Discussion  of  modes  of  producing  beat-effects:  flames,  friction,  conduc- 
tion, radiation,  etc 

Radlatlcm  and  absorption:    KlrchhofTs  law,  "  Black  Body." 

Measurement  of  energy. 

a.  Rise  In  temperature. 

b.  Mechanical  means. 

c  Electrical  method:    Bit 

*  The    lectures    are    based    upon    the    author's    text-book    "  General 
Physics,"  published  by  the  American  Book  Co.,  New  York. 

>y  Google 

Illuminatinq  Enqineebinq 

Spectra  of  radiation. 

Dlsperelye  apparatus. 

Detecting  and  meaBurlng  apparatus. 

Visible,  ultra-violet,  tofra-red  radiation. 

ContlDuous,  dlscontlttuouB.  and  absorption  spectra. 
Modes  of  producing  radiation. 

a.  "  Temperature-radiation." 

b.  Luminescence:    fluorescence,  electrical  discharge,  etc. 
Color  sensation. 

Canse  of  color  of  natural  objects. 

a.  Body  absorption. 

b.  Surface  absorption. 

c.  Exceptional  caaes. 

Elxtenalon  or  temperature  acaloe  b;  radiation  methods. 

LeCTdee  I 
Physical  Qvantities  and  Measurements 

Matter.  Through  out  various  seuBes,  Buch  as  those  of  sight  aud 
hearing,  we  are  constantly  receiving  seneationa  which  we  interpret 
objectively;  i.  e.,  we  locate  the  cauee  of  a  sensstion  in  a  definite 
jwrtion  of  Bpace.  We  picture  to  ourBelvee  the  existence  there  of 
something  which  we  call  "matter";  and  to  a  limited  portion  of 
space  which  contains  matter  we  give  the  name  "  physical  body." 
Matter  may  be  divided  into  two  great  claeses:  that  which  is  living, 
such  as  plants  and  animals,  and  that  which  is  not,  such  as  pieces 
of  wood  and  glass,  water  and  air.  Physics  is,  broadly  speaking, 
'  the  science  concerned  with  this  second  division  of  matter,  which 
may  be  called  "  ordinary  matter  " ;  and  phenomena  occurring  in  con- 
nection with  matter  of  this  kind  are  called  "physical  phenomeiia," 

The  scientific  study  of  a  subject  involves  three  distinct  ideas; 
the  discovery,  the  investigation,  and  the  explanation  of  phenomena. 
The  first  two  require  no  discussion  here;  but  it  may  be  well  to 
state  that  by  the  words  "  to  explain  a  phenomenon  "  is  meant  to 
determine  its  exact  connection  with  other  phenomena,  to  describe 
it  in  terms  of  simpler  ones,  and  in  this  manner  to  reduce  the 
number  of  fundamental  ideas  as  far  as  possible. 

In  seeking  for  explanations  of  phenomena  we  assume  either 
directly  or  indirectly,  that  there  is  a  definite  connection  between 
consecutive  events,  of  such  a  nature  that  if  we  are  able  to  reproduce 
exactly  a  definite  condition,  the  same  effect  will  follow  regardless 

>y  Google 

The  Physical  Basis  op  the  Productiok  of  Light        3 

of  the  epoch  of  time  or  the  location  in  space.  We  are  justified  in 
this  belief  by  all  of  our  esperience  and  observationB. 

Ether.  The  careful  study  of  the  phenomena  of  light  led  philoso- 
pherE,  many  years  ago,  to  believe  that  there  is  present  in  space 
another  medium  for  phenomena  than  that  furnished  by  ordinary 
matter.  It  has  become  an  accepted  fact  that  tliroughout  the  vast 
regions  of  space,  in  the  solar  system  and  beyond,  there  is  a  medium 
permeating  all  ordinary  matter  and  having  many  properties  in 
common  with  matter  and  yet  not-  identical  with  it.  This  is  called 
"  the  ether."  In  order  to  explain  many  electrical  and  magnetic 
phenomena,  and  even  to  describe  the  phenomena  of  radiation,  it 
is  necessary  to  assume  its  existence.* 

Vhynai.  The  object  of  physics  may  therefore  be  defined  to  be 
the  attempt  to  determine  the  exact  connection  between  phenomena, 
both  in  ordinary  matter  and  in  the  ether,  and  to  express  these 
relations  with  as  few  hypotheses  as  possible  concerning  the  nature 
and  properties  of  either. 

Physioal  Qnantitiea.  A  physical  quantity  is  one  which  we  can 
imagine  as  capable  of  changing  in  amount,  something  to  which  we 
can  assign  a  numerical  value.  Some  quantities  can  be  measured, 
others  cannot.  To  measure  a  quantity,  another  similar  one  must 
first  be  chosen  as  a  standard  or  i^nit,  and  then  the  number  of 
times  this  is  contained  in  the  original  quantity  is  its  measure. 
Thus,  a  length  can  be  measured  in  terms  of  an  inch,  a  yard,  a 
centimeter,  etc.,  depending  upon  the  choice  of  unit.  It  is  possible 
to  understand  the  meaning  of  a  zero  value  of  any  measurable  quan- 
tity; further,  two  or  more  measurable  quantities  of  the  same  kind, 
for  instance  two  lengths,  may  be  added.  On  the  other  hand  there 
are  many  physical  quantities  which  cannot  be  measured;  and  yet 
it  is  possible  to  give  them  numerical  values.  Thus,  the  temperature 
of  a  body  cannot  be  measured,  although  it  is  possible  by  measuring 
the  change  in  volume  of  mercury  in  a  thermometer  to  give  a  num- 
ber to  temperature. 

Simple  Quantities.  To  most  physical  quantities  e^act  definitions 
can  be  given,  but  there  are  a  few  for  which  this  is  impossible;  there 
are  no  simpler  ideas  in  terms  of  which  we  can  describe  them.  The 
question  as  to  the  exact  number  of  these  need  not  be  discussed 
here,  and  in  what  follows  the  philosophy  based  upon  Kant  will  be 

*  One  Bbould  add  that  a  new  scbool  of  philosophy  ezleta  which  looks 
at  nature  from  a  different  Btandpolnt 

>y  Google 

4  Illithinatinq  Enoinebeinq 

accepted.  According  to  this  we  divide  our  simple  ideas  into  two 
dBBses;  intuitive  and  experimental.  The  two  intuitive  ideas  are 
those  concerned  with  epace  and  time. 

1.  A  straight  line,  a  polygon,  or  a  solid  figure  bodnded  by  plane 
faces,  together  with  the  ideas  involved  in  assigning  numerical  values 
to  lengths,  areas  and  volumes  are  conddered  intuitive.  That  is, 
it  is  impossible  to  define  what  is  meant  by  length ;  and  the  idea  of 
two  equal  lengths  admits  of  no  ambiguity.  We  can  choose  a  unit 
length  arbitrarily  and  then,  making  use  of  a  method  of  super- 
position, determine  the  number  to  be  given  any  length.  The  same 
general  metiiod  may  be  applied  to  areas  and  volumes. 

2.  In  regard  to  time,  we  have  a  definite  conception  of  what  is 
meant  by  two  equal  intervals  of  time;  certain  physical  phenomena 
appear  to  ub  to  repeat  themselves  at  intervals  of  time  apparently 
equal,  e.  g.,  the  vibrations  of  a  pendulum  or  the  balance  wheel  of 
a  watch.  We  have  no  way  by  which  we  can  prove  that  these  inter- 
vals are  equal,  yet  there  is  every  reason  for  believing  that  these 
motions  of  a  pendulum  and  of  the  balance  wheel  of  a  watch  are 
exactly  periodic;  for  at  any  instant  the  external  conditions  affecting 
the  motion  are  exactly  the  same,  so  far  as  we  can  tell,  as  they  were 
at  a  definite  interval  of  time  before.  In  order  to  give  a  number 
to  an  instant  of  time  one  must  choose  some  periodic  motion  such 
as  just  described,  e.  g.,  a  certain  pendulum  vibrating  under  definite 

.  conditions,  and  some  arbitrary  epoch  of  time  from  which  to  count 
the  number  of  vibrations;  the  number  of  vibrations  between  the 
epoch  and  the  instant  for  which  a  number  is  desired  is  this  number. 
Among  the  fundamental  ideas  of  which  we  learn  by  means  of  our 
sens^  may  be  mentioned  temperature,  pitch  of  sound,  and  what  we 
call  "  force."  For  instance,  through  our  muscular  sense  we  become 
conscious  of  certain  definite  sensations  when  with  our  hands  or 
arms  or  bodies  we  perform  certain  experiments  on  matter.  Thus, 
if  a  large  stone  is  held  in  the  hand  we  become  conscious  of  a  cer- 
tain property  of  matter  called  its  "weight";  if  we  chsnge  the 
motion  of  a  body  by  means  of  our  arms,  e.  g.,  if  we  throw  a  ball 
or  stop  one  in  motion,  we  become  conscious  through  the  same  chan- 
nel of  a  property  of  matter  called  "  inertia."  It  is  possible,  of 
course,  to  bold  a  body  suspended  from  the  earth  and  to  set  a  body 
in  motion  or  to  atop  it  if  moving,  by  other  means  than  by  our 
muscles;  thus  a  weight  can  be  suspended  from  a  spiral  spring  and 
hang  at  rest  with  reference  to  the  earth,  a  compressed  spiral  ^ring 

>y  Google 

The  Phtsioal  Basis  of  thb  Pboimjotion  of  Light        6 

may,  as  in  &  toy  gan,  produce  the  acceleration  of  a  bullet,  etc. 
Under  all  these  conditionB  which  are  in  their  nature  identical  vith 
tlioee  brou^t  about  by  our  mnsdeB  -we  say,  in  ordinary  language, 
that  "a  force  ie  acting  on"  the  body;  but  it  should  be  borne  in 
mind  that  this  ie  simply  a  deBcription,  nothing  more.  In  order  to 
assign  a  numerical  value  to  a  force  one  follovs  the  natural  way 
of  studying  the  simplest  cases  of  forces  one  can  have,  and  then 
using  definitions  and  methods  based  upon  these  observations.  The 
discussion  of  this  subject  forms  that  branch  of  mechanics  known 
OS  dynamics. 

The  simplest  mode  of  obtaining  a  unit  or  standard  force,  at  least 
from  the  standpoint  of  the  inhabitants  of  this  earth,  is  nndoubt^ 
edly  as  follows:  1.  Select  arbitrarily  a  certain  piece  of  matter. 
2.  Suspend  it  from  a  fiied  support  by  a  cord.  3.  Call  the  tension 
in  tJiiB  cord  a  unit  force.  It  is  easy  to  see  how,  by  means  of  a 
pulley,  it  is  possible  to  balance  this  force  by  an  equal  one  obtained 
by  su^ending  from  the  other  end  of  the  cord,  passing  over  the 
pulley,  another  body  which  is  added  to  gradually  until  there  is  a 
balance.  Having  thus  obtained  two  equal  forces  one  can  obtain 
a  force  twice  as  great  by  balancing  one  body  against  the  two  used 
in  the  first  experiment,  etc.  In  this  way  a  set  of  standard  bodies 
may  be  obtained  whose  weights  give  forces  equal  to  1,  2,  3,  4,  5, 
etc.,  and  then,  if  it  is  desired  to  give  a  number  to  an  unknown 
force,  this  may  be  done  by  balancing  it  against  a  selection  of  these 
known  forces. 

One  can  discuss  in  a  similar  manner  methods  of  giving  numbers 
to  temperature,  etc.,  and  this  will  be  done  in  a  later  lecture. 

TTsiti.  The  science  of  mechanics  is  based  upon  our  ideas  of 
loigth,  time  and  of  force,  and  methods  have  been  discussed  showing 
how  we  can  give  numbers  to  all  these  quantities.  It  is  seen,  how- 
ever, that  in  each  of  these  methods  certain  steps  are  arbitraiy,  and 
that  the  number  finally  obtained  depends  upon  the  nature  of  this 
arbitrary  step. 

a.  Length.  In  giving  a  number  to  a  length  the  first  step  is  to 
select  a  length  to  which  we  give  the  number  1  (if  we  use  the  inch, 
we  have  one  value  for  the  length,  if  we  use  the  centimeter  we  have 
a  different  value,  etc.).  The  scientific  world  agrees  to  adopt  as  its 
unit  of  length  the  one-hundredth  portion  of  the  length  of  a  certain 
platinum  rod,  kept  in  Paris,  when  this  rod  is  at  the  temperature  of 
melting  ice.     The  length  of  this  rod  under  these  conditions  is 

>y  Google 

6  Illduinatiho  Gnoineebino 

called  a  "  meter " ;  and  one-hundredth  of  this  length  ie  called  a 
"centimeter."  There  are  other  unit  lengths  in  daily  nse  in  this 
country  and  in  England,  but  it  ie  not  neceBsarj  to  diecuBs  them. 

b.  Time.  In  assigning  a  number  to  an  inatant  of  time  we  saw 
that  it  was  necessary  to  select  a  "  time-keeping  mechanism,"  such 
Bs  a  clock,  and,  secondly,  to  agree  upon  some  definite  instant  from 
which  to  begin  counting.  The  scientific  world  has  agreed  to  adopt 
as  its  time-keeping  instniment  the  earth  itself  as  it  rotates  on  its 
axis,  and  to  use  as  the  unit,  in  terms  of  which  interrals  of  time  are 
expressed,  the  "  mean  solar  second."  This  quantity  is  the  second  of 
time  referred  to  the  "  mean  solar  day,"  which  is  the  average  length 
for  one  year  of  the  lengths  of  the  solar  days  during  that  interval, 
a  solar  day  being  the  interval  of  time  between  the  two  instant* 
when  the  sun  crosses  the  earth's  meridian  at  any  point.  It  is  known 
that  solar  days  differ  in  length,  but  pendulums  may  be  made  whose 
periods  are  such  that  they  agree  exactly  with  the  earth  in  its  rota- 
tions at  intervals  a  year  apart,  and  these  clocks  are  used  ordinarily 
as  time-keeping  instruments.  Different  epochs  are  chosen  in  dif- 
ferent localities;  these  usually  differ  by  one,  two,  etc.,  hours. 

c.  Force.  In  assigning  a  number  to  a  force  it  was  seen  that  the 
essential  step  was  to  select  an  arbitrary  piece  of  matter;  and  here 
the  scientific  world  has  agreed  to  use  a  certain  piece  of  platinum 
kept  in  Paris.  When  this  body  is  suspended  and  allowed  to  hang 
vertically  there  is  said  to  be  "a  force "  in  the  string  equal  to  the 
"  weight  of  one  kilogram."  The  thousandth  portion  of  this  force 
is  called  the  weight  of  "  one  gram."  In  England  and  this  country 
other  unit  forces  are  sometimes  used,  commonly  what  is  called  the 
weight  of  a  "  pound." 

The  unit  force  on  the  "  centimeter-gram-second  "  (C.  G.  S.)  sys- 
tem, as  used  in  all  scientific  laboratories,  is  the  force  required  to  pro- 
duce an.  acceleration  of  one  centimeter  per  second  per  second  in  a 
piece  of  matter  whose  mass  is  one  gram.  This  force  is  called  one 
"  dyne."  The  weight  of  one  gram  is  very  closely  980  dynes — it 
is  not  the  same  at  all  points  on  the  earth. 

d.  Pressure.  From  these  fundamental  properties — length,  time 
and  force — numerous  other  quantities  are  derived,  one  of  which 
should  be  mentioned  here:  pressure.  By  pressure  we  mean  the 
force  per  unit  area,  and,  of  course,  the  number  we  obtain  for  any 
pressure  depends  upon  our  selection  of  units  of  force  and  of  area. 

>y  Google 

The  Physical  Basis  op  the  Pboduction  op  Light        7 

Ueunrementa.  It  is  necesBary  to  say  a  few  words  in  regard  to 
the  actual  meaBurement  of,  or  methoda  of  assigning  numbers  to, 
the  phyeical  quantities  so  far  discuBsed;  but  it  is  easily  understood 
that  for  any  satisfactory  discuaeion  of  the  subject  reference  should 
be  made  to  some  laboratory  hand-book. 

a.  Length.  In  the  measurement  of  small  lengths  two  methods 
are  in  general  use;  one,  depending  upon  the  use  of  a  screw  and 
divided  head,  the  other  upon  the  use  of  a  vernier.  In  the  measure- 
ment of  greater  lengths  special  precaution  must  be  taken  against 
changes  due  to  temperature,  flexure,  etc. 

b.  Volume.  Measurements  of  volume  are  made  in  one  or  two 
ways;  if  the  volume  to  be  measured  has  the  shape  of  a  simple  geo- 
metrical figure,  its  linear  dimensions  are  measured  and  its  volume 
calculated ;  if  the  volume  is  irregular,  or  if  it  is  that  of  an  inacces- 
sible space,  a  method  is  used  depending  upon  our  knowledge  of 
the  volume  of  mercury  which  is  required  to  produce  a  definite 
weight  at  a  definite  temperature;  e,  g.,  the  volume  of  a  bulb  may 
be  determined  by  filling  it  with  mercury,  expelling  the  mercury, 
noting  its  temperature,  and  then  weighing  it. 

c.  Time.  Methods  of  accurate  measurement  of  time  are  too 
complicated  to  be  discussed  here.  It  is  sufGcient  to  note  that  there 
are  several  metJiods  which  give  an  accuracy  of  a  minute  fraction 
of  a  second, 

d.  Force.  The  general  method  of  measuring  a  force  is,  as  stated 
before,  to  balance  it  against  a  known  force,  or  a  combination  of 
such  forces.  It  is  possible  to  buy  sets  of  weights,  or  a  spiral-spring 
balance,  which  will  give  results  sufficiently  accurate  for  all  purposes. 

e.  Pressure.  It  is  ouatomary  to  measure  pressures  such  as  those 
of  the  atmosphere,  of  boilers,  of  water  mains,  etc.,  by  balancing  the 
pressure  against  a  vertical  column  of  mercury.  An  illustration  of 
this  method  is  furnished  by  the  ordinary  mercury  barometer.  Since 
this  is  the  accepted  method,  the  unit  in  terms  of  which  pressures 
are  most  often  expressed  is  that  of  "  one  centimeter  of  mercury," 
hy  which  is  meant  the  vertical  pressure  required  to  balance  a  column 
of  mercury,  at  the  temperature  of  melting  ice,  one  centimeter  in 
height,  when  the  force  of  gravity  is  that  which  exists  at  sea-level 
at  latitude  45  degrees.  This  is  a  perfectly  definite  unit,  and  its 
value  is  known  in  terms  of  the  other  units. 

Erron  of  lutminents  and  Obserratlons.  In  this  brief  refer- 
ence to  the  measurements  of  these  five  quantities  it  is  seen  that 

>y  Google 

8  Illdhinatinq  Enoinebrinq 

reliaooe  miiBt  alwajB  be  placed  upon  an  inBtniment  fumiahed  b; 
some  instrument  maker;  e.  g.,  a  micrometer  screw,  a  vernier  scale, 
a  Bet  of  weigbts,  a  clock,  etc.,  and  it  ehould  not  be  neceesaiy  to 
empbasize  two  facts  in  connection  with  these  instruments.  First, 
every  inatmment  mnst,  of  course,  be  compared  with  the  original 
standard,  or  with  copies  of  it  whose  errors  are  known.  It  is  for 
this  purpose  that  in  all  civilized  countrieB  Bureaus  of  Standards 
exist  where  such  comparisons  may  be  made.  Thus  every  testing 
laboratory  in  America  has  or  should  have  standards  of  length  and 
of  mass,  whose  values  are  known  accurately  in  terms  of  the  Paris 
standards.  But,  even  granting  that  the  testing  laboratory  has  tliese 
standards,  there  are  many  errors  or  uncertaintieB  inherent  in  the 
use  of  every  ioBtrument,  and  a  thorough  study  must  be  made  of 
it  before  it  can  be  used  for  purposes  of  measurement.  Thus  no 
screw  has  an  absolutely  uniform  pitch,  and  the  variations  in  this 
must  be  determined  by  known  methods;  no  set  of  wei^ts  is  ac- 
curate, and  its  errors  must  be  learned;  and  similar  statements  are 
true  in  regard  to  every  instrument.  The  first  precaution  therefore 
in  the  measurement  of  any  qnantity  is  to  determine  the  true  scale 
of  the  instrument,  which  is  not  by  any  means  in  all  cases  that 
assigned  to  it  by  the  instrument  maker,  and  also  to  learn  the  varia- 
tions is  this  scale  in  different  parts  of  the  instrument. 

Second,  when  an  instrument  is  to  be  used  for  purposes  of  meas- 
urement it  is  not  sufficient  to  simply  make  one  observation,  e.  g., 
to  observe  onoe  the  reading  on  a  micrometer  of  the  diameter  of  a 
wire.  It  is  necessary  to  repeat  the  measurement  often.  To  begin 
with  it  is  always  possible  that  an  error  may  be  made  In  reading 
the  figures  on  the  instrument  or  in  recording  them.  Again,  when 
the  same  measurement  is  repeated,  the  measuring  instrument  being 
removed  and  then  replaced,  it  is  noted  that  as  a  rule  a  different 
reading  is  obtained.  This  does  not  mean  that  the  quantity  measured 
has  changed  or  that  the  instrument  used  is  defective,  but  simply 
that  in  the  use  of  the  instrument  there  are  certain  inherent  errors 
which  limit  the  accuracy  to  which  it  may  be  trusted,  errors  coming 
in  part  from  the  individual  using  the  instrument,  in  part  from 
the  instrument  iteelf,  and  in  part  from  other  causes.  When  a 
sufficient  number  of  observationB  have  been  made  one  may  calculate 
by  known  methods  the  most  probable  value  to  be  attached  to  the 
quantity,  and  also  learn  something  concerning  the  certainty  with 
which  this  number  may  be  regarded  as  approaching  the  true  value. 

>y  Google 

Thb  Physicai,  Basis  op  thb  Peoduotioh  of  Light        9 

The  confidence  felt  in  their  meaBuremente  by  cerUin  obBeireTB,  «Bd 
their  entire  lack  of  appreciation  of  the  need  of  ascertaining  tiie 
probable  errors  and  nncertaintiee  inTolved,  is  little  short  of  astoond- 
ing  to  one  accustomed  to  ordinary  laboratory  methods. 

Eleotrioal  Qaantities.  It  seems  necessary  in  this,  the  first  lecture 
of  the  conrse,  to  give  a  brief  diseussjon  of  some  quantities  which 
will  not  be  fully  explained  until  later  in  the  coarse.  These  are  the 
Tarioua  electrical  quantities ;  and,  of  oourfle,  to  most  engineers  they 
are  all  well  known.  In  the  history  of  electric  currents  many  unitB 
have  come  to  the  front  at  different  periods,  and  even  at  the  present 
time  the  definitions  are  not  the  same  in  all  coontriea.  The  differ- 
ences, however,  are  so  slight  as  to  justify  ua  in  neglecting  them  in 
all  ordinary  cases.  The  definitions  glren  in  what  follows  are  those 
in  terms  of  which  practically  all  the  measuring  instruments  now  in 
use  are  calibrated.  The  unit  of  resistance — the  ohm — is  defined 
to  be  equal  to  the  resistance  of  a  column  of  mercury  at  zero  degrees, 
of  uniform  croaa-Bection,  of  length  106.3  cms.,  and  having  the 
weight  of  14.4521  grams.  (Thia  column  then  has  a  crose-section 
of  almost  exactly  one  square  millimeter.) 

The  ampere — the  unit  of  current — is  defined  to  be  auch  a  current 
as  flowing  in  a  silver  voltameter  of  a  specified  pattern  deposits  per 
second  .001118  grams  of  silver. 

The  volt — the  unit  of  e.  m.  f. — is  defined  to  be  such  a  difference 
of  potential  as  will  produce,  when  applied  to  a  conductor  whose 
resiatance  is  one  ohm,  a  current  of  one  ampere. 

One  of  the  fundamental  properties  of  current  when  flowing  in  a 
conductor  is  to  develop  heat  in  this  condoctor,  and  it  is  well  known 
that  a  simple  formula  connects  the  heat  developed  and  the  electrical 
characteristics  of  the  system.  This  matter  will  be  discussed  more 
fully  in  the  second  lecture. 

In  order  to  give  numbers  to  the  resistance  of  a  conductor  the 
current  flowing  in  it  and  the  difference  of  potential  at  any  two 
points,  various  methods  have  been  devised  and  instrumente  per- 
fected. At  the  present  time  there  are  no  instrumente  in  common 
use  in  laboratories  which  have  attained  accuracy  to  euch  a  remark- 
able degree  as  these.  Thia  is  owing  in  large  part  to  the  epoch- 
making  inrentions  of  Siemens  and  Lord  Kelvin  in  Europe,  and  of 
Weston  in  this  country.  Thanks  to  the  efforts  of  tiiese  scientists 
we  now  have  instruments  for  the  measurement  of  volts,  amperes 
and  watts  which  are  sufficiently  accurate  for  roost  purposes.    I  may 

>y  Google 


be  pardoned  if  I  again  emphasize  the  fact,  however,  that  all  instru- 
ments are  imperfect  and  that  uncertainty  is  attached  to  every  ob- 

Lecture  II 

Energy  and  Thermal  Phenomena 
Work  asd  Ener^.  We  are  all  familiar  with  the  use  of  the 
words  "  work  "  and  "  energy  "  in  every-day  language.  They  have 
been  adopted  in  physics  as  names  of  certain  physical  quantities 
which  admit  of  exact  definition.  Naturally  these  definitions  have 
been  made  so  as  to  coincide  as  nearly  as  possible  with  those  every- 
day experiences  which  gave  rise  to  the  names  originally.  Thus, 
if  a  man  raises  a  weight  vertically  from  the  ground,  if  he  com- 
presses a  spring,  if  he  throws  a  base-ball,  he  knows  that  he  is  doing 
work.  The  essential  ideas  in  all  cases  of  work  are,  first,  the  action 
of  a  force,  and,  secondly,  a  displacement  in  the  direction  of  this 
force.  Corresponding  to  these  ideas  the  numerical  value  of  work 
is  defined  to  be  the  product  of  these  two  quantities,  i.  e.,  the  value 
of  the  force  by  that  of  the  displacement  in  the  direction  of  the 
force.  It  is  easily  seen  that  in  all  cases  in  mechanics  the  results 
of  a  force  are  either  to  overcome  another  force  or  to  produce  accel- 
eration (i.  e.,  change  of  velocity  of  a  piece  of  matter).  Correspond- 
ing to  these  two  types  of  fwces  there  are  two  ways  in  which  work 
may  be  done;  first,  when  a  force  or  opposition  is  overcome,  as  when 
a  weight  is  lifted,  a  spring  is  wound  up,  a  bow  is  bent,  etc. ;  second, 
when  acceleration  is  produced,  as  when  a  ball  is  thrown,  a  fly-wheel 
or  grindstone  is  set  in  motion,  etc.  It  is  common  experience  that 
in  all  cases  when  work  is  done  on  a  body,  as  when  a  weight  is 
raised  from  the  earth,  a  spring  is  wound,  a  body  given  accelera- 
tion, the  body  as  a  result  gains  the  power  o(  doing  work  itself.  It 
is  said  to  have  gained  "  energy."  If  the  work  done  on  the  body 
has  been  done  in  overcoming  an  opposing  force,  the  body  is  said 
to  have  gained  "  potential "  energy ;  whereas,  if  the  work  has  been 
done  in  producing  acceleration,  the  body  is  said  to  have  gained 
"  kinetic  "  energy.  Potential  energy  is  therefore  always  associated 
with  a  body  in  a  strained  or  "unnatural"  condition;  kinetic  en- 
ergy, with  motion,  either  translation  or  rotation.  It  is  a  matter 
of  common  experience  also  that  in  all  cases  of  mechanical  work  one 
body  loses  energy  and  a  second  body  gains  it.  Thus,  if  a  bullet 
is  expelled  from  a  toy  gun  by  means  of  the  sudden  relaxation  of  a 

>y  Google 

The  Physical  BA6ia  of  the  Prodcction  op  Light      11 

compressed  spring,  the  buUet  gains  energy  and  the  spring  loses  it. 
It  is  easy  to  show  tiiat  for  all  types  of  ordinary  mechanical  forces 
the  amount  of  energy  lost  by  one  part  of  the  system — namely,  that 
which  is  doing  work,  is  numerically  equal  to  the  energy  gained  by 
another  portion  of  the  system,  that  on  which  work  is  being  done; 
and,  as  a  coneequence,  therefore,  the  total  amount  of  energy  in  the 
system  remains  unchanged.  It  was  recognized  many  years  ago 
that  there  were  certain  apparent  exceptions  which  were  associated 
with  friction.  Thus,  if  a  fly-wheel  in  motion  is  disconnected  from 
the  driving  shaft,  its  energy — as  shown  by  ita  motion — gradually 
decreases,  as  it  comes  to  rest  under  the  action  of  friction.  Here, 
then,  is  a  case  of  an  apparent  disappearance  of  energy.  It  was 
noted,  however,  that  in  all  cases  like  this  there  were  certain  heat- 
eSects  produced ;  and  it  has  been  established  that  there  is  an  inti- 
mate connection  between  the  loss  of  mechanical  energy  and  the 
resulting  heat-phenomena.  Before  stating  this  connection,  how- 
ever, it  may  be  well  to  say  a  few  words  in  regard  to  our  ideas  of  heat. 
Heat-Phenomena.  Onr  attention  is  called  to  tiiermal  phenomena 
by  means  of  our  temperature  sense.  We  possess  in  certain  portions 
of  the  surface  of  our  bodies  nerve  endings  which  are  sensitive  to 
thermal  changes  in  our  environment.  That  is,  if  we  expose  our 
hands  tc  sunshine  or  bring  them  near  a  stove  in  which  there  is  a 
fire,  or  to  a  flame,  we  experience  a  definite  sensation,  and  we  say 
that  we  feel  warm.  Whereas,  if  we  put  our  hands  on  a  block  of 
ice,  or  if  we  allow  some  volatile  liquid  to  evaporate  from  them, 
we  experience  a  different  sensation  and  say  that  we  feel  cold.  The 
first  step  in  the  scientific  investigation  of  these  phenomena  must 
be  taken  by  exposing  a  piece  of  inanimate  matter,  such  as  a  rod  of 
iron,  to  the  same  conditions  as  those  under  which  we  felt  warm  or 
cold.  When  this  is  done,  it  is  found  that  the  piece  of  matter 
undergoes  various  changes;  and  these  are  called  thermal  effects. 
In  ordinary  language  we  speak  of  a  change  from  a  condition  when 
we  feel  cold  to  a  condition  when  we  feel  hot  as  being  a  change 
from  low  "  temperature  "  to  high  temperature.  Experiments  show 
that  when  the  temperature  of  a  body  is  changed,  all  of  its  physical 
properties,  with  the  exception  of  its  mass  and  weight,  are  also 
changed.  We  select  ordinarily  from  these  thermal  effects  a  few 
of  the  most  obvious  and  the  most  important  for  purposes  of  study 
and  observation.  Among  these  may  be  mentioned  change  in  volume, 
change  in  electrical  resistance,  and  change  in  state,  as,  for  instance. 



Then  a  piece  of  ice  melts  and  becomes  liquid.  On  examination  it 
is  found  that  whenever  work  is  done  against  friction,  heat-effects 
are  produced,  and  the  investigations  of  Joole  led  him  to  believe 
that  the  comiection  between  these  two  phenomena  waa  an  exact 
one,  which  conld  be  stated  by  saying  that  the  amount  of  heat-effect 
produced  depended  simply  upon  the  amount  of  work  done  against 
friction,  i.  e.,  upon  the  apparent  loss  of  energy,  and  upon  nothing 
else,  cot  upon  the  time  taken  for  the  change,  nor  the  temperature 
of  the  working  parte,  etc.  As  a  matter  of  fact,  if  we  consider 
variouB  cases  in  which  heat-effects  are  being  produced,  we  see  that 
in  them  all  work  is  being  done  against  the  smaller  parts  of  the 
body  which  experiences  the  heat-effect,  in  such  a  manner  that  the 
energy  of  these  smaller  parts  is  altered.  Aa  a  consequence  of 
various  experiments,  but  notably  those  of  Joule,  the  scientific  world 
has  accepted  the  belief  that,  when  we  are  dealing  with  friction  or 
similar  phenomena,  there  is  no  loss  of  energy,  but  that  simply  the 
poriiions  of  matter  with  which  it  becomes  associated  are  too  minute 
for  observation  with  our  eyes,  and  therefore  ve  do  not  observe  by 
this  means  the  effect  produced,  but  tiiat  this  effect  ia  shown  to  us 
through  our  temperature  sense  or  by  some  heat-effect.  This  state- 
ment means  that  one  can  apply  a  numerical  value  to  the  heat-effects 
produced,  in  such  a  manner  that  if  it  is  introduced  into  the  total 
value  of  the  energy  of  a  system,  this  total  value  remains  unchanged 
no  matter  how  much  friction  may  take  place  in  the  system. 

Conaerration  of  Energy.  This  coDstaocy  of  a  certain  number 
when  applied  to  the  energy  of  a  system,  including  in  that  the  proper 
figure  to  take  into  account  heat-phenomena,  ia  an  illustration  of 
what  is  meant  by  the  principle  of  the  conservatiQn  of  energy.  This 
principle  was  extended  by  Joule,  Mayer  and  Helmboltz  to  include 
other  phenomena  than  those  of  mechanics  and  heat.  For  instance, 
we  know  that,  if  we  place  some  granules  of  zinc  in  a  test  tube  and 
pour  sulphuric  acid  upon  them,  there  is  a  violent  evolution  of  gas 
and  the  teat  tube  gets  warm.  This  experiment  can  be  described  in 
terms  of  energy  by  saying  that  the  internal  energy  of  the  molecules 
of  the  zinc  and  of  the  acid  furnish  the  supply  necessary  for  the 
formation  of  the  new  molecules  and  also  for  the  production  of  the 
rise  in  temperature.  This  experiment  forms  one  of  thousands 
coming  under  the  head  of  Thermo-Chemistry,  and  all  of  these  have 
'  resulted  in  justifying  the  above  description  of  the  experiment  in 
terms  of  the  internal  energy  of  the  various  substances.     We  also 

>y  Google 

The  Phybioax  Babis  of  the  Pkodtjotiok  of  Lzaht      13 

know  that,  if  we  take  a  test  tube  containing  snlphnric  acid  and 
insert  into  it  a  strip  of  zinc  and  a  strip  of  some  other  metal  like 
copper,  the  two  being  joined  outside  the  teet  tube  by  means  of 
some  wire,  we  ^all  theu  have  what  we  call  an  electric  current. 
This  is  an  iUastration  of  a  primary  cell.  In  this  particular  type 
of  cell  the  zinc  diasolvea  in  the  acid,  and  there  la  an  evolution  of 
gas;  the  chemical  aide  of  the  experiment  is  exactly  the  same  as  in 
the  pieviouB  teat-tube  experiment  just  described.  It  is  observed, 
however, .  that  in  the  second  ezperimentj  that  with  the  primary 
cell,  there  is  practically  no  change  in  temperature  of  the  test  tube. 
This  means,  in  general  language,  that  the  energy  previously  nsed 
in  causing  a  change  in  temperature  is  consumed  in  this  case  in 
producing  the  electric  current  As  a  matt«r  of  fact,  we  all  know 
that,  when  an  electric  current  is  passing  in  a  c<mductor,  the  tem- 
perature of  the  latter  is  raised;  and,  if  the  conservation  of  energy 
can  he  extended  to  the  phenomena  of  electric  currents,  we  would 
expect  to  fnd  on  investigation  that  the  energy  consumed  in  the 
heating  of  the  conductor  by  the  current  is  exactly  the  same  as 
that  which  is  not  accounted  for  in  the  heating  of  the  test  tube 
where  the  chemical  reactions  are  going  on.  Complete  investigations 
on  this  point  justify  this  belief.  Joule  performed  many  interesting 
experiments  to  see  if  in  return  for  a  given  amount  of  work  he 
always  obtained  the  same  heat-effect  regardless  of  the  method  and 
mechanism  by  which  the  latter  was  caused  by  the  former;  thus,  by 
meana  of  a  steam  engine,  it  Is  possible  to  turn  a  paddle  in  water 
and  one  can  note  the  rise  in  temperature  of  tiie  water,  or  by  meana 
of  the  same  engine  one  can  turn  a  dynamo,  thus  producing  a  cur- 
rent which  can  be  made  to  flow  in  a  wire  immersed  in  water,  and 
again  the  final  effect  is  the  riee  in  temperature  of  water.  In  all 
cases  like  this  it  is  found  that  the  conservation  of  energy  is  fully 
justified.  As  a  consequence  of  these  and  countless  other  experi- 
ments it  has  become  an  accepted  belief  that  the  conservation  of 
energy  can  be  extended  to  all  phenomena  of  both  matter  and  ether. 
Temperatnre  and  Ihennometcn.  Before  discussing  questicms  of 
radiation  and  absorption  as  heat-phenomena  it  is  necessary  to  say 
something  in  regard  to  temperature  and  the  methods  by  which  we 
are  able  to  give  a  number  to  the  temperature  of  a  body.  As  we 
use  the  words  hot  and  cold  and  speak  of  high  temperature  snd 
low  temperature  in  ordinary  language,  we  are  making  use  of  ideas 
which  come  from  our  temperature  senaea,  and  therefore  the  tem- 

>y  Google 

14  Illuminating  Engineeeino 

perature  of  a  body  is  a  tenn  which  refers  to  its  relaiive  hotness. 
It  is  easily  seen  that  this  quantity  canoot  be  measured,  i.  e.,  we 
cannot  regard  otherwise  than  as  absurd  such  an  idea  as  selecting 
a  unit  of  hotnese  and  determining  how  many  times  it  ia  contained 
in  the  hotnees  to  which  we  wiah  to  give  a  number.  The  words 
themselTes  are  nonsense.  It  is,  however,  evident  that  we  can  choose 
such  a  measurable  property  of  some  body  as  changes  when  the  tem- 
perature of  the  body  changes,  and  make  use  of  the  measured  change 
in  this  as  a  means  of  giving  a  number  to  the  temperature  itself. 
For  instance,  we  can  select  arbitrarily  a  certain  cop^r  rod,  measure 
its  length  under  some  condition  which  can  be  easily  repeated,  such 
as  at  the  temperature  of  melting  ice,  again  measure  ita  length  when 
it  is  as  another  definite  temperature,  for  instance,  when  it  is  im- 
mersed in  steam  under  standard  conditions,  then  measure  its  length 
at  the  temperature  for  which  a  number  is  desired.  We  can  assign 
arbitrarily  a  certain  ntunber  of  steps  or  degrees  to  the  interval 
between  tiie  temperatures  of  meltiug  ice  and  of  steam,  say,  100; 
then  an  obvious  method  of  giving  a  number  to  the  temperature 
would  be  to  take  a  proportion  of  100  equal  to  the  ratio  of  the  change 
in  length  of  the  rod  between  melting  ice  and  the  unknown  tempera- 
ture to  the  change  in  length  between  melting  ice  and  steam,  i.  e., 

t  =  100  Y^ 

which  are  justified  by  observations;  namely,  that  the  temperature  of 
melting  ice  and  of  boiling  water  under  standard  conditions  are  the 
same  at  all  points  on  the  earth's  surface,  and  at  all  times  (thia  may 
be  shown  by  proving  that  a  body  will  always  return  to  the  same 
length  when  placed  in  a  bath  of  ice  and  water,  etc.) ;  further,  that 
the  copper  rod  we  have  selected  always  attains  the  same  length 
under  the  same  thermal  conditions.  It  should  be  noted,  too,  that 
this  scale  gives  the  number  0  to  the  temperature  of  melting  ice 
and  100  to  that  of  boiling  water.  (It  is  clear  that  this  method  of 
giving  a  number  to  temperature  is  practically  the  same  as  that  which 
anyone  would  follow  if  called  upon  to  give  a  street  number  to  a 
house  erected  at  some  point  in  a  block  otherwise  vacant.)  It  can- 
not be  emphasized  too  often  that  we  have  devised  a  method  for 
giving  a  number  to  temperature,  and  that  we  have  not  in  any  sense 
tried  to  measure  temperature. 

Some  other  observer  might  decide  to  take  as  his  thermometer, 
or  instrument  tor  numbering  temperatures,  an  iron  rod  and  meaa- 

>y  Google 

The  Physical  Basis  op  the  Production  op  Lioht      16 

nre  iU  change  in  length ;  or  a  glase  bnlb  containing  mercury  and 
measare  the  apparent  change  in  Tolnme  of  the  mercnry ;  or  a  glaee 
bulb  containing  some  gas  and  measure  the  change  in  pressure  of 
the  gas,  its  volume  being  kept  constant;  or  a  platinum  wire  and 
measure  the  change  in  its  electrical  resistance;  and  so  on.  One 
of  these  methods  is  as  good  as  another;  eadi  gives  consistrait  re- 
sults by  itself ;  and,  if  several  observers  use  instruments  of  the  same 
kind,  their  readings  are  concordant.  But  the  readings  obtained 
for  any  one  temperature  by  the  use  of  different  methods  and  instru- 
ments would  all  be  different;  and  it  is  necessary  for  workers  in 
scientific  laboratories  to  come  to  an  agreement  as  to  which  instru- 
ment they  will  use.  The  scientific  world  has  agreed  to  adopt  as 
the  instrument  for  giving  numbers  to  temperature  the  constant 
volume  hydrogen  thermometer.  In  various  bureaus  of  standards 
throughout  the  world  ordinary  mercury  thermometers  may  be  com- 
pared with  the  standard  instruments,  so  that  the  former  may  be 
used  for  ordinary  purposes,  as  they  are  much  more  convenient. 

It  is  clear  that  this  definition  of  temperatare  applies  only  through, 
the  range  of  temperature  over  which  we  can  make  use  of  the  hydro- 
gen thermometer.  When  we  come  to  temperatures  so  low  or  so 
high  that  there  are  serious  defects  in  the  use  of  tiie  instrument, 
it  is  necessary  to  define  other  scales  of  temperature.  For  instance, 
at  extremely  low  temperatures  a  helium  thermometer  may  be  used, 
or  a  platinum  resistance  instrument;  and  at  high  temperatures  a 
scale  of  temperature  based  upon  certain  empirical  laws  of  radiation 
may  be  adopted.  In  both  these  cases  of  the  introduction  of  new 
scales  of  temperature  the  attempt  is  made  to  define  them  so  that 
they  agree  with  the  gas  temperatures  at  those  moderately  low  and 
moderately  high  temperatures  over  which  this  gas  scale  can  be 
used  at  the  same  time  as  the  two  new  ones.  In  this  way  a  certain 
continuity  is  obtained,  but  it  mlist  not  be  thought  that  we  are 
extending  the  hydrogen-gas  scale;  on  the  contrary,  we  are  intro- 
ducing new  scales. 

Badiation  and  Abaorptioa. — In  text-books  on  physics  one  finds 
a  full-  description  of  methods  of  producing  heat-effects  such  as 
flames,  frietion,  etc.,  and  also  a  description  of  the  various  methods 
by  which  in  general  these  effects  are  distributed  from  one  point  to 
another,  as  by  conduction  or  radiation.  In  this  course  of  lectures 
special  emphasis  must  be  laid  upon  the  radiation  process.  This  is 
illustrated  when  we  expose  our  hands  to  sunshine  and  in  many 

>y  Google 

16  Iii.nHiifATiNo  EnqiitiBbing 

other  similar  ways.  It  is  known  as  a  result  of  experimenta,  which 
need  not  be  diacussed  here,  that  the  eraential  features  of  the  procees 
are:  firat,  an  emission  from  one  body  of  energy  in  the  form  of 
ether  disturbances,  second,  the  absorption  of  this  energy  by  another 
body.  It  is  known  further  that  all  bodies  in  the  nniveTse  are 
emitting  this  energy.  As  a  consequence,  therefore,  of  these  two 
facta  the  question  as  to  whether  there  will  be  any  heat-effect  pro- 
duced in  a  body  owing  to  radiation  processes  depends  upon  two 
things ;  first,  how  much  energy  the  body  is  losing ;  second,  how  moch 
it  is  gaining.  The  phenomma  of  radiation  and  absorption  of  many 
bodies  under  differ^t  conditions  have  been  carefully  studied  by 
many  observers,  and  in  the  middle  of  the  last  century  at  about  the 
same  time  a  very  important  law  was  announced  by  Balfour  Stewart 
in  England,  and  by  Kirchhoff  in  Germany.  '  The  statement  is 
ordinarily  called  "  Kirchhoff's  Law."  One  form  of  it  is  to  say 
that  the  radiating  power  and  absorptive  power  of  a  body  are  iden- 
tically the  same  in  all  respects  at  any  one  temperature;  i.  e.,  if  a 
,body  under  certain  conditions  radiates  a  certain  type  of  energy 
more  intensely  than  a  second  body,  then  the  first  body  under  the 
same  condition  will  absorb  that  same  type  of  energy  more  intens^y 
than  the  second.  (In  the  end  this  principle  is  an  illustration  of 
resonance.)  In  connection  with  this  discussion  of  radiation  and 
absorption  Kirchhoff  introduced  the  idea  of  a  "  black  body,"  mean- 
ing by  that  a  body  which  absorbs  completely  all  radiations  falling, 
upon  it;  for,  of  course,  in  general,  when  radiation  is  incident  upon 
a  body  part  is  reflected,  part  is  transmitted,  and  only  part  is 

Temparatnn  Sadiation.  When  the  radiation  from  bodies  was 
more  carefully  studied  it  was  found  necessary  to  make  certain  limi- 
tations in  the  application  of  Kirchhoff's  law.  Kirchhoff  himself 
applied  it  only  to  those  cases  where  radiation  was  to  be  considered 
simply  as  a  heat  process,  not  as  a  chemical  or  electrical  one,  and 
recent  experiments  appear  to  prove  that  we  are  justified  in  using 
Kirchhoff's  law  only  in  the  case  of  certain  particular  bodies  under 
definite  conditions.  One  way  of  defining  this  is  to  say  that,  if 
there  is  no  change  in  the  molecular  conGtitution  of  a  body  when  it 
is  radiating  energy,  its  temperature  being  maintained  constant, 
then  it  obeys  Kirchhoff's  law;  and  the  radiation  from  it  ia  called 
"  pure  temperature  radiation."  Other  types  of  radiation  will  be 
discussed  in  the  following  lecture. 

>y  Google 

The  Phtsioal  Basis  of  the  Pbodcotion  of  Light      17 

It  follows,  thee,  that  since  a  "  black  body  "  la  the  best  absorber 
possible  it  is  also  the  best  radiator;  i.  e.,  at  a  given  temperature 
it  radiates  moie  energy  of  any  particular  kind  than  any  other  radia- 
tor which  obeys  Kirchhoff's  law ;  and  it  also  follows,  therefore,  that 
all  "  black  bodies  "  radiate  alike  and  obey  the  same  laws.  If  we 
can  secnre  such  a  body,  then,  we  have  an  instmment  of  great  im- 
portance. E!irchhoff  himself  showed  that,  if  a  hollow  body,  sadi 
as  a  cast-iron  shell,  be  maintained  at  a  constant  temperatnre,  the 
radiation  inside  the  space  wbb  that  which  is  characteristic  of  a 
"  black  body  "  at  the  given  temperature.  If  a  small  opening  is 
made  from  without  to  the  interior  of  such  a  shell,  some  radiation 
will  escape;  but  the  type  of  radiation  inside  will  not  be  seriously 
affected ;  and,  since,  through  the  opening  we  receive  on  the  outside 
the  random  radiation  which  is  characteristic  of  the  interior,  we 
can  secure  in  this  manner  what  is  practically  a  "  black-body " 
radiator.  The  various  laws  which  have  been  deduced  for  the  radia- 
tion from  Bocb  a  body  will  be  discussed  in  the  next  lecture. 

KeMnniiiuit  of  liatTgj  and  Power.  So  far  nothing  has  been 
said  in  regard  to  the  measurement  of  energy  or  the  units  in  terms 
of  which  it  is  expressed.  If  we  use  the  C.  G.  S.  system  of  units, 
the  standard  of  energy  or  its  units  is  called  the  "erg" — i. e,,  the 
work  done  by  a  force  of  1  dyne  acting  through  1  cm. — which  is  an 
extremely  small  quantity,  bo  small  that  it  is  more  customary  to 
nse  10'  ergs  as  the  unit.  This  amount  is  called  a  "Joule."  If 
we  are  interested  not  simply  in  the  amount  of  energy  hut  in  the 
rate  at  which  it  is  delivered,  we  introduce  the  word  "  power "  to 
signify  the  energy  delivered  per  unit  of  time,  and  if  the  amount 
of  work  is  one  Joule  per  second  the  power  is  said  to  be  one  "  watt" 
(On  the  English  system  the  uDit  of  work  is  the  "foot-pound"; 
and  the  unit  of  power  is  a  "  horse-power,"  which  is  defined  to  be 
33,000  foot-pounds  per  minute — this  equals  approximately  746 

There  are  three  standard  ways  of  measuring  energy;  by  rise  in  . 
temperature,  by  mechanical  means,  by  electrical  methods.  A  few 
words  should  be  said  in  regard  to  the  first  and  third.  By  experi- 
ments performed  by  Joule,  by  Rowland  and  by  others  we  know 
accurately  the  amount  of  energy  requirad  to  raise  the  temperature 
of  water;  and  by  the  experiments  of  B^piault  and  many  others  we 
knojr  the  ratio  between  the  amount  of  energy  required  to  raise  the 
temperature  of  water  and  that  required  to  raise  the  temperature 

>y  Google 

18  Illuminating  Engikbbbinq 

of  other  eubetancee.  ConBequently,  if  we  can  observe  the  rise  in 
temperature  owing  to  heat-caasea  of  any  body  of  known  character, 
and  of  known  weight,  we  know  accurately  the  amount  of  energy 
supplied.  ThuB,  if  radiation  falls  upon  a  body  and  is  totally 
absorbed,  we  have  a  meane  of  measuring  the  amount  of  energy 

In  the  case  of  experiments  with  electric  cuirente  we  know  that 
the  energy  consumed  per  second  is  equal  to  the  product  of  the 
electro-motive  force  and  the  current;  and  the  units  of  tlie  ampere, 
the  volt  and  the  watt  are  bo  chosen  that,  if  the  electro-motive  force 
as  measured  in  volts  is  multiplied  by  the  value  of  the  current  in 
amperes,  the  product  is  the  number  of  watts  of  power  furnished  by 
the  current.  It  is  easy  to  see  how  by  having  this  simple  means 
of  determining  power  through  the  operation  of  the  electric  current, 
we  can  make'  use  of  it  for  the  general  measurement  of  energy. 

Lectdse  III 

Radiation.  By  radiation  we  mean  those  disturbances  in  the 
ether  which  are  being  emitted  by  matter  of  all  kinds  and  at  all 
times.  For  a  proper  study  of  its  nature  we  require  instruments 
'  which  analyze  the  radiation  and  which  measure  the  quantity  of 
energy  in  the  radiation.  It  was  observed  by  Newton  that  when 
the  radiation  from  a  small  source  of  light  was  allowed  to  pass 
through  a  prism  of  glass  it  was  broken  up  or  "  dispersed,"  bo  that 
the  white  light  of  the  sun,  for  inetance,  was  divided  into  many 
colore,  each  particular  color  corresponding  to  radiation  leaving  the 
prism  in  a  definite  direction.  This  process  of  analysis  of  radiation 
by  means  of  a  prism  is  called  "  dispersion  ";  aud  the  investigations 
of  Freanel  and  others  showed  that  what  takes  place  is  this;  the 
prism  transmits  in  definite  directions  trains  of  waves  of  definite 
wave-length ;  bo  that,  whatever  the  nature  of  the  incident  radiation, 
that  which  is  transmitted  is  distributed  into  regular  groups,  each 
group  having  a  definite  wave-length  and  leaving  the  prism  in  a 
definite  direction.  It  was  shown  by  Frannhofer  and  otherB  that 
one  could  secure  dispersion  by  other  means  than  by  the  use  of  a 
prism,  as,  for  instance,  by  the  use  of  a  dispersion  grating. 

The  apparatus  by  which  the  dispersion  of  light  is  studied  is 
called  a  "spectroscope."     It  consists  essentially  of  three  parts:  a 

>y  Google 

The  Physical  Babis  op  the  Phodtiction  op  IjIbht      19 

narrow  Blit  through  which  the  light  eotere;  a  priem  or  grating  to 
cause  the  dispersion ;  a  lens  or  concave  mirror  to  focus  the  different 
streams  of  radiation  on  a  suitable  screen,  where  the  detecting 
or  measuring  inBtrument  is  placed. 

Spectra.  When  the  radiation  from  any  very  hot  source  such  as 
the  sun  or  the  carbona  in  an  arc  light  is  thus  analyzed  and  spread 
out  according  to  its  wave-lengths,  it  is  observed  that  only  a  small 
portion  affects  the  eye.  This  is  called  "  the  visible  spectrum."  We 
see  a  broad  band  of  light,  colored  red  at  one  end,  and  violet  at  the 
other.  In  between  these  there  are  different  colors,  each  merging 
imperceptibly  into  its  neighbors.  Certain  colors  have  definite 
names;  and  we  often  speak  of  red,  orange,  yellow,  green,  blue, 
indigo,  violet,  as  being  the  "  colors  of  the  spectrum  ";  yet  we  must 
remember  that  these  colors  are  not  isolated;  the  transition  from 
red  to  violet  is  a  gradual  one.  If  a  photographic  plate  is  held  in 
the  region  beyond  the  violet,  it  is  affected  intensely;  and,  if  a 
thermometer  ie  held  in  the  region  beyond  the  red,  it  shows  by  its 
rise  in  temperature  that  energy  is  falling  upon  it.  We  are  thus 
accustomed  to  apeak  of  the  "  uUra-violet  spectrum  "  and  the  "  infra- 
red." When  the  wave-lengths  of  the  radiations  causing  in  our 
eyes  the  color  sensations  are  measured,  it  is  found  that  a  deSnite 
color  is  associated  with  a  definite  wave-length;  and  so  we  often 
speak  of  "  red-light,"  etc.,  meaning  radiation  of  such  a  wave-length 
as  produces  in  our  eyes  the  sensation  of  red,  etc.  The  wave-length 
of  the  radiation  in  the  extreme  ultra-violet  is  the  shortest  of  all; 
then,  as  the  wave-lengths  become  longer,  the  blue  end  of  the  spec- 
trum is  approached;  as  it  becomes  still  longer,  the  color  gradually 
changes  from  blue  to  green,  to  red,  etc.,  down  into  the  infra-red. 

Becordin^  IiistmmeBta.  It  is  not  easy  to  find  an  instrument 
which  will  respond  to  waves  of  all  wave-lengths,  i.  e.,  which  will 
absorb  them  or  will  indicate  the  amount  of  the  incident  energy. 
For  waves  which  are  extremely  short,  much  shorter  than  those  which 
affect  our  sense  of  sight,  we  may  use  a  jAotographic  or  a  photo- 
electric process;  through  the  visible  spectrum  we  may  also  use  a 
photographic  process  for  the  detection  of  the  radiation,  but  for  ita 
quantitative  measurement,  either  here  or  in  the  infra-red,  we  must 
use  some  modification  of  a  thermometer.  Various  types  of  instru- 
ments have  been  devised  and  the  problems  are  now  fairly  well  un- 
derstood. The  four  forms  of  instruments  in  general  use  are:  a, 
the  bolometer,  which  is  a  thin  strip  of  blackened  platinum  whose 

>y  Google 

20  Illdminatimg  Enoinbbeiko 

change  in  electrical  reBietance  produced  by  the  radiation  ia  raeaa- 
nred;  b,  the  thermo-couple,  or  junction  of  two  metala.  forming  a 
closed  circuit,  whose  E.  M.  F.  aa  altered  by  the  radiation  is  meaa- 
ured ;  c,  the  radio-micrometer,  an  instrument  in  which  the  thermo- 
electric current  product  by  the  radiation  flows  through  a  small 
circuit  suspended  between  tiie  poles  of  a  magnet,  and  can  therefore 
be  measured  by  the  deflection  produced ;  d,  the  radiometer,  a  modi- 
fication in  Crookes'  original  form  of  the  instrument,  depending 
upon  the  repulsion  produced  by  incident  radiation  in  a  blackened 
disk  suspended  in  a  partial  vacuum.  Any  one  of  theee  instruments, 
when  properly  calibrated,  may  be  used  to  measure  the  energy  of 

ClauM  of  Spectra.  If  the  Spectra  of  solids  and  liquids  are 
studied,  it  is  found  in  almost  every  case  that  there  is  a  continuous 
spectrum,  having  its  mazimuni  in  a  region  depending  primarily 
upon  the  temperature  of  the  source.  On  the  other  hand,  if  a  gas 
is  made  luminous  by  the  discharge  through  it  of  an  electric  current 
or  by  any  other  means,  it  is  noted  that  its  spectrum  is  discon- 
tinuous, i.  e.,  is  made  up  of  isolat^  trains  of  waves.  When  the 
light  from  a  white-hot  Bolid  is  allowed  to  fall  upon  any  body  sucli 
aa  a  piece  of  glass  or  a  tank  containing  some  liquid,  a  certain 
amount  of  the  radiation  ie  absorbed  by  the  body,  and  if  the  trans- 
mitted radiation  is  analyzed  by  a  prism  or  a  grating  tlie  resulting 
spectrum  is  called  "  the  absorption  spectrnm  "  of  the  body.  It  ia 
obvious  that  the  nature  of  this  spectrum  depends  not  simply  on 
the  body  itself  but  also  on  the  character  of  the  source. 

letnperatnre  Radiation.  In  the  preceding  lecture  some  time  was 
devot«d  to  the  discussion  of  the  conditions  under  which  Kirchhoffs 
law  of  radiation  and  absorption  could  be  applied.  It  may  be  re- 
membered that  Uiese  conditions  were  aa  follows :  If  a  body  is 
emitting  radiation  and  if  its  temperature  is  maintained  constant 
by  suitable  means,  then,  provided  there  are  no  permanent  changes 
produced  in  the  body,  it  obeys  Kirchhoff's  law  and  the  radiation 
which  it  emits  is  called  "pure  temperature  radiation."  The  im- 
portance of  this  discuseion  and  definition  comes  from  the  fact  that 
for  bodies  which  are  emitting  such  radiations  it  ia  possible  by  apply- 
ing certain  general  principles  of  physics  to  deduce  theoretically 
certain  relations  between  the  temperature  of  the  body  and  its  radia- 
tion. Further,  if  the  radiation  from  a  "  black  body "  is  studied 
experimentally,  certain  empirical  laws  connecting  gas  temperature 

i.y  Google 

Thb  Phibicai,  Babis  or  the  Pboddotiow  op  Liqut      81 

and  energy  of  radiation  may  be  learned,  and  all  "black  bodies" 
radiate  alike.  This  matter  will  be  referred  to  more  in  detail  to- 
wards the  end  of  the  lecture.  It  is  extremely  difflcnlt  to  obtain 
pure  temperature  radiation,  though  we  can  approximate  closely  to 
it  by  the  use  of  a  "  black  body  "  such  aa  described  in  the  last  lecture. 

LoBlineHcaiae.  In  general,  however,  when  a  body  is  emitting 
radiation  there  are  changee  going  on  in  it  even  if  ite  temperature 
is  maintained  constent  by  beating  it  from  withont;  such  bodies 
are  said  te  be  "  luminescent."  We  hare  many  types  of  luminescence 
and  it  may  be  wortii  while  to  say  a  few  words  concerning  some  of 
these.  There  is  what  is  called  "  chemical  ItunineBcenee,"  which  is 
illustrated  by  tiie  alow  ozidatton  of  phosphorus;  there  is  "electro- 
lomineecence "  ^hich  we  have  when  a  gas  is  made  laminons  by 
an  elecb-ical  discharge;  ihere  is  "  flnorescence,"  which  is  observed 
in  many  bodies  and  constBte  in  the  absorption  of  light  of  a  certain 
wave-length,  and  in  the  emission  of  light  of  a  different  wave-length. 
The  exact  energy  relation  for  the  various  cases  of  luminescence  are 
not  clear  in  all  cases;  nor  ie  it  poedble  to  stete  any  relatione  which 
connect  the  radiation  with  the  physical  properties  of  the  source. 

Fhotometry.  The  most  obrious  property  of  radiation  is,  of 
course,  its  power  to  affect  our  sense  of  eight  in  case  the  source  haa 
a  temperature  snfflciently  high,  or  in  case  it  is  emitting  waves  suf- 
ficiently short.  As  has  been  said,  we  associate  different  colors  witti 
different  wave-lengths,  and  the  question  therefore  as  to  our  color 
sensation  depends  primarily  upon  two  things;  the  nature  of  the 
radiating  source  and  the  power  of  our  eyes  to  recognize  color.  The 
physiolt^cal  action  of  the  eye  is  to  be  discussed  in  later  lectures; 
and  it  may  be  sufficient  to  note  here  that  the  eyes  of  most  people 
are  competent  to  distinguish  colors  with  great  accuracy,  provided 
the  illumination  is  sufficiently  intense. 

The  moat  important  matter  connected  with  radiation  is  the  ques- 
tion of  the  energy  carried  by  the  trains  of  waves  of  definite  wave- 
length. This  can  be  inveetagated  obviously  by  means  of  a  suitable 
dispersive  apparatus  and  a  sensitive  recording  instxument,  such  as 
a  bolometer  or  radio-micrometer  properly  standardized.  But  this 
is  largely  of  theoretical  importance.  What  we  are  most  closely 
concerned  with  is  the  question  as  to  the  intensity  of  the  effect  of 
radiation  upon  our  eyes.  The  investigation  of  the  various  problems 
connected  with  this  forms  the  science  of  photometry.  We  must 
find  suitable  methods  of  comparing  the  eJSciency  of  various  sources 

>y  Google 

2S  Illuhikatiko  Engikeekinq 

of  light  in  producing  light  eeosation ;  this  implies  a  study  of  the 
intensity  of  the  light  Bensation,  of  the  energy  required  for  this, 
and  of  that  portion  of  the  energy  of  the  source  which  is  radiated  in 
the  invieible  portions  of  the  spectrum. 

Colon  of  Objeotl.  We  are  concerned  most  often,  however,  not 
with  the  color  of  the  source  of  light  itaelf  but  with  the  color  which 
natural  objects  appear  to  have  when  viewed  in  a  certain  light. 
'We  ordinarily  call  a  leaf  gieen,  a  brick  red,  etc.,  meaning  simply 
that  when  viewed  in  sunlight  tiiese  objects  have  these  colors.  If  we 
study  carefully  many  cases  of  colored  objects  we  soon  recognize 
that  their  color  is  in  general  due  to  one  of  two  causes.  The  com- 
monest of  all  causes  is  what  is  called  "  body  absorption,"  and  is 
illustrated  perfectly  by  a  piece  of  colored  glass,  a  tank  of  colored 
water,  flowers,  etc.  The  process  is  as  follows:  The  incident  light 
penetrates  into  the  body,  where  certain  trains  of  waves  of  definite 
wave-lengths  are  absorbed,  and  where  the  rest  of  the  light  is  either 
transmitted  or  is  scattered  in  all  directions  by  small  inequalities  or 
dust  particles.  Consequently,  if  one  looks  at  the  object  either  by 
transmitted  light  or  from  any  direction,  he  will  receive  in  his  eye 
only  that  portion  of  the  incident  light  which  is  left  over  after  the 
absorption  in  the  interior  of  the  body.  If  the  incident  light  is 
white,  and  if  red  light  is  absorbed  by  the  body,  it  will  appear  blue, 
because  when  white  light  loses  its  red  constituent  it  becomes  blue. 
It  is  evident  therefore  that  the  nature  of  the  color  which  an  object 
appears  to  us  to  have  depends  vitally  upon  the  nature  of  the  light 
in  which  it  is  viewed,  because  we  see  in  the  end  that  light  which 
is  the  result  of  eubtraction  from  the  incident  light  owing  to  ab- 
sorption. The  same  body  will  appear  to  ns  of  a  different  color, 
if  the  color  of  the  source  is  changed.  If  the  light  after  passing 
through  one  colored  object  is  allowed  to  fall  upon  a  second,  and  it 
we  view  this  transmitted  light  we  have,  of  course,  a  double  sub- 
traction. This  is  the  process  which  we  have  ordinarily  in  the  mixing 
of  paints.  The  explanation  of  the  color  of  a  painted  object  is  ex- 
actly that  just  given;  the  light  enters  a  short  distance  and  is 
scattered  out,  so  that  if  two  paints  are  mixed  we  have  a  double 
subtraction.  It  is  hardly  necessary  to  emphasize  the  importance  of 
this  general  discussion  of  color  in  the  question  of  the  illumination 
in  a  room,  i.  e.,  the  effect  of  the  color  of  the  walls,  curtains,  etc., 
upon  the  general  illumination,  etc. 

>y  Google 

The  Phtsioal  Basis  of  the  PttODncTiON  or  Light      33 

There  are  certain  objects,  however,  which  owe  their  color  to  a 
process  different  from  this,  ae,  for  instance,  metala  and  the  aniline 
dyes.  In  their  case  the  incident  light  suffere  absorption  at  the 
Burface,  not  in  the  interior,  and  so  their  color  is  said  to  be  due  to 
"  Burface  absorption." 

There  are  many  other  exceptional  cases  of  color  about  which  noth- 
iog  need  be  said  at  the  present  time,  such  as  the  colors  associated 
with  luminescence,  interference,  the  scattering  due  to  fine  parti- 
cles, etc. 

LavB  of  TemperatUTe  Badiatian,  The  most  important  type  of 
radiation  is,  as  has  been  said  repeatedly,  pure  temperature  radia- 
tion; and  for  many  years  many  competent  obaerrers  have  been  in- 
vestigating the  connection  between  the  temperature  of  the  "  black 
body"  emitting  such  radiation  and  the  nature  of  the  spectrum 
and  the  amount  of  the  energy.  It  has  been  shown  that,  if  all  the 
energy  emitted  is  measured  by  using  a  suitable  absorbing  instru- 
ment, the  connection  between  the  temperature  of  a  source  and  the 
total  quantity  of  the  energy  may  be  expressed  by  an  extremely 
simple  formula,  namely, 

energy  emitted=a(tH-273)*, 
where  t  is  temperature  on  the  gas  scale,  and  a  is  a  measurable  con- 
stant, independent  of  temperature.  This  is  called  "  Stefan's  Lew." 
This  evidently  furnishes  a  means  of  defining  a  scale  of  temperature 
in  a  re^oD  where  a  gas  thermometer  could  not  be  used,  since  we 
can  measure  the  energy  emitted  by  bodies  at  all  temperatures.  The 
method,  of  course,  is  to  take  the  law  as  given,  which  states  the 
relation  between  gas  temperature  and  energy  over  the  extreme  range 
to  which  a  gas  thermometer  can  be  used,  and  define  the  tempera- 
ture for  regions  of  higher  temperature  by  the  formula  itself.  That 
is,  we  would  measure  the  enei^  from  a  certain  source  and  by  the 
use  of  the  formula  deduce  the  value  of  the  temperature.  It  should 
be  clearly  understood  that  there  is  no  assumption  involved  in  this; 
it  is  a  matter  of  definition. 

It  has  been  found  further  that,  when  the  energy  of  a  "black 
body"  has  been  dispersed  into  its  spectrum,  and  tfae  amounts  of 
energy  carried  by  trains  of  waves  of  definite  wave-length  are  meas- 
ured, there  is  also  a  connection  between  the  distribution  of  this 
energy  as  a  function  of  the  wave-length  and  the  temperature  of  the 
source,  as  measured  on  the  gas  scale.    Several  formulas  have  been 

>y  Google 

34  Illuminating  Enoineekino 

derived  from  these  experiments;  and  here  again  we  have  a  meaBS 
of  defining  a  temperature  acale  which  can  be  applied  to  extremely 
high  temperatures.  All  these  scales  defined  by  radiation  formulas 
seem  to  agree  to  a  high  degree  of  accuracy. 

One  of  these  relations,  known  as  Planck's  law,  may  be  written 

where  Ea  is  the  energy  carried  by  waves  whose  wave-lengths  lie 
between  A  and  A+dA,  T  is  written  for  t+373,  e  is  the  base  of  the 
natural  system  of  logarithms,  Ci  and  Ci  are  constants. 
Two  other  relations  are: 

Xm«,T=:  const, 

where  T  is  again  written  for  t  +  273;  Xnoc  is  the  wave-length  cor- 
responding to  the  maximum  value  of  E*  for  the  temperature  T; 
and  Em  is  the  value  of  Ei,  at  this  wave-length  AmM- 

=  const. 

>y  Google 




Bt  Edwabd  p.  Htdb 

Lecture  I 

1.  IntrodncUon. 

A.  What  U  llgbtr 

B.  The  condltlonB  to  be  fuUUled  by  light  Murces. 

C.  The  BoarceB  of  inpply  ftnd  loss  of  energj'. 

2.  Lnmlnoiis  eOcleinT. 

A.  Sensibility  of  the  eye  to  enersy  of  different  vave-tenKths. 

a.  Time  relation  between  stlmalas  and  Mnsatlon. 

b.  Senelblllty  a  function  of  aboolute  Intensity  of  illtunlnatlon 

(Parklnje  effect). 

B.  Luminosity  curves  for  various  fllumlnants. 

C.  Mechanical  equivalent  of  light. 

a.  UnaatlBfactory  nature  of  ordinary  deflnltton. 

b.  Mechanical  equivalent  of  most  elllclent  mmiochromatlc  radia- 

tion (M  =  800  lumens  per  vatt). 

D.  Highest  poSBlble  efficiency  of  white  light  (about  800  lumens  per 

B.  Highest  possible  efficiency  of  black  tody  radiation  (about  140 

lumens  per  watt). 
F.  Quantities  entering  In  dlscuaslon  of  efltcienev. 

a.  Power  Bupplled  to  lamp  (Q). 

b.  Power  radiated  by  lamp  (R). 

c  Power  dlBBlpated  by  convection  (C,,). 

d.  Power  dleslpated  by  conduction  (C^). 

e.  Power  radiated  In  visible  spectrum  (L). 

f.  Luminous  flux  in  lumens  (#}. 
8.  Quality  of  light. 

A.  Integral  color  of  composite  light. 

B.  Spectral  dletrlbutlon. 
4.  Temperature  radiation. 

A.  BIoclC  body  radiation. 

a.  Properties  of  the  theoretical  blacfe  body. 

b.  Quantity  and   qnsll^  of  black   body  radiation  at  various 


>y  Google 

26  Illuminatiso  Engineehiso 

c  Ratioa  of  energy  radiated  In  visible  spectrum  to  total  cnergr 
radiated  [  ~ir  I  ^^  various  temperatures. 

d.  Ratios  of  luminous  flux  to  energ;  radiated  in  visible  spec- 

trum [>T~ }  ''  various  temperatures. 

e.  Ratios  of  luminous  Dux. to  total  energy  radiated  l^j  at 

various  temperatures. 

f.  Temperature   of   highest    possible    efflclency    of    black    hod^/ 

about  SOOO"  absolute. 
B.  Selective  radiation. 

a.  No  natural  body  Is  absolutely  "black." 

1.  Difference  In  emiaiivitj^—"  gray  "  bodies. 

2.  Difference  In  spectral  distribution — "  selective  "  bodies. 

b.  Qrajf  bodies  have  same  efflclency  as  blocft  bodies  at  same 

C.  Selective  bodies  may  have  higher  etBclency  than  blacle  body 

at  same  temperature. 
d.  Metallic    filaments    as    a    rule    owe    efficiency    In    part    to 

6.  LnmlnescBncfl. 

A.  Accepted  definition  of  luminescence. 

B.  Query  as  to  significance  of  term  "  luminescence." 

C.  Employment  of  terms  In  present  lectures. 

D.  Types  of  luminescence. 

a.  Ctaeml-lumlnescance. 

b.  Fboto-luminescence  or  phosphorescence, 
c  Electro-luminesoence. 

Lectdsc  II 

1.  Introduction. 

2.  The  physics  of  the  electric  Incandescent  lamp. 

A.  CH  loss  In  leadtng-ln  wires. 

B.  Loss  by  thermal  conduction  and  convection  of  gas  negligible  In 

commercial  lamps. 

C.  Relation  between  loss  through  gas  and  pressure  of  gas  for  special 

platinum  filament  lamp  at  about  1700°  absolute. 

D.  Loss  by  thermal  conduction  along  leadlng-ln  wires  and  anchor 

wires  not  more  than  5^  for  commercial  tungsten  and  7%  tor 
tantalum  lamps. 

E.  Radiation  arises  from  temperature  and  not  luminescence. 

F.  Efflclency  of  metal  filament  lamp  partly  due  to  temperature  of 

operation  and  partly  to  favorable  selectivity.  Osmium  prob- 
ably most  selective  of  ordinary  filaments. 

O.  Values  of  ^  for  incandescent  lamps. 

H.  Relations  between  voltage,  cnrrent  and  candle-power  for  In- 
candescent lamps. 

>y  Google 

Physical  CHABACTESisnoa  of  Lumikous  Soubobs       37 

3.  Tbe  pbTtki  of  the  arc  lamp. 

A.  DeflnlUoD  ot  "  arc." 

B.  CharacterlBtlc8  of  arc  dlacharge. 

C.  Distribution  of  potential  in  the  arc 

a.  Pall  of  potential  at  anode. 

h.  Fall  ot  potential  alovE  vaporous  path. 

c  Pall  of  potential  at  cathode. 

D.  Sources  of  lumluouB  flux  In  the  arc 

a.  Anode  principal  source  of  luminous  flux  In  direct  current 

open  and  enclosed  arcs. 

b.  Tbe  two  electrodes  equally  the  principal  sources  of  lomlnons 

flux  In  altematluE  current  open  and  enclosed  arcs, 
c  The  luminous  vapor  the  principal  source  of  lumtnoue  flux 

In  "  luminous  "  and  "  flaming  "  arcs. 
B.  The  dtlTerence  between  "  luminous "  and  "  flaming "  arcs  lm> 

portant  from  physical  standpoint. 
F.  Is  luminosity  of  gas  to  be  ascribed  to  selective  temperature 

radiation  or  to  so-called  "  luminescence  "T 
O.  Probable  temperatures  of   anode,   cathode  and   vapor  Id  open 

carbon  arcs. 
H.  Conduction  and  convection  lossea  In  arc  lamp  not  accurately 


I.  Values  of-n-  and  ^for  various  types  of  arc  lampa. 

4.  Tbe  physics  of  low  pressure  arcs  and  vacuum  tubes. 

A.  Distinction  between  arc  and  vacuum  tube  discharge. 

B.  The  ordinary  mercury  vapor  lamp  an  enclosed  luminous  arc  at 

low  pressure. 
a.  Efficiency  ascribed  to  luminescence  with  large  percentage  ot 
radiation  In  tbe  visible  spectrum. 

C.  The  mercury  arc  In   quartz  tube  operated  at  higher  current 

density  and  increased  efficiency, 
a.  Temperature  radiation  supposed  to  supplement  luminescence 
In  quartz  mercury  arc. 

D.  Data  on  conduction  and  convection  losses,  and  on  values  of  -g- 

for  mercury  arcs  meager. 

E.  In  vacuum  tube  discharge  the  character  of  the  light  depends  on 

nature  of  gas  between  electrodes. 

F.  Owing  to  distribution  of  potential  In  vacuum  tubes,  long  tubes 

are  necessary  for  high  luminous  efficiency. 
Q.  Luminous  efficiency  of  vacuum  Vxb%  sources  aecrlbed  to  lumi- 

5.  The  physics  of  open  flames,  and  of  tbe  Incandescent  mantle. 

A.  The  ordinary  open  flame  owes  Ite  luminosity  to  the  temperature 

of  carbon  particles  heated  to  Incandescence 

B.  Tbe  temperature  of  Bunsen  flame  about  2100'  abaolute  at  Its 

hottest  part. 

>y  Google 


C.  Tile    pecallar    rmdlatlng   propntlw   ol    rare    eartha   and    th«ir 


D.  HTpotheses  ttaat  hare  been  adranced  to  acconnt  tor  hl^  eflt 

clencf  of  nuntlea. 

a.  Lamlnescence. 

b.  Localized  blgli  temperature  due  to  catalTtls. 

c.  Selective  emlaalon  at  temperature  coaaUtent  irltb  tliat  of 

Bunsen  flame. 
H.  Uoet  geaerallr  accepted   tlieor?  at  present  that  giTen   nnder 

D — c.  but  quest  Ion  still  in  doubt 
F.  Peculiar  pbeoomMia  of  mixtures  of  tboria  and  cerla  explained 

on  baalB  of  relative  emlettTltlea  and  selectivitles  of  the 

two  Bubatances. 
O.  Eatimates  of  temperature  of  Incandescent  mantle. 
H.  The  luminous  efllciencf  o(  mantle  and  values  of  -j^-  ■ 

I.  Temperature  of  acetylene  flame. 

J.  The  luminous  efficiency  of  acetylene,  and  the  value  of  -g-  ■ 

6.  The  physics  of  the  Nernst  glower. 

A.  The  glower  a  "  solid  electrolyte,"  composed  of  oxides  of  rare 


B.  Conduct lOD,  convection  and  other  losses. 

C.  Probable  temperature  of  glower. 

D.  The  luminous  efflclency  of  the  glower  and  the  value  of  -^  ■ 

7.  The  physics  of  the  flre-fly  and  other  light-producing  organisms. 

A.  The  high  efBciency  of  the  flre-fly  due  to  extremely  selective 

luminescent  radiation. 

B.  Light-giving  properties  of  bacteria  and  other  organisms. 

S.  The  distribution  of  energy  In  the  spectra  of  the  various  luminous 

A.  Spectra  of  gases,  liquids  and  solids. 

a.  Unique  spectra  of  rare  earths. 

B.  Energy  dlBtrlbutlon  In  visible  spectrum  of  ordinary  illuminants. 

C.  Energy    distribution    in    infra-red    spectrum   of  ordinary    Illu- 

ft.  The  quality  of  light  from  the  various  luminous  sources. 

A.  Integral  color  and  continuity  of  visible  spectrum. 

B.  Coiorlmetric  measurements  of  ordinary  Uluminanta. 

Lectdrb  I 
1.  Introduction 
The  sensation  of  ligkt  is  produced  normally  when  radiant  energy 
transmitted   through   the   luminiferons  ether  in   electro-magnetic 
waves  of  sufficieDt  amplitude,  and  within  certain  limits  of  wave- 
length impinge  upon  the  retina  of  the  eye.     It  is  necessary  to 

>y  Google 


keep  Id  mind  that  the  nltiinate  object  of  every  luminoufl  source  is 
to  produce  the  eenution  of  light,  and  that  iiierefore  the  relation 
between  the  psycho-phyBiological  sensation  and  the  physical  stimulus 
furnishes  a  fundamental  criterion  in  an  anaiyaiB  of  the  physical 
characteristics  of  luminous  source. 

Hoverer,  the  first  condition  to  be  fulfilled  by  a  laminous  source 
is  that  it  radiate  energy  within  the  limits  of  the  Tisible  spectrum. 
This  is  the  initial  condition,  but  there  are  many  other  conditions, 
physical  and  non-physical,  scientific  and  aesthetic,  which  determine 
the  real  efficiency  of  a  luminons  source,  where  by  efficiency  is  meant 
the  d^ree  of  adaptability  to  the  required  end.  From  a  physical 
standpoint,  the  energy  relations  in  the  production  of  luminous 
energy  are  of  prime  importance.  The  interest  centers  in  the  ef- 
ficiency of  the  tran^ormation  of  the  energy  supplied  to  the  lamp 
into  the  Ugkt  received  from  it. 

A  definite  amount  of  what  is  familiarly  termed  chemical  energy 
is  stored  up  in  the  molecules  of  acetylene  and  oxygen.  After  com- 
bustion a  smaller  amount  of  energy  is  stored  up  in  the  resultant 
molecules  of  COj  and  water  vapor,  a  part  of  the  residue  becoming 
available  as  light.  The  gross  efficiency  of  the  combustion  of  acety- 
lene as  a  source  of  light  is  the  ratio  of  the  light  produced  to  the 
energy  stored  up  in  tiie  molecules  of  acetylene  and  oxygen  before 
combustion,  the  two  being  measured  in  appropriate  units.  The 
energy  stored  up  in  the  resultant  molecules  of  CO2  and  water  vapor 
may  be  considered  as  waste  ao  far  as  the  present  transformation 
is  concerned. 

This  example  illustrates  chemical  rather  than  physical  relations 
in  transformation  of  energy,  but  serves  to  show  that  in  many  cases 
the  two  are  intimately  interconnected.  Judged  from  a  purely 
physical  aspect  the  efficiency  of  the  acetylene  lamp  depends  en- 
tirely upon  the  ratio  of  the  light  produced  to  the  energy  liberated 
in  the  chemical  transformation.  Thus  some  of  the  energy  is  dis- 
sipated by  conduction,  some  by  convection  and  some  by  radiation. 
Of  the  latter  a  relatively  small  part  is  available  as  light.  The 
matta%  of  fundamental  importance  to  the  physicist,  therefore,  are 
the  relations  of  the  energy  dissipated  by  conduction  and  convection 
to  that  radiated,  tlie  spectral  distribution  in  the  radiant  energy,  and 
the  causes  which  determine  these  relations. 

The  incandescent  electric  lamp  furnishes  an  interesting  illustra- 
tion.   A  definite  amount  of  energy  per  second  is  supplied  electrically 

>y  Google 

30  Illdminatinq  Enqineerinq 

to  the  terminals  of  the  lamp.  A  part  of  this  is  transfoTmed  into 
heat  by  the  CK  loss  in  the  leading-in  wires  and  jnnctionB.  The 
remainder  is  traneformed  into  heat  by  the  passage  of  the  current 
through  the  high-reaistanee  filament.  That  which  is  transformed 
into  heat  by  the  C^R  losa  in  the  leading-in  wires  is  completely  lost, 
as  far  as  its  direct  influence  on  the  luminous  efBcieney  of  the  lamp 
is  concerned.  This  loss  in  the  ordinary  types  of  lamps  manufac- 
tured at  the  present  time  is  negligibly  small,  amounting  in  most 
cases  to  leas  than  1  per  cent. 

The  energy  which  is  transformed  into  heat  in  the  filament  ia 
dissipated  in  various  ways,  only  a  small  part  of  it  ultimately  be- 
coming available  for  the  production  of  light  A  part  of  the  energy 
ia  dissipated  by  conduction  and  convection  by  the  gasea  in  the  bulb 
in  cases  where  the  vacuum  is  not  high,  but  this  loss  in  a  good  lamp 
is  entirely  negligible.  Another  portion  of  the  energy  is  dissi- 
pated through  heat  conduction  by  the  leading-in  and  anchoring 
wires.  Thus,  owing  to  the  high  temperature  of  the  filament  com- 
pared with  that  of  the  leading-in  and  supporting  wires  with  which 
it  comes  into  contact,  there  is  a  continual  heat  conduction  away 
from  the  filament  at  these  points,  thus  cooling  the  filament  locally 
and  decreasing  its  luminous  ^ciency. 

The  remainder  of  the  energy  transformed  in  the  filament  is 
radiated,  the  apectral  distribution  depending  upon  the  temperature 
of  the  filament.  Only  that  portion  which  is  radiated  in  waves 
within  the  limits  of  wave-length  of  the  visible  spectrum  is  pro- 
ductive of  light.  As  stated  above,  the  loss  due  to  conduction  and 
convection  by  the  gas  in  a  normal  lamp  must  be  negligibly  small. 
It  is  quite  a  simple  matter,  however,  to  show  what  a  saving  is 
effected  in  the  case  of  an  ordinary  incandescent  lamp  through  the 
use  of  an  evacuated  bulb.  If  a  lamp  is  constructed  having  a  fila- 
ment of  some  material,  such  as  platinum,  which  can  be  operated 
either  in  air  or  in  a  vacuum,  the  difference  in  power  supplied  to  the 
lamp  when  evacuated  and  when  filled  with  air,  the  temperature 
of  the  filament  being  the  same  in  the  two  cases,  is  quite  large. 
Thus  a  platinum  filament  of  0.1mm.  diameter  and  15  cm.  length, 
mounted  in  a  pear-shaped  bulb  of  8  cm.  maximum  diameter  and 
13  cm.  length,  when  operated  at  a  temperature  of  approximately 
1700°  Ahs.  (Centigrade-|-373°),  requires  4.75  watts  when  the  bulb 
is  evacuated,  and  24.3  watts  when  filled  with  air  at  atmospheric 
pressure.     In  other  words,  the  loss  by  convection  and  conduction 

>y  Google 

Physical  Chabactebistics  of  Luuinous  Sources        31 

of  the  gas  la  400  per  cent  of  the  total  power  required  to  operate 
the  filament  in  a  vacuum. 

The  loBses  by  couductioa  at  the  leading-in  and  anchoring  wireB 
have  been  varionBly  estimated,  the  valnee  found  ranging  from  an 
almost  negligible  quantity  to  as  high  as  25  or  50  per  cent  in  various 
types  of  standard  lamps.*  Attcmpte  at  direct  measurement  of  the 
energy  radiated  seem  to  indicate  comparatively  high  figures  for 
the  thermal  conduction  losses,  whereas  the  conclusion  from  prac- 
tical experience  in  lamp  manufacture  points  to  rather  small  losses. 
Preliminary  measurementa  by  a  new  direct  method  gave  for  these 
losses  for  normal  carbon,  tantalum  and  tungsten  lamps  values  in 
all  cases  of  the  order  of  magnitude  of  5  per  cent,  which  would  seem 
to  be  more  consistent  with  the  experience  of  lamp  manufacturers 
than  the  much  larger  losses  found  by  other  investigators. 

If  then  the  losses  by  convection  and  conduction  amount  to  but  a 
small  percentage  of  the  total  energy  supplied  to  the  filament,  ez- 
planatioD  of  the  relatively  low  luminous  efficiency  of  the  lamp  must 
be  sought  in  the  spectral  distribution  of  the  radiated  energy. 

2.  Luminous  Efficiency 

Of  the  energy  radiated  by  a  luminous  source  only  that  portion 
which  lies  within  the  wave-length  limits  of  visihility  produces  the 
sensation  of  light.  Even  within  these  narrow  limits  the  Intensity 
of  the  sensation  varies  greatly  with  the  wave<length  when  the  retina 
is  excited  with  equal  quantities  of  energy.  Thus  a  quantity  of 
energy  which  in  the  deep  red  or  extreme  violet  is  scarcely  sufficient 
to  be  visible,  would  in  the  yellow  or  green  regions  of  the  spectrum 
produce  a  moderately  strong  sensation. 

The  extreme  wave-lengths  which  mark  the  limits  of  the  visible 
spectrum  are  somewhat  variable,  depending  on  the  individual.  For 
normal  eyes  radiant  energy  between  the  limits  of  wave-lengths  of 
0.8  fi  (fi=0.001  mm.)  on  the  red  side  to  a  little  less  than  0.4  fi  on 
the  violet  side  produces  the  sensation  of  light.  With  moderately 
intense  sources  the  eye  can  perceive  rays  of  wave-lengths  down  to 
0,38  p.,  but  there  is  no  sense  of  color  beyond  0.4  fu 

The  energy  contained  in  the  visible  spectrum  of  the  radiation 
from  an  ordinary  solid  at  ordinary  temperatures  comprises  but  a 
very  small  fraction  of  the  totel  energy  radiated.  Beyond  the  visible 
on  the  red  side,  the  infra-red  spectrum  extends  from  0.8  /x  to  in- 
definitely longer  wave-lengths,  which  have  been  isolated  and  stndied 

>y  Google 

38  Illduinatiko  Enojnbkumo 

np  to  96.7  pJ  It  is  in  this  region  that  in  most  cases  tte  great  bulk 
of  radiant  energy  is  emitted.  Thus,  in  the  ctM  of  the  tungBteo 
lamp  about  95  per  cent  of  the  energy  radiated  by  the  filament  ia 
emitted  in  the  form  of  heat  rays  of  wave-lengths  too  long  to  eicite 
the  human  retina. 

Beyond  the  visible  spectrum  on  the  violet  aide  the  vltra-violet 
Bpectnim  estenda  from  about  0.4  fi  or  0.38  ;i  to  indefinitely  shorter 
wave-lengths  which  have  been  isolated  and  studied  down  to  0.1  ^ 
The  energy  radiated  in  the  ultra-violet  region  of  the  spectrum  ia 
for  all  ordiiury  sources  very  small,  even  compared  with  that  radiated 
in  the  visible  spectrum,  and  may  generally  be  neglected  in  the  fol- 
lowing discoeaion. 

It  has  been  stated  that  the  energy  radiated  in  the  infra-red  and 
ultra-violet  regions  of  the  spectrum  doea  not  conduce  to  tiie  sensa- 
tion of  light,  and  that  even  within  the  narrow  limits  of  wave- 
length comprising  the  visible  spectrum  equal  quantities  of  energy 
in  different  portions  of  the  visible  spectrum  do  not  produce  the 
same  intensity  of  sensation.  It  is  of  much  interest,  therefore,  and 
most  pertinent  to  the  question  of  the  efficiency  of  light  sources^  to 
consider  briefly  the  relation  between  the  energy  of  the  stimulua  and 
the  intensity  of  the  resultant  sensation  for  the  various  wava-lengths 
lying  within  the  limits  of  the  visible  spectrum. 

At  the  outset  it  is  necessary  to  note  that  the  intensity  of  the 
saisaticm  does  not  depend  solely  on  the  intensity  of  the  stimulus, 
even  for  any  one  wave-length.  The  time  interval  during  which 
the  stimulua  acts  determines,  to  some  extent,  the  intensity  of  the 
sensation.  There  ia  a  lower  limit  to  the  duration  of  the  stimulus, 
below  which  no  sensation  is  produced.  As  this  time  interval  ie 
increased  the  sensation  rises  rapidly  for  some  wave-lengths  even 
beyond  that  of  perman^t  regime  and  then  falls  again  to  what  has 
been  termed  the  permanent  regime,  or  normal  sensation.  All  of 
this  occurs  within  a  fraction  of  a  second.  After  the  retina  has 
been  exposed  for  a  long  time  to  a  constant  stimulus,  the  sensation 
gradually  decreases  owing  to  fatigue.  The  elemrat  of  time,  there- 
fore, plays  an  important  rSle  in  determining  the  intensity  of  sen- 
sation for  a  giren  stimulua. 

There  is  a  second  element  which  should  be  mentioned  at  the 
beginning  as  determining  the  relation  between  the  intensity  of 
the  sensation  and  the  intensity  of  the  etimulns  for  different  wave- 
lengths.   If  there  have  been  found  two  quantities  of  energy  in  the 

>y  Google 


red  and  blue  ends  of  the  visible  Bpectrum,  respectively,  which  pro- 
duce equivalent  inteneitiee  of  eenBation  where  the  absolute  intensity 
of  sensation  is  low,  it  does  sot  follow  that  the  two  sensatione  will 
remain  equivalent  if  the  quantities  of  energy  are  greatly  increased, 
even  though  each  is  increased  by  the  same  relative  amount.  The 
red  sensation  at  the  higher  intensity  would  be  relatively  larger. 
This  phenomenon  is  familiarly  known  as  the  Purkinje  effect,  and 
may  be  stated  in  general  as  follows:  The  relative  intensitiee  of 
sensation  for  equal  energy  excitation  in  different  portions  of  the 
visible  spectrum  d^nd  upon  the  absolute  magnitude  of  the  ^lergy 
stimuli.  In  other  words,  the  relation  between  the  increase  in  sen- 
sation and  the  iDcrease  in  stimulus  is  not  the  same  for  different 
wave-lengths  in  the  visible  spectrum. 

In  addition  to  these  two  elements  of  interval  of  duration  and 
absolute  magnitude  of  the  stimulue  in  determining  the  relative 
sensationE  produced  by  equal  quantities  of  energy  in  the  different 
portions  of  the  visible  spectrum,  there  are  other  psycho-phy Biological 
elements  which  will  not  even  be  mentioned  here.  Moreover,  the 
two  elements  which  have  been  described  briefly  will  not  be  con- 
sidered farther  in  the  discussion.  It  will  be  assumed,  (1)  that 
in  every  case  the  stimuli  act  over  a  suiBciently  long  interval  to 
produce  the  normal  sensations  of  permanent  regime;  (3)  that  the 
absolnte  magnitudes  of  the  stimuli  are  always  moderately  large, 
since  it  is  only  at  relatively  low  intensities  of  illumination  that 
the  Purkinje  effect  is  distinctly  noticeable. 

What,  then,  under  normal  conditions,  is  the  relation  between  the 
intensity  of  the  stimulus,  and  the  intensity  of  the  sensation  in 
different  portions  of  the  visible  spectrum?  The  answer  is  given  in 
Figure  1. 

The  BO-called  sensibility  curve  which  gives  this  relation  is  com- 
monly obtained  by  determining  the  quantity  of  energy  per  second 
necessary  in  different  portions  of  the  spectrum  to  produce  the  same 
luminosity,  i.  e.,  the  same  intensity  of  sensation.  The  reciprocals 
of  these  quantities  of  energy  are  then  plotted  as  the  sensibility 
curve.  The  curve  obtained  in  this  way  is  shown  in  Figure  1. 
Neglecting  the  variations  caosed  by  the  Purkiaje  phenomenon,  the 
relative  candle-powers  of  two  sources  may  be  computed  by  multi- 
plying the  ordinates  of  the  spectral  energy  curves  of  the  two  sources 
by  the  ordinates  of  the  sensibility  curve,  and  comparing  the  areas 
enclosed  by  the  two  luminosity  curves  thus  obtained. 

>y  Google 

34  Illuuinattng  Enqineebtnq 

Luminoeity  curres  obtained  in  this  way  foi  a  nnmbcr  of  common 
light  sources  are  given  in  Figure  3,  Curves  a,  b,  c,  etc.,  are  the  Bpec- 
tral-energy  cnrrea  for  the  3.1  w.  p.  c.  carbon  lamp,  the  1.25  w.  p.  c, 
tungsten  lamp,  the  Nernst  lamp,  and  the  Welsbach  mantle  {99.35 
per  cent  thoria,  0.75  per  cent  ccria)  and  curves  a',  b',  e',  etc.,  are 
the  corresponding  luminosity  curves,  i.  e.,  the  cun'es  showing  the 
relative  intensities  of  sensation  produced  in  different  parts  of  the 
epectnim.  The  energy  curves  are  so  drawn  that  the  total  energy 
in  the  visible  spectrum  (taken  arbitrarily  for  this  particular  il- 
lustration as  extending  between  the  limits  of  wave-length  X=0.?0  p. 

Fra.  1. — So-Called  Sensibility  Curve. 
(Luminoeity  Curve  for  E^]ual  Elnergy  Distribution.) 

on  the  red  side  to  A=0.43  /i  on  the  violet  side)  is  the  same  for  all. 
In  other  words,  the  areas  enclosed  by  the  energy  curves  and  the 
asis  of  abscissas,  between  the  two  limiting  ordinates,  are  equal. 

It  is  seen  from  an  inspection  of  the  luminosity  curves  a',  b',  etc., 
that  although  the  eye  has  its  maximum  sensibility  at  A  — 0.545/1, 
the  wave-length  of  maximum  luminosity  for  most  sources  is  shifted 
well  toward  the  red  end  of  the  spectrum,  owing  to  the  predominance 
of  energy  in  the  longer  wave-lengths.  Moreover,  the  wave-lengths 
of  maximum  luminosity  for  the  various  sources  are  somewhat  differ- 
ent, as  are  also  the  shapes  of  the  luminosity  curves,  owing  to  the 
different  distributions  of  energj'  in  the  spectra  of  the  various 

>y  Google 

Physical  Chaeacteristics  of  Ldminods  Sodbces        35 

The  literature  on  the  efBciency  of  various  light  sources  contains 
many  reports  of  determinations  of  the  mechanical  equivalent  of 
tight '  where  by  this  term  is  meant  the  energy  per  second  within 
the  limits  of  the  visible  spectrum  which  will  produce  a  unit  flux 
of  tight,  measured  photometrically — in  other  words  the  watts  per 

Fig.  2. — Energ?  and  Luminosity  Cnrres  for  Various  Light  Sources. 

lumen.  The  determination  of  the  mechanical  equivalent  of  light 
is  an  attempt  to  eorrellate  flux  of  energy,  measured  in  watts,  with 
the  resultant  sensation  produced,  "measured  in  light  unit!).  As  ordi- 
narily determined,  however,  it  is  subject  to  criticism  in  two  re- 
spects: (1)  the  value  found  for  any  light  source  depends  upon  the 
wave-lengths  arbitrarily  chosen  as  limiting  the  visible  spectrum; 
(2)  for  any  definitely  chosen  limits  of  wave-length,  the  value  de- 

>y  Google 


pends  on  the  light  sources  used.  Both  deficiencies  arise  fanda- 
mentally  from  the  same  canse,  viz.,  that  the  mechanical  equivalent 
of  light  ia  different  for  every  color  or  wave-length,  and  therefore 
has  definite  significance  only  as  applied  to  light  of  some  one  wave- 

Mkchahical  EqmvAiBiTTfl  or  Ltort  as  Otten  fob  Ssnoui,  Illumihartb 

(Wave-length  limits  taken  as  0.38  /i  and  0.76  p.) 
8ou«e  Authorii,  ^t^^'^ 

Hefner  Ingitr&m  .0096 

Arc  Drysdale  .0064 

Nernet  "  ,0095 

Black  bodr  at  6000°  Aba Ives  .0080 

Ideal  yellow-green  light "  .0012 

In  Figure  3,  curves  a',  b',  etc.,  are  luminosity  curves  correspond- 
ing to  equal  quantities  of  energy  between  the  wave-lengths  A=0.70/i 
and  A  =  0.43;i.  The  areas  enclosed  by  these  luminosity  curves  and 
the  axis  of  abscissas  taken  between  the  two  limiting  ordinates  might 
be  taken  as  giving  tiie  relative  values  for  the  mechanical  equivalent 
of  light  obtained  from  these  sources.  The  values  thus  obtained, 
however,  are  not  comparable  with  those  usually  given,  because 
ordinarily  the  limits  of  wave-length  are  taken  0.76  n  and  0.38  p.. 
But  although  it  is  true  that  visibility  extends  to  these  wave-length 
limits,  the  luminosity  at  the  two  ends  is  so  small  that  it  might  be 
neglected.  On  the  other  hand,  energy  at  the  red  end  between 
0.70  p.  and  0.76  p.  would  constitute  for  most  sources  a  large  per- 
centage of  tie  total  energy  in  the  visible  spectrum.  In  Figure  2 
the  narrow  limits  are  taken,  because  for  some  of  the  sources  given 
accurate  energy  curves  to  the  larger  limits  were  unobtainable. 

In  Table  I  some  values  of  the  mechanical  equivalent  of  light 
for  vaiiouB  sources,  and  between  the  customary  limits  of  0.76 1* 
and  0.38^1  are  quoted. 

Although  the  uncertainty  in  the  actual  values  given  may  be 
great  owing  to  the  difficulty  of  measuring  accurately  by  objective 
methods  the  small  quantity  of  energy  in  the  visible  spectrum,  the 
differences  in  the  relative  values  obtained  for  different  light  sources 
are  largely  due  to  the  fact  that  the  mechanical  equivalent  for  every 
different  light  source  having  a  different  spectral-energy  distribu- 
tion ia  necessarily  different. 

>y  Google 

Physical  CHAEACTBSianoe  of  Luminous  Soubcss        37 

A  much  better  definition  of  the  tenn  "  mechanical  equivalent  of 
light"  would  be  the  flax  of  energy  (in  watts)  for  some  definite 
ware-length — preferably  the  wave-length  of  mazimiim  sensibility 
(A=0.&45^) — that  produces  a  unit  flux  of  Light,  measured  photo- 
metrically (one  lumen).  From  the  best  determinationa  of  ihvB 
quantity  up  to  the  present  time  the  most  probable  value  for 
X=0.5i5fi.  is  of  the  order  of  magnitude  of  800  lumens  per  watt, 
or,  as  more  conunouly  expressed,  0.015  watt  per  mean  spherical 
candle,  though  these  values  may  be  in  error  many  per  cent  Hence 
the  moat  eflOcient  light  source  that  could  be  imagined  would  be  one 
in  which  all  the  energy  supplied  to  the  lamp  is  transformed  into 
radiation,  and  all  this  radiant  energy  is  concentrated  in  light  of 
that  wave-length  (A=0.545/«)  to  which  the  human  eye  responds 
most  intensely.  The  efl&ciency  of  this  source  would  be  800  lumens 
per  watt,  or  0.015  watt  per  mean  spherical  candle. 

This  light,  of  the  single  wave-length,  A=0.545  n,  would,  of  course, 
not  be  white.  Its  color  would  be  yellowish-green,  and  it  would  be 
very  unsuitable  for  ordinary  illumination  both  on  account  of  its 
own  color  and  also  because  of  the  unnatural  appearance  objects 
illuminated  by  it  would  assume.  The  question  naturally  arises, 
"What  would  be  the  highest  possible  efficiency  of  white  light,  if 
all  the  energy  supplied  to  the  source  was  transformed  into  radiation, 
and  all  this  radiant  energy  was  concentrated  within  the  limits  of 
the  visible  spectrum  in  such  a  way  a&  to  produce  white  light,  where 
by  white  light  is  meant  a  distribution  of  energy  in  the  visible  spec- 
trum similar  to  that  in  the  spectrum  of  average  noon-time  sun- 
light?" The  answer  to  this  question  is  approximately  300  lumens 
per  watt,  or  about  one-third  the  eflSciency  of  the  most  efficient  mono- 
chromatic light  Expressed  in  watts  per  candle  Hie  specific  con- 
sumption of  the  most  efficient  white  light  would  be  about  0.04  watt 
per  mean  spherical  candle.  If  the  limits  of  the  visible  spectrum 
were  taken  as  0.70  fx  and  0.43  ^  the  corresponding  flgures  would  be 
400  lumens  per  watt  or  0.03  watt  per  mean  spherical  candle. 

Compared  with  this  the  efficiency  of  those  ordinary  illuminants 
for  which  we  can  measure  the  power  supplied  in  iratts  is  extremely 
low.  The  flaming  arc  has  an  efficiency  of  about  50  lumens  per 
watt,  or  0.25  watt  per  mean  spherical  candle.  The  tungsten  lamp 
has  an  efficiency  of  8  lumens  per  watt,  or  approximately  1.6  watts 
per  mean  spherical  candle  (1.25  watts  per  mean  horizontal  candle). 
Expressed  in  another  way,  if  the  efficiency  of  the  most  efficient 

>y  Google 


monochromatic  Bource  is  taken  as  100  per  cent,  the  efficiency  of  the 
most  efficient  white  light  is  approximately  15  per  cent,  the  efficiency 
of  the  flaming  arc  and  tungsten  incandeBcent  lamp  are,  respectively, 
(j  per  cent  and  0.9  per  cent. 

The  reasons  for  the  relatively  low  efficiencies  of  artificial  sources 
compared  even  with  the  most  efficient  white  light  are  threefold,  as 
indicated  in  the  illustration  of  the  incandescent  lamp  given  in  an 
earlier  paragraph.  (1)  Not  all  the  energy  supplied  to  the  lamp  is 
transformed  into  radiation;  soibe  ia  lost  by  conduction  and  con- 
vection. (3)  Only  a  small  part  of  that  radiated  is  contained  in 
the  visible  spectrum.  Much  is  emitted  in  waves  too  long  to  affect 
the  human  eye.  (3)  That  part  of  the  radiant  energy  which  is 
contained  within  the  visible  spectrum  is  not  distributed  most  ad- 
vantageously. Granting  that  conduction  and  convection  losses  could 
he  eliminated,  the  spectral  distribution  of  the  radiant  energy  is  an 
ontstanding  factor  to  be  reckoned  with. 

If  we  confine  our  attention  to  the  case  of  the  simplest  radiating 
solid,  viz.,  the  black  body  {which  see),  the  only  opportunity  offered 
to  change  the  spectral  distribution  is  the  variation  of  the  tempera- 
ture of  the  radiator.  It  can  readily  be  shown  that  if  all  the  energy 
supplied  to  a  lamp  was  radiated  in  the  continuous  spectrum  of 
black-body  radiation  corresponding  to  the  temperature  of  highest 
efficiency,  the  efficiency  of  the  lamp  would  be  approximately  140 
lumens  per  watt,  or  16  per  cent  of  the  highest  possible  efficiency  of 
the  most  efficient  monochromatic  light,  to  which  we  assigned  the 
arbitrary  value  of  100  per  cent  efficiency.  In  other  words,  this 
most  efficient  black-body  radiator  would  be  IS  times  as  efficient  as 
the  tnngsten  lamp.  In  passing,  it  is  significant  that  the  tempera- 
ture of  the  black-body  under  this  condition  is  that  corresponding 
roughly  to  the  temperature  of  the  sun.  In  other  words,  a  black 
body  at  the  temperature  to  produce  white  light,  is  at  the  temperature 
of  maximum  efficiency  for  pure  temperature  radiation. 

In  the  literature  on  the  general  subject  of  luminous  efficiency,' 
various  phrases. indicating  different  ratios  of  power  and  light  have 
been  invented  and  used.  The  confusion  that  has  resulted  from  the 
use  of  quite  similar  terms  to  signify  distinctly  different  quantities 
suggests  in  the  present  treatment  the  confinement  to  an  explana- 
tion of  the  important  quantities  involved,  without  any  extended  use 
of  the  complicated  nomenclature.  One  of  the  more  common  ex- 
pressions, the  mechanical  equivalent  of  light,  has  been  referred  to 

>y  Google 

Physical  Chabactkhisticb  op  Ldmisous  Sodrceb        39 

already.  But  even  this  term,  as  was  pointed  out,  is  indefinite  and 
unsatisfactory  as  ordinarily  used.  By  the  mechanical  equivalent 
of  light  is  meant  the  light  value  of  radiant  energy,  where  only  that 
Tadiant  energy  is  included  which  may  call  forth  the  seneation  of 
light,  i.  e.,  that  portion  of  the  radiant  energy  which  lies  within 
the  limits  of  the  visible  spectrum. 

As  has  been  stated  in  a  previous  paragraph,  the  indefiniteneas 
in  the  mechanical  equivalent  of  light  arises  from  the  fact  that 
the  mechanical  equivalent  for  every  wave-length  of  light,  and  hence 
for  every  different  composite  light,  is  different.  The  light  value 
for  any  one  wave-length  is  much  more  definite,  and  the  determina- 
tion of  the  light  value  for  energy  of  the  wave-length,  X=0.645p, 
of  maximum  sensibility  (at  high  intensities)  is  of  prime  importance 
aa  indicating  the  upper  limit  of  efBciency  theoretically  obtainable. 
This  quantity,  which  may  be  denoted  by  M,  is  not  known  accurately, 
but  has  an  approximate  value  of  800  lumens  per  watt,  or  0.015 
watt  per  mean  spherical  candle. 

The  principal  relations  which  excite  interest  in  a  study  of  the 
efficiency  of  light  sources  may  be  stated  briefly.  Given  a  definite 
quantity  of  energy  supplied  to  a  lamp:  (1)  What  proportion  of 
that  energy  is  transformed  into  radiation,  and  what  part  is  dissi- 
pated in  other  ways,  being  thus  lost  so  far  as  ita  light-producing 
power  is  concerned?  (2)  Of  that  energy  transformed  into  radia- 
tion, what  proportion  is  contained  within  the  limits  of  the  visible 
spectrum,  and  is  thus  productive  of  light  in  varying  degrees? 
(3)  What  is  the  ligh1>giving  power  of  that  energy  radiated  within 
the  wave-length  limits  of  the  visible  spectrum,  or,  in  other  words, 
what  is  the  mechanical  equivalent  of  light  tor  that  particular  lamp? 
Of  these  three  relations  the  first  is  quite  definite  and  of  consider- 
able importance,  whereas  the  second  and,  consequently,  the  third 
are  more  or  less  indefinite  owing  to  the  ill-deSned  limits  of  the 
visible  spectrum.-  If  the  red  end  of  the  spectrum  is  taken  as  0.8  pi 
instead  of  0.76  /t  no  appreciable  difference  would  be  observed  in 
the  light  flni  owing  to  the  almost  negligible  luminosity  of  energy 
between  these  limits  of  wave-length.  But  the  amounts  of  energy 
ascribed  to  the  visible  spectrum  in  the  two  cases  would  be  different 
by  many  per  cent  for  most  ordinary  light  sources.     ' 

The  greatest  interest,  from  a  practical  standpoint,  centers  not 
in  the  individual  steps  of  the  above  analysis,  but  in  the  resultant 
ratio  of  luminous  flax  available  from  a  lamp  in  proportion  to  the 

>y  Google 

40  iLLOHiNATtxa  Enginesbikq 

power  Bupplied  to  the  lamp.  Th«  various  steps  in  the  analyata  are, 
however,  of  considerable  importance  in  indicating  for  any  light 
source  ita  most  pronounced  deficiency. 

The  various  relations  can  be  represented  briefly  by  the  use  of 
symbols.  Let  Q  be  the  power  supplied  to  a  lamp,  meaeaied  in 
watte;  B  the  power  radiated,  measured  in  watta;  L  the  power  radi- 
ated within  the  viaible  spectrum  (from  X=:0.38^  to  A=t0.76f« 
taken  arbitrarily),  measured  in  watts;  and  ^  the  luminous  Bux 
from  the  lamp  measured  in  lumens  (ir  spherical  candles).    The 

first  of  the  three  ratioa  given  above  is  represented  by  -^  ,  the  second 
by  -p- ,  and  the  third  by  5-  .    Under  Q,  the  power  supplied  to  the 

lamp,  would  come  the  power  lost  by  conduction  C*,  the  power  lost 
by  convection  Cb,  and  the  power  radiated  R,  so  that  Q  =  C*-|-Cp+R. 
If  the  analj'sis  is  carried  further,  as  in  the  illustration  afforded  by 
the  combustion  of  acetylene,  the  total  power  involved  in  the  re- 
action may  be  represented  by  Q',  where  Q'  =  Q  +  C»,  the  latter 
symbol  indicating  the  rate  at  which  energy   is  stored  up  in  the 

resultant  mdlecules  of  CO,  and  11,0.     Although  the  ratio -^  of 

resultant  luminous  fiux  to  total  power  involved  in  the  transforma- 
tion would  give  the  ultimate  efficiency  of  the  light  process,  such  a 
definition  would  be  comparable  with  that  for  an  incandescent  lamp 
in  which  not  only  the  heat  and  other  losses  in  the  generation  of 
electric  power,  but  even  the  chemical  reactions  in  the  fire-box  under 
the  boiler  "  are  included  in  the  energi-  supplied.  Such  an  elaborate 
analysis  would  take  us  beyond  the  logical  limits  of  a  discussion  of 
the  physical  characteristics  of  light  sources,  and  hence  will  not  be 
attempted  in  these  lectures. 

This  general  discussion  of  the  elements  entering  to  determine 
the  efficiency  of  light  sources  is  intended  to  prepare  the  way  for 
the  more  detailed  discussion  of  definite  light  sources  in  the  second 
lecture.  Under  the  treatment  of  each  source  the  data  on  efficiency 
will  be  given,  in  all  possible  cases.  Apart  from  the  mere  analysis 
of  the  efficiency  or  inefficiency  of  light  sources,  our  interest  should 
carry  U8  further  into  the  study  of  the  causes  which  underlie  the 

*  To  make  the  analogy  complete  It  would  be  necessary  to  consider  the 
energy  relations  In  the  generation  of  acetylene,  since  this  la  a  manu- 
factured product. 

>y  Google 

Physical  CHAEAOTBEiwioa  op  Ldmihocs  Sodbcbs       41 

pbetKHnena  exhibited  bj  the  tampe.  Much  valuable  knowledge 
is  gained  from  a  etud;  of  radiation,  the  laws  of  radiation  and  the 
radiating  properties  of  matter. 

3.  Qvaiiiy  of  Light 

In  the  previous  lectures  in  this  coarse,  a  discnssioii  of  color 
of  natural  objects  was  given.  In  this  discnssion  it  waa  asstuned 
that  the  incident  light  was  white  light,  nonnally  produced,  as 
in  the  case  of  Bunlight.  A  quite  different  question,  and  one  of 
distinct  importance  to  the  illuminating  engineer,  is  that  of  the 
quality  of  the  light  furnished  by  various  types  of  luminonB  eourcts.' 
The  qoalitj  of  the  light  manifests  itself  in  two  wa^:  (1)  in  the 
color  of  the  light  itself,  when  the  lamp  is  viewed  directly,  or  in 
the  apparent  color  of  white  objects  when  seen  illuminated  by  the 
light;  (S)  in  the  apparent  colors  of  various  differently  colored  ob- 
jects when  seen  illuminated  by  the  light.  Both  of  these  mani- 
festations can  be  predicted  for  any  light  when  its  spectral  composi- 
tion is  known. 

Two  lights  may  both  appear  white  and  yet  have  quite  different 
spectral  compositions.  Taking  average  mid-day  sunlight  as  stand- 
ard for  white  light,  an  ordinary  solid  body,  euch  as  carbon,  would 
emit  a  white  light  if  it  could  be  heated  to  a  temperature  of  5000* 
or  6000°.  The  spectrum  of  such  a  body  would  be  continuous,  and 
approximately  the  same  as  that  of  a  theoretical  black  bod^  (which 
see)  at  the  same  temperature. 

On  the  other  hand,  a  white  light  can  be  obtained  by  the  admix- 
'ture,  in  the  proper  proportions,  of  red,  green  and  blue  light,  if 
for  these  three  colors  the  right  wave-lengths  are  chosen,  or  by  the 
admixture  of  properly  chosen  pairs  of  spectral  colors.  From  a 
mere  visual  inspection  of  the  luminom  source  itself,  or  of  a  white 
surface  illuminated  by  it,  it  would  be  impossible  to  tell  the  true 
nature  of  the  white  light.  But  if  objecte  of  various  colors,  when 
viewed  nnder  normal  daylight,  are  illuminated  successively  by  the 
light  from  the  two  apparently  white  sources,  (hey  would  appear 
Haite  different  under  the  two  lights.  Illuminated  by  the  white 
light  from  the  incandescent  carbon  at  high  temperature,  the  colored 
object  would  appear  the  same  as  when  viewed  in  daylight.  But 
when  illuminated  by  the  white  light  composed  of  three  primary 
colore  they  would  assume  new  and  strange  tints.  It  is  not  sufficient 
then  to  adjudge  a  light  good  or  bad  on  the  basis  of  its  composite 

>y  Google 


appearance.  A  spectroBCopic  analyBis  is  necessary  to  show  whether 
the  spectrum  is  continuouB  or  diacontinuouB,  and  if  discontinuous 
whether  the  discontinuity  consists  of  a  few  bright  linea  scattered 
through  the  spectrum,  as  in  the  case  of  the  mercury  arc,  or  of  a 
very  large  number  of  bright  lines  distributed  throughout  the  entire 
spectrum,  as  in  the  spectrum  of  CO,  at  low  pressure.  A  source 
with  a  discontinuous  spectrum  of  the  latter  type  is  for  most  prac- 
tical purposes  equivalent  to  a  source  having  a  continuous  spectrum 
of  the  same  composite  quality.  In  addition  to  the  knowledge  of 
the  spectral  distribution  in  the  light  from  two  sources  of  the  same 
composite  quality,  it  is  of  interest  to  study  the  composite  qualities 
of  the  various  illuminants.  These  differ  greatly  among  themselves, 
in  most  cases  the  light  being  distinctly  more  yellowish  than  average 
daylight  The  composite  quality  of  any  light  may  be  expressed  in 
terms  of  {he  quantities  of  tliree  primary  colors,  red,  green  and  blue, 
necessary  for  a  match  in  color  with  the  light  under  investigation, 
taking  the  quantities  necessary  to  produce  the  white  light  of  average 
daylight,  as  red  33  per  cent,  green  33  per  cent,  and  blue  33  per  cent. 
In  the  detailed  discussion  of  the  various  artificial  illuminants 
in  the  nest  lecture,  data  will  be  given  in  all  cases  where  such  ob- 
servations have  been  published,  on  the  quality  of  the  composite  light, 
as  determined  by  colorimetric  measurements,  and  also  on  the  dis- 
tribution  of  energy  in  the  visible  spectrum  as  given  by  spectro- 
photometric  analysis. 

i.  Temperature  Radiation 
Frequent  reference  has  been  made  in  the  previous  paragraphs  to 
the  various  elements  which  enter  to  determine  the  ultimate  lumi- 
nous efficiency  of  any  light  source.  As  a  first  criterion  for  high 
efficiency  it  was  foimd  that  the  losses  of  energy  by  conduction  and 
convection  should  be  as  small  as  possible,  in  order  that  most  of 
the  energy  supplied  the  lamp  should  be  transformed  into  radiation. 
But  even  though  all  the  energy  supplied  to  a  lamp  were  trans- 
formed into  radiant  energy,  the  resultant  luminous  efficiency  might 
range  from  0  per  cent  to  100  per  cent,  depending  upoin  the  distri- 
bution of  the  energy  in  the  spectrum  of  the  radiating  body.  The 
study  of  radiation — the  laws  of  radiation  and  the  radiating  prop- 
erties of  matter — is  therefore  of  prime  interest  and  -importance  in 
considering  the  physical  characteristics  of  luminous  sources. 

>y  Google 

Fhtsical  Chabaatbristios  of  Luumous  Sodrces       43 

It  is  necessary  to  distinguiph  two  kinds  of  radiation:  (1)  tern- 
peratvre  radiation,  and  (2)  lumineacence.  Every  body  radiates 
energy  at  least  in  the  form  of  heat  radiation  of  long  wave-lengthB. 
If  a  body,  during  this  process  of  radiation  does  not  change  its 
nature,  it  would  continue  to  radiate  in  the  same  way  if  its  tem- 
perature were  maintained  constant  through  the  addition  of  heat. 
Such  radiation  is  ordinarily  known  as  temperature  radiation.  On 
the  other  hand,  if  a  body  undergoes  change  during  the  process  of 
radiation,  it  wonld  not  in  general  continue  to  radiate  in  the  same 
way,  even  though  its  temperatnre  were  maintained  constant  throngh 
the  addition  of  heat.  Such  a  process  of  radiation  is  known  as 

Considering  first  temperature  radiation,  which  plays  an  important 
role  in  determining  the  lurainoua  efficiency  of  practically  all  ordi- 
nary illuminants,  and  which  determines  entirely  the  efficiency  of 
electric  incandescent  lamps,  it  is  necessary  to  introduce  the  idea 
of  the  theoretical  blaci;  body,'  or  complete  radiator,  as  it  is  some- 
times called.  All  natural  bodies  show  individual  peculiarities  in 
their  radiation,  and  it  is  therefore  desirable  to  refer  back  to  some 
simple  standard  radiator. 

We  ordinarily  call  an  object  "  black  "  which  seemingly  reflects 
little  or  none  of  the  light  incident  on  it.  Exact  measurement  would 
show  that  each  such  object  actually  does  reflect  some  light,  and, 
moreover,  in  general  that  it  reflects  i-elatively  more  energy  of  some 
wave-lengths  than  of  others.    In  other  words,  it  reflects  selectively. 

A  theoretical  black  body  absorbs  all  the  energy  of  every  wave- 
length throughout  the  entire  extent  of  the  whole  spectrom,  i.  e., 
it  reflects  none  of  the  energy  incident  on  it. 

According  to  a  law  first  formulated  by  Eirchhoff,  and  known  by 
his  name,  the  quantity  of  enet^  radiated  per  second  by  any  body 
at  any  temperature  is  proportional  to  the  absorptive  power  of  the 
body  at  that  temperature.  Thus,  given  two  bodies,  A  and  B,  such 
that  at  some  definite  temperature  the  corfBcient  of  absorption  for 
body  A  for  energy  of  some  definite  wave-length  is  double  the  cor- 
responding coefficient  for  B ;  then  at  the  same  temperatnre  body  A 
would  radiate  pes  unit  area  per  second  twice  the  amount  of  energy 
of  the  given  wave-length  radiated  by  B.  Since  the  black  body 
absorbs  all  the  energy  incident  on  it,  it  will  conversely,  at  any 
temperature,  radiate  more  energy  of  every  wave-length  per  second 
than  any  natural  body. 

>y  Google 

44  Illuuinatino  Enoinebrinq 

The  relatively  simple  properties  of  the  theoretical  b!ade  body 
have  inspired  Beveral  attempts  at  theoretical  deductions  of  the  laws 
of  black-body  radiation.  Moreover,  in  recent  years  the  theoretical 
Hack  body  has  been  quite  closely  approximated  by  the  use  of  a 
hollow  cylinder  insulated  as  far  as  possible  from  the  surrounding 
air,  and  having  a  small  aperture  at  one  end  fiirough  which  the 
radiation  from  the  interior  walls  of  the  cylinder  escapes.  Such  a 
body,  heated  uniformly  by  an  electric  current,  emits  radiation  ap- 
proximating quite  closely,  both  in  quality  and  in  quantity,  that 
emitted  by  a  true  black  body  at  the  same  temperature.  Experi- 
ments carried  out  with  radiators  of  this  type  have  corroborated  in 
a  general  way  the  black-body  radiation  laws  deduced  theoretically. 
There  still  remains,  however,  considerable  uncertainty  as  to  the 
exact  values  of  the  constants  entering  in  the  mathematical  expres- 
sions of  the  laws,  and  in  the  case  of  the  law  of  spectral  distribution 
of  energy  at  any  given  temperature  the  exact  form  of  the  law  is 
not  yet  satisfactorily  established.  As  a  discussion  of  the  laws  of 
black-body  radiation  is  included  in  another  lecture  of  this  course, 
they  will  not  be  given  here.  Constant  reference  will  be  made,  how- 
ever, to  the  properties  of  black-body  radiation  as  a  convenient 
standard  with  which  to  compare  the  radiatson  from  natural  bodies. 
Even  though  the  radiation  from  a  natural  body  may  be  due  entirely 
to  the  temperature  of  the  body,  the  quantity  and  quality  of  the 
energy  radiated  by  material  bodies  at  the  same  temperature  depend 
on  the  nature  of  the  bodies  themselves.  Only  in  the  case  of  an 
absolutely  black  body  is  the  radiation  simply  a  function  of  the 

Without  introducing  mathematical  analysis  it  is  instmctive  to 
consider  briefly  the  changes  produced  in  the  quantity  and  quality 
of  energy  emitted  by  a  black  body  corresponding  to  change  in  tem- 
perature. Such  a  consideration  will  conduce  to  a  proper  apprecia- 
tion of  the  importance  of  attaining  the  highest  possible  tempera- 
tures if  high  luminous  efficiency  is  to  be  secured.  It  has  already 
been  stated  that  the  highest  possible  efficiency  obtainable  from  a 
black  body  is  but  16  per  cent  of  the  highest  possible  efficiency  of 
monochromatic  light.  Moreover,  even  this  is  only  obtainable  at  the 
extremely  high  temperature  of  5000°  or  6000°,  a  temperature  far 
in  excess  of  any  that  has  as  yet  been  realized  in  any  lamp.  It  is 
of  interest,  therefore,  to  investigate  the  relation  between  the  lumi- 
nous eflSciency  and  the  temperature  of  a  black  body  at  various  tem- 

>y  Google 

PnTBicAL  Chahactebistiob  of  Luhinodb  Soubces       45 

peraturee.  Ttiis  relation  depends  on  the  relative  amount  of  energy 
radiated  in  the  visible  spectrum  compared  with  the  total  radiation, 
and  on  the  way  in  which  the  energy  in  the  visible  spectrum  ia 

Flo.  3. — EnerK7  Curvea  (or  a  Btack  Boay  at  Varloua  Temperatures 

At  very  low  temperatures  the  energy  radiated  per  second  in  the 
visible  spectrum  is  too  email  to  affect  the  eye.  As  the  temperature 
ia  increased  the  total  energy  radiated  increases  rapidly  and  the 
rate  of  increase  is  most  rapid  for  the  shorter  wave-lengths  such 
as  affect  the  eye.     Thus,  at  temperatures  in  the  neighborhood  of 

>y  Google 


1900°  and  2100°  Aba.,  when  the  temperature  is  increased  1  per 
cent  the  total  energ>-  radiated  is  increaBed  4  per  cent,  wbereag  the 
energy  radiated  in  the  visible  spectrum  is  increased  aboat  10  per 
cent  or  15  per  cent.  Consequeotly,  as  a  result  of  1  per  cent  rise 
in  temperature  there  is  an  increaee  of  8  per  cent  or  10  per  cent 
in  efficiency,  i.  e.,  in  lumens  per  watt  radiated. 




Fio.  4. — VKlues  o(  : 

id  g  .  for  a  BlacA:  Body  at  Various  Tern- 

In  Figure  3  are  plotted  curves  showing  the  relation  between 
energy  radiated  and  wave-length  for  various  temperatures  ranging 
from  1000°  to  8000°  Aba.  Below  1000°  Abs.  the  energy  in  the 
visible  spectrum  is  practically  negligible.  As  the  temperature  is 
increased,  relatively  more  and  more  of  the  energy  is  radiated  in 
the  visible  spectrum,  the  position  of  the  maxitnum  emisaion  shifting 

>y  Google 

Physical  CHABAOTEsiaTics  of  Luuikods  Sources       47 

constaDtiy  toward  shorter  wave-lengths.  At  a  temperatare  of 
about  6000°  the  maximum  lies  in  the  visible  spectrum,  and  ap- 
proximately at  that  wave-length  in  the  visible  spectrum  corre- 
spondiug  to  the  maximum  sensibility  of  the  eye. 

As  the  temperature  is  increased  beyond  6000°  the  maKimum  is 
displaced  still  further  toward  shorter  wave-lengths,  and  the  pro- 
portion of  energy  in  the  visible  spectrum  begins  to  decrease.  There 
is,  therefore,  for  a  black  body,  a  temperature  of  maximum  etBciency 
beyond  which  the  efBciency  falls  off  again.  Inasmuch,  however, 
as  all  illuminante  in  present  use  which  depend  on  temperature  for 
their  efficiency  are  operating  at  temperatures  very  much  below  that 
of  maximum  efficiency,  any  improvement  which  would  make  pos- 
sible the  use  of  higher  temperatures  would  conduce  to  higher 

If  the  limits  of  the  visible  spectrum  are  taken  as  X=0,76fi  on 

the  red  side,  and  A=0.38 1^  at  the  violet  end,  the  ratio-p^  of  energy 
radiated  in  the  visible  ^>ectrum  to  total  energy  radiated  by  a  bla£k 
body  at  various  temperaturee  can  readily  be  computed.  The  resulte 
of  such  a  computation  are  plotted  in  curve  "a,"  Figure  4.  The 
abscissae  are  temperatures,  and  the  ordinates  are  the  corresponding 

values  of  the  ratio  _- .  It  is  thus  seen  that  for  the  arbitrarily 
chosen  limits  of  wave-length  of  the  visible  spectrum  given  above, 
the  greatest  proportion  of  energy  is  radiated  in  the  visible  spectrum 
when  the  temperature  of  the  black  body  is  a  little  over  6000°  Aba. 

Even  at  this  temperature,  however,  the  absolute  value  of  p^  is  only 
50  per  cent. 

It  cannot  be  assumed  a  priori,  however,  that  the  highest  luminous 

efficiency  corresponds  to  the  greatest  value  of  -  - ,  for  the  dietribn- 
tioo  of  the  energy  in  the  visible  spectrum  as  compared  with  the 
sensibility  curve  of  the  eye  determines  to  a  large  extent  the  lumens 

per  watt  radiated,  i.  e.,  the  ratio  -—- .     In  other  words,  ■—-  is  the 
*-^  X  Y  >  ""^  ^^  maxi- 

mum of  the  product  will  agree  with  the  maximum  of  one  of  the 
terms  only  in  case  the  two  ratios  ^p-and—  have  their  maxima  at 
the  same  temperature. 


48  iLLnunTATiNo  Enoinekbinq 

Ab  has  already  been  explained,  -^  is  the  ratio  of  the  luminouE 
flux  measured  photometrically  to  the  coireBpoDding  energy  flnx 
(between  the  wave-lengths  A.=  0,38;i  and  A=0.76fi),  measured  in 
watts,  and  depends  for  its  value  on  the  distribution  of  the  energy 
in  the  risible  Bpectmm.  If  all  the  energy  were  concentrated  at 
the  wave-length  of  maximum  sensibility  (X=0.545;i)  the  values 
of  the  ratio  f-  vroalA  be  M=800  lumens  per  watt  Calling  this 
nuurimum  value  unity,  the  values  of  ^  corresponding  to  a  black 

body  at  various  temperatures  are  given  in  curve  *'b,"  Figure  4. 
From  an  examination  of  this  curve  it  is  seen  that  for  black-body 
radiation   the   ratio  -^  reaches  a   maximum  at  a  temperature  of 


has  its  maximum.  In  other  words,  it  bo  happens  that  for  black- 
body  radiation  the  temperature  at  which  the  largest  proportion  of 
tiie  radiant  energy  lies  in  the  visible  spectrum,  ia  also  the  tempera- 
ture at  which  the  distribution  of  energy  in  tiie  visible  ^ectrum  is 
most  favorsble  for  the  production  of  light.  Since  this  tempera- 
ture, 5000°-6000°,  is  approximately  that  of  the  sun  and  corre- 
sponds to  white  light,  the  conclusion  follows  that  owing  to  the 
peculiar  sensibility  curve  of  the  human  eye,  possibly  due  to  inherited 
ancestral  adaptation,  the  sun  is  at  the  temperature  of  highest  pos- 
sible luminous  efficiency  for  a  black-body  radiator. 

From  an  inspection  of  curve  "  b  "  in  Figure  4,  it  is  seen  that 
even  at  the  temperature  of  maximum  efficiency  the  lumens  produced 
by  one  watt  radiated  in  the  visible  spectrum  is  only  33  per  cent 
of  the  lumens  that  would  be  produced  if  all  of  the  energy  were 
concentrated  at  the  wave-length  (X=0.545fi)  of  maximum  swsi- 
hility.     Multiplying  the  ordinates  of  curves  "  a "   and  "  b "   in 

Figure  4,  there  is  obtained  the  value  of  ~  ,  I  ^  x  ■?■  =  ^  ) .  f*"" 
a  black  body  at  various  temperatures.  This  new  curve  thus  ob- 
tained is  plotted  in  curve  "  c,"  Figure  4.  It  is  seen  that  the 
maximum  efficiency  obtainable  from  a  black  body  (at  the  tempera- 
ture of  the  sun)  is  only  16  per  cent  of  the  highest  possible'  efficiency 
of  monochromatic  radiation. 

Inasmuch  as  temperature  plays  so  important  a  part  in  deter- 
mining luminous  efficiency,  it  is  of  interest  to  consider  briefly  the 

>y  Google 

PHYfliGu.  CKABAomiBTioa  OF  UnnBOCS  Soowns        49 

possibilities  in  the  emploTmeot  of  high  temperatures  ae  fixed  by 
the  melting  points  of  available  substances.  The  melting  points  of 
elementary  substances  sewn  to  follow  a  periodic  function  of  the 
atomic  weights/  Thus,  carbon,  tungsten,  tantalom  and  thorium 
all  lie  at  periodic  maxima  of  melting-point  temperatures.  The 
element  which  has  the  hi^^test  melting  point  is  carbon,  which  has 
figured  prominently  in  artificial  illuminatioQ  from  the  earliest  days, 
as  in  fiames  and  in  the  carbon  incandescent  and  arc  lamps.  But 
the  employment  of  carbon  suggests  another  consideration  which 
enters  in  the  choice  of  a  filament  for  use  in  an  incandescent  lamp. 
High  melting  point  does  not  avail  much  if  the  vapor  tension  of 
the  material  is 'so  high  that  the  filament  evaporates  rapidly  at 
moderately  low  temperatures.  This  element  conditions  the  tempera- 
ture  practicable  in  an  incandescent  lamp,  and  although  "  flawing  " 
and  "  metallizing  "  have  tended  toward  reducing  the  vapor  tension, 
and  thus  made  possible  higher  working  temperatures,  the  tempera- 
ture that  can  be  employed  is  still  low  compared  with  the  melting 
point  of  carbon. 

The  oxides  and  silicates,  particularly  of  the  rare  earth  metals, 
form  a  group  of  highly  refractory  substances  which  have  been  em- 
ployed in  lamps  (e.  g.,  the  incandescent  mantle  and  the  Nemst 
glower)  partly  because  of  their  refractoriness  and  partly  on  account 
of  the  peculiar  nature  of  Aeir  emission  spectra. 

We  have  seen  that  there  is  an  upper  limit  to  the  possible  effi- 
ciency of  a  Hack  body,  and  that  this  upper  limit  is  relatively  low. 
Moreover,  at  present  there  are  no  means  known  for  even  approach- 
ing the  temperature  at  which  the  highest  efficiency  is  secured.  What 
other  possibilities  are  there,  then,  of  obtaining  high-efficiency  lamps  ? 
Etcluding  luminescence,  which  will  be  diBcussed  later,  and  con- 
fining our  attention  to  temperature  radiation,  the  answer  to  thiE 
question,  if  there  is  an  answer,  most  be  found  in  the  phenomenon 
of  selective  radiation.' 

So  far  as  is  known,  no  body  in  nature  radiates  precisely  as  a 
black  body.  No  material  body  absorbs  all  of  the  energy  of  any 
wave-length  incident  on  it;  hence  a  black  body,  from  its  very 
definition,  must,  at  a  given  temperature,  emit  more  energy  of  every 
wave-length  than  any  other  body  at  the  same  temperature.  Con- 
sequently, a  material  body  may  differ  from  a  black  body  in  that  it 
emits  per  unit  ares  at  a  given  temperature  a  smaller  quantity,  as, 
for  example,  one-half  or  one-third  of  the  energy  of  every  wave- 
length of  that  emitted  by  the  black  body  at  the  same  temperature. 



The  energy  curw  of  ench  a  body  would  be  identical  with  that  of 
a  black  body  except  that  ite  ordinates  would  be  reduced  proportion- 
ally throughout  the  entire  apectrum.  Such  a  body  is  flometimea 
known  as  a  gray  body.  It  can  be  realized  experimentally  by  inter- 
posing between  an  experimental  blach  body  and  the  screen  on  which 
the  radiation  falls  a  rotating  sectored  disk.  If  the  total  aperture 
of  the  disk  were  180°  it  would  reduce  by  one-half  the  energy  of 
every  waTc-length  received  from  the  black  body.  If  the  aperture 
were  90°  but  one-fourth  of  the  energy  of  the  black  body  would  be 
received.  In  both  caaea  the  radiation  received  on  the  screen,  i.  e., 
the  radiation  emitted  by  the  black  body  and  sectored  disk,  con- 
sidered as  a  unit,  would  be  that  of  a  gray  body,  that  is,  the  same 
in  quality  but  less  in  quantity  than  that  of  a  black  body  at  the 
same  temperature.  It  is  evident  that  there  can  be  an  infinite 
number  of  gray  bodies  corresponding  to  a  black  body  at  any  given 
.  temperature.  The  various  gray  bodies  would  differ  from  one 
another  and  from  the  black  body  in  total  emissivity. 

The  importance  of  the  distinction  between  the  black  body  and 
the  gray  body  arises  from  the  fact  that  not  infrequently  the  emis- 
sivity of  a  substance  is  cited  in  partial  explanation  of  high  efficiency. 
It  is  true  that  some  Bubstances  which  have  low  emissivities  exhibit 
also  the  property  of  seleciivity,  which  will  be  discassed  presently, 
but  it  should  be  emphasized  that  mere  graynesa  or  difference  in 
total  emissivity  has  no  direct  influence  on  the  efficiency  of  the 
radiation  from  a  aubstance  possessing  this  property.  A  gray  body 
is  no  more  or  less  efficient  than  a  blac^  body  at  the  same  tempera- 
ture.    The  quality  of  the  radiation  is  the  same  for  both.     The 

ratios  ^  and  j   would  be  identical  for  troth.     In  the  case  of  two 

filaments  of  the  same  size,  one  black  and  one  gray,  it  would  be 
necessary,  in  order  to  bring  both  to  the  same  temperature,  to 
supply  more  energy,  say  two  or  three  times  as  much  energy  to  the 
black  filament  as  to  the  gray  one.  But  the  luminous  flux  obtained 
from  the  black  filament  would  be  twice  or  three  times  that  emitted 
by  the  gray  filament.  It  is  a  question  of  the  difference  between  a 
32  c.  p.  and  a  16  c.  p.  lamp  at  the  same  watts  per  candle. 

An  indirect  practical  advantage  in  the  use  of  substances  having 
low  emissivities  in  the  manufacture  of  lamps  is  that,  owing  to  the 
lower  emissivity,  filaments  of  larger  size  for  any  given  candle- 
power  may  be  used,  thus  making  possible  stouter  and  stronger 

>y  Google 

Physical  CEABAormisncs  of  Lumikocs  Sodbobs       51 

lamps.  Another  advantage  ia  that  the  filament  or  mantle  of  lower 
emissirity  woidd  have  a  lower  intrinnc  brightness  than  a  blaek 
filament  or  mantle  at  the  same  temperatnre. 

But  a  material  body  can  differ  from  a  black  bod;  in  its  radiating 
properties  in  another  way.  Not  only  may  the  quantity  of  energy 
emitted  be  different  from  that  of  a  black  body  at  the  same  tempera- 
ture, but  the  quality  may  also  be  different.  Thus,  if  a  body  emitted 
one-fourth  aa  much  red,  and  one-third  as  much  green,  and  one-half 
aa  much  blue  ae  a  black  body  at  the  same  temperature,  it  voyld 
not  correepond  to  a  gray  body,  but  would  radiate  in  a  way  that  is 
known  as  selective;  that  is,  it  would  radiate  relatively  more-  energy 
at  one  wave-length  than  at  another  compared  with  a  black  body 
at  the  same  temperatare.  This  type  of  selectivity  is  to  be  dis- 
tinguished from  that  kind  of  selective  radiation  exhibited  in  the 
bright  line  spectra  of  luminous  gases. 

All  Bubstances  which  have  been  investigated  show  deviations 
from  the  ideal  black  body  in  respect  both  to  the  quantity  and  the 
quality  of  the  radiation.  It  is  therefore  a  matter  of  great  interest 
and  importance  in  the  consideration  of  the  phj-eica  of  light  produc- 
tion to  study  the  radiating  properties  of  matter,  and,  if  possible, 
to  correlate  these  with  the  other  properties  of  elementary  substances.. 
Although  much  investigation  has  been  directed  toward  the  solution 
of  these  problems,  the  results  obtained  thus  far  are  relatively 
meager.  TTniversal  agreement  on  the  laws  of  black-body  radiation, 
the  simplest  case,  has  not  yet  been  peached,  and  the  investigation 
of  the  peculiarities  of  the  radiation  from  matter  is  but  just  begun. 

It  ia  beyond  the  scope  of  this  lecture  to  discuss  in  detail  the 
methods  that  have  been  employed,  and  the  results  that  have  been 
obtained  in  the  investigation  of  the  radiating  properties  of  matter. 
Closely  linked  with  the  radiating  properties  of  a  substance  are  its 
refieeting  properties,  as  formulated  in  KiTchhoff*s  law,  and  much 
valuable  information  regarding  the  radiating  properties  has  beai 
obtained  by  this  indirect  method,  supplementing  the  results  of 
direct  investigation.  The  fundamental  difficulty  in  arriving  at 
definite  conclusions  regarding  the  radiating  properties  of  matter 
at  high  temperatures  is  that  of  measuring  the  temperature.  Meth- 
ods have  been  employed,  however,  which  indicate  quite  certainly 
in  a  qualitative  way  the  relative  selectivity  of  the  radiation  from 
various  substances,  and  which,  with  the  aid  of  some  simple  and 
probable  assumptions,  give  an  idea  of  the  magnitude  of  the  dif- 

>y  Google 

S&  Illduinativg  Engikb^so 

It  is  perhaps  worth  while  to  suggest  at  this  time  the  practical 
importance  of  eelective  radiation.  We  have  seen  ttiat  differencee 
in  emissivity  have  only  an  indirect  effect  in  the  production  of 
efQcient  lampg.  A  gray  body  ia  no  more  or  less  efficient  than  a 
black  body  at  the  same  temperature.  Selectivity,  on  the  other  hand, 
plays  a  direct,  and  possibly  in  certain  cases  an  important,  rSle  in 
determining  the  efficiency  of  a  lamp.  Given  a  number  of  aoh- 
stances  which  can  be  operated  at  the  same  temperature,  that  anb- 
stance  would  be  most  efficient  as  a  luminous  radiator,  other  things 
being  equal,  which  radiated  the  largest  percentage  of  energy  in  the 
viaible  spectrum  and  with  the  energy  in  the  visible  spectrum  so 
distributed  as  to  produce  the  greatest  light  effect.  If  a  substance 
could  be  found  which  radiated  aU  the  energy  in  the  visible  spec- 
trum, and  so  distributed  as  to  produce  white  light,  an  ideal  lamp 
would  result.  No  substance  has  been  found  in  which  these  condi- 
tions are  approached,  but  investigation  has  shown  that  for  some 
substances,  e.  g.,  the  filament  of  the  osmium  lamp  (which  see), 
selectivity    is    of    quite    appreciable    significance   in   determining 

From  the  standpoint  of  luminous  efficiency  that  type  of  selec- 
tivity is  of  interest  in  which  the  emission  in  the  visible  spectrum 
is  exaggerated  compared  with  that  radiated  in  other  wave-lengths. 
On  the  other  hand,  substances  exist  which  would  be  much  less 
efBcieot  as  luminous  radiators  than  a  black  body  at  the  same  tem- 
perature. For  example,  ordinary  glass,  if  it  could  withstand  the 
temperature  of  carbon  filaments,  would  be  much  less  efficient  than 
a  carbon  lamp,  aFsuming  that  the  radiating  properties  of  glass 
would  not  tuidergo  serious  change  at  higher  temperatures.  At 
ordinary  temperatures  glass  absorbs  very  little  energy  in  the  visible 
spectrum  compared  with  that  absorbed  in  the  deep  infra-red.  Con- 
versely, glass  emits  relatively  much  less  energy  in  the  visible  spec- 
trum than  a  black  body  at  the  same  temperature.  Such  a  sub- 
stance is  ill-fitted  to  serve  as  a  luminous  radiator. 

At  the  present  time  there  are  not  sufficient  data  on  the  radiating 
properties  of  substances  to  justify  any  extensive  classification.  As 
a  rule  metals  show  relatively  low-reflecting  powers  in  the  visible 
spectrum  and  uniformly  high-reflecting  powers  in  the  infra-red. 
Conversely,  such  metals  would  show  a  relatively  higher  emission 
in  the  visible  as  compared  with  the  infra-red  spectrum,  and  would 
hence  he  more  efficient  luminous  radiators  than  a  black  body  at 
the  same  temperature. 

>y  Google 

Physical  Ghasaotkbistics  of  Lnuiiroirs  Sousces        53 

In  etDdyin^  the  incandescent  mantle  and  the  Nemst  glower  is 
tiie  neit  lecture  the  pecoliar  fonn  of  the  spectra  of  the  oxides 
composiiig  these  radiabHB  will  be  diecueaed.  It  has  been  proposed 
bj  some  inTeatigstors  that  the  radiation  from  these  substances  ie 
not  to  be  aacribed  entirely  to  the  temperature,  but  is  due  in  part 
to  lominescence. 

5.  Luminescence 

If  a  body  during  the  process  of  radiation  undergoes  a  change  *in 
nature,  it  would  not  in  general  continue  to  radiate  in  the  same 
way  even  t^on^  its  temperature  were  maintained  constant  through 
the  addition  of  heai  Such  a  process  of  radiation  has  been  defined 
as  luminescence.  The  cause  of  the  radiation  in  this  case  is  con- 
sidered to  lie  not  in  the  temperature  of  the  system,  but  in  some 
other  source  of  energy.  It  is  a  simple  matter  to  adduce  illustra- 
tiona  of  apparently  typical  luminescence,  and  of  typical  tempera- 
ture radiation,  but  in  many  cases  the  distinction  is  ditScult  if  not 
impoBsible.  Even  in  apparently  typical  cases  of  luminescence  it  is 
a  question  as  to  whether  the  ultimate  cause  of  the  radiation  may 
not  be  temperature — not  the  average  temperature  of  the  system,  but 
the  high  localized  temperature  in  isolated  portions  of  the  system. 

The  deBnition  of  luminescence  that  has  been  given  is  taken 
from  Drude,  and  differs  slightly  from  the  original  definition  of 
Wiedemann,  who  first  introduced  the  term.  In  the  light  of  more 
recent  experiments  and  more  modern  theory  it  is  questionable 
whether  either  definition  is  illuminating  or  helpful  to  a  better 
understanding  of  the  phenomena.  In  the  process  of  light  produc- 
tion by  the  passage  of  an  electric  current  through  the  filament  of 
incandescent  lamps  we  commonly  say  that  the  electric  energy  is 
transformed  into  \eai,  and  that  the  filament  is  heated  to  such  a 
temperature  tl}at  it  becomes  incandescent.  On  the  other  hand,  in 
the  case  of  the  luminous  vapor  in  the  arc  discharge  we  frequently 
say  that  the  vapor  radiates  by  luminescence,  and  that  there  ie  a 
direct  transformation  of  electric  energy  into  radiation  without  the 
intermediate  form  of  heat  energy.  And  yet  in  this  case,  as  in 
the  other,  there  is  some  intermediate  form  of  energy,  viz.,  the 
kinetic  energy  of  the  corpuscles  or  ions  which  are  atcelerated  by 
the  electric  force. 

Whether  or  not  there  is  any  ultimate  difference  between  tempera- 
ture radiation  and  luminescence  in  true  physical  significance  there 

>y  Google 


is  unqueBtionably  a  marked  difference  in  the  apparent  phenomena 
exhibited  in  the  two  eaaes.  In  the  discussion  of  tiie  various  light 
sources  in  the  second  lecture  following  general  custom  the  term 
luminescence  will  be  used  to  describe  that  type  of  radiation  which 
it  is  claimed  by  some  has  never  been  produced  by  heating  the  gys- 
tem  *  as  a  whole,  but  throughout  the  term  will  be  used  with  re- 
serve ae  to  its  exact  significance. 

Drude"  classifies  under  the  general  term  luminescence  the  fol- 
lowing phenomena:  (1)  Ckemi-luminescence,  as  in  the  glow  from 
slowly  oxidizing  phosphorus;  (2)  jAoto-luminesoence,  ordinarily 
known  as  phosphorescence,  which  is  the  after-glow  resulting  from 
prerious  radiant  excitation;  (3)  electro-luminescence,  aa  in  the  glow 
from  Geiasler  tubes.  Under  electro-luminescence  would  also  come 
the  luminescence  from  the  vapor  in  the  arc  discharge. 

Lectore  II 
1.  Introduction 

In  the  preceding  lecture  a  general  discussion  of  the  various  ele- 
ments which  enter  in  a  stndy  of  the  physical  characteristics  of 
luminous  sources  was  undertaken.  In  the  present  lecture  the 
different  types  of  illuminaots  will  be  discussed  in  regard  to  the 
various  elements  presented  in  the  first  lecture,  so  far  as  the  peculiar 
natures  of  the  different  illuminants  and  the  available  data  will 
permit  It  ia  of  interest  to  notice  how  investigation  has  been  pur- 
sued along  different  lines  for  the  different  sources,  making  a  well- 
balanced  analysis  difiBcult,  if  not  at  times  impossible.  It  has  been 
the  aim  in  this  lecture  to  consider  each  illuminant  in  the  same 
general  way,  and  at  the  same  time  to  emphasize  the  peculiar  and 
interesting  physical  properties  of  each. 

Although  in  general  the  various  physical  properties  of  each  il- 
luminant are  presented  in  the  discussion  of  that  particular  illumi- 
nant, exceptions  have  heen  made  in  the  matter  of  spectral  distri- 
bution and  quality  of  light,  since  it  seemed  that  this  information 
would  be  of  more  value  when  collected  in  a  comparable  manner. 
These  questions  therefore  are  discussed  in  separate  sections  at  the 
end  of  the  lecture. 

■  In  a  recent  oote  In  tbe  Comptes  Rendna  (Vol.  130,  No.  26,  p.  1747; 
June  27,  1910)  Bauer  reports  an  experiment  In  which  the  characteristic 
bright  line  spectrum  of  sodlnm  vapor  wu  in  his  opinion  ohtalned  bj 

>y  Google 

FuTsicAL  Chakacteeibtics  of  Ldminods  Soubces        55 

S.  The  Physicg  of  the  BUctrie  Incandescent  Lamp  * 

In  many  ways  the  electric  iacandeecent  lamp  is  the  simplest  lamp 
that  could  be  conetnicted,  speaking  from  a  physical  standpoint 
This  is  particularly  true  of  the  older  form  of  untreated  carbon 
lamp,  since  untreated  carbon  approaches  quite  close  to  the  theoret^ 
ical  black  bodtf  in  its  radiating  properties.  Untreated  carbon  haa 
a  high  emisBivity,  and  exhibits  very  slight  indication  of  any  selectiv- 
ity in  its  radiation. 

For  the  electric  incandescent  lamp  in  general  it  may  be  said  that 
all  the  energy  supplied  to  the  lamp  is  transformed  into  heat,  prac- 
tically all  of  the  energy  into  heat  in  the  filament  itself,  since  the 
resistance  of  the  Icading-in  wires  is  in  all  practical  cases  quite  small. 
Of  the  energy  transformed  into  heat  in  the  filament  nearly  all  is 
radiated.  As  we  shall  see  later  the  losses  by  thermal  conduction 
along  the  leading-in  and  anchoring  wires,  and  the  conduction  and 
con-vection  by  the  enclosed  gas  at  very  low  presBure,  are  very  small 
in  commercial  lamps.  Finally,  the  radiation  is  pure  temperature 
radiation,  there  being  no  luminescence,  and,  in  the  case  of  the  un- 
treated carbon  filament,  the  efficiency  is  due  almost  entirely  to  the 
temperature,  there  being  no  appreciable  selectivity.  The  metal- 
filament  lamps  show  marked  evidence  of  selective  emission,  which 
in  part  determines  their  efficiency,  but  with  this  exception  are  quite 
similar  to  the  untreated  carbon-filament  lamp  in  all  respects. 

In  the  introduction  to  the  first  lecture  the  incandescent  lamp  was 
cited  to  illustrate  the  principles  involved  in  the  physics  of  a  lamp. 
In  this  illustration  an  experiment  was  described  showing  the  tre- 
mendous losses  that  would  result  from  conduction  and  convection 
by  gas  in  the  bulb.  Under  the  conditions  of  the  experiment  which 
was  performed  upon  a  platinum  filament  (0.1  mm.  diameter,  15  cm. 
long)  in  a  pear-shaped  bulb  (13  cm.  long  and  8  cm.  maximum 
diameter),  the  power  required  to  bring  the  filament  to  a  tempera- 
ture of  about  1700°  Abs.  when  the  bulb  was  filled  with  air  at  at- 
moapheric  preaaure  was  found  to  be  about  five  times  that  required 
to  bring  the  fi.Iament  to  the  same  temperature  in  a  vacuum. 

The  curves  in  Figure  6  show  the  relation  between  the  required 
power  and  the  pressure  inside  the  bulb  for  the  same  platinum  lamp 
and  the  same  temperature.  It  is  seen  that  the  required  power 
changes  most  rapidly  at  moderately  low  pressures,  and  that  the 
loss  due  to  the  gas  is  still  quite  appreciable  at  pressures  as  low  as 
0.05  mm.  mercury.     It  should  be  emphasized  that  the  numerical 

>;  Google 


results  found  depend  largely  on  the  conditions  of  the  experiment. 
A  difference  in  size  of  bulb  or  composition  of  filftment,  or  tempera- 
ture of  filament  or  composition  of  gas,  would  probably  affect  the 
numerical  results  to  a  marked  e^itent.  The  curves  of  Figure  5  are 
given  to  show  the  general  nature  of  the  phenomenon. 

As  stated  in  a  previous  paragraph,  this  losa  by  conduction  and 
convection  of  the  enclosed  gas  in  a  commercial  lamp  is  negligibly 
small.    The  vacuum  that  is  secured  in  a  good  lamp  is  probably  of 

1  1  n  1  1  1  1  1  M  1  1  1  M 

1  1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 

"                                                niHin      ■>    n<i>    K   * 

"^  "*(->.;?     •"     •" 

i     PrlT 

^,          A                                                                  l< 

*^, : : .  :  : :        : ^ :  :  : 

i.i.l  1  l,i  1  1  1  1  1  ^   \  \  \     1  1  IJm 

Fia.  E. — Watte  Required  at  Varying  Air  PresBures  to  Maintain  a 
Platinum  Pllamflnt  at  a  Constant  Temperature  (that  of  a  Color  Match 
with  a  Black  Body  at  1690°  Abe.). 

the  order  of  magnitude  of  0.001  mm.  mercurj',  which  is  smaller 
than  any  vacuum  measured  in  the  described  experiment. 

The  loeees,  on  the  other  hand,  by  conduction  away  of  heat  at  the 
leading-in  and  anchoring  wireB,  though  relatively  small,  are  not 
negligible.  Measurements  of  the  total  energy  radiated,  compared 
with  that  supplied,  have  led  several  investigators  to  the  conclusion 
that  aa  much  as  20  per  cent  or  30  per  cent  of  the  supplied  energy 
ia  lost  in  this  way.    Recent  measurements  on  the  temperature  gradi- 

>y  Google 

Physical  Ceabactehistios  op  Luminous  SonBCES        57 

ent  of  filamentB  near  the  leadmg-in  and  oDchoring  wires  have  led 
to  the  conclusion  that  the  loss  by  conduction  in  an  ordinary  1.25 
V.  p.  c,  llO-volt,  25-watt  Mazda  lamp  it  not  more  than  5  per  cent. 
For  a  2w.  p.  c,  110-voIt,  40-watt  tantalum  lamp  the  loss  is  less 
than  7  per  cent.  The  larger  iosB  for  the  tantalum  lamp  is  due 
probably  to  the  relatively  larger  number  of  anchor  wires. 

The  diCFerences  in  efficiency  among  the  various  incandescent  lamps 
are  to  be  ascribed  therefore  to  the  differences  in  the  quality  of  the 
radiation."  Since  there  is  no  luminescence  the  qualiti^  of  the  radia- 
tion results  from  (1)  the  temperature  at  which  the  filament  oper- 
ates, and  (2)  the  selectivity  of  the  radiation  eorreaponding  to  the 
filament  materia!.  "Unfortunately,  it  is  impossible  to  separate  these 
two  elements  entirely.  The  meaaurement  of  temperature  involves 
in  general  certain  assumptions  regarding  the  nature  of  the  radia- 
tion, and  the  measurement  of  selectivity  depends  upon  temperature 
rdations.  Methods  have  been  devised,  however,  which  give  quali- 
tative indication  of  the  relative  selectivity  of  the  various  filaments, 
and  which  on  the  basis  of  probable  assumptions  indicate  the  lower 
quantitative  limits  to  the  selectivity. 

It  is  beyond  the  scope  of  these  lectures  to  enter  upon  a  discusBion 
of  the  methods  employed  or  of  the  detailed  results  obtained.  The 
principal  conclusions  regarding  the  selectivity  of  lamp  filaments 
which  have  been  reached  up  to  the  present  time  may  be  stated  as 
follows:  If  the  various  metal-filament  lamps  were  operated  at  the 
same  temperature  as  an  ordinary  untreated  carbon  filament  at  4 
watts  per  candle,  then  it  is  probable  that  owing  to  selectwity  the 
tantalum  lamp  would  have  a  higher  efficiency  than  that  of  the 
carbon  by  more  than  10  per  cent  or  12  per  cent,  the  tangsten  lamp 
would  have  a  higher  efficiency  by  more  than  35  per  cent  or.  30  per 
cent,  and  the  osmium  lamp  would  have  a  higher  efficiency  by  more 
than  40  per  cent.  Of  any  of  the  metels,  platinum,  tantelum,  tung- 
sten or  osmium,  the  last  seems  to  differ  most  widely  in  the  quality  of 
its  radiation  from  the  blacle  body.  Both  tentalum  and  tungsten 
give  evidence  of  greater  selectivity  than  platinum.  Platinum  devi- 
ates far  from  the  black  body  in  its  emissivity,  but,  as  explained 
above,  this  haa  only  an  indirect  influence  on  its  fitness  for  use  as 
a  lamp  filament. 

The  efficiency  of  all  the  metal  lamps  is  therefore  probably  due  in 
part  to  selectivity.  Only  in  the  case  of  the  osmium  filament,  how- 
ever, is  the  selectivity  so  great  that  ite  high  efficiency  {1.5  w.  p.  c), 

>y  Google 

58  Illouinating  Enginebbino 

as  compared  with  a  3.1  w.  p.  e.  carbon  lamp,  is  almost  entirely  ex- 
plained on  the  basis  of  the  eelectivity  of  its  radiation.  If  the 
conclusions  cited  above  are  correct  the  temperature  •  of  the  osmium 
lamp  is  probably  not  very  different  from  that  of  a  flashed-carbon 
lamp  at  3.1  watte  per  caudle. 

Inasmuch  as  practically  all  of  the  energy  supplied  to  iDcan- 

accurately  about  0,95)  for  the  ordinary  commercial  lamps.  Hence, 
-^  may  be  taken  as  -Sr-,  and  is  the  ordinary  luminous  efficiency 
for  the  vaiionB  types  of  lamps  at  normal  burning  expressed  in 

as  ranging  between  0.03-0.03  for  a  4-watt  carbon  lamp,  to  0.08-0.10 

for  a  1.25-watt  tungsten  lamp.    Of  course,  the  ratios  =-  and  -^ 

can  be  modified  at  will  by  operating  the  lamps  at  different  tem- 
peratures, i,  e.,  at  different  voltages,  but  these  cases  do  not  have 
any  interest  to  ns  at  present  except  in  so  far  as  the  relations  be- 
tween voltage,  resistance  and  candle-power  are  of  great  importance 
in  practical  operation.  On  account  of  their  importance  in  opera- 
tion these  characteristics  will  probably  be  considered  in  another 
lecture.  Suffice  it  to  say  here  that  the  temperature  coefficients  of 
resistance  of  the  various  filaments  differ  widely  among  themselves 
both  in  sign  and  amount,  ordinary  carbon,  for  example,  having  & 
negative  coefficient  at  normal  teroperatnres  of  operation,  whereas 
tungsten  and  the  other  metals  have  relatively  large  positive  co- 
efBcients.  Moreover,  owing  to  the  peculiar  radiating  properties  of 
the  various  filaments,  there  are  marked  differences  in  the  changes 
in  candle-power  correaponding  to  a  given  change  in  total  energy 

S.  The  Physics  of  the  Are  Iximp 
The  varioua  types  of  arc  lamps  have  been  subject  to  much  in- 
vestigation, both  as  to  their  physical  and  operating  characteristics. 
Many  of  the  physical  characteristics  of  the  arc  are  so  intimately 

•Owing  to  the  great  dlBcrepancies  In  the  publlabed  values  for  the 
true  temperaturea  ot  the  different  Incandescent  lamps,  no  attempt  U 
made  to  give  In  this  lecture  a  table  of  moat  probable  values.  Refer- 
ences to  original  pnblloatlons  on  the  subject  are  given  in  the  bib- 

>y  Google 

Physioal  Chabaotebistiob  of  LuuiNons  Sodbobb       59 

connected  with  its  operation  that  they  will  nndonbtedly  receiye 
treatment  in  another  flection  of  this  Beries  of  lectures.  I  shall  there- 
fore confine  myself  to  a  brief  etatement  of  some  of  the  more  purely 
phyaical  characteristics  of  the  arc. 

The  arc  may  be  defined  for  out  pnrposee  aa  a  portion  of  the 
drciiit  consisting  of  a  pair  of  electrodes  of  solid  or  liquid  material, 
electrically  connected  by  a  body  of  vapor  which  results  from  the 
TolatilizatioD  of  material  from  one  or  both  of  the  electrodes.  The 
term  "  arc  "  is  Bometimes  nsed  in  a  more  restricted  senee  to  apply 
to  the  bridge  of  vapor  alone,  and  also  in  a  still  different  sense  to 
the  process  of  the  discharge  or  flow  of  current  between  the  elec- 
trodes, rather  than  to  the  part  of  the  circuit  where  it  occurs. 

If  we  connect  a  voltameter  to  the  terminals  of  an  arc  aa  used  in 
lighting,  we  find  that  the  volt-meter  indicates  a  difference  of  poten- 
tial of  40  volts  or  more,  depending  upon  the  type  of  lamp  used. 
The  difference  of  potential  is  determined  by  the  current,"  tiie  dis- 
tance between  the  electrodes,  the  materials  of  the  electrodes  and  the 
pressure  of  the  surrounding  atmosphere.  The  fall  of  potential 
across  the  arc  is  not  due  to  mere  ohmic  resistance,  as  in  a  wire 
carrying  a  current.  Under  certain  special  conditions  an  arc  may 
be  obtained  in  which  the  voltage  increases  with  the  current,  but 
such  is  not  a  true  arc  as  used  in  lighting.  This  kind  of  discharge 
is  sometimes  called  a  "  glimm-strom."  A  distinct  characteristic  of 
a  true  arc  is  that  as  the  current  increases  the  voltage  decreases. 
For  this  reason  it  is  essential  that  are  lamps  operated  on  constant 
potential  supply  circuits  should  be  provided  with  series-ballast  re- 
sistances to  prevent  the  arc  from  short-circuiting." 

In  the  case  of  an  ordinary  incandescent  lamp  the  fall  in  potential 
is  practically  distributed  uniformly  along  the  entire  length  of  the 
filament,  so  that  if  we  should  measure  the  voltage  drop  along  each 
centimeter  of  the  filament  we  would  find  it  to  be  the  same  through- 
out, and  equal  to  that  fraction  of  the  total  applied  voltage  which 
one  centimeter  bears  to  the  total  length  of  the  filament.  Such  is 
not  the  case  with  the  arc  The  fall  in  potential  takes  place  in  three 
distinct  stepS:  (1)  at  the  anode,  or  where  the  current  passes  from 
the  positive  electrode  to  the  gas;  (2)  along  the  vaporous  path;  and 
(3)  at  the  cathode,  or  where  the  current  passes  from  the  gas  or 
vapor  to  the  negative  electrode.  The  proportion  of  the  total  applied 
voltage  which  is  taken  up  at  each  of  these  three  places  depends 
on  the  nature  of  the  arc.    In  the  ordinary  short  carbon  arc  "  the 

>y  Google 

60  Illuminating  Enginbhuno 

greater  part  of  the  fall  of  potential  is  at  the  anode,  whereas  in 
flaming  arcs  moat  of  the  electrical  energy  is  transformed  in  the 
conducting  vapor. 

Correspondiiig  to  the  three  regions  where  the  fall  of  potential 
takes  place,  with  the  corresponding  transfonnation  of  electric  en- 
ergy, are  the  three  distinct  regions  of  luminous  radiation,  vis.,  the 
anode,  the  cathode  and  the  vapor.  And  just  as  the  distribution  of 
the  potential  drop  depends  on  the  natnre  of  the  arc,  bo  the  distri- 
bution of  the  radiant  energy  varies  greatly.  In  the  case  of  the  short, 
open,  direct-current,  carbon  arc  the  gas  or  vapor  contribntes  but  a 
very  small  part'  (several  per  cent)  of  the  total  luminous  flus. 
The  anode  and  cathode  both  are  raised  to  a  high  temperature  and 
consequently  radiate  energy,  but  the  temperature  of  the  anode  is 
much  higher  than  that  of  the  cathode,  and  is  to  be  considered 
practically  as  the  light  source  in  the  open  carbon  arc.  In  the  en- 
closed arc  the  luminous  vapor  plays  a  larger  part  in  determining 
the  luminous  efficiency,  but  the  principal  source  of  light  in  the 
direct-current  arc  is  again  the  anode.  In  alternating-current  arcs 
the  two  electrodes  play  equal  parts  in  producing  the  luminous  flux, 
but  as  the  two  terminals  are  alternately  positive  and  negative,  and 
as  relatively  little  heat  is  produced  at  the  negative  terminal,  the 
average  temperature  of  the  carbon  electrodes  of  an  alternating-cur- 
rent arc  is  lower  than  the  temperature  of  the  anode  in  a  direct- 
current  are,  and  hence  the  luminous  efficiency  of  the  former  is  less 
than  that  of  the  latter. 

In  direct-current  and  alternating-current  open  and  enclosed  car- 
bon arcs  the  electrodes  supply  the  larger  part  of  the  luminous  flux, 
which  therefore  may  be  said  to  be  due  to  pure  temperature  radia- 
tion. In  the  case  of  the  luminous  and  flaming  arcs  the  principal 
source  of  the  luminous  flux  is  the  luminous  vapor,  the  electrodes 
adding  little  to  the  luminous  efficiency.  The  difference  between 
these  two  types  of  arcs,  Ivminotis  and  flaming^  is  significant  in  its 
bearing  on  the  physics  of  light  production  in  these  arcs.  The  es- 
sential difference  between  the  two  arcs  consists  in  the  two  distinct 
processes  by  which  the  light-giving  vapors  find  their  way  into  the 
arc,  where  they  perform  simultaneously  the  two  functions  of  con- 
ducting the  current  and  radiating  the  light — two  functions  that  are 
doubtless  intimately  connected.  (1)  A  carbon  arc  may  be  used  as 
a  basis,  the  anode  being  impregnated  witii  salts  or  pierced  with  a 
longitudinal  hole  through  which  a  metal  wire  is  threaded.    This 

>y  Google 

Physical  Cuab^lotebistios  of  Ldhinous  Socboes       61 

gives  the  flaming  arc.  The  anode  ia  used  because  in  the  case  of  the 
carbon  arc  it  ie  the  hotter.  The  vapor  is  the  restilt  of  the  evapora- 
tion of  the  salts  or  metal,  and  takes  part  in  the  conduction  of  the 
current  through  the  are.  The  anode  bums  away  rapidly  as  a  result 
of  the  high  temperature  and  consequent  evaporation.  (3)  The 
vapor  comes  from  the  cathode.  Such  an  arc  is  called  a  luminous  arc. 
In  it  the  anode  may  be  entirely  free  from  burning  or  melting  away, 
being  quite  cool.  There  is  necessarily  a  consumption  of  the  cathode, 
but  it  may  be  rather  slow. 

The  importance  of  the  diatinction  between  these  two  types  of 
arcs  enters  in  the  consideration  of  the  possible  explanation  of  the 
high,  luminous  efficiency  of  the  radiating  vapor.  Ib  this  efficiency 
to  be  ascribed  to  pronounced  selective  temperature  radiation  at  a 
high  temperature  or  to  luminescence?  This  is  a  much-mooted 
question,  though  the  consensus  of  opinion  at  the  present  time  eeenu 
to  be  that  the  efficiency  is  to  be  ascribed  to  luminescence  rather  than 
to  pure  temperature  radiation. 

The  probability  of  this  theory  may  be  seen  from  the  phenomena 
of  the  luminous  arc.  In  the  case  of  the  flaming  arc  the  metallic 
vapors  get  into  the  arc  through  evaporation  at  the  anode,  indicating 
at  least  as  high  and  possibly  a  higher  temperature  than  that  of  the 
anode.  According  to  VioUe"  the  temperature  of  the  carbon  flame 
is  higher  than  that  of  the  anode.  With  such  temperatures  it  ia  not 
unthinkable  that  with  vapors  showing  selective  emission  in  favor 
of  the  visible  region  of  the  spectrum  the  luminosity  might  be  due 
to  selective  temperature  radiation  rather  than  to  luminescence.  In 
the  case  of  the  luminous  arc,  however,  the  conditions  are  quite  dif- 
ferent The  anode  may  be  entirely  cold,  the  vapor  being  carried 
into  the  arc  stream  by  the  process  of  conduction.  It  is  probable 
that  the  cathode  is  always  fairly  hot,  but  the  evaporation  of  the 
cathode  is  unquestionably  below  that  of  the  anode  in  the  flaming 
arc.  In  the  case  of  the  luminous  arc,  therefore,  there  are  perhaps 
greater  difficulties  in  explaining  the  luminosity  by  temperature  ra- 
diation, and,  on  the  other  hand,  more  cogent  reasons  for  accepting 
the  theory  of  luminescence. 

Unfortunately,  no  data  are  available  on  the  actual  temperatures 
of  the  luminoua  gas  in  luminous  and  flaming  arcs.  On  the  other 
hand,  numerous  measurements  of  the  temperature  of  the  anode 
crater  of  a  direct-current  carbon  arc  have  been  made,  and  some  little 
data  on  the  temperature  of  the  cathode  and  vapor  have  been  pub- 

>y  Google 

63  Illuhinatino  Encineerikq 

liBhed."  The  most  probable  temperatnre  deduced  from  these  obser- 
vationa  is  about  3800''-4000''  Abs.  (Centigrade+STS").  The  tem- 
perature of  the  cathode  has  been  found  by  Rosetti  to  be  3300° 
Abe.,  bat  as  the  same  observer  determined  the  temperature  of  the 
positive  carbon  to  be  about  4150°  Abs.  it  is  probable  that  his  value 
for  the  cathode  is  800°  or  300°  too  high.  For  the  vapor  itself  the 
value  5000°  Aba.  has  been  given.  In  the  alternating-current  arc 
the  two  electrodes  have  the  same  temperature,  intermediate  between 
the  temperaturea  of  the  anode  and  cathode  of  the  direct-current  arc. 

The  arc  lamp,  like  many  other  practical  sources,  is  subject  to 
loeaeB  of  energy  by  conduction  and  convection.  In  addition,  arc 
lamps  operated  on  constant  potential  circuits  are  subject  to  still 
further  losses  owing  to  the  necessity  of  series  resistance  or  resistance 
as  ballast  On  constant-current  circuits  this  ballaat  ia  unnecessary. 
The  amount  of  loss  in  cases  where  lamps  are  operated  on  constant 
potential  varies  greatly  in  practice.  Aa  a  rule  the  aeries  resistance 
ia  not  the  least  which  would  give  stability,  but  is  also  made  use 
of  to  adapt  the  arc  to  the  existing  supply  voltage.  Thus,  in  the 
case  of  a  carbon  are  operating  on  a  HO-volt  cireuit,  more  than  half 
the  total  energy  is  wasted  in  resistance.  On  account  of  the  nn- 
eertainty  of  this  loss  and  of  the  entire  absence  of  it  in  constant- 
current  circuits  no  account  is  taken  of  it  in  the  data  on  efficiency 
given  below. 

With  regard  to  the  conduction  and  convection  losses  no  data  ex- 
ist, so  far  aa  I  know,  which  permit  an  accurate  estimate  of  the 
magnitude  of  these  losses.  It  is  a  well-known  fact  that  the  efficiency 
is  increased  by  diminishing  the  thicknesa  of  the  carbons,  but  the 
percentage  loss  through  thennal  conduction  by  the  electrodes  has 
probably  never  been  determined.  The  fact  that  an  are  is  in  contact 
with  the  air  at  atmospheric  pressure  would  tend  to  make  the  loss 
by  air  conduction  and  convection  great,  but  doubtless  such  an  effect 
is  much  reduced  by  the  fact  that  the  radiation  ia  proportional  to  a 
higher  power  of  the  temperature  than  the  air  loss  probably  m,  bo 
that  in  percentage  it  probably  is  much  less  than  the  loss  with  the 
Nernst  lamp,  for  example. 

Owing  to  the  complicated  nature  of  the  production  of  light  from 
the  various  arc  lamps  the  available  data  do  not  permit  the  exact 
analysis  of  the  energy  relations  such  as  is  posaible,  for  example,  in 
the  incandescent  lamp.  Thus  we  have  seen  that  the  temperature 
of  the  crater  of  a  direct-current  open  carbon  are,  which  is  chiefly 

>y  Google 

Physical  Ghabacteeistics  of  Luuinous  SoosoEa        63 

respoDsible  for  the  efficieccy  of  this  type  of  arc  lamp,  is  approxi- 
mately 3800''-4000''  AbB.  A  black  body  at  this  temperatuie,  with 
no  losses  by  conduction,  convection,  etc.,  would  be  aboat  8  or  10 
per  cent  as  efficient  aa  the  moBt  efficient  monochromatic  source. 
Compared  with  this  value  the  open  carbon  arc  has  an  efficiency" 
of  only  about  2  per  cent.  In  otiier  words,  if  there  were  no  losses 
except  in  the  infra-red  radiation  from  a  ilach  body  at  3800°  Abs., 

the  ratio  f-y-  =  -^]  of  the  luminons  flux  to  the  power  input 
would  be  approximately  70  or  80  lumens  per  watt,  whereas  for  the 
direct-current  carbon  open  arc  the  ratio  -^  equals  approximately  12 
or  15  lumens  per  watt.  It  must  be  borne  in  mind,  however,  that  al- 
though the  crater  of  the  anode  is  at  3800''-4000''  Aba.,  other  parts  of 
the  anode  near  the  tip,  as  well  as  the  tip  of  the  cathode,  are  radi- 
ating at  much  lower  and  hence  much  lees  efficient  temperatures, 
BO  that  a  relatively  large  proportion  of  the  supplied  energy  is 
radiated  in  the  infra-red,  compared  with  the  radiation  from  carbon 
at  3800''-4000''  Abs. 

Similarly,  whereas  for  carbon  at  3800''-4000''  Abs.  the  ratio  -g- 
would  be  approximately  0,30,  the  actual  values  foimd  for  various 
types  of  carbon  arcs  by  Marks  and  by  Nakano  range  from  0.08 
to  0.17. 

The  efficiency  of  the  alteniatingH3urrent  arc  is  roughly  one-half 
to  three-quarterB  that  of  the  direct-current  arc,  whereas  with 
flaming  and  luminous  arcs  efficiencies  from  three  to  Ave  times  that 
of  the  direct-current  carbon  arc  have  been  obtained.  Thus  lunainous 
efficiencies  of  40-60  lumens  per  applied  watt  are  found.  Still  higher 
efficiencies  are  possible  in  the  use  of  electrodes  exhibiting  more  pro- 
nounced selectivity  in  the  visible  spectrum. 

ITcder  the  head  of  arc  lamps  should  properly  come  the  mercury 
arc,  or  mercury-vapor  lamp,  as  it  is  more  commonly  called,  but  since 
in  practice  it  is  considered  as  a  distinct  type  of  lamp,  and  moreover 
resembles  in  some  ways  the  vacuum-tube  lamps,  it  will  be  consid- 
ered in  a  separate  section  devoted  to  these  two  types  of  illuminants. 

Jf.  The  Physics  of  Low-pressure  Arcs  and  Vacuum  Tvbes 

The  mercury-vapor  lamp  "  is  an  arc  at  low  pressure,  and  is  to  be 

distinguished  physically  from  the  nitrogen,  carbon  dioxide  and  other 

vacuum-tube  sources  with  which,  from  its  appearance,  it  might  be 

>y  Google 


fioufnsed.  The  chief  poiot  of  reEemblance  between  the  two  types 
of  lamps  is  that  in  each  caee  the  light  is  emitted  by  a  luminescent 
gag  or  vapor  at  presanres  considerably  below  that  of  the  atmosphere. 
The  dietingnishing  characteristic  is  the  process  by  which  the  dis- 
charge through  the  tube  takes  place,  with  its  effect  on  the  nature  of 
the  radiation  emitted.  In  the  low-pressure  mercnry  arc  the  con- 
dncting  material  is  mercnry  vapor  supplied  by  the  hot,  mercury 
cathode,  and  the  character  of  the  light  is  given  by  the  emission 
spectrum  of  mercury  vapor,  just  as  iu  the  luminous  arc  the  cathode 
material  enters  the  arc  and  determines  the  character  of  the  radia- 
tion. The  difference  between  the  mercury  arc  and  the  ordinary 
Inminons  arc  is  mainly  one  of  pressure  of  the  surrounding  gas. 

In  an  ordinary  vacuum-tube  discharge,  on  the  other  hand,  the 
conducting  material  is  the  gas  between  the  electrodes,  air,  nitrogen, 
carbon  dioxide,  etc.,  and  the  character  of  the  radiation  depends 
on  the  emission  spectra  of  these  gases.  The  material  of  which  the 
electrodes  are  composed  plays  no  large  part  in  determining  the  char- 
acter of  the  light  emission. 

Considering  first  the  low-pressure  arc,  as  in  the  mercury-vapor 
lamp,  the  phenomena  exhibited  are  in  general  the  same  as  those 
presented  in  the  discussion  of  ordinary  arcs.  There  is  the  same  fall 
of  potential  *  at  the  anode  and  at  the  cathode,  but  owing  to  the  low 
pressure  in  the  tube  the  conductivity  of  the  mercury  vapor  is  much 
greater,  permitting,  or  even  necessitating,  a  much  longer  arc  for 
high  efficiency. 

The  temperature "  of  the  mercury  arc  in  a  glass  tube  is  ap- 
parently quite  low,  and  the  explanation  of  the  efficiency  is  ordinarily 
ascribed  to  luminescence,  with  a  relatively  large  part  of  the  energy 
in  the  visible  spectrum.  The  efficiency"  of  the  arc,  as  ordinarily 
operated,  is  variously  given  as  ranging  between  13  and  24  lumens 
per  applied  watt,  corresponding  to  0.5-1.0  watt  per  mean  spherical 
candle.  By  using  quartz  instead  of  glass  it  is  possible  to  operate 
the  lamp  with  a  much  higher  current  density  and  greatly  increased 
efficiency."  It  is  probable  that  an  efficiency  of  about  50  or  60 
lumens  per  watt,  corresponding  to  0.20  or  0.25  watt  per  mean 
spherical  candle,  may  be  reached  in  the  case  of  the  quartz  are.  It 
is  believed,  in  this  case  that  at  high  temperatures  pure  temperature 
radiation  of  increasing  efficiency  supplements  the  decreasing  effi- 
ciency of  the  luminescent  radiation. 

>y  Google 

Phtsioal  Chabactebistics  op  Luminous  SouscEa        66 

Data  as  to  the  losses  by  conduction,  convection,  etc.,  and  as  to  the 
ratio  [-5']  °^  tlifi  energy  in  the  visible  spectrum  to  the  total  energy 
emitted  are  somewhat  meager  and  indefinite.  Lux"  finds  for  the 
Uviol  (special  kind  of  glass)  and  quartz  lamps  the  values  0.068 

and  0.176,  respectively,  for  the  ratio  ^ .  Moreover,  for  a  Uviol 
lamp  operating  at  about  0.66  watt  per  mean  spherical  candle,  be 
finds  that  about  one-half  the  total  energy  supplied  to  the  lamp  is 
radiated,  the  other  half  being  dissipated  at  the  electrodes,  and  by 
conduction  and  convection. 

In  the  mercury  arc,  as  in  other  arcs,  the  material  of  one  or  both 
of  the  electrodes  determines  the  character  of  the  light  emission, 
whereas  in  an  ordinary  vacuum-tube  discharge  the  nature  of  the 
gas  between  the  electrodes  determines  the  spectrum,  modified,  how- 
ever, by  such  conditions  "  as  pressure,  potential  gradient,  etc.  In 
the  arc,  after  the  gap  between  the  electrodes  is  bridged,  i.  e.,  after 
the  are  is  "  struck,"  the  supply  of  "  ions  "  or  carriers  of  electricity 
is  furnished  by  the  negative  electrode,  and  the  conduction  of  cur- 
rent is  continuous.  The  fall  of  potential  at  the  electrodes  of  a 
vacuum  tube  is  always  very  high,  of  the  order  of  magnitude  of 
several  hundred  or  a  thousand  volts,  so  that  the  applied  voltage 
must  always  be  great.  The  fall  in  potential  per  centimeter  length 
of  tube  is  small  compared  with  the  fall  of  potential  at  the  electrodes, 
and  consequently  very  long  tubes  must  be  used  in  order  that  a 
moderate  amount  of  the  supplied  energy  may  be  radiated  by  the 
gas  rather  than  practically  all  lost  at  the  electrodes.  The  con- 
ductivity of  .the  gas  varies  with  the  pressure,"  reaching  a  maximum 
at  pressures  of  the  order  of  magnitude  of  tenths  of  a  millimeter 
of  mercury. 

As  in  the  mercury  arc,  the  radiation  is  considered  to  be  electro- 
luminescence. The  efficiency  depends  on  the  distribution  of  energy 
in  the  emission  spectrum,  which  varies  from  gas  to  gas.  Angstrom  " 
found  for  nitrogen  a  maximum  of  91  per  cent  of  the  radiated  en- 
ergy lying  in  the  visible  spectrum,  69  per  cent  for  carbon  dioside, 
and  60  per  cent  for  hydrogen.  Commercial  installations,  as  in  the 
Moore  tubes,  have  an  efficiency  *  of  6  or  6  lumens  per  watt  for  the 
nitrogen  tube,  and  about  one-third  or  one-quarter  that  value  for 
the  carbon-dioxide  tube.  The  great  discrepancy  between  the  lumi- 
nous ^ciency  of  the  radiation,  and  the  actual  luminous  efficiency 

>y  Google 


of  the  lamp  is  to  be  ascribed  to  loeees,  of  which  those  at  the  elec- 
trodes are  by  far  the  largest.  Owing  to  the  comparatively  low  tem- 
perature of  the  tube  the  conduetioii  and  convectioB  loeses  are 
Telatirely  small. 

5.  The  Physics  of  Open  Flames  and  of  the  Incandescent  Manth 
The  LDcandeecent  mantle  lamp  presents  some  of  the  most  difB- 
cnlt  problems  in  the  physics  of  luminous  sources.  In  addition  to 
the  problems  connected  with  the  mantle  itself  are  those  of  the 
Bunsen  flame,  and  these  latter  are  so  intimately  interwoven  with 
the  inter-molecular,  or  so-^»illed  chemical  processes  in  the  flame, 
that  it  is  impossible  to  undertake  a  complete  discnssion  of  the  flame 
in  a  lecture  of  this  nature.  And  yet  there  are  certain  physical 
properties  of  flame  BOurces  which  must  be  mentioned  briefly  as 
auxiliary  to  a  consideration  of  the  incandescent  mantle. 

The  open  luminous  flame  was  the  earliest  form  in  which  gas 
was  used  as  an  illuminant,  but  the  physics  as  well  as  the  chemistry 
of  the  open  flame  has  been  the  subject  of  much  dispute,  even  in 
recent  years.  Various  theories  of  the  chemical  transformations 
within  the  flame  have  been  proposed  with  accompanying  explana- 
tions of  the  light-giving  properties  of  the  flame."  The  ultimate 
source  of  energy  is  chemical,  but  it  has  been  a  mooted  question 
whether  the  radiation  from  the  fiarae  is  dependent  solely  on  the  • 
temperature  or  is  due,  at  least  in  part,  to  chemi-luminescence.  Ac- 
cording to  the  theory  generally  accepted  at  the  present  time  the 
light  from  the  open  luminous  flame  is  due  to  the  temperature 
radiation  from  finely  divided  carbon  particles  heated  to  incan- 
descence by  conduction  from  the  hot  gases  of  the  flame.  The 
spectral  distribution "  of  the  radiation  is  that  which  would  be 
emitted  by  carbon  at  a  temperature  well  within  the  accepted  limits 
of  temperature  of  the  luminous  zone,  viz.,  1500"  Aba.,  at  the 
beginning  of  the  luminous  zone  and  S100°  Abs.  at  the  outer  zone 
of  complete  combustion.  The  luminosity  of  the  flame  will  depend 
on  the  number  of  carbon  particles  present,  and  the  temperature 
which  they  attain.  The  causes  which  tend  to  increase  or  decrease 
the  luminosity  of  flames  may  therefore  be  divided  into  two  classes, 
(1)  those  that  aSect  the  formation  and  quantity  of  the  carbon, 
and  (2)  those  that  determine  the  temperature. 

The  efficiency  "  of  the  open  flame,  considered  from  a  physical 
standpoint  is  very  low,  but  it  is  necessary  to  keep  in  mind  the 

>y  Google 

Phtsical  Chaeacteristics  op  Ldminods  Sodbces        67 

eseential  diiference  between  the  conditionB  prevailing  in  the  pro- 
duction of  light  in  the  open  flame  and  in  the  electric  incandescent 
lamp,  for  example.  Id  the  former  the  chemical  tranafoTmations 
with  the  generation  of  heat  take  place  in  the  Same  itself,  and  it  is 
difficult  to  separate  the  efficiency  of  the  heat  production  from  that 
of  the  incandescent  carbon  particles  rendered  Inminons  by  the  heat. 
In  the  case  of  the  electric  incandescent  lamp  the  chemical  trans- 
formations, with  the  reeultant  generation  of  heat,  take  place  under 
the  boiler  where  the  adduced  gas  bums  (supposing  a  gas  engine), 
and  there  are  large  heat  losaes  even  in  the  most  efficient  systems. 
Moreover,  there  is  a  second  loss  when  the  heat  energy  is  trans- 
formed into  electrical  energy  which  must  also  be  considered.  It 
is  unquestionably  true,  however,  that  the  efficiency  of  the  open 
lominotis  flame,  even  in  its  most  efficient  form  in.  the  regenerative 
burner,  is  still  very  low.  Owing  to  the  large  conduction  and  con- 
vection loeeea  the  heat  available  for  rendering  incandescent  the 
carbon  particles  is  not  large,  and  the  radiant  efficiency  of  these, 
because  of  the  relatively  low  temperature,"  is  comparatively  small. 

The  open  luminous  flame  has  been  very  generally  supplanted  by 
the  incandescent  mantle,  heated  in  a  Bunsen  flame.  In  the  latter, 
which  is  particularly  non-luminous,  a  mixture  of  gaa  and  air  is 
burned  with  the  result  that  a  more  complete  combustion  takes  place 
in  the  body  of  the  flame.  The  temperature  *  of  that  portion  of  the 
flame  between  the  slightly  luminous  bluish-green  surface  of  the 
inner  zone  and  the  outer  limits  of  the  outer  zone  ranges  from  about 
1800°  Abs.  at  the  inner  zone  to  about  2000°-3150''  Abs.  at  the 
outer  zone.  Although  the  maximum  temperature  of  the  Bunsen 
flame  ia  perhaps  hut  slightly  if  any  higher  than  that  of  the  open 
luminous  flame,  the  average  temperature  of  a  large  portion  of  the 
former  is  much  greater  than  that  of  the  latter,  and  the  temperature 
to  which  finely  divided  solids  placed  in  the  Bunsen  flame  may  be 
raised  is  much  higher  than  any  temperature  available  with  the 
Inminons  flame. 

Coming  now  to  the  incandescent  mantle  in  its  most  common 
form,  consisting  principally  of  the  oxides  of  thorium  and  cerium, 
various  hypotheses  have  been  proposed  to  account  for  its  high 
luminous  efficiency.  Following  numerous  attempts  at  the  use  of 
metallic  mantles,  such  as  the  platinum  gauze  of  Gillard,"  and  in 
one  or  two  cases  of  mantles  of  infusible  osides,  as  the  basket  mantle 
exhibited  by  Claymond  in  1880,"  and  the  Fahnehjelm  "  comb  pat- 

>y  Google 


ented  in  1885,  Auer  von  WelsbaCh  brought  ont  hi8  first  mantle  in 
1886."  In  his  original  patent  application  Aner  mentioned  rariouB 
rare  earths  aa  particularly  useful  in  securing  light  of  Tarious  huea. 
Subsequently,"  the  mantle  of  approximately  99  per  cent  thoria  and 
1  per  cent  ceria  as  used  to-day  was  developed. 

Many  years  before  the  introduction  of  the  Auer  mantle  the  re- 
markable properties  of  certain  of  the  rare  earths  when  heated  to 
incandescence  were  known.  Bunsen  in  1864 "  discovered  that 
didymium  earth  when  heated  gives  not  only  a  continuous  Bpectrum, 
but  also  superimposed  bright  bands.  Babr  in  1865  *  found  a  similar 
phenomenon  in  the  case  of  erbium  earth.  Eahr  and  Bunsen  "  con- 
jointly in  1S66  made  a  further  careful  study  of  erbium  oxide  and 
came  to  the  conclusion  that  the  bright  bands  were  emitted  by  the 
solid  and  not  its  vapor.  Higginaon,  1870,*'  confirmed  these  con- 
clusions, investigating  besides  erbium  a  large  number  of  other 
materials,  and  found  these  bright  lines  and  bands  in  the  spectra  of 
lime,  magnesia,  etc. 

The  practical  use  of  Auer  mantles  stimulated  further  research 
into  the  properties  of  the  rare  earths.  In  1891  Haitinger"  studied 
neodymium  and  praeseodymitim,  using  mantles  saturated  with  the 
nitrate  solutions.  He  found  that  pure  neodymium  shows  the  phe- 
nomenon very  weakly  and  praeseodymium  not  at  all,  but  that  the 
addition  of  1  per  cent  or  less  of  aluminum  oxide  brings  out  the 
bright  banda  in  both  cases.  The  marked  effects  produced  by  the 
addition  of  small  quantities  of  one  earth  to  another  in  greatly  in- 
creasing the  luminous  radiation  have  led  to  the  widely  different 
views  that  have  been  taken  in  explanation  of  the  efficiency  of  the 
incandescent  mantle. 

Without  attempting  a  chronological  treatment  of  the  suggested 
hypotheses,  it  is  interesting  and  important  to  mention  briefly  some 
of  the  theories "  that  have  been  proposed,  principally  because  at 
the  present  time  no  one  theory  is  universally  accepted.  One  of  the 
earliest  theories  accounted  for  the  high  efficiency  on  the  basis  of 
phosphorescence,  and  there  are  those  of  -the  present  time  who  hold 
that,  although  temperature  radiation  enters,  the  peculiarly  high 
efficiency  is  to  be  ascribed  to  some  form  of  luminescence.  For  the 
most  part,  however,  the  theories  have  been  based  on  temperature 
radiation,  but  the  variations  have  arisen  in  attempting  to  explain 
the  observed  phenomena  on  tiiis  basis.  According  to  one  theory 
which  held  sway  for  a  time  the  high  temperature  was  produced 

>y  Google 

Physical  CHAEACTEMSTiofl  of  Luminous  Sookchb       69 

locally  by  a  catalytic  action  of  the  particles  of  the  mixture  of  ceria 
and  thoria  composing  the  mantles.  Even  at  the  present  time 
catalysis  in  one  form  or  another  is  suggested  as  the  cause  of  the 
high  ^ciency. 

The  most  probable  theory,  as  accepted  at  the  present  time,  waa 
first  proposed  by  Nemst  and  BOse,  and  afterward  further  elaborated 
by  F6ry.  It  is  based  on  pure  temperature  radiation  with  selective 
emission,  and  suggests  an  explanation  of  the  peculiar  effects  of 
mixtures  which  have  made  the  problem  so  difficult  of  solution. 
Before  outlining  the  theory  the  principal  facta  which  seem  to  be 
fairly  well  established  should  first  be  pregented.  These  facts  are: 
(1)  The  temperature  in  the  ordinary  Bunsen  flame  probably  does 
not  exceed  3130°  or  3140°  Abs.  at  the  region  of  highest  tempera- 
ture, and  consequently  could  not  account  for  the  high-luminous 
^ciency  of  the  mantle  if  the  latter  radiated  as  a  black  body;  (3) 
the  spectra  of  the  rare-earth  oxides  are  in  general  peculiar  in  ex- 
hibiting banded  spectra;  (3)  when  a  small  quantity  of  certain  of 
the  rare-earth  oxides,  as  ceria,  is  intimately  mixed  with  some  other 
such  rare-earth  oxide  as  thoria,  and  the  mixture  in  a  finely  divided 
state,  as  in  the  incandepcent  mantle,  is  heated  in  a  Bunsen  flame, 
the  mixture  has  a  much  higher  luminosity  than  either  constituent 
separately;  (4)  the  luminwity  of  the  mixture  of  ceria  and  thoria 
depends  greatly  on  the  proportions  of  the  two  constitueuta  present 
As  the  ceria  is  increased  from  0  per  cent  to  1  per  cent  the  luminos- 
i^  riees  to  about  10  or  15  times  its  initial  value,  but  rapidly  de- 
creases again  as  the  proportion  of  ceria  is  increased  beyond  1  per 
cent;  (5)  the  pure  thoria  mantle  is  probably  at  a  temperature  be- 
tween 100°  and  150°  lower  than  that  of  the  flame,  and  the  addi- 
tion of  ceria  causes  a  still  further  decrease;  (S)  thoria  has  a  low 
emisaivity,  and  no  favorable  selective  emission  in  the  visible  spec- 
trum, whereas  ceria  has  a  much  higher  emisaivity  and  pronounced 
selectivity  in  the  visible  spectrum. 

The  theory  most  generally  accepted  aa  accounting  for  the  phe- 
nomena exhibited  by  the  mantle  with  varying  proportions  of  thoria 
and  ceria  depends  on  the  radiating  properties  of  the  two  substances 
as  given  in  (6).^  The  mantle  of  pure  thoria,  owing  to  its  low  emis- 
sivity,  assumes  a  temperature  but  slightly  lower  than  that  of  the 
flame,  but  its  luminosity  is  not  great  because  the  temperature  is 
not  very  high,  and  thoria  shows  no  favorable  selective  emission  in 
the  visible.     The  introduction  of  a  very  small  quantity  of  ceria 

>y  Google 

70  Illdminatino  Enoinebeino 

lowers  the  temperature  slightly  because  of  the  greater  emisaivity 
of  ceria,  bat  this  decrease  in  temperature  is  much  more  than  com- 
pensated for  bj  the  pronouDced  selective  emissioii  of  the  ceria  Id 
the  visible  Bpectrum.  The  integral  effect  therefore  is  to  increase 
greatlj  the  luminosity  of  the  mantle.  As  the  proportion  of  ceria  is 
increased  the  luminoeilry  constantly  irises  untU  the  composition  of 
the  mantle  is  99  {ier  cent  thoria  and  1  per  cent  ceria,  when  the 
maximum  luminosity  is  obtained.  Further  increments  of  oeria 
produce  decreases  in  the  luminosity  because,  owing  to  the  high 
emiBsivity  of  ceria,  the  temperature  of  the  mantle  drops  so  low  that 
the  selective  emisaion  of  the  ceria  is  no  longer  sufficient  to  com- 
pensate for  the  decrease  in  temperature. 

Although  this  theory  is  probably  the  one  most  generally  accepted 
at  present,  it  is  still  open  to  question,  and  certain  facts  point  to 
the  existence  of  catalysis,  or  luminescence,  or  perhaps  both.  The 
peculiar  nature  of  the  spectra  of  the  rare  earths  makes  the  problem 
difficult,  as  ordinary  optical  pyrometry  is  likely  to  give  quite  er- 
roneous results.  Thus  the  temperature  of  the  mantle*  has  been 
variously  estimated  from  1920°  to  2470°  Abs.,  and  it  is  difficult 
to  assign  the  correct  value.  The  temperature  of  the  Bunaen  flame 
is  by  no  means  definitely  established,  and  even  if  it  were  there 
would  still  be  difficulties  in  arriving  at  the  temperature  of  the 
mantle.  Those  physicists  who  subscribe  to  the  catalytic  theory 
would  see  no  objection  to  assigning  to  the  mantle  a  temperature  in 
excess  of  that  of  the  flame.  If  there  is  no  excess  temperature,  the 
question  still  remains  as  to  what  extent  the  temperature  of  the 
mantle  is  lower  than  that  of  the  flame.  The  use  of  very  small 
thermo-couples  by  White  and  Travers  has  led  to  the  value  3020" 
to  2120°  Abs.  as  the  maximum  temperature  of  the  Bunsen  flame, 
and  the  same  method  applied  to  the  mantle  has  indicated  tempera- 
tures 100°  or  150°  lower.  The  escesaively  high  values  that  have 
been  suggested  have  been  obtained  from  the  use  of  optical  methods 
which  are  subject  to  large  errors  in  cases  of  such  selective  radiation 
as  that  exhibited  by  the  Auer  mantle. 

In  a  similar  way  there  is  difficulty  in  determining  the  luminous 
efficiency  of  the  incandescent  mantle.  White  and  Russell  give  as 
the  consumption  for  the  most  efficient  mantle  containing  1  per  cent 
cerium,  35  British  thermal  units  per  hour  per  candle  (presumably 
measured  horizontally).  Since  IB. t.u.  per  hour  equals  approxi- 
mately ^jT — ^  calories  per  second,  and  1  calory  per  second  equals 

>y  Google 

Phtbioal  Charaotbbistiob  of  LuuiNons  Sodboes        71 

4.19  watts,  the  watts  per  mean  sj^erical  candle  (taking  the  spherical 

reduction  factor  equal  to  0.88  as  given  by  Lux)  are  ^Jr^^  X  s^ 

=  11  watta  per  mean  spherical  candle,  or  approximately  1.1  Inmens 
per  watt.  According  to  Fulweiler  the  moat  efficient  incandescent 
mantle  can  be  operated  at  20  B.  t.  n.  per  candle,  which  would  reduce 
the  above  values  to  approximately  €  watts  per  mean  spherical  candle, 
or  about  2  lumens  per  watt.  Lux  gives  for  a  mantle  containing 
0.8  per  cent  cerium  practically  the  same  values  as  those  found  by 
White  and  Sussell  for  a  1  per  cent  cerium  mantle,  and  for  a  mantle 
containing  0.1  per  cent  cerium  he  finds  an  efficiency  of  about  three- 
fifths  that  of  the  O.S  per  cent  cerium  mantle. 

Aa  stated  in  an  earlier  paragraph  regarding  fiomes,  the  effici^cy 
obt&iiied  as  the  ratio  of  the  light  produced  to  the  heat  energy  sup- 
plied is  not  entirely  comparable  with  the  efficiency  derived  for  an 
electric  incandescent  lamp  by  dividing  the  lumens  emitted  by  the 
watts  supplied,  because  in  the  generation  of  tiie  electric  power  the 
efficiency  of  heat  transformation  is  not  100  per  cent.  In  a  similar 
way  the  values  given  for  the  incandescent  mantle  are  not  com- 
parable with  those  ordinarily  given  for  electric  lamps.  If  the 
analysis  were  carried  back  to  the  coal  in  each  case  a  more  accurate 
comparison  could  be  made,  but  such  an  analysis  is  beyond  the  scope 
of  this  paper. 

The  ratio  ^ ,  i.  e.,  the  ratio  of  energy  radiated  in  the  visible 

spectrum  to  total  energy  radiated  has  been  found  by  White  and 
Travere"  to  be  0.045,  being  quite  close  to  the  value  0,06  for  the 
Nenist,  as  obtained  from  several  determinations.  The  ratio  of  L 
to  the  total  energy  supplied  is  given  by  Lux  "  as  0.005,  indicating 
that  only  about  one-tenth  the  energy  supplied  is  radiated  by  the 
mantle.  The  remainder  must  be  lost  by  conduction  and  convection. 
These  figures  are  given  merely  to  show  the  order  of  magnitude  of 
the  various  energy  losses  as  far  as  the  published  results  may  be 

Although  no  attempt  has  been  made  to  give  numerical  values  for 
open  gas  flames  of  ordinary  illuminating  gas,  it  may  be  well  to 
mention  briefly  some  of  the  characteriatica  of  acetylene.  Various 
values  have  been  assigned  for  the  temperature*  of  the  acetylene 
flame,  but  it  is  probable  that  the  temperature  is  not  far  from 

2300°  Abs.     For  p-  *  the  average  of  several  determinations  by 

>y  Google 


Angstrom,  Nichols  and  Cobleutz,  and  Stewart,  is  about  0.045,  ihe 
eame  as  that  given  for  the  iiic&ndesrant  mantle.  Similarly  for  the 
efiBciency  "  Liebenthal  quotes  for  ordinary  burners  an  average  spe- 
cific conBumption  of  1.1  liters  of  gas  per  candle-honr  with  a  possible 
minimum  of  0.65  liter  per  candle-honr.  Taking  for  acetylene  the 
heating  value  given  by  Morehead  "  these  figures  lead  to  an  average 
specific  consumption  of  19.3  watts  per  mean  spherical  candle  with 
a  minimum  specific  consumption  of  11.6  watts  per  candle,  corre- 
sponding to  0,65  and  1.1  lumens  per  watt,  respectively.  Lux  gives 
for  acetylene  the  specific  consumption  of  17.7  watts  per  candle, 
corresponding  to  an  efficiency  of  0.7  lumen  per  watt. 

It  is  to  be  borne  in  mind  that  in  all  discussions  of  flames  and 
mantles  large  discrepancies  may  arise  owing  to  the  nature  of  the 
burner  used,  or  to  the  exact  nature  of  the  gas,  or  to  the  regulation 
of  the  gas  in  the  burner.  For  these  reasons  only  approximate 
values  are  attempted. 

6.  The  Nemat   Glower 

Closely  akin  to  the  incandescent  mantle  and  suggested  by  it, 
the  N'emst  glower  nevertheless  stands  out  uniquely  from  any  other 
practical  illnminant.  Like  the  incandescent  mantle  it  is  com- 
posed of  oxides  of  the  rare  earths,  but  unlike  the  mantle  it  is  heated 
to  incandescence  by  the  passage  of  an  electric  current.  The  Nemst 
glower  is  what  is  known  as  a  solid  electrolyte,  i.  e.,  a  substance 
which  conducts  electrolytically  (as  distinguished  from  metallically) 
when  at  a  sufficiently  high  temperature.  At  ordinary  temperatures 
it  is  an  electric  insulator. 

The  work  of  Nemst,  which  led  to  a  patent  application  in  1897, 
was  probably  anticipated  to  a  certain  extent  by  Jablochkoff,"  who 
in  1879  made  an  electric  lamp  whose  radiating  body  was  made  of 
a  small  plate  of  kaolin,  a  portion  of  which  was  rendered  incan- 
descent by  the  spark- discharge  current  of  an  induction  coil.  The 
detailed  patents  of  Nernst,  given  out  in  1901,  covered  a  number  of 
combinations,  involving  the  oxides  of  zirconium,  thorium,  cerium, 
erbium  and  yttrium,  which  may  be  used  to  make  satisfactory 
glowers.  According  to  an  analysis  given  by  Beebe"  several  years 
ago  (1905),  the  regular  commercial  glower  as  manufactured  in  this 
country  consists  normally  of  85  per  cent  of  zirconia  and  15  per  cent 
of  yttria,  but  from  other  descriptions  that  have  been  given  it  seems 
probable  that  erbia,  thoria  and  ceria  have  at  times  been  included. 

>y  Google 

Physical  Charactsristios  of  Luminous  Soubces       73 

It  iB  an  interesting  fact  that  the  pure  oxides  are  not  as  satisfactory 
as  are  mixtures  of  tro  or  more  oxides,  either  from  the  standpoint  of 
electrical  conductivity  or  luminous  radiation. 

The  energy  relations  in  the  Nernat  glower  are  still  to  a  great 
extent  a  matter  of  speculation.  The  destructive  electro-chemical 
decomposition  at  the  electrodes  which  accompanies  electrolytic  con- 
duction is  supposed  in  the  case  of  the  Nemst  glower  to  be  counter- 
acted by  the  oxidizing  action  of  the  sir.  The  glower  consequently 
will  not  operate  in  a  vacuum,  and  is  hence  suhject  to  losses  of 
energy  by  thermal  conduction  and  convection  of  the  air.  Further- 
more, it  has  been  sugg^ted**  that  since  the  atmosphere  surround- 
ing the  glower  is  ionized  and  there  is  present  a  very  appreciable 
potential  gradient  some  of  the  energy  supplied  the  lamp  may  be 
conducted  electrically  through  the  surrounding  air  and  not  through 
the  glower  body. 

Regarding  the  losaea  through  thermal  conduction  and  convection 
in  the  air  there  are  several  published  estimates."  According  to 
Hartmann  these  losses  amount  to  anywhere  from  5  per  cent  to  70 
per  cent,  depending  upon  the  assumption  on  which  the  estimate 
is  made.  Lux  gives  the  loss  as  about  30  per  cent,  and  Leimbach 
as  approximately  GO  per  cent.  A  recent  experiment  made  to  indi- 
cate roughly  the  order  of  magnitude  of  these  losaea  gave  as  a  result 
a  loss  of  approximately  50  per  cent  to  within  ±  10  per  cent.  From 
the  measured  losses  in  a  platinum  lamp  when  burning  in  air  it 
would  scarcely  seem  probable,  notwithstanding  the  distinct  char- 
acteristics of  the  two  filaments,  that  the  losses  in  the  Nemst  should 
be  as  low  as  5  per  cent  or  10  per  cent- 
Due  to  the  pronounced  negative  temperatare  coefficient  for  the 
material  of  the  glower  at  ordinary  temperature,  it  is  necessary  to 
place  in  series  with  the  glower  a  ballast  resistance.  The  magnitude 
of  this  temperature  coefficient  is  evidenced  by  the  fact  that  for  a 
110-volt,  44-watt  lamp  the  resistance  of  the  glower  which  is  320 
ohms  at  normal  voltage  (110  volta),  drops  to  240  ohms  at  135 
volts,  and  rises  to  600  ohms  at  92  volts."  The  necessary  ballast,  for 
which  is  chosen  a  material  with  high-positive  temperature  coeffi- 
cient, must  have  a  resistance  such  that  at  normal  burning  10  per 
cent  of  the  supplied  energy  is  lost  in  the  ballast.  There  is  another 
known  loss  of  2  per  cent,  which  arises  from  the  necessity  of  a 
magnetic  cut-out  to  throw  the  heating  coil  out  of  circuit  when  the 
glower  begins  to  conduct." 

>y  Google 

74  Illumtnatinq  En(}I(ibebinq 

The  question  of  lossee  is  intimately  connected  with  that  of  the 
explanatioQ  of  the  efficiency  "  of  the  Nemst,  for  which  values  rang- 
ing from  2.0  to  3.0  watte  per  mean  spherical  candle  have  heen  given. 
It  is  probable  that  the  average  value  lies  between  3,4  and  2.8  watts 
per  mean  spherical  candle  corresponding  to  4  or  5  lumens  per  ap- 
plied watt.  If  there  is  any  large  loss  of  energy  by  conduction  or 
convection,  this  efficiency  could  only  result  from  moderately  high 
temperature  or  markedly  selective  emission.  The  estimates  of 
temperature  "  that  have  heen  given  range  from  1800°  Abs.,  made 
from  extrapolated  thermo-conple  measurements,  to  2450°  Abs.,  de- 
termined by  optical  methods.  It  is  quite  probable  that  the  true 
temperature  is  at  least  above  2000°  Abs.,  and  hence  several  hundred 
degrees  higher  than  that  of  Uie  incandescent  mantle. 

To  what  extent  selective  emission  determines  the  efficiency  is 
not  known.  In  the  discussion  of  the  mantle  it  was  seen  that  the 
rare-earth  oxides  frequently  exhibit  selectivi^  to  the  extent  of 
pronounced  bands,  particularly  at  low  temperatures  and  in  a  finely 
divided  condition.  Such  a  banded  spectrum  has  been  observed  for 
the  Nemst  glower,  both  in  the  visible  and  infra-red  regions  at 
abnormally  low  temperatures,  but  the  bands  disappear  at  higher 
temperatures  so  that  in  the  neighborhood  of  the  temperature  of 
normal  operation  the  spectrum  is  practically  continuous.  At  this 
temperature,  if  there  is  any  selective  emission,  at  least  in  the  visible 
spectrum,  it  is  of  the  nature  of  that  found  for  metals,  being  merely 
an  exaggerated  relative  emiBsion  in  the  shorter  wave-lengths,  as 
compared  with  the  radiation  of  a  black  body  at  the  same  temperature. 

Various  estimates  have  been  made  of  the  ratio  [-=-  j  of  the 
energy  radiated  per  second  in  the  visible  spectrum  to  the  total 
emission  per  second.**  Ingersoll,  using  Angstrom's  method,  obtained 
for  a  110-volt,  89-watt  glower,  0,046;  Drude  quotea  a  value  of 
0.065  for  a  glower  at  1  watt  per  hefner,  which  perhaps  corresponds 
to  a  slightly  higher  temperature  than  normal  operation.  Coblentz, 
by  integration  of  an  energy  curve  obtained  0.055  for  a  filament  at 
83   watts    (presumably   88   watts  normal).     It  is  probable   that 

-=-  =0.05  expresses  approximately  the  relative  amount  of  energy 

radiated  in  the  visible  spectrum  between  A=0.38^  and  X=0.76^ 
Without  a  knowledge  of  the  losses  in  the  glower  it  is  imposMble  to 

>y  Google 

Petsioal  Chabaoteribtios  of  LuumocB  Sources       75 

per  secood  to  the  power  supplied  to  the  lamp.     If  we  aBsume  the 
condnctioii  and  conTection  losses  to  be  50  per  cent,  as  indicated  by 

Q  ~  R  -^  Q  " 

tion,  in  conjunction  with  the  average  value  of  2.6  watts  of  power 
supplied  per  mean  spherical  candle,  or  4.8  lumens  per  watt,  in  a 

normal  lamp,  we  obtain  the  value  of  ~- ,  the  ratio  of  luminous  flui 

to  total  energy  radiated  as  about  9.6  lumens  per  watt,  which  would 
be  obtained  from  a  black  body  at  g300°-3300°  Abs.  From  this  it 
would  seem  that  the  temperature  of  the  Kemst  is  either  quite 
high  or  else  that  there  is  selective  emission  to  partly  account  for 
the  efSciency. 

The  same  result  can  be  arrived  at  from  Drysdale's  "  value  of  the 

so-called  mechanical  equivalent  {-$-]  for  the   Nerast,  which  ia 

lows  that  -R-  =  L  ^  ^  =150x0.05  =  7.5  lumens  per  watt  radi- 
ated. Indeed,  the  extension  of  this  method  leads  indirectly  to  the 
energy  emission  in  the  Nemst.    For,  if  the  efficiency  is  7.5  lumens 

per  watt  radiated  f^j,  and  4.8  lumens  per  watt  suppliedf-—^] 

it  would  follow  that  -^  =0.64,  and  so  the  loss  (Q  — B)  is  approxi- 
mately 36  per  cent  of  the  power  supplied. 

7.  The  Pliynca  of  the  Fire-Fly  and  Other  lAghi-Prodvcing 

Although  the  fire-fly  can  scarcely  be  considered  as  a  commercial 
illuminant,  its  interest  and  the  attention  which  it  has  received 
merits  its  brief  consideration  here.  It  is  peculiarly  fitting  that 
this  natural  illuminant  should  be  discussed  at  the  end  of  the  series 
of  human  attempts  at  light  production  because,  from  the  stand- 
point of  radiant  efficiency,  it  surpasses  any  other  known  source. 
In  the  first  lecture  we  found  that  if  all  the  energy  supplied  to  a 
lamp  were  radiated,  and  if  all  the  radiant  energy  lay  at  that  wave- 

>y  Google 


length  in  the  fieible  spectrum  to  which  the  average  httman  eye  is 
most  BeDsitive,  the  highest  poBsible  efficiency  vould  be  obtained. 

In  the  fire-fly  we  have  practically  an  exemplification  of  at  leaat 
one  of  theee  requirementa,"  According  to  the  best  information  at 
present  it  would  appear  as  though,  from  the  standpoint  of  radia- 
tion, the  efliciency  of  the  fire-fly  is  almost  as  great  (estimated  at 
96-97  per  cent)  as  that  of  the  most  efficient  radiator  possible. 
Unfortunately,  we  do  not  know  what  chemical  and  biological  trans- 
formations occur  in  the  process  of  "glowing,"  and  without  thia 
knowledge  we  can  form  no  idea  of  the  real  efficiency  of  trans- 

The  process  of  light  production  in  the  fire-fly  is  called  lumines- 
cence, and  seems  to  depend  on  the  presence  of  oxygen  and  water. 
Other  living  organisms,  such  as  glow  worms,  certain  bacteria  and 
numerous  fishes,  exhibit  the  property  of  light  production,  but  oar 
knowledge  of  these  at  present  is  quite  meager.  Further  investi- 
gation of  these  natural  lamps  may  disclose  processes  of  light  pro- 
duction which  could  with  profit  be  copied  by  man  in  the  conatruetion 
of  artificial  sources. 

8.  The  Dittribviion  of  Energy  in  the  Speatra  of  the  Various 
Luminous  Sources 

One  of  the  most  interesting  of  the  physical  characteristics  of 
luminous  sources  is  the  distribution  of  energy  in  their  spectra.  The 
spectral  distribution  determines  the  ratio  (-p)  of  the  energy 
radiated  per  second  in  the  visible  to  that  radiated  per  second  in  the 
complete  spectrum,  and  also  the  ratio  f^j  giving  the  photo- 
metric value  of  the  visible  radiant  energy.  Eliminating  conduc- 
tion, convection  and  other  incidental  losses,  the  energy  distribution 
determines  the  commercial  efficiency  of  practical  sources,  accounta 
for  the  quality  of  the  composite  light,  and  explains  the  appearance 
of  colored  objects  illuminated  by  it.  We  are  interested  to  know 
whether  the  spectrum  of  a  source  is  continuous,  discontinuous  or 
banded;  what  proportion  of  the  energy  is  in  the  visible  spectrum, 
and  whether  in  cases  of  continuous  spectra  the  distribution  is  that 
of  black-body  radiation  at  some  temperature  or  distinctly  different 
from  it  owing  to  pronounced  selectivity. 

It  is  ordinarily  considered  that  gases  snd  liquids  when  incan- 

>y  Google 

Physical  Chasaotbbibticb  of  Lchinous  Soubces       77 

descent  emit  diseontinuoue  apectra,  but  that  Bolidfl  in  general  emit 
continiioiis  spectra  in  which  the  enei^  ia  difltributed  very  much 
ibe  same  as  in  the  spectrum  of  a  blaeh  body.  But  even  for  luminous 
gases  and  vapors  certain  distinctions  must  be  made  in  the  light  of 
recent  experiment  and  theory.  The  bright-line  epectriun,  as  in 
the  spectrum  of  sodinm  when  common  salt  is  heated  in  a  Bunsen 
flame,  has  generally  been  considered  ns  intimately  connected  with 
some  chemical  reaction,  in  the  course  of  which  the  sodium  atoms 
are  brought  into  a  radiating  state,  which  cannot  be  reproduced  by 
mere  heating  of  sodium  vapor.  When  gases  or  vapors  are  heated, 
it  has  usually  been  agreed  that  only  banded  spectra  are  obtained, 
except  when  the  temperature  is  so  high,  as  in  the  quartz-mercury 
arc,  that  there  is  a  continuous  ppectrum  as  background  to  the  bright 
lines.  The  bright  lines  are  obtained  only  whep  chemical  or  elec- 
trical excitation  is  employed,  and  not  when  the  gas  or  vapor  is 
merely  heated.  Some  recent  experiments  by  Bauer  seem  to  indi- 
cate an  exception  to  this  rale  that  bright-line  spectra  are  always 
associated  with  so-called  luminescence,  but  this  work  is  too  recent 
to  justify  a  reveraal  of  opinion  in  this  regard  at  the  present  time." 

For  solids  which  radiate  approximately  as  a  bJach  body  the  lumi- 
nous efficiency  increases  rapidly  with  the  temperature,  and  it  is  to 
the  temperature  influenced  to  some  extent  probably  in  all  cases, 
and  to  a  very  comiderable  extent  in  some  cases  by  selectivity  in  the 
emissionj  that  the  luminous  efBciency  of  many  sources  is  to  be 
ascribed.  There  is  a  class  of  solids,  however,  illustrated  by  the 
rare-earth  oxides  of  the  incandescent  mantle  and  the  Nernst  glower 
which,  though  solid,  exhibit  at  least  at  low  temperatures,  peculiar 
banded  spectra  superimposed  upon  a  continuous  background. 

It  is  perhaps  not  appropriate  to  discuss  in  these  lectures  theories 
of  spectral  energy  distribution,  however  attractive  such  a  com- 
parative discussion  might  seem  in  the  light  of  the  varied  spectral 
phenomena  exhibited  by  the  difFerent  commercial  light  sources.  The 
presentation  of  the  spectra  of  all  the  sources  in  one  section,  bow- 
ever,  does  not  contemplate  such  a  discussion,  but  is  intended  rather 
to  give  a  better  comparative  idea. 

In  Figure  6  are  plotted  the  spectral  energy  curves  in  the  visible 
region  for  various  common  light  sources.  These  curves  show  the 
absolute  distribution  of  energy  for  each  source,  but  not  the  absolute 
amount  of  energy  radiated  per  unit  area,  being  plotted  in  arbitrary 
units  BO  chosen  that  the  ordinate  at  A=0.S9/i  is  the  same  for  all. 

>y  Google 



The  curves  were  taken  from  the  best  published  values"  in  most 
cases,  but  were  partly  determined  by  the  author  to  fill  in  gaps  in 
the  literature  of  the  subject.  The  variations  in  the  cnrrea  obtained 
by  different  obeeirerB  for  some  of  the  sources  were  found  to  be 
strikingly  large,  o?ing  partly  to  the  conditions  of  the  experiment 

T  CurveB  Tor  Different  Sources  In  the  Visible 

and  partly  to  variations  in  the  exact  nature  of  the  illuminant.    In 
such  cases  what  seemed  to  be  the  beet  average  value  was  taken. 

Spectral  energy  curves  in  the  visible  spectrum  are  given  for  the 
following  light  sources:  ordinary  carbon  lamp  at  3.1  watts  per 
mean  horizontal  candle,  tantalum  lamp  ^t  S  watts  per  mean  hori- 
zontal candle,  and  tungsten  lamp  at  1.2S  watts  per  mean  horizontal 
candle;  Welsbach  regular  commerctKl  upright  mantle  (average  of 

>y  Google 

Phi^sical  Chasa,otsbi8ti08  of  Ldhikocs  Sources       79 

various  epecimene),  Nemst  glower  (average  of  Tariona  typee),  and 
acetylene  (average  of  various  determiiiatioiia  with  different  burners, 
and  under  different  conditiona) .  In  every  case  the  spectrum  is  con- 
tinuous. But,  although  the  spectra  are  continuous,  the  distribu- 
tion is  not  in  every  case  that  which  could  be  obtained  from  a  ilack 
body  at  the  proper  temperature.  Thus  the  incandescent  mantle 
shows  evidence  of  a  pronounced  selectivity  in  the  green. 

The  spectrum  of  an  incandescent  mantle  depends  greatly  on  the 
compoBition  of  the  mantle  and  on  the  temperature  to  which  it  is 
heated  in  the  Bnnsen  flame,  which  latter  depends  on  the  heating 
value  and  composition  of  the  gns  used,  and  on  the  adjustment  of 
the  flame.  The  spectrum  oi  acetylene  depends  on  the  burner  and 
on  the  thickness  of  the  flame,  containing  relatively  more  blue  in 
the  thin  than  in  the  thicker  flameB.  It  also  depends  to  some  extent 
on  atmoepberic  conditions.  Hence,  the  curves  given  for  these 
sources,  tiie  incandescent  mantle  and  acetylene,  can  only  be  con- 
sidered as  representative  of  the  general  type  of  curve  obtained. 
Indeed,  inveetigatoi^  frequently  fail  to  give  the  exact  specifications 
of  the  sources  employed,  making  accurate  comparisons  impossible. 
It  would  seem  as  though,  despite  the  immense  amount  of  work 
done  on  commercial  light  Bources,  there  is  still  lacking  a  compre- 
hensive comparative  study  of  the  exact  spectral  compositions  of 
these  sources  under  carefully  defined  conditions. 

It  was  hoped  that  a,  curve  for  the  open  arc  might  be  included, 
but  when  the  literature  was  searched  for  data  on  the  energy  dia-' 
tribution  in  the  visible  spectrum  of  the  arc,  very  few  curves  were 
found,  and  these  showed  such  enormous  discrepancies  that  no  value 
could  be  attached  to  a' mean  curve  derived  from  them.  The  curve 
is  contioQOUB  in  the  visible  spectrum  with  a  superimposed  band  in 
the  blue  due  to  the  arc  flame. 

The  spectra  of  the  Moore  tube,  the  mercury  arc  and  the  various 
luminous  and  flaming  arcs  have  not  been  given  in  Figure  6  because 
their  spectra  are  discontinuous,  consisting  mainly  of  distinct  bright 
lines,  which  it  would  be  diEBcult  to  represent  to  scale  of  intensity. 
For  such  sources  the  most  important  facta  for  us  to  know  are  (1) 
how  closely  the  bright  lines  occur,  and  (2)  the  integral  color  of 
the  light,  i.  e.,  the  color  which  a  white  surface  would  assume  when 
illuminated  by  the  light.  These  questions  will  be  discuesed  sub- 
sequently in  considering  the  quality  of  the  light  from  the  various 
sources  as  determined  by  the  use  of  the  colorimeter. 

>y  Google 


Illuminatixq  Enqinbkbinq 

Compar&ble  in  importance  with  the  spectral  energy  distribntion 
in  the  visible  epectrnm,  that  in  the  infra-red  region  demands  our 
conaideration."  In  fact  it  is  the  relative  amount  of  energy  in  the 
visible  as  compared  with  that  in  the  infra-red,  taken  together  with 
the  distribution  in  the  visible  spectrum,  which  determines  the 
candle-power  for  each  watt  radiated.  The  difficulties  in  the  way 
of  making  accurate  infra-red  measurements  are  in  some  ways  greater 
than  thoee  encountered  in  the  visible  spectrum,  which  no  doubt 
expliiius  the  paucity  of  available  data  on  infra-red  energy  curves. 
Goblentz  gives  as  the  infra-red  curve  for  a  tungsten  lamp,  pre- 





PiQ.  7. — a.  Energy  Curve  (or  Tungsten  lAmp  at  Normal  Tottage. 
b.  Energy  Curve  for  a  Stack  Body  at  2200°  Atw. 

sumably  under  normal  conditions,  that  shown  In  Figure  ?,  curve 
"  a."  In  general  form  it  resembles  the  energy  curves  of  a  black 
body,  but  differs  somewhat  from  the  latter.  Thus,  if  we  plotted 
the  energy  curves  of  a  hJack  body  at  such  a  temperature  Uiat  the 
distribution  in  the  visible  spectrum  was  the  same  as  that  of  the 
tungsten,  the  two  curves  would  differ  in  the  infra-red,  that  of  the 
tungsten  lying  below  that  of  the  f/iade  body.  Such  a  hlack-bodg 
curve  is  given  in  curve  "  b,"  corresponding  to  a  black  body  at 
2300°  Abs.,  at  which  temperature  the  black  body  has  approximately 
the  same  distribution  of  energy  in  the  visible  spectrum  as  the 
tungsten  lamp  at  normal  efficiency.     The  two  curves  are  plotted 

>y  Google 


to  the  Bame  oidinate  at  the  eame  wave-length  of  the  visible  apec- 
tnun,  ea;  k—0,7fi. 

For  the  electric  incandeecent  lamps  at  nonnal  operation,  curves 
somewhat  similar  to  that  of  tungsten  would  be  found.  For  carbon 
the  curve  would  correspond  more  nearly  to  that  of  a  blach  body, 
but  the  temperature  of  tlie  black  body  for  the  same  spectral  distri- 
bution in  the  visible  would  be  lower  than  that  for  tungsten.  For 
osmium  probably  the  greatest  deviation  from  the  black  body  would 
be  found. 

The  infra-red  curve  for  an  incandescent  mantle  (composition 
99.3  per  cent  thoria,  0.8  per  cent  eeria)  has  been  found  by  Kubens, 
by  subtracting  the  radiation  of  tlie  open  Bunsen  burner  from  the 
combined  radiation  of  the  burner  and  mantle.  This  curve  is  given 
in  Figure  8.  The  peculiar  broken  form  of  the  curve  seems  to 
he  characteristic  of  the  radiation  from  rare  earths  at  moderate  tem- 
peratures. The  curve  for  the  Nemst  found  by  Coblentz  (Figure  9) 
is  quite  smooth,  although  the  glower,  lilce  the  incandescent  mantle, 
is  composed  of  rare-earth  oxides.  But  at  lower  temperatures  the 
Nemst  glower  also  shows  evidence  of  a  banded  spectrum.  Whether, 
in  the  case  of  the  Nemst  at  normal  operation,  the  smoothness  is 
to  be  ascribed  entirely  to  a  higher  temperature  than  the  mantle, 
or  to  its  more  compact  form,  remains  to  be  determined. 

>y  Google 

Illuminating  Enqinbbbinq 

In  Fignre  10  is  given  the  infra-red  curye  of  acetylene  as  found 
by  Stewart.  For  the  same  reasone  aa  those  given  for  omitting  the 
visible  spectra  of  the  various  fiaming  and  luminous  arcs  and  vacnnm- 
tube  lamps,  as  the  Moore  tube,  no  attempt  will  be  made  to  give 
here  the  infra-red  energy  curves  of  these  a 

Fio.  9. — Energy  Distribution  of  a  110-Tolt  Nemst  Qlower  Operated  at 
77.7  Watts — According  to  Coblentz. 

9.  The  Qwility  of  the  Light  from  the  Various  Luminous  Sources 

Closely  associated  with-  the  question  of  spectral  energy  distribu- 
tion la  that  practical  one  of  the  quality  of  the  light  from  illumi- 
nante  and  the  appearance  of  colored  objects  when  illuminated  by 
these  illiiminants,  as  eicplained  at  length  in  the  first  lecture.  When 
the  energy  distribution  in  the  visible  spectrum  is  continuous  and 
represented  by  a  smooth  curve,  the  integral  color  of  a  source,  i.  e., 
the  color  which  a  white  surface  illumiuated  fay  it  assumes,  is  a  fair 
indication  of  the  variation  in  color  values  which  will  occur  when 
the  source  is  substituted  for  average  daylight,  taken  as  normal 
white  light.    But  if  the  spectrum  of  an  illuminant  is  discontinuous. 

>y  Google 

Phtsicai,  Charaoiebistics  of  Luminous  Souboes       83 

composed  of  a  number  of  distinct  lines,  the  distribution  of  these 
lines,  together  with  the  integral  color,  must  be  examined. 

The  integral  color  of  the  light  from  any  source  can  readily  be 
measured  by  determining  the  relative  amounts  of  red,  green  and 
blue  light  which  when  mixed  give  a  resultant  color  which  matches 
in  hue  that  from  the  source  under  investigation.     Such  measure- 

ments carried  out  with  tho  F.  E.  Ives  colorimeter  have  been  pub- 
lished by  H.  E.  Ives  "  for  a  number  of  illuminants.  These  results 
are  given  in  Table  II.  White  light  is  taken  as  that  emitted  by  a 
hlach  body  at  6000"  Abs.,  for  which  the  sensatioii  values  are  red 
33.3  per  cent,  green  33.3  per  cent  and  blue  33.3  per  cent  The 
color  values  of  the  varioue  illominants  are  expressed  in  terms  of 
red,  green  and  blue  senaationa,  such  that  the  three  values  given 
add  up  to  100  per  cent. 

From  a  consideration  of  this  table  it  is  seen  that  the  carbon- 
dioxide  vacuum  tube  approaches  most  nearly  to  average  daylight. 

>y  Google 

84  Illchinatino  Enoinberiko 

Althoiigh  the  spectrum  of  the  vactmm-tube  sonrce  is  always  die- 
continuous,  the  number  of  bright  lines  in  the  Bpectmm  of  carbon 
dioxide  is  very  large,  and  the  lines  are  distributed  throughont  the 
entire  Tuible  spectrum,  being  thus  equivalent  for  practical  purposes 
to  a  continuous  spectrum.  The  other  sources  which  show  discon- 
tinuous spectra,  as  stated  in  the  discussion  of  spectral  energy  dis- 
tributions, are  the  low-pressure  mercury  arc  and  the  ordinary  lumi- 
nons  and  flaming  arcs.  In  the  case  of  the  mercury  arc  the  effect 
of  the  visible  spectrum  being  composed  of  a  few  lines  widely  sep- 
arated is  plainly  shown  in  the  unnatural  appearance  of  certain 
colored  objects  illuminated  by  its  light. 

One  sLgnificiint  feature  in  regard  to  the  integral  color  of  light 
sources  is  the  relatively  different  impressions  produced  by  two  lights, 
each  slightly  different  from  average  daylight,  when  the  direction 
of  the  difference  is  one  way  or  another.  If  the  color  of  a  light  is 
approximately  that  which  a  black  body  gives  at  some  temperature, 
it  does  not  appear  nearly  so'  strikingly  different  from  daylight, 
although  the  hue  may  be  diatinctly  reddi^,  as  a  light  which  differa 
from  daylight  in  such  a  way  as  not  to  lie  on  the  scale  of  color 
which  a  hlach  body  assumes  as  the  temperature  is  varied.  The 
explanation  of  this  phenomenon  comes  rather  within  the  province 
of  physiological  optics  than  that  of  physics. 

GuAsincATTON  OF  Light  Soubcbb  Accqbotno  to  Coutt  Values 

^"""-  Bed.  Greeo.  Blue. 

1.  Black  body  at  BOOO"  Abs 33.3%  33.3%  83.3% 

2.  Blue  sky   32,0  32.2  3B.8 

3.  Overcast  sky   34.6  33.9  31.6 

4.  Afternoon  eun    37.7  37.3  25.0 

6.  Hefner    55.0  38.8  6.2 

6.  3.1  w.  p.  c.  carbon  lamp 51.3  40.4  8.3 

7.  Acetylene   49.1  40.5  lO.S 

8.  Tungsten,  1.26  w.  p.  c 48.7  40.5  10.9 

9.  Nernst    49.2  40.7  11.1 

10.  WelBbach,  %%  ceria   42.S  40.8  16.7 

11.  WelBbach,  %%  cerIa  45.6  42.0  12.5 

12.  WelBbach,  114%  certa  47.8  41.8  11.0 

13.  Direct  current  arc   41.0  36.3  22.7 

14.  Mercury  arc    29,0  30.3  40.7 

15.  Yellow  flame  arc 52.0  37.5  10.6 

16.  Moore  carbon  dioxide  tube 31.3  31.0  37,7 

>y  Google 

Physical  CHAEAOTBHisnca  of  Luminous  Sodeces       85 


1.  R.  Ton  Helmholtz,  Belblaetter,  14,  p.  689.  1890. 
U  Hartman,  Ph^B.  Zelt.  5.  p.  679,  1904. 

L.  Hartman,  Phre.  Rev.  20,  p.  322,  1906. 

H.  Liu.  IlL  E^.  (Lond.)  1,  p.  98,  1808. 

C.  V.  Drysdale,  Proc.  Roy.  Soc.  A.  80,  1907. 

T6tj  et  ChflDeveau,  Bui.  Soc  Int.  dee  Bllec.,  2d  Ser.  9,  p.  683,  1909. 

Q.  Lelmbach,  Za.  f.  wlsa.  Pbotog.  8,  p.  333,  1910. 

2.  H.  Rubens  and  H.  Hollnagel,  Preuaa.  Akad.  WIm.,  B^lln,  Slta.  Ber. 

i,  p.  26.  1910.    See  also  Scl.  Aba.  A.  13,  p.  182, 1910. 

3.  J.  TbomBen,  Pogg.  Aon.  JUS,  p.  348,  1866. 

O.  Tumllra  u.  A.  Krug,  Wien.  B«r.  97,  pp.  1621  and  1626,  1888. 

O.  Tumllrz.  Ann.  der  Phra.  3S,  p.  640.  1889. 

C.  Bering,  Elec.  World,  S7,  p.  631,  1901. 

K.  AngEtrdm,  Pbr&  Zelt.  a.  p.  267,  1902. 

K.  jlngBtrfim,  Pbya.  Rer.  17,  p.  303,  1903. 

H.  anmaebl,  Z.  f.  d.  Phyalk,  u.  Cbem.  Unterr.  16,  p.  210,  1903. 

K.  Schaum,  Zelt.  f.  win.  Pbotog.  2.  p.  389.  1904. 

W.  Wedding,  Jour.  t.  Gaabeleucb.  1^05.     See  also  Ueber  den  Wlr- 

kungsgrad  und  die  Praktiecbe  Bedeutung  der  gebraflchlichtBten 

Llcbtquellen.  Hflnchen.  1905. 

B.  L.  Nichola.  III.  Eng.  (N.  Y.)  1,  p.  427,  1906. 
J.  Russner,  Pbya,  Zeit.  8,  p.  120,  1907. 

W.  Toege,  Pbys.  Zelt  8,  p.  306,  1907. 
H.  Lux,  Zeit  fur  Beleuch.  No.  16,  1907. 

C.  V.  Drysdale,  Proc.  Royal  Soc  A.  SO,  1907.     See  also  IlL  Eng 

London,  1,  p.  164,  1908. 
H.  E.  Ivea,  Trans.  lU.  Eng.  Soc  5,  p.  113,  1910. 

4.  H.  8.  Weber.  Phye.  Rev.  2,  p.  112,  1894. 
B.  L.  Nichols,  Phys.  Rev.  11.  p.  221.  1900. 
K.  AngatrOm,  Pbys.  Rev.  17.  p.  302.  1903. 

B.  Liebentbal,  Praktlscbe  Pbotometrle,  p.  SO,  1907. 
H.  Lux.  III.  Eng.  (Lond.)  1.  p.  99,  1908. 

C.  Drysdale.  III.  Eng.  (Lond.)  1,  p.  460,  1908. 
H.  E.  Ives,  Tram.  111.  Eng.  Soc  5,  p.  113,  1910. 

E.    G.  H.  Stlckney,  Trans.  111.  Bag.  Soc  2,  p.  282,  1907. 
P.  Bauder,  Jour.  SYank.  Inat:  ie9.  p.  223,  1910, 
H.  E.  Ives.  Trans.  111.  Eng.  Soc  5,  p.  189,  1910. 
8.    G.  R.  Klrchhoff,  Fogg.  Ann.  109.  p.  292,  1860. 

0.  Lummer,  NaturwlaBenscbaftllcbe  Rundacbau.  11.  p.  65.  1895. 

0.  Lnmmer  u.  E.  Prlngsbelm,  Wled.  Ann.  es,  p.  395.  1897. 

O.  Lummer  u.  F,  Kurlbaum.  Verb,  der  Pbya.  Gesell.  zu  Berlin,  17, 

p.  106,  1898. 
O.  Lummer  u.  E.  Prlngsbelm,  Verb,  der  Pbys.  Gesell.  1.  pp.  23  and 
216,  1899. 
7.   W.  Wbitney.  Gen.  Elec  Rev.  S.  p.  101,  1910.    See  also  Elec.  Age,  il. 
p.  70,  1910. 

>y  Google 

86  Illuminating  Enginkebinq 

8.  r.  Knrlbaum  und  O.  Schulz^,  Ber.  der  Deuticbe.  PhTS.  Oeaell.  1, 

p.  429,  1903. 
C.  Watdner  and  Q.  Burgess,  Bui.  Bur.  of  Stds.  t,  p.  319,  1906. 
F.  Leder,  Ann.  der  Phre.  24,  p.  305,  1907. 
W.  Coblentz,  Bui.  Bur.  of  Stds.  5,  p.  339,  1909. 

E.  P.  Hyde,  Jour.  Frank.  Inst.  169,  p.  439,  1910.    See  also  HI.  Bng. 

(Lond.)   S,  p.  241,  1909. 
8a.  Drude,  Lehrbuch  der  Optik,  Leipzig,  p.  480,  1906. 

9.  H.  a.  Weber,  Phys.  Rbt.  B.  p.  112,  1894. 

10.  B.  Merrltt,  Sill.  Jour.  37,  p.  167.  1889. 

H.  Crew  and  O.  H.  Baequin.  Hep.  of  Brit,  Assoc.,  p.  677,  1897. 
B.  Raach,  Elek.  Zelt  28,  p.  155,  1601. 
W.  Nernst,  Elek.  Zelt  S2,  p.  256,  1901. 

B.  Rasch,  Elek.  u.  Mascb.,  Heft  7,  1903. 

F.  a.  Bailey,  Elec.  (Lond.)  .52,  p.  646,  1904. 

E.  Ascbkluaes,  Ann.  der  Pbys.,  Ser.  i,  17,  p.  960,  1906. 

C.  B.  Mendenhall,  Phys.  Rev.  20,  p.  160,  1905. 

G.  Waldner  and  O.  Burgess,  Bui.  Bur.  of  Stds.  S,  p.  327,  1906. 
J.  Russner,  Pbys.  Zelt.  S,  p.  120,  1907. 

W.  W.  Coblentz,  Bnl.  Bur.'of  Stds.  5,  p.  339,  1909. 
R.  E.  Nysander.  Pbys.  Rev.  S8,  p.  438,  1909. 

B.  P.  Hyde,  Jour.  Frank.  Inst.  169,  p.  439.  1910. 

11.  M.  Lucas.  C.  R.  100,  p.  454,  1885. 

M.  Le  Chateller.  Jour.  6.  Pbys.  1,  p.  203,  1892. 

H.  S.  Webor.  Pbys.  Rev.  S.  p.  112,  1894. 

P.  Janet.  C.  R.  ItS.  p.  734,  1898. 

O.  Lummer  and  B.  Pringslielm,  Verb,  der  Deutsch.  Pbys.  Oes.  1,  n 

235,  1899. 
L.  Lombardl,  Elek.  Zelt  £5.  p.  41,  1904. 
G.   Clerlcl,   Elettrlclta,   Milan,    SS,    p.   156,   1907.     See  also   Elec 

(Lond.)  59,  p.  226.  1907. 
A.  Qrau.  Elek.  u.  Mascb.  25,  p.  296,  1907. 
Morris,  Stroude  and  Bllis.  Elec.  (Lond.)  S9,  p.  584,  1907. 
P«ry  et  CbSneveau,  Bui.  Soc.  Int.  des  Elec.,  2d  Ser.  9.  p.  683,  1809, 
M.  V,  Plranl,  Ber.  der  Deiitscb.  Pbys.  Qes.  IS.  p.  801,  1910. 
a.  Scbulze,  see  E.  Llebenthal,  Praktiache  Pbotometrle,  p.  836,  1907. 

12.  H.  Th.  Simon,  Pbys.  Zelt.  6.  p.  297,  1905. 

13.  W.  Kautmann,  Ann.  der  Phys.  S,  p.  158.  1900, 

C.  StelnmeU,  Trans.  Int  Elec.  Cong.,  St  Louis,  t,  p.  726,  1904. 
C.  Stelnmetz,  Radlat  Light  ft  111.,  p.  137,  1909. 

14.  Q.  Scbultz,  Ann.  d.  Phys.  H).  12.  p.  828,  1903. 

C.  D.  Cblld,  Trans.  Int.  Elec.  Cong.,  St.  I«uts,  1,  p.  193,  1904. 
J.  3.  Thomson.  Conduction  of  Electricity  through  Oases,  p.  610, 1606. 
16.   Abney  and  FesUng,  Phil.  Trans,  m.  II,  p.  423,  1S86, 

16,  Ttolle,  0.  R.  119.  p.  949,  1894. 

17.  Rosettl.  La  Lumlfire  Elec.  1,  p.  159, 1879. 
H.  Le  Cbateller,  a  R..  IH,  p.  737,  1862. 
Wilson  and  Gray.  Proc.  Roy.  Soc.  5B,  p.  24,  1896. 

>y  Google 

pHxeicAL  Chaoaotbristios  of  Lduikocs  SoimcES       87 

H.  Wanuer,  Ann.  der  PhyBlk,  2,  p^  141,  1900. 

O.  Lummer  and  B.  Pringabelm,  Verb.  d.  Deutsch.  Fhjt.  Q«s.  1,  pp 

23  and  215,  1396. 
P.  W.  Very,  Astrophys.  Jour,  10,  p.  208,  1899, 
P«i7,  C.  R.  ISi.  pp.  977  (Ud  1201.  1902. 

C.  Waiduer  and  G.  Burg«s8,  Bui,  Bur.  or  Stds.  1,  p.  109,  1904. 
M.  RosenmUller,  Ann.  d.  PbyB.  29,  p,  394,  1909. 

18.  Nakano,  Trana.  A.  I.  E.  E.  6,  p.  308,  1889. 
Marks,  Trans.  A.  I.  E.  E.  7,  p.  170,  1890. 
W.  Wedding,  Blek.  Zelt..  p.  717,  1897. 

W.  CzudnocbowBkl,  Daa  Elektrlsche  Bogenlfght,  p,  76,  1906. 
H.  Clifford,  Proc  Nat.  Elec.  Light  Aaaoc.  1,  p.  561.  1906. 
HerEog  u.  Petdmann,  Handbuch  der  ICiektrlacben  Belencbtnns,  pD 
117  and  144.  1907. 

19.  L.  Arona.  WIed.  Ann.  J7,  p.  767.  1892. 

P.  C.  Hewitt,  Trans.  A.  I.  B.  B.  J8,  p.  936.  1901. 
P.  C.  Hewilt.  Bl«.  (Lond.)  52.  p.  447.  1904. 
SO.   L.  ArcHis.  Wied.  Ann.  58,  p.  73,  1896. 

M.  T.  Reckllnghauaen,  Blek.  Zelt.  SS,  v.  1102,  1904. 
A.  P.  Wills,  Pbys.  Rev.  19,  p.  66.  1904. 
J.  Stark  et  al.,  Ann.  d.  Phya.  (i)  IS,  p.  213,  1905. 
J.  Pollak,  Ann.  d.  Pbys.  <.*)  IB.  p.  217,  1906. 
C.  Cblld.  Jabr.  d.  Radio,  u.  Elek.  S.  p.  189, 1906. 

21.  Ii.  Arona.  WIed.  Ann.  SS.  p.  73, 1896. 

A.  WUla,  Phya.  Rev.  19,  p.  66.  1904. 

22.  W.  Geer,  Elec.  World.  40.  p.  86,  1902. 

H.  Clifford,  Proc.  Nat.  Blec.  Light  Aaaoc.  1,  p.  573,  1906. 
H.  Boaa.  Blek.  Zelt  S7.  p.  867,  1906. 
K.  Stockhausen,  Blek.  Zelt.  27,  p.  868,  1906. 
L.  Bloch,  Belencbtungatecknlk,  p.  131,  1907. 

Herzog  u.  Feldinann,  Handbuch  der  Blektrlachen  Belencbtung,  p.  91, 

B.  Lewis.  111.  Bng.  (N.  Y.)  2.  p.  427.  1907. 
J.  Polak.  Blek.  Zelt.  28,  p.  656.  1907. 

0.  Vogel,  Zelt  t.  Beleucht  15,  p.  149, 1909. 

23.  O.  BuBsmann,  Blek.  Zelt.  3S,  p.  932,  1907. 

24.  H.  Lux,  111.  Bng.  (Lond.)  1.  p.  90,  190S. 

26.  Wlnkleman,  Handbuch  der  Physlk,  1.  p.  698,  1904. 

20.  D.  McP.  Moore.  Proc.  A.  I.  B.  B.,  p.  530,  1907. 

27.  K.  AugBtriini.  Ann.  der  Phya.  iS.  p.  493.  1893. 

B.  Drew,  Pbys.  Rev.  17.  p.  321,  1903. 

28.  B.  P.  Hyde  and  J.  Woodwell.  Trans.  III.  Bng.  Soc.  4,  p.  871, 1909. 

C.  Sharp  and  P.  Millar,  Trans.  111.  Bug.  Soc  i,  p.  886,  1909. 
W.  Wedding.  Elek.  Zelt  SI.  p.  691.  1910. 

29.  J.  Draper,  Phil.  Mag.  iS)  42.  p.  100,  1848. 

a.  KlTchhoff  u.  R.  Bunsen;  Pogg.  Ann.  110.  p.  160.  1860. 
P.  Hoppe-Seyler,  Pogg.  Ann.  m,  p.  101.  1S72. 
W.  Siemens,  Ann.  der  Phys.  18,  p.  311,  1883. 


88  iLLmnKATiMQ  Enginbebino 

A.  SmltbelU,  Fha  Hag.  S7,  p.  245,  1894. 

H.  Kayser,  Handbuch  der  Spectroscople,  t.  p.  137,  1902. 

B.  8.  LaC7.  Za.  f.  Pbya.  Chem.  64.  p.  G33.  1908. 

W.  H.  Pulweller,  Trana.  lU.  Eng.  Soc.  i,  p.  76,  1909. 

30.  H.  C.  Dlbbltta,  Pogg.  Ann.  182,  p.  497,  1864. 

A.  Crova,  C.  R.  S7.  p.  322,  1878. 

a.  P.  Langley,  Phil.  Mag.  SO.  p.  278,  1890. 

O.  Lununer  u.  B.  Prlngshelm,  Verb.  d.  Deutach.  Phya.  Oea.  i,  p. 

230,  1899;  and  S,  p.  36,  1901. 
H.  Lux,  111,  Bng.  (Lond.)  1,  p.  99,  1908. 

B.  Nlcbola  and  B.  Marrltt,  Plirs.  Rev.  iO,  p.  338.  1910. 

31.  Preece,  B.  A.  Rep.,  1888;  also  Nature,  98,  p.  496. 
S.  P.  I^ngley,  PhlL  Mag.  30.  p.  261,  1890. 

W.  Wedding,  Ueber  den.  Wlrkungsgrad,  etc.,  Mflnchen,  1906. 
E.  L.  Nichols,  lU.  Bng.  (N.  T.)  1,  p.  427,  1906. 
H,  Lux.  III.  Bng.  (Lond.)  I.  p.  99,  1908. 

C.  V.  Dryadale,  lU.  Bng.  {Lend.)  1,  p.  164,  1908. 

32.  W,  M.  Watta,  Pill.  Mag.  M,  p.  327,  1870. 
Le  Cbateller,  C.  R.  ISt.  p.  1144,  189G. 

A.  Smlthells,  Jour.  Chem.  Soc.  67,  p.  1050,  1S96. 
Mc  Crae,  Ann,  der  Phya.  55,  p.  97,  1S9&. 

E.  L.  NlcholB,  Phya.  Rev.  10.  p.  234,  1900, 
Q.  W.  Stewart,  Phye.  Rev.  15,  p.  306,  1902, 

F,  Kurlbaum.  Phya.  Zelt  S.  p.  187,  1902. 
R.  Ladenburg.  Pbys,  Zelt.  7,  p.  697,  1906. 
A.  Becker,  Ann.  der  Fbya.  2S,  p.  1017,  1909. 

33.  W.  Waggener,  Wled.  Ann.  5S.  p.  679,  18S6. 
F.  BerkenbuBch.  Wled,  Ann,  67,  p.  649.  1899. 
F6ry,  C.  R.  J37,  p.  909,  1903. 

F.  Haber  u.  F.  Rlchardt,  Zelt  f.  Anorg.  Chemie,  SB,  p.  60.  1904. 
H.  Schmidt,  Ann.  der  Phya.  20.  p.  3EE,  1S09. 
B  Bauer,  C.  R.  HB.  p.  90S,  1909. 

34.  Jour.  Gaa  Lt  Lond.,  p.  318,  18G0. 
36.   Jour.  Gaa  Lt.  Loud.,  p.  1002,  1887. 

36.  Jour.  Gaa  LL  Lond.,  p  22,  18S6. 

37.  Vienna  Pbarma.  Centiacbe,  2,  1886. 

38.  Beiblaetter,  19,  p.  423.  1S9S. 

39.  R.  Bunsen,  Lleblg'a  Ann.  d.  Chem.  u.  Pharm.  ISl.  p.  266, 1864. 

40.  J.  Babr,  Lleblg'a  Ann.  d.  Chem.  u.  Pbarm.  1S5,  p.  376,  1866. 

41.  J.  Bahr  and  R.  Bunaen,  Lleblg'a  Ann.  d.  Cbem.  u.  Pharm.  15"!,  p.  1, 


42.  W.  Hugglns,  Proc.  Roy.  Soc.  IB.  p.  646,  1870.    See  alao:  Phil.  Mag. 

(4)  iO,  p,  302,  1870. 

43.  L.  Haltinger,  Monata,  f.  Chemie,  12,  p.  362,  1891. 

44.  E,  L,  Nlchola  and  B.  W.  Snow,  Phil.  Mag.  SS.  p.  19,  1892. 
Ch.  St.  John,  Wled.  Ann.  5S.  p.  433.  1896. 

C.  Killing.  Jour.  (.  Gaabeleucb.,  p.  697,  1896. 

V.  B.  Lewes,  Jour.  Qaa  Lt  (Lond.).  p.  1104,  1896. 

>y  Google 

Physical  GHASAOTSBiaiios  of  Luuinoub  Soobobs        S9 

DKMBbach,  Jour.  f.  Qasbeleucb.  iO,  p.  174, 18ST;  H.  P.  3GS,  189S. 
H.  Bunte,  Ber.  Ctaem.  a«aell.  SI,  p.  6,  1S9T. 
MoBcbell,  Zeit.  f.  Beleuch.  11,  1897. 

H.  Le  Chatelter  et  O.  Boudouard,  C.  R.  Ite,  p.  1861,  1S98. 
BeL  tu  den  Ann.  der  FhyB.  X2,  p.  313,  1898. 
A,  A.  Swlotau,  Pfoc.  Roy.  Soc  6S.  p.  116,  1S99. 
W.  Nernst  and  K  Bose,  Pbys.  Zelt.  I,  p.  289,  1900. 
H.  Tbtote,  Ber.  Cbem.  OeaeU.  SS.  p.  183.  1900. 
H.  KfiTser,  SpectroBCDpl«,  S,  p.  161,  1603. 
C.  PSry,  C.  R.  IH.  p.  977,  1803. 
M.  Solomon,  Nature,  in,  p.  82,  1902. 
H,  Bunte,  Ber.  Int.  Cong.  d.  Cbemle,  Berlin,  May,  1903. 
St.  Clair  DevUle.  C.  R..  1903. 
H.  Rubens,  Pbys.  Zelt.  «,  p.  790.  1906. 
J.  Swinburne,  Blec  (Lond.)  57,  p.  744,  1906. 
H.  Kayier,  Bpectroscople,  p.  4GZ.  1906. 
Foiz,  C.  R.  m,  p.  686,  1907. 

R.  J.  Meyer  and  A.  AuBcbtttz,  Scl.  Abs.  10  A,  p.  53S,  1907. 
III.  Ene.   (Lend.)  1,  pp.  173  and  968,  190S. 
A.  Slmonlnl.  Trana.  IlL  Eng.  Soc.  ^,  p.  647,  1909. 
46.    H.  Le  Chatelier  et  O.  Boudouard,  C.  R.  126,  p.  1861,  1898. 

A.  White  and  A.  Travers,  Jour.  Soc.  Cbem.  Ind.  Si,  p.  1012,  1902. 

Holborn  u.  Kurlbaum,  Ann.  der  Pbys.  10.  p.  237,  1903. 

H.  Rubene,  Pbys.  Zelt  7.  p.  1S7,  1906. 

H.  Rubens,  Ann.  der  Pbys.  iO.  p.  673,  1906. 

H.  Lux,  Zeit.  f.  Beleucbt.  SS.  p.  375.  1909. 

46.  A.  Wbite  and  A.  Travera.  Jour.  Soc.  Cbem.  Ind.  tl,  p.  1012,  1902. 

47.  H.  Lux.  Zeit  t.  Beleucbt.  SS,  p.  375.  1906. 

48.  Le  Cbateller.  C.  R.  Itl,  p.  1144,  1895. 
T.  Lewes,  Cbem.  News,  11,  p.  181,  1896. 
Smltbells,  Jonr.  Cbem.  Soc.  67,  p.  1060,  1896. 
E.  L.  NlcbolB.  Pbys.  Her.  10.  p.  234,  1900. 

R.  lAdenburg.  Phys.  Zelt.  7,  p.  697,  1906. 

49.  E.  Nicbols.  Phya.  Rev.  11,  p.  216.  1900. 

K.  AngstriSm,  Astrophys.  Jour.  15,  p.  223,  1902.    See  also  Pbys.  Zelt. 

S.  p.  267,  1902. 
E.  NlcbolB  and  W.  Coblentz.  Pbys.  Rev.  n,  p.  267.  1903. 
O.  Stewart,  Pbys.  Rev.  16,  p.  126.  1903. 

60.  B.  Llebenthal.  PraktlBCbe  Photometrle,  p.  357,  1907. 
H.  Lux,  111.  Eng.  (Lond.)   1,  p.  99,  1908. 

61.  J.  Morebead,  Acet.  Jour.  11.  p.  261. 1910. 

%i.  Buaeman  und  Boehm,  Blek.  Zelt.  2^.  p.  2S1,  1903.     See  also  E.  De- 

Fodor,  Scl.  Abs.  g.  p.  713,  1899. 
63.   M.  C.  Beebe.  Scl.  Abe.  B,  8,  p.  3S8.  1906. 
Elec.  World.  i9.  p.  981,  1904. 

54.  H.  N.  Potter,  Proc.  Inter.  Elec.  Cong.,  St  Louis,  i.  p.  852.  1904. 

55.  A.  J.  Wurtz,  Trans.  A.  1.  E.  E.  IS,  p.  511.  1901. 
L.  Hartman,  Phys.  Rev.  22,  p.  353,  1906. 

>y  Google 

90  Illuhikathtq  £)i7QiiiEESiifQ 

H.  Lnz,  Zelt  t.  Beleuch.,  1907. 

O.  Lelmbach,  ZelL  f.  wlsa.  Pbot.  8,  p.  39G.  1910. 
GS.  R  HlrBCbaoer,  Elek.  Zelt  29,  p.  87.  1908. 
67.   W.  Nernat  and  W.  Wild,  Za.  t.  Blektrochem.  7,  p.  373. 1900. 

Herzog,  n.  Feldman,  Handbuch  d.  Elek.  BeleuclL,  p.  70,  1907. 
S8.  W.  Wedding,  Elek.  Zelt  22,  p.  631,  1901. 

Zelt  f.  Instr.  iS,  p.  178,  1S03. 

Blec.  World,  43,  p.  981. 1904. 

M.  C.  Beebe,  Elec.  Rev.  ie,  p.  6G7,  1905. 

J.  H«-zog  n.  C.  Feldmauii.  Handbuch  der  Elek.  Beleach..  p.  144,  1907. 
G9.   O.  Lnmmer  u.  E.  Prlngstieim,  Verb,  der  Deutacb.  Phys.  Gea.  I,  p. 
236.  1899. 

F.  Eurlbaum  und  O.  Scbulze,  Ber.  der  Deatscb.  Fhys.  Oe«.  1.  p.  488, 

L.  R.  Ingeraoll,  Phya.  Rev.  17,  p.  376,  1903. 

L.  Hartman,  Phys.  Rev.  tt,  p.  363,  190B. 

Mendenhall  and  Ingeraoll,  Phya.  Rev.  ti.  p.  230,  1M7;  tS,  p.  IS,  1907. 

W.  Coblentz,  Bui.  Bur.  of  8tde.  i,  p.  636,  1907. 

W.  Coblentz.  Bu).  Bur.  of  Stda.  5,  p.  183,  1908. 

60.  L.  R.  Ingeraoll,  Phys.  Rev.  J7,  p,  371,  1903. 
Drude,  Lehrbucb  der  Optik,  p.  474,  1906. 

W.  W.  CoblenU,  Bui.  Bur.  of  Stda.  i,  p.  653,  1907, 
W.  W.  CoblenU,  Bui.  Bur.  of  Stda.  5,  p.  184.  1908. 

61.  C.  DryBdale,  111.  Eng.  (Lond.)  1.  p.  643,  1908. 

«2.   8.  LAngley  and  F.  Very,  PbU.  Mag.  90.  p.  260.  1890. 

Broomall,  Scl.  Amer.  Nov.  5.  1898. 

H.  E.  Ivee  and  W.  Coblentz,  Trana.  111.  Eng.  Soc  4,  p.  657,  1909. 
63.   A.  Krug,  Astrophys.  Jour.  2S,  p.  800.  1908. 

M.  B.  Bauer,  C.  H.  ISO,  p.  1747,  1910. 
04.   B.  Nlchola  and  Franklin  Amer.  Jour,  of  Scl.  S6,  p.  100,  1889. 

F.  Gtaud,  C.  R.  1S9,  p.  7B9,  1899. 

Blaker,  Phya.  Rev.  13,  p.  346,  1901. 

P.  Valllant,  C.  H.  US.  p.  81,  1906. 

E.  NlcbolB,  Trana.  III.  Eng.  Soc.  3,  p.  322,  1908. 

H.  Kayaer,  Spectroacople,  3,  p.  427,  1905, 

B.  KOttgen,  Ann.  der  Pbya.  S3,  p.  801,  1894. 

Nernat  and  Boae,  Pbya.  Zelt  i,  p.  289,  1900. 

H.  B.  Ivea,  Bui.  Bur.  of  Stda.  6,  p.  234.  1909. 

O.  Stewart.  Pbya.  Rev.  IC.  p.  126,  1903. 

L.  Hartman,  Phya.  Rev.  17.  p.  66.  1903. 

B.  Nlcbola.  Phys.  Rev.  90,  p.  333,  1910. 

Kurlbaum  und  Scbuize,  Ber.  d.  Deutacb.  Pbya.  Gee,  J,  p.  428,  1903. 

8.  P.  lABgley,  PbIL  Mag.  29,  p.  52.  1890. 

H.  E.  Ivee,  Trana.  111.  Bng.  Soc.  S.  p.  208.  1910. 
6E.   E.  Nlchola.  Phya.  Rev.  f,  p.  260,  1894. 

B.  P.  Hyde,  Trana.  111.  Eng.  Soc  i,  p.  334,  1909. 

W.  CoblenU,  Bui.  of  Bur,  of  Stda.  5.  p.  360,  1908. 

W.  Coblentz,  Jour.  Frank.  Inat.  170,  p.  174,  1910. 

>y  Google 

Vhtsical  Ceabactebi8tic8  op  LuhiiiOds  Soukoes       91 

Le  CbateUo-  et  Boudonard,  C.  R.  116.  p.  ISGl,  189S. 

O.  L.iin]mOT  and  Q.  pFlngBhelm,  Yerh.  d.  DentBch.  PtiTS.  Oea.  1.  9. 

235.  1899. 
Rubens,  FhTB.  Zelt.  B,  p.  790,  1906. 
Rubens,  FLts.  Zfllt.  7,  p.  186.  1906. 
W.  Coblentz.  Bnl.  Bur.  of  Stds.  6.  p.  173,  1910. 
a.  Stewart,  Phya.  Rev.  16,  p.  125.  1903. 
W.  Coblentz.  BuL  Bur.  of  Stds.  i.  p.  533,  1907. 
W.  Coblentz,  Bui.  Bur.  of  Stds.  S,  p.  184,  1908. 
E.  Drew,  Pliys.  Rev.  17,  p.  321,  1903. 
.   W.  Yoege,  Jour.  t.  Gas  Beleuch.  48.  p.  513, 1905. 
H.  B.  Ives.  Trans.  111.  Eng.  Soc  S,  p.  208,  1910. 
D.  McF.  Hoore,  Trans.  111.  Eng.  Soc.  5,  p.  209,  1910. 

>y  Google 



By  Wniia  E.  WHnmsz 

Peculiar  poaltlon  of  the  elemoit  carbon  In  almost  all  lighting  Byatems. 

Carbon  heated  to  lamlneaceDce  Id  oil,  lUumlnatlng  gaa  and  acetylene 

Arc  llgbtiDg  and  IncandeBceot  llgbting:. 
Sabatltution  of  other  materials  tor  luminous  carbon  In  flames. 

DrnmmoDd  light. 

Welsbach  maotle. 
Carbon  arc  lighting — History  of. 

Blectrochemtstry  of  the  arc. 

Combustion  and  electrical  migration. 

Enclosed  arc  and  air  control. 

Direct  and  alternating  current  arcs. 

Arcs  of  other  material  than  carbon. 

Solids  heated  by  arc. 
Non-carbon  arcs. 

Iron,  magnetite,  titanium  carbide  arcs. 
Etfllclency  and  size  of  light  unit. 
The  mercury  arc. 

Its  ultra-Tlolet  light  and  production  of  ozone. 
Vacuum  tube  lighting. 
I^e  Incandescent  lamp. 
'  Carbon  filament. 

Chemistry  of  the  methods  of  manufacture. 
Forming,  baking,  firing,  coating  and  metalllEing. 

Osmium  filament. 

Tantalum  filament. 

Tungsten  filament 

Only  a  few  years  ago  anyone  studying  the  chemistry  proper  of 
the  sources  of  artificial  illumination  might  well  have  been  led  to 
conclude  that  he  could  confine  his  efforts  to  a  single  element,  i.  e., 
carbon.  This  was  owing  to  its  general  and  peculiar  applicability 
in  all  types  of  artificial  lighting,  no  matter  how  widely  they  differed 
in  their  methods  of  employment  of  this  interesting  element,    I  even 

>y  Google 

94  Illdminating  Engineebinu 

thiok  he  might  have  been  forgiven  for  aseuming  that  in  relation 
to  light  carbon  occupied  pome  such  particular  place  among  the  ele- 
ments, as  it  does  in  the  chemical  relations  of  life.  Carbon,  of  all 
the  elements,  is  tiie  basis  of  organic  chemistry  and  the  one  funda- 
mental element  without  which  organic  substance  and  life  itaelf  are 
impossible.  All  artificial  light  was  at  that  time  due  to  carbon 
heated  to  incandescence.  The  rfBciency  of  the  light  sources  de- 
pended on  the  efficiency  of  maintaining  carbon  at  a  high  tempera- 
ture. In  the  various  types  of  oil  lamps  which  were  in  use  several 
thousand  years  ago,  the  light  is  due  to  the  incandescence  of  carbon. 
This  carbon  is  a  product  of  decomposition  of  the  vapors  of  the  oil. 
It  can  easily  be  deposited  from  the  flame  and  be  kept  from  burning 
by  introducing  a  cooled  surface  into  the  flame.  This  service  of 
the  carbon  is  a  double  one  in  the  case  <Tf  oil  and  ordinary  gas 
illumination.  Here  an  element  is  needed  which  forms  readily 
vaporizable  compounds  or  gases,  and  compounds,  too,  which  are 
decomposed  by  the  moderate  heat  produced  by  the  reaction  of  the 
compound  with  the  air,  and,  finally,  the  element  must  itaetf  be  non- 
vaporizable  at  the  temperature  of  the  continuing  reaction.  In  these 
respects  carbon  is  apparently  the  only  element  which  posaesses  the 
needed  properties.  It  did  not  follow  of  necessity  that  this  same 
element  should  be  beet  suited  for  electric  arc  lights  and  for  incan- 
descent filaments,  and  yet  for  half  a  century  it  was  the  mainstay 
for  both  methods  of  illumination.  Possibly  it  is  this  apparent 
selective  fitness  of  carbon  among  the  77  elements  that  caused  post- 
ponement of  attempts  at  discovery  of  other  methods  of  illumination. 

In  an  address  of  this  kind  on  the  chemistry  of  luminous  sources 
(a  subject  selected  to  properly  fit  into  a  comprehensive  scheme 
covering  illuminating  engineering),  it  seems  best  to  spare  special 
emphasis  of  selectod  kinds  as  much  as  possible,  and  to  consider  in 
something  of  a  co-ordinating  way  the  chemistry  of  all  the  prac- 
tical methods  of  lighting. 

In  such  a  consideration  one  is  soon  impressed  with  the  fact  that 
the  several  different  types  of  illumination  differ  relatively  little  in 
their  net  efficiency.  The  labor  and  material  involved  in  the  pro- 
duction of  the  light  of  a  candle  does  not  seem  to  differ  much  by 
whatever  methods  one  employs  to  produce  the  li^t.  A  candle- 
power  from  a  modem  oil  lamp,  an  alcohol  lamp,  from  a  gas 
lamp,  or  from  an  electric  lamp  is,  speaking  quite  generally,  a  matter 
of  about  the  same  order  of  magnitude  of  cost     This  would  not 

>y  Google 

The  Chbmibtby  of  Lcminocs  Sodhcbs  95 

be  BO  remarkable  if  they  were  all  nearly  perfect  illuminants,  or  if 
they  were  all  of  very  high  degree  of  energy  efficiency — i.  e.,  if  they 
were  all  nearly  perfect — but  they  are  not. 

That  they  are  nearly  alike  in  cost  is  due  to  the  fact  that  they 
are  all  ao  far  removed  from  the  perfect  artificial  illnminant  that 
the  large  proportion  of  wasted  energy  practically  determines  the 
cost  The  kerosene  oil  lamp  uses  a  few  tenths  of  1  per  cent  of  the 
energy  of  the  combustion  of  the  oil  in  the  prodnction  of  visible 
light  waves.  The  temperature  at  which  the  carbon  is  heated  in  this 
flame  is  so  low  that  almost  any  other  way  of  heating  the  carbon 
will  give  more  light.  In  the  case  of  the  very  intense  acetylene 
flame  we  probably  see  the  effect  of  much  higher  temperature  of  the 
ehrbon  particles,  as  this  is  a  hotter  flame  than  that  produced  by 
common  gas.  It  is  known  that  the  luminous  radiation  rises  ex- 
ceedingly rapidly  with  rise  of  temperature  at  burning  tempera- 
tures, BO  that  the  carbon  does  not  have  to  be  heated  very  much 
hotter  in  order  to  give  off  a  very  much  greater  light.  Probably 
the  range  of  temperature  within  which  carbon  is  heated  in  the 
various  kinds  of  lamps,  excepting  the  arc  and  acetylene  flame,  lies 
below  ISOO"  C. 

When  ordinary  illuminating  gas  is  used,  the  maximum  light  is 
gained  by  a  selected  composition  of  the  gas  and  construction  of 
the  burner. 

This  is  almost  equal  to  saying  that  the  gas  is  so  mixed  with  the 
air  which  combines  with  it  that  none  of  the  carbon  produced  by 
decomposition  of  the  gas  is  allowed  to  escape  as  soot,  but  is,  on  the 
othef'  hand,  kept  heated  without  combustion  within  the  flame  as 
long  and  at  as  high  a  temperature  as  possible.  If  more  air  were 
introduced  into  the  flame,  less  light  would  be  produced,  but  a 
locally  higher  temperature.  This  is  due  to  the  increased  rapidity 
of  combustion  of  the  carbon.  This  fact  led  to  the  introduction  of 
other  materials  than  carbon  into  the  flame  to  be  heated  by  the 
burning  gas.  Naturally,  very  little  advance  was  made  along  this 
line  until  a  scheme  for  making  total  and  rapid  combustion  of  the 
gas  was  developed.  This  was  the  work  of  Bunsen,  who  found 
that  air  mixed  with  the  gas  in  suitable  proportions  brought  about 
the  effect  of  raising  the  temperature  of  the  gas  flame.  In  this  appli- 
cation the  carbon  is  immediately  consumed  and  does  not  lend  any 
luminosity  to  the  flame.  The  industry  waited  at  least  a  decade 
for  some  suitable  substitute  for  the  luminous  carbon.    It  was  the 

>y  Google 

,96  Illdmikatinq  Enqineerinq 

ezhaoBtive  work  of  Dr.  Aner  von  Welsbach  which  produced  the 
mantles  of  metallic  oxides  which  we  know  to-day.  These,  when 
heated  to  the  high  temperature  produced  by  the  combustion  of 
mixed  air  and  gas,  give  a  much  greater  light  for  a  given  rate  of 
gas  supply  than  the  previous  method  of  use  of  the  same  gas.  This 
increased  light  efficiency  is  also  greatly  angmented  by  the  proper 
selection  of  the  components  of  the  mantle  mixture.  It  would,  at 
first  thought,  seem  probable  that  any  white  mantle  capable  of  with- 
standing the  high  temperature  of  the  flame  would  give  the  same 
definite,  constant  quantity  of  light  under  the  same  conditions 
of  heating  gas  flame.  That  this  is  not  so  is  readily  shown  by  a 
study  of  the  efficiency  of  various  oxide  mixtures  when  used  as 
mantle  compounds.  There  are  a  number  of  metallic  oxides  which 
do  not  melt  or  vaporize  at  the  temperature  of  the  flame,  but  the 
most  refractory  is  not  the  most  satisfactory.  Each  mixture  of 
oxides  seems  to  have  its  own  characteristic  light-giving  power,  and 
to  possess  also  some  considerable  selective  power  in  producing  color 

Thia  has  led  to  an  immense  quantity  of  purely  experimental 
research,  in  order  to  discover  what  particular  compound  or  mixture 
would  give  the  most  efficient  and  satisfactory  light.  As  an  illustra- 
tion of  this  fact,  it  is  worth  noting  that  Welsbach  discovered  that 
pure  thorium  oxide,  when  used  in  a  mantle,  will  not  give  a  tenth 
of  the  light  that  will  be  produced  under  the  same  conditions  by  a 
mantle  made  of  a  mixture  of  99  parts  of  thorium  oxide  and  1  part 
of  cerium  oxide.  This  very  interesting  phenomenon  will  doubtless 
be  taken  up  by  Mr.  Whitaker,  and  is  therefore  only  referred  to 
at  this  point.  An  instructive  article  on  this  subject  was  published 
in  the  April,  1909,  number  of  the  Journal  of  Industrial  and  Engi- 
neering Chemistry.  It  is  the  one  discovery  which  has  apparently 
given  the  illuminating-gas  industry  the  help  it  needed  to  keep  in 
competition  with  methods  of  electric  lighting. 

Just  as  no  story  of  incandescent  electric  lighting  can  be  properly 
started  without  at  least  a  reference  to  the  enormous  contn,butioQ 
of  Edison,  so  also  any  history  of  arc  lighting  properly  commences 
with  Sir  Humphry  Davy.  In  1809  he  waa  experimenting  with 
phenomena  produced  by  a  battery  of  2000  primary  celis,  and  pub- 
licly showed  that  a  very  luminous  arc  was  produced  when  the  cur- 
rent passed  across  the  gas  between  carbon  pointa.  While  he  may 
not  have  been  the  discoverer  of  the  are,  he  was  one  of  the  first  to 

>y  Google 

The  Chbuistrt  of  Luuinods  Socucbs  97 

Bee  a  use  for  it.  For  a  great  many  years  thereafter  no  practical 
application  was  made  of  this  discovery,  because  there  had  not  beai 
developed  any  satisfactory  devices  for  generating  the  large  amount 
of  electrical  energy  consumed  by  even  a  small  carbon  arc.  In  1870 
the  Gramme  generator  was  devised.  Carbon  arc  lamps  were  oper- 
ated from  this  machine,  in  place  of  batteries.  Some  of  the  first 
attempts  at  practical  use  of  these  machines  and  lamps  were  made 
in  connection  with  light-houses  on  the  English  and  French  coasts. 
Soon  thereafter  the  Jablochkoff  electric  candle  came  into  use.  This 
is  an  arc  lamp  with  parallel  carbons.  These  were  kept  separated  by 
a  thin  wall  of  clay,  or  a  mixture  of  sand  and  glass,  which  gradually 
vaporized  during  the  burning  of  the  arc.  At  one  time  several 
thousand  of  these  were  in  use  in  Europe.  At  the  Paris  Exhibition, 
in  1878,  the  iUumination  produced  by  these  candles,  operated  by 
Gramme  machines,  marked  an  epoch  in  lighting  which  the  previous 
30  years  of  laboratory  experiment  with  arcs  had  but  dimly  fore- 

Somewhat  later  the  simple  carbon  arc  was  commercially  realized, 
and  the  clay  part  of  Jablochkoff  candles  disappeared  from  the  elec- 
tric lamp  for  a  time. 

The  phenomenon  of  this  direct-current  carbon  arc  is  still  quite 
far  from  being  perfectly  understood.  From  the  chemical  stand- 
point, the  arc  presents  two  pure  carbon  pencils,  each  of  which  is 
slowly  consumed.  In  the  ordinary  lamp  the  consumption  of  the 
positive,  which  is  usually  the  upper  electrode,  is  much  more  rapid 
than  that  of  the  lower  or  negative  electrode.  It  was  long  evident 
that  the  wasting  away  of  the  carbon  electrodes  was  largely  due  to 
simple  combustion  by  the  air,  and  many  attempts  were  made  to 
prevent  this  combustion,  while  retaining  the  characteristics  of  the 
carbon  arc.  This  led  to  the  discovery  that  the  upper  electrode  is 
heated  much  hotter  than  the  lower  during  the  passage  of  the  cur- 
rent, that  carbon  actually  distills  from  this  positive  electrode,  and 
when  this  carbon  cannot  bum  it  will  deposit  upon  the  cooler  parts 
of  the  electrode.  This  property  of  building  out  mushroom  growths 
on  the  electrodes  when  operated  in  vacuo  or  in  inert  gases  seemed 
to  stand  in  the  way  of  economizing  in  such  a  lamp  by  practically 
separating  the  ordinary  combustion  of  the  electrodes  from  the 
proper  electric-arc  phenomena.  It  was  finally  found,  however,  that 
by  properly  controlling  the  current  and  voltage,  and  by  admitting 
only  a  very  small  quantity  of  air  to  the  globe  of  a  carbon  arc  lamp, 

>y  Google 

98  Illuminatino  Enqinebbino 

the  combustion  of  the  electrodes  could  be  greatly  reduced.  This 
air  rate,  which  is  controlled  by  the  openings  in  the  enpports  of 
the  inner  globe  of  the  enclosed  arc  lamp,  so  greatly  reduces  the 
burning  of  the  electrodes  that  the  life  is  increased  ten-fold  or  more. 
This  gives  us,  then,  the  two  primary  types  of  carbon  arc  lamps, 
the  open  and  enclosed.  In  the  closed,  as  in  the  open,  it  is  the 
positive  electrode  which  wastes  or  bums  away  the  more  rapidly 
of  the  two;  it  is  the  hotter  and  is  the  source  of  most  of  the  light 
from  the  arc.  In  the  pure  carbon  arc  only  a  few  per  cent  of  the 
light  is  due  to  the  flame  or  arc  proper.  This  arc  stream  is  far 
from  dense,  and  most  of  the  carbon  in  t^e  space  is  already  present 
as  carbon  monoxide. 

While  it  is  out  of  place  here  to  go  very  deeply  into  the  con- 
ceptions of  theories  which  have  been  formed  to  cover  the  action  in 
the  arc,  it  may  not  be  amiss  to  point  out  that  the  simplest  ideas 
are  not  applicable.  For  example,  it  is  quite  apparent  that  a 
motion  of  positively  charged  particles  across  the  gap  of  the  arc 
does  not  account  for  all  the  phenomena.  As  will  be  seen  more 
clearly  later,  the  negative  electrode,  at  least  in  most  cases,  is  the 
one  which  determines  the  character  of  the  arc,  and  a  carbon  arc 
is  still  a  carbon  arc  when  the  positive  electrode  ie  some  other  con- 
ducting substance,  while  it  is  usually  no  longer  a  characteristic 
carbon  arc  when  the  negative  electrode  ie  another  substance.  There 
is  no  simple  quantitative  relation  known  between  the  current  car- 
ried in  an  are  and  the  waste  or  loss  at  either  electrode.  In  this 
respect  the  arc  diffei-s  from  the  passage  of  current  through  a  gap 
within  a  soliition,  for  example.  Attempts  made  to  determine  the 
minimum  loss  of  electrode  for  a  given  arc  current  have  only  led 
thus  far  to  the  conclusion  that  if  any  quantitative  consumption  of 
electrode  takes  place  of  necessity  when  an  arc  is  passing,  the  quan- 
tity of  material  corresponding  to  a  given  current  is  at  least  a  thou- 
sand times  smaller  than  migrates  when  equal  current  passes  throng 
a  solution  or  an  electrolyte.  Moreover,  it  seems  that  this  motion 
within  the  arc  is  usually,  if  not  always,  made  up  of  material  from 
the  negative  electrode.  This  general  subject  has  led  to  a  great 
deal  of  quantitative  work  in  which  arc  electrodes  of  other  mate- 
rials than  carbon  have  been  used.  In  most  cases,  as  with  cartran, 
tiie  results  are  affected  by  the  simultaneous  oxidation  of  the  elec- 
trodes. Copper  and  iron  electrodes,  when  used  as  arc  terminals, 
dow  such  irregularities  that  it  has  been  impossible  to  accurately 

>y  Google 

The  Chehistrt  of  Ldminods  Sources  99 

determine  values  of  loss  at  cathode  or  anode  which  might  corre- 
spond in  some  waj  to  the  Faradaj  equivalents  in  electrolysis.  Even 
when  such  arcs  are  operated  in  inert  atmosphere  or  under  water, 
one  usually  finds  that  the  material  of  either  electrode  has  passed 
in  some  irregular  degree  to  the  other  electrode  and  deposited  upon 
it.  Such  effects  may  be  largely  accredited  to  simple  distillation. 
Some  cases  have,  however,  been  found  in  which  the  processes  of 
combustion  may  be  fairlj  well  separated  from  those  of  current 
action,  and  here  again  it  seems  proven  that  in  an  arc  it  is  essential 
that  material  pass  from  the  cathode  into  the  arc  space  only,  and 
that  a  consumption  of  the  anode  or  positive  electrode  is  always  an 
accidental  accompanying  effect.     This  will  be  referred  to  later. 

We  have  thus  far  considered  only  the  chemistry  of  the  pure 
carbon  arc.  Modification  of  this  arc  of  interest  to  illuminating 
engineers  have  been  many.  It  seems  necessary  to  refer  briefiy  to 
a  few  of  them  before  considering  other  arcs.  The  direct-current 
carbon  are  owes  its  efficiency  to  the  highly  heated  crater  or  are 
terminal  on  the  positive  carbon.  When  an  alternating-current 
carbon  lamp  was  measured,  it  was  found  that  not  quite  so  great 
efficiency  was  possible,  though  by  the  alternating  position  of  the 
crater  with  each  change  in  current  direction,  the  distribution  of 
the  light  is  somewhat  improved. 

Many  inventors  have  attempted  to  increase  the  light  from  a 
given  are  energy  by  introduction  of  suitable  chemical  compounds 
into  the  arc.  Some  of  these  have  led  to  successful  commercial 
lamps.  If  a  small  piece  of  a  very  refractory  material,  such  as 
zirconia,  be  brought  into  the  carbon  arc,  it  is  heated  to  a.  tempera- 
ture at  which  it  is  very  luminous.  This  is  quite  like  the  use  of  a 
rod  of  lime  in  the  Drummond  gas  lamp.  The  difficulties  in  the 
way  of  stability,  of  mechanism,  ignition  and  control,  may  account 
for  the  failure  to  develop  this  device  in  its  simplest  form. 

A  small  zirconia  rod  placed  between  the  two  carbon  electrodes 
(when  arranged  as  ordinarily,  one*  above  the  other),  although 
patented  as  an  arc  lamp,  has  not  been  commereially  developed.  A 
modification  of  this  scheme,  whereby  a  fecial  form  of  Welsbach 
mantle  is  placed  about  the  carbon  arc  to  be  heated  by  the  are,  has 
also  not  advanced  very  far.  A  considerable  difficulty  in  such  schemes 
lies  in  the  fact  that  the  hot  path  of  the  are  stream  is  usually  of 
very  small  cross-section,  and  in  lamps  of  moderate  energy  consump- 
tion ia  not  easily  confined  to  a  limited  position,  so  that  it  is  not 

>y  Google 

100  Illuminating  Enoineebiko 

easy  to  keep  interposed  material  heated  to  iocaiideBceiic^  by  this 

Countlees  schemes  for  contmuously  mtrodueing  powders  or  va- 
pors into  the  arc  have  also  been  tried.  It  was  found  many  years 
ago  that  the  addition  of  such  salts  as  carbonate  of  soda  to  carbon- 
arc  electrodes  gave  added  luminosity  to  the  arc,  reduced  the  volt- 
age across  the  arc  and  also  permitted  the  arc  to  be  lengthened  with- 
out extinguishing  it.  Very  smail  quantities  of  such  salts  are 
effective.  This  general  knowledge  did  not  produce  the  modem 
flame  arcs  at  once,  as  the  effect  of  such  salts  as  were  used  a  quarter 
of  a  century  ago  was  probably  not  greatly  marked  or  appreciated. 
About  10  years  ago  inventions  involving  this  principle  became  quite 
common.  Perhaps  best  known  among  them  are  those  of  Blonde! 
in  Prance  and  Bremer  in  Germany.  They  and  othera  made  use 
of  very  considerable  proportions  of  salt^  added  to  the  carbon  during 
tiie  manufacture  of  the  electrode.  Usually  10  per  cwt  or  more  of 
mineral  substance  was  added,  and  many  different  salts  were  pro- 
posed. Most  successful  seem  to  be  the  fluorides  and  chlorides  of 
calcium  and  magnesium.  Some  inventors  found  they  were  able 
to  construct  an  operative  electrode  by  using  a  homogeneons  rod  of 
carbon  and  the  s^ta.  Others  preferred  to  confine  the  salt  to  a  core 
inside  one  or  both  electrodes.  In  most  cases  this  core  also  contained 
some  special  form  of  carbon,  and  in  some  cases  there  were  two 
concentric  cylinders  of  various  composition  about  the  central  core. 
It  has  been  quite  common  to  iise  carbon  electrodes  with  a  core  of 
soft  carbon,  as  the  arc  by  this  means  is  kept  centered  on  the  elec- 
trode. The  present  so-called  carbon  flame  arcs,  which  are  usually 
characterized  by  great  luminosity,  with  predominance  of  reddish- 
yellow  color,  are  made  in  the  above  way.  The  electrodes  usually 
contain  so  much  mineral  matter  that  they  cannot  be  used  in  en- 
•closed  lamps  of  the  ordinary  types.  The  mineral  matter,  after 
passing  into  the  arc,  must  be  carried  from  the  lamp  by  a  good 
draught,  otherwise  it  will  deposit  on  the  globe  and  soon  greatly 
reduce  the  luminosity  of  the  lamp.  The  necessary  draught  involves 
also  the  rapid  consumption  of  the  electrodes,  so  that  such  lamps 
usually  have  to  be  trimmed  or  supplied  with  new  electrodes  daily. 
The  presence  of  the  salts  insures  low  voltage  for  the  lamp,  so  that 
they  are  usually  burned  two  in  series  oh  the  110-volt  circuit. 

The  most  useful  future  application  of  chemistTy  to  this  type  of 
flame  arc  lamp  will  doubtless  be  along  the  lines  of  producing  as 

>y  Google 

Tttw  Chemistrt  of  LuMiMons  SouacES  101 

great  an  efficiency  in  white  light  as  is  now  produced  in  the  reddish 
■tint  Taken  aa  an  dectric-Iight  source  alone,  these  reddiBh-flame 
arcs  are  the  most  efficient  of  any  of  the  commercial  lampe.  I  attach 
a  table  of  efficiencies  of  various  kinds  of  electric  lamps  for  com- 
parison. Such  a  table,  taken  alone,  may  be  very  misleading.  No 
indication  of  color,  convenience,  size  of  unit,  and  other  practical 
considerations,  appear  in  such  a  table. 


Carbon  (open  arc)  ... 

...  D.C. 

10     A. 

43  V. 

1.43  (spherical) 

(encloaed)   .. 

...  D.C. 

5     A. 

80  V. 


(enclosed)   .. 

...   A.C. 

7.5  A. 

80  V. 


Carbon  flame  arc  . . . 

...   D.C. 

10     A. 



MagneUte  arc  

...   D.C. 

4     A. 

80  V. 



...   D.C. 


110  V. 

1.7    (horlM»ntal) 

Hetalllsed  carbon   . . . 


110  V. 




110  V. 



...   D.C. 



(prewure)  .. 

...  D.C. 


lloore  tubes  

...  A.C. 



...   Both 





...      •■ 



It  is  particularly  in  the  arcs  tliat  the  chemical  nature  of  the 
electrodes  plays  a  determining  part.  When  a  simple  carbon  arc 
is  considered,  the  quality  of  the  carbon  is  of  the  greatest  im- 
portance. Pure  graphite  is  not  acceptable,  but  a  hard,  dense  carbon, 
quite  low  in  ash  and  of  very  fine  physical  structure,  is  most  eat^ 
factory.  For  many  years  these  were  imported  from  Germany,  and 
th^  still  are  to  some  extent. 

In  the  introduction  of  new  substances  to  the  carbon  arc  there 
are  many  chemical  and  physical  properties  which  unite  to  determine 
the  value  of  the  added  substance.  The  salts  of  many  elements  add 
more  or  less  intense  colors  to  the  arc,  in  accord  with  the  spectrum 
lines  of  the  particular  element.  This  effect  is  greatly  influenced  by 
the  degree  of  volatili^  of  the  salt  and  by  the  nature  of  the  other 
elements  or  compounds  vaporizing  at  the  same  time.  Calcium 
oxide  doefl  not  greatly  affect  the  luminosity  of  the  carbon-arc  stream, 
while  calcium  fluoride  does. 

During  the  past  10  yeare  some  advances  have  been  made  in  the 
practical  use  of  other  arcs  than  carbon.  The  best  known  are  the 
magnetite  and  the  mercury  arcs. 

>y  Google 


The  magnetite  differs  chemically  from  the  carboD  in  being  much 
less  combustible,  as  it  bums  only  in  changing  from  FCjO,  to  FcjO,, 
in  giving  non-volatile  oxides  and  in  giving  to  the  arc  flame,  to  a 
high  degree  of  intensity,  the  characteristic  colors  of  the  iron  &pec- 
trum.  The  iron  spectrum  is  one  of  those  metal  spectra  which, 
while  made  up  of  defined  lines,  contain  such  a  great  number  of 
them  (over  2000  have  been  mapped)  that  the  effect  is  practically 
that  of  a  continuous  spectruin.  In  the  magnetite  arc  practically 
all  of  the  light  is  due  to  the  arc  or  flame.  The  luminous  positive 
of  the  carbon  arc  i&  in  this  lamp  replaced  by  a  large  block  of  copper 
or  other  metal,  which  does  not  contribute  to  the  consumption  in  the 
arc,  so  that  this  lamp  is  an  arc  lamp  with  only  a  single  consuming 
electrode.  The  quality  of  the  arc  is  greatly  influenced  by  tiie 
quality  of  the  magnetite  electrode.  It  might  seem  probable  at  first 
that  iron  itself  would  be  preferable  to  magnetite,  but  long  series  of 
experiments  seemed  to  show  that  a  compound  and  rather  complex 
mixture,  containing  large  proportions  of  pure  magnetite,  gave  the 
best  results.  Such  arcs  must  burn  Btea,dily  and  the  electrode  must 
contain  a  small  amount  of  relatively  volatile  matter,  such  as  the 
common  salts  of  potash  or  soda.  For  a  given  current  the  rate  of 
waste  of  the  electrode  can  be  very  materially  altered  by  the  addi- 
tion of  otherwise  inactive  materials,  such  as  alumina  and  chromium 
oxide,  without  any  considerable  reduction  in  the  light  produced. 
This  effect  is  probably  due  to  the  reduction  of  vapor  pressure  of 
the  iron  oxide  in  the  molten  top  of  the  electrode.  This  corre- 
sponds  to  vapor-pressure  reduction  in  case  of  simple  solutions. 
Finally,  it  was  found  that  the  intensity  of  the  arc  is  greatly  in- 
creased by  the  addition  of  another  element  which  has  its  own  rich 
spectrum,  such  as  titanium.  So  that  the  magnetite  arc  is  really 
the  arc  spectra  of  iron  and  titanium  superposed.  Such  strictly  arc 
flames  have  one  advantage  over  carbon  arcs,  in  that  they  can  operate 
economically  in  small  units.  The  eflSciency  of  the  carbon  arc  is 
greater  the  larger  the  unit  within  a  wide  range,  but  units  below 
500  watts  begin  to  be  relatively  inefficient.  On  the  other  hand,  the 
efficiency  of  the  strictly  luminous  arcs  is  maintained  high  as  low  as 
250  or  300  watts.  This,  to  the  illuminating  engineer,  means  that 
he  has  greater  elasticity  in  the  distribution  of  his  lighting  energy. 

The  mercury  arc  may  be  said  to  differ  but  little  from  the  other 
arcs.  It  is  greatly  lengthened  by  being  conflned  to  a  glass  tnbe, 
and  thus  any  combustion  or  loss  of  material  is  obviated.    Its  color 

>y  Google 

The  Chemistry  op  LoMiNone  SouRcea  103 

and  light  are  determined  b8  in  the  case  of  other  arcs,  by  the  nature 
of  its  cathode  electrode.  The  anode,  ae  in  the  other  arcs,  may  be 
made  of  almost  any  conducting  material.  The  vapors  which  arc 
produced  at  the  cathode  condense  to  liquid  state  and  return  by 
gravity  to  the  cathode.  If  the  chemical  elements  had  more  fluid 
members  among  those  of  hi^ly  luminous  spectra,  the  principle  of 
the  enclosed  mercury  lamp  would  probably  quickly  yield  more  new 
and  useful  lighting  methods.  The  light  of  the  mercury  lamp,  when 
broken  down  by  the  pmm,  is  seen  to  be  composed  of  only  a  few 
widely  separated  lines.  Among  them  is  no  red.  For  this  reason 
red  articles  appear  black  under  this  light,  and,  for  this  reason, 
many  other  colors  fail  to  appear  natural  under  the  mercury  arc. 

There  are  two  interesting  facte  concerning  the  mercury  arc  which 
may  well  ultimately  be  utilized  in  a  practical  manner.  The  arc 
ie  very  rich  in  ultra-violet  light.  This  is  not  particularly  noticeable 
when  tiie  arc  is  surrounded  by  glass,  but  when  pure  quartz  is  sub- 
stituted for  the  glass  the  ultra-violet  light  pe&etrates  into  the  sur- 
rounding air.  This  produces  ozone  in  a  very  marked  manner,  and 
this  unaltered  light  has  a  very  eerious  and  injurious  effect  on  the 
eyes.  It  is  highly  probable  that  this  modified  mercury  lamp  is  to 
be  the  most  readily  applicable  form  of  ultra-violet  light  for  thera- 
peutic purposes.  Secondly,  it  has  been  discovered  that  when  the 
arc  is  operated  under  two  or  three  atmospheres  of  mercury  pressure 
the  efficiency  is  high  and  the  color  more  nearly  approaches  day- 
light. Glass  tubes  will  not  withstand  the  temperature  of  the  arc 
at  this  pressure,  but  quartz  will.  Such  quartz  mercury  lamps  are 
being  made  and  sold  abroad  at  the  present  time. 

Any  considerable  practical  improvement  in  the  color  of  the  mer- 
cury arc  has  not  been  made  by  the  amalgamation  of  other  elements 
with  the  mercury.  An  element  like  copper  or  iron  fails  to  vaporize 
from  the  cathode  of  the  mercury  arc.  Some  of  the  alkali  metals 
Bomewhat  alter  the  light,  but  most  of  them  also  attack  the  glass 
of  the  lamp.  It  is  worthy  of  note  that  some  fluorescent  dies,  rhoda- 
mine,  for  example,  are  capable  of  absorbing  the  green  and  blue 
spectral  lines  and  returning  in  their  place  some  considerable  red, 
but  this  has  not  proven  an  efficient  process. 

The  luminosity  of  gases  and  vapors  has  always  seemed  a  very 
promising  field  of  artificial  illumination.  In  the  case  of  heated 
aolids,  the  laws  of  radiation,  convection  and  conduction  are  well 
enough  known,  so  that  a  field  in  which  less  is  known  is  apt  to  seem 

>y  Google 


promising.  The  Oeissler  or  Pliicker  tubes,  in  which  attenuated 
gases  are  rendered  luminona  by  relatively  high  voltage  and  low- 
current  discharge,  are  well  known  to  all.  It  seems  very  probable 
that  future  developments  of  importance  will  be  made,  and  already, 
in  the  McFarlane-Moore  System,  very  considerable  advances  have 
been  made.  Here  the  chemical  composition  of  the  gasea  and  their 
pressnre  are  the  detenniniog  factors  of  the  color  and  Efficiency. 
A  peculiar  phenomenon  in  these  tamps  is  the  apparent  conaumptioD 
of  the  gas  or  air  in  the  tubes.  Gradually,  in  such  apparatus,  the 
gas  disappears,  as  though  driven  into  or  combined  with  the  glass. 
For  this  reason  the  inventor  of  this  system  has  devised  an  automatic 
inlet  valve  which  operates  to  let  gas  into  the  lamp  when  the  vacuum 
rises  to  a  certain  degree.  This  seems  to  be  a  similar  effect  to  the 
well-known  "  hardening  "  of  X-ray  bulbs  from  continued  use,  which 
is  an  improvement  in  vacuum,  and  is  also  noted  in  the  case  of  the 
vacuum  of  an  ordinary  incandescent  lamp. 

Without  wishing  to  go  deeply  into  the  history  of  the  incandescent 
lamp,  it  is  necessary  to  point  a  moment  to  the  work  of  Mr.  Edison. 
The  fact  that  electric  current  flowing  through  a  conductor  could 
heat  it  to  incandescence  had  long  been  known.  That  carbon  in 
filament  form,  when  preserved  from  combustion  by  a  vacuum,  would 
make  a  lamp  was  clear.  J.  W.  Starr  bad  patented  such  a  lamp  in 
1845,  and  Swan,  in  England,  had  exhibited  one  in  1879.  But 
between  this  point  and  a  satisfactory  incandescent  lamp  was  a  great 
gulf,  which  needed  the  untiring  energies  of  such  an  inventor  as 
Mr.  Edison  to  help  bridge.  A  piece  of  carbonized  thread,  confined 
in  such  a  vacuum  as  was  known  when  he  undertook  the  work,  did 
not  constitute  a  practical  lamp  at  all.  In  the  poor  vacuum  produced 
by  methods  used  in  those  days,  even  a  good  filament  of  the  present 
time  would  have  produced  but  a  very  imperfect  lamp.  The  simpler 
methods  of  producing  carbon  filaments  are  capable  of  yielding  only 
very  imperfect  lamp  filaments.  There  are  few  artificial  products 
which  excel  the  filament  in  the  divergence  between  apparent  sim- 
plicity and  actual  complexity. 

The  choice  of  elements  for  incandescent-lamp  filaments  may  be 
said  to  be  more  nearly  a  physical  than  a  chemical  problem,  but  in 
the  manufacture  of  all  of  them  chemistry  plays  a  dominant  r6le. 
The  best  carbon  filaments  now  in  use  may  be  described  as  con- 
sisting of  a  core  of  pure  carbon,  not  graphite,  covered  with  a  coat 
or  shell  of  pure  graphite,  which  has  been  so  changed  by  an  electric- 

>y  Google 

The  Chbhistby  of  Luminods  Sotihobs  lOfi 

luruace  treatment,  under  atmoepheric  pressure,  that  it  has  a  posi- 
tive-reBiBtance  temperature  coefficient  instead  of  a  negative  one. 
This  graphite  coating,  to  which  the  name  metallized  graphite  haa 
been  given,  has  the  appearance  of  having  been  melted  or  Bintered 
iogether,  and  thus  dijfen  from  all  other  graphite. 

The  chemical  and  physical  processee  bj  which  these  carbon  fila- 
ments are  produced  are  as  follows: 

High-grade  cotton  is  dissolved  in  a  strong  solution  of  sine  chloride, 
which  is  then  squirted  through  a  small  hole  into  dilute  alcohol.  The 
alcohol  coagulates  the  viscouB  solution  of  cellulose  so  that  a  trans- 
psrent  thread  is  the  product,  and  by  washing  this  in  running  water 
the  zinc  chloride  is  removed. 

Another  equally  satisfactory  method  for  reaching  the  same  end 
is  to  squirt  a  thick  solutioa  of  nitro-cellulose,  dissolved  in  acetic 
acid,  into  a  container  holding  water.  Washing  with  ammonia 
sulphide  and  water  changes  the  nitro-cellulose  into  non-explosive 
hydro-cellulose.  This  product  is  then  dried  in  the  air  while  stretched 
on  drums.  It  is  then  cut  to  desired  lengths,  formed  into  the  nec- 
essary loops  on  brass  frames,  and  finally  packed  in  graphite  boxes 
in  a  packing  material  such  as  baked  peat,  and  very  gradually  heated 
nntil  carbonization  takes  place.  In  this  process  the  carbonized 
filaments  are  heated  to  as  high  a  temperature  as  can  be  obtained 
by  gas  or  oil-heated  muffles. 

The  product  at  this  point  is  dense,  hard  carbon,  which,  even 
under  the  microscope,  is  far  from  having  the  appearance  of  charcoal, 
and  seems  almoat  free  of  pores.  The  carbon  filament  in  this  form 
would  make  a  very  inferior  lamp.  The  color  or  quality  of  its  sur- 
face, and  probably  the  volatility  of  its  material,  is  not  nearly  so  fa- 
vorable to  lamp  making  as  the  corresponding  properties  of  graphite. 
At  any  definite  operating  energy  the  amount  of  light  produced 
by  a  gray-graphite  surface  is  greater  than  that  produced  by  a 
black-carbon  surface,  so  that  the  carbon  filaments  are  graphite- 
<wat«d.  This  is  done  by  heating  them  by  the  current  in  an  atmos- 
phere of  hydro-carbon,  such  as  benzine,  at  low  pressure.  The  quality 
and  thickness  of  the  coat  may  be  controlled  by  the  duration  and 
temperature  of  the  treatment.  Until  a  few  years  ago  the  gt^ater 
part  of  all  carbon  filaments  were  made  in  this  way.  It  was  then 
found  thfft  the  effect  of  subjecting  the  graphite-coated  filaments  to 
temperatures  above  3000°  C  for  a  few  minutee  changed  the  graphite 
very  materially  in  its  properties.    Those  which  are  of  interest  to 

>y  Google 

106  Illuhinatino  Enqinbehino 

nB  now  are  the  resistance,  its  temperature  coefficient  and  the  sta- 
bility at  operating  lamp  i«mperatureB.  Briefly,  the  resistance  of 
the  graphite  coat  is  reduced  to  about  30  per  cent  of  its  original  re- 
Bietance.  Its  temperature  coefficient  is  reversed  and  its  lasting 
powers  in  the  lamp  increased  nearly  three-fold. 

This  point  Beems  a  proper  one  at  which  to  mention  the  standard 
of  use  for  incandescent  lamps  as  determined  by  practical  conditions. 
Burning  at  a  low  efBciency,  an  incandescent  lamp  has  practically 
an  indefinite  life.  At  3  watts  per  candle-power  it  may  have  1300 
hours*  life  and  at  2.5  about  500  hours  to  80  per  cent  of  its  original 
candle-power.  It  has  been  foand  by  use  that  about  500  hours'  life 
for  a  carbon  lamp  is  most  practical,  this  500  hours  being  the  length 
of  time  the  lamp  remains  above  80  per  cent  of  its  starting  candle- 
power.  The  metallized  filament  lamps,  therefore,  instead  of  being 
burned  at  the  former  efficiency  of  3.1  watts  per  candle,  are  made 
to  burn  at  about  2i  w.  p.  c,  at  which  they  have  about  500  hours' 
life.  Evidently  the  higher  the  cost  of  the  lamp  the  more  stress  has 
to  be  laid  upon  long  life,  while  with  very  cheap  lamps  there  is  an 
advantage  gained  by  burning  them  at  unusually  high  efficiency  and 
replacing  them  at  the  end  of  much  less  than  500  hours. 

The  history  of  the  development  of  the  various  metallic  filament 
lampB  is  particularly  interesting  from  the  chemical  standpoint.  In 
ihe  early  days  of  incandescent  lighting  Mr.  Edison  and  others  rec- 
ognized the  peculiar  value  of  metallic  filaments  because  of  their 
flexibility  and  electrical  conductivity.  At  that  time  platinum  and 
iridium  were  the  metals  which  offered  most  promise.  They  were 
the  metals  of  highest  melting  point,  bo  far  as  then  known.  It  was 
soon  apparent  that  these  metals  could  not  be  run  at  high  enon^ 
temperature  to  make  a  practical  lamp,  though  they  were  very  nearly 
suitable.  Mr.  Edison  then  carried  out  a  great  number  of  experi- 
ments in  an  attempt  to  raise  the  melting  point  of  the  platinum. 
The  effect  of  the  occluded  gases  was  carefully  studied,  but  a  com- 
■  mercial  lamp  did  not  result  For  over  a  quarter  of  a  century  there- 
after, it.  remained  unknown  that  at  least  sir  or  seven  of  the  then 
known  metals  had  higher  melting  points  than  platinum.  The  en- 
tering' wedge  into  this  field  was  driven  by  Dr.  Auer  von  Welsbach, 
who  had  acquired  a  personal  and  almost  exclusive  knowledge  of  a 
large  group  of  more  or  less  rare  chemical  elements  in  connection 
with  his  extensive  researches,  which  were  crowned  by  his  gas-mantle 
inventions.    At  this  time  probably  none  of  the  metals  which  melt 

>y  Google 

The  Chemistbx  op  Luminods  Soubces  107 

higher  than  platinum  had  ever  been  produced  in  any  other  form 
than  that  of  a  fine  black  powder.  Osmium  vas  the  first  of  a  trio  of 
metals  to  become  a  nearly  practical  filament.  It  occutb  in  nature 
in  metallic  state,  usually  alloyed  with  iridium,  platinum,  rhodium 
and  mtJienium.  It  is  found  only  as  veiy  small  grains  or  plates, 
and  nowhere  in  any  considerable  quantity.  By  mixing  powdered 
metallic  osmium  with  a  suitable  ftarch  or  sugar  binder,  Welsbach 
squirted  a  thread  which,  after  drying  and  baking,  could  be  freed 
of  carbon  by  heating  in  a  mixed  atmosphere  of  hydrogen  and  water- 
vapor.  The  resulting  metallic  filament  was  quite  soft  when  hot, 
but  was  well  suited  for  incandescent  lamps,  as  it  withstood  tem- 
peratures necessary  to  produce  a  lamp  burning  satisfactorily  at 
about  1^  watts  per  candle-power.  The  world's  known  supply  of 
osmium  is  very  small,  and  to  conserve  this  supply  the  lamps  were 
usually  rented  instead  of  being  sold. 

In  1901  Dr.  Werner  von  Bolton  announced  the  discovery  of 
ductile  tantalum.  Operating  in  an  incandescent  lamp,  it  could  be 
burned  at  about  1.7  watts  per  candle-power  for  a  thousand  or  more 
hours.  The  metals  tantalum  and  niobium  are  a  pair  usually  occur- 
ring together  and  formerly  quite  difficult  of  separation.  They 
occur  in  small  quantities  in  Connecticut,  in  the  Black  Hills  of 
Dakota,  in  Sweden  and  in  Australia,  the  mineral  being  usually 
tantalite  (a  compound  of  the  oxides  of  tantalum  and  iron,  with 
or  without  manganese  or  tin)  or  some  combination  of  tantalum  and 
niobium  oxides  with  iron,  etc.,  as  columbite,  samarskite,  fergu- 
sonite,  etc.  It  was  necessary  first  to  perfect  methods  of  preparing 
the  pure  metals,  and  of  these  the  tantalum  was  found  to  have  the 
higher  melting  point.  It  is  about  3100°,  while  that  of  niobium  is 
about  2900°,  or  still  well  above  platinum. 

Until  this  investigation  it  had  apparently  been  known  only  as 
powder.  This  powder  was  melted  together  into  large  buttons  in 
an  electric  arc  and  then  drawn  to  wire  in  the  usual  manner  through 
diamond  dies. 

Probably  most,  if  not  all,  of  the  tungsten  filaments  in  the  lamps 
on  tiie  market  are  made  by  some  method  of  squirting  through  a 
die  tungsten  powder  mixed  with  a  binding  agent.  The  metal,  in 
■finely  divided  state,  is  usually  obtained  by  the  reduction  of  tungstic 
oxide  at  a  red  heat  by  hydrogen.  This  oxide  is  in  turn  obtained 
from  the  minerals  Wolframite,  which  is  a  tungetate  of  iron  or  iron 
and  manganese,  and  Scbeelite,  a  tungstate  of  calcium.     Several 

>y  Google 

108  Illduinatinq  ENOiN&sRixa 

thousand  tons  of  ore,  averaging  over  50  per  cent  tmigstic  oxid^ 
are  mined  annually,  largely  for  use  in  high-speed  tool  steel. 

Some  of  the  succeaeful  proceeeea  for  making  the  filamente  are  as 
follows : 

The  powdered  metal  is  mixed  with  a  proper  carbonaceous  binder, 
then  formed  into  threads  by  being  forced  through  a  Buitable  die, 
dried  and  haked  at  about  red  heat.  They  are  then  heated  by  pas- 
sage of  current  through  them  in  a  suitable  atmosphere  of  hydrogen 
or  mixture  of  hydrogen  and  nitrogen.  By  this  treatment  a  shrink- 
age of  the  filament  takes  place,  it  becomes  dense  and  metallic  in 
appearance,  and  at  the  same  time  the  carbon  present  is  removed. 
The  product  is,  therefore,  pure  tungsten. 

Similarly,  a  metallic  binding  agent  may  be  used.  The  finely 
divided  metal  in  one  such  process  is  mixed  with  a  cadmium-bismuth 
amalgam  and  the  resulting  mixture  is  pressed  through  a  die.  A 
thread  not  unlike  a  fine,  lead  fuse  wire  is  the  result.  On  heating 
this  in  in  vacuo  all  metals  but  the  tungsten  are  vaporized,  and  at 
the  final  temperature  this  is  also  sintered  t<^ther  into  a  compact 

In  the  case  of  tantalum,  nature  seems  to  supply  just  aboxit  enough 
of  the  ore  to  satisfy  the  demand,  and  probably  this  element  would 
have  been  a  more  successful  competitor  in  the  incandescent-lamp 
field  if  it  only  had  to  contend  against  carbon  and  osmium.  It  was 
more  efficient  than  the  former  and  much  more  plentiful  than  the- 
latter.  It  is  interesting  to  recognize  the  fact  that  the  most  recent 
successful  metal  filament,  tungsten,  occurs  in  nature  in  abundance. 
It  was  discovered  by  Scheele  in  1781.  For  over  200  years  it  was 
known  in  the  pure  state  only  as  an  infusible  gray  and  heavy  metallic 
powder.  Its  melting  point,  as  determined  by  Pirani,  is  3350°,  and 
is  the  highest  melting  point  of  which  we  have  measurement.  The 
only  measurement  of  higher  temperature  on  the  earth  is  that  of  the 
carbon-arc  crater,  said  to  be  about  3500°  C,  by  Burgess  and  Waid- 
ner.  In  all  types  of  incandescent  lamps  there  lies  a  promise  that 
continued  study  will  give  continued  advance  in  the  art.  This  is 
sought  usually  as  higher  efficiency.  A  carbon  lamp  will  bum  a  few 
moments  at  an  efficiency  10  times  as  great  as  its  normal  value.  In 
other  words,  from  the  materials  at  hand,  this  increase  in  efficiency 
is  possible  for  a  short  time.  It  seems,  therefore,  not  impossible 
that  this  limiting  time  feature  may  be  better  controlled  when 
better  understood. 

>y  Google 



By  Chablbs  Pbotbus  Steinuitiz 



1.  The  dlflerent  forms  of  radiators  and  different  Mndi  of  radiatltm. 

Claaa location  of  electric  llluinlnaiits. 

2.  Importance  of  the  volt-ampere  characteristic  and  the  resistance- 

temperature  characteristic  of  the  conductor  used  in  electric 
illumlaante.  Discussion  of  the  multiple  or  constant  potential, 
and  the  series  or  constant-cnrrent  electric  distribution  system. 

Solid  Cokddctobb 

3.  Volt-ampere  and  reslatance-temperature  characteristic  of  incandes- 

cent lamp  filaments.    Positive  and  negative  temperature  coelB- 

clentB,  -^  >0.  Stability  of  operation  on  constant  potential  and 
on  constant  cnrrent  circuits.  [Fig.  1:  Volt-ampere  characteristics 
of  Incandescent  lamp  filaments.  Fig.  2:  Resistance-character- 
istics of  incandescent  lamp  filaments.] 

4.  Volt-ampere    characteristic    of    pyroelectrolytlc    conductors.      The 

Nemst  lamp  glower  as  pyroelectrolyte.     The  Instability  range, 

-^  <0.  of  pyroelectrolytes  on  constant  potential  suwly.  and  the 
necssBlty  of  steadying  reelstance  or  reactance.  The  Nemst  lamp. 
[Fig.  3;  Volt-ampere  characteristic  of  low  resistance  pyroelectro- 

5.  The  light  radiation  of  solid  conductors,  ss  Incandescent  lamps  and 

the  Nemst  lamp  glower.  Black-body,  gray-body  and  colored- 
body  radiation.  Effect  on  the  efficiency  of  the  Incandescent  lamp 
filament  and  the  Nemet  lamp  glower.    Limitation  of  eOlciency. 

6.  Relation  of  refractoriness  and  vapor  tension  or  disintegration,  to 

the  possible  efficiency  of  the  Incandescent  lamp.    Comparison  of 

the  carbon  filament  with  the  metal  filaments. 
T.  The  production  of  the  carbon   filament  lamp.     Base  carbon  and 

treated  carbon,  and  their  stability. 
S.  Metallized  carbon,  Its  resistance  and  temperature  coeUclent,  and  the 

gem  lamp. 
9.  Metol-fllament  Incandescent  lamps.     Osmium  lamp,  tantalum  lamp, 

tungsten  lamp.    Their  eUclenclee. 

>y  Google 

110  Illominatino  Ekoiheebino 

10.  The  manutacture  of  tbe  tungBten  lamp. 

11.  Thlnnesa  and  length  of  metal  fllamenta.    Fragllltr. 

12.  Efflclencles  of  the  different  incan descent  lampe.    Conventional  rat- 

ing In  horizontal  candle-power.  Relation  of  efficiency  to  useful 

13.  Relation  of  tbe  efficiency  ot  tbe  incandescent  lamp  to  tbe  size  of  the 

unit,  or  the  power  consumption.  Limitation  by  supply  voltage  at 
small  units,  by  size  of  tbe  lanip  globe  at  lar(»  unlta.  Wide  range 
of  units  with  fairly  uniform  efficiency. 
H.  Inferiority  of  the  incandescent  lamp  In  efBclency,  to  the  flame  arc 
and  luminous  arc  Superiority  in  amall  unlta.  Haln  Seld  ot 
application  ot  Incandescent  lamps  and  Nemet  lamps  In.anutll 
nnlts,  where  no  other  electric  tUnminant  exists. 

Oasbous  CoNDucniBs 
IG.  Difference  between  dlaruptlve  or  Geissler-tnbe  conduction,  and  con- 
tinuous or  arc  conduction. 

Qbissleb-Tubi  Condl'ctioh 

16.  Electric  characteristics  ot  Gelssler-tube  conduction:    total  voltage, 

terminal  drop  and  stream  voltage  as  function  of  gas  pressure. 
[Fig.  4:  Volt-pressure  characteristic  ot  Gelister  tube  with  air  as 
conductor.  Fig.  6:  Volt-preasure  characteristic  ot  the  Oelssler 
tube  with  mercury  vapor  as  conductor.] 

17.  Performance,  efficiency  and  color  ot  light.    The  Moore  tube 

ABC  Conduction 

18.  Nature  of  the  arc  conductor.    Tbe  arc  as  unidirectional  conductor. 

Rectification  by  the  arc  Tbe  alternating  current  arc  Constant- 
pressure  and  varylng-pressure  arcs. 

19.  Volt-ampere   and    volt-length    characteristics   of   the   arc:    ^  <0. 

[Fig.  6:  Volt-ampere  characteristic  of  magnetite  arc  ot  .5,  1.5  and 
2.5  cm.  length.  Fig.  7:  Volt-length  characteristic  ot  magnetite 
arc  at  2,  4,  8  and  16  amperes.] 

20.  Dependence  ot  tbe  arc  voltage  on  two  independent  variables,  current 

and  arc  length.  Instabltty  of  the  arc  on  conatant  voltage  supply. 
Necessity  of  steadying  resistance  or  reactance.  Tbe  stablUtr 
curve  of  the  arc.  [Fig.  S:  Stability  curve  of  the  1.6  cm.  magna* 
tlte  arc] 

21.  Instability  of  parallel  operation  ot  arcs  without  stsadying  resis- 

tances. Instability  due  to  non-lnductlve  resistance  sbunt.  BX' 
tlnctlon  by  shunted  capacity.  Tbe  arc  aa  Interrupter,  The 
singing  arc 

22.  Stream  voltage  and  terminal  drop  of  tbe  arc    HeaUng  of  the  termi- 

nals by  the  terminal  drop.  The  carbon  arc  aa  Incandescent  radi- 
ator. Relation  between  tbe  efficiency  of  the  carbon  arc,  and  the 
alze  and  the  life  of  tbe  terminals. 

>y  Google 

Electric  Illuminakts  111 

23.  The  open  carbon  arc  or  sbort  burning  arc  lamp.     The  enclosed 

carbon  arc  or  long  burning  lamp.     Its  Inferiority  tn  efficiency. 

24.  nDec<»ii>mlcaI  operation  ot  contlnuoue-current  series  arc  circuits. 

Tbe  series  alternating  enclosed  arc  lamp.    lie  very  low  efSdency. 

25.  Replacement  of  tlie  enclosed  alternating  carbon  arc  by  tbe  magnetite 

arc  lamp  In  street  Itgbtlng,  by  tbe  intensified  arc  or  the  tungsten 
Incandescent  lamp  fn  Indoor  lighting.    Tbe  Int^uiifled  arc  lamp. 

26.  The  luminous  arc  and  the  flame  arc.     Their  character tatic  differ- 

ences, advantages  and  disadvantages.    The  magnetite  arc. 

27.  The  flame  carbon  arc.    Relation  between  size  of  electrodes  and  effi- 

ciency. The  Bhort-burnlng  and  the  long-burning  flame  carbon 
arc.  The  yellow  color  of  the  flame  carbon  arc.  Titanium,  calcium 
and  mercury  as  the  three  most  efficient  arc  stream  radiators. 

28.  The  mechanism  of  the  arc  lamp:   starting  device,  feeding  device, 

steadying  device,  shunt  protective  device,  damping  devices. 
Series  lamp,  shunt  lamp,  differential  lamp.  - 

29.  The  effective  resistance  of  the  arc.    Relation  between  arc  length  and 

efllclency.  The  abort  carbon  arc  and  tbe  long  luminous  and  flame 

30.  Regulation  of  arc  lamp  for  constant  light  flux.    Tbe  floating  system 

of  control  of  the  carbon  arc  and  its  advantages.  Fixed  arc  length 
required  by  the  luminouB  arc.  Its  difflcuttlee  In  constant  potential 
tamps.    Tbe  compromise  control  of  the  flame  carbon  lamp. 

31.  Classiflcation  ot  arc  lamps;  the  most  Important  forma  of  arc  lamps: 

Tbe  open  carbon  arc  on  9.G  amperea  aeriea  direct  current 

The  enclosed  carbon  arc,  for  multiple  and  series  circuits,  on 
alternating  and  on  direct  current 

The  Intenslfled  carbon  arc,  on  alternating  and  on  direct  cur- 
rent circuits. 

The  yellow  flame  carbon  arc,  on  alternating  and  on  direct 
current  circuits. 

The  magnetite  arc. 

The  mercury  arc. 

32.  Increase  of  tbe  efficiency  of  the  arc  with  Increasing  size  of  the 

light  unit.  Relation  between  the  efficiency'  of  the  arc  lamp  and 
the  current,  arc  length  and  power,  at  constant  arc  length,  con- 
stant current  and  constant  power.  The  condition  of  maximum 
efficiency.  [Fig.  9:  Efficiency  and  power  consumption  of  the  4- 
ampere  magnetite  arc  tor  different  arc  lengths.  Fig.  10:  Effi- 
ciency and  power  ot  the  .T-lncb  magnetite  arc  for  different  cur- 
rents. Fig.  11:  Efficiency,  arc  length  and  voltage  of  the  300-watt 
and  the  500-watt  magnetite  arc,  tor  different  currents.  Fig.  12: 
Relation  between  voltage,  current,  arc  length  and  efficiency  ot 
the  magnetite  arc,  under  the  condition  ot  maximum  efficiency, 
tor  various  powers.] 

33.  Comparison  of  tbe  arc  lamp  and  the  Incandescent  lamp. 

>y  Google 


Vacuvk  ABC8 

34.  The  low-presBure  mercur;  arc  In  tbe  glaaa  tube.    Ite  lilgh-preMure 
mercury  arc  la  tlie  quartz  tube.    Htdr  riuracterlBtlCi. 

BwiWiHii  I  TO  Tabiationb  or  tri  'BixcnHc  Pow^  Supply 

36.  Ckimpu-isOD  of  various  (ormji  of  Incandescent  lamps  and  arc  luniw 
regarding  their  sensltwity  to  variations  of  the  electric  power 

1.  The  Different  Forms  of  Eadiatora  and  DiffererU  Kinds  of  Badi- 
aiion.    ClagsificaHon  of  Electric  Hhuninants 

lo  the  production  of  light  from  electric  {fewer,  solids,  liquids  or 
gases  (the  latter  including  vapors)  may  be  used  as  conductors  of 
electric  power,  and  the  radiation  may  be  due  to  incandescence  of 
the  radiator,  that  is,  temperature  radiation  (black-body,  gray-body 
or  colored-body  radiation),  or  it  may  be  the  result  of  a  more  or 
less  direct  conversion  of  the  electric  power  into  radiation,  aa 

Solids  as  conductors  of  electric  power  are  nsed  in  the  various 
forms  of  incandescent  lamps :  the  different  types  of  carbon-filament 
lamps  and  the  metal-filament  lamps,  aa  the  osmium  lamp,  the 
tantalum  lamp  and  the  tungsten  lamp,  and  also  in  the  Nemst 
lamp.  Liquids  are  not  used  as  conductors,  due  to  their  difficulty 
of  application,  but  gases  and  vapors  are  extensively  used  in  the 
various  forms  of  arc  lamps,  as  the  open  and  the  enclosed  carbon 
arcs,  the  fiame  arcs  and  the  luminous  arcs,  which  latter  include 
the  vacuum  arcs,  and  in  the  Geissler  tube  as  illnminant  (Moore 
light).  In  the  former,  the  arc  lamps,  the  vapors  of  the  electrode 
material  are  used ;  in  the  latter,  the  Moore  light,  the  gas  which  fills 
the  space  between  the  electrodes. 

In  all  solid  conductors,  and  also  in  the  plain-carbon  arc  lamp, 
the  light  production  is  due  to  temperature  radiation  or  incandes- 
cence, either  black-body  or  gray-body  radiation,  or  colored-body 
radiation.  In  the  flame  arcs,  luminous  arcs  (including  vacuum 
arcs)  and  Geissler  tubes  luminescence  plays  an  essential  part  in  the 
light  production. 

>y  Google 

Electric  Illohinints  113 

S.  Importance  of  the  Volt-Ampere  Ckaracterigtic  and  the  Reaiat- 
ance-Tempemture  Characteristic  of  the  Conductor  Used  in 
Electric  Illuminantt.    Diaeuation  of  the  Mvltiple  or  Con- 
stant Potential,  and  the  Series  or  Constant-Current  Electric 
Distribution  Systems 
Since  in  electric  illuminante  the  light  is  given  by  electric  con- 
duction, the  properties  of  the  electric  conductor,  which  is  used  in 
the  illuminant,  are  of  the  foremost  and  fundamental  importance, 
that  is,  the  relation  of  current  and  roltage  to  each  other,  or  the 
so-called  "volt-ampere  characteristic "  of  the  conductor;  and  the 
relation  of  the  ratio  of  volta  and  amperes,  that  is,  the  effective 
resistance,  to  the  temperature,  that  is,  the  "  resistance  character- 
istic "  of  the  conductor.    This  is  obvious,  since  the  illuminant  must 
be  capable  of  use  in  the  existing  electric-power  distribution  systema. 
Electric  power  is  distributed  in  two  different  forms :  by  the  con- 
stant-potential or  multiple-distribution  system,  that  is,  at  the  con- 
stant voltage  of  110  or  230  volte,"  ot  by  the  constant-current  or 
series  system. 

'  In  the  constant-potential  system  all  apparatus  are  connected  in 
parallel  between  the  same  supply  mains,  and  thereby  receive  the 
same  voltage,  but  each  takes  a  different  part  of  the  supply  current. 
All  the  illnminants  must  therefore  be  designed  to  operate  at  the 
same  constant-terminal  voltage  of  110  or  230,  and  within  such 
variations  of  this  voltage  as  may  be  met  in  a  constant-potentiBl 
distribution  system,  which  varies  from  1  per  cent  to  5  per  cent  or 
more,  depending  on  the  character  of  the  system.  The  different 
illuminants,  however,  may  be  designed  for  different  currents.  The 
multiple  system  has  the  advantage  of  permitting  practically  un- 
limited extension :  with  increase  of  the  number  of  illuminants,  the 
current  in  the  supply  feeders  and  mains  increases,  and  larger  con- 
ductors become  necessary,  but  the  voltage  remains  the  same.  When 
the  number  of  illuminants  becomes  so  large  that  the  size  of  supply 
conductors  becomes  uneconomical,  more  sources  of  supply  become 
necessary.    Since,  however,  these  sources  of  supply  are  usually  sec- 

*  110  volts  here  means  any  constant  voltage  between  about  lOG  and 
12G.  and  220  volte  twice  tbli  value:  not  the  same  voltage  la  used  in  dl{- 
terent  distributing  systemB,  but  sllghtlr  different  voltages,  lor  tlie 
purpose  of  making  tbe  economical  production  of  exactly  rated  Incandes- 
cent lamps  possible.  (See  "General  Lectures  on  Electrical  Engineer- 
ing," by  the  author,  p.  12.) 

>y  Google 


ondary  stations,  that  is,  transfonDera  or  converters  receiving  their 
power  from  a  primary  generating  system  at  high  voltage,  this  intro- 
duces no  serious  limitation.  The  constant-potential  system  of  dis- 
tribution therefore  is  now  generally  used,  with  the  exception  of 
those  few  cases,  vhere  it  is  not  economical :  at  the  lov  voltage  of 
110  or  S30  volts,  the  distance  to  which  electric  power  can  be  sent 
is  rather  limited.  When  numerous  illuminants  are  scattered  over 
a  wide  area  this  difficulty  is  met  by  secondary  stations,  as  trans- 
formers, as  stated  above.  If,  however,  individual  illuminants  are 
scattered  over  a  wide  area,  as  in  street  lighting,  the  individual 
illuminants  cannot  be  reached  from  one  110-  or  230-volt  feeding 
point,  while  the- installation  of  a  transformer  at  every  lamp  is 
uneconomical,  and  in  this  case  the  constant-potential  system  becomes 
uneconomical  and  the  constant-current  system  is  used.  For  street 
lighting  the  series  system  is  therefore  universally  employed,  with 
the  esception  of  those  few  places  in  large  cities  where  the  street 
lamps  can  be  reached  by  a  multiple  system  installed  for  general 

In  the  constant-current  or  series  system  all  apparatus  are  con- 
nected in  series  with  each  other,  and  thereby  receive  the  same 
current,  and  the  voltages  consumed  by  the  different  illuminants 
add.  The  illuminants  therefore  are  designed  for  the  same  current, 
but  may  consume  different  voltages.  Since  the  voltage  of  a  dis- 
tribution circuit  cannot  be  indeiinitely  increased  without  involving 
difficulties  with  insulation  and  danger  to  life  and  fire  risks,  the 
number  of  apparatus  which  can  be  connected  into  one  series  circuit 
is  rather  limited;  a  aeries  circuit  is  a  very  small  unit  of  electric 
power,  from  our  present  point  of  view,  and  as  economy  requires  the 
use  of  the  largest  possible  units  series  circuits  are  used  only  in 
those  cases  where  they  are  economically  necesaaiy,  that  is,  for 
.  street  lighting.  It  was,  however,  with  series  arc  circuits  that  electric 
lighting  started  in  the  early  days. 

Series  circuits  are  usually  operated  at  4,  5,  6.6  or  7.5  amperes, 
some  of  the  old  open  carbon  arc  circuits  at  9.6  amperes,  and  with 
voltages  ranging  usually  from  4000  to  6000. 

Not  all  conductorB,  and  therefore  not  all  illuminants,  can  be 
connected  promiscuously  into  multiple  circuits  or  into  series  cir- 
cuits, even  if  designed  for  the  proper  voltage  respectively  current, 
and  the  study  of  the  electric  characteristics  of  the  conductors  which 
Are  used  in  illuminants  is  therefore  of  importance  for  their  design 
and  operation. 

>y  Google 

Elbctbic  Illpminants 



S.  VaU-Ampere  and  Resistance-Temperature  Characteristic  of  In- 
candescent Lamp  Filaments.  Positive  and  Negative  Tem- 
perature Coefficients,  —  >0.  Stability  of  Operation  on 
Constant-Potential  and  on  Constant-Current  Circuits 

The  conductors  of  incandcBcent  lamps  are  ohmlc  reeistaiiceB) 
that  is,  conductors  in  which  the  resiBtance  does  not  directly  de- 
pend on  curreot  or  Toltage,  bnt  is  constant  at  constant  tempera- 
tare,  and  if  it  varies  with  a  change  of  temperature,  in  case  of  a 
negative  temperature  coefficient,  that  is,  a  decrease  of  resistance 
with  increase  of  temperature,  the  decrease  of  resistance  with  in- 
crease of  temperature  is  less  than  the  increase  of  current  required 
to  cause  the  increase  of  temperature.  That  is,  such  conductors  are 
characterized  by  tiie  relation: 

d^  . 

In  other  words,  an  increase  of  enrrent  always  causes  an  increase 
of  terminal  voltage.  If  the  resistance  were  perfectly  constant, 
that  is,  the  temperature  coefficient  sero,  the  voltage  would  be  pro- 
portional to  the  enrrent,  and  the  volt-ampere  characteristic  given 
by  a  straight  line  going  through  the  origin,  I  in  Fig.  1,  and  the 
resistance  characteristic  given  by  a  horizontal  straight  line,  I  in 
Pig.  2.  No  conductor  exists  which  has  zero  temperature  coefficient 
over  more  than  a  limited  range  of  temperature. 

If  the  temperature  coefficient  is  positive  the  resistance  increases 
with  increase  of  temperature,  and  the  voltage  thus  increases  more 
than  proportional  to  the  current  j  that  is,  an  increase  of  current  i 
causes  an  increase  of  temperature  and  thereby  of  resistance  r,  and 
thus  an  increase  of  the  voltage  e=ir,  which  is  greater  than  pro- 
portional to  i,  as  shown  in  curves  II  to  IV  in  Figs.  1  and  8. 
Inversely,  if  the  temperature  coefficient  is  negative  the  resistance 
decreases  with  increase  of  current,  and  therefore  of  temperature; 
but  the  voltage  still  increases  with  increase  of  current,  though  less 
than  proportional  to  the  current,  as  shown  in  curves  V  and  VI 
in  Figs.  1  and  2. 

As  illustrations  are  shown  in  Fig.  1  the  volt-ampere  character- 
istic, and  in  Fig.  2  the  resistance  characteristic  of  the  conductors 
or  filaments  of  various  types  of  incandescent  lamps.     In  Fig.  1 

>y  Google 

Illduinatiko  Enginsbbiho 

the  co-ordinates  have  been  choeen  bo  ae  to  start  all  curves  at  tlie 
slope  of  4S°  at  the  origin.  In  Fig.  3  the  co-ordinates  have  been 
choeen  so  as  to  give  10  at  the  operating  point  of  the  lamp.     In 

n^             ^^     - 

7       - 


-A                    7 

7               / 

I               7 

1               L               /... 

t            ^/               ^  - 

1^7    .,,. 

l            i            ^      .'.. 

^         / 

4         4      ^^ 

1         t    ^7 

r      ^      Z           ^^1 

t        t     y        .y^^        .. 

/  /  .^^ 

1  J:7  ^^-^ 

t  77^^ 

L  ///^ 



r  .,     -i     Ji      0     ^     .t     >     .a     ^      to    /,/     /Hi     1.)     iA    IS 

Pio.  1. — Volt-Ampere  CharacterletlcB  of  Incoudescent  Lamp  Filamenta. 

Fig.  2  as  abscissae  have  been  used  'J'w,  which  with  a  black-body 
radiator  would  be  proportional  to  the  absolute  temperature  (for 
high  values  of  w).    It  is: 

I.  The  theoretical  conductor  of  constant  resistance. 
II.  The  tungsten  lamp  filament 

>y  Google 


III.  The  oamiTim  lamp  filament. 

lY.  The  metallized  carbon,  or  gem  lamp  filament. 

V.  The  treated  carbon,  or  3.1-watt  carboii'filament  lamp. 
VI.  The  untreated  caTbon,  or  base  filament. 

.    \ 











































\  \ 


operated  eatisfactorily  on  constant-potential  as  well  as  on  constant- 
cnrrent  circnite,  provided,  obviously,  that  ita  resistance  is  choceD 
so  as  to  consume  the  rated  power  at  the  constant  voltage  respectively 
current  of  the  circuit;  on  constant-potential  supply  the  current, 
and  thereby  the  power  consumed  by  the  conductor,  is  limited  to 

>y  Google 

lib  iLLOMiNATixa  Ekodjbbrinq 

that  corresponding  to  the  supply  voltage;  on  constant-current 
supply  the  terminal  voltage,  and  thus  the  power  consumed  by  the 
conductor,  is  limited  to  that  corresponding  to  the  snpply  current. 

Jf.    Volt-Ampere    Characteristic    of    Pyroelectrolytic    Conductors. 

The  Nemst  Lamp  Qlower  as  Pyroelectrolyte.     The  Inata- 

bility  range,     .  <0,  of  Pyroelectrolytes  on  Conatant-Poteur 

tial  Supply,  and  the  Necessity  of  Steadying  Resistance  or 

Reactance.  Tke  Nemst  Lamp 
Very  different  are  the  conditions  in  the  conductor  of  the  Nemst 
lamp,  the  Nemst  lamp  glower.  This  belongs  to  a  class  of  con- 
ductors, the  pyroelectrolytes,  in  which  the  temperature  coefficient 
within  a  certain  range  of  temperature,  and  thus  of  current,  is  so 
greatly  negative,  that  with  increase  of  current  the  terminal  voltage 
decreases.  That  is,  with  increase  of  tempei;ature  the  resistance 
drops  faster  than  the  increase  of  current  required  to  produce  the 
increase  of  temperature,  and  the  voltage  e=ir  thus  decreases  with 
increase  of  i.    In  this  range,  it  therefore  is : 

Such  pyroelectrolytic  conductors  are  many  metal  oxides,  silicates, 
sulphides,  etc.  A  typical  volt-ampere  characteristic  of  such  a  con- 
ductor (magnetite)  is  given  in  Fig.  3,  with  Vi  as  abscissae,*  the 
terminal  voltage  e  as  ordinates.     As  seen,  from  i=0  to  i„  it  is 

^f->0;fromi,  to  iritis  ^,! 
di         '  '        '  di 

With  most  pyroelectrolytes  the  voltage  peak  at  i,  is  so  high  that 

the  conductor  cannot  be  carried  beyond  it  by  the  mere  application 

of  voltage,  but  artificial  heating  is  required,  and  the  resistance 

bdow  i,  is  usually  extremely  high,  usually  near  i,  fusion  occurs, 

and  beyond  that  the  conductor  is  an  ordinary  electrolytic  con- 


The  operating  point  of  the  Nernst  glower  is  in  the  range  between 

i,  and  i,,  where -,t-<0. 

■  For  the  purpose  of  better  ahowlng  the  initial  part  of  the  curve,  VI 
Is  used  as  abscissae,  [nstead-of  i. 

tSee  "Electric  Conduction."  paper  read  before  the  Electrochemical 
Society,  1807,  by  the  author. 


Electric  lLLUMiN4.NTa 

,  de 

A  conductor,  in  which— p-<0,  can  be  operated  on  constant- 
current  supply,  but  cannot  be  operated  on  constant-voltage  supply ; 
but  at  constant  terminal  voltage  it  is  unstable  within  the  entire 
range  from  ij  to  i^,  in  Fig,  3 ;  on  constant-voltage  supply  an  in- 
crease of  current,  by  lowering  the  voltage  consumed  by  the  con- 
ductor, causes  a  further  increase  of  current  and  power,  and  thus 
further  decrease  of  voltage,  increase  of  current  and  power,  etc.. 



1  1  1  1  1 .  1 




1     1     1  *"[     1     1     1 


















and  the  conductor  destroys  itself  by  melting ;  a  slight  decrease  of 
current  causes  an  increase  of  the  voltage  required  by  the  con- 
ductor, and  since  this  is  not  available  on  constant- voltage  supply 
a  still  further  decrease  of  current,  increase  of  required  voltage, 
etc.,  and  the  conductor  open-circuits,  that  is,  the  lamp  goes  out. 
On  constant-potential  supply,  such  a  conductor  therefore  either 
open-circuits  or  short-circuits,  and  to  operate  it  at  constant  power 
on  a  multiple  ciicait  a  resistance  or  reactance  is  required  in  series 

>y  Google 

120  Illuuinatinci  Enoimeebinq 

to  the  pyroelectroljte  sufSciently  large  so  that  the  voltage  cousumed 
by  pyroelectrolyte  (glower)  plus  steadying  resistance  increases 
with  increase  of  current,  that  is,  fulfils  the  conditionB  of  operation 
on  constant-potential  supply, -p->0. 

The  Nemst  lamp  thus  requires  a  "steadying  resistance"  in 
series  to  the  glower.  To  reduce  this  resistance,  and  thereby  the 
waste  of  power  caused  by  it,  to  a  minimum,  iron  wire  is  used, 
operated  in  hydrogen  or  in  a  vacuum  at  that  range  of  tempera- 
ture at  which  the  temperature  coefficient  of  the  iron  is  abnormally 
high,  and  with  increase  of  the  current  i  the  resistance  r  very 
rapidly  increases,  thus  causing  an  abnormally  rapid  increase  of  ir. 

In  arc-  and  Geissler-tube  conduction,  a  similar  instability  on 
constant  potential  will  he  discussed. 

5.  The  Light  Radiation  of  Solid  Conductors,  as  Incandescent 
Lamps  and  the  Nemst  Olower.  Black-Body,  Oray-Body  and 
Colored-Body  Radiation.  Effect  on  the  Efficiency  of  the 
Incandescent  Lamp  Filaments  and  the  Nemst  Glower.  Limi- 
tation of  Efficiency 

The  light  production  by  solid  conductors  as  radiators  is  temper^ 
tare  radiation.  That  is,  by  the  renistance  of  the  conductor,  tiie 
electric  power  i*r  is  converted  into  heat,  causing  a  rise  of  tempera- 
ture which  produces  the  radiation. 

Normal-temperature  radiation,  that  is,  black-body  or  gray-body 
radiation,  as  given  very  closely  by  the  varioup  types  of  carbon- 
iilament  lamps,  is  a  very  inefficient  light  producer.  The  efficiency 
of  light  production  increases  with  increase  of  temperature,  but  is 
still  very  low  at  the  highest  temperatures  at  which  solids  can  be 
operated.  The  selective  radiation  of  a  colored  body  which  is  de- 
ficient in  radiating  power  in  the  ultra-red  gives  a  higher  efficiency 
of  light  production.  The  radiation  of  some  of  tlie  metal  filaments, 
and  that  of  the  Nemst  lamp  glower,  is  such  a  colored-body  radia- 
tion, and  thereby  gives  a  light  efficiency  higher  than  corresponds 
to  the  temperature  of  the  radiator.  However,  the  selectivity  seems 
to  decrease  with  increase  of  temperature,  that  is,  with  increasing 
temperature  the  body  seems  to  approach  more  a  gray  body.  For 
instance  the  Nemst  glower  radiates  strongly  selective  at  low  tem- 
perature, at  its  operating  temperature  the  radiation  curve  has 

>y  Google 

Elecisic  Illuhinants  121 

greatly  smoothed  out,*  and  while  there  is  probably  a  gain  io 
efficiency  in  some  metal  filaments  and  the  N^emat  glower  oyer 
nonnal-temperatare  radiation,  the  gain  does  not  seem  to  be  bo 
large  as  to  bring  the  efficdency  of  light  production  much  beyond 
that  reached  by  normal-temperature  radiation,  and  it  does  not 
appear  probable  that  we  shall  be  able  to  reach  very  much  higher 
efSciencies  by  colored-body  temperature  radiation.t 

6.  Relation  of  Refractoriness  and  Vapor  Tension  or  Disintegra- 
tion to  the  Possible  Efficiency  of  the  Incandescent  Lamp. 
Comparison  of  the  Carbon  Filaments  with  the  lUetal  Fila- 
ments. ■ 

Since  temperature  radiation  reaches  fair  values  of  light  efficiency 
only  at  very  high  temperatures,  only  the  most  refractory  bodiea 
come  into  consideration  as  radiators  in  incandescent  lamps. 

The  moat  refractory  substances  are  carbon,  tungsten,  osmium, 
tantalum,  etc4 

However,  refractoriness  ie  not  the  only  requirement,  but  the 
vapor  tension,  or  rate  of  disintegration  of  the  material  below  the 
melting  point,  ie  equally  of  importance,  since  on  it  depends  how 
far  we  can,  in  the  operating  temperature  of  the  radiator,  approach 
its  melting  point.  This  is  well  illustrated  by  the  relation  between 
tungsten  and  the  different  forms  of  carbon.§ 

Carbon  is  the  most  refractory  body,  and  has  been  the  first  em- 
ployed in  commercially  Guccessful  incandescent  lamps,  and  the 
carbon-filament  lamp  still  is  the  one  used  in  the  largest  quantities. 
Carbon  has  the  disadvantage  of  a  relatively  rapid  evaporation  or 
disintegration  far  below  its  boiling  point,  and  this  limits  the  oper- 
ating temperature  of  the  carbon  filament  so  that  we  cannot  get 
the  full  benefit  of  the  high  refractoriness  of  carbon;  but  metals,  as 
tungsten,  which  are  lees  refractory  than  carbon,  can  give  a  higher 
efficiency  by  being  operated  at  higher  temperature.  Great  differ- 
ences in  stability,  however,  exist  between  different  modifications 
of  carbon. 

'BnUetlna  of  the  National  Bureau  of  Standards. 

t  See  "  Radiation,  Llsht  and  Illnmliiatton,"  by  the  author,  p  TO. 

f  See  "  Radiation,  lAgbt  and  Illumination,"  p.  77. 

I  See  "  RadlaUon.  Light  and  IllumlnaUoo,"  p.  79. 

>y  Google 


7.  Tke  Production  of  the  Garbon-Filament  Lamp.    Base  Carbon 
and  Treated  Carbon,  and  their  StabUity 

The  first  commercial  carbon-filament  incandescent  lamps  were 
made  of  carbonized  bamboo  fiber.  Verj  soon  this  was  replaced  by 
the  squirted  filament,  which  could  be  produced  more  uniformly. 
A  solution  of  cellulose  in  zinc  chloride  (or  eupric  ammon),  or  of 
nitro-celluloee  in  glacial  acetic  add,  is  squirted  through  a  fine 
hole  into  a  hardening  solution:  methyl  alcohol  with  zinc-chloride 
solution,  diluted  acid  with  cupric-ammon  solution,  water  with 
nitro-cellulose.  The  filament  ia  then  washed,  put  into  the  desired 
shape  (in  the  case  of  nitro-cellulose,  after  reduction  to  cellulose) 
and  dried.  It  then  consists  of  a  structureless  cellulose,  in  appear- 
ance very  similar  to  horn.  This  is  now  carbonized  in  a  gas  furnace 
at  high  temperature,  and  constitutes  what  is  now  known  as  a  "  base 
filament,"  because  it  is  mainly  used  as  a  base  on  which  to  deposit 
a  better  form  of  carbon.  The  base  carbon  is  not  very  stable  at 
high  temperature,  and  early  lamps  made  of  it,  therefore,  had  only 
a  relatively  low  efficiency.  It  has  a  high  resistance  and  a  high 
negative- temperature  coefficient,  as  shown  by  its  characteristic  in 
Figs.  1  and  2.  Somewhat  later  a  considerable  improvement  in 
efficiency  resulted  from  the  introduction  of  the  "treated  filament." 
The  base  filament  is  eleotricaliy  heated  in  an  atmosphere  of  hydro- 
carbon vapor  (gasolene)  in  a  vacuum,  and  by  the  dissociation  of 
the  vapor  a  shell  of  a  different  modification  of  carbon  is  deposited 
on  the  base.  This  shell  carbon  has  a  far  greater  stability  at  high 
temperature,  thereby  allowing  the  operation  of  the  lamp  at  higher 
temperature  and  thus  higher  efficiency.  It  is  of  lower  resistance, 
and  in  the  treated  filament  lamp  most  of  the  current  thus  Sows 
in  the  shell;  less  in  the  inner  core  or  base  of  the  filament.  The 
temperature  coefficient  of  the  shell  carbon  is  still  negative,  but 
decreases  with  increasing  temperature,  and  finally  begins  to  rise, 
so  that  the  compound  structure  of  the  treated  filament  gives  a 
characteristic  as  shown  in  Fig,  S. 

S.  Metallized  Carbon,  its  Resistance  and  Temperature  Coefficient, 
and  the  Oem  Lamp 

A  few  years  ago  a  further  advance  was  made  by  discovering  a 
form  of  carbon  of  still  much  higher  stability,  the  metallized  carbon 
used  in  the  so-called  "gem  lamp."     The  shell  carbon   (but  not 

>y  Google 

Elkcthio  Illuhinants  123 

the  base  carbon)  converts  at  tlie  highest  temperature  of  the  electric 
furnace  into  a  modificatioli  of  carbon  of  nearly  metallic  character; 
it  has  a  very  low  reaistance,  lower  than  some  metals,  and  a  positive- 
temperature  coefficient,  like  metals,  though  lower  than  that  of 
pure  metals,  as  shown  by  the  characteristic  of  the  carbon  tilament 
with  metallized  shell,  in  Figs.  1  and  2.  In  the  production  of  the 
gem  lamp  the  haae  filament  is  heated  in  the  electric  furnace  to 
expel  all  impurities,  then  treated  in  gasolene  vapor,  and  thereby 
a  layer  of  shell  carbon  deposited  on  it,  and  then  is  once  more  heated 
in  the  electric  furnace.  The  filaments  are  then  sealed  in  glass 
bulbs  with  platinum  leading-in  wires  and  exhausted.  It  gives  an 
efficiency  of  about  3.3  watts  per  candle-power. 

Apparently,  the  electric  resistance  and  its  temperature  coefficient 
are  indications  of  the  stability  of  carbon  at  high  temperature; 
the  lower  the  cold  resistance  and  the  higher  its  temperature  co- 
efficient the  more  stable  is  the  carbon  at  high  temperature,  and  the 
higher  efficiencies  can  thus  be  reached. 

9.  Metal-Filament  Incandescent  Lamps.  Osmium  Lamp,  Tantalum 
Lamp,  Tungsten  Lamp.    Their  Efficiencies 

In  recent  years  metal-filament  incandescent  lamps  have  been 
developed,  and  are  rapidly  replacing  the  carbon-filament  lamps  by 
their  higher  efficiency. 

First,  the  osmium-filament  lamp  was  developed,  giving  an  effi- 
ciency of  about  1.9  watts  per  candle-power.  Its  filament  was  made 
by  some  squirting  process,  similar  to  the  carbon  filament.  It 
found  a  limited  use  only,  since  osmium  is  a  very  rare  metal,  exist- 
ing in  very  limited  quantities,  and  was  soon  replaced  by  the  tanta- 
lum filament.  Tantalum  is  a  ductile  metal,  and  the  tantalum 
lamp  is  made  by  winding  drawn  tantalum  wire  on  a  glass  frame. 
The  tantalum  lamp  gives  an  efficiency  of  about  2.6  watts  per 
candle-power,  hence  lower  than  the  osmium  lamp  but  higher  than 
the  gem  lamp.  Tantalum,  while  a  rare  metal,  exists  in  fairly  large 
quantities,  and  the  tantalum  lamp  appeared  very  promising  until 
the  development  of  the  more  efficient  tungsten  lamp  of  1,5  to  1.7 
watts  per  candle-power. 

The  tantalum  lamp  was  the  first  incandescent  lamp  made  of 
drawn  metal,  and  showed  the  features  of  a  much  better  life  with 
direct  current  than  with  alternating  current;  with  alternating  cur- 

>y  Google 

184  Illduinatino  Enginbbbino 

rent  the  dravn  filament  loses  ita  ductility  and  gradually  offeetsi 
that  is,  breaks  up  into  numerons  short  lengths,  which  aie  wdded 

10.  The  Manufacture  of  the  Tungsten  Lamp 

The  highest  efScienciee  of  incandescent  lamps  have  been  realized 
by  the  tungsten  filament.  Tungsten,  or  wolfram,  is  a  fairly  com- 
mon metal ;  is  extremely  refractory,  more  than  osmium  or  tantalum, 
but  less  than  carbon,  but  fairly  difficult  to  produce  in  such  purity 
ae  necessary  as  filament.*  Several  methods  of  manufacture  of 
tungsten  filaments  hare  been  devised  and  are  still  in  commercial 
development,  though  many  millions  of  tungsten  lamps  have  been 
made.  One  series  of  processee  consists  of  squirting  the  metal  as 
powder,  or  in  the  colloidal  state,  with  some  binder,  and  then  burn- 
ing out  the  binder  by  electric  heating  in  a  suitable  gas;  another 
by  squirting  a  filunent  of  tungsten  ozide  with  some  reducing  ma- 
terial, reduce  by  heat,  and  then  eliminate  the  excess  of  reducing 
material  and  of  oxide  by  electrically  heating  in  a  suitable  gas  at 
reduced  pressure.  A  third  process  consists  of  squirting  or  drawing 
a  wire  of  some  tungsten  alloy,  and  by  electrically  heating  evaporate 
the  alloying  metal  and  sinter  the  tungsten,  and,  finally,  methods 
have  been  found  to  draw  the  pure  tungsten  metal  into  wire  of 
sufficiently  small  size  for  use  in  filaments.  All  these  methods 
except  the  last  give  a  filament  which  is  not  ductile,  but  brittle,  like 
the  osmium  and  carbon  filament,  and,  therefore,  due  to  its  ex- 
treme thinness,  is  very  fragile. 

11.  Thinness  and  Length  of  Metal  Filaments.    Fragility 

All  these  metals  have  a  much  lower  resistance  than  the  base 
carbon,  which  constitutes  the  main  part  of  the  carbon  filament, 
and  since  they  are  more  efficient,  that  is,  at  the  same  supply  voltage 
require  less  current  for  the  same  light  flux,  they  must  be  of  ex- 
treme thinness  and  considerable  length.  Therefore,  in  these  lamps 
a  number  of  squirted  filaments  are  used  in  series,  or  with  drawn 
wire  a  considerable  length  of  wire  wound  zigzag  on  a  frame. 
This  difficulty  does  not  exist  with  the  metallized  carbon  filament; 

■  For  Instance,  a  contamination  by  0.3  per  cent  ot  carbon  woold  rep- 
resent an  impurity  of  10  per  c«nt  tungsten  carbide  WOiC. 

>y  Google 

Elsotbio  Illdiokanis  126 

Thjle  the  metallized  carbon  aleo  faaa  a  very  Iot  resiBtance,  it  la 
need  only  as  a  thin  shell  on  the  baae  carbon,  which  practically  does 
not  carry  any  cnrrent,  in  the  gem  lamp,  while  the  metal  filamentB 
are  solid  conductors  in  which  the  whole  croBS-Bection  conducts. 

IS.  Efficiencies  of  the  Different  Incandescent  Lamps.  Conven- 
tional Bating  in  Horizontal  Gandle-Power.  Relation  of  Effi- 
ciency to  Useful  Life 

The  approximate  efficiencies,  or  rather  specific  consumptions,  of 
the  different  types  of  incandescent  lamps  are: 

Base  carbon  filament  (not  used  any  more) 6  watts  per  c  p. 

Treated  carbon  fllam«it  4 

Metalllced  carbon  (gem  fllamettt) 3.3 

Tantalum  lamp  2.6 

Osmium  lamp    1.9 

TungBten  lamp    l.G  to  1.7* 

Light  flux  is  measured  in  lumens,  and  light  efRciency  thus  in 
lumens  per  watt,  specific  consumption  in  watts  per  lumen.  Usually 
instead  of  the  lumen  as  measure  of  the  light  output  of  an  illumi- 
nant  the  mean  spherical  candle-power  is  used,  which  is  -^  times 

as  much,  and  the  efficiency  then  given  in  mean  spherical  candles 
per  watt,  the  specific  consumption  in  watts  per  mean  spherical 

By  convention,  incandescent  lamps  are  usually  rated  in  mean 
horizontal  candles,  and  their  specific  consumption  expressed  by 
giving  the  watts  per  mean  horizontal  candles  and  the  spherical 
reduction  factor.  Thus,  above  lamps  are  commercially  rated  at: 

Treated  carbon  fllament . . .  3.1  watts  per  mean  borlzontal  candle-powar 

Oem  lamp  2.6  watts  per  mean  taorfEOatal  candle-power 

Tantalum  lamp  S.O  watts  per  mean  borlxontal  candle-power 

Osmium  lamp 1.5  watts  per  mean  borlzontal  candle-power 

Tungsten  lamp  1.15  to  1.33  watts  per  mean  horlsontal  candle- 

At  the  spherical  reduction  factor  0.78,  this  gives  above  values. 
In  comparison  with  other  illuminants,  obviously,  the  horizontal 
candle-power  has  no  meaning,  but  the  total  fiux  of  light,  that  is, 
the  mean  spherical  candle-power,  has  to  be  used. 

*See  "itadlatlon,  Light  and  Illumination,"  p.  179. 

>y  Google 

126  Illcminatinq  Enginbehinq 

When  coQflidering  efficiency,  however,  the  useful  life  of  the  lamp 
must  also  be  considered.  Obviously,  higher  or  lower  eflu^endes 
may  be  reached  by  operating  the  same  lamp  at  higher  or  at  lower 

When  speaking  of  the  efficiency  of  a  carbon-fiiaraent  lamp  it  is 
understood,  by  general  convention,  that  the  lamp  is  operated  at 
such  a  voltage  as  to  give  a  useful  life  of  500  hours.  As  useful  life 
is  understood  the  time  during  which  the  lamp,  on  constant- voltage 
supply,  decreases  by  20  per  cent  in  candle-power." 

With  metal  filaments  no  such  convention  has  yet  been  generally 
established,  but  due  to  the  higher  efficient^  and  higher  cost  of 
the  lamp  probably  a  useful  life  of  1000  hours  or  more  will  be 
economical,  t 

EfBciency  tests  of  incandescent  lamps  therefore  are  meaningless 
if  not  accompanied  by  life  tests  at  that  efficiency. 

IS.  Relation  of  the  Efficiency  of  the  Incandescent  Lamp  to  the  Size 
of  the  Unit  or  the  Power  Consumption.  Limitation  by  Sup- 
ply Voltage  at  Sviall  Units,  by  Size  of  the  Lamp  Qlobe  at 
Large  Units.  Wide  Range  of  Units  ivttk  Fairly  Uniform 

Characteristic  of  the  incandescent  lamp  is,  that  its  efficiency  is 
(theoretically)  independent  of  the  unit  of  light;  filaments  of  large 
diameter  and  great  length,  consuming  large  power  and  giving  a 
large  unit  of  light,  give  the  same  efficiency  when  operating  at  the 
same  temperature  as  filaments  of  small  diameter  and  short  length, 
that  is,  filaments  which  consume  small  power  and  give  small  units 
of  light,  and  operating  at  the  same  temperature,  should  have  the 
same  life.  Thus  incandescent  lamps  give  a  wide  range  of  sizes 
of  illuminants  of  nearly  the  same  efficiency. 

A  limitation  of  the  possible  size  of  incandescent'  light  units 
appears  with  small  sizes  in  the  voltage  of  the  system  of  electric 
power-supply.  At  the  same  supply  voltage — 110  or  S20 — a  smaller 
light  unit  requires  a  filament  of  smaller  diameter,  and  finally  a 
point  is  reached  where  the  small  diameter  makes  the  filament  so 
delicate  that  either  the  life  of  the  lamp  would  be  materially  short- 

*  See  '■  RadiatloD,  Light  and  Illumination,"  p.  79. 
t  See  "  General  Lecturee  on  Electrical  E^ne:lIle«^lIlg,"  by  the  author, 
p.  209. 

>y  Google 

Electbio  Illcminants  127 

ened,  or  a  lower  operating  temperatnre,  that  is,  lower  efficiency, 
mast  be  allowed.  Thns,  with  the  carbon-filament  lamp  on  110- 
volta  supply,  50  watts  (or  16  horizontal  candle-power  with  the 
treated  filament,  30  horizontal  candle-power  with  the  gem  fila- 
ment), are  the  smallest  units  at  which  full  efficiency  can  be  reached. 
Carbon-filament  lamps  of  lees  than  50  watte  for  110-volt  circuits, 
therefore  mast  he  made  for  lower  efficiency,  and  the  efficiency  low- 
ered the  more  the  smaller  the  unit  is.  ObTiouely,  for  a  55-volt 
circuit,  an  8-candle-power  lamp  could  be  made  of  the  same  effi- 
ciency as  the  16-candle-power  lamp  on  the  110-volt  drcnit,  and 
the  320-ToIt,  16-candle-powep  lamp  cannot  be  built  any  more  for 
the  same  efficiency  as  the  110-volt  lamp,  other  things  being  equal. 

The  same  applies  still  more  to  metal-filament  lamps,  as  in  these 
the  iilamenta  are  thinner  and  longer  than  in  carbon-filament  lamps 
of  the  same  voltage  and  candle-power.  Thus  in  the  tungsten  lamps 
higher  effidenciea  are  given  to  the  larger  units. 

For  low-voltage  lamps,  obviously,  this  limitation  of  miuimnm 
size,  by  the  mechanical  structure  of  the  filament,  does  not  exist, 
and  lamps  of  1-  or  2-watts  consumption,  or  even  less,  at  4-  to  10- 
volts  supply,  can  he  made  of  the  same  efficiency  as  the  50-watt  lamp. 

With  increasing  size  of  the  unit,  a  practical  limitation  is  also 
reached;  the  useful  life  of  the  carbon-filament  lamp  is  limited 
largely  by  the  blackening  of  tie  globe  by  carbon  deposits,  and  to 
give  equal  blackening  the  surface  of  the  lamp  globe  should  be 
proportional  to  the  power  consumed  in  the  lamp.  This,  however, 
gives  for  large  units  impracticably  large  glob^,  and  the  use  of 
smaller  globes  leads  to  a  shorter  life. 

This  limitation  exists  less  with  metal-filament  lamps.  In  these 
it  seems  that  the  life  is  not  so  much  limited  by  the  gradual  black- 
ening of  the  globe  as  by  impairment  of  the  vacuum,  and  for  equal 
performance  only  the  volume  of  the  globe  and  not  the  surface,  as 
with  the  carbon  filament,  should  increase  proportional  to  the  power 
consumption.  This  makes  metal-filament  lamp  units  of  several 
hundred  watts  feasible,  while  carbon-filament  lamps  of  such  power 
consumption  are  impracticable. 

The  gem  lamp,  due  to  the  metallic  properties  of  the  filament, 
stands  intermediate  between  the  treated  carbon  filament  and  the 
metal  filament  in  this  respect,  and  lamp  units  of  250  watte  have 
been  fairly  successful. 

>y  Google 

128  Illuminatino  Ekqineebihg 

ii.  Inferiority  of  the  Incandescent  Lamp  in  Efficiency  to  the  Flame 

Arc  and  Luminous  Arc.    Superiority  in  Small  Units.    Main 

Field  of  Application  of  Incandescent  Lamps  and  Nemst 

Lamps  in  Small  Units  where  no  Other  Efficient  Electric  II- 

luminant  Exists 

The  iiicandeeeent  lamp  thus  gives  unite  of  light,  of  pr&cticaU; 

the  same  efficiency,  from  a  fraction  of  a  candle-power  to  sereral 

hundred  candle-powers,  covering  a  nider  range  than  any  other 

electric  illuminant. 

However,  the  efficiency  of  light  production  is  of  lower  magnitude 
than  that  of  some  other  electric  illaminants;  even  in  the  moat 
efficient  incandescent  lainp,  the  tungsten  lamp,  the  specific  con- 
sumption of  1.6  to  1.7  watts  per  candle  is  of  fai  higher  magnitude 
than  the  specific  c-onsumption  reached  in  some  flame  arcs  and 
luminooB  arcs,  of  half  a  watt  or  less  per  candle-power. 

Thus,  in  efficiency,  the  incandescent  lamp  cannot  compete  with 
the  flame  arc  or  the  luminous  arc,  and  is  therefore  excluded  from 
economical  use  in  those  cases  where  these  arcs  can'  be  used,  but 
must  find  its  field  of  application  in  those  cases  where  the  more 
efficient  illuminanta  cannot  be  used,  and  especially  ie  this  the  case 
with  smaller  units  of  light,  since  the  efficiency  of  the  arc  rapidly 
decreases  witli  decreasing  power  consumption,  while  that  of  the 
incandescent  lamp  remains  the  same,  and  the  incandescent  lamp 
(including  the  N'cmst  lamp)  is  therefore  the  only  one  available 
for  smaller  units  of  light,  of  100  candle-power  or  less. 

15.  Difference  between  Disruptive-  or  Oetssler-Tube  Conduction 
and  Continuous  or  Arc  Conduction 
Two  forms  of  conduction  of  gases  or  vapors  exist:  diaruptive- 
or  Geissler-tube  conduction,  and  continuous  or  arc  conduction.  The 
distinction  is,  that  in  the  former  the  gas  which  fills  the  space  is 
the  conductor;  in  the  latter  conduction  takes  place  by  a  moving 
stream  of  electrode  vapor.  Gas  or  vapor  conduction  is  accom- 
panied by  luminescence  of  the  conductor,  and  thus  can  be  used 
for  light  production.  In  Geissler-tube  conduction  the  light  gives 
the  spectrum  of  the  gas  which  fills  the  space  between  the  electrodes; 
in  arc  conduction  the  spectrum  is  that  of  the  electrode  material.* 

*  See  "  RftdiatlOD,  Ltght  and  Illumination,"  p.  9S. 

>y  Google 

Eleothic  Illcmisants 


The  condactor  may  be  at  atmospheric  presBure,  as  in  the  carbon 
arcs,  flame  area  and  most  luminous  arcs;  or  in  a  vacuutn,  as  in 
the  Geieeler  tube  or  the  vacuum  arc  (of  which  the  only  industrially 
important  exponent  is  the  mercuTy  arc). 


10.  Electrical  Characteristics  of  Oeissler-Tube  Conduction:   Total 
Yoltage,  Terminal  Drop  and  Stream  Voltage  as  Function  of 
Oaa  Pressure 
.  Very  little  is  known  on  the  electrical  characteristics  of  Qeiflsler- 

tnbe  conduction.    The  only  commercial  illuminant  of  tliis  class  is 

the  Moore  tube. 


1   1 
































Pro.  4. — Volt-Preaaure  Cbaracterletlc  o 

OelBBler  Tube. 

It  seems  that,  at  constant  temperature  and  constant  gas  pressure, 
the  voltage  consumed  by  the  Geissler  tube  is  approximately  constant 
and  independent  of  the  current,  that  is,  ~if-  =0.  The  volt-ampere 
characteristic  of  the  Geissler  tube  thus  would  be  a  straight  horizon- 
tal line.  As  result  hereof,  a  Geissler  tube  cannot  be  operated  on 
constant-supply  voltage,  but  requires  a  steadying  resistance  or  re- 
actance to  fulfil  the  conditions  of  stability,--,^  >0.  The  reactance 
of  the  step-up  transformer  is  used  for  this  purpose  in  the  Moore 

>y  Google 



The  Toltage  consimied  b;  the  OeisBler  tube  conBiBta  of  a  potential 
drop  at  the  tenainals,  the  "  terminal  drop,"  and  a  voltage  con- 
sumed in  the  luminous  stream,  the  "  stream  voltage,"  which  latter 
is  proportional  to  the  length  of  the  tube.  Both  greatly  depend  on 
the  gas  pressure,  and  vary  with  varying  gas  pressure  in  opposite 
directions:  with  increasing  gas  pressure  the  terminal  drop  de- 
creases and  the  atteam  voltage  iDcreaeea,  and  the  total  voltage 
consumed  by  the  tube  thug  gives  a  minimmn  at  some  definite  gas 
pressure.  This  pressure  of  minimum  total  voltage  depends  on 
the  length  of  the  tube,  and  the  longer  the  tube  is  the  lower  is 
the  gas  pressure  of  minimum  total  voltage. 


1    1    1 







































Fra.  5. — ^Tolt-Preasure  Characterlatlc  of  Qelsaler  Tube 

In  Fig.  4  is  shown  the  voltage-presaure  characteristic,  at  constant 
current  of  0.1  and  of  0.05  ampere,  of  a  Qeissler  tube  of  1.3  cm. 
diameter  and  200  cm.  length,  using  air  as  conductor;  and  in  Fig.  6 
the  characteristic  of  the  same  tube  with  mercury  vapor  as  con- 
ductor.* Figs.  4  and  5  also  show  the  two  component  voltages, 
the  terminal  drop  and  the  stream  voltage.  As  abscissae  are  used 
the  logarithms  of  the  gas  pressure,  as  measured  by  McLeod  gau^ 
at  the  moment  of  taking  current  and  voltage  readings. 

■  It  la  Interesting  to  note,  tbat  total  voltage,  terminal  drop  and 
stream  voltage  In  tbe  Qelaaler  tube  using  mercury  vapor  as  conductor, 
are  nearly  tbe  aame  as  with  air,  and  entirely  different  from  tbe  terminal 
drop  and  the  atream  voltage  of  the  vacuum  mercury  arc  The  spectrum 
is  the  same,  the  mercurv  apectrum. 

>y  Google 

Elboteio  Illuminants  131 

With  increasing  preeGure  the  discharge  finally  stops,  dne  to  the 
limited  supply  voltage;  with  decreasing  pressure,  finally  the  gas 
density  becomes  so  low  that  a  tendency  to  arc  condnction  appears, 
and  the  beginning  of  arc  formation  naually  destroys  the  tube. 

17.  Performance,  Efficiency  and  Color  of  Light.    The  Moore  Tube 

As  seen,  the  values  of  terminal  drop  are  very  high,  and  as  this 
voltage  gives  no  equivalent  of  light,  efficiency  requires  the  use  of 
such  a  long  tube  as  to  make  the  terminal  drop  a  small  part  of  the 
total  voltage.  In  consequence  hereof,  the  Moore  tube  is  a  very 
large  unit  of  light  and  does  not  allow  economical  subdivision.  It 
requires  high-voltage  alternating  current,  which  is  usually  pro- 
duced by  a  step-up  transformer  attached  to  the  terminals  of  the 
tube.  Intermittent  direct  current  may  equally  well  be  used,  but 
continuous  direct  current  is  not  suitable,  as  the  Geissler-tnbe  con- 
duction rapidly  changes  to  arc  conduction,  and  as  the  latter  re- 
quires much  lower  voltage,  leads  to  short-circxdt. 

In  the  Geissler  tube  the  terminals  disintegrate  and  the  gas 
pressure  falls  fairly  rapidly,  possibly  by  absorption  of  the  gas  by 
disintegrated  electrode  material.  As  commercial  illuminant,  the 
Geissler  tube  therefore  requires  means  of  feeding  gas  intermittently 
into  the  tnbe.  This  is  done  in  the  Moore  tube  by  an  automatic 

As  far  ae  known,  the  most  efficient  Geissler-tnbe  conductor  U 
nitrogen.  It  gives  a  reddish-yellow  light,  of  an  efficiency  which 
in  very  long  tubes  reaches  values  of  2.6  watts  per  candle-power, 
that  is,  about  the  same  as  the  tantalum  lamp,  but  of  lower  magni- 
tude than  the  Same  arc  and  the  luminous  arcs.  Carbon  dioxide 
CO,  is  also  used  as  conductor.  It  gives  a  white  light,  but  a  lower 
efficiency.  Mercury  vapor  gives  it  green  light,  but  also  at  low 

The  great  advantage  of  the  Moore  tube  is  its  low  intrinsic  bril- 
liancy, and  in  the  CO,  tnbe  its  white  color. 


IS.  Nature  of  the  Are  Conductor.    Th/:  Arc  as  Unidirectional  Con- 
ductor.   Rectification  by  the  Arc.    The  .iUernating-Current 
Arc.    Constant-Prfamre  and  Varying-Pressure  Arcs. 
In  the  electric  arc  the  current  is  carried  across  the  apace  between 
the  electrodes  or  arc  terminals  by  a  stream  of  electrode  vapor  which 

>y  Google 


issues  from  a  spot  on  the  uegative  tenainsl,  the  so-called  negative 
spot,  as  a  high-Ydocity  blast  (probably  of  a  velociij  of  several 
thoiiBand  feet  per  second).  If  the  n^ative  terminal  is  finid  the 
negative  spot  causes  a  depression,  which  is  in  a  more  or  less  rapid 
motion,  depending  on  the  fluidity.  Before  arc  conduction  can  take 
plaee  the  vapor  stream  has  to  be  produced,  that  is,  an  arc  baa  to 
be  started.  This  is  done  by  bringing  the  electrodes  into  contact 
and  then  separating  them,  or  by  a  high-voltage  spark  or  a  Qeisaler 
discharge,  or  by  the  vapor  stream  of  another  arc,  or  by  heating  the 
space  between  the  electrodes,  for  instance,  by  an  incandescent 

The  arc  stream  is  conducting  only  in  the  direction  of  its  motion, 
that  is,  any  body  whicii  is  reached  by  the  arc  stream  is  conductively 
connected  with  it,  if  electro-positive  regards  to  it,  but  is  not  in 
conductive  connection  if  negative  or  isolated.  The  arc  thus  is  a 
unidirectional  conductor,  and  as  such  has  found  an  extensive  use 
for  the  rectification  of  alternating  current.f 

Since  the  arc  is  a  unidirectional  conductor,  it  usually  cannot 
exist  with  alternating  current,  since  at  the  end  of  every  half  wave 
the  vapor  stream  extingiUEbes,  and  at  the  beginning  of  the  next 
half  wave  a  new  vapor  stream  in  opposite  direction  has  te  be 
sterted.  An  alternating-current  arc  exists  only  if  the  conditions 
are  such  that  at  every  half  wave  a  new  arc  starts.  This  is  the 
case  if  the  voltege  in  the  circuit  ia  sufGciently  high  to  send  a  dis- 
ruptive spark  across  the  gap  at  every  half  wave,  or  if  the  arc 
temperature  is  so  high  as  to  start  the  arc,  as  is  the  case  with  the 
carbon  arc-J 

In  their  industrial  application  we  may  distinguish  between  con- 
stant-pressure arcs  and  varying-preseure  area,  that  is,  arcs  in  an 
enclosed  space,  usually  a  vacuum,  in  which  the  gas  or  vapor  pres- 
sure varies  with  the  current,  etc.  The  only  industrially  used  arc 
of  the  latter  class  is  the  mereury  are. 

*  See  "Radiation,  Light  and  Illumination,"  p.  106. 

t  On  the  arc  as  unldlrectlonai  conductor,  see  "  Radiation,  Light  and 
Ilium loation,"  p.  111.  On  the  electric  characteristics  of  the  mercury 
arc  rectifier,  see  "  Theory  and  Calculation  of  Transient  Electrical 
Phenomena  and  Oeclllatione,"  by  the  author,  p.  249. 

X  See  "  Radiation,  Light  and  Illumination,"  p.  115. 

>y  Google 

Electuo  Illdmikahtb 



19.  'Volt-Amperes  and  VoH-Length  Characteristics  of  the  Are, 


f  1- 






KX  A 



















— n 

— . 
















J  . 




,     ^ 

Characteristic  of  the  arc  as  conductor  ia,  that  the  Toltag«  de- 
creases with  increase  of  current,  that  i8-,?<0  oyer  the  entire 

range.    The  volt-ampere  characteriBtica  of  the  arc  therefore  are 
,  cnrrea  of  Che  shape  shown  in  Fig.  6  for  the  magnetite  arc,  for  the 

>y  Google 


Illdiunatinq  Enqinesbino 

arc  lengths  of  0.6,  1.6  and  2.5  cm.  With  increasmg  current  the 
arc  voltage  decreaeee  and  approaches  a  finite  limiting  value,  which 
with  the  magnetite  arc  is  about  30  volts  (about  36  volts  with  the 
carbon  arc,  13  volta  with  the  mercury  arc,  etc.).  Inversely,  with 
decreasing  current  the  voltt^  increases,  and  tends  towards  infinity. 










































f ' 


t     * 

'    ' 

1*     J 

'  .i 



Pw.  7. 

or  rather  probably  the  voltage  required  by  the  electrostatic  spark, 
that  is,  by  Geisaler-tube  conduction  serosa  the  arc  gap. 

At  constant  current,  with  increasing  arc  length,  the  arc  voltage 
increases  very  nearly  proportional  to  the  arc  length,  and  the  volt- 
length  characteristics  of  the  arc  thus  are  practically  straight  lines, 
as  shown  in  Fig.  7  for  the  magnetite  arc  of  %,  4,  8  and  16  amperes.* 

•  See  "  RadtattoQ.  Light  and  Illumination,"  p.  137. 

>y  Google 

Elbotsio  Illduinakts  135 

80.  Dependence  of  the  Arc  Voltage  on  Two  Independent  Variables, 
Current  and  Arc  Length,  Inet(Aiiiti/  of  the  Are  on  Oonatant- 
Voltage  Supply.  Necessity  of  Steadying  Resistance  or  Re- 
actance.   The  Stability  Curve  of  the  Arc 

The  arc  as  conductor  in  indnBtrial  illumlnants  thus  differs  from 
the  solid  conductjirB  discussed  in  the  preceding  by  two  main  char- 

a.  In  the  solid  conductors  the  relation  between  e  and  i  is  fixed, 
that  is,  e  is  determined  by  i,  and  inyersely.  In  the  arc,  however, 
two  independent  variables  exist,  the  current  or  voltage  and  the 
arc  length.  That  is,  e  is  a  function  of  i  as  well  ae  of  1  which  can 
be  expressed  with  fairly  good  approximation  (except  for  very  small 
currents,  for  which  the  voltage  is  higher  than  given  by  the  equa- 
tion) by  the  formula: 

where  eo,  c  and  S  are  constants,  depending  on  the  material  of  the 
electrodes,  and  more  particularly  on  the  negative  electrode. 

Least  cloee  ie  the  agreement  with  above  formula  in  the  carbon 
arc,  which  in  many  other  properties  shows  an  exceptional  character 
as  result  of  the  physical  properties  of  carbon.* 

'f  • 


Herefrom  follows: 

An  arc  is  unstable  and  cannot  be  operated  on  constant-voltage 
supply,  but  with  constant  voltage  at  the  arc  terminals  a  slight 
momentary  increase  of  the  arc  resistance,  by  requiring  a  higher 
voltage,  decreases  the  current  and  thereby  still  further  increases 
the  required  voltage  and  the  arc  goes  out.  Or,  a  slight  momentary 
decrease  of  the  arc  resistance  increases  the  current,  thus  lowers  the 
arc  voltage,  thereby,  at  constant-supply  voltage,  increases  the  cur- 
rent and  still  further  lowers  the  arc  voltage,  etc.,  and  the  arc 
ehort-drcuite.  The  arc,  however,  is  stable  on  constant-current 

The  arc  thus  ie  essentially  a  constant-current  phenomenon,  its 
operation  more  steady  on  constant-current  circuits,  and  additional 
apparatns  is  required  for  its  operation  on  constant-potential  cir- 

*  See  "  Radiation,  Ugbt  and  lUomlnatlon,"  p.  140. 

>y  Google 



GuiU.  That  ie,  a  resistance  or  Teactaoce  (with  alternating  arcs) 
moat  be  inserted  in  series  snfScientJy  large  so  that  for  the  total 
voltage  consumed  by  the  arc  with  its  steadying  resistance --^>0. 
Thus,  while  in  Fig.  8  the  lower  curve  is  the  volt-ampere  chaiy 


l<vl     1 






























* ' 

*  / 

Fig.  S. 

acteristic  of  a  1.5  cm.  magnetite  arc,  to  operate  such  an  arc  on  a 
constant-potential  supply  a  much  higher  voltage  is  required:  the 
supply  voltage  muet  be  greater  than  that  given  by  the  upper  curve 
in  Fig.  8  to  give  stable  operation,  and  the  more  so  the  greater  the 
required  stability.  This  curve  thus  is  called  the  "stability  curve" 
of  the  arc.* 

•  See  "  Radlatton.  Light  and  H lamination,"  p.  142. 

>y  Google 

Elbctbic  Illumimants  137 

!B1.  Inatability  of  Parallel  Operation  of  Arcs  Without  Steadying 
Resistances.  Instability  Due  to  Non-inductive  Resistance 
Shunt  Extinction  by  Shunted  Capacity.  The  Arc  as  In- 
terrupter.   The  Singing  Arc 


Several  arcs  cannot  be  operated  in  parallel  except  by  giving 
each  of  tbem  a  eteadjing  resistance  or  reactasce  ae  large  as  would 
be  required  for  its  operation  on  oonetant-potentia]  circnit.  With- 
out this  ail  the  arcs  go  ont  but  one. 

Shunting  the  arc  by  ■  a  non-inductive  resistance  decreases  its 
Btabilily,  and  with  decreasing  resistance  a  definite  value  is  reached 
at  which  the  arc  becomes  unstable,  that  is,  goes  out.  The  stability 
of  an  arc  thus  can  be  measured  by  the  current  which  can  be  shunted 
around  it  by  a  non-inductive  resistance. 

A  condenser  in  shunt  to  the  arc  makes  it  unstable  and  interrupts 
it;  a  momentary  increase  of  arc  resistance,  and  thereby  increase  of 
are  voltage,  increases  the  current  shunted  momentarily  by  the 
condenser,  thereby  decreases  the  arc  current,  and  still  further  in- 
creases the  arc  voltage  and  shunts  still  more  current  into  the  con- 
denser, etc.  Even  a  small  condenser  in  shunt  to  the  arc  thus  puts 
it  out.  If  the  supply  voltage  la  sufficiently  high  to  restart  the  are, 
after  it  is  put  out  by  a  shunted  condenser,  the  arc  with  shunted 
condenser  then  acta  as  an  interrupter,  causing  rapid  successive 
interruptions  of  the  circuit  with  fairly  constant  frequency.  The 
lower  the  stabiliiy  of  the  arc  the  more  sudden  are  the  interruptions, 
and  low-temperature  ares,  as  the  mercuiy  arc,  thus  give  inter- 
nptions  of  extreme  suddenness.  Inversely,  if  the  capacity  is  very 
small  and  the  gas  filling  the  space  around  the  arc  stream  of  low 
dielectric  strength,  as  hydrogen  or  light  hydrocarbons,  the  arc 
may  start  again,  through  the  residual  are  vapor,  before  completely 
extinguished,  and  the  arc  current  becomes  pulsating,  the  so-called 
"singing  arc." 

22.  Stream  Voltage  and  Terminal  Drop  of  thp.  Arc.    Heating  of 

the  Terminals  by  the  Terminal  Drop.     The  Carbon  Arc  at 

Incandescent  Radiator.    Relation  between  the  Efficiency  of 

the  Carbon  Arc  and  the  Size  and  the  Life  of  the  Terminals 

Voltage,  and  therefore  power,  is  consumed  in  the  arc  stream  and 

at  the  arc  terminals.    The  power  consumed  in  the  arc  stream  is 

>y  Google 

138  Illuuinatino  Enqinbskino 

converted,  more  or  leas  directly,  into  radiation,  and  if  a  large  part 
of  this  radiation  ie  in  the  visible  range,  as  ia  the  case  vith  titanium, 
caldmn  and  mercury  vapor  as  oondiictore,  the  arc  stream  may  be 
used  as  iUamlnant.  If  very  little  of  tlie  radiation  is  in  the  visible 
range — as  is  the  case  with  carbon  vapor  as  conductor — the  arc 
stream  does  not  contribute  appreciably  to  the  light  given  by  the 

The  power  consumed  at  the  electrodes  is  partly  converted  intA 
the  latent  heat  of  evaporation  and  the  kinetic  energy  of  the  moving 
vapor  stream  (which  is  the  arc  conductor)  largely  into  heal^ 
especially  at  the  positive  terminal.  If  the  arc  terminals  then 
are  sufficiently  small  to  rednce  the  heat  conduction  away  from 
them,  and  of  sufficiently  refractory  materia]  to  reach  very  high 
temperature,  they  may  be  used  as  radiators  in  giving  light.  The 
radiation  then  is  due  to  incandescence  or  temperature  radiation. 

The  latter  is  the  case  with  the  plain  carbon  arc  lamp.  When 
using  pure  carbon  as  arc-lamp  electrodes  the  arc  stream  gives  very 
little  light,  and  that  of  a  useless,  violet  color.  Considerable-  heat 
is,  however,  produced  at  the  positive  electrode,  and  if  this  is  not 
t«o  large  its  tip  reaches  a  very  high  temperature :  the  boiling  point  . 
of  carbon,  and  then  gives  hght  by  temperature  radiation,  practically 
black-body  radiation.  The  plain  carbon  arc  therefore  gives  light 
by  incandescence,  just  like  the  carbon-filament  incandescent  lamp, 
and  the  arc  stream  in  the  former  ie  merely  the  heater  which  raises 
the  temperature  of  the  radiator,  the  positive-electrode  tip,  to  a 
high  temperature,  and  the  much  higher  radiation  efficiency  and 
white  color  of  the  carbon  arc,  compared  with  the  carbon  filament, 
is  due  to  the  higher  temperature  of  the  former.  Nevertheless, 
while  the  radiation  efficiency  of  the  carbon  arc  is  the  highest  which 
can  be  reached  by  black-body  radiation,  it  ie  very  much  lower  than 
the  efficiencies  available  by  lumineecence  of  the  arc  etream. 

Of  the  heat  produced  at  the  positive  terminal  of  the  carbon  arc 
only  a  part  becomes  useful  as  incandescent  radiation;  the  rest  is 
conducted  away  through  the  electrode,  carried  away  by  air  currents, 
etc.  The  lower  this  loss,  that  is,  the  smaller  the  electrodes,  the 
higher  is  therefore  the  efficiency,  and  with  very  large  electrodes 
the  heat  conduction  becomes  so  large  that  the  electrode  tips  do  not 
reach  any  more  the  temperature  of  efficient  radiation,  and  the 
efficiency  vanishes.  The  efficiency  of  the  carbon  arc  lamp  thus 
depends  on  the  size  of  the  electrodes,  and  increases  with  decreasing 

>y  Google 

Ei^BOTEio  Illvuinants  139 

size.  However,  with  decieasiog  size,  the  cooBtimptiou  of  the  elec- 
trodee  by  combustion  increases,  aod  thtia  reqoires  more  frequent 
trimming  of  the  lamp,  that  is,  higher  cost  of  maintenance. 

2S.  The  Open  Carbon  Arc  or  Short-Burning  Arc  Lamp.  The 
Enclosed  Carton  Arc  or  Long-Burning  Lamp.  Its  Inferiority 
in  Efficiency 

The  first  carbon  arc  lamps  were  operated  vith  high  current: 
on  9.6  amperes  constant  direct-crirrent  circuits,  with  electrodes, 
which  were  fairly  small  relative  to  the  current,  and  therefore  gave 
fairly  good  efficiencies:  about  1  watt  per  candle-power.  However, 
under  these  conditions,  the  rate  of  consumption  of  the  electrodes 
was  very  rapid,  and  electrodes  of  the  greatest  lehgth,  which  could 
conveniently  be  used  in  a  lamp,  lasted  only  a  few  hours.  As  result 
thereof,  twin  carbon  lamps  were  designed,  and  were  in  extensive 
use.  The  high  cost  of  operation,  due  to  the  required  daily  trim- 
ming, of  these  so-called  "open  arc  lamps"  or  "short-burning  arc 
lamps"  led  to  the  development  of  the  enclosed  carbon  arc  lamp. 
In  this  type  of  lamp  the  arc  is  enclosefl  in  a  small,  nearly  air-tight 
glass  globe,  and  the  rate  of  consumption  of  the  electrodes  thereby 
greatly  reduced  and  a  longer  life  of  electrodes  secured.  As  the 
retarded  combustion  of  the  electrodes  resulted  in  their  assuming' 
a  more  flattened  shape,  the  aic  length  had  to  be  increased  to  limit 
the  obstruction  of  the  light  issuing  from  the  positive  electrode  by 
the  shadow  of  the  negative  electrode.  The  higher  arc  voltage  ro- 
sulting  herefrom  required  a  decrease  of  current  to  retain  the  same 
power  consumption,  and  while  the  open  arc  operated  at  40  to  45 
voli«  on  9.6  amperes  circuits,  the  enclosed  arc  lamp  consumes  70 
to  75  volta  on  6.6  or  7.5  amperes  circuits.  As  the  same  size  of 
electrodes  was  retained,  or  the  size  even  increased,  to  get  a  long 
life,  while  the  current  and  thereby  the  luminous  area  of  the  elec- 
trodes was  reduced,  the  heat  losses  by  conduction  and  convection  - 
were  greater  in  the  enclosed  arc,  and  the  efficiency  therefore  lower 
than  in  the  open  arc.  Nevertheless,  the  advantage  of  lower  main- 
tenance cost  resulting  from  the  less  frequent  trimming,  weekly 
with  the  enclosed  arc  lamp  against  daily  with  the  Apen  arc  lamp, 
has  led  to  the  entire  abandonment  of  the  latter,  and  while  open 
arcs  have  survived  in  a  few  cities  they  have  practically  ceased  as 
an  article  of  manufacture. 

>y  Google 

140  Illduinatiho  Enoikeebino 

34.  UTieconomiail  Operation  of  Continuous-Ourrent  Series  Are 
Circuiit.  The  Series  Alternating  Enclosed  Arc  Lamp.  lit 
Very  Low  Efficiency 

In  regard  to  the  electrical-power  supply,  the  enclosed  arc  lamp 
is  inferior  to  the  open  are  lamp,  since  with  the  former  the  higher 
voltage  and  lower  current  gives,  with  the  same  maximum  voltage 
of  the  constant-current  circuit,  a  smaller  unit,  and  with  direct 
current  an  arc  machine  was  required  for  each  circuit.  This  waB 
such  an  economical  disadvantage  that  the  direct-current  series 
enclosed  carbon  arc  lamp  is  used  to  a  limited  extent  only  in  such 
places  vhere  efficiency  of  light  production  is  essential,  and  the 
illuminant,  which  is  most  universally  used  for  street  lighting,  is 
the  constant-current  alternating  enclosed  carbon  arc  lamp.  With 
this  lamp,  operating  from  constant-current  transformers,  the  small 
size  of  the  individual  arc  circuit  is  not  such  a  serious  handicap. 

The  economic  disadvantage  of  numerous  small  machine  units, 
which  handicapped  the  series  direct-current  arc  lamp,  has  been 
eliminated  by  the  development  of  the  constant-current  mercury 
arc  rectifier  system,  which  permits  operation  of  constant-direct 
current  arc  circuits  from  constant-current  transformers.  This  de- 
velopment, however,  was  too  late  to  help  the  direct-current  carbon 
arc,  but,  coming  after  the  development  of  the  luminous  arc,  it  led 
to  the  rapid  introduction  of  the  latter  in  place  of  the  carbon  arc. 

The  efficiency  of  the  alternating-current  carbon  arc  lamp,  how- 
ever, is  much  lower  than  that  of  the  direct-current  lamp:  in  the 
alternating-current  lamp  the  losses  of  heat  through  the  electrodes 
are  more  than  doubled :  while  the  heat  loss  by  conduction  and  con- 
vection is  continuous,  heat  is  produced  at  either  electrode  mainly 
during  that  half  wave  of  current  where  the  electrode  is  positive, 
and  then  only  during  that  part  of  this  half  wave  where  the  current 
is  high.  Thus,  while  the  alternating-current  carbon  arc  lamp  gives 
light  from  both  electrodes,  its  efficiency  of  light  production  is  much 
lower,  and  with  the  standard  series  enclosed  alternating-current 
arc  lamp  at  70  to  75  volts  per  lamp,  on  6.6  and  7.5  amperes  con- 
stant alternating-current  circuits,  the  specific  consumption  Is  up 
to  3.5  to  3  watts  per  candle-power,  and  even  higher,  that  is,  the 
efficiency  has  dropped  down  below  that  reached  with  modem  in- 
candescent lamps. 

In  spite  of  its  very  low  efficiency,  the  small  amount  of  attention 
required  by  it,  and  the  convenience  of  operation  from  altemating- 

>y  Google 

ElECTEIO  iLlDMIiTANia  141 

cnrreot  supply  circuits,  through  constant-current  tTansformers  in 

street  lighting,  has  led  to  the  almost  nnivenal  adoption  of  the 

alternating-current  enclosed  carbon  arc  lamp,  and  probably  more 

lamps  of  this  type  are  used  in  street  lighting  than  of  all  other  types 


96.  Replacement  of  the  Enclosed  Alternating  Carbon  Arc  by  the 

Magnetite  Arc  Lamp  in  Street  Lighting,  by  the  Intensified 

Arc  or  the  Tungsten  Incandescent  Lamp  in  Indoor  Lighting. 

The  Jntengified  Arc  Lamp 
However,  with  the  development  of  high-efficiency  incandescent 
lamps,  the  position  of  the  standard  enclosed  alternating  carbon  arc 
lamp  became  untenable,  and  while  it  is  still  being  used  in  enormous 
numbers  it  is  being  rapidly  replaced  by  the  magnetite  arc  lamp 
in  street  lighting,  and  by  the  intensified  arc  lamp  and  the  tungsten 
incandescent  lamp  in  indoor  lighting,  and  the  manufacture  of  the 
enclosed  alternating  carbon  arc  lamp  has  greatly  decreased. 

While  thus  the  enclosed  carbon  arc  lamp  is  rapidly  disappearing 
from  the  streets,  before  the  luminous  arc,  for  indoor  lighting, 
where  the  luminous  arc  and  the  flame  arc  are  handicapped  by  being 
too  large  units  of  light,  and  by  producing  eraoke  and  gases,  and 
the  tungsten  lamp  is  the  only  competitor,  the  enclosed  carbon  arc 
lamp  is  retaining  its  field  as  the  "  intensified  arc  lamp."  Since 
the  efliciency  of  the  carbon  arc  lamp  increases  with  decreasing  size 
of  carbons,  by  the  use  of  very  small  carbons  in  an  enclosed  type 
of  lamp,  a  very  good  efficiency,  about  1  candle-power  per  watt,  is 
reached  in  the  so-called  "intensified  arc  lamp"  on  direct  current 
as  well  as  on  alternating  current.  The  life  of  the  electrodes  of  the 
intensified  arc  lamp  is  shorter  than  that  of  the  enclosed  arc  lamp 
of  old,  but  aa  this  lamp  is  mainly  used  indoors,  where  usually  the 
daily  operation  is  only  a  few  hours,  the  life  is  sufficient  to  reduce 
the  frequency  of  trimming  satisfactorily,  and  the  higher  efficiency 
and  white  color  of  light  gives  to  the  intensified  arc  an  advantage 
over  the  tungsten  lamp  in  those  cases  where  large  units  of  light 
are  permissible. 
S6.  The  Luminous  Arc  and  the  Flame  Arc.    Their  Characteristic 

Differences,  Advantages  and  Disadvantages.    The  Magnetite 

The  carbon  arc  is  an  illuminant  using  a  solid  radiator  and  pro- 
ducing light  by  incandescent  radiation,  like  the  incandescent  lamps. 

>y  Google 

14S  Illduihatino  Enoxneesinq 

In  all  other  arcs  luminesceiice  plays  an  eseential  part,  and  all  or 
most  of  the  light  is  given  by  the  arc  Same  a&  vapor  conductor. 

These  luminescent  arcs  can  be  divided  into  two  classes:  the 
luminous  arcs  and  the  fiame  ares.*  In  the  luminous  arcs  the  lumi- 
nescent material  is  introduced  into  the  arc  stream  by  electro-con- 
duction from  the  negative,  that  is,  is  used  as  arc  conductor.  Typical 
arcs  of  this  class  are  the  so-called  magnetite  arc  and  the  mercury 
arc.  The  latter,  as  vacuum  arc,  will  be  discussed  later.  In  the 
flame  arcs  the  luminescent  material  is  Introduced  into  the  arc 
stream  by  heat  evaporation,  either  from  the  positive  as  the  hotter 
terminal,  or  from  both  terminals.  The  characteristic  difference 
resulting  herefrom  is,  that  in  the  luminous  arc  the  temperature 
of  the  electrode  has  no  direct  relation  to  the  ^Sciency,  and  the 
electrodes  thus  can  be  maintained  at  such  low  temperature  as  to 
consume  very  slowly.  The  luminous  arc  thus  lends  itself  to  the 
production  of  long-burning  arc  lamps,  that  is,  lamps  requiring  very 
infrequent  trimming,  and  the  size  of  the  electrodes  is  usually  made 
such  as  to  give  a  life  of  100  to  200  hours  as  the  longest  time  which 
it  is  advisable  to  allow  a  lamp  to  bum  without  cleaning  the  globe, 
and  other  attention.  The  positive  electrode  of  the  luminous  arc 
is  entirely  immaterial,  and  usually  made  of  some  metal  of  high 
heat  conductivity  so  as  not  to  consume  appreciably,  that  is,  of  a 
life  of  some  thousand  hours. 

At  the  same  time  the  number  of  materials  which  can  be  used 
in  the  luminous  arc  ia  much  more  limited,  the  di£Bculties  of  design 
80  as  to  get  steady  operation,  greater  than  with  the  flame  arc,  and 
no  succesBfu]  luminous  arc  has  yet  been  commercially  developed 
for  alternating-current  circuits,  but  the  luminous  arc  has  been 
developed  for  direct-current  circuits  in  the  so-called  "  magnetite 
arc  lamp,"  also  occasionally  called  "  metallic-oxide  arc  lamp  "  and 
"  ferro-titanium  arc  lamp."  In  this  the  negative  electrode  is  a 
mixture  of  the  oxides  of  iron,  titanium  and  chromium  (magnetite, 
illmenite,  rutile,  chromite),  usually  enclosed  by  a  thin  iron  shell. 
The  positive  electrode  is  a  permanent  part  of  the  lamp. 

The  magnetite  arc  lamps  are  operated  on  constant  direct-current 
zircuits  of  4  amperes  and  of  6.6  amperes,  with  about  75  volts  per 
lamp,  usually  from  constant-current  transformers  through  mercury 
arc  rectifiers. 

•See  "Radiatloa.  Light  and  IlluminaUon,"  p.  123. 

>y  Google 

Electric  iLLnMiNANTS  143 

27.  The  Flame  Carbon  Arc.  Relation  between  Size  of  Electrode 
and  Efficiency.  The  Short-Burning  and  the  Long-Burning 
Flame  Carbon  Arc.  The  Yellow  Color  of  the  Flame  Carbon 
.ire.  Titanium,  Calcium  and  Mercury  as  the  Three  Most 
Efficient  Arc-Stream  Radiators 

Id  the  flame  arcs  the  luminescent  material  is  introduced  into 
the  arc  stream  largely  by  heat  eTaporation,  a  high  temperature 
of  the  positive  electrode  thus  is  essential,  and,  to  Bome  extent, 
similar  relations  exist  between  the  size  and  therefore  temperature 
of  the  electrodes  and  the  efficiency.  Carbon  is  always  need  as  the 
main  electrode  material,  since  carbon  gives  the  hottest  arc,  and 
also  the  steadiest  arc,  and  the  inherent  8t«adinese  of  the  carbon, 
arc  has  made  the  development  of  the  flame  carboD  arc  lamp  lees 
difficult  and  therefore  more  rapid  than  that  of  the  luminous  arc, 
and  made  it  possible  to  operate  such  ares  on  alternating-current 
circuits  as  well  as  on  direct-current  circuits. 

Since,  however,  carbon  rapidly  consumes,  and  the  size  of  the 
electrodes  cannot  be  materiaUy  increased  without  loss  of  efficiency, 
the  flame  carbon  arc  lamp  is  essentially  a  short-burning  arc  lamp, 
requiring  daily  trimming.  This  has  in  this  country  excluded  its 
use  for  general  street  illnmination,  and  restricted  it  largely  to 
decorative  lighting. 

To  make  the  flame  carbon  arc  long  burning  requires  enclosing 
it  similar  as  with  the  enclosed  plain  carbon  arc  to  reduce  the 
access  of  air.  Since,  however,  by  the  consumption  of  the  electrodes 
the  luminescent"  materials  contained  therein  escape  as  a  smoke, 
means  are  required  to  deposit  this  smoke,  by  a  circulating  system, 
at  some  place  where  it  does  not  obstruct  the  light  by  deposition 
on  the  globe.  A  number  of  such  long-burning  flame  lamps  have 
been  designed,  but  none  of  them  has  yet  found  an  extended  in- 
dustrial introduction,  probably  largely  due  to  conditions  outside 
of  the  lamp  mechanism :  the  yellow  color  of  the  light,  the  large 
unit  of  light,  the  expense  of  the  electrodes,  lack  of  steadiness,  etc. 

The  only  materials  which  thus  far  are  used  in  flame  carbon  arcs 
as  luminescent  matter  are  calcium  compounds,  as  fluorides,  borates, 
phosphates,  tungstates,  etc.  They  give  a  very  high  efficiency,  but 
a  yellow  light.  White-flame  carbons  have  not  yet  been  introduced 
of  an  efficiency  comparable  with  that  of  the  more  efficient  yellow- 
flame  carbons. 

>y  Google 

X44  Illuhinatino  Enoinesrinq 

To  some  extent  the  flame  carbon  arc  etanda  intermediate  between 
the  luminons  arc  and  the  plain  carbon  arc:  the  plain  carbon  arc 
givee  light  only  1^  incandescence  of  the  electrode  terminals,  the 
InminoTiB  arc  only  by  Imninescetice  of  the  arc  stream,  and  the 
flame  carbon  arc  gives  most  of  its  light  by  luminescence  of  the  are 
stream,  bnt  also  some  light  by  incandescence  of  the  positiYe  carbon 

It  is  interesting  to  note,  that  thus  far  only  t^ree  materials  have 
been  found  which  in  the  arc  give  very  high  efficiencies  of  light 
production,  reaching  in  large  units  values  of  3  to  4  candles  pec 
watt;  titanium,  calcium  and  mercuiy.  The  first  gives  a  white 
light,  and  is  used  in  the  magnetite  arc;  the  second  gives  a  yellow 
light,  and  is  used  in  the  flame  carbon  arc;  and  the  third  is  re- 
stricted to  the  vacuum  arc. 

S8.  The  Mechanism  of  the  Arc  Lamp:  Starting  Device,  Feeding 
Device,  Steadying  Device,  Shunt  Protective  Device,  Damping 
Device.    Series  Lamp,  Shunt  Lamp,  Differential  Lamp 

Due  to  the  nature  of  the  arc,  as  discussed  above,  all  arc  lamps 
require  an  operating  mechanism. 

Since  the  arc  does  not  start  spontaneously,  a  starting  device  is 
required.  This  consists  of  a  mechanism  which  brings  the  electrodes 
together,  thereby  closes  the  circuit  between  them,  and  then  sepa- 
ratee them  and  so  starts  an  arc. 

Since,  in  supplying  the  vapor  conductor  of  the  arc  stream, 
the  electrodes  consume,  more  or  less  rapidly,  a  feeding  device  is 
necessarj-,  that  is,  a  mechanism  which  gradually  ■  moves  the  elec- 
trodes together  so  as  to  maintain  the  proper  length  of  the  arc 

In  constant-potential  or  multiple  lamps  a  steadying  device  is- 
necessary  since,  as  seen,  the  arc  is  unstable  on  constant  potential. 
This  consists  of  a  resistance  in  direct  current,  of  a  reactance  in 
alternating-current  arc  lamps,  which  is  connected  in  series  to  the 
arc,  and  usually  made  adjustable  so  as  to  accommodate  the  lamp 
to  the  different  supply  voltages  met  in  electric-supply  systems. 

In  constant-current  or  series  lamps  a  shunt  protective  device  is 
necessary  to  close  the  circuit  around  the  arc  in  case  the  circuit  in 
the  lamp  opens  by  breakage  or  consumption  of  the  electrodes.  This 
usually  consists  of  a  shunt  resistance,  connected  across  the  lamp 
terminals  by  a  potential  magnet. 

>y  Google 

Elsctbio  Illduinants  145 

In  addition  thereto  dashpots  or  other  retarding  devices  are  nec- 
essary to  slow  down  the  motion  of  the  operating  mechanism  so  as 
to  draw  the  arc  sufficiently  slowly  not  to  break,  and  to  gnard  against 
OTer-Teaohing  of  the  feeding  mechanism. 

The  operating  mechanism  is  actuated  by  electromagnets  or 
solenoids,  frequentiy  in  combinatioa  with  weights,  and  rarely 
springs,  as  the  latter  hare  uenally  proved  unreliable  in  continuoue 

If  only  series  magnets  are  used  the  lamp  is  called  a  series  lamp ; 
if  only  shunt  magnets  are  used  it  is  called  a  shunt  lamp;  if  series 
and  shunt  magnets  are  used,  a  differential  lamp.  The  aeries  lamp 
regulates  for  constant  current  in  the  lamp,  thus  is  not  applicable 
where  several  lamps  are  connected  in  series ;  the  shnnt  magnet 
regulates  for  constant  voltage,  irrespective  of  the  current,  and  the 
*  differential  lamp  r^ulates  for  constant  relation  of  current  and 
voltage.    The  latter  type  is  most  commonly  used. 

The  different  forms  of  arc-lamp  mechanisms  which  are  in  in- 
dustrial use  cannot  be  described  here,  but  may  be  studied  from  the 
pubUcations  of  the  various  arc-lamp  manufacturers,  which  give 
detailed  information,  or  by  inspection  of  the  exhibit  of  typical 
arc  lamps  shown  here,*  and  only  some  general  principles  can  be 
diacnesed,  which  may  enable  a  judgment  of  the  correctness  of 
individual  operating  mechanisms. 

IS0.  The  Effective  Beaistance  of  the  Arc.    Relation  between  Arc 
Length  and  Efficiency.    The  Short'Carbon  Arc  and  the  Long 
Luminous  and  Flame  Arcs 
The  effective  resistance  of  the  arc  is  not  constant,  but  continu- 
ously and  often  rapidly  varies  or  pulsates  somewhat.     The  arc 
conductor  is  a  vapor  stream  of  a  temperature  very  much  higher 
than  the  surrounding  air,  and  thus,  even  when  well  screened,  more 
or  less  affected  by  air  currents,  drafts,  etc.     In  the  plain  carbon 
arc  lamp,  in  which  the  heated  terminals  are  the  radiator,  and  the 
voltage  consumed  by  the  arc  stream  is  wasted,  the  arc  length  is 
made  as  short  as  possible,  witiiont  obstructing  the  light  by  the 
shadow  of  the  electrodes,  and  the  fluctuations  of  the  arc  resistance 
therefore  are  moderate.     In  the  Same  arcs  and  luminous  arcs, 
however,  in  which  the  light  is  given  by  the  arc  stream,  and  the 

•  AUo  sw  "  Radiation,  Light  and  Illumination,"  p.  161. 

>y  Google 

146  Illuuinatino  Ekoikeerinq 

potential  drop  at  the  terminalB  represents  largely  waited  power, 
efficiency  requires  a  long  arc  stream,  and  tliis  is  more  sensitive 
to  air  curreDts,  thns  the  fluctuations  of  the  are  resistance  are 
greater,  especially  when  very  small  currenta  are  used,  as  necessary 
in  smaller  nnite  of  light. 

The  problem  in  arc-lamp  design  thus  is  to  devise  an  operating 
mechanism  which  regnlates  as  closely  as  possible  for  constant 
production  of  light  by  the  arc  lamp,  and  at  the  same  time  permits 
the  use  and  economical  operation  of  the  are  lamp  on  existing  dis- 
tribution systems. 

SO.  Regulation  of  the  Arc  Lamp  for  Constant  Light  Flux:    the 
Floating  Sgstem  of  Control  of  the  Carbon  Arc  and  its  Ad- 
vantages.   Fixed  Arc  Length  Required  &y  the  Luminous  Arc. 
Its  Difficulties  in  Constant-Potential  Lamps.     The  Compro- 
mise Control  of  the  Flame  Carbon  Lamp 
In  the  plain  carbon  are  the  light  production  depends  on  the 
current,  but  not  on  the  arc  length,  provided  the  latter  is  sufficient 
to  minimize  the  shadow  of  the  electrodes.    Regulation  for  constant 
light  flux,  therefore,  ie  closest  by  control  for  constancy  of  current. 
Thns  the  series  magnet,  which  varies  the  are  length  to  maintain 
constant  current,  is  most  satisfactory  in  constant-potential  lamps, 
while  in  constant-current  lamps  the  controlling  mechanism  merely 
has  to  maintain  the  arc  length  sufficient  for  reducing  the  electrode 
shadows,  and  not  too  long  to  give  too  much  waste  of  power.     As 
the  arc  length  has  no  direct  elTect  on  the  light,  a  floating  system  of 
control  thus  can  be  used,  and  is  always  used,  as  being  easiest  to 
operate.    That  is,  one  or  both  carbons  are  held  floating  by  the 
counteracting  forces  of  shunt  and  series  magnet,  or  of  series  magnet 
and  weight,  and  continuously  move  in  adjusting  the  arc  length  to 
the  fluctuations  of  are  resistance.    The  voltage  at  the  terminals  of 
such  a  constant-current  are  lamp  thus  shows  very  small  fluctuations. 
Entirely  different,  however,  are  the  conditions  in  the  luminous 
ere.     In  this  the  light  flux  is  proportional  to  the  current  and  the 
arc  length,  and  any  fluctuation  of  the  are  length,  by  a  floating 
system  of  control,  would  give  a  corresponding  fluctoation  of  light 
flux,  and  is  therefore  objectionable.    A  fluctuation  of  are  resistance 
is  accompanied  by  a  change  of  luminosity,  such  that  an  increase 
of  are  resistance  and  therefore  of  arc  voltage  at  constant  arc  length 
usually  gives  a  decrease  of  luminosity,  and  thereby  of  light  flux; 

>y  Google 

Elbotbic  Illuminants  147 

and  regulation  for  constant  light  flux  therefore  would  require  an 
increase  of  arc  length  at  an  increase  of  arc  voltage  caused  by  an 
increafied  arc  resistance.  The  floating  system  of  control,  by  short- 
ening the  arc  at  an  increase  of  arc  voltage,  thus  in  this  case  controlB 
in  the  wrong  direction  and  accentuates  fluctuations  of  light,  and 
the  nearest  approach  to  proper  regulation  of  light  flux  is  given  by 
maintaining  constant  are  length.  In  luminous  arc  lamps  there- 
fore always,  as  far  as  it  is  possible,  a  control  for  fixed  arc  length 
is  used.  That  is,  the  arc  terminals  are  locked  in  position  at  a  fixed 
distance,  and  at  intervals,  depending  on  the  rate  of  consumption, 
this  distance  is  adjusted  by  resetting  the  arc.  This  fixed  arc-length 
control  gives  a  curve  of  terminal  voltage,  which  fluctuates  con- 
siderably, following  the  fluctuations  of  the  are  resistance.  In  con- 
stant-current circuits  this  ia  not  objectionable,  as  the  voltage  fluc- 
tuations of  the  numerous  lamps  in  series  with  each  other  superpose 
to  a  constant  total  voltage.  In  multiple  or  constant-potential  lamps, 
however,  the  fluctuatious  of  arc  voltage  may  interfere  with  the 
operation,  and  thus  either  a  very  large  inductance  has  to  be  used 
in  series  to  the  arc,  to  steady  the  current,  or  regulation  for  con-- 
stant  light  flux  more  or  less  sacrificed  by  the  use  of  a  floating 
system  of  control,  and  as  the  result,  the  multiple-luminous  are 
lamp  is  less  steady  than  the  series  arc  lamp. 

In  flame  are  lamps,  usually  lai^er  currents  and  thus  longer  arcs 
are  employed,  and  a  sluggish  floating  mechanism,  if  limited  to 
work  over  a  moderate  range  only,  is  leas  objectionable,  but  never- 
theless the  light  flux  of  the  lamp  is  less  steady  than  in  the  plain 
carbon  lamp,  and  one  of  the  main  objections  of  the  flame  arc  is  its 
inferiority  in  steadiness  of  the  light  flux. 

31.  Classification  of  Arc  Lamps.    The  Most  Important  Forms  of 

Arc  Lamps 

In  claeaifying  the  different  types  of  arc  lamps  we  have : 

By  the  nature  of  the  light  production ;  the  plain  carbon  are,  the 

flame  carbon  are  and  the  luminous  are,  the  latter  including  the 

mercury  are  as  vacuum  arc. 

By  the  life  of  the  electrodes :  the  short-burning  arc  and  the  long- 
burning  arc.  The  former  giving  a  life  of  electrodes  of  from  8  to  20 
hours,  depending  on  the  current  and  the  size  of  electrodes,  the 
latter  a  life  of  50  to  250  hours,  or  even  much  more,  as  with  the 
mercury  are. 

>y  Google 


By  the  protection  of  the  arc  against  the  accem  of  air :  the  open 
arc  and  the  encloeed  arc. 

By  the  nature  of  the  supply  circuits:  the  constant-potential  or 
multiple  arc  lamp  and  the  constant-cnrrent  or  series  arc  lamp. 

By  the  arrangement  of  the  electrodes :  the  vertical  arc  and  the 
horizontal  arc.  In  the  former  the  electrodes  are  arranged  Yertically 
above  each  other,  and  the  msjdmnm  light  flux  thne  isBoes  in  the 
horizontal  direction,  except  in  the  direct-carrent  plain  carbon  arc, 
in  which  the  maximum  light  flux  is  downwards  from  the  upper 
positire  electrode  as  radiator.  In  the  horizontal  arc  the  electrodes 
are  converging  downwards,  and  the  maximum  light  flux  tiius  is 
in  the  downward  direction. 

The  most  important  forms  of  arc  lamps  thus  are : 

The  open  plain  carbon  arc.  A  short-burning  arc,  which  has 
survived  in  a  few  cities  on  9.6  amperes  series  circuits. 

The  enclosed  plain  carbon  arc.  Long  burning,  for  multiple  and 
for  aeries  circuits,  on  alternating  and  on  direct  current.  The  ma- 
jority of  the  arc  lamps  now  in  use  are  series  alternating  enclosed 
■  carbon  arcs,  on  6.6  amperes  and  on  7,5  amperes  series  circuits. 
This  iype  of  arc  is,  however,  rapidly  disappearing,  due  to  its  low 

The  intensified  arc.  It  ie  an  enclosed  plain  carbon  arc,  medium 
long  burning,  gaining  its  efficiency  by  the  small  size  of  the  elec- 
trodes. It  is  mainly  used  for  indoor  lighting  of  high  efficiency  and 
white  color,  on  constant-potential  direct-  and  altematinff-carKnt 

The  yellow-flame  arc.  Usaally  an  open  and  ahort-buming  arc, 
with  converging  carbons  for  downward  distribution  of  light,  used 
mainly  for  outdoor  decorative  lighting,  and  to  some  extent  for 
second-class  interior  lighting.  Its  disadvantage  is  the  yellow  color 
of  the  light. 

The  magnetite  arc,  mainly  used  on  4  amperes  and  6.6  amperes 
direct-current  series  circuits,  for  street  light,  where  it  is  taking  the 
place  of  the  series  enclosed  carbon  arc.  It  is  an  open,  long-bumiDg 

The  mercury  arc  or  vacuum  arc,  mainly  used  for  indoor  lighting 
of  high  efficiency  and  steadiness.     Its  disadvantage  is  the  green    ■ 
color  of  the  light. 

>y  Google 

Electric  Illdminantb 


39.  Increase  of  the  Efficiency  of  the  Arc  with  Increasing  Size  of 
the  Light  Unit.  Relation  between  the  Efficiency  of  the  Arc 
Lamp  and  the  Current,  Arc  Length  and  Power,  at  Constant 
Arc  Length,  Constant  Current  and  Constant  Power.  The 
Conditions  of  Maximum  Efficiency 

Unlike  the  incaiideBcent  lamp,  in  which  the  efficiency  of  light 
production  remains  practically  constant  over  a  wide  range  of  units 
of  light,  the  efficiency  of  the  arc  lamp  increases  with  increasing 
power  consumption  and  thus  increasing  size  of  unit  of  light,  but 

— 1 

























1  1  1 

f   f   7   ■'   -r   ' 


Pro.  9. 

falls  off  with  decrcaaing  size  of  the  light  unit,  and  the  arc  lamp 
thus  is  essentially  a  large  unit  of  light,  but  for  small  units  does 
not  have  the  efficiency  to  compete  with  the  modern  incandescent 
lamps,  while  iaversely  for  large  units  it  reaches  efScienciee  of 
higher  magnitude  than  possible  with  incandescent  lamps. 

The  relation  of  the  efficiency  of  light  production  by  the  arc  to 
the  power  consumption  can,  with  fair  approximation,  be  calculated, 
especially  for  the  luminous  arc. 

For  instance,  in  the  series  direct-current  magnetite  arc,  the  ap- 
proximate equation  of  the  arc  voltage  is 




where  the  arc  length  1  is  given  in  inches,  and  the  approzinuite  ez- 
preeeioD  of  the  light  flux  *,  in  mean  spherical  candle-power,  is : 



(asenming,  as  approximately  the  case,  the  light  flux  as  proportional 
to  the  arc  length  and  the  current). 

For  constant  arc  length  1  then  follows,  from  equations  (1)  and 
(2),  for  different  values  of  current  i,  the  power  confiomption  p  =  ei 
and  the  efficiency  i;.  Curves,  for  the  arc  length  I  =  .7  inches,  are 
given  in  Fig.  9. 



























For  constant  current,  i=4  amperes,  curves  of  the  power  con- 
sumption and  of  the  efficiency  for  different  arc  lengths  are  given 
in  Fig.  10. 

As  seen,  with  increasiog  current  at  constant  arc  length,  and 
with  increasing  arc  length  at  constant  current,  the  efficiency  in- 
creases, but  the  piower  coneumption  also  increases. 

For  constant  power  consumption,  p  =  ei,  then  follow,  from  equa- 
tions (1)  and  (2),  values  of  arc  length,  arc  voltage  and  efficiency. 
They  are  plotted,  for  300  watts  and  500  watt«  power  consumption 
in  the  arc,  in  Fig.  H  as'  function  of  the  current.     As  seen,  with 

>y  Google 

Eleotsio  Illcuinantb 


increasing  current  at  constant  power  consumption,  the  efficiency 
increases  to  a  maximum — which  is  higher  with  the  500-watt  arc 
than  with  the  300-watt  arc — and  then  decreases  again. 
Determining  then  the  condition  of  maximum  efficiency,  as  f uuc- 



■   \ 













\  Y 





)  > 




\ ' 


,  1 



































tion  of  power  consumption  in  the  arc,  gives  the  curves  shown  in 
Fig.  13.  As  seen,  to  operate  at  maximum  efficiency,  with  increasing 
power  consumption  the  current  in  the  arc  and  the  arc  length  has 
to  be  increased,  while  the  arc  voltage  remains  nearly  constant. 
The  efficiencfy  rises  rapidly  with  increasing  power  consumption. 

>y  Google 


Illuminating  Enqineehikq 

SS.  Comparison  of  the  Arc  Lamp  and  the  Incandescent  Lamp 
As  seen  from  Fig.  13,  the  eSScienc;  of  the  timgsteii  incaDdescent 

lamp,  of  approximately  0.66  candle-power  per  watt,  is  reached  at 

70  watts  power  consumptioii. 

1   ! 




























^ — 















' — 







































Consideriag,  however,  that  the  efficiency  is  not  the  only  factor 
in  the  coat,  but  that  the  cost  of  attention,  trimming,  etc.,  also  en- 
terB,  furthermore,  that  at  the  lower  consumption  some  efficiency 
would  have  to  be  sacrificed  to  steadiness  by  increasing  the  current 
beyond,  and  therefore  reducing  the  afc  length  below  that  corre- 
sponding to  maximum  efficiency,  the  dividing  line  between  tungsten 

>y  Google 

Elbcthic  Illtjminants  153 

incaDdeBcent  lamp  and  maguetite  arc  lamp  for  aee  in  Btreet  ligbtiiig 
probably  lies  at  about  100  to  150  watts  power  consumption,  depend- 
ing on  the  individual  conditions;  below  tbia  the  tungsten  lamp 
above  the  magnetite  arc  is  more  efficient,  other  things  being  equal. 
Similar  relatione  exist  with  other  types  of  arcs :  with  the  flame 
carbon  arcs,  approximately  the  same  relations  would  exist — except 
that  the  numerical  values  of  efficiency  are  proportionally  changed — 
provided  that  the  size  of  the  flame  carbons  is  changed  proportional 
to  the  current.  If  the  same  size  of  flame  carbons  is  retained,  the 
^ciency  falls  of!  more  rapidly  with  the  decrease  of  current,  and 
increases  more  rapidly  with  its  increase,  due  to  the  change  of  the 
rate  of  evaporation.  However,  in  economical  comparison  with  the 
tungsten  lamp,  the  very  much  higher  cost  of  trimming,  with  the 
short-burning  flame  lamp,  would  probably  shift  the  dividing  line 
of  economical  use,  between  the  tungsten  lamp  and  the  arc  lamp, 
to  higher  values  of  power,  while  more  efficient  long-burning  lumi- 
nous arcs  would  shift  it  to  lower  values  of  power. 

Si.  The  Low-Pressure  Mercury  Arc  in  the  Glass  Tube.    The  High- 
Pressure  Mercury  Arc  in  the  Quartz  Tube,    Their  Charac- 

The  only  industrially  used  vacuum  arcs  are  the  mercury  arcs; 
the  low-pressure  mercury'arc,  operated  in  a  glass  tube,  and  the 
high-pressure  mercury  arc,  operated  in  a  quartz  tube. 

In  the  mercury  arc  the  terminal  drop  is  constant,  and  about 
13  volts,  while  the  stream  voltage  is  proportional  to  the  arc  length 
and  independent — ^within  a  certain  range — of  the  current,  but 
depends  upon  the  diameter  of  the  arc  tube,  and  on  the  vapor  pres- 
sure ;  it  increases  with  decreasing  tube  diameter  and  with  increasing 
vapor  pressure,  so  that  in  an  arc  tube  of  about  8  em.  diameter  and 
a  high  vacuum  it  is  as  low  as  0.5  volte  per  centimeter,  and  rises  to 
8  to  10  volts  per  centimeter  in  a  tube  of  1  cm.  diameter  at  a  mer- 
cury-vapor pressure  about  equal  to  atmospheric  pressure. 

The  mercury  arc  is  a  luminous  arc  and  stands  at  the  one  end  of 
a  series,  of  which  the  carbon  arc  stands  at  the  other  end ;  while  the 
latter  is  the  hottest  arc  the  former  is  the  coldest,  and  in  the  low- 
pressure  mercury  arc  in  a  glass  tube  the  temperature  of  the  arc 
stream  is  only  about  200°  to  350°  G. 

>y  Google 

164  Illdminatino  Enqikezring 

Like  all  arcs,  it  requires  a  Btarting  mechanism;  the  feeding  is 
done  by  tMndenBing  the  mercury  vapor  in  a  condensing  chamber, 
and  returning  it  to  the  negative  electrode  by  gravity. 

A  valuable  characteristic  of  the  mercury  arc  ia,  that  it  can  be 
bnilt  of  very  good  ^ciency  in  smaller  units  than  any  other  arc : 
as  low  as  80  to  100  watts. 

In  the  high  vacuom  of  the  mercury  arc  in  the  glass  tube  the 
arc  length  is  very  great  at  moderate  voltages,  and  mercury  arc 
tubes  of  over  3  feet  length  are  operated  on  110-volt  circuits;  in 
the  quartz-tube  arc,  due  to  the  high  vapor  pressure,  the  arc  length 
is  short  and  comparable  with  that  of  other  arcs  of  the  same  voltage; 
an  arc  length  of  8  inches  requires  a  330-volt  supply. 

Like  all  arcs,  the  mercury  arc  requires  a  steadying  resistance  on 
constant-potential  supply  circuits. 

The  light  of  the  mercury  arc  has  the  advantage  of  great  steadi- 
ness and  higlt  efBciency,  but  the  disadvantage  of  a  green  color, 
which  is  almost  entirely  deficient  in  red  rays,  and  therefore  greatly 
distorts  colors. 


35.  Comparison  of  ike  Variovs  Forms  of  Incandescent  Lamps  and 
Arc  Lamps  Regarding  Their  Sensitivity  to  Variation  of  the 
Electric-Power  Supply: 

The  various  forms  of  electric  illnminants  must  find  their  place 
in  existing  electric  distribution  systems,  either  constant-potential 
or  constant  current.  TSo  electric  circuit,  however,  maintains  abso- 
lutely constant-potential  respectively  constant  current,  but  fluctua- 
tions of  greater  or  lesser  extent  occur,  and  it  thus  is  of  importance 
to  know  the  sensitivity  of  the  illuminants  to  variations  of  the  sup- 
ply circuit.  Since  the  limit  of  sensitivity  of  the  human  eye  for 
changes  of  light  flux  is  not  much  below  2  per  cent,  a  sudden  change 
of  light  flux  of  5  per  cent  is  not  seriously  objectionable,  and  a  grad- 
ual change  even  of  30  per  cent  is  hardly  appreciable.  The  per- 
missible range  of  sudden  and  of  gradual  variation  of  the  electric- 
supply  system,  and  inversely,  in  a  system  of  given  regulation,  the 
degree  of  satisfactorinese  of  an  illuminant  would  then  he  deter- 
mined by  the  ratio  of.  the  change  of  light  flux  to  the  change  of  the 
electric  supply  causing  it. 

In  the  following  arc  given  a  number  of  approximate  values  of  the 

>y  Google 

Elbcieic  Ilitjminamtb  155 

percentage  change  of  light  fluz  of  various  electric  iUumiDants, 
reflultiiig  irom  a  change  of  the  electric-supply  voltage,  current  or 
power  by  1  per  cent. 

In  the  calculation  of  the  incaadeecent  lamp  values,  the  curves 
of  Fig.  3  have  been  used;  the  arc-lamp  values  are  calculated  from 
the  characteristic  curves  of  the  arc,  equations  (1)  and  (2).  They 
depend  to  a  greater  or  less  extent  on  arc  length,  per  cent  of  steady- 
ing resistance,  etc.,  and  thus  can  be  approximate  only. 

AppBOxncATB  VAsuTion  or  Casn^'Pqweb,  nt  Fib  Ckht 

For  1  per  cent  variation  of —  Power  Voltage   Current 

Incandescent  lampe: 

Treated  carbon  filament 2.S  5.6  5.6 

Q«m  filament  2.5  4.45  5.7 

Tungsten  filament    2.33  3.75  6.25 

Constant  current  arcs — 75  volts  per  lamp: 

Magnetite  arc  lamp 1.42  ...  1.0 

Flame  carbon  arc,  dUterentlal  control...  1.7  3.4  3.4 

Flame  carbon  arc,  sbnnt  control 1.56  . . .  1.56 

Constant  potential  arcs — 110-voIt  supply,  33 
per  cent  steadying  resistance: 

Mercnrr  arc  76  8.0  1.0 

Magnetite  arc  (constant  arc  lengtb) 88  7.5  1.0 

Flame  carbon  arc,  diOerentlal  control. .  1.7  3.4  3.4 

Flame  carbon  arc,  sbunt  control.......  1.17  4.7  1.55 

Flame  carbon  arc,  series  control 2.66  2.65 

Incandescent  lamps  in  general  are  much  more  sensitive  to 
changes  of  supply  than  are  lamps,  that  is,  require  a  closer  regula- 
tion of  the  electric  supply. 

Especially  the  arcs  with  constant  fiied  arc  length,  as  the  mag- 
netite arc  and  the  mercury  arc,  are  very  little  sensitive  to  changes 
of  current,  while  the  arc  lamp  with  floating-feed  and  differential 
control  is  most  sensitive  to  current  changes,  though  less  so  than 
the  incandescent  lamp. 

Inversely,  on  constant-potential  supply,  the  constant-pressure 
arc  with  fixed  arc  length  shows  tiie  greatest  sensitivity  to  voltage 
variations.  This  depends  on  the  amount  of  steadying  resistance, 
and  decreases  with  increasing  steadying  resistance,  while  with  less 
than  33  per  cent  steadying  resistance  the  sensitivity  increases  so 
that  the  arc  soon  becomes  inoperative.  The  least  aenaitivity  on 
multiple  circuit  is  afforded  by  series  control. 

>y  Google 

166  Illdhinatino  Enoinb^nq 

It  thus  vould  foUow,  that  the  incandescent  lamps  with  higb- 
poeitire  temperature  coefficient  have  an  advantage  on  constant- 
potential  supply,  but  a  corresponding  disadvantage  on  constant- 
cnirent  supply.  On  constant-current  circuits  the  arc  lamps  with 
filed  arc  length,  as  the  magnetite  and  mercury,  would  be  most  con- 
stant in  their  light  production,  and  next  thereto  the  lampe  with 
shunt  control,  vhile  inversely  on  coDBtant'pot«ntial  circuits  these 
two  operating  mechanisms  are  most  sensitive  to  voltage  variations. 

>y  Google 



By  Alex.  C.  Hdmpheets 



Scope  ot  lecture. 

Brief  reference  to  the  special  character  of  the  lllnmlnants  con- 

Petroleum  and  by-producta — kerosene. 
Illumlmmta  considered. 
Plntsch  gaa. 

Brief  history. 
Present  extent  of  use. 
How  made. 

Eixternally  beated  retorts — low  pressure. 
Internally  heated  generators,  low  pressure  and  high  press- 
Special  characteristics. 
How  employed. 

Lighting  of  railroad  cars. 
Lighting  of  huoys,  beacons,  lightships,  etc. 
Special  appliances — especially  pressure  regulators,   car   lamps 

and  buoy  lanterns. 
Plntsch  system  provides  for  sclentlflc  diatrlbnt^on  of  light. 
Carburetted-alr  gas. 
Brief  history. 
Present  extent  of  use. 
Produced  from  certain  hydrocarbons. 
Special  charscterlstlcs. 
How  employed. 

Isolated  plants. 
Town '  plants. 
Special  appliances. 

Brief  history. 
Present  extent  of  use. 
Carbide  of  calcium — CaC,. 
How  produced  from  CaCt. 
Special  characteristics. 
Special  precautions. 

>y  Google 

158  Illdminatimg  Enqineebino 

How  etnplored. 

Isolated  pluits. 

Town  plants. 

Portable  lamps  and  lanterns. 
Special  appliances. 

Those  who  unselfiably  have  token  the  initiative  in  arranging  for 
this  course  of  lectures  on  the  science  and  art  of  illuminating  engi- 
neering, sparing  neither  time,  thought  nor  efFori:,  should  be  exempt 
from  all  unfriendly  mticism.  Then  it  should  be  understood  that 
in  36  lectures,  treating  of  19  dirieions  of  the  subject,  not  more 
can  be  done,  certainly  with  regard  to  some  of  the  divisions  and  sub- 
divisions, than  point  the  way  to  those  who  desire  to  devote  them- 
selves seriously  to  the  study  and  practice  of  illuminating  eogi- 

This  lecture  is  one  of  two  which  are  expected  to  cover  "  Oas  and 
Oil  lUuminonts."  To  Professor  Whltaker  has  been  assigned  the 
open  flame  and  tlie  incandescent  mantle,  and  to  me  Fintsch  gas, 
carburetted-air  gas  and  acetylene.  It  must  be  apparent  that  the 
hour  and  a  half  allotted  to  each  lecture  is  entirely  inadequate  for 
a  comprehensive  -consideration  of  the  three  systems  named. 

It  should  be  understood  that  this  lecture,  notwithstanding  the 
Bub-title  of  Fart  V,  is  not  intended  to  cover  coal  gas,  water  gas 
or  natural  gas,  which  are  the  gas  illuminants  most  generally  dis- 
tributed and  which,  especially  if  taken  together,  furnish  more 
artificial  illumination  than  electric  light  together  with  the  three 
illuminants  here  to  be  coiisidered. 

As  we  proceed  it  will  be  seen  that  Fintech  gas,  carburetted-air 
gas  and  acetylene  do  not  compete  with  coal  gas,  water  gas  or 
natural  gas,  but  are  employed  where  these  are  not  commercially 
available  or  obtainable,  or  where  a  special  character  of  service  is 

Of  theee  three  sources  of  artificial  illumination,  two,  namely, 
Pintsch  gas  and  air  gas,  are  made  from  oil.  Fintsch  gas  is  pro- 
duced by  the  destructive  distillation  of  petroleum  oil.  It  is  not 
to  be  understood  that  the  manufacture  of  oil  gas  is  confined  to  this 
process.  Oil  gas  has  been  employed  to  a  considerable  extent  in 
the  United  States  and  Europe  for  the  iltumination  of  small  towns, 
factories,  etc.     Oil  gas  was  bo  employed  before  it  was  applied  in 

>y  Google 

0a5  AMD  Oil  Illuicjnants  159 

compressed  form  to  car  lighting,  and  patents  for  oil-gas  manu- 
facture were  granted  in  the  earlj  part  of  the  last  century. 

In  some  few  cases  compressed  oil  gas  has  been  employed  for 
the  lighting  of  small  towns,  the  compressed  gas  being  delivered 
in  cylisders  instead  of  through  mains  laid  in  the  thoroughfares; 
these  uudertakingB  were  short-lived.  In  Scotland,  prior  to  the 
production  of  petroleum  in  large  quantities,  a  high  candle-power 
gas  was  made  from  oil  distilled  from  rich  shales. 

By  far  the  moat  extensive  use  of  oij  in  the  making  of  illumi- 
nating gas  has  been  in  the  manufacture  of  carburetted  water  gas. 

Although  the  title  of  Division  V  includes  oil  as  an  illuminant, 
neither  Professor  Whitaker  nor  I  are  expected  to  consider  it  as 
a  direct  source  of  illomination.  When  we  realize  that  refined 
kerosene  oil  is  used  throughout  the  whole  civilized  world,  com- 
peting with  all  other  sources  of  artificial  illumination  covered  in 
these  lectures  and  relied  upon  where  these  are  not  to  be  found, 
this  well  serves  to  illustrate  the  fact  that  the  36  lectures  cannot 
be  made  to  cover,  even  superficially,  the  whole  field  of  artificial 

In  this  connection,  let  me  refer  you  to  "Petroleum  and  Its 
Products,"  by  Sir  Boverton  Redwood,  Sd  Edition,  1906,  two  vol- 
umes, published  by  Charles  Griffen  &  Company,  London.  This 
work  is  most  valuable  in  itself,  and  also  for  the  extensive  bibli- 
ography annexed. 

Redwood  gives  an  interesting  and  instructive  history  of  the 
petroleum  iuduBtry,  beginning  with  a  reference  to  an  account  writ- 
ten by  Herodotus,  450  B,  C,  of  a  well  producing  "  asphalt,  salt 
and  oil,"  Petroleum  is  now  being  produced  in  all  parts  of  the 
world,  and  in  many  places  in  large  quantities.  Vast  quantities 
have  been  discovered  recently  in  Mexico,  this  oil  being  unusually 
rich  in  asphalt. 

The  production  of  oil  from  coal  and  shale  is  of  interest,  especially 
as  much  of  scientific  and  practical  value  was  learned  in  the  course 
of  the  evolution  of  the  process  and  apparatus  employed. 

I  presume  that  other  of  the  lectures  included  in  this  course 
win  give  some  account  of  the  several  by-products  from  the  dis- 
tillation of  petroleum  which  have  been  and  still  are  employed  in 
manufacturing  water  gas  and  enriching  coal  gas,  and  of  the 
commercial  utilization  of  these  and  other  by-producte  of  kerosene 

>y  Google 

160  Illomimatiho  Enginbbbino 

manofactare  which  has  enabled  the  great  oil  companies,  eepeciaU; 
the  Standard,  to  produce  kerosene  at  a  minimum  cost. 

The  internal  combustion  engine,  particularly  as  used  in  motor 
cars  and  motor  boats,  has,  within  the  last  few  years,  developed 
an  extensive  and  rapidly  growing  market  for  the  more  volatile  of 
the  distillates  which,  together  with  the  so-called  gas  oil,  were 
almost  a  drug  on  the  market  25  years  ago. 

Some  idea  of  the  magnitude  and  growth  of  the  production  of 
kerosene  oil  is  found  in  the  records  for  1906  and  1909.  In  the 
former  year  the  total  production  of  kerosene  is  estimated  by  the 
Standard  Oil  Company  to  have  been  48,000,000  barrels  or  2,016,- 
000,000  gallons;  and,  in  1909,  53,000,000  barrels  or  2,336,000,000 
gallons,  an  increase  in  3  years  of  more  than  10  per  cent. 

Pintsch  Qas 

Pintsch  gas  is  bo  named  after  Julius  Pintsch,  of  Germany,  the 
founder  of  the  great  finu  of  that  name. 

Pintsch  gas  is  made  by  the  destructive  distillation  of  petroleum 
or  other  mineral  oil  in  retorts  (cast  iron  or  clay)  externally  and 
continuoosly  heated,  or  in  generators  filled  with  fire-brick  checker- 
work,  internally  and  intermittently  heated.  The  product  is  in 
great  measure  a  fixed  gas,  principally  methane  (CH4)  and  heavy 
hydrocarbons  with  a  very  small  volume  of  hydrogen.  The  oil  gas  as 
so  made,  unlike  water  gas,  is  not  diluted. 

The  Pintsch  system  was  originally  developed  for  the  lighting 
of  railway  passenger  cars.  In  the  early  days  of  railroading  some 
trains  were  not  run  after  dark,  and  in  many  cases  where  the  trains 
were  run  through  the  night  hours  it  was  not  considered  necessary 
to  furnish  artificial  illimiination.  The  illuminants  first  employed 
were  candles  and  oil  lamps. 

In  1866  experiments  were  begun  in  Germany  in  the  lighting  of 
railway  carriages  with  coal  gas.  It  happened  that  in  the  TTnited 
States  the  Beading  Railroad  also  began  to  light  some  of  its  cars 
with  coal  gas  in  the  same  year. 

By  reason  of  the  limited  space  available  on  railrojid  cars  for 
the  storage  of  the  illuminant,  city  gas  was  found  to  be  too  bulky, 
and  this  suggested  that  the  gas  should  be  of  comparatively  high 
candle-power  and  be  compressed  into  a  greatly  reduced  volume. 
This  led  Pintsch  to  turn  his  attention  to  gas  made  from  coal  oil 
and  petroleum. 

>y  Google 

Gas  and  Oil  Illdminants  161 

As  compared  with  coal  gas  a  doable  advaatage  was  secured  b; 
the  eubBtitntion  of  compressed  oil  gas  for  railroad  lighting  and 
similar  service,  for  the  oil  gas,  in  addition  to  an  initial  illumi- 
nating  power  three  or  four  times  higher  than  that  of  coal  gas, 
snffere  a  loss  in  illuminating  power  due  to  compression  of  only 
one-third  to  one-half  of  that  of  coal  gas.  This  loss  in  compressing 
Fintsch  gas  to  10  atmospheres  is  only  about  10  per  cent. 

The  advantages  of  compressed  oil  gas  so  markedly  apparent  in 
its  application  to  the  lighting  of  railway  passenger  cars  were  in 
even  greater  degree  found  to  be  applicable  to  the  lighting  of  buoys, 
beacons,  stake  lights  and  lightships.  In  the  late  seventies  Fintsch 
tomed  his  serious  attention  to  the  development  of  a  system  to 
satisfy  the  varying  demands  of  lighthouse  authorities  and  met  with 
prompt  success. 

For  the  storage  of  compressed  gas  at  the  works  Fintsch  developed 
a  process  of  welding  by  which  were  produced  storage  cylinders  of 
large  capacity  free  from  eeams  or  rivets.  These  seamless  cylinders 
are  now  manufactured  to  a  maximum  size  of  8  feet  in  diameter 
by  33  feet  in  length.  For  lighthouse  work  welded  buoys  were 
made  of  the  several  required  shapes,  the  body  of  the  buoy  serving 
as  a  holder  for  the  compressed  gas.  DifGcult  as  was  the  welding 
of  the  storage  cylinders,  the  welding  of  the  buoy  bodies  was  far 
more  difBcult.  The  application  of  this  welding  system  to  the  manu- 
facture of  buoys  was  particularly  useful,  because  by  eliminating 
riveted  joints  there  was  obtained  the  necessary  strength  and  ca* 
pacity  with  the  minimum  of  weight,  and  consequently  the  maxi- 
mum of  buoyancy. 

Fintsch  also  devised  a  wind-  and  wave-proof  lantern  which 
demonHtrated  its  ability  to  maintain  a  steady  and  constant  light 
under  the  severest  weather  conditions. 

In  the  use  of  compressed  gas  for  car  lighting,  and  still  more  for 
lighthouse  service,  it  was  necessary  to  develop  a  pressure  regulator 
capable  of  receiving  the  gas  at  a  pressure  of  from  150  pounds  to 
1  pound  per  square  inch,  and  delivering  it  constantly  to  the  burner 
supply  pipe  at  such  a  reduced  pressure  as  might  he  required  for 
the  most  efficient  operation  of  the  particular  burner  employed. 
To  meet  this  requirement  Fintsch  invented  a  regulator  which,  prac- 
tically without  change,  has  met  snccessfally  all  the  requirements 
of  nearly  40  years  of  the  most  varied  and  exacting  service. 

>y  Google 

168  Ili.lmjnatino  Enqineebino 

As  far  as  I  know,  and  I  had  a  very  pereonal  experience  with  this 
regulator  from  the  Iatt«r  part  of  the  year  1881  to  the  end  of  1884, 
no  g»B,  compressed  or  uncompreeaed,  is  sapplied  to  the  point  of 
ignition  tinder  more  uniform  preaenre  than  the  gas  supplied  by  the 
Pintsch  system.  I  lay  particular  Btress  on  this  point  because  I 
know  that  qnesdons  have  frequently  been  raised  as  to  the  com- 
plete reliability  of  such  an  instrument  for  constant  and  accurate 
regulation  within  narrow  limits  of  outlet  pressure. 

I  will  describe  bri^y  a  couple  of  tests  which  occurred  under 
my  own  eye  about  the  year  1883.  The  first  was  a  test  by  the 
representatives  of  the  ITnited  States  Lighthouse  Board  of  a  Pintsch 
regulator  and  buoy  lantern  in  competition  with  similar  appliances 
of  a  rival  system.  The  claim  was  made  for  the  latter  system  that 
operating  under  600  pounds  pressure  a  decided  advantage  was 
secured  by  reason  of  the  longer  supply  of  light  thus  obtained  from 
the  one  filling  of  the  gas  reservoir.  Although  the  Pintsch  gov- 
ernor was  only  tested  and  guaranteed  for  a  pressore  of  150  pounds, 
to  meet  the  claims  of  the  competitor,  the  Pintsch  Ck>mpany'B  rep- 
resentatives offered  to  subject  this  governor  to  the  600  pounds 
pressure.  Upon  examination  it  was  found  that  the  storage  holder 
of  the  rival  concern  was  charged  only  to  300  pounds  instead  of 
600  pounds  as  claimed.  U-water  gauges  were  connected  to  the 
pipes  connecting  the  governor  outlets  to  the  lanterns.  The  inlet 
preeeuree  to  both  governors  were  first  adjusted  at  1  pound,  and  the 
corresponding  outlet  pressures  as  indicated  by  the  U  gauges  were 
accurately  observed  and  marked.  By  a  quick  movement  of  the 
hand  the  full  pressure  of  300  pounds  was  admitted  to  the  inlet  of 
each  of  the  governors.  In  the  case  of  the  Piniach  governor  the 
fiuctuation  of  the  governed  pressure,  as  indicated  by  the  U  gauge, 
was  found  to  be  less  than  one-tenth  of  an  inch  of  water  and  the 
fiames  were  not  affected ;  whereas  in  the  other  case  the  water  was 
blown  out  of  the  U  gauge  and  struck  the  ceiling  of  the  room  in 
which  the  test  was  being  made,  and  the  light  was  extinguished. 
In  this  test  the  lafitems  were  also  subjected  to  conditions  repre- 
senting a  hurricane,  the  wind  effect  being  obtained  by  the  use  of 
an  air  blower  and  the  washing  of  the  waves  by  water  delivered 
from  a  2-inch  hose  under  heavy  pressure  against  all  parts  of  the 
lanterns.  The  Pintsch  lantern  remained  lighted  white  the  other 
was  extinguished. 

>y  Google 

Gas  and  Oil  Illdminants  163 

The  other  case  also  eerved  to  show  the  reliability  of  the  governor 
and  the  buoy  lantern  under  extraoTdinaiily  severe  conditione.  Fol- 
lowing a  heavy  Btonn  it  was  reported  that  one  of  the  buoya  recently 
anchored  in  New  York  Harbor  had  been  extinguished.  With  the 
Ligbtiioase  Board's  district  inspector,  I  made  a  personal  investi- 
gation. When  we  arrived  at  the  buoy,  from  the  tender  it  appeared 
that  the  light  was  extinguished.  Determined  that  there  should  bo 
no  question  as  to  the  accuracy  of  the  record  I  climbed  into  the 
cage  surrounding  the  lantern  of  the  buoy.  Opening  the  lantern 
I  found  that  the  set-ecrew  which  regulates  the  size  of  the  flames 
bad  been  screwed  down  hard  so  that  the  amount  of  gas  leaking  by 
was  only  sufficient  to  produce  flames  practically  non-luminous, 
with  the  result,  even  after  the  lantern  was  opened,  that  those  on 
the  lishthonse  tender  could  not  see  the  flames.  That  the  record 
should  not  depend  upon  my  word  I  demonstrated,  by  lighting  a 
piece  of  paper  at  the  flames,  that  the  light  was  not  extinguished. 
The  delicacy  of  action  of  the  governor  and  the  efficiency  of  the 
lantern  can  be  understood  when  I  say  that  the  flames  were  so 
small  that  after  lighting  the  paper  I  extinguished  them  by  fanning 
them  with  a  single  motion  of  my  hand. 

While  the  use  of  pressure  regulators  in  connection  with  the  dis- 
tribution of  city  gas  introduces  unnecessary  complications,  in  the 
case  of  such  special  service  as  that  which  the  Pintsch  system  has 
to  perform,  which  necessarily  demands  special  appliances  designed 
and  constructed  to  operate  with  mathematical  accuracy,  no  addi- 
tional complication  is  introduced  provided  the  regulator  is  com- 
pletely dependable.  Given  a  gas  delivered  at  a  pressure  well  above 
that  required  for  maximum  efficiency  with  any  illuminating  bamer, 
an  important  economic  advantage  is  secnred  by  the  use  of  a  gov- 
ernor which  can  be  relied  upon  to  reduce  this  excessive  pressure 
to  any  desired  point.  This  is  well  illustrated  in  the  application 
of  the  Pintech  system  to  mantle  lighting,  as  later  to  be  explained. 

Between  the  years  1870  and  1880  the  Pintsch  system  of  lighting 
was  introduced  to  a  very  considerable  extent  on  the  Prussian  State 

Pintsch's  first  United  States  patents  were  taken  out  between 
the  years  1870  and  1880. 

In  the  year  1880  the  Pintsch  system  was  brought  to  the  United 
States,  being  first  applied  in  lighting  tlie  sound  steamers  of  the 
Stonington  Line  and  the  cars  of  the  connecting  line  of  the  New 

>y  Google 


York,  ProTidence  and  Boston  Bailroad,  now  part  of  the  Nev 
Haven  system.  The  Pintech  plant  for  supplying  the  boats  and 
cars  vas  located  at  Stonington,  Conn. 

The  next  railroad  to  adopt  the  light  was  the  Erie,  the  worla 
for  making  and  compreBsing  the  gas  being  built  in  &e  railroad's 
yards  in  Jersey  City.  Shortly  thereafter  a  similar  plant  was  built 
at  Weehawken  for  the  West  Shore  Railroad,  and  practically  all 
of  its  passenger  cars  then  being  built  were  equipped  for  the  new 

At  first  the  policy  of  the  United  States  Pintsch  Company  was 
to  induce  each  railroad  adopting  the  system  to  own  and  operate 
its  own  gas  works,  one  or  more.  This  would  have  led  to  unneces- 
sary multiplication  of  gas  works  throughout  the  country.  The 
policy  was  persisted  in  for  a  number  of  years,  and  in  this  is  to  be 
found  the  reason  why  the  system  made  but  little  progress  in  the 
United  States  during  the  first  years  of  the  American  company's 
existence.  It  was  not  until  a  new  element  came  into  control  of 
the  United  States  Pintsch  Company  that  this  policy  was  aban- 
doned and  more  rapid  progress  made,  the  company  undertaking 
the  building  of  gas  works  and  the  supply  of  compressed  oil  gas 
to  the  railroads  adopting  the  system.  While  now  some  of  the 
railroads  own  and  operate  plants  built  for  them  by  the  company, 
the  Pintsch  Company  owns  and  operates  works  of  its  own  through- 
out the  United  States,  Canada  and  Mexico,  in  many  cases  sup- 
plying several  roads  from  the  same  plant.  In  a  number  of  cities 
Pintech  gas  is  manufactured  and  distributed  to  the  railroads  by 
the  local  gas  company  operating  in  partnership  with  the  Pintsch 
Company  and  the  railroads  served. 

The  Pintsch  system  is  in  use  practically  throughout  the  civilized 
world.  Up  to  date  about  180,000  cars  in  all  are  equipped  for 
Pintsch  light. 

Up  to  April  30, 1909,  there  were  in  service  in  the  following  coun- 
tries, namely.  Great  Britain,  Germany,  Holland,  Belgium,  France, 
Portugal,  Denmark,  Euasia,  Tunis,  Sweden,  Austria,  Italy,  United 
States,  Brazil,  Argentine  Republic,  Uruguay,  Egypt,  India,  South 
Africa,  Canada,  Australia,  Wew  Zealand,  Algiers,  Spain,  Japan  and 
China,  buoys,  1947;  beacons  and  stake  lights,  485;  lightships,  96; 
these  being  supplied  from  77  charging  stations. 

A  later  return,  covering  the  lighthouse  service  for  the  United 
States  and  Canada,  ihows  that  on  August  10,  1910,  the  number 

>y  Google 

Qab  asd  Oil  Iixuuihants  165 

of  buoys  in  service  in  these  two  cotmtries  was  461,  an  increase  of 
30  in  the  15  months. 

Up  to  June  1,  1910,  there  were  in  the  United  States,  Canada 
and  Mexico  93  Fintech  gas  works,  supplying  compiessed  gas  to 
360  railway  Btatione.  In  the  same  territory,  up  to  Jannaiy  1, 
1910,  the  number  of  cars  equipped  for  the  Pintsch  system  was 

During  the  last  few  years  the  Fintsch  system  has  been  further 
developed  to  secure  the  additional  advantages  to  be  obtained 
through  tiie  use  of  incandescent  mantle  burners.  Up  to  date  there 
are  mantle  lamps  installed  in  railway  cars  as  follows :  France,  about 
96,000;  Great  Britain,  about  61,000;  other  European  countries, 
about  159,000;  United  States,  Canada  and  Mexico,  about  80,000; 
total,  about  388,000.  These  figures  represent  mantle-lamp  equip- 
ment for  about  55,400  cars. 

Pintsch  gas,  as  has  been  stated,  is  obtained  by  the  destructive 
distillation  of  oil.  In  the  early  days  of  the  system  oil  produced 
by  the  distillation  of  coal  or  shale  was  used.  Of  late  years  crude 
petroleum  oil  and  its  distillates  have  been  employed,  markut  con- 
ditions controlling  the  choice.  The  crude  oil  can  be  satisfactorily 
employed  and  was  at  one  time  largely  used.  To-day  market  con- 
ditioDf  generally  lead  to  the  use  of  a  distillate. 

At  first,  and  until  recent  years,  the  gas  was  manufactured  only 
in  cast-iron  retorts  externally  heated.  Much  of  the  gas  is  still 
80  made.  Two  retorts  are  set  in  each  "  bench  "  or  furnace,  the 
two  retorts  being  so  connected  at  their  back  ends  that  the  gas  passes 
from  one  to  the  other.  The  oil  is  introduced  at  the  front  of  the 
upper  retort  and  falls  upon  a  removable  sheet^iron  tray  which  col- 
lects most  of  the  carbonized  oil.  The  gas  and  vapor  produced  in 
the  upper  retort  pass  down  to  the  back  of  the  lower  retort,  and  so 
through  to  the  front  of  the  bench,  passing  by  a  decension  pipe  to 
the  hydraulic  main  located  below  the  fioor  of  the  house.  Issuing 
from  the  hydraulic  main  the  gaa  and  vapor  pass  through  a  dry 
scrubber,  condenser,  purifiers  and  station  meter,  and  are  collected 
in  the  low-pressure  storage  holder.  From  the  storage  holder  the 
gas  is  drawn  by  a  compressor,  compressed  into  one  or  more  of 
the  welded  cylinders  before  described,  and  is  then  ready  for  dis- 
tribution through  the  high-pressure  pipes  to  the  cars  or  transport 
holders.  All  necessary  precautions  are  taken  to  trap  the  liquid 
hydrocarbon  thrown  down  by  the  process  of  compression,  the  object 

>y  Google 

166  Illduinaiinq  Enouibbbinq 

being  to  obtain  a  thoroughly  dry  gaa,  which  resnlt  is  eecnred  to  a 
remarkable  degree. 

The  early  German  practice  limited  the  compression  between  8 
and  10  atmospheres.  The  more  recent  practice,  eapedally  in  the 
United  States,  is  between  12  and  14  atmospheres. 

Particularly  in  connection  with  the  larger  plants,  day  retorta, 
as  used  in  coal-gas  manufactare,  came  into  use.  This  change,  by 
reason  of  the  porosity  of  the  day  retorts,  has  made  necessary  the 
employment  of  exhausters  to  draw  the  gas  from  the  retorts  and 
push  it  on  through  the  other  parte  of  the  plant  to  the  storage 

When  the  gas  is  distilled  in  day  retorts  the  distillation  is  com- 
pleted in  a  single  retort,  the  oil  being  introduced  through  a 
wrought-iron  pipe  carried  through  the  front  of  the  retort,  extending 
nearly  its  entire  length  and  open  at  the  end.  The  oil,  gas  and 
vapor  issue  from  the  open  end  of  the  pipe  and  return  through 
the  retort  to  the  front.  The  gas  and  vapor  issue  from  the  front 
of  the  retort  and  pass  by  an  ascension  pipe  to  the  hydraulic  main 
located  on  the  top  of  the  bench,  and  from  there,  as  before  described, 
t«  the  storage  holder. 

Some  years  ego  experiments  were  undertaken  to  determine  if 
greater  economy  could  be  secured  by  distilling  the  oil  in  generators 
internally  fired.  This  is  necessarily  an  intermittent  process  and 
so  is  markedly  differentiated  from  the  continuous  retort  process. 

The  generator  consists  of  a  steel  shell  6  feet  in  diameter  and 
about  12  feet  in  height.  It  is  lined  with  fire-brick  and  the  interior 
is  divided  into  two  compartments,  a  smaller  lower  compartment 
which  serves  as  a  combustion  chamber,  and  a  larger  upper  com- 
partment which  is  filled  with  fite-brick  checker-work  nearly  up  to 
the  top  of  the  shell ;  this  upper  diamber  terminates  in  a  cone,  upon 
the  top  of  which  is  a  stack  valve. 

The  tar,  obtained  as  a  by-product  from  the  distillation,  is  used 
as  fuel.  This  is  injected  into  the  combustion  chamber  bdow  the 
checker-work  by  means  of  a  liquid-fuel  burner.  A  mechanical 
blower  produces  the  necessary  forced  draft. 

As  soon  as  the  generator  has  been  "  blown  "  to  its  proper  working 
temperature  the  tar  fuel,  steam  and  air  are  shut  off  and  the  stack 
valve  is  closed.  Gas  oil  under  pressure  is  then  injected  throu^ 
three  oil  nozzles  located  in  the  top  of  the  generator  and  the  finely 
divided  oil  is  thrown  upon  the  checker  brick.     The  oil  vapor  so 

>y  Google 

Gas  akd  Oil  Illuhikantb  167 

fonned  paeees  down  through  the  heated  checker  bricks  and  is  so 
decomposed,  the,  gas  produced  finally  issuing  from  the  generator 
through  the  take-off  pipe  located  in  the  side  of  the  combustion 

The  cycle  is  ^vided  into  a  heating  period  ("blow")  of  about 
5  minntes,  and  a  gas-making  period  ("run")  of  from  6  to  8 

The  rate  of  flow  of  oil  is  regulated  by  the  so-called  trovel  test 
which  the  gas  maker  applies  at  short  intervals.  This  test  con- 
sists in  permitting  s  fine  jet  of  the  hot  gas  to  impinge  upon  the 
polished  blade  of  a  mason's  trowel,  the  figure  made  upon  the  trowel 
by  the  condensed  tar  indicating  to  the  practiced  eye  the  amonnt 
of  condensable  vapoi  in  the  gas.  With  care  in  operation  the  gas 
is  obtained  of  quite  uniform  quality  in  spite  of  the  gradually  de- 
croaaing  tranperature  of  the  generator. 

About  1600  feet  of  the  gas  are  made  per  "  run." 

The  gas  after  leaying  the  generator  is  dry  scrubbed  and  cooled, 
and  is  then  collected  in  a  "  relief  "  holder.  From  the  holder  it  is 
drawn  by  the  compressor  through  the  purifiers  and  station  meter, 
and  then  compressed  into  the  high-presenre  storage  holders  at  a 
pressure  of  about  14  atmospheres. 

It  is  found  that  by  this  .method  of  intermittent  distillation  in 
internally  fired  generators  a  gas  can  be  obtained  about  10  per  cent 
higher  in  candle-power  than  by  the  retort  process,  with  the  at- 
tendant advantages  of  largely  reduced  floor  space,  reduced  cost  of 
construction,  and  lower  manufacturing  cost  due  to  economy  in 
fuel,  labor  and  repairs. 

In  order  to  simplify  the  apparatus  and  reduce  the  investment  at 
stations  where  the  output  is  small,  a  still  later  development  is  the 
generation  of  the  gas  under  the  pressure  required  for  delivery. 
(See  Fig.  1.) 

To  withstand  this  heavy  pressure,  the  generator  shell  is  con- 
structed of  heavy  steel  plat^.  The  shell  is  divided  as  follows; 
At  the  bottom,  a  combustion  chamber;  above,  a  chamber  filled  with 
fire-brick  checker- work ;  above  this,  a  space  for  the  oil  sprays;  and 
above  this,  another  chamber  filled  with  checker-work. 

The  "  blow  "  and  "  run  "  occur  as  in  the  low-pressure  generator. 
In  order,  however,  to  check  the  rapid  decomposition  of  the  oil, 
which  would  otherwise  occur  when  operating  under  heavy  pressure, 
steam  is  injected  with  the  oil  into  the  generator.    The  ateam  so 

>y  Google 


Fio.  1. — PIntach   Gas  Hlgti-Presaure   Generator  Plant. 

>y  Google 

Gas  and  Oh.  iLLOMiNANTa 


Fia.  la. — PlntBch  Gas  High-Pressure  Generator  Plant. 

>y  Google 


Illdminatinq  Ehqineehinq 

used  acta  as  a  carrier  and  protector  for  the  oil  vapor  and  gas,  and 
does  not  react  with  the  carbon  of  the  oil  to  produce  water  gas,  the 
temperature  of  the  generator  being  too  low  for  this  reaction. 

The  eteam  enters  the  generator  at  the  top,  being  superheated  in 
passing  down  through  the  upper  checker-work.     Coining  to  the 

Fra.  lb. — Plntscb  Gaa  Hlgh-FresBure  Oenerator  Plant 

intermediate  chamber  it  meets  and  mingles  with  the  finely  divided 
oil,  and  then  steam  and  oil  vapor  pass  down  through  the  second 
checker-work  wherein  the  oil  vapor  is  decomposed.  The  gas  and 
superheated  steam  finally  leave  the  bottom  of  the  generator  through 
the  take-off  pipe  in  the  side  of  the  combustion  chamber.    The  gas 

>y  Google 

Gas  and  Oil  iLLCMiNAKTa  171 

and  highly  Buperheated  steam  are  then  dry  scrubbed  and  cooled, 
and  the  gas,  tar  and  water  are  charged  into  the  first  storeholder 
imder  a  preesnre  of 'about  14  atmospheres.  Passing  from  the  first 
storeholder  the  gae  is  purified  under  pr^sure  and  is  then  stored 
in  other  high-presenre  storeholders. 

Before  the  next  "  blow  "  the  gse  and  oil  vapor  which  remain  in 
the  generator  under  presanre  are  displaced  by  means  of  eteam  at 
sufficient  pressure,  being  thus  forced  through  the  scrubbing  and 
cooling  devices  and  into  the  storeholders. 

This  high-pressure  plant  is  more  simple  and  compact  because 
the  low-pressure  gas  holder  and  compressing  apparatus  are  not 

Three  of  these  high-pressure  generator  installations  have  been 
pnt  into  operation,  and  one  of  these  has  been  operating  satisfac- 
torily  for  2  years;  three  more  are  now  in  process  of  construction. 

In  both  low-  and  high-pressure  systems  the  generators  are  oper- 
at«d  at  a  temperature  of  about  1200"  F. 

The  average  of  analyses  of  25  samples  of  compressed  Fintsch 
gas  was  as  follows,  and  furnishes  a  representative  indication  of  i^ 
composition : 

Methane  CH,  60% 

Heavy  lllDmlnants: 

Benzene  CA ** 

Proprlene  C  jl.  ...  L 3G 

Ethylene  C,H«  etc.  J 

CO    . . ; S 

Hydr<%en   -,      4.6 


Specific  gravity  .80  to  .85. 

Ignition  temperature,  determined  by  Mtlton  L.  Hersey,  chemist  and 
chief  engineer  of  tests  of  Canadian  Pacltlc  Railway,  made  at  HcGlU 
University  1562"  T.  or  860°  C. 

Ciploslve  limit  between  about  4  per  cent  and  10  per  cent  of  the  gas. 

The  horizontal  candle-power  of  the  compressed  gas,  tested  in  open 
£at-flame  burner  sufficiently  small  to  avoid  smoking  and  calculated 
to  the  5  feet  per  hour  consumption,  is  about  40.  By  reason  of  the 
necessarily  smaU  rate  of  consumption  this  does  not  furnish  a  re- 
liable indication  of  the  candle-power.  The  spherical  illuminating 
power  of  the  lamps,  naked  fiame  and  mantle,  as  later  to  be  stated, 
are  the  values  to  be  considered  for  purpcraes  of  comparison. 

>y  Google 

178  IixuuiNATiKo  Enoinkebino 

Most  of  the  Pintscb  gas  is  used  for  the  lighting  of  railroad  carg. 
While  a  relatively  small  amoant  ie  used  in  lighting  buoys,  beacons, 
etc.,  the  service  performed  is  one  of  commanding  importance.  In 
the  early  days  steamers  and  ferry  boats  were  satdsfactorily  lighted 
by  this  system.  Many  of  the  ferry  boats  plying  in  New  York 
Harbor  were  at  one  time  lighted  by  coal  gas,  uncompressed  or  com- 
pressed. In  some  cases  these  methods  were  superseded  by  the 
Fintsch  system.  The  advance  in  the  art  of  electric  lighting,  coupled 
with  the  special  adaptability  of  electric  lighting  to  the  illumina- 
tion  of  vesseb  equipped  for  steam  power,  led  naturally  to  the  re- 
placement of  compreesed  gas  by  the  electric  light. 

Pio.  2. — PinUch  Gas  Regnilator. 

The  Pintech  regulator  deserves  more  than  a  passing  notice,  (See 
Fig.  S.)  The  essential  parts  of  the  regulator  are  a  needle  valve 
of  special  form  and  a  large  diaphragm  made  of  leather  so  treated 
as  to  be  gas  proof  and  extremely  flexible.  The  diaphragm  is  sub- 
jected only  to  the  reduced  or  regulated  pressure  and  controls  the 
movement  of  the  valve  through  a  lever  of  such  proportions  that  the 
pressure  of  the  valve  against  its  seat  is  11  times  the  total  pressure 
against  the  diaphragm.  A  pair  of  springs  acting  on  the  lever 
through  a  knife  edge  oppose  the  pull  of  the  diaphragm  and  can 
be  regulated  so  as  to  give  the  required  outlet  pressure.  The  needle 
valve,  BO  controlled,  is  relied  upon  to  exclude  the  high  pressure 
from  the  interior  of  the  regulator,  no  auxiliary  valve  being  em- 
ployed for  that  purpose  when  the  lamps  are  shut  off. 

For  the  illumination  of  railroad  care  many  forms  of  naked-flame 
lamps  have  been  employed.    These  have  all  been  designed  to  meet 

>y  Google 

Gas  and  Oil  Illuminants  173 

the  exacting  reqniremente  necessarily  involved  in  the  lighting  of 
cars  running  at  varying  speeds,  subject  to  abrupt  stops,  so  ven- 
tilated that  the  lamps  are  required  to  resist  strong  air  draughts, 
and  under  the  care  of  trainmen  who  cannot  be  relied  upon  to  give 
the  lamps  expert  attention. 

As  tiiis  is  one  of  a  number  of  lectures  on  illuminating  engi- 
neering it  is  in  order  that  1  should  call  particular  attention 
to  the  fact  that  the  engineers  of  the  Pintsch  Companies  here 
and  abroad  have  recognized  constantly  that  they  were  required  to 
solve  their  problems  from  the  standpoint  of  the  engineer  of  il- 
lumination. Not  only  has  the  effort  been  to  secure  the  greatest 
amount  of  light  from  a  minimum  of  material  and  at  a  minimum 
cost,  but  the  effort  has  been  to  distribute  this  light  so  as  best  to 
serve  the  travelling  public.  It  has  always  been  recognized  that  an 
important  element  in  the  problem  was  to  secure  an  effect  which 
would  be  pleasant  and  restful  to  the  eye.  All  the  problems  in- 
volved have  been  under  discussion  and  subject  to  experimentation 
constantly.  It  was  recognized  that  the  first  step  was  to  obtain  a 
steady  flame,  free  from  flicker,  and  that  this  must  be  secured 
through  the  design  of  a  draug^t-proof  lamp  and  a  pressure  regu- 
lator at  once  sensitive  and  reliable. 

I  know  that  some  hold  that  illuminating  engineering  was  not 
the  subject  of  scientific  study  by  gas  engineers  until  the  electric 
light  engineers  led  the  way.  I  am  inclined  to  think  that  some 
of  our  electric  light  associates  in  the  Illuminating  Engineering 
Society  are  of  this  number.  Many  facts  in  regard  to  gas  engineer- 
ing practice  could  be  cited  against  this  proposition.  In  addition 
to  the  record  made  by  the  Fintsch  engineers,  let  me  refer  to  one 
example,  which  is  notable  in  this  connection.  Some  few  years  ago, 
at  a  meeting  of  a  committee  of  our  Society,  I  learned  that  tiie 
electrical  engineers  present  were  of  the  opinion  that  a  notable 
advance  in  the  science  of  illumination  was  made  when  rooms  were 
first  illuminated  by  light  reflected  from  sources  hidden  from  the 
eye,  and  that  this  advance  was  to  be  credited  to  the  electric  light 
engineers.  I  then  described  the  lighting  of  the  Liverpool  Fhil- 
hannonic  Hall  by  naked  gas  fiames  placed  eo  as  to  be  hidden  from 
view  by  the  plaster  cornice,  the  light  being  reflected  down  into  the 
hall  from  the  curved  surface  of  the  ceiling.  This  installation  wag 
made  50  years  before  I  first  saw  it,  which  was  over  10  years  ago. 

I  trust  I  may  be  pardoned  for  this  little  digression,  and  especially 
by  my  electric  light  s 

>y  Google 

174  iLLDHINATItia  Enginxebikq 

Figure  3  sbowe  a  flat-flame  fonr-barner  raOrosd  car  lamp.  It 
is  here  to  be  borae  in  mind  that  the  methods  of  hanging  and  the 
design  of  tbe  body  of  the  lamp  have  been  varied  to  meet  practical 
conditions  and  the  demands,  sometimee  artistic  and  sometimes  not* 
of  the  railroads'  managers.  In  this  lamp  the  air  supply  to  the 
burners  passes  through  the  upper  portion  of  the  body  and  so 
into  the  cylinder  enclosing  the  four  chimneys,  down  into  the  lower 
portion  of  the  lamp  and  so  into  the  globe,  where  it  reaches  the 
flames.     The  products  of  combustion  go  up  past  the  central  re- 

Fio.  3. — Fonr-Buroer  Flat-Flame  Railroad  Car  Lamp 

flector,  and  so  on  up  through  the  chimneys,  some  of  the  sensible 
heat  of  the  products  of  combustion  being  transferred  to  the  in- 
coming air. 

The  four  burners  together  consume  about  3^  feet  of  gas  an 
hour,  and  give  30  to  35  mean  hemispherical  (lower)  candle-power. 

Of  recent  years  the  Pintsch  Companies  have  devoted  much  at- 
tention to  the  application  of  incandescent  mantles  to  car  lighting 
and  buoy  lighting.  Experiments  with  vertical  mantles  were  not 
successful,  by  reason  of  frequent  breakages.  After  the  trial  of 
many  devices  to  reduce  the  effect  of  shock  the  engineers  of  the 
United  States  Company  solved  the  problem  by  means  of  a  strong 
inverted  mantle  rigidly  fixed  to  the  burner.     To  secure  increased 

>y  Google 

G&s  AND  Oil  Illdhinaitts  175 

strength  these  maotles  ate  made  heavier  than  the  ordinary  mantle, 
and  to  compensate  for  the  loss  in  illuminating  power  dne  to  this' 
increase  in  mass  the  gas  is  supplied  to  the  bumerB  at  a  pressure 
of  3  pounds.  This  advantage  is  secured  by  the  use  of  a  coiapressed 
gas  controlled  by  a  reliable  governor.  It  is  a  rather  remarkable 
fact  that  the  lamps  are  not  provided  with  means  of  adjustment. 
The  gas  orifices  and  air  inlets  are  drilled  to  standard  sizes,  and. 

Pio.  4.— Single  Mantle  Car  Lamp. 

having  passed  the  calibration  tests,  the  lamps  are  erected  as  turned 
out  from  the  factory. 

These  mantle  burners  conaume  2  feet  of  gas  an  hour  and  give 
(the  mantles  alone)  90  to  100  horizontal  candle-power  without  the 
aid  of  reflectors.  As  arranged  in  the  car  lamp,  they  give  a  mean 
hemispherical  candle-power  of  90  to  100.  Comparing  with  the 
flat-flame  lamps  already  described,  the  lighting  effect  is  about  4 
to  1,  and  with  the  same  gas  storage  capacity  the  length  of  period 
between  fillings  is  practically  increased  60  pet  cent. 

>y  Google 


Illuuikatixq  Ekqinebbino 

These  inverted  mantles  as  now  need  have  established  a  satis- 
factoi?  life  record.  Some  little  time  ago  a  careful  obserration  was 
made  of  their  service  on  25  steam  railway  cars  engaged  in  New 
York  snbarban  traffic.  These  cars  were  equipped  with  125  lamps. 
The  ears  were  handled  in  the  regular  way  by  the  trainmen,  who 
were  not  informed  that  the  lamps  were  under  special  observation. 
The  Pintech  employees,  however,  renewed  all  broken  mantles  so 
that  an  accurate  record  of  the  mantles  used  might  be  obtained. 
The  result  of  this  test  for  the  135  lamps  was  an  average  mantle 


Equipment  for  Platach  Lighting. 

life  of  376  days.  This  shows  a  notable  improvement  even  over 
the  old  inverted  mantle  as  first  made. 

The  construction  of  the  lamp  is  shown  in  Fig.  4.  The  regu- 
lated gas  at  2  pounds  pressure  is  admitted  through  fitting  No.  3146, 
and  passes  down  to  a  strainer  of  peculiar  construction  placed  is  the 
vertical  channel.  The  gas  issuing  therefrom  is  met  by  the  air 
pulled  in  at  the  sides  by  the  gas,  and  the  gas  and  air  mixture  then 
passes  down  unobstructed  to  the  burner,  which  consists  of  a  metal 
disc  accurately  drilled  with  seven  orifices. 

Fig.  5  shows  car  equipment  for  Pintsch  lighting;  Fig.  6  is  an 
interior  view  of  a  railway  coach  lighted  with  mantle  lamps,  and 
Fig.  7  is  an  illumination  diagram  for  such  a  coach. 

I  cannot  conclude  this  section  of  my  lecture  without  describing, 
at  least  briefly,  the  Pintsch  buoy,  a  very  beautiful  example  of 

>y  Google 

Gas  and  Oil  Illominants  177 

specialized  engineering  akin  to  illuminating  engineering,     (See 

Fig.  8.)  The  buoy  body  is  a  seamless  wclded-steel  shell  designed 
and  constructed  to  withstand  the  high  pressure  of  the  gas  stored 
therein,  and  to  afford  ample  buoyancy  for  the  support  of  the  anchor 
ciiain,  lautem  and  other  parte.     The  buoy  bodies  are  made  in 

no.  6. — Interior  View  ot  Coach  with  Mantle  Lamps. 

different  shapes  to  meet  varying  conditions  as  to  depth  of  water, 
anchorage,  tideways,  etc.  A  suitable  tower  surrounded  by  a  cage 
supports  and  protects  the  lantern  and  carries  a  platform  to  afford 
a  footing  for  the  attendant  when  lighting  or  adjusting  the  flames. 
The  lantern  is  designed  and  constructed  to  protect  the  light  from 

>y  Google 


rain,  waves  and  wind  under  the  severegt  possible  conditions  to  be 
found  close  to  the  surface  of  the  sea.  The  base  of  the  lantern 
forms  the  case  for  the  pressure  governor. 

In  the  original  lantern  the  burner  was  placed  inside  a  Fresnel 
dioptric  iixed-Hght  lens  which,  by  bending  the  light  rays,  confined 
them  approximately  between  two  horizontal  planes,  thus  increasing 

Fig.  T.— Illumlaatlon  Diagram  for  Coach  witb  Mantle  Lamps. 

the  power  and  range  of  the  light.  Home  of  the  lanterns  thus 
equipped  are  still  in  use.  In  orcler  to  give  the  light  a  specific  char- 
acteristic and  at  the  same  time  reduce  tlie  gas  consumption  and  so 
increase  the  interval  between  gas  ehargings,  a  further  improvement 
was  made  by  the  addition  of  a  flashing  mechanism.  This  is  a 
simple  and  reliable  device  which  controls  the  flow  of  gas  to  the 
burner  so  that  the  gas  is  ignited  and  extinguished  at  intervals,  the 

>y  Google 

Oas  asd  Oil  Illuminaktb 


lengths  of  vMcIi  are  predetermined  to  meet  the  particular  condi- 
tlona  of  each  case.  This  antomatic  mechanism  is  enclosed  in  a 
chamher  located  immediately  above  the  governor,  and  is  actuated 
by  the  gss  flowing  through  this  chamber  on  its  way  from  the 
governor  to  the  flaeh-li^t  burner,  which  is  ignited  by  a  pilot  light 
burning  continaoasly  and  receiving  its  gas  supply  direct  from  the 

Fra,  8,— PIntoch  Buoy. 

The  relative  periodicity  of  light  and  darknesB  can  be  varied  by 
the  adjustment  of  the  mechanism  to  meet  varying  requirements. 
The  etaudard  adjustment  gives  periods  of  equal  lengths,  usually 
5  seconds  or  10  seconds  each.  If  desired  the  periods  can  be  made 

The  latest  form  of  this  mechanism  provides  for  the  buoy  being 
need  either  as  a  fixed  or  flash  light,  ae  required  for  any  location; 
all  the  buoys  as  now  built  and  supplied  are  so  equipped.    Nearly 

>y  Google 


all  the  baojB  sow  in  oae  are  equipped  with  the  flash-li^t  mechan- 
iam,  and  most  of  these  ate  of  the  convertible  type. 

Ab  the  fixed-light  lene  permits  the  Tsye  to  radiate  horizontally 
through  360  d^reee,  to  still  further  increase  the  power  and  range 
of  the  bnoy  lights  a  "  bnll's-eye "  or  flash  lens  can  be  employed 
instead  of  the  flashing  mechanism  just  described.  If  desired,  a 
series  of  these  lenses  can  be  grouped  in  a  circle  around  the  light 
source.  That  the  light  may  be  Tisible  at  all  points  in  the  horizon 
the  bull's-eye  lens,  or  series  of  lenses,  must  be  revohed.  This  is 
effected  by  a  motor  driven  by  the  gas  flowing  to  the  burner.  This 
lens  arrangement  delivers  a  light  at  least  20  times  as  powerful  as 
that  from  the  fixed-light  tens. 

An  additional  advantage  is  that  the  characteristic  of  the  buoy 
light  can  be  further  determined  by  the  design  and  the  reUtive 
positions  of  the  lenses  of  the  series.  There  are  comparatively  few 
of  the  revolving  lenses  in  service. 

Until  recently  flat-flame  burners  were  used  exclusively  in  the 
Piotsch  buoys,  but  mantle  burners  are  now  displacing  the  flat 
flames.  The  older  lanterns  are  being  remodeled  for  mantle  burners, 
and  all  new  lanterns  are  of  this  type.    (See  Fig.  &.) 

As  compared  with  the  flat  flame  the  mantle  burner  gives  a 
caudle-power  three  times  as  great,  and  its  intrinsic  brilliancy  is 
ten  times  as  great,  resulting  in  greatly  increased  power  for  the 
same  consumption  of  gas.  The  flat-flame  burners  are  made  for 
different  rates  of  consumption,  while  the  mantle  burners  are  made 
for  one  rate  only. 

Bells  operating  either  above  or  below  the  surface  of  the  water 
and  actuated  by  the  flow  of  gas  supplying  the  burner  are  in  some 
cases  attached  to  these  buoys. 

With  one  gas  charge  these  buoys  wilt  run  from  55  to  538  days ; 
the  size  of  the  buoy  body,  whether  flat-flame  or  mantie  burner, 
whether  fixed  or  flash  light,  and  if  flat  flame,  the  size  of  burners, 
determining  the  number  of  days. 

Stationary  beacons  and  Hght  ships  are  also  equipped  for  and 
operated  with  Piotsch  gas. 

Oas  under  a  pressure  of  100  atmospheres  is  now  being  used  ex- 
tensively for  tliis  marine  work.  For  beacons  and  light  ships  it 
is  burned  direct  from  the  cylinders  in  which  the  gas  is  courted. 
In  the  case  of  buoys  the  high-pressure  cylinders  obriate  the  neces- 
sity for  large  storage  holders  and  compressors  on  the  supply  tender. 

>y  Google 

0a3  ahd  Oil  Tlluhinants 

Fio.  B. — ^Pintacli  Buoy  Lantern  with  Mantle  Barn«r. 

>y  Google 

183  Illcuinatinq  Enqinebbino 

the  buoys  being  charged  direct  up  to  10  atmoBpheres  from  the  100- 
atmosphere  cylinders. 

It  IB  fovmd  that  about  133  volumes  of  the  gas  can  be  stored  under 
a  pressure  of  100  atmospheres,  and  that  little  or  no  additional  loss 
in  candle-power  is  suffered  in  carrying  the  compression  from  14  to 
100  atmospheres.  There  is  an  additional  deposit  of  liquid  hydro- 
carbon, as  indicated  by  the  increased  storage  volume,  but  if  the 
outlet  pipe  is  sealed  in  this  liquid  the  liquid  revaporizes,  and  at 
the  reduced  pressure  of  14  atmospheres  and  below  it  is  carried 
through  the  appliances  to  the  burner  practically  as  a  dry  gas. 

Let  me  conclude  by  pointing  out  two  features  of  the  Fintsch 
system  of  great  practical  advantage  to  its  patrons. 

In  connection  with  the  filling  of  the  care  it  is  important  that  the 
amount  of  gas  delivered  to  each  car  ^ould  be  readily  ascertainable 
for  record.  It  is  even  more  important  that  the  attendant^  should 
be  able  to  tell  by  inspection  at  any  time  how  many  hours  of  lighting 
are  provided  for  by  the  gas  in  the  cylinder.  Both  of  these  re- 
quirements are  met  by  making  the  cylinders  of  standard  sizes,  the 
cubical  contents  in  feet  being  marked  and  recorded.  A  high-pres- 
sure gauge  showing  the  pressure  in  atmospheres  is  attach^  at 
each  car-filling  valve.  The  simple  calculation  of  multiplying  the 
gauge  reading  by  the  capacity  of  the  cylinder  gives  the  available 
volume  of  gas  contained. 

Another  important  feature  is  that  throughout  the  territory  cov- 
ered by  the  United  States  Pintseh  Company  all  parts  of  machinery 
and  all  fittings  are  interchangeable.  The  design  of  the  smallest 
and  apparently  most  insignificant  part  has  been  carefully  con- 
sidered. The  engineers  have  from  the  first  recognized  that  they 
were  offering  to  perform  a  special  service  involving  many  difB- 
cultiee.  As  a  result,  a  system  has  been  developed  that  provides  for 
the  supplying  of  Fintsch  gas  to  any  railroad  car  equipped  with 
Fintsch  standardized  appliances,  no  matter  how  far  that  car  may 
be  from  its  home  territory,  provided  it  is  within  reach  of  any  one 
of  the  93  gas  works  or  any  one  of  the  360  Fintsch  gas-supplied 
railway  stations  located  in  the  United  States,  Canada  or  Mexico. 

Carburetted-Air  Oas 
Carburetted-air  gas  consists  of  atmospheric  air  to  which  hydro- 
carbon vapor  has  been  added,  the  proportions  of  air  and  vapor  vary- 
ing with  the  process  employed. 

>y  Google 

Gas  and  Oil  Illumikants  1S3 

The  application  of  carburetted-air  gas  as  an  illuminating  and 
heating  agent  to  meet  certain  special  conditions  has  been  an  in- 
dustry for  about  40  years.  As  a  source  of  energy  in  the  internal 
combustion  engine  its  use  has  been  of  late  greatly  increased  and 

Carburetted-air  gas  machines  can  be  grouped  in  two  classes,  those 
operated  without  heat  and  those  operated  with  heat.  Those  of  the 
former  group  have  been  more  generally  employed,  especially  where 
the  principal  eerrice  has  been  lighting.  In  operating  with  cold 
air  it  is  necessary  to  use  refined  highly  volatile  gasoline;  but  if 
steam  or  other  heat  source  is  employed  to  assist  evaporation,  the 
somewhat  less  volatile  and  less  expensive  naphthas  are  used. 

Carburetted-air  gas  differs  fundamentally  from  coal  gas,  water 
gas  or  oil  gas  through  the  fact  that  whereas  in  the  process  of  mixing 
the  liquid  hydrocarbon  is  vaporized  it  is  not  changed  chemically, 
while  in  the  case  of  the  other  three  gases  the  manufacturing  process 
to  which  the  coal  or  oil  is  subjected  converts  the  hydrocarbon  into 
fixed  gases  in  major  proportion  and  certain  vapors  in  minor 

In  the  distillation  of  crude  petroleum,  as  the  temperature  of  the 
still  rises,  the  several  distillates  are  driven  off  successively  accord- 
ing to  the  following  approximate  classification : 

90°  and  above  0.6363  and  below  Rbigolene  ft  cymogene 

90°  to  80°  0.G363  to  0.6667  Gasoline 

80*  to  70°  0.6667  to  0.7000  Llsht  napbtha 

70°  to  60°  0.7000  to  0.7368  Heavy  naphtba 

■  Following  these  distillates  come  the  kerosenes,  lubricating  oils,  gas 
oil,  solid  hydrocarbons,  tars  and  solid  carbons  or  hydrocarbons. 

While  refined  gasoline  of  90°  B.  (sp.  gr.,  .6363)  is  obtainable  in 
this  country,  the  price  and  the  extra  difficulty  in  holding  it  against 
evaporation  have  operated  to  prevent  the  development  of  a  wide 

'  market  for  this  grade. 

Hefined  gasoline  lighter  than  86°  B.  (sp.  gr.,  ,6481)  is  not  gen- 
erally obtainable  in  this  country.  This  distillate  consists  mainly 
of  hexane'(CgHi,)  and  pentane  (C,H,,),  with  some  still  lighter 
and  some  heavier  hydrocarbons.  It  should  evaporate  under  condi- 
tions of  use  without  giving  off  at  first  an  excess  volume  of  light 
vapors  or  leaving  unvaporized  heavy  residues.    A  distillate  capable 

>y  Google 

184  Illohinatino  Enqineebino 

of  meeting  these  conditiom  can  be  obtained  only  by  repeated  dis- 
tiUationa  in  the  refinery  to  isolate  in  the  gasoline  those  cloaely 
related  hydrocarbons  which  will  evaporate  in  approximately  the 
same  Tolumee  under  the  same  conditions.  The  refining  process 
must  also  provide  for  the  removal  of  all  traces  of  tar  which  other- 
wise would  deposit  in  the  smaller  pipes,  gum  the  floats  and  clog 
the  burners.  For  the  making  of  air  gas  the  more  general  practice 
has  been  to  use  a  gasoline  of  about  84°  B.  (sp,  gr.,  .6542).  While 
this  distillate  leaves  unvaporized  a  little  residne,  the  amount  is 
small,  and  as  a  rule  does  not  have  to  be  pumped  out  oftener  than 
every  6  to  12  months.  It  is  interesting  to  note  that  the  residue 
is  about  63°  B.  (sp.  gr.,  .7254).  The  nomenclature  which  developed  . 
to  identify  the  distillates  of  petroleum  in  many  cases  is  based  only 
upon  a  commercial  or  industrial  suggestion.  As  the  names  given 
to  several  of  these  distillates  have  been  the  occasion  for  considerable 
confusion,  a  few  words  of  explanation  may  not  be  out  of  place. 

^Vhen  these  lighter  distillates  from  petroleum  were  first  obtained 
uses  for  them  in  the  arta  were  still  to  be  found.  In  manufacturing 
kerosene,  for  which  there  was  a  ready  market,  the  refiners  were 
embarrassed  to  find  storage  for  these  distillates  produced  as  by- 
producte,  and  for  which  there  was  little  or  no  market.  At  that 
time  benzene — a  hydrocarbon  having  the  chemical  formula  C,H„ 
obtained  principally  from  the  distillations  of  coal-tar — possessed 
a  considerable  value  in  the  industries  as  a  solvent  for  fats  and 
greases  and  an  enricher  for  gaa.  It  soon  became  clear  that  some 
of  the  lighter  distillates  of  petroleum  could  be  used  as  a  substitute 
in  part  for  benzene,  and  thus  a  commercial  reason  was  furnished 
for  designating  these  distillates  by  the  name  benzine.  In  the  same 
way  other  distillates  of  coal-tar,  known  as  light  and  heavy  naphthas, 
had  their  names  pre-empted  for  other  petroleum  distillates.  As  the 
nomenclature  thus  developed  fails  to  meet  the  requirements  of  a 
technical  terminology  the  result  naturally  has  been  a  most  em- 
barrassing confusion  in  technical  and  industrial  literature.  As  an 
example,  there  are  uses  for  benzene  and  coal-tar  naphthas  for  which 
the  petroleum  distillates  cannot  be  substituted;  hence  the  need  to 
be  sure  whether  the  substance  under  consideration  is  benzine  or 
benzene  in  the  first  case,  or  naphtha  or  petroleum  "  naphtha "  in 
the  second  case.  Another  feature  of  commercial  practice  which 
has  led  to  confusion  is  that  of  designating  the  specific  gravity  of 
petroleum  distillates  by  the  Baum4  hydrometer  readings,  even  to 

>y  Google 

Qas  and  Oil  Illuuinants  165 

the  extent  in  some  cases  of  calling  that  reading  the  specific  gravity. 
This  is  all  the  more  unfortunate  for  the  reason  that  in  the  upper 
part  of  the  scale  as  the  Bauin^  reading  increases  the  distillate  is 
of  a  lighter  specific  gravity,  and  in  the  lower  part  of  the  scale  as 
the  Baum^  reading  decreases  the  distillate  is  of  a  heavier  specific 
gravity,  the  Banm^  reading  of  70"  indicating  a  specific  gravity 
of  .70. 

An  additional  complication  arises  from  the  fact  that  the  Baum^ 
scale  for  liquids  lighter  than  water  is  calculated  on  more  than  one 
formula,  and  therefore  the  tables  used  in  converting  Baum^  degrees 
to  specific  gravity  do  not  always  agree.    The  values  here  given  are 

calculated  on  the  formula  sp.  gr.  equals    ,...     „ ; jr — 

'^    °       ^  130  +  Baume  reading 

which  is  the  American  standard.  Another  formula  more  often  fol- 
I„«d  ia  Englidi  books,  » 146T3  +  Ssfnl  FeiaEj '  U-^ort^'tely, 
the  tables  are  frequently  given  without  the  formula  and  the  unwary 
may  be  deceived.  Some  authorities  are  careful  to  state  in  the  title 
of  the  table,  "  American  Standard,"  In  the  majority  of  books  of 
reference  the  tables  do  not  go  above  80°  B.,  and  in  some  the  tables 
are  even  more  limited.  For  these  reasons  for  American  prac- 
tice  it   is   convenient   to   remember   the  formula  ep.  gr.   equals 


130+Banme  reading ' 

The  volume  of  gasoline  vapor  that  can  be  carried  by  a  given 
quantity  of  air  depends  upon  the  temperature,  the  pressure  remain- 
ing constant.  The  ability  of  air  to  take  up  and  hold  in  suspension 
gasoline  vapors  increases  very  rapidly  with  the  increase  in  tempera- 
ture. Professor  Leslie  says  in  this  connection  that  while  the 
temperature  itself  advances  uniformly  in  arithmetical  progression 
the  increased  dissolving  power  thus  communicated  to  the  air  ad- 
vances with  the  accelerating  rapidity  of  a  geometrical  progression. 

While  experiments  that  have  been  made  to  test  this  theory  have 
not  agreed  in  confirming  its  truth,  they  suggest  that  it  may  be  at 
least  approximately  true. 

Sir  Boverton  Redwood  states  with  regard  to  86°  B.  (sp.  gr., 
.6481)  gasoline  that 

100  volumes  of  air  at  32°  F.  vrill  retain  10,7  per  cent  of  vapor, 
(9.7  per  cent  of  the  mixture). 

>y  Google 

186  Illdhinating  Enoinberinq 

100  voluiuee  of  air  at  50°  F.  will  retain  17.5  per  cent  of  vapor, 
(14.9  per  cent  of  the  mixture). 

100  volumes  of  air  at  68°  P.  will  retain  27  per  cent  of  vapor, 
(21.3  per  cent  of  the  mixture). 

In  this  connection  Redwood  goee  on  to  Bay  that  "  air  charged 
with  735  grains  of  gasoline  per  cubic  foot  has  been  found  to  pos- 
sess an  illuminating  power  of  16.5  candles  when  consumed  at  the 
rate  of  31^  cubic  feet  an  hour  in  a  15-hole  Ai^and  burner." 

If  we  assume  the  gasoline  vapor  to  have  a  specific  gravity  of  3., 
it  follows  that  the  mixture  has  31^  per  cent  of  gasoline  vapor 
by  volume. 

Redwood  goes  on  further  to  describe  a  series  of  experiments, 
which  he  carried  on  with  the  assistance  of  Mr,  Blunderstone,  to 
determine  "  the  planner  in  whicli  crude  petroleum  and  certain 
volatile  petroleiim-distillates  evaporate  when  subjected  to  a  current 
of  dry  air.  ...  In  these  experiments,  dry  air  was  caused  to  bubble 
slowly  through  the  liquid  in  a  series  of  graduated  tubes  maintained 
at  a  constant  temperature.  ...  A  set  of  determinations  being  made 
at  temperatures  of  40",  60%  80°  and  100°  F." 

At  60°  three  determinations  were  made  with  gasoline  of  a  sp.  gr. 
of  .639,  44.7  c.  c.  of  the  liquid  being  used.  In  the  first,  0.9  liter  of 
air  was  passed  through  the  six  tubes;  in  the  second,  2.15  liters,  and 
in  the  third,  3.55  liters.  The  first  gave  a  total  evaporation  of  .66 
volume  of  liquid  to  100  volumes  of  air;  the  second  gave  ,59  and  the 
third  .51  volume. 

It  is  thus  seen  that  the  relatively  small  amount  of  air  took  up  the 
largest  amount  of  gasoline.  The  result  of  the  first  teat,  if  calculated, 
shows  that  the  mixture  contained  53  per  cent  by  volume  of  the 
gasoline  vapor — certainly  an  extraordinary  result.  The  probabilities 
are,  at  least  in  this  last  series  of  experiments,  that  the  small  quan- 
tity of  air  slowly  bubbling  through  the  liquid  in  six  small  streams 
resulted  in  a  selective  evaporation.  If  so,  this  does  not  truly  repre- 
sent the  result  from  a  liquid  of  .639  sp.  gr.  Certainly,  we  are  not 
warranted  in  believing  that  any  such  percentages  of  gasoline  can 
be  carried  in  air-gas  practice  as  are  indicated  in  the  two  cases  lost 

The  limits  between  which  gaEoline  vapor  and  air  form  an  ex- 
plosive mixture  are  3  per  cent  of  vapor  with  98  per  cent  of  air  and 
5  per  cent  of  vapor  with  95  per  cent  of  air  by  volume.  This  fact 
furnishes  a  reason  for  dividing  carburetted-air  gas  into  two  classes: 

>y  Google 

QiB  AND  Oil  Illuiiinants  187 

First,  that  Id  which  the  proportion  of  gasoline  vapor  to  air  ie  lesB 
than  2  per  cent;  and,  second,  that  in  which  the  proportion  of  gaso- 
line vapor  to  air  is  more  than  5  per  cent. 

The  former  presents  siHne  very  interesting  features.  A  carburet- 
ted-air  gas  containing  ly^  p^r  cent  of  gasoline  vapor  is  low  in 
heating  value,  is  noB-explosive,  is  non-asphyxiating,  and  yet,  when 
used  with  a  Welsbach  mantle,  furnishes  a  Batisfactory  light.  It 
would  appear  that  such  a  gas  hoe  much  to  recommend  it.  This 
class  of  air  gas  has  been  adopted  in  England  to  a  considerable  ex- 
tent for  lighting  country  estates,  audience  halls,  summer  hotels, 
and  the  like.  As  yet  it  has  received  little  recognition  in  this  coun- 
try.   A  company  is  now  presenting  its  claims  for  recognition. 

The  specific  gravity  of  the  vapor  of  gasoline,  as  now  generally 
used  for  air  gas,  is  about  3. 

The  calorific  value  of  gasoline  is  variously  quoted.  In  this  con- 
nection it  is  to  be  remembered  that  "  gasoline  "  is  not  a  substance 
of  constant  chemical  composition.  Furthermore,  the  statements  do 
not  always  show  whether  the  value  quoted  is  gross  or  net  heating 
value.  The  United  States  Geological  Survey  gives  19,300  B,  t.  u. 
per  pound  as  the  net  value  of  gasoline  of  ,71  to  .73  specific  gravity. 
Bulletin  Xo.  191  of  the  United  States  Department  of  Agriculture, 
on  the  authority  of  Lucke  &  Woodward,  gives  21,130  gross,  19,660 
net,  B,  t.  u.  per  pound. 

Redwood,  in  discussing  vapor  tensions,  says: 

"Salleron  ft  Urbain  give  also  the  following  aa  the  determined  vapor- 
preSBures  (vapor-tens Ions)  of  petroleum  products  of  various  denaltieB." 

He  then  goes  on  to  say  that  the  values  given  are  "  founded  on 
a  belief  not  in  all  cases  correct." 

This  table,  so  rather  guardedly  quoted  by  Sedwood,  gives  as  the 
vapor  tension  of  distillate  of  sp.  gr.  .65  (B.  85.38),  2110mm.  of 
water.  This  can  be  accepted  at  least  as  approximately  correct,  and 
would  then  show  that  air  would  be  saturated  when  30.42  per  cent 
by  volume  of  gasoline  was  present. 

In  this  country  the  use  of  carburetted  air  has  been  confined  for 
many  years  to  machines  that  produce  a  mixture  containing  over 
5  per  cent  of  gasoline  vapor,  and  it  has  been  the  practice  to  use 
5y2  to  6^  gallons  of  gasoline  to  1000  feet  of  the  mixture,  the  con- 
tent of  gasoline  vapor  then  showing  a  wide  margin  of  safety  above 
the  5  per  cent  explosive  limit.  A  oVi>-gallon  gas  burned  in  an 
Ai^and  burner,  gives  from  15  to  16  candie-power,  it  contains  about 

>y  Google 

188  Illuminating  Ekoikeerin'o 

1314  per  cent  gasoline  vapor,  and  its  specific  gravity  is  about  1.26. 
The  Bpecific  gravity  of  the  mixture  is  important ;  for  being  heavier 
than  air,  in  case  of  leak,  not  possessing  the  tendency  to  rise,  it  is 
less  rapidly  dissipated  by  the  ordinary  means  of  ventilation.  This 
necessitates  increased  precautions  against  explosion  and  asphyxia- 
tion. Such  a  gae  cannot  be  subjected  to  a  temperature  below 
43°  F.  without  depositing  gasoline  in  the  pipes;  therefore,  it  must 
be  protected  against  cold  either  by  wrapping  the  pipes  or  by  ex- 
ternal heat.  This  gas  will  have  a  calorific  value  of  about  570 
B.  t.  u.  per  foot,  and  therefore  can  be  employed  to  advantage  for 
lighting  (cBpecially  by  mantles)  and  heating. 

Four  principal  systems,  the  first  substituting  hydrogen  for  air, 
are  used  in  the  application  of  gasoline  vapor  to  gas  making,  and 
these  are  as  follows: 

1.  Although  not  an  air-gas  system,  it  may  be  convenient  to  men- 
tion here,  by  reason  of  similarity  of  method,  the  process  of  forcing 
manufactured  hydrogen  gas  over  or  through  gasoline  by  which  the 
hydrogen,  which  has  no  illuminating  value  of  its  own,  becomes 
saturated  with  the  rich  hydrocarbon  vapors.  This  mixture  has  a 
high  heating  value ;  and  especially  when  used  with  the  incandescent 
mantle,  a  high  illuminating  value.  This  system  is  seldom  found 
in  general  practice  and  is  principally  used  in  metallurgical  labo- 

3.  The  employment  of  devices  by  means  of  which  a  current  of 
air  is  forced  over  or  through  a  body  of  gasoline  or  some  porous 
or  fibrous  material  saturated  or  impregnated  with  gasoline,  by 
which  means  the  air  becomes  carburetted  with  the  hydrocarbon 
vapors  to  such  an  extent  that  the  mixture  can  be  used  advan- 
tageously for  illuminating  and  beating  purposes.  This  method  is 
called  the  eold-air  procese,  and  is  the  one  most  used  in  small  private 
installations  and  town  plants. 

Fig.  10  shows  such  an  installation.  It  consists  of  a  blower  "  A," 
carbureter  "L"  and  mixer  "  M."  The  blower,  operated  by  sus- 
pended weights,  as  shown  in  the  drawing,  or  water  power,  takes 
in  air  and  forces  part  through  the  carbureter  and  part  into  the 

Fig.  11  shows  a  sectional  view  of  a  box-^rpe  carbureter,  the  kind 
generally  used  in  plants  of  moderate  Bi;!e.  It  is  a  flat,  rectangular 
box  made  of  sheet  metal,  having  partitions  running  longitudinally 
and  parallel  to  each  other  through  the  box,  but  leaving  a  connect- 

>y  Google 

Qas  USD  Oil  Illdhinants 

>y  Google 


Illdminatiko  Enoineerino 

ing  opening  between  eacli  two  adjacent  compartmeiite  sequentially 
at  alternate  ends.  In  these  compartments  are  hung  or  stretched, 
as  shown,  strips  of  Canton  flannel.  There  is  an  opening  for  filling, 
an  inlet  for  air  from  the  blower  at  one  end,  and  at  the  other  end  an 
outlet  for  the  carhnretted  air.  The  carbureter  is  about  15  inches 
deep  but  is  BUed  vith  gasoline  to  a  depth  of  only  6  inches.  It  is 
buried  in  the  ground.    The  air  entering  through  the  top  at  one  end 


Fig.  11, — Carbureter,  BO-L!ght  Air  Gas  Machine. 

traverses  all  the  passages,  flowing  through  the  flannel  which,  by 
capillarity,  is  kept  wetted  with  gasoline;  the  carburetted  air  then 
passing  on  to  the  mixer.  The  box  must  be  of  sufficient  .size  to  per- 
mit a  very  slow  movement  of  the  air  and  the  recovery  from  the 
surrounding  earth  of  the  heat  rendered  latent  by  the  evaporation 
of  the  gasoline. 

As  stated  before,  commercially  refined  gasoline  is  a  mixture  of 
several    hydrocarbons,   though   largely   composed   of   hexane   and 

>y  Google 

Gas  ano  Oil  Illcminants  191 

pentane.  Under  the  conditions  of  slow  evaporation  here  presented, 
there  is  a  selective  evaporation,  the  lower  boiling  fractions  going 
off  in  large  volumes  first,  gradually  decreasing  in  volume  as  the 
gravity  of  the  remaining  gasoline  increases,  until  finally  a  mini- 
mum permieeible  candle-power  is  reached,  when  the  carbureter 
must  be  recharged.  The  cycle  is  then  repeated.  To  minimize  this 
fluctuation  in  candle-power  and  heating  value,  a  tank  containing 
a  considerable  supply  of  gasoline  is  sometimes  connected  to  the 
carbureter  with  a  ball  and  float  valve,  by  which  the  height  of  the 
gasoline  in  the  carbureter  is  replenished  as  fast  as  evaporated.  It 
is  claimed  that  this  is  an  unnecessary  refinement  when  a  separate 
-  mixer  ie  employed. 

When  air  is  brought  so  intimately  in  contact  with  highly  volatile 
gasoline  the  quantity  of  gasoline  vapor  that  passes  off  with  the  air 
may  be  considerably  in  excess  of  that  required  to  saturate  the  air . 
at  the  final  temperature.  The  gas  from  the  carbureter  is,  there- 
fore, not  then  in  condition  to  use ;  it  is  too  rich  and  unstable  as  to 
condensibility.  As  has  been  shown,  and  perhaps  explained.  Bed- 
wood  is  authority  for  the  apparently  contradictory  statement  that 
while  air  will  require  only  28  per  cent  of  gasoline  vapor  to  saturate 
at  60°  ¥.,  yet  when  the  air  is  bubbled  slowly  through  a  series  of 
six  tubes  containing  gasoline  of  the  same  gravity  at  the  same 
temperature,  the  mixture  of  vapor  and  air  passing  off  consisted 
of  more  than  50  per  cent  vapor. 

The  carburetted  air  from  the  carbureter  is  passed  into  the 
mixer  (Fig.  10),  The  mixer  consists  of  a  small  holder  rising  and 
falling  above  the  water  in  an  enclosing  metal  cylinder.  The  holder 
has  trips  which  open  and  close  cocks  at  its  lowest  and  highest 
points,  thereby  operating  automatically  by  the  flow  of  the  gas. 
There  is  a  test  light  and  an  adjusting  cock  for  regulating  the  pro- 
portion of  air  to  be  mixed  with  the  highly  carburetted  air  from 
the  carbureter,  and  a  valve  which  is  designed  to  control  within 
certain  limits  the  proportions  of  air  from  the  blower  and  carburet- 
ted air  from  the  carbureter. 

This  is  known  generally  as  the  cold-air  process;  under  proper 
and  reasonable  supervision  it  affords  a  safe  and  practical  means  of 
illumination  and  heating.  When  installed  so  as  to  comply  with  the 
underwriters'  requirements  it  involves  no  increase  in  insurance 
rates.  While  designed  and  intended  only  for  a  mixture  above  the 
explosive  limits  it  could  he  mechanically  adapted  to  yield  a  mixture 
below  the  explosive  limits. 

>y  Google 

192  Illuminating  Ekoineebino 

3.  To  coDveit  gasoline  into  a.  vapor  by  the  appUcstion  of  ex- 
ternal heat  and  then  I^  snitable  mechanical  means  to  mix  the  gas 
or  vapor  so  formed  with  any  desired  proportion  of  air.  This 
process  has  been  applied  in  a  number  of  types  of  air-gas  machines. 
Oenerally,  the  heating  device  is  in  the  form  of  a  coil  through 
which  the  gasoline  passes  and  which  is  heated  by  a  burner. 

Machines  of  this  class  are  simpler  as  to  number  and  complexity 
of  paris,  but  the  direct  application  of  ilames  to  a  coil  containing 
gasoline  has  not  been  considered  safe  by  most  insurance  companies, 
and  their  use  is  therefore  restricted. 

4.  The  fourth  method  consists  of  inducing  a  current  of  air 
into  a  small  tube  by  a  jet  of  steam  and  at  the  same  time  allowing 
sufficient  gasoline  or  naphtha  to  enter  to  condense  the  steam  and 
combine  with  the  air.    The  latent  heat  of  the  steam  in  this  process 

.  is  intended  to  compensate  for  the  refrigerating  action  of  the  gaso- 
line or  naphtha  in  passing  to  the  stat£  of  vapor. 

With  both  the  third  and  fourth  methods  petroleum  naphtha  of 
a  considerably  lower  gravity  may  be  used,  say  73°  to  68°  B.  (sp. 
gr.,  .6931  to  .7071) ;  while  with  the  cold-air  process  gasoHne  not 
heavier  than  8S°  B.  (sp.  gr.,  .6604)  can  be  used  without  the  neces- 
sity of  pumping  the  residue  from  the  carbareter  oftener  than 
once  in  6  months. 

The  fourth  method,  one  of  the  earlier  inventions  of  Hiram 
Maxim,  is  probably  best  for  a  large  output  of  gas.  Fig.  12  fihows 
one  of  these  machines  with  a  sectional  view  of  the  steam  injector 
for  air  and  naphtha.  Steam  at  about  60  pounds  gauge  pressure, 
controlled  by  a  regulator,  is  supplied  to  chamber  "  A,"  from  which 
it  issues  at  high  velocitj'  through  injector  nozzle  "L"  into  tube 
"  G,"  drawing  in  air  from  "  C  "  by  the  injector  action.  At  the 
other  end  of  tube  "  G "  a  secondary  injector  action  takes  place, 
naphtha  entering  by  the  adjustable  valve  "  D."  The  latent  heat 
of  the  steam  vaporizes  Ihe  naphtha  and  by  doing  so  the  st^om 
itself  becomes  condensed.  The  naphtha  vapor  and  air  unite  and 
pass  into  the  gas  holder,  while  the  condensed  steam  is  trapped  away. 
The  operation  of  this  machine  is  entirely  automatic.  When  work- 
ing close  to  its  capacity  very  little  of  the  gas  remains  in  the 
holder,  but  when  the  consumptioD  of  gas  is  reduced  to  a  minimum 
the  holder  fills  with  gas,  and  by  means  of  a  system  of  trips  and 
levers  the  process  is  interrupted  by  the  closing  of  the  steam  noz- 
zle; when  the  holder  descends  the'operation  is  reversed,  the  steam 

>y  Google 

Gas  and  Oil  iLLOMiNANTa  193 

nozzle  is  opened  and  the  making  of  gas  contioucB  as  before.  By 
regulating  the  adjastable  air  and  naphtha  valves  any  desired  mix- 
ture of  vapor  and  air  can  be  obtained,  and  in  larger  quantities 
than  wifh  any  of  the  cold -air  processes. 

The  simplicity  of  carbnretted-air  processes  is  evident;  no  puri- 
fying of  the  delivered  gas  is  required,  and  all  the  heat  of  the 
liquid  fuel  is  directly  transferred  to  the  air  and  vapor  mixture. 

The  burners  used  for  securing  illumination  through  the  agency 

of  carburctted  air  are  the  ordinary  flat-flame  lava  tip,  and  the 
various  forms,  both  upright  and  inverted,  of  mantle  burners. 

Where  no  mixer  is  installed  and  the  gas  is  consumed  directly 
from  the  carbureter  Ihe  lavn-tip  burner  has  a  small  set-screw, 
by  which  the  gas  can  be  adjusted  in  its  flow  so  as  to  prevent  heavy 
and  sinoky  flames. 

The  specific  gravity  of  the  gas  being  much  greater  than  that 
of  coal  or  water  gae,  it  requires  a  larger  opening  in  the  check  for 
the  same  quantity  of  gas  to  flow  through,  and  in  some  eases  larger 
openings  for  the  air  through  the  Bunsen  are  required. 

When  a  well -designed  mixer  is  installed  with  the  machine,  as  it 

>y  Google 


always  should  be,  there  is  no  incoiiTement  fiuctuation  in  the  candle- 
power  of  the  light  from  the  mantle  bomer. 

When  bumjog  a  mixture  containing  less  than  3  per  cent  gasi>- 
line — below  the  range  of  explosibility — the  Bnnaen  burner  on  the 
Welabach  burner  is  omitted  entirely,  as  the  gas  contains  sufficient 
air  for  a  non-luminous  flame. 

The  extent  to  which  carburetted-air  gas  is  used  for  lighting 
cannot  be  determined  accurately  from  available  statistics.  It  occu- 
pies a  field  similar  to  acetylene — that  of  isolated  plants  and  plants 
for  the  general  supply  of  small  towns  and  villages.  From  many 
of  the  plants,  especially  those  operated  by  municipalities,  no 
answers  are  received  to  applications  for  information ;  in  many 
other  cases  the  answers  are  vague  and  ambiguous.  Brown's  Gas 
Directory  shows  that  in  the  United  States  there  are  134  town 
plants.  It  is  claimed  that,  including  the  smaller  plants,  there  are 
twice  this  number.  A  fair  estimate  of  the  amount  of  gas  made 
and  distributed  by  the  124  town  plants  is  not  less  than  166,000,000 
cubic  feet  a  year.  The  gas  is  used  for  street  lighting  as  well  as 
for  domestic  consumption.  In  some  cases  the  gas  is  distributed 
through  a  considerable  mileage  of  mains.  The  prices  charged 
vary  from  $1.25  to  $2.50  per  1000  feet.  One  of  the  largest  com- 
panies reports  a  total  annual  sale  of  35,000,000  cubic  feet  sold 
through  1S6  meters  and  44  public  lamps  and  distributed  through 
81^  miles  of  mains. 

All  things  considered,  perhaps  the  field  in  which  carburetted- 
air  gas  can  demonstrate  its  greatest  economic  efficiency  is  in  that 
of  factories  using  various  special  heating  devices  of  comparatively 
■small  individual  capacity.  The  plant  being  installed  primarily 
for  this  special  heating,  it  can  also  be  employed  economically  for 

Acetylene  is  one  of  the  group  of  hydrocarbons  covered  by  liie 
general  formula  C,hH,„,  its  own  formula  being  C,H,;  that  is, 
its  one  molecule  contains  two  atoms  each  of  carbon  and  hydrogen. 
This  gas  has  long  been  known  to  the  chemists ;  and  even  as  pro- 
duced synthetically,  by  uniting  the  elements  in  the  compound,  the 
record  goes  back  to  183G,  though  the  reaction  was  not  then  fully 
understood.     In   1868  Woehler  announced  the  discovery  of  the 

>y  Google 

Gas  and  Oil  iLLOMnrAHTe  196 

production  of  acetylene  from  calcium  carbide  made  by  heating  to 
a  Tery  high  temperature  a  mixture  of  charcoal  with  an  alloy  of 
zinc  and  calcium.  Acetylene  was  known  by  chemiBta,  and  gae 
engineere  also,  ae  one  of  the  heavy  illnininanta  analytically  pro- 
duced in  small  percentages  during  the  destructive  distillation  of 
coal  in  the  making  of  coal  gas  and  in  the  generation  of  water  gas, 
and  its  high  value  as  an  enricher  was  understood. 

Acetylene  polymerizes  at  about  600°  C.  {1112°  P.),  that  is,  at 
elevated  temperatures  it  is  converted  into  other  hydrocarbons  hav- 
ing the  same  percentage  composition,  but  containing  more  atoms 
of  carbon  and  hydrogen  in  their  molecules.  Acetylene  readily 
polymerizes  to  benzene,  C,H,.  This  change  is  indicated  by  the 
equation  3CjH,  =  C,H,.  Benzene,  like  acetylene,  contains  by 
weight  almost  exactly  92.3  per  cent  carbon  and  7.7  per  cent  hydro* 
gen,  but  its  molecule  contains  six  atoms  of  each  element  instead 
of  2,  as  in  the  case  of  acetylene.  It  will  be  seen  later  that  this 
instability  of  acetylene,  tt^ether  with  its  other  characteristics,  has 
a  most  important  bearing  upon  its  treatment  and  application  and 
the  precautions  to  be  taken  against  accidents. 

In  1892  Thomas  M.  Willson,  an  electrical  engineer,  while  ex- 
perimenting on  the  production  of  metallic  calcium,  employing 
therefor  an  electric  furnace  of  high  voltage  in  which  was  a  mix- 
ture of  lime  and  coal-tar,  obtained  a  mass  which  he  accidentally 
discovered  contained  calcium  carbide,  and  which  gave  off  ace^lene 
when  immersed  in  water.  Willson  was  the  first  to  demonstrate 
that  acetylene  could  be  obtained  from  calcium  carbide  in  sufficient 
quantities  and  at  a  cost  that  would  secure  it  a  place  in  the  in- 
dustrial arts. 

This  discovery  of  Willson's  undoubtedly  increased  and  intensified 
the  interest  in  electro-chemical  research  and  in  synthetic  chemistry, 
which  two  fields  of  research  hold  out  much  of  pronuse  for  the  bene- 
fit of  mankind.  It  has  also  served  to  strengthen  the  theory  or  sur- 
mise that  metallic  carbides  exist  ui  the  earth's  interior,  and  are 
the  origin  of  petroleum  and  natural  gas.  Calcium  carbide  is  com- 
posed of  one  part  of  calcium  and  two  parts  of  carbon,  as  shown 
by  the  formula  CaCf.  It  is  a  hard,  crystalline  substance,  dark  gray 
in  color,  specific  gravity  about  2.32.  One  cubic  foot  of  compact 
carbide   therefore  weighs  about   138   pounds. 

The  two  highly  refractory  substances,  lime  and  carbon,  are 
forced  to  combine  under  the  action  of  excessively  high  tempera- 

>y  Google 

196  Illcminatino  Engikeerino 

turea,  as  most  readily  obtained  in  the  electric  furnace.  The  re- 
action ia  shown  by  the  equation 

CaO      +     3C       ^  CaC,  +  CO 

(Quicklime)    (Carbon)        (Calcium  Carbide)    (Carbon  Monoxide) 

which  shows  that  56  pounds  of  Hme  combine  with  36  pounds  of 
carbon  to  form  64  pounds  of  calcium  carbide  and  28  pounds  of 
carbon  monoxide,  Eoughly,  then,  for  the  making  of  a  long 
ton  of  the  carbide,  there  is  required  a  short  ton  (2000  pounds)  of 
lime  and  1375  pounds  of  carbon. 

In  the  manufacture  of  the  carbide  the  purity  of  the  raw  material 
ie  of  prime  importance.  Those  forms  of  carboniferous  material 
in  which  there  is  a  low  percentage  of  fised  carbon  are  to  be  avoided 
as  the  rapid  evolution  of  gaseous  products  therefrom  is  likely  to 
lead  to  explosions. 

The  calcium  carbonatea,  such  as  limestone,  marble,  etc.,  from 
which  the  lime  or  calcium  oxide  is  prepared,  must  be  low  in  con- 
tent of  magnesia,  alumina,  silica,  sulphur  and  pboephoruB.  The 
ordinary  limekiln  cannot  be  used  because  of  the  impurities  that 
would  be  introduced  therefrom.  As  it  takes  about  100  pounds  of 
carbonate  of  lime  to  yield  56  pounds  of  the  oxide,  those  impurities 
not  driven  off  with  the  carbonic  acid  would  be  nearly  doubled. 
These  necessary  precautions  led  to  the  general  practice  of-  cal- 
cining the  carbonate  at  the  carbide  factory. 

After  mixing  the  lime  and  carbon  in  proper  proportions  they  are 
fused  by  a  powerful  electric  current.  Resistance  and  arc  furnaces 
are  both  used.  The  furnace  must  be  operated  under  uniform 
■  heating.  For  Hie  generation  of  the  heavy  currents  required  re- 
course may  now  be  had  to  more  or  less  remote  water  powers  if  other- 
wise desirable,  as  railroad  transportation  ■  of  the  carbide  is  no 
longer  hampered  by  onerous  restrictions.  The  carbide  is  neces- 
sarily packed  in  tightly  sealed  cans  to  protect  from  moisture. 

While  a  generation  has  not  yet  elapsed  since  the  first  introduc- 
tion of  acetylene  to  the  commercial  world  the  files  of  the  patent 
officea  contain  such  a  multiplicity  of  applications,  granted  and 
rejected,  that  it  would  be  futile  at  this  time  to  touch  on  this  branch 
of  the  subject.  Many  of  these  applications  show  that  the  inventors 
neither  understood  the  principles  involved  nor  the  progress  of 
the  art,  an  ignorance  frequently  accompanying  much  so-ca!let! 

>y  Google 

Gab  and  Oil  Illuminants  197 

The  production  of  carbide  in  Euroi)e  in  1908  is  approximated 
as  follows: 

Sweden  and  Norway  36,000 

France    26,000 

Switzerland    30.000 

Italy    31,000 

Anatrla 20,000 

Germany    40,000 

Scattered 10,000 

Total 1B2.000 

Practically  all  of  this  carbide  was  nsed  for  the  production  of 

Coming  now  to  the  mannfactnre  of  acetylene,  it  is  to  be  regretted 
that  more  complete  and  accurate  data  cannot  be  had  as  to  its  use 
as  an  illuminant,  and  especially  in  the  United  States.  Brown's 
Directory  of  Gas  Companiee  records  184  acetylene  town  plants  in 
operation  the  first  of  this  year.  These  works  report  a  total  output 
of  18,500,000  cubic  feet.  A  paper  read  before  the  Illuminating 
Engineeriog  Society  in  1909  is  authority  for  the  statement  that 
there  were  at  that  time  390  towns  lighted  with  acetylene.  It  can 
be  understood  readily  that  the  record  in  Brown's  Directory,  de- 
pending for  its  facte  as  it  does  upon  answers  to  question  sheets, 
may*  be  quite  incomplete  by  reason  of  the  indifference  of  those 
in  control,  and  especially  so  in  case  of  the  municipal  plants. 

In  addition  to  the  acetylene  so  distributed,  the  total  is  con- 
siderably increased  by  that  used  in  private  houses,  contractors' 
plants,  car  lighting  and  portable  lamps,  particularly  automobile 

The  rate  charged  for  acetylene  bj  the  town  eompauies  seems  to 
run  from  1%  to  2  cents  per  cubic  foot,  or  $15  to  $20  per  1000 
cubic  feet.  Under  efficient  management,  as  to  installation  and 
operation,  these  rates  are  said  to  afford  a  fair  return  on  the  in- 

To  comprehend  the  precautions  to  be  taken  in  the  use  of  calcium 
carbide  and  acetylene,  there  must  be  borne  in  mind  the  difference 
between  esoihermic  and  endothermic  reactions. 

Exothermic  compounds  are  those  whose  formation  from  ele- 
mentary substances  is  attended  with  liberation  of  heat,  and  whose 
decomposition  into  simpler  compounds  or  elementary  substances 
is  attended  with  absorption  of  heat. 

>y  Google 

198  Illuminating  Enoineebino 

Endothenuic  compounds  are  those  whose  formation  from  ele- 
mentary aubstanceg  ie  attended  with  abtorpdon  of  heat,  and  whose 
decomposition  into  other  compounds  or  elementary  substances  is 
attended  with  liberation  of  heat. 

These  latter  compounds  are  not  very  numerous,  they  are  more 
or  less  unstable,  and  some  of  them  are  resolved  into  their  elements 
with   explosive   force. 

Acetylene  is  an  endothermic  compound. 

Acetylene  is  obtained  from  calcium  carbide  through  a  double 
decomposition.    The  first  step  is  shown  by  the  equation 

CaC,        +   H,0 
(Calcium  carbide)   (Water) 

C,H,     +  CaO  , 

(Acetylene)  (Calcium  oxide  or  lime)*       *■*' 

But  the  quicklime,  CaO,  in  the  presence  of  an  excess  of  water, 
will  be  found  in  the  form  of  slaked  lime,  or  calcium  hydroxide, 
Ca(OH)„  as  shown  by  the  equation 

CaO-t-H,0  =  Cft(OH)..  (2) 

As  these  reactions  in  the  presence  of  sufBcient  water  may  occur 
simultaneously,  the  double  reaction  can  be  shown  by  the  equation 

CaC,  +  2H,0  =  C,H,  +  Ca(0H)j.  (3) 

This  is  an  exothermic  reaction  because  the  quantity  of  heat  lib- 
erated exceeds  the  quantity  of  heat  absorbed.  There  is  some  little 
question  as  to  the  heat  of  formation  of  calcium  carbide,  authori- 
ties varying  from  —0.65  calories  (large)  to  +3.9.  But  these 
difFerencee  of  opinion  do  not  affect  the  question  as  to  whether  the 
reaction  as  a  whole  results  in  absorption  or  liberation  of  heat; 
it  only  affects  in  minor  degree  the  quantity  of  heat  liberated. 

The  heat  of  formation  of  Ca(OH),  (exothermic  substance)  is 
+  160.1  large  calories;  the  heat  of  formation  of  water  (exothermic 
substance)  is  +69,  and  hence  for  decomposition  is  —69;  taking 
heat  of  formation  of  calcium  carbide  as  +3.9,  for  decomposition 
it  is  —3.9.  The  heat  of  formation  of  acetylene  is  —58,1.  As  the 
formation  of  Ca(OH),  is  obtained  by  the  decomposition  of  the 
water  and  the  carbide  and  the  formation  of  the  acetylene,  we  have 
heat  liberated  in  the  formation  of  the  Ca(OH)j  160.1,  and  the 
heat  absorbed  as  follows : 

>y  Google 

Oas  and  Oil  Illuuinants  199 

Formation  of  acetylene  —  58.1  Endothermlc  substance. 

Decomposition  of  vater        —   69.     Exothermic-. 
Decomposition  of  carbide       —     3.9  Exothermic. 
Total        -131.0. 

Deducting  the  131  absorbed,  from  the  160.1  set  free,  we  have  as 
a  net  result  39.1  large  calories  liberated. 

While  this  reaction  as  a  whole  is  esothennic,  acetylene  as  a 
substance  is  seen  to  be  decidedly  endothermlc,  and  bo  is  ready  to 
liberate  large  quantities  of  heat  whenever  the  conditions  for  de- 
composition obtain. 

While  this  reaction  may  be  modified  it  should  be  pointed  out 
that  the  reaction  where  there  is  no  excess  of  water,  as  indicated 
in  equation  (1),  produces  in  practice  results  which  are  quite  dif- 
ferent from  those  obtained  where  there  is  excess  of  water,  as  indi- 
cated in  equation  (3). 

In  the  acetylene  generators  of  the  most  modern  and  usual  pat- 
tern, some  of  the  surplus  water  is  evaporated  by  the  heat  liberated, 
and  some  of  this  water  vapor,  even  at  low  temperatures,  is  carried 
away  with  the  escaping  gas.  If  the  heat  liberated  during  the  de- 
composition of  the  carbide  is  not  otherwise  absorbed,  it  is  sufficient 
in  amount  to  vaporize  almost  exactly  three  parts  by  weight  of 
water  for  every  four  parts  of  carbide  attacked.  But  if  this  quan- 
tity of  heat  were  expended  upon  some  substance,  such  as  acetylene 
or  calcium  carbide,  which,  unlike  water,  cannot  absorb  an  extra 
amount  by  changing  its  physical  state,  as  from  liquid  to  gas,  the 
heat  thus  generated  during  the  decomposition  of  the  carbide  would 
be  in  evidence  to  a  far  greater  extent.  For  reasons  that  can  be 
indicated  only  within  the  time  allowed  me,  it  is  essential  for  good 
working  that  the  temperature  of  both  the  acetylene  and  the  carbide 
shall  be  prevented  from  rising  to  any  considerable  extent. 

Experiments  were  conducted  by  Caro  and  by  Lewes  to  determine 
the  temperature  of  the  carbide  due  to  decomposition.  Caro's  ex- 
periments showed  a  maximum  temperature  of  280°  C.  (536°  F.). 
Lewes'  esperimenta  gave  a  maximum  temperature  of  807°  C, 
{1480°  F.).  The  temperature  attained  is  in  part  dependent  upon 
the  time  elapsed  in  the  reaction,  for  the  longer  the  time  the  greater 
the  opportunity  for  the  escape  of  heat  liberated.  The  divergence 
in  the  results  obtained  by  Caro  and  Lewes  is  explained  by  the 
difference  in  the  design  of  the  generators  and  the  speed  at  which 
they  were  operated.    In  Lewes'  generator  little  or  no  provision  was 

>y  Google 

200  Ili-umikating  Engineeriso 

made  against  overheating,  and  it  is  not  to  be  Bupposed  that  such 
temperatures  as  were  observed  by  Lewee  are  found  in  a  c»mmer- 
cial  generator.  But  his  determination  is  important  as  showing  the 
danger  to  be  avoided,  for  the  temperature  he  found  is  considerably 
above  that  at  which  acetylene  decomposes  into  its  elements  in  the 
absence  of  air,  namely,  780°  C.  or  1436°  F.  Excessively  high  tem- 
peratures in  the  generator  must  be  avoided,  because  whenever  the 
temperature  in  the  immediate  neighborhood  of  a  mass  of  calcium 
carbide  which  is  evolving  acetylene  under  the  attack  of  water  rises 
materially  above  the  boiling  point  of  water,  one  or  more  of  three 
objectionable  effecta  is  produced ;  namely,  upon  the  gas  generated, 
upon  the  carbide  decomposed,  or  upon  the  general  chemical  re- 
action then  taking  place.  Time  does  not  permit  a  full  discussion 
of  the  questions  here  involved,  but  a  few  hints  may  be  given. 

Lewes  points  out  that  not  only  does  acetylene  decompose  at 
780°  C,  but  it  begins  to  polymerize  at  600°  C.  (1112°  F.).  Sup- 
pose acetylene  polymerizes  into  benzene,  the  burner  adapted  to  the 
efficient  utilization  of  the  former  will  not  be  bo  adapted  for  ben- 
zene. Furthermore,  under  certain  conditions,  the  benzene  liquefies 
and  deposits  with  wat«r  vapor  in  the  pipes.  An  additional  trouble 
from  polymerization  occurs  when  the  temperature  rises  above  the 
point  at  which  benzene  is  formed,  for  then  other  hydrocarbons  may 
be  formed  having  a  higher  proportion  of  carbon  than  is  present 
in  acetylene  and  benzene,  setting  free  non-luminous  hydrogen,  and 
thus  reducing  the  illuminating  value  of  the  gaseous  mixture.  In 
certain  experiments  by  Lewes  the  loss  in  candle-power  was  found 
to  be  a  reduction  from  240  to  126.  Another  effect  of  heat  upon 
acetylene  has  already  been  indicated.  Being  an  endothermic  sub- 
stance it  gives  out  heat  upon  decomposing.  It  decomposes  at 
780°  C.  when  free  from  air,  a  spark,  or  shock,  or  pressure  of  30 
pounds  or  more  being  sufficient  to  effect  the  change.  This  change 
raises  the  temperature  and  so  increases  the  pressure  of  the  disasso- 
ciated hydrogen,  and  may  cause  the  containing  vessel  to  explode.  If 
air  is  present,  as  it  may  be  through  bad  design  of  apparatus  or 
incompetent  attendance,  the  acetylene  can  be  ignited  at  480°  G. 
(896°  F.).  Under  certain  conditions  25  per  cent  of  air  and  75 
per  cent  of  acetylene  are  explosive. 

The  extreme  limits  of  explosibility  of  acetylene  mixed  with  air 
are  variously  stated.  Clowes  gives  the  extremely  wide  range  of 
explosibility  from  3  per  cent  to  83  per  cent  of  acetylene.     Le 

>y  Google 

(iks  AKD  Oil  Illuhimants  SOI 

Chatelier  givca  8.9  per  cent  to  64  per  cent.  Eitner  made  exhaustive 
teste  vith  several  gaeee,  in  each  case  the  mutnre  being  saturated 
with  aqueovs  vapor,  thus  reducing  tke  limits  of  explostbUity.  For 
aoetylene  he  gives  from  3,35  to  52.30  per  cent.  Teclu,  experi- 
mentiDg  with  a  dry  mixture,  determiDed  the  limits  as  1.53  to  59 
per  cent.  These  results  naturally  are  changed  if  the  mixture  con- 
tains other  gases  besides  acetylene  and  air,  but  enough  has  been 
said  to  show  that  acetylene  cannot  be  handled  carelessly.  This 
is  emphasized  by  Eitner's  ezperimentfl,  comparable  bat  not  giving 
extreme  limits,  which  gave  as  the  limits  for  coal  gas  7.90  to  19.10 
per  cent,  or  a  range  of  only  II.SO  per  cent  against  acetylene  range 
of  48.95  per  cent,  as  shown  above. 

In  the  generator  the  effect  of  heat  on  the  carbide  itself  may  be 
troublesome.  If  part  of  the  gas  polymerizes  part  may  so  be  re- 
solved into  tar,  which  coats  the  carbide  still  nnattacked  and  so 
protects  it  more  or  lees  from  further  attack,  thus  reducing  the 
output  and  leaving  the  residue  with  a  content  of  acetylene,  which 
may  later  occasion  trouble  during  or  after  removal. 

The  effect  of  accumulating  heat  in  the  generator  itself  has  to 
be  guarded  against.  For  example,  at  a  temperature  ae  low  as 
800°  C.  (398°  F.),  if  the  ordinary  solder  were  used  in  the  joints 
it  would  be  melted  and  the  vessel  become  unsafe.  This  serves  to 
point  to  the  fact  that  the  materials  used  and  the  minor  details  of 
construction  in  a  generator  may  be  such  as  to  condemn  a  design 
generally   commendable. 

Having  indicated  most  superficially  some  of  the  conditions  to 
be  considered  in  the  design  and  construction  of  acetylene  genera- 
tors, with  the  aid  of  diagrams  taken  from  Leeds  and  Bntterfield's 
work  entitled  "  Acetylene,  Its  Generation  and  Use,"  I  shall  show 
in  a  general  way  how  these  conditions  are  met,  but  without  at- 
tempting to  discuss  the  relative  advantages  and  disadvantages  of 
the  several  types. 

Acetylene  generators  may  be  roughly  claasified  as  follows: 
1st.  Carbide  to  water. 

(a)  Non-automatic. 

(b)  Automatic. 
2d.   Water  to  carbide. 

(a)  Non-automatic. 

(b)  Automatic. 

>y  Google 


Illumikatinq  Enqineerikq 

In  general,  the  type  having  the  widest  limits  of  safety  is  that  in 
which  a  small  quantity  of  carbide  is  introduced  into  a  considerable 
body  of  water,  the  acetylene  as  it  bubbles  through  the  water  passing 
directly  ont  and  into  a  holder.  If  this  holder  has  ample  capacity 
for  the  maximum  night's  demand,  it  can  be  filled  with  gas  during 
the  day  and  the  generator  locked  for  the  night.  This  non-auto- 
matic form  may  be  criticized  on  the  ground  of  tlrst  cost. 

Fio.  13.— Acetylene  Generator.    Non-Automatle.    Carbide  to  Water  Type. 
Fio.  14.— Acetylene  Generator.    Automatic.    Carbide  to  Water  Type. 

If  the  introduction  of  carbide  is  controlled  by  an  automatic 
device  which  admits  carbide  automatically  as  the  acetylene  is  con- 
sumed, a  smaller  generator  and  holder  can  be  employed. 

Figs.  13,  14  and  15  show  types  of  carbide  to  water  generators. 

Fig.  13  represents  the  non-automatic  type.  The  carbide  is  fed 
by  hand  through  the  chute  A  into  the  generator  B.  The  generator 
is  filled  with  water  above  the  opening  of  the  chute  to  prevent  the 
gas  from  escaping  through  the  chute.  Grids  D  and  E  catch  and 
support  the  lumps  of  carbide,  permitting  the  acetylene  to  be  com- 

>y  Google 

Oas  and  Oil  Illuhinants  203 

pletely  liberated  before  permittiiig  the  mass  to  mix  with  the  gtndge 
of  slaked  lime  in  the  bottom  of  the  tank.  The  carbide  cannot  be 
used  in  small  lamps,  as  then  the  generation  of  acetylene  would  be 
sufficiently  active  to  blow  the  seal  and  allow  the  gas  to  escape 
through  the  chate. 

Fig.  14  shows  an  automatic  generator  of  the  £rst  class.  The 
carbide  is  held  in  a  hopper  which  is  supported  by  holder  bell  I, 
which  rises  and  falls  according  to  the  volume  of  acetylene  con- 
tained.    The  hopper  is  closed  at  the  bottom  by  a  valve  Q,  from 





\    P3 


\    P2 


\      CtCM 


\    ^ 


PlO.   IS. 

Pto.  16. 

Fio.  1G.~ Acetylene  Generator.    Automatic  Dipping.    CarblOe  to  Water 

Fid.  16. — Acetylene  Generator.    Water  to  Carbide  Type.     Water  Inlet  at 

which  depends  a  rod  H.  As  acetylene  is  withdrawn  from  the  bell 
the  bell  falls  until  the  rod  strikes  the  bottom  of  the  tank,  the  valve 
is  thus  forced  open  permitting  more  carbide  to  fall  into  the  water, 
more  acetylene  la  released,  the  bell  again  rises  until  the  valve  seat 
and  valve  engage,  when  the  supply  of  carbide  is  again  stopped. 

Fig.  15  shows  a  dipping  generator.     The  carbide  is  held  in  a 
perforated  vessel  which  hangs  from  the  inside  of  the  crown  of  the 

>y  Google 


Illuminatiko  Enoineebinu 

holder  bell.    As  the  acetylene  is  consumed  the  beU  falls  until  the 
carbide  dips  in  the  water,  when  acetylene  is  again  liberated. 

Figs.  16,  17  and  18  show  types  of  water  to  carbide  generators. 
Fig.  16  shows  a  generator  in  which  the  carbide  is  contained  in  a 
series  of  pans,  P,  Fl,  PS  and  P3,  a  small  quantity  in  each  pan. 
Water  is  admitted  at  the  bottom  through  pipe  M,  As  each  pan 
is  flooded  the  acetylene  rises  to  the  top  of  the  tank  and  passes  out 



\  ^ 



J     "  c,c,  1 


J  n  c,c.  1 

i-tCl    1 

Fia.  17. — Acetrleae  Generator.     Water  ti 
at  Top. 

Carbide  Type.    Water  Inlet 

at  B.  The  gas  passing  out  is  charged  with  water  vapor,  and  this 
water  acting  upon  the  carbide  in  the  upper  pans  produces  "  after 
generation,"  which  is  an  objectionable  feature. 

Fig.  17  shows  a  better  type.  The  carbide  is  contained  in  pans 
as  in  the  previous  case.  Here  the  water  is  admitted  at  the  top 
of  the  tank  and  first  acts  on  the  carbide  in  the  top  pan.  The  gas 
passes  oS  wi^out  coming  in  contact  with  the  carbide  in  the  other 
pans.  As  the  first  pan  is  flooded  the  water  overflows  throng^  the 
pipe  S  to  the  second  pan.  This  is  repeated  until  the  carbide  in  the 
last  pan  is  attacked.  The  acetylene  escapes  from  the  pipe  at  the 
top  of  the  tank,  as  shown. 

>y  Google 

Oas  and  Oil  Illuminants  305 

Fig.  18  shows  a  generator  not  to  be  commended.  The  carbide 
is  contained  in  the  tank.  T.  Water  enters  at  the  top  in  drops  or 
a  fine  Etream.  ThiB  tjrpe  produces  "after  generation"  and  dan- 
gerous overheating. 

Generally  speaking,  in  the  water  to  carbide  generators  the  gen- 
erator is  opened  to  the  air  while  being  charged  with^freah  carbide; 
this  is  a  decided  disadvantage,  for,  as  already  shown,  acetylene 
should  be  guarded  from  mixing  with  air  on-  account  of  its  wide 
range  of  eiplosibility. 


Fio.  18. — Acetylene  Generator.    Water  to  Carbide  Type.    Crude  Form. 

What  I  have  said  fails  to  show  the  great  variety  of  apparatus 
actually  employed  or  the  manner  in  which  the  several  types  merge 
into  eacli  other.  I  have  not  attempted  to  show  the  complete  acety- 
lene installation,  including  t)ie  parts  for  generation  and  governing. 
It  should  be  pointed  out  tliat  it  is  necessary  cither  to  use  a  pure 
carbide  or  provide  means  for  purifying  the  acetylene,  as  otherwise 
compounds  of  phosphorus,  silicon,  ammonia  and  sulphur  might 
be  present  rendering  the  gas  objectionable  on  the  score  of  spon- 
taneous inflammability  or  non-hygienic  qualities.  Leeds  and  But- 
terfield's  work  give  the  rules  and  regulations  adopted  by  govern- 
ments and  insurance  oonipanics  for  the  construction  and  installa- 
tion of  acetylene  plants. 

The  carbide  is  sold  in  several  sizes.  For  generators  the  size 
varies  from  31^  inches  by  2  inohes  down  to  Vj  inch  by  1/12  inch. 
For  lamps,  from  1  inch  by  V-^  inch  down  to  dust.     The  rate  of 

>y  Google 

206  Illdminatinq  Enqineekino 

evolutioii  IB  invereely  proportioD&l  to  the  aiie  of  the  lump.  Lumps 
coated  with  dust  nuj  pve  insularity  in  operation. 

Acetylene  liquefies  at  0°  C.  and  about  21%  atmoepherea  pres- 
sure. It  is  then  most  unstable,  Bpontaneous  disasBociation  with 
explosive  force  being  due  to  occur  on  the  application  of  a  spark  or 
when  shocked.  After  quite  a  number  of  disastrous  accidenta  it  is 
now  generally  understood  that  liquid  acetylene  is  too  dangerous 
to  use.  As  before  mentioned,  the  gas  is  liable  to  explode  if  heated 
to  780°  C.  or  if  held  under  a  pressure  of  2  atmospheres  absolute, 
or  above. 

Acetylene  le  readily  soluble  'in  many  liquids,  and  this  property 
is  utilized  to  bring  the  acetylene  into  small  compass.  Acetone,  at 
ordinar}'  temperature  and  atmospheric  pressure,  will  dissolve  about 
25  volumes  of  acetylene,  and  at  13  atmospheres  vdll  dissolve  about 
300  volumes.  Acetone  is  an  exothermic  substance  with  a  composi- 
tion shown  by  the  formula  C,H,0,  and  hence  combustible,  and 
within  certain  limite  of  pressure  its  presence  tends  to  decrease  the 
severity  of  explosion.  At  20  atmospheres  pressure  the  acetone  adds 
to  the  danger  from  explosion.  Acetylene  dissolved  in  acetone  car- 
ried up  to  a  pressure  of  10  atmospheres  is  safely  employed,  but 
tbere  are  practical  objections  to  its  use  in  this  liquid  form.  To 
overcome  these  objections  the  cylinders  are  filled  with  some  porous 
material  which  does  not  react  on  the  acetone.  A  material  is  used 
which  has  a  porosity  of  80  per  cent,  that  is,  when  the  vessel  is 
apparently  full  of  the  material  about  SO  per  cent  only  of  the  space 
is  really  occupied.  The  portable  cylinders  for  this  serrice  cannot 
be  filled  with  acetone,  for  the  reason  that  ample  space  must  be  left 
for  expansion  as  the  liquid  takes  up  the  gas.  A  cylinder  having  a 
normal  capaci^  of  100  volumes  will  have  say  20  volumes  taken 
up  by  the  porous  filling,  and  can  safely  be  charged  with  40  volumes 
of  acetone.  This  40  volumes  of  acetone  dissolves  40x25  =  1000 
volumes  of  acetylene;  and  by  compression  to  10  atmospheres  this 
is  increased  to  10,000  volumes.  In  this  form  acetylene  is  sold 
under  various  trade  names  and  used  for  automobile  head  lights 
and  similar  service  where  limited  storage  capacity  is  of  decided 

Acetylene,  under  favoring  conditions  including  moisture,  will 
combine  with  copper  to  form  aeetylide  of  copper,  an  explosive 
compound.  As  acetylene  is  now  generally  produced  and  used  these 
conditions  are  not  apt  to  obtain,  so  the  danger  from  this  source  is 

>y  Google 

Gas  and  Oil  Illukikakts  207 

now  not  regarded  serionslj.  Copper  alloys  and  compounds  should 
not  be  employed  in  the  conBtruction  of  parts  of  plant  exposed  to 
the  gfa  oi  in  the  process. 

Straight  acetylene,  burned  in  an  open-flame  burner  of  a  char- 
acter and  size  beat  adapted  to  give  the  highest  illumioating  value, 
the  burner  being  so  placed  as  to  carry  to  the  photometer  disc  the 
strongest  horizontal  rays,  the  bar  readings  being  calculated  pro  rata 
to  a  consumption  of  5  feet  an  hour,  gives  a  candle-power  of  from 
240  to  250. 

The  general  practice  of  selecting  the  burner  and  rate  of  con- 
sumption so  as  to  develop  best  efBciency  of  the  gas  instead  of  being 
confined  to  one  type  of  burner  and  a  rated  consumption  of  5  feet 
an  hour,  has  received  the  approval  of  the  Oas  Referees  of  London 
acting  under  Parliamentary  powers. 

The  specific  gravity  of  acetylene  is  .9056,  usually  taken  as  .91. 

For  self-luminous  fiamea,  lava-tip  burners  are  employed,  the 
gas  issuing  either  from  a  slot  or  two  holes,  both  producing  fiat 
flames.  With  the  latter  form  the  flat  flame  is  produced  by  the 
impinging  of  the  two  currents  of  gas  against  each  other,  the  plane 
of  the  flame  being  produced  at  right  angle  to  the  plane  of  the  two 
holes.  The  burners  are  made  in  many  different  forms  in  the  effort 
to  overcome  difi^culties  due  to  the  richness  of  the  gas  and  its 
instability.  The  richness  of  the  gas  made  it  necessary  to  employ 
small  burners  or  to  make  extra  provision  for  injecting  air  into  the 
body  of  the  flame  by  the  action  of  the  issuing  gae.  This  was  best 
accomplished  by  some  form  of  two-jet  burner,  which  dragged  in 
the  air  at  a  point  between  the  jets  and  below  the  flame.  To  better 
accomplish  this  result  burners  were  devised  with  two  tips  so  as  to 
separate  farther  the  two  jete  of  gas.  Further,  to  assist  in  the  in- 
jection of  air,  ace^lene  is  burned  at  a  pressure  far  in  excess  of 
that  employed  with  coal  gas.  Its  high  specific  gravity  also  calls 
for  additional  pressure.  The  design  of  acetylene  burners  well  il- 
lustrates that  burners  must  be  designed  to  supply  such  a  quantity 
of  air  to  the  fiame  as  will  produce  a  maximum  incandescence.  If 
one  of  these  burners  were  used  with  coal  gas,  so  much  air  would 
be  dragged  in  that  the  carbon  particles  of  this  thinner  gas  would 
be  consumed  with  little  or  no  preliminary  incandescence. 

The  comparatively  high  efBciency  of  the  acetylene  flame  is  due 
not  alone  to  the  high  carbon  content;  an  important  factor  is  the 
high  flame  temperature,  which  is  in  part  the  result  of  liberation 

>y  Google 

808  Illumixatixo  Enoixeebixq 

of  heat  at  time  of  disassociation  of  this  cndothermic  gas.  Mahler 
gives  t!ie  flame  temperature  at  0°  C.  and  760  mm.  as  2350°  G.  or 
4642°  F.     Le  Chatelier  gives  2100°  C.  to  2400°  C. 

Acstylene  is  also  employed  with  incandescent  mantles,  TesttltJDg 
in  a  considerable  increase  in  candle-power,  this  gain  according  to 
different  authorities  being  from  160  to  200  per  cent.  For  certain 
special  applications  a  still  lai^r  gain  has  been  secured.  There 
are  decided  difficulties  to  be  overcome  and  advantages  to  be  aban- 
doned in  using  acetylene  for  incandescent  lighting,  and  the  high 
efficiency  and  the  whiteness  of  the  self-luminous  flame  make  it 
less  necessary  or  desirable  to  overcome  these  difficulties. 

Acetylene  is  also  employed  for  illumination  in  the  form  of  car- 
buretted  acetylene  or  carburylene,  and  in  this  form  it  is  more  ad- 
vantageously applied  to  incandescent  lighting,  but  time  does  not 
permit  a  discussion  of  this  branch  of  the  subject. 

In  connection  with  illuminating  engineering,  the  color  of  the 
acetylene  flame  is  of  great  importance.  A  comparison  with  sun- 
light and  other  light  sources  will  be  given  in  another  of  these 

Acetylene  can  also  be  used  for  heating.  Its  calorific  value  per 
foot  is  363  large  calories,  or  1440  B.  t.  u.,  which  is  about  two  and 
one-hnlf  times  that  of  city  gas.  The  comparison  is  not  favorable 
to  acetylene,  however,  when  relative  costs  are  considered. 

Within  the  limits  of  a  single  lecture,  inordinately  long,  it  is 
true,  I  have,  according  to  instructions,  endeavored  to  cover  three 
sources  of  illumination.  Many  lectures  could  be  devoted  advan- 
tageously to  each  of  these.  Acetylene  alone  could  not  be  covered 
completely  in  many  lectures. 


Pi.NTScii  Gab 
King's  Treatise  on  the  Manufacture  of  Gae.    Volume  HI, 
The  Comparative  Merita  of  Varioua   Systems  of  Car  Lighting:     A,  M. 

Wellington.  W.  B.  D.  Penniman,  Charles  Whftlng  Baker.    Engineer- 

Ing  News  Publishing  Company,  New  York,  1S92. 
Engineering  Chemistry:     Thomas  B.  Stillman.    The  Chemical  Publlsb- 

Ing  Co.,  Baston,  Pa..  1910. 
Car.  Lighting:     R.  M.  Dixon.     Stevens  Institute  Indicator,  Vol.  XXV, 

No.  1.  Jan.,  1908. 
Lighting  of  Railway  Cars:     Geo.  E.  Hulae.    Transactions  of  the  iiiumt- 

natlng  Engineering  Soc,  Vol.  V,  No.  1,  January,  1910. 

>y  Google 

Gas  akd  Oil  Illuhinants  209 

Car  Lighting:     L.  R.  Pomeroy.     Froceedlnga  Canadlaii  Railway  Club, 

Vol.  IX.  No.  2,  February,  IBIO. 
Leuchtf«uer  uod  NflbelalgDal:     E.  Klebert.    Journal  fur  Oasbeleuchtung 

und  WaBMirersorgung,  May,  1909. 
OelgaaauBtalt  mlt  Qeneratorbetrlsb:     PrlU  I^indaberg.    Zeltscbrltt  dea 

Verelnes  DeutacbeV  iDcenletire,  Nr.  37,  Band  62,  September,  1909. 
Oelgashentellung   In    Oeneratoreu    nod    GasTermverBorgung   In    Hocb- 

druckleltung:     Fritz  L&ndsberg.    Qlaser'a  Annual,  AugUBt  1,  1910. 
Llgbtlng  of  Paaaenger  Can:     Max  Buettner.     Publlahed  by  Springer, 

Berlin,  1901. 
Fetrotoum  and  tta  Products:    Vol.  II,  Sir  Boverton  Redwood.    Published 

by  Charles  Qrlffln  A  Co.,  Ltd.,  London,  England,  ISOG.    Brief  descrip- 
tion under  oU  gas.    , 

Petroleum  and  Its  Prodacta:  2  Vols.  Sir  Boverton  Redwood.  Pub- 
llahed by  Charles  GrlAn  A  Co.,  Ltd.,  London,  England,  1906.  Tbls 
worb  contains  a  very  lull  bibliography. 

Petrol  Alr-Qas;  Henry  O'Connor.  Publlahed  by  Crosby  Lockwood  C 
Son,  London,  England,  1909. 

Acetylene:     The  Principles  of  Its  Generation  and  Use  by  F.  H.  Leeds 

and  W.  J.  AtklDBon  Butterfleld.    Published  by  Charles  GrllSn  4  Co., 

Ltd.,  London,  England,  1910. 
Calcium  Carbide  and  Acetylene  by  Geo.  Gilbert  Pond.    Bulletin  of  tbe 

Department  of  Chemistry  of  the  Pennsylvania  State  College.  1909. 

This  work  contains  a  full  bibliography. 

>y  Google 




>y  Google 



By  M.  C.  Whitakeh 


Heat  Bourcee:    chemical,  electrical. 

Substance;  gaa,  wood,  coal,  etc 

Supporter  of  combuatlon;    oxrgen,  air. 

Kindling  temperature:   electric  spark,  lighted  match,  etc. 
Chemlstrr  of  combustion. 

Marsh  gas  -i-  oxygen  =:  water  vapor  +  c&rboa. 

Carbon  -|-  oxygen  ^  carbon  dioxide. 
Open  tip  combuHtlm. 
Bunsen  burner  combustion. 

iNCuiDcecEKT  Gas  Illumination 
Priaclplea  InTOlved, 
History;      Hare,   Drummond    and   Claymond   lights.   Slemens-Lungrsn 

lamp.  Auer  von  Welsbach  lamp. 
Bunsen  burner:    historjr,  construction  and  chemistry  ot  operation. 
Adaptations  tor  use  with  incandescent  mantles. 
Upright  and  Inverted. 
Single,  cluster  and  arc. 
Inside  and  outside. 
Upright  burner  construction. 
Bunsen  tube. 

Check  for  gas;  plate,  needle,  multiple  hole,  check;  nlr  adjustment; 
ganzes;  gallery. 
Inverted  bnrner  construction. 

Types:     vertical,    horizontal    and    goose-neck    burners;     velocity. 

gravity  and  buoyant  action  In  dovraward  flow  of  mixture. 
Checks  for  gas;  air  adjustment;  means  [or  overcoming  flash-backs; 
crown  for  glassware. 

Oas  Mantles 
Process  of  manufacture:    history,  knitting,  washing,  saturating,  incin- 
erating,  shaping,  collod ionizing,  trimming  and   inspecting,   packing 
and  shipping. 
Physical  structures  of  mantlee. 

Basic  flhers:  cotton,  ramie,  artificial  silk. 
Threads,  weaves,  stitches,  etc. 

>y  Google 

£18  Illuuinatino  Ehoisebbing 

Cbem teals  and  sonreefl. 

Lighting  principles;  thorium  and  cerium. 

Thorium;  source  (monazite,  tborianite) ;  manufacture,  market,  use. 
Collodion;  compoaltimi,  manufacture,  use. 
Types  of  mantles:    upright  and  Inverted,  railroad  train,  j^ces,  pressure, 

ng,  acetylcQe,  kerosene,  etc. 
Quality  and  serrlce  characteristics  as  determined  by  process  of  manu- 
Cotton:     shrinkage,    depreciation    In    candle-power;    color    value; 

phyBlcal  strength. 
Ramie:  ditto,  etc. 
Artlflclal  silk:    ditto,  etc. 


Assuming  that  the  illuminatiDg  power  of  a  gas  flame  is  derived 
from  the  heating  of  solid  particles  to  incandescence,  the  practice  of 
gas  illumination  divides  itself  into  two  general  principles: 

First.  Where  the  solid  incandescent  material  is  supplied  by  the 
decomposition  of  the  gas  in  the  process  of  combuation.  (Open-tip 

Second.  Where  the  gae  is  completely  consumed  in  a  Bunsen 
burner  for  the  production  of  the  maximum  amount  of  heat  and 
a  pennanent  incandescent  material  is  supplied  as  a  part  of  the 
apparatus.    (Incandescent  gas  system.) 

The  first  steps  toward  the  improvement  of  the  efficiency  of  gas 
for  lighting  was  made  on  the  first  of  these  principles  by  pre- 
heating the  gas  before  it  reached  the  point  of  combuation  in  the 
so-called  regenerative  burner  of  the  Siemens-Lungren  or  Gordon- 
Mitchell  type  (Kg.  1).  There  are  some  of  these  regenerative 
lamps  in  use  at  the  present  time.  The  regenerative  burner  was 
the  most  effective  ever  produced  by  following  the  first  principle 
mentioned  above,  and  gave  the  most  efficient  results  up  to  the  in- 
troduction of  the  incandescent  mantle  system,  which  is  based  on 
the  second  principle. 

Professor  Robert  Hare  (Philadelphia  Chemical  Society,  1802) 
first  fully  described  a  form  of  "incandescent"  gas  light,  which  is 
the  basic  principle  now  utilized  in  this  industry. 

At  a  meeting  of  the  Philadelphia  Chemical  Society,  held  in 
December,  1801,  he  showed  experiments  and  described  this  in- 
candescent lime  light  as  follows: 

"  The  cock  of  the  pipe  communicating  with  the  hydrogen  gas  was 
tben  turned  until  as  much  was  emitted  from  the  orifice  of  the  cylinder 
as  when  lighted  formed  a  flame  smaller  In  size  than  that  of  a  candle. 

>y  Google 

Incandescent  Qas  Mantles 


Under  tbls  flame  was  placed  the  body  to  be  acted  on,  supported  either 
by  charcoal,  or  by  some  more  Bolld,  and  incombustible  subatance.  The 
cock  retaining  the  oxygen  gaa  was  then  turned  until  the  light  and  beat 
appeared  to  have  attained  the  greatest  Inteneity.  When  thla  took  place, 
the  eyes  conld  scarcely  sustain  the  one,  nor  could  the  most  refractory 
substances  reelat  the  other." 

Fig.  1. — Regenerative  Lamp. 

Fw.  3.— Drummond  Calcium  Light 

Henry  Dniinmond,  in  1826,  made  use  of  the  incandescent  lime 
light,  similar  to  that  euggested  by  Professor  Hare,  for  signaling. 
Dnimmond's  application  of  thiB  principle  of  producing  an  illnmi- 
nation  of  high  intensity  was  adopted  generally,  and  he  is  usually 
credited  with  the  invention.  The  lime  tight  i&  sometimes  called 
the  "Drummond  light"  (Fig.  2). 

At  the  Crystal  Palace  Exposition  in  Paris  in  1833  a  lamp  of  the 
inverted  type  was  shown  in  which  the  illumination  was  produced 
by  a  platinum  basket  suspended  in  a  blast  flame.  The  life  of  the 
basket  was  limited  to  50  or  60  hours. 

Various  other  lamps  for  the  application  of  the  principle  of  sup- 
plying the  incandescing  material  were  suggested,  such  as  cones 

>y  Google 

214  iLLnillNATINQ    ENQINKKtI»0 

made  from  platinum  wires  covered  with  a  refractory  coating,  per- 
forated baskets,  grids  placed  above  the  flame  of  the  fish-tail 
burner,  etc. 

The  greatest  step  in  the  development  of  a  commercial  incan- 
descent gas  light  was  made  by  Dr.  Carl  Auer  von  Welsbach,  In 
1886,  Dr.  Auer,  while  a  student  in  the  laboratory  of  Professor 
Robert  Bunsen,  in  Heidelberg,  discovered  that  the  ash  formed  by 
saturating  a  cotton  fabric  in  a  solution  of  erbium  salts  and  burning 
out  the  organic  matter  would  take  the  shape  of  the  original  fibers, 
and  would  adhere  to  form  a  mesh  of  considerable  strength.  This 
finely  divided  ash  fabric,  when  suspended  in  the  fiame  of  a  Bunsen 
burner,  became  intensely  luminous.  Erbium,  however,  produces 
green  light.  Nevertheless,  the  principle  of  forming  an  attenuated 
but  closely  adhering  ash  was  established  by  Dr.  Aner  in  these 
experiments,  and  he  immediately  proceeded  to  develop  this  idea 
with  a  view  to  producing  a  commercially  desirable  light  by  heating 
oxide  webs  which  he  called  "  stockings  "  or  mantles. 

His  earfy  mantles  were  made  from  a  mixture  of  lanthanum  and 
zirconium  oxides.  The  light  given  by  this  mixture  was  not  sat- 
isfactory, and  the  investigation  was  continued  until  he  discovered 
the  wonderful  luminescence  obtained  with  a  mantle  made  from 
the  rare  oxides  of  thorium  and  cerium. 

Incandescent  Burners 

The  earlier  burners  constructed  to  use  Auer's  invention  were  de- 
signed for  use  with  the  lanthaniim -zirconium  mantle,  which  did 
not  give  the  high  candle-power  given  by  the  present  mantle.  These 
burners  were  consequently  very  large  and  clumsy  and  bore  a  very 
remote  resemblance  to  the  modern  types. 

Modem  Types 

The  present  practice  in  incandescent  burner  construction  should 
be  divided,  for  clear  discussion,  into — 

First.     Individual  upright  burners. 

Second.    Individual  inverted  burners. 

Third.    Gas  arc  lamps  (upright  burnera). 

Fourth.     Gae  arc  lamps  (inverted  burners). 

Fifth.  Lamps  for  special  application  (pressure  oil  lamps,  rail- 
way coach  lamps,  kerosene  lamps,  etc.). 

>y  Google 


Upright  Burners 
The  functional  parts  of  the  upright  incandescent  burner  may  be 
divided  into  (Fig.  3) : 

(a)  Bunsen  tube. 

(b)  Bunsen  base. 

(c)  QaB-adjuBtmeut  means. 

(d)  Air-adjustment  meanB. 

(e)  Mixing  chamber. 

(f)  Gallery  to  support  chimneys,  glassware,  reflectors,  etc. 

Ou  tdluatniHit. 
Air  idjuatment. 

Fio.  3. — Upright  Burner  Gut  to  Show  Interior  Gonetructlon. 

The  Bunsen  tube  is  carefully  designed  to  meet  a  wide  range  of 
gas  conditions,  such  as  fluctuations  in  pressure,  gravity,  etc.,  and 
still  produce  a  mixture  which  has  entrained  the  proper  quantity 
of  air  to  produce  complete  combustion  at  the  gauze  line.  The 
dimeiisions  of  this  tube  have  been  carefully  determined  in  ex- 
perience, and  are  comparatively  uniform  in  all  styles  of  standard 

The  Bunsen  base  is  usually  turned  from  solid  brass  bar  and 
threaded  internally  to  fit  the  average  run  of  i/^-indi  gas  nipples 

>y  Google 

816  Illdminatino  Enoineeeino 

rather  than  any  stanrJard  thread.  This  base  is  also  adapted  to 
carry  the  gas-adjusting  device  and  to  form  an  assembly  base  for 
the  entire  lamp. 

A  gas  adjustment  is  an  essential  feature  of  the  standard  burner 
used  in  this  country.  Some  foreign  burners,  and  many  of  the  early 
burners  in  this  country,  had  a  fixed  gas  orifice,  but  the  variation 
in  density  and  pressure  of  the  gas  in  different  localities  have  com- 
pelled the  modem  gas  burner  to  include  some  means  for  adjusting 
the  gas  flow. 

There  are  several  prevailing  types  of  gas  checks,  some  of  which 
fulfil  the  function  of  regulating  the  flow  of  gas,  but  fall  far  short 
of  meeting  other  essential  requirements. 

The  efficiency  of  burners  of  this  type  is  lai^ely  dependent  upon 
the  velocity  of  the  gas  jet,  and  its  consequent  ability  to  entrain  the 
amount  of  air  necessary  to  produce  complete  combustion.  Any 
construction  which  tends  to  cut  down  this  jet  velocity  seriously 
affecte  the  efficiency  and  proper  operation  of  the  burner  unless  the 
initial  gas  pressure  is  high  enough  to  produce  a  proper  jet  velocity 
in  spite  of  the  design  of  the  check.  Low  and  variable  pressures  are 
common  conditions  and,  as  a  consequence,  must  be  provided  for 
in  all  designs  intended  for  general  sale  and  use. 

A  single  round  hole  through  a  thin  plate  offers  the  minimum 
amount  of  resistance  to  the  flow  of  the  gas  stream  and,  as  a  con- 
sequence, gives  the  maximum  jet  velocity  in  the  Bunsen  tube.  An 
iris  diaphragm,  similar  to  the  device  iieed  in  a  camera,  has  been 
suggested  as  the  ideal  way  to  construct  an  adjustable  single-hole 
check,  but  the  cost  of  construction  and  the  mechanical  difficulties 
involved  in  making  it  gas-tight  prevent  its  general  adoption. 

Among  the  adjustable  chficks  in  general  use  the  preferred  types 
seem  to  be  included  in  the  following  general  designs: 

First.  The  Mason  check,  which  is  a  aeries  of  round  holes  in 
superimposed  plates,  one  of  which  may  be  rotated  upon  the  other 
in  such  a  way  as  to  bring  more  or  fewer  holes  into  action,  depending 
upon  the  direction  of  rotation.  While  the  number  of  small  holes 
offers  somewhat  more  friction  to  the  flow  of  the  gas  than  the  ideal 
single  bole,  it  is  thought  that  this  device,  which  ia  capable  of 
economical  and  reliable  mechanical  construction  gives  the  most 
efficient  results  over  the  widest  range  of  conditions.' 

Second.  The  annular-orifice  check  is  produced  by  inserting  a 
needle  in  a  single  round  hole  and  providing  a  mechanical  construc- 

>y  Google 

Incandbbcent  Gas  Maittlbs  217 

tioii  which  permits  the  needle  to  be  drawn  in  or  out  in  relation  to 
a  stationary  hole,  or  a  stationary  needle  with  a  cap-orifice  arranged 
to  be  raised  and  lowered.  Obviously,  the  annular  orifice,  which 
may  give  satisfactory  results  with  favorable  conditions,  will  offer 
unfavorable  resietance  to  the  flow  of  the  gas  on  lower  pressures  and 
thereby  affect  the  mixture. 

Flo.  4. — Upright  Burner. 

Adjustment  of  the  air  supply  is  usually  automatically  taken  care 
of  hy  the  regulation  of  the  gas  flow  when  the  composition  and 
other  conditions  are  normal.  Certain  gases  require  some  extra 
provision  for  air  adjustment,  and  most  upright  burners  are  now 
80  constracted  as  to  permit  of  this  equipment,  if  necessary. 

The  mixing  chamber  is  the  enlarged  portion  &t  the  top  of  the 
Bunsen  tube,  and  exercises  an  important  function  in  producing 
a  more  intimate  mixture  of  gas  and  air,  and  also  serves  as  a 
mounting  for  the  mantle. 

The  function  of  the  gallery  is  obviously  for  supporting  the 
chimney,  globe,  reflector  or  other  equipment. 

Modern  burner  design  is  carried  out  on  the  best  scientific  lines 
with  a  view  to  producing  a  burner  satisfactory  for  all  gas  condi- 

>y  Google 


Illdminatinq  Snoineebinq 

tions.  Adjustable  gas  checks,  automatically  mixing  air  supply, 
properly  proportioned  Bunsen  tubes  and  mixing  chambers,  a  shapely 
exterior  construction  vith  the  finest  material  and  workmanship, 
make  the  modern  burner  a  very  effective  and  artistic  appliance 
(Fig.  4). 

Fio.  5, — Clarmottd  Inverted  Lamp. 

Inverted  Burners 

The  most  important  step  in  the  improvement  of  gas  illnmina- 

tion,  embodying  the  use  of  the  Welsbach  mantle,  has  been  the 

>y  Google 

Tncandiscbnt  Gas  Mantles  219 

commercial  introduction  of  the  inverted  incandescent  lamp.  The 
reasons  which  underlie  the  rapid  development  of  this  lamp  appear 
to  be  improved  efficiency,  direct  downward  distribution  of  the  light, 
decorative  possibilities  and  more  durable  mantles. 

The  first  inverted  incandescent  lampe  were  made  by  Claymond 
and  exhibited  in  1883-1883  (Fig.  5).  CoDsiderable  activity  was 
subsequently  shown  by  inventors,  and  numerous  forme  were  ex- 
ploited without  commercial  succese.  With  the  advent  of  the  Wels- 
bach  upright  mantle,  this  line  of  research  was  abandoned  and  no 
developments  of  any  consequence  were  made  for  8  years. 

Interest  in  the  inverted  light  was  renewed  by  the  exhibition  in 
Germany  of  a  burner  for  the  thorium-cerium  mantle  in  1900, 
These  inverted  lamps  did  not  meet  vrith  marked  commercial  appli- 
cation in  this  country,  because  their  designers  failed  to  take  into 
account  the  principles  which  modern  inverted-burner  builders  recog- 
nize as  basic. 

The  history  of  the  modern  inverted  burner  is  confined  almost 
entirely  to  the  development  of  methods  of  overcoming  the  compli- 
cated conditions  of  inverted  combustion. 

Types.  Two  general  divisions  may  be  made  which  involve  dif- 
ferent applications  of  the  principles  of  combustion. 

The  first  is  based  on  a  burner  calculated  to  pre-heat  the  gas 
or  air,  or  the  mixture;  and  the  other  is  a  type  where  the  con- 
struction is  arranged  to  avoid,  as  far  as  possible,  increasing  the 
temperature  of  the  gases  before  they  reach  the  point  of  combustion. 

The  advantage  of  pre-hoating  the  gases  before  combustion  is 
questionable,  and  prominent  authorities  may  be  quoted  for  and 
against  the  increased  lighting  efficiency  to  be  obtained  by  this 
method.  It  might  be  inferred  that  the  incandescence  of  the  mantle 
would  be  increased  by  raising  the  initial  temperature  of  the  gases 
before  entering  the  combustion  chamber,  but  practical  results  show 
conclusively  that  the  abnormal  rarefaction  of  the  gases  due  to  the 
increased  temperature  of  the  mixture  tends  to  produce  the  oppo- 
site effect. 

On  the  other  hand,  artificially  cooling  the  gaseous  mixture  be- 
fore combustion  produces  a  decrease  in  the  efficiency.  Further- 
more, the  pre-heating  or  extreme  cooling  of  the  mixture  complicates 
the  burner  construction. 

Inverted-burner  designers  are  adopting  the  medium  system,  and 
are  constructing  a  burner  so  that  the  gaseous  mixture  will  main- 

ly GoOglc 


tain  a  temperature  ranging  between  the  extremely  hot  gases  pro- 
duced in  the  regenerative  type  and  the  cold  gases  produced  in  the 
cooling  type. 

In  modem  practice,  inrerted-bumer  construction  falls  under 
three  general  designs. 

First.  The  upright  Bnnsen,  with  the  tube  carrying  the  gas  and 
air  mixture  curved  through  half  the  arc  of  a  circle  (Fig.  6). 

Second.  The  horizontal  Batuen  tube,  with  the  mixing  chamber 
curved  through  one-quarter  of  an  arc  (Fig.  7). 

Fis.  6. — Upright  BuDsen  Inverted  Bamer. 

Thv-d.  The  vertical  Bunsen,  with  a  straight  tube,  for  the  de- 
livery of  the  mixture  to  the  point  of  combustion  (Fig.  8). 

Designers  recognize  as  an  essential  feature  of  design  and  function 
the  following  general  points: 

First.  The  production  of  a  proper  mixture  of  gas  and  air  tinder 
all  conditions  of  operation,  to  insure  perfect  combustion. 

Second.  Means  for  positively  preventing  flaeh-backe  tinder  all 
GonditionB  of  operation. 

Third.  Special  conBtruction  of  the  Bunsen  tube  designed  to 
project  the  gas  and  air  mixture  downward  to  the  point  of  com- 
bustion with  maximum  velocity,  in  order  to  overcome  the  ascend- 
ing tendency  of  the  mixture. 

>y  Google 

Incandescent  Q-as  Mantlbs  221 

Fourth.  The  elimination  of  obetructioiifi,  long  circuitous  pas- 
sageB  for  the  mixture,  or  any  other  features  offering  frictional 
resistance  to  the  projection  of  the  gases  toward  the  point  of  com- 

Fifth.  The  protection  of  the  fresh-air  supply  from  vitiation  by 
the  products  of  combustion. 

Sixth.  An  efBcient,  well-constructed  and  reliable  adjustable 
gas  check. 

Seventh.    Refractory  construction  at  the  burner  head. 

Eighth.    Good  mantles. 

Ninth.  Glassware  and  reflectors  specially  selected  and  adapted 
for  the  effective  and  economical  distribution  of  the  light. 

Fia.  7. — Horizontal  BuDBen  Inverted  Burner. 

Combustion,  as  applied  to  the  Bunsen  burner,  must  recognize 
three  basic  essentials:  (a)  the  combustible,  represented  by  the  gas; 
(b)  the  supporter  of  combustion,  represented  by  the  oxygen  of  the 
air,  and  (o)  the  kindling  temperature  necessary  to  start  the  com- 
bustion, applied  through  the  medium  of  a  lighted  match,  electric 
spark,  or  some  other  heating  means.  Eliminate  any  one  or  more 
of  these  three  essentials  and  combustion  ceases. 

When  a  certain  amount  of  gas  is  admitted  to  the  mixing  tube 
of  the  Buneen  burner,  a  definite  amount  of  oxygen  (air  containing 
oxygen)  must  be  entrained  and  mixed  with  it  in  order  to  completely 
consume  the  combustible  constituents  of  the  gas.  If  the  air  supply 
is  insufficient  to  meet  these  requirements,  unconsumed  or  par- 
tially consumed  constituents  of  the  gas  will  be  discharged  from  tiie 
burner,  either  in  the  form  of  solid  particles  of  carbon  or  as  noxious 

>y  Google 

222  Illuminating  ENOiNEBBiNa 

gases.  Od  the  other  hand,  an  excess  in  the  supply  of  air  results  in 
a  great  refluction  in  the  heating  power,  with  its  consequent  de- 
crease of  light,  or  produces  a  mixture  below  the  critical  point,  which 
mav  result  in  a  "  flash-back." 

FiQ.  8. — Vertical  BuDsen  Inverted  Burner. 

The  direct  cause  of  a  "  flash-back  "  is  an  e.\plosi\'c  action  which 
carries  the  flame  into  tiie  mixing  tube  and  sets  up  combustion  at 
the  gas  orifice.  The  usual  method  used  for  overcoming  the  ten- 
dency to  "flash-back"  is  by  placiug  a  gauze  in  the  burner  tube 
at  or  near  the  point  of  combustion.  This  gauze  serves  to  maintain 
the  mixed  gases  in  the  burner  tube  at  a  temperature  somewhat 
below  the  kindling  temperature,  and  thereby  prevents  combustion 

>y  Google 

Incandescent  Gas  Mantles  233 

withiQ  the  zone  it  protects.  This  will  be  recognized  as  the  prin- 
ciple involved  in  the  safety  lamp  invented  by  Sir  Humphry  Davy, 

The  nee  of  the  gauze  for  preventing  "  flash-backs  "  is  sometimes 
objected  to  on  the  ground  that  it  obstinicts  and  materially  retards 
the  downward  projectioQ  of  the  mixture  in  the  burner  tube,  and 
that  it  becomes  clogged  with  dust  and  materially  cuts  off  the 

The  thermostat  {Fig.  9)  is  a  device  placed  in  the  lower  Bunsen 
tube  of  one  type  of  the  inverted  burner,  and  performs  the  function 
of  a  gauze  without  introducing  its  objectionable  features.     When 

Closed— Cold.  Open— Hot. 

Flo.  9.— Thermostat. 

the  lamp  is  cold  the  fingers  of  the  thermostat  are  closed,  forming 
a  slitted  cone  which  prevents  a  "  flash-back"  on  lighting.  As  the 
lamp  becomes  heated  the  thermostat  opens,  leaving  the  Bunsen 
tube  clear  for  the  unobstructed  flow  of  the  gas  and  air  mixture.  It 
is  so  placed  in  the  tube  that  it  is  not  corroded  by  the  action  of  the 
flame,  and  its  automatic  movements  prevent  it  from  collecting  dust. 
This  thermostat  is  made  from  a  double  metal  in  which  each  side 
possesses  a  difi'ereut  coefficient  of  expansion;  for  example,  brass  and 
iron.  The  brass  is  placed  inside,  and  due  to  its  own  rapid  expan- 
sion when  heated  causes  the  curved  fingers  of  the  thermostat  to 
straighten  out  and  lie  against  the  wails  of  the  Bunsen  tube.  On 
cooling,  the  brass  contracts  more  than  the  iron  and  the  fingers 
resume  their  original  curved  position,  forming  the  slitted  cone 
(Fig-  9). 


324  Illuminatimg  Ehgineerino 

The  Buuseu  tube,  even  in  its  highly  developed  form,  now  used  in 
upright  bumers  fails  in  Bome  essential  features  when  applied  to  the 
inverted  burner.  In  considering  this  problem,  it  should  be  noted 
that  the  ordinary  illuminating  gases  are  lighter  than  the  air  and 
possess  a  marked  ascending  tendency  even  at  the  normal  tempera- 
ture.    When  considered  in  connection  with  the  heated  condition 

Ou  idjuitinent 

Crown  lor  hoId< 

Fio.  10. — Inverted  Burner  Cut  to  Sbow  Interior  Conatrnctlon. 

of  the  inverted  mixing  tube  it  is  seen  that  this  ascending  tendency 
is  thereby  greatly  increased. 

The  method  used  for  overcoming  the  ascending  tendency  of  tbe 
mixture  is  to  project  it  downward  with  sufficient  velocity  to  cany 
it  to  the  point  of  combustion  without  regard  to  the  specific  gravity. 

The  only  force  available  for  projecting  the  mixture  downward 
is  that  obtained  from  the  velocity  of  the  gas  at  the  check  orifice. 
When  it  Je  considered  that  in  many  cases  the  initial  gas  preaniie 

>y  Google 

Incandescent  Gas  Mantlbs  225 

ie  very  low,  thereby  greatly  reduciDg  the  available  force,  and  also 
that  a  certain  portion  of  this  force  mnst  be  given  to  entraining  the 
air  for  the  mixture,  it  is  obvious  that  great  importance  attaches 
to  this  function  of  the  inverted  burner. 

To  meet  the  conditions  of  varying  composition  and  pressure,  or 

Pro.  11. — Qw  Arc  Lamp.    Upright  for  Inside  Llgbtlng. 

uniformly  lov  pressure  in  the  gas  supply,  a  construction  is  re- 
quired embodying  all  the  features  of  a  highly  efficient  projector 
for  gases.  This  requires  an  adjustable  check  which  will  give  the 
greatest  jet  velocity  to  the  gae  as  it  is  admitted  to  the  Bunsen; 
air  porte  properly  placed  to  give  the  most  efficient  entraining 
capacity ;  a  "  raceway "  of  correct  diameter  and  length  to  give 
the  mixed  gases  the  velocity  necessary  to  carry  them  through  the 
mixing  chamber  and  to  the  point  of  combustion. 

>y  Google 

SS6  Illuminating  Engineebino 

An  analysis  of  the  large  variety  of  inverted  burners  on  the 
market,  in  the  light  of  these  facts  and  principles,  will  show  a 
number  which  do  not  conform  to  any  specifications  except  cheapness. 

Bapid  progress  is  being  made,  however,  and  standardization  will 
ultimately  be  reached  on  a  combination  basis  of  efficiency,  relia- 
bility, convenience,  durability,  pleasing  appearance — all  with  a  fair 
and  reasonable  cost. 

Fio.  12. — Gaa  Arc  Lamp.    Upright  for  Instde  Lighting. 

Gas  Arc  Lamps  and  Clusters 

Following  the  introduction  of  the  upright  burner,  high  candle- 
power  unit  requirements  were  met  by  forming  a  cluster  of  indi- 
ridual  burners,  with  separate  gas  cocks  and  chimneys,  gathered 
under  a  common  reflector.  These  groups  of  burners  were  next 
simplified  by  the  introduction  of  a  cluster  of  burners  controlled 
hy  a  single  gas  cock  and  surrounded  by  a  single  globe  to  replace 
the  individual  chimneys.  This  design  of  lamp  was  called  a  gas 
arc  lamp,  and  it  met  with  success  on  account  of  its  simplicity  of 
construction  and  easy  maintenance. 

The  principal  aims  in  the  development  of  the  gas  arc  lamp  have 
been  to  produce  a  unit  (Figs.  11  and  12)  : 

>y  Google 

Incandescent  Gas  Mantles  327 

First.    With  a  concentrated  eoarce  of  light. 

Second.    With  high  efficiency. 

Third.    Simplicity  of  operation. 

Fourth.    MinimiiiD  cost  of  maintenance. 

Fifth.    Individual  gas  adgiistraent  for  each  burner. 

Xo  principles  differing  from  those  encountered  in  the  individual 

Fig.  13.— Gas  Arc  for  Outside  Liehtlng. 

burner  were  involved  in  the  development  of  this  upright  arc  lamp, 
although  some  perplexing  conditions  were  met  with. 

It  was  found  that  in  order  to  approximate  the  efficiency  of 
the  individual  burner,  the  arc  would  have  to  be  conetnicted  with  a 
"stack"  to  induce  more  active  combustion  at  the  burner  heads. 
Theae  stacks  are  made  from  fused  enamel  on  steel,  or  from  brass 
in  various  finishes.  Mechanical  devices  have  been  evolved  for  con- 
venient methods  for  renewing  and  replacing  mantles,  removing  and 
cleaning  glassware,  and  innumerable  other  methods  of  simplifying 
and  economizing  maintenance  and  up-keep. 

Upright  arcs  have  been  succ-essfuliy  developed  for  use  outside  in 
places  exposed  to  the  action  of  wind  and  rain  ^Fig.  13). 

>y  Google 

2S8  Illuhinatimo  Enginebbimo 

Inverted  Gas  Arcs 

Arcs  of  the  inverted  type  for  both  inBide  (Figs.  14  and  15)  and 

outside  (FigB.  16  and  17)  lighting  are  juBt  coming  into  nse,  and 

are  being  rapidly  improved  and  developed  with  every  prospect  of 

great  success . 

Fio.  14. — Inverted  Oas  Arc  for  Inelde  Ughtlng. 

Two  different  methods  of  construction  are  utilized  in  the  most 
prominent  types  of  inverted  arc  lamps.  One  in  which  an  indi- 
vidual Bunsen  is  provided  for  each  mantle  (Fig.  14),  and  the  other 
where  a  single  common  Bunsen  leads  into  a  manifold  head  from 
which  outlets  are  provided  for  each  mantle  (Fig,  15).  Both  of 
these  types  are  now  appearing  in  various  sections,  and  eiperienee 
alone  will  demonstrate  the  wisdom  of  the  design. 

Incandescent  Mantles 
The  incandescent  gas  mantle  was  invented  by  Dr.  Carl  Aner  von 
Welsbach  in  1885  and  1886. 

>y  Google 

Inoandescent  Qa8  Mantles  229 

The  basis  of  Dr.  Auer's  invention  is  the  refractory  hood  or 
mantle  made  from  an  attenuated  mixture  of  the  oxide  of  thorium 
with  a  small  percentage  of  cerium  oxide.  The  cerium,  which  is 
present  in  quantities  varying  from  '^  to  2  per  cent,  is  not  an  acci- 
dental impurity  as  has  been  inferred,  but  is  an  essential  constituent 
exerting,  by  very  small  variations  in  amount,  a  marked  effect  upon 
the  candle-power  and  quality  <^f  the  light.  The  candle-power-cerium 

Fia.  16. — Inverted  Gas  Arc  for  Inelde  Lighting. 

relation  is  best  illustrated  by  the  curve  shown  in  Fig.  18,  in  which 
the  candle-power  is  plotted  vertically  and  the  per  cent  of  cerium 
horizontally.  It  will  be  noted  that  the  maximum  candle-power  ia 
obtained  with  1  per  cent  of  cerium,  and  that  a  small  amount  of 
cerium  more  or  less  than  1  per  cent  causes  the  candle-power  to  fall 
off  very  rapidly. 

This  peculiar  result  may  be  attributed  to  the  existence  at  the 
1-per-cent  point  of  a  solid  solution  or  a  definite  componnd  which 
possesses  a  higher  emissivity  than  either  the  thorium  alone  or  the 
cerium  alone. 

>y  Google 

230  Illdminatinq  Enoineerino 

The  manufacture  of  incaDtlescent  gaa  mantles  is  a  most  inter- 
esting and  complicated  chemical  process,  and  by  a  peculiar  coioci- 
dence  resembles  in  the  delicacy  of  the  hand  work  invoWed  the 
close  attention  to  details  and  the  technical  supervision  required  in 
the  manufacture  of  the  incandescent  electric  lamps. 

A  brief  outline  of  the  materials  and  processes  involved  in  the 
mantle  manufacture  may  be  of  interest. 

The  first  step  consists  in  knitting  a  tubular  fabric  of  open  mesh 
from  threads  of  some  combustible  organic  substance  which,  after 
being  properly  saturated  with  the  thorium  solution,  may  be  cob- 
veniently  burned  out,  leaving  the  ash  in  a  more  or  less  adherent 
mass  reproducing  the  physical  form  of  the  original  fiber.  The 
selection  and  preparation  of  the  original  fiber  is  therefore  a  matter 
of  vital  importance.  Imperfect  fibers  or  threads,  mineral  impuri- 
ties, irregular  knitting,  etc.,  all  directly  affect  the  quality  of  the 

The  present  practice  is  to  use  threads  made  from  natural  cotton 
fiber,  natural  ramie  fiber  or  artificial  silk  fiber. 

>y  Google 

Incandescent  Gas  Mantles 

Pio.  17. — Inverted  Arc  Lamp  tor  Inside  Lighting. 

>y  Google 

939  Illuuinatino  Enoinsbsino 

The  cotton  thread  muet  be  made  of  a  high  grade,  long  staple, 
Sea  Island  fiber,  unifonn  in  size  and  free  from  knota  or  flaivB.  The 
tensile  strength  of  the  resulting  mantle  fiber  depends  largely  opon 
the  length  and  physical  characteristics  of  the  basic  fiber.  Farther- 
more,  if  any  knots,  flaws,  thin  places,  etc.,  exist  in  the  threads  they 
are  reproduced  in  the  mantle. 

Bamie  is  a  natural  v^etable  fiber  made  from  a  sabstance  known 
as  "  China  grass."  The  commercial  supply  of  ramie  is  obtained 
almost  entirely  from  China,  India  and  Italy.  In  its  crude  form 
the  ramie  fiber  contains  large  amounts  of  reeins  and  mineral  matter, 
and  its  purification  is  a  very  difficult  and  complicated  chemical 















'*-            1 


rio.  18. 

Bamie  fibers  are  long  compared  with  cotton  and  possess  greater 
tensile  strength  and  wonld  naturally  be  expected  to  make  a  stronger 
mantle.  While  mantles  made  from  a  ramie  base  do  not  shrink  as 
badly  as  mantles  made  from  cotton  their  tensile  strength  is  some- 
what disappointing,  especially  after  being  used  for  a  time. 

Artificial  silk,  as  the  name  implies,  is  an  artificial  fiber.  It  is 
made  by  dissolving  cellulose  in  some  suitable  solvent  to  form  a 
thick  viscous  solution,  squirting  this  syrup  through  very  fine  dies, 
by  great  pressure,  into  some  fixing  bath.  The  resultant  continuous 
filaments  are  then  twisted  into  a  thread.  This  thread  may  be 
knitted  into  mantle  fabric  and  subjected  to  a  special  process  of 
treatment  for  the  production  of  a  remarkably  improved  product. 
Mantles  made  from  artificial  fibers  show  improved  physical 
strength,  no  tendency  to  shrink,  no  change  in  quality  of  light. 

>y  Google 

Inoandescknt  Gas  Mantles  333 

and  a  remarkably  small   candle-power  depreciatioii,   even   after 
several  thOTiBand  hours  of  continuous  burning. 

Saturating  is  a  comparativelj  simple  process,  where  the  thor- 
oughly dried  fabric  is  placed  in  a  suitable  vessel  and  covered  with 
the  lighting  fluid.  As  soon  as  it  is  thoroughly  saturated,  the  ei- 
cess  of  fluid  is  drawn  off  and  the  fabric  ia  put  through  an  equal- 
izing machine  piece  by  piece,  in  order  to  bring  each  mantle  to  a 
Dniiorm  degree  of  saturation. 

In  the  highest  grades  of  mantles  the  amount  of  lighting  fluid 
used  is  based  upon  a  earful  consideration  of  the  amount  of  oxide 
required  to  produce  a  mantle  of  the  highest  physical  and  candle- 
power  life. 

The  lighting  fluid  is  composed  of  a  solution  of  approximately 
99  per  cent  nitrate  of  thorium  and  1  per  cent  nitrate  of  cerium  in 
distilled  water.  This  solution  is  usually  of  about  3  parts  by  weight 
of  water  to  1  part  by  weight  of  mixed  nitrates.  While  the  formula 
varies  somewhat  with  different  manufacturers,  the  limits  are  not 

The  commercial  source  of  the  nitrates  of  thorium  and  cerium 
is  from  a  mineral  known  as  monazite.  This  mineral  occurs  in 
commercial  quantities  only  in  North  and  Soutti  Carolina  and  in 
Brazil.  The  Carolinas'  monazit«  is  found  as  a  sand  in  the  stream 
beds  of  the  old  mountainoug  districts,  white  in  Brazil  it  occurs 
as  a  beach  sand. 

Monazite  sand  is  mined  on  the  principle  involved  in  placer 
mining  for  gold.  The  gravel  and  associated  minerals  are  shoveled 
onto  screens  and  worked  through  into  sluice  boxes,  where  the  min- 
erals of  lower  specific  gravity  are  carried  away  by  the  water  cur- 
rente,  while  the  heavy  monazite  remains  behind.  This  crude  con- 
centrate, carrying  from  20  to  40  per  cent  monazit«,  is  shipped  to 
central  plants,  where  it  is  further  concentrated  by  the  use  of 
Wilfley  tables  and  magnetic  concentrators.  The  final  product,  as 
it  is  delivered  to  the  manufacturer  of  lighting  fluid,  contains  about 
90  per  cent  of  monazite  of  the  following  average  compoeition : 

FhOBphortc  anhydride    28;^ 

Cerium  oxide  30 

Lanthannm  oxide   14 

NeodTmlum  and  praseodrmlum  16 

Thorium  oxide   6 

Tttrlum  oxide 2 

Iron  oxide,  calcium  oxide,  etc S 

Total 100% 


Digitized  oyGOOglC 


The  manufacture  of  nitrate  of  thorium  from  monazite  sand  is 
B  very  difficult  and  complicated  chemical  process.  It  requires 
from  4  to  6  months  to  recover  the  small  percentage  of  thorium  and 
render  it  sufficiently  pure  to  be  used  in  the  manufacture  of  lighting 
fluid.  The  by-producta  from  this  proeees  have  great  scientific  and 
chemical  interest  but  no  commercial  value,  and  the  thorium  must 
stand  the  entire  expense.  The  refined  thorium  nitrate  must 
be  chemically  pure — free  from  all  traces  of  mineral  impurity  and 
the  other  constituents  of  the  monazif*  sand. 

The  saturated  fabric  is  now  fixed  for  suspension  by  using  as- 
bestos thread  to  form  a  loop,  then  shaped  up  preparatory  to  burn- 
ing out  the  organic  material  and  conTeri;ing  the  nitrates  into  oxides. 

The  burning-out  process  is  accomplished  by  igniting  the  fabric 
with  a  torch  and  waiting  until  the  organic  matter  slowly  oxidizes. 

After  the  fabric  is  completely  consumed  the  ash  of  thorium  and 
cerium  oxides  hangs  in  a  soft,  shapeless,  flabby  condition,  and  pr&- 
sents  a  very  remote  resemblance  to  a  mantle. 

When  Dr.  Auer  first  explained  his  idea  for  making  a  mantle 
to  Professor  Bunsen  that  famous  teacher  replied :  "  It  is  extremely 
doubtful  if  the  ash  can  be  made  to  hold  together."  This  opinion 
was  based  upon  Professor  Bunsen'a  knowledge  of  the  general  char- 
acteristics of  metallic  oxides,  but  the  oxides  with  which  Dr.  Auer 
was  working  were  notable  exceptions.  The  incandescent  gas  light- 
ing industry  depends  upon  this  remarkable  exception. 

After  the  organic  matter  is  completely  burned  out  in  the  process 
just  described,  the  soft,  flabby  ash  is  carefully  adjusted  over  a 
blow-pipe.  The  operator  of  this  device  controls  levers  which  raise 
and  lower  the  mantle,  and  which  adjust  the  gas  and  the  air  supply 
to  the  blow-pipes.  In  some  cases  the  gas  is  used  under  a  pressure 
of  several  pounds  to  produce  the  intense  flame  required,  but  in 
either  event  the  adjustment  of  the  flame  and  the  control  of  the 
position  of  the  mantle  is  entirely  in  the  hands  of  the  operator. 

Under  the  influence  of  this  intense  blast  flame  the  flabby  ash, 
left  when  the  organic  fabric  was  burned  out,  is  blown  (by  the 
proper  control  of  the  flame)  into  the  required  shape,  and  is  changed 
from  its  soft,  pliable  state  into  a  hard,  resilient  form.  This  opera- 
tion requires  greater  skill  and  experience  than  any  other  work  con- 
nected with  mantle  manufacture. 

Coating.  The  object  of  this  process  is  to  form  a  protecting 
elastic  skin  over  the  ash  to  carry  it  while  the  mantle  is  going 

>y  Google 

Incandescent  Gas  Mantles  235 

through  the  inspecting,  trimming,  packing,  transportation  and  in- 
stallation stages. 

This  coating,  or  collodion,  as  it  is  usually  called,  is  made  from 
soluble  cotton.  Soluble  cotton  is  made  by  the  so-called  nitrating 
process  in  'which  the  loose  cellular  fiber  is  treated  with  a  mixture 
of  sulphuric  and  nitric  acids,  and  a  product  is  formed  closely 
allied  to  gun-cotton. 

This  nitrated  cotton,  after  being  thoroughly  washed  and  dried, 
is  dissolved  in  some  of  the  numerous  solvents  such  as  alcohol- 
ether,  acetone,  etc.,  and  a  thick,  viscous  liquid  is  produced. 

The  collodion  is  placed  in  suitable  vessels,  over  vhich  the  mantles 
are  suspended  and  into  which  they  are  dipped,  then  transferred 
to  hoods  to  dry.  The  mantles  are  then  inspected  and  packed  to 
meet  the  great  variety  of  needs  of  the  established  markets. 

It  is  estimated  that  the  American  market  oonsumee  60,000,000 
mantles  per  year,  most  of  which  are  standard-sized  upright  and 
inverted  mantles.  Large  quantities  of  mantles  are  also  produced 
for  railroad-coach  lighting  with  Pintsch  gas,  kerosene  lamps,  gaso- 
line systems  and  high-pressure  oil  lamps. 

In  the  limited  allotment  of  time  for  this  subject,  this  review 
must  necessarily  be  brief  and  superficial,  but  I  have  attempted  to 
make  it  clear  to  you  that  the  development,  growth  and  future  of 
the  incandescent  gas-lighting  industry  is  a  matter  of  immense 
scientific  and  economic  interest. 

>y  Google 





By  John  B.  Whitbhead 



1.  Interl<Hr  Illumination. 

a.  S^Btema  of  power  anpply:   generating  planta;  conatant  potential; 

direct  current,  2-  and  3-wlre;  alternating  current,  2-  and  8-vire; 
alternating  current,  falgh  voltage  alngle  and  polyphon;  trans- 
formers;  Isolated  power  plants. 

b.  S7>tema  of  distribution:     2-,  ^  and  6-wlre  parallel  dlstrlbstlon. 

tor  Incandescent  glower,  vapor  and  arc  lighting;  series  pftraltel 
distribution;  low  voltage  Incandtjscent  lamps  on  direct  and 
alternating  current  circuits. 

c.  Design  of  electrical  syston:   Choice  of  sTStem;  regulation  of  sup- 

ply srstem;  voltage  drop  In  direct  ssd  alternating  circuits; 
permlsBtble  voltage  variation;  size  of  feeders;  diversity  factor; 
number  and  sizes  of  broncbes. 

2.  Exterior  and  street  illumination. 

a.  Systems  of  power  supply:    Constant  potential  and  constant  cur- 

rent, high  and  low  voltage,  direct  and  alternating;  constant- 
current  generator;  constant-current  regulators  and  rectifiers. 

b.  Systems  of  distribution:     Parallel  and  series  parallel  constant 

potential,  for  arcs  and  In  can  descents.  Constant-current  series 
systems  for  arcs  and  incandescents.  Alternating  current  to 
direct  current  systems. 

c.  Design  of  electrical  system:    Choice  of  system.    Constant  voltage 

and  constant  current  regulation.  Size  of  feeders.  Power  loss 
In  aeries  circuits;  underground  and  overhead  systems. 

3.  Meters. 

a.  Types  of  meter. 

b.  Accuracy,  calibration  and  Inspection  of  meters. 

Tkc  iKBTAix&Tion  or  BtMeraic  Lishtino  Ststeics 
1.  Interior  tllnmlnatlon. 

a.  Type  of  Installation.     Two-  and  three-wire.     Exposed  and  con- 

cealed wiring.    Conduit  systems  and  outlet  hoxes. 

b.  Control.     Service  connections.     Distributing  centers.     Switches. 

Protective  devices.    Subdivision  of  total  copper. 
&  Relative  costs. 

>y  Google 


d.  Fire  and  Insuraiice  control.    Groiind  connections. 

e.  Speclflcatlone,  drawings  and  contracts  for  work  of  Installation, 

Including  materials. 
f.  Teats. 
2.  Exterior  Jllumlnatlon. 

a.  Type  of  installation,  arc  or  incandesceat,  parallel  or  series.  Over- 
head or  underground  Bysteme.    Insulation, 
b-  Control.       Automatic     cut-outs.      Protective     devlcea,     ligbtJng 

c  Municipal  restrictions.    Underground  construction  and  cables. 
2.  Coet  of  operation, 

a.  Cost  of  electric  power.. 

b.  Systems  of  rates  of  aale  of  power;  flat  rates;  maximum  demand; 

two-rate  a;stema. 
c  Contracts  for  purchase  of  power. 

Principle*  and  Design 

Electricitj  for  lighting  may  be  taken  from  any  type  of  gen- 
erator. The  earliest  types  of  generator  were  developed  to  meet 
the  requirements  of  lighting  apparatus.  With  the  introduction 
of  other  applications  of  electricity  generators  have  been  designed 
with  characteristics  to  meet  particular  purposes,  but  it  is  probable 
that  every  operating  generator  furnishes  more  or  less  current  for 
lighting.  In  modem  installations,  in  which  a  large  portion  of  the 
total  capacity  is  consumed  in  lighting,  the  generators  are  designed 
with  special  reference  to  the  regulation  required  by  lighting  ser- 
vice. Such  generators  are  of  various  types,  the  extremes  being 
the  smallest  continuous-current  dynamo  of  the  isolated  plant  for 
a  single  building,  and  the  modern  high-power  (20,000  kw.)  alter- 
nator of  the  city  central  station. 

The  proper  circuit  conditions  for  electric  lighting  are  either 
constant  potential  or  constant  current.  The  general  problem  of 
central-station  design  to  meet  these  conditions  involves  a  knowl- 
edge of  the  various  sources  of  energy,  types  of  prime  movers,  gen- 
erators, control  and  regulating  apparatus,  and  is  distinctly  within 
the  province  of  the  present-day  electrical  engineer.  The  electrical 
phase  of  the  problem  of  the  illuminating  engineer  will  only  in 
extremely  rare  instances  contain  the  questions  of  prime  movers, 
generators  and  station  design.  In  general  his  concern,  certainly 
in  interior  illumination,  need  go  no  further  back  than  the  avail- 
able service  mains.  At  this  point  he  need  only  recognize  the  type 
of  service,  know  what  regulation  he  may  demand,  and  be  able  to 

>y  Google 

Geseeation  and  Dibteibdtion  of  Electsicitt        239 

draft  a  service  contract  for  Ms  client.  From  this  point  inward  he 
must  be  able  to  deeign  the  diBtribntiDg  system  electrically  and 
mechanically,  vith  due  regard  to  fire  hazard  and  conformity  to 
local  regulations.  He  must  be  able  to  draft  a  specification  and 
prepare  drawings  for  an  installation  which  shall  amply  secure 
for  his  client  a  completed  and  tested  lighting  system  for  a  definite 
price.  The  eiterior  problem  requires  a  somewhat  wider  knowl- 
edge of  the  principles  of  distribution,  but  will  rarely,  if  eyer,  ap- 
proach the  station  nearer  than,  say,  a  constant-current  regulator. 
In  brief,  the  illuminating  engineer  can  generally  assume  that  the 
electrical  engineer  will  furnish  him  with  constant  pressure  or  con- 
stant current.  The  electrical  problem  of  the  former  is  to  know 
the  limits  of  this  constancy,  and  to  be  able  to  design,  install  and 
test  the  proper  distributing  system.  Should  the  illuminating  engi- 
neer eTer  desire  to  extend  his  knowledge  to  the  engineering  of 
■  generating  equipment,  many  excellent  treatises  on  the  subject  are 
readily  available,  and  it  does  not  appear  desirable,  in  the  short 
space  allotted  here  to  the  electrical  problem  of  the  illuminating 
engineer,  to  devote  more  than  occasional  mention  to  a  kindred 
topic  of  wide  extent  and  well  treated  in  the  literature  of  the  subject. 

1.  Interior  Illamination 
(a)  Systems  of  Power  Supply.  The  commonest  class  of  public 
power  supply  for  interior  lighting  is  at  constant  potential.  In 
the  hearts  of  cities  it  is  usually  in  the  form  of  continuous  current 
supplied  by  an  underground  three-wire  interconnected  network 
of  mains.  This  network  is  fed,  over  underground  feeders  con- 
nected to  the  mains  at  various  points,  from  rotary  converters  or 
motor  generators  in  one  or  more  substations.  The  general  plan 
of  such  a  network  is  indicated  in  Fig.  1.  These  machines  are 
operated  by  alternating  current  which  is  generated  at  voltages  up 
to  15,000,  or  even  higher,  in  central  stations  at  some  distance 
from  the  substation  centers  of  distribution.  The  voltage  of  these 
networks  is  230  to  240  between  two  so-called  "  outside  "  wires,  and 
110  to  130  volts  between  either  outside  wire  and  a  third  or 
"  neutral "  wire  which  is  usually  kept  at  the  potential  of  the 
eartli,  or  "  grounded  "  by  connecting  to  an  underground  system  of 
water  pipes,  or  by  other  methods.  Most  interior  lighting  devices 
are  designed  for  voltages  in  the  neighborhood  of  110,  and  the  aggre- 
gate load  is  divided  as  uniformly  as  possible  between  the  two  sides 

>y  Google 


Illuminating  Engineering 

of  the  three-wire  network.  In  thia  way  the  two  halves  of  the  load 
are  connected  in  eerieg,  and  the  distribntion  for  llO-volt  service 
is  accomplished  at  330  volt«,  with  great  saving  in  the  amount  of 
copper,  since,  at  a  given  loss  and  distance,  the  amount  of  copper 
necessary  varies  inversely  as  the  square  of  the  voltage.  The  use 
of  the  neutral  conductor,  however,  reduces  the  amount  of  this 
theoretical  saving.  The  neutral  conductor  is  made  necessary  by 
the  facte  that  the  component  parts  of  the  load  on  the  two  sides  of 
the  system  are  often  separated  by  gome  distance,  and  especially 
that  the  two  aides  of  the  system  are  never  exactly  equally  loaded. 








'"         -J 



Pio.  1. — Direct-Current  Underground  Network, 

Fio.    2. — Outlying   Alternating-Current   DUtrlbntlon. 

The  escess  current  of  the  more  heavily  loaded  side  flows  back 
to  the  substation  over  the  neutral  conductors  of  the  mains  and 
feeders.  This  conductor  therefore  only  carries  the  difference  in 
the  current  of  the  two  sides  of  the  circuit,  and  in  a  large  system 
with  average  balance  of  load  between  the  two  sides  of  from  3  to 
5  per  cent,  its  cross-section  may  be  considerably  less,  say  one-half 
that  of  the  outer  wires.  This  system  therefore  requires  a  genera- 
tor connection  at  a  point  midway  between  the  potentials  of  the 
outer  terminals.  This  may  be  accomplished  by  operating  two 
generators  in  series  and  connecting  the  neutral  to  their  junction. 
By  the  use  of  various  auxiliary  devices  single  machines  may  be 
constructed  for   supplying   three-wire   service. 

>y  Google 

GbNEEATION  and  DiSXfilBCTION  OP  Elrcteicitt 


The  croes-BectioD  of  the  main  conductors  of  such  a  network  may 
aggr^;ate  several  million  circular  mils,  divided  into  lead-covered 
cables  of  1,000,000  or  2,000,000  c.  m.  each.  The  feeder  cables  are 
usnall;  somewhat  smaller,  with  neutral  one-half  the  section  of 
the  outside  conductors.  These  feeders  are  provided  with  g  small 
insulated  strand  leading  back  to  the '  station,  which  serrea  to 
indicate  in  the  station  the  potential  at  the  network.  The  voltage 
drop  in  the  feeders  varies  from  time  to  time  and  may  be  as  great 
as  10  per  cent. 

In  locations  at  some  distance  from  the  central  station  or  sab- 

ijlf-- -=" 

Phi.  3. — AlternBtlng-Current  Secoodary  Network. 
Fig.  4. — Two-Phase  Three-  and  Four- Wire  SrstemB. 

station  power  is  transmitted  as  high-voltage  alternating  current, 
and  the  voltage  lowered  by  transformers  which  feed  into  the  con- 
sumers' circaits.  For  the  extreme  outlying  districtfi  with  widely 
scattered  consumers,  each  is  often  fed  from  a  single  small  trans- 
former located  at  the  property  line,  and  supplying  power  over 
two  wires  only-  In  intermediate  regions  where  the  consumption 
is  fairly  dense  several  consumers  may  be  fed  from  the  same  trans- 
former, as  indicated  in  Fig.  2.  For  still  denser  regions,  beyond  the 
reach  of  the  continuous-current  network,  a  secondary  alternating- 
current  network  fed  by  several  transformers  at  different  points  is 
sometimes  formed  (Fig.  3).  In  each  of  these  cases  the  three- wire 
system  is  commonly  used  with  320  to  240  volts  on  the  outer  wires, 
and  the  neutral  connected  to  the  middle  point  of  the  transformer 

>y  Google 



secondsries.  Both  the  primary  and  secondary  circuits  in  this 
class  of  supply  are  usually  carried  overhead,  though  not  invariably. 

The  modem  large  central  station  generates  at  25  cycles,  three- 
phase.  This  frequency  is  the  best  for  transmiEsion  and  for  trans- 
formatlcn  to  mechanical  power.  It  is  not,  however,  well  adapted 
to  either  arc  or  incandescent  lighting,  although  there  are  many 
instances  in  which  it  is  need  for  the  latter.  Economy  of  trans- 
mission  copper  and  the  euperiority  of  the  polyphase  motor  for 
power  service  result  in  the  general  use  of  three-phase  instead  of 
single-phase  primary  circuits. 

For  lighting  from  such  systems  motor  generators  are  often  used 
for  changing  from  35  to  60  cycles,  the  latter  being  the  standard 
frequency  for  alternating-current  lighting. 

Fia.  e. — Three-Phaae  Tbree-Wfre  SyBtem. 
Fio.  6. — ^Three-Phase  Four-Wire  STBtem. 

The  60-cycle  generator  for  city  lighting  operates  usually  at 
some  voltage  between  2200  and  2600.  Since  there  is  always  a 
market  for  power  also,  it  is  commonly  of  two-  or  three-phase  type. 
Secondary  lighting  circuits  at  220  to  240  volts  are  obtained  from 
the  polyphase  primaries  by  transformers  connected  in  various  ar- 
rangements. If  the  service  is  for  lighting  only,  single-phase  sec- 
ondaries only  are  needed,  and  single  transformers  for  the  separate 
loads  are  connected  to  the  various  phases,  and  in  such  manner 
that  the  aggregate  load  is  as  nearly  as  poedble  divided  evenly 
among  the  several  phases.  In  some  jnetances,  however,  there  is 
a  power  load  requiring  small  motors  which  cannot  be  operated  at 
the  high  primary  voltage.  These  moderate-size  motors  are  also 
moet  satisfactory  in  the  polyphase  type.  The  secondary  distrib- 
uting system  must  therefore  be  polyphsse,  and  this  is  accomplished 

>y  Google 

Generation  and  Distbibution  of  Eleotsioity        343 

by  varioufi  transformer  combinations,  resulting  in  tbree-,  foor- 
and  even  five-wire  secondary  systems.  Figs.  4,  5,  6  and  7  show 
several  ot  these  combinations.  Lighting  circuits  may  be  taken 
from  any  one  of  the  branches  of  snch  a  symmetrical  system.  This 
results,  however,  in  placing  both  lamps  and  motors  on  the  same 
transformer.  Since  the  altemating-cunent  motor  often  takes  a 
starting  cnrrent  several  times  as  great  as  its  normal  rmming  cur- 
rent, the  starting  of  the  motors  frequently  resnlts  in  a  momentary 
finctnation  of  voltage  which  is  noticeable  at  the  lamp.  For  the 
most  satisfactory  results  the  lighting  and  power  loads  should  be 
on  separate  transformers,  as  the  greater  part  of  the  voltage  dis- 
turbance occurs  in  the  secondary  distributing  circuit  and  in  th« 
transformer  itself. 



-t i 

An  important  type  of  supply  system  is  the  so-called  isolated 
plant  of  a  single  large  building  or  factory.  These  plants  are  either 
steam-,  gas-  or  water-driven,  and  generate  current  of  the  class  and 
voltage  required  at  the  lamp.  Thus  many  of  the  earliest  plants  are 
equipped  with  110-volt,  two-vrire  continuous-current  compound 
generators.  Those  of  more  modem  design,  however,  have  S20-volt 
three-wire  generators,  as  the  extent  of  the  distributing  system  is 
usually  sufficient  to  demand  the  resulting  economy  in  copper.  This 
type  of  plant  for  lighting  alone  constitutes  as  reliable  a  source  of 
supply  as  can  be  obtained.  Properly  chosen,  the  generating  plant 
will  give  as  constant  voltage  regulation  as  may  be  desired,  and 
satisfactory  performance  of  the  lamp  vrill  then  depend  only  on 
the  design  of  the  distributing  system.     Usually,  however,  these 

>y  Google 

244  iLLDUINATCtO  £NaiNKraiNO 

plants  muat  also  fanuah  power.  Little  trouble  Ib  caused  if  the 
motors  are  of  moderate  size  and  the  power  load  fairly  constant. 
The  elevator  or  other  similar  motor,  however,  with  its  large  starting 
current  will  asnally  cause  serious  voltage  fluctuations  unless  special 
provision  is  made  for  its  fluctuating  load.  The  most  approved 
equipment  for  this  purpose  is  the  storage  battery  with  "  booster  " 
geoeraton;  the  battery  automatically  charges  at  values  of  load 
under  the  average  and  discharges  when  the  elevator  motors  re- 
quire their  greatest  currents.  This  equipment  also  serves  to  equal- 
ize the  demand  on  the  generators  between  the  light  and  heavy 
portions  of  the  daily  load  curve,  the  battery  charging  during  the 
morning  hours  and  discharging  as  the  lighting  peak  comes  on. 


-* zzo\ 

'jL L 


Pia.  9.— Three- Wire  Distribatlon. 

(b)  Syitenu  of  Conneotion.  Except  in  rare  instances  all  in- 
terior electric  lighting  is  obtained  from  lamps  designed  for  constant 
voltage.  The  range  of  this  voltage  between  110  and  130  volts 
is  that  within  which  the  incandescent  lamp  can  be  moat  satis- 
factorily constructed.  This  has  probably  been  the  most  important 
influence  in  fixing  this  practically  standard  value  for  the  voltage 
of  low-tension  distributing  circuits.  The  various  other  types  of 
lamp  for  interior  service  have,  therefore,  naturally  developed  with 
conformity  to  this  voltage,  or  to  double  its  value,  as  obtained  from 
the  outer  wires  of  the  three-wire  distributing  system. 

Incandescent  lamps,  therefore,  for  interior  illumination  may 
be  fed  from  the  simplest  type  of  two-wire  distribution,  shown  in 
Pig.  8,  or  they  may  be  connected  between  either  outer  wire  and 
the  neutral  of  a  three-wire  system,  as  in  Fig.  9,  or  iiiey  may  be 

>y  Google 

Generation  and  Disteibotion  of  Eleotbicitt        846 

connected  across  any  branch  of  a  eecondarj  potyphaee  lighting  and 
power  network,  as  already  described.  The  nsual  voltage  in  these 
cases  is  between  110  and  130.  Incandescent  lamps  for  operatiott 
on  320  Tolts  are  obtainable,  and  are  soinetimes  used  on  a  440-volt 
three-wire  system  to  secure  the  benefits  of  the  higher  voltage  for 
distribution.  These  lamps  are  inefficient  and  less  rugged  than 
those  of  lower  voltage,  and  this  system  has  not  beeo  generally 

Obviously  the  incandescent  lamp  may  be  supplied  from  either 
alternating  or  continuous-current  circuits.  Several  exceptions  to 
this  statement  must  be  noted,  however.  The  Hfe  of  the  tantalum 
lamp,  for  reasons  not  yet  understood,  is  shortened  when  used  on 
alternating  circuits;  the  amount  of  this  shortening  is  about  50 
per  cent  when  operated  at  60  cycles.  In  many  instances  it  is 
jweaible  to  defect  a  Sicker  in  lamps  operated  from  25-cycle  circuits; 
this  is  most  noticeable  in  lamps  of  low  candle-power  and  high 
voltage  in  which  the  filament  is  necessarily  of  small  diameter. 
Xemst  lamps  operate  on  110-  and  230-volt  constant-potential 
alternating  circuits.  A  glower  adapted  to  continuous  current  has 
been  developed  but  has  not  met  with  success.  Such  lamps,  there- 
fore, are  adapted  to  the  several  types  of  altemating-distribnting 
^tems  only. 

The  mercury  vapor  lamp  is  best  adapted  to  constant-potential 
continuous-current  systems,  but  is  also  manufactured  for  alter- 
nating service.  It  may  be  constructed  for  a  wide  range  of  voltage, 
but  is  commonly  manufactured  for  110-  and  SSO-volt  circtiits. 

Interior  illumination  of  stores,  factories,  etc.,  by  means  of  arc 
lamps  is  quite  common,  and  in  such  instances  the  lamps  are  oper- 
ated from  constant-potential  circuits.  Although  the  arc  lamp  in 
its  best  form  is  a  constant-current  device,  constant-current  cir- 
cuits are  usually  of  a  voltage  too  high  for  introduction  into  build- 
ings. Multiple  or  constant-potential  arc  lamps  have,  therefore, 
been  developed,  and  it  is  now  possible  to  secure  an  arc  lamp  suit- 
able for  any  type  of  available  supply  system.  Thus  single  arc 
lamps  may  be  supplied  from  either  110-volt  side  or  from  the 
230-volt  outer  wires  of  either  a  continuous  or  an  alternating 
three-wire  distributing  system.  Lamps  for  110  volts  continuous 
current  may  be  operated  singly  or  two,  four  and  five  in  series 
from  1I0-,  220-,  440-,  550-  (railway)  volt  circuits.  Lamps  may 
be  operated  singly,  by  means  of  compensators  or  transformers  from 
alternating  circuits  of  any  voltage  or  frequency. 

>y  Google 

246  Illithinatinq  EiraiNBEBiNa 

Incandescent  lamps  are  frequently  connected  several  in  series, 
vhere  the  available  voltage  is  higher  than  that  of  the  lamps.  The 
most  familiar  instance  of  this  method  of  connection  ie  fonnd  in 
the  cam  and  buildings  of  the  street-rail^ra;  eyetem  operating  at 
550  volte.  The  introduction  of  the  efGcient  metallic  filament  has 
led  to  an  extension  of  this  method  of  connection,  aa  applied  to  low- 
voltage,  low  candle-power  incandescent  lamps  operating  from  110- 
volt  circuits.  Tids  method  of  connection  has  arisen  from  the 
desire  for  a  lamp  of  lower  rating  and  more  durable  construction 
than  the  S5-watt,  llO-volt  tungsten.  Standard-base  tungeten 
lamps  may  now  be  bad  for  any  voltage  below  130,  and  lamps 
of  1.25-watt  consumption  in  10-,  15-  and  SO-watt  sizes  and  at 
voltages  from  10  volte  upward  are  standard  with  manufactnrerB. 
The  obvious  objection  to  this  series  parallel  method  of  connection 
ie  the  fact  that  the  failure  of  one  lamp  cute  out  the  others  in 
series  with  it.  These  lamps  are  especially  hardy,  however,  owing 
to  their  short,  stout  filaments,  and  when  once  in  place  in  rigid 
sockets  the  plan  is  well  adapted  to  long  passageways  and  other 
areas  requiring  four  or  more  units  of  low  intensity  without  inde- 
pendent control.  Four  28-volt,  10-watt  lamps  in  series  on  a  115- 
volt  circuit  is  a  very  satiafactoij  instance  of  this  type  of  connection. 
Ten  10-  or  13-volt,  5-watt  lamps  in  series  are  Bometimes  used 
for  sign  lighting,  but  the  arrangement  is  not  satisfactory  owing 
to  the  result  consequent  upon  the  failure  of  one  lamp. 

Multiple  operation  and  independent  control  of  low-voltage  lamps 
on  existing  multiple  wiring  ie  possible  on  alternating  circuits  by 
the  use  of  transformers  and  "  economy  coils "  or  auto-trans- 
formers. With  the  use  of  the  latter  it  is  possible  to  operate  the 
lamps  in  series,  and  the  failure  of  one  lamp  will  not  affect  the 

(c)  Design  of  Electrical  System.  The  designer  of  interior  il- 
lumination will  rarely  if  ever  find  it  necessary  to  extend  the  system 
of  electrical  conductors  beyond  the  property  line.  In  cities,  for  ex- 
ample, the  source  of  supply  will  be  either  an  underground  continu- 
ous-current three-wire  network  with  connections  from  a  nearby 
manhole  available  at  the  building  line,  or  an  overhead  secondary 
alternating-current  line  at  a  greater  or  less  distance,  or  in  many  of 
the  larger  problems  the  power  supply  will  be  in  or  very  near  the 
building  itself  in  the  form  of  an  isolated  plant. 

Considering,  first,  the  source  of  supply,  the  careful  engineer 

>y  Google 


will  consideT  (a)  its  capacity,  (b)  ite  reliability,  (c)  its  voltage 
legalatioD,  and  (i)  its  diebuice  and  the  class  of  current,  i.  e., 
whether  coDtinuous  or  alternating  carrent,  and  if  the  latter,  its 
frequency  and  voltage.  The  capacity  oi  the  soorce  of  supply,  in- 
dading  the  tranamlBaion  conductors  up  to  the  point  from  which 
the  new  lighting  load  is  to  be  taken,  must  be  sufficient  to  ensure 
satisfactory  and  continnous  operation  of  all  the  loads  upon  it. 
Moreover,  the  application  or  removal  of  any  individual  load  should 
be  without  disturbing  effect  on  any  of  the  others.  The  questions 
of  capacity  and  reliability  are  not  apt  to  arise  in  connection  with 
an  underground  continuous-current  network,  oi  with  an  individual 
isolated  plant  constructed  especially  for  the  system  under  design. 
Also,  in  the  commoner  instances  of  alternating-current  secondary- 
distributing  systems,  since  the  transformers  are  the  property  of 
the  supply  company,  the  question  of  capacity  will  be  taken  care 
of  by  them.  The  question  of  reliability  in  such  systems  assumes 
importance  when  the  transformers  are  at  the  end  of  long  feeders, 
particularly  if  the  latter  pass  through  open  country  for  any  dis- 
tance. Suburban  lighting  from  the  best  class  of  supply  company 
is  often  subject  to  interruption  from  line  troubles  due  to  wind, 
snow  and  deet,  and  lightning.  For  large  installationB,  where  even 
a  short  interruption  is  to  be  avoided,  these  considerations  may  be 
sufficient  to  justify  an  isolated  plant.  Generally,  however,  occa- 
sional brief  interruptions  in  localities  where  this  class  of  service 
is  the  only  one  available  can  be  tolerated,  and  the  engineer  need 
only  satisfy  himself  that  the  overhead  lines  are  of  approved  con- 
struction and  protected  by  lightning  arresters  of  design  and  loca- 
tion dictated  by  the  beet  present-day  knowledge  of  this  imperfectly 
understood  portion  of  the  problem.  A  committee  of  the  National 
Electric  light  Association  is  now  engaged  in  an  effort  to  standard- 
ize the  methods  of  installing  the  various  types  of  distributing 
system.  At  this  time  definite  recommendations  have  covered  sec- 
ondary-distributing systems  only,  and  are  contained  in  a  report  to 
the  Association  at  its  convention  in  May,  1910.  The  work  of  this 
committee  when  completed  will  furnish  an  excellent  reference  for 
all  questions  of  overhead  lighting  wiring.  Should  the  engineer 
find  it  necessary  to  specify  and  install  his  own  transformers,  the 
questions  as  to  the  methods  of  installing  them,  together  with  their 
fuse  blocks,  are  considered  at  length  in  the  report  mentioned 
above.    The  transformer  capacity  and  subdivision,  in  its  relation 

>y  Google 


to  the  install atioQ,  depend  on  the  time  dlBtiibution  and  concen- 
tration of  the  connected  load.  These  two  elements  are  usually 
combined  in  the  "  diversity  factor,"  or  the  ratio  of  the  sum  of 
the  maxituum  demands  of  the  several  consumers  to  the  maximum 
demand  which  actually  results  from  their  combined  service.  For 
transformers  in  residence  lighting  this  factor  is  about  3,  in  com- 
mercial lighting  from  1.6  to  1.1, 

Speaking  generally,  transformers  may  be  operated  for  an  hour 
or  two  at  50  per  cent  over  their  rated  capacity,  and  for  short 
intervals  at  75  per  cent  or  100  per  cent.  On  accotint  of  the  short 
distances  to  which  low-voltage  altematizig  current  may  be  trans- 
mitted, transformers  on  poles  rarely  exceed  15  kw.  in  capacify, 
and  30  kw.  is  about  the  limit  in  size  for  transformers  for  lighting 

The  voltage  regulation  of  the  supply  Hystem,  next  to  constancy 
of  service,  is  the  most  important  factor  for  satisfactory  lighting. 
Too  often  the  engineer  has  to  be  content  in  this  particular  witii 
what  be  can  get.  In  the  present  state  of  the  art  it  is  rarely  pos- 
sible to  secure  from  a  supply  company  any  statement  or  guarantee 
as  to  the  limits  of  fluctuation  of  its  voltage.  Probably  the  most 
constant  voltage  obtainable  js  that  in  the  best  type  of  isolated 
continuous-current  plant,  as  found  in  a  few  modem  office  buildings 
with  special  provision  for  motor  loads.  In  this  case  the  feeders 
are  all  short,  and  the  regulation  approzimateB  the  practical  con- 
stancy obtainable  in  compound  generators.  Often,  however,  the 
isolated  plant  has  too  little  capacity,  and  carries  both  motor  and 
lamp  loads  without  special  regulating  apparatus.  In  such  cases 
the  regulation  is  very  poor. 

The  underground  230-volt,  three-wire  continuous-current  net- 
work of  the  best  type  of  city  plant  yields  excellent  regulation. 
Such  a  system  comprises  a  close  network  of  mains,  often  compris- 
ing several  1,000,000  circularrmila  cables.  The  voltage  in  this 
network  is  maintained  constant  by  connecting  it  at  various  points 
with  feeders  from  the  station.  Potential  wires,  as  already  men- 
tioned, are  also  run  to  the  station  and  indicate  the  voltage  through- 
out the  network.  The  voltage  on  the  feeders  is  varied  according 
to  the  needs  by  connection  to  several  sets  of  bus-bars  of  different 
voltages,  operated  from  separate  machines,  or  through  boosters 
and  other  regulating  devices.  The  method  is  indicated  in  Pig.  10. 
The  load  changes  of  such  a  system,  owing  to  its  size,  are  quita 

>y  Google 

Gbkeoation  and  DiBTBiBnTioif  OF  Electricity        249 

uniform,  so  that  voltage  adjustment  at  any  point  of  the  network 
is  simple.  In  Buch  a  system  a  daily  constancy  within  1  or  3  volte 
is  obtainaMe. 

Alternating-current  Becondary-distributing  syatfims  do  not,  as 
a  Ufiaal  thing,  afEord  ae  Batisfactory  voltage  regulation  as  the  con- 
tinaous-CDrreiit  eyBtem.  The  drop  in  the  primary  wires  is  rarely 
a  distorbing  factor,  since  this  is  compensated  fot  in  the  station, 
and  that  in  the  transformer  may  be  less  than  2  per  cent  on  non- 
inductive  incandescent-lamp  load.  A  serious  drop  due  to  induc- 
tive reactance,  however,  occurs  in  the  low-voltage  secondary  cir- 
cuits, and  limits  their  length  to  comparatively  short  distances. 
For  this  reason  secondary  networks,  commensurate  in  size  with 
thoBe  of  the  continuous-current  system,  are  not  used.     In  such 


Fia.  10. — Station  ConnectlonB  Direct-Current  Feedere. 

an  alternating-current  network  a  transformer  fed  from  a  separate 
pair  of  primary  wires  constitutes  a  feeder  corresponding  to  that 
of  the  continuous-current  network,  and  since  altemating-vottage 
regulation  is  simpler,  the  station  apparatus  of  this  system  is  less 
elaborate  and,  therefore,  cheaper.  The  transformers  must  be  close 
together,  however,  owing  to  the  drop  in  the  secondary  circuits, 
and  this  condition  is  greatly  aggravated  in  the  fairly  common 
event  of  one  transformer  getting  into  trouble.  These  facte  are 
sufficient  to  have  restricted  the  alternating  network  to  compara- 
tively limited  areas.  When  several  transformers  are  connected  to 
tiie  same  primary  circuit,  station  control  compensates  for  the 
variable  drop  of  changing  load;  obviously  that  this  arrangement 
be  satisfactory  to  all  conBumers,  they  should  alt  have  approximately 
the  same  type  of  load  variation.  It  will  be  thus  seen  that  the 
voltage  regulation  of  alternating  sources  of  supply  may  be  good 

>y  Google 

260  iLLDuiNAnNG  Enoinbebinq 

or  bad  depending  on  the  type  of  control  at  the  station,  the  number 
of  consumers  on  a  line,  the  particular  way  in  which  these  con- 
sumers vary  their  demand,  etc  If  the  lighting  circuits  are  also 
uAed  for  motors,  it  is  etill  more  difficult  to  eecure  good  regulation. 
In  the  better  classee  of  Berrice  momentary  variations  of  1  or  2 
volts  in  110  should  not  be  cause  for  complaint.  As  the  load  goes 
on  the  voltage  is  raised  at  the  station  either  automatically  or  by 
hand,  and  this  may  cause  an  extreme  daily  variatioB  of  3  or  4 
volts.  Departing  from  the  beat  claes  of  service)  it  is  possible  to 
find  almost  any  degree  of  poor  regulation  in  lighting  drcoita. 
In  these  days,  however,  a  total  daily  variation  greater  tiian  5  per 
cent  should  not  be  tolerated  from  a  company  professing  to  give 
first-class  service. 

With  the  voltage  variation  of  the  supply  system  given,  the  engi- 
neer must  design  his  distributing  system  so  as  to  add  as  little 
voltage  variation  as  possible,  and  so  keep  the  voltage  at  the  lamp 
as  nearly  constant  as  the  source  of  supply  will  permit.  The  prin- 
ciples involved  in  this  design  are  simple,  and  the  problem  is 
usually  the  very  indefinite  one  of  a  decision  as  to  what  additional 
drop  to  allow  in  order  to  secure  a  low  cost  of  the  distributing 
system.  With  continuous  current  the  application  of  Ohm's  law  in 
one  of  its  several  forms  will  determine  the  size  of  conductor  for 
the  choseu  voltage  drop.  Temperature  variation  of  resistance, 
however,  must  be  duly  considered.  If  many  calculations  are  to 
be  made  it  is  usually  worth  while  to  make  use  of  tables  giving 
relations  between  current,  voltage  drop,  distance,  etc.,  such  as  ma; 
be  found  in  Hering's  Wiring  Computer  and  other  like  works.  In 
most  cases,  however,  it  will  be  more  satisfactory  to  make  calcula- 
tion using  resistance  tables  with  temperature  factors  clearly  given. 

The  loss  or  drop  in  voltage  in  alternating-current  circuits  is 
due  to  resistance  and  reactance.  The  resistance  may  usually  be 
that  given  by  any  wire  table  with  temperature  correction.  The 
reactance  drop  is  caused  by  the  electromotive  force  induced  in  the 
circuit  by  its  own  alternating  magnetic  field.  This  electromotive 
force  is  therefore  proportional  to  the  current,  and  to  the  frequency, 
and  to  the  self-inductance,  which  depends  on  the  length,  the  sepa- 
ration and  the  size  of  the  conductors.  The  mathematical  expres- 
sion for  the  reactance  in  ohms  is  2irNIj,  and  for  the  reactance 
volts  SwNLi,  N  being  the  frequency,  L  the  self-inductance  and  i 
the  current.    The  resistance  and  reactance  volts  are  both  propor- 

>y  Google 

Generation  and  Distkibdtion  op  Electbicitt        261 

tional  to  the  current,  but  differ  in  phase  by  one-quarter  of  a 
period  Eo  that  the  total  drop  in  volte  is  the  square  root  of  the 
Bum  of  the  squares  of  the  resistance  and  reactance  volte.  While 
the  resistauce  decreases  rapidly  with  increasing  eize  of  wire  the 
reactance  decreases  very  sUghtly,  consequently  there  is  in  ali«r- 
nating-current  distribution  an  early  limit  to  the  improvement 
in  voltage  regulation  by  increase  in  the  size  of  conductor.  It  is 
for  this  reason  that  low-tension  alternating  circuits  must  be  short. 
The  resistance  and  reactance  at  60  cycles  per  mile  of  a  circuit  of 
two  No.  5  wires,  24  inches  apart,  in  ohms,  are  3.24  and  1.4;  for 
No.  00  the  values  are  .8  and  1.24 ;  it  is  seen  that  for  sizes  in  this 
neighborhood  little  is  gained  by  increasing  the  size  of  wire.  Ccon- 
plet«  tables  of  resistance,  reactance  and  impedance  volts  for  vari- 
ous sizes  of  wire,  separation,  frequency,  are  now  readily  available, 
so  that  calculations  may  be  quickly  made.  Attention  to  the  re- 
actance drop  is  especially  necessary  in  designing  overhead  service 
connections  with  eptne  separation,  and  must  not  be  lost  sight  of 
even  in  interior  wiring  where  the  two  sides  of  the  circuit  are 
close  together  inside  one  conduit  For  example,  two  No.  0  wires 
in  a  2-inch  conduit  may  easily  have  an  average  interazial  separa- 
tion of  1  inch ;  the  reactance  per  1000  feet  is  about  .1  ohm  or 
one-half  as  great  as  the  resistance;  the  impedance  is  therefore 
.284  or  25  per  cent  greater  than  the  resistance. 

Secondary-distributing  networks  must  therefore  have  trans- 
formers connected  at  fairly  frequent  intervals.  A  common  method 
is  to  run  three-phase  primaries,  supplying  three  or  four  city  blocks 
from  one  phase  through  three  transformers  with  their  secondaries 
connected  to  a  common  three-wire  main.  The  next  three  blocks  go 
on  the  next  phase,  etc.,  preserving  the  balance  as  far  as  possible. 
Speaking  broadly  high-class  secondary  distributing  or  service  cir- 
cuits should  not  exceed  400  to  600  feet  in  length,  or  between 

In  calculating  the  wiring  for  any  installation  two  types  of  volt- 
age drop  must  be  considered,  viz.,  that  due  to  the  gradual  daily 
increase  of  the  total  load,  and  that  due  to  the  sudden  cutting  in 
or  out  of  a  portion  of  the  total  load.  The  former  occurs  gradually 
and  principally  in  the  mains  from  the  supply  system,  and  in  the 
"  risers  "  and  distributing  feeders.  If  the  maximum  load  on  all 
branches  is  deiinitely  known,  the  voltage  drop  due  to  this  cause 
may,  of  course,  be  kept  within  any  limits  by  proper  choice  of  con- 

>y  Google 

26S  Illuuinatino  Enoinbbrino 

ductors.  A  wide  range  of  Tariatdon  o(  thiB  kind  ie  very  objection- 
able. If  the  lamps  are  chosen  foi  Hie  high  Toltage,  their  InminoOB 
efBcienc]'  is  impaired  as  the  load  goeB  on;  if  for  the  low  voltage 
the  life  of  the  lamps  used  at  light  loads  is  shortened.  As  this 
type  of  variation  ie  gradual  it  is  not  noticeable,  and  too  little  at- 
tention is  given  it  in  design.  The  second  type  of  voltage  variation 
is  necessarily  less  in  amount  than  the  first,  and  ie  principally  ob- 
jectionable in  causing  a  momentary  fluctuation  of  Hgbt  from  other 
burning  lamps.  This  disturbance  ie  reduced  by  designing  so  that 
the  smaller  part  of  the  total  permissible  drop  takes  place  in  the 
mains  and  principal  distributing  feeders,  and  by  increasing  the 
number  of  branch  feedere.  A  change  of  1  per  cent  in  the  voltage 
on  a  tungsten  tamp  causes  a  variation  of  4  per  cent  in  its  candle- 
power.  Fluctuations  of  this  nature,  therefore,  are  to  be  par- 
ticularly guarded  against  in  those  cases  where  there  is  frequent 
cutting  in  or  out  of  large  numbers  of  lights. 

It  is  difficult  and  scarcely  necessary  to  fix  an  absolute  limit  to 
the  permissible  voltage  variation  on  an  incandescent  lamp.  Sat- 
isfactory illumination  is  given  by  the  120-volt  tungsten  lamp  over 
a  range  of  4  volts  or  more  than  3  per  cent ;  in  fact,  a  given  lamp 
is  now  rated  for  3  voltages  covering  this  range.  The  important 
consequences  of  this  variation  are  the  effects  on  the  efficiency  and 
life  of  the  lamp,  rather  than  on  the  illumination.  Speaking  gen- 
erally, with  a  supply  system  constant  to  within  1  or  2  per  cent, 
very  satisfactory  service  will  be  given  if  the  maximnm  voltage 
drop  inside  the  service  connection  be  limited  to  3  per  cent.  Of 
this  the  smaller  part  should  be  in  the  service  wires  and  larger 
branches.  A  greater  drop  than  this  may  be  allowed  when  the 
greater  proportion  of  the  lamps  are  operated  together,  and  so 
cause  approximately  a  fixed  drop  in  the  service  wires.  With  the 
entire  load  connected  as  one  unit,  i.  e.,  with  no  independent  opera- 
tion of  single  lamps,  any  amount  of  drop  in  the  service  connection 
may  be  allowed,  by  a  proper  choice  of  lamp.  The  calculation  of 
the  size  of  service  wires,  feeders  and  branches  to  meet  the  require- 
ments of  voltage  is  thus  a  simple  matter  of  distances  as  soon  as  the 
location  of  the  individual  outlets  and  sizes  of  lamps  are  fixed. 

The  usual  installation  begins  at  the  service  wires,  which  may 
be  either  overhead  or  underground;  these  are  generally  320-volt, 
three-wire,  carried  directly  to  a  center  of  distribution,  which  may 
be  of  any  degree  of  elaboration.     A  simple  iron  box  containing 

>y  Google 


a  main  ^vitcti  and  fused  branch  cut-outa  aiifBceg  for  a  small  resi- 
dence. For  large  installations  a  switchboard  havlDg  panels  for 
the  main  connection  and  individnal  feeders,  as  found  in  the 
largest  buildings,  may  be  required.  From  this  center  feeders 
run  to  distributing  centers  in  various  portions  of  the  building. 
From  these  distributing  centers  sub-feeders  are  often  taken  to 
smaller  local  centers,  though,  more  commonly,  so-called  branch 
circuits  lead  directly  to  the  lamp  outlets.  The  number  of  feedera 
and  sub-feeders  is  regulated  by  the  height  and  the  door  area  of 
the  building.  For  great  heights  individual  feeders  for  one  or 
more  floors  may  be  necessary.  Generally,  however,  several  floors 
may  be  fed  from  one  riser.  For  large  areas,  sub-feeders  from  the 
distributing  to  local  centers  may  often  be  used  to  advantage.  The 
three-wire  system  is  carried  to  the  centers  where  branch  circuits 
are  connected.  Probably  the  most  important  factor  in  determin- 
ing the  number  of  feeders  is  the  permisBible  length  of  branch 
circuit.  The  "  National  Electrical  Code  "  limits  the  lamp  capacity 
of  a  single  branch  circuit  to  660  watts.  This  figure  was  chosen 
as  representing  twelve  55-watt  carbon  lamps.  Under  this  rule 
it  is  now  possible  to  install  twenty-six  S5-watt  tungsten  lamps, 
although  such  a  plan  is  not  advisable.  Further,  no  wire  smaller 
than  a  No.  12  B.  &  S.  sbotUd  be  used  for  the  branch  circuits.  At 
115  volta,  660  watts  represent  about  5.5  amperes,  and  the  re- 
sistance of  No.  12  wire  is  1.62  ohms  per  1000  feet.  An  average 
length  of  50  feet  of  this  circuit  would  therefore  cause  a  drop  of 
1  Tolt.  With  good  regulation  of  supply  system  and  ample  copper 
in  service  wires  and  feeders,  branch  circuits  may  sometimes  have 
a  length  of  100  feet,  but  this  should  be  the  maximum  of  conserva- 
tive practice.  Exceptional  cases  may  be  met  by  increasing  the 
axe  of  the  branch  circuit.  A  radius  between  60  and  100  feet, 
therefore,  marks  the  area  to  be  fed  from  one  center  or  feeder  con- 
nection. In  small  buildings,  such  as  residences,  therefore,  no 
feeders  are  required,  all  branch  circuits  starting  from  a  suitable 
distributing  board  where  the  service  wires  ent^r.  In  larger  build- 
ings the  density  of  the  load  on  various  floors  will  determine 
whether  more  than  one  floor  may  be  fed  from  one  feeder.  The 
low  consumption  of  tunget«n  lamps  will  usually  permit  two  or 
three  floors  to  a  feeder  or  a  riser,  with  10  or  13  branch  circuits 
to  a  floor.  In  such  a  case  it  is  advisable  to  place  a  distributing 
board  on  each  floor,  and  not  extend  the  branch  circuits  from  a 
center  on  one  floor  to  outlets  on  floors  above  and  below. 

>y  Google 

254  Illuiunating  Ekoineebinq 

With  all  feeders  brought  back  to  the  service  connectioii,  vhich 
ahotild  be  located  as  near  the  mean  center  of  load  aa  possible,  it  is 
a  aimple  matter  to  calculate  the  voltage  variation  on  any  lamp 
with  all  lamps  burning.  With  a  fixed  limit  of  variation,  tfais 
condition  vill  uBually  call  for  larger  feeders  and  service  wires  than 
neceeaary.  A  study  of  the  particular  problem  only  can  determioe 
the  probable  maximum  number  of  lights  to  be  operated  at  one 
time.  For  residences  this  number  will  rarely  exceed  one-third  to 
one-half  of  the  total,  while  in  office  buildings,  churches,  theaters, 
etc,  the  total  connected  load  may  often  be  in  operation  at  one 
time.  It  is  only  in  exceptional  cases  that  the  cost  of  copper  in 
the  feeders  is  a  sufficiently  large  proportion  of  the  total  cost  to 
warrant  the  reduction  of  their  size.  No  great  increase  in  cost 
will  generally  result  from  designing  service  wires  and  feeders  to 
the  end  that  the  operation  of  the  maximum  connected  load  will 
only  cause  the  permissible  voltage  variation  on  the  lamp  most  un- 
favorably located. 

S.  Exterior  and  Street  Illumination 
(a)  System!  of  Supply.  Exterior  illumination  may  be  taken 
from  any  available  source  of  supply.  The  uae  of  constant-poten- 
tial continuouB-carrent  service  is  limited  to  loads  concentrated 
within  a  small  area,  owing  to  the  fact  that  the  distributing 
loeaes  mount  very  rapidly  for  any  considerable  distance.  Instances 
of  this  type  of  supply  are  electric  signs  from  110-  to  880-volt 
mains,  multiple  arc  lamps  in  front  of  buildings  or  stores,  and 
street  arches  supplied  from  650-volt  railway  circuits,  the  lamps 
being  connected  in  series-pardlel. 

Constant-potential  alternating  current  at  l^SOO,  4400  or  6600 
volts  is  probably  the  most  common  type  of  exterior  supply  circuit, 
and  it  may  be  utilized  in  various  ways  for  exterior  lighting.  It 
may  be  simply  transformed  to  low-voltage,  constant-potential  ser- 
vice or  to  high-voltage,  constant-alternating  current  for  series  arc 
and  incandescent  circuits,  or  to  high-voltage  continuous  current 
by  means  of  the  mercury  rectifier.  The  last  mentioned  is  perhaps 
the  moat  satisfactory  of  all  methods  of  arc  lighting. 

For  many  years  constant-continuoUB  current,  series  arc  circuits 
were  supplied  from  Thomson-Houston  and  Brush  constantKnirrent 
generators.  Many  instances  of  the  latter  type  of  Installation  are 
stilt  in  operation,  and  these  machines  operate  with  as  high  voltage 

>y  Google 

Generation  and  Distbibdtion  of  Electricitt        255 

as  13,000  with  cnTrente  of  5  to  10  amperes  supplying  upwards 
of  320  lamps.  These  excellent  machines,  after  a  highly  honorable 
record,  are  now  being  rapidly  gupplanted  by  constant-potential  to 
constant-carrent  transformers  fed  from  3300  volts  constant-poten- 
tial alternating  circmta  and  Bupplying  on  the  secondary  side  coa- 
stant-alternating  current.  These  transformers  are  equipped  with 
one  stationary  coil  and  a  movable  coil  which  automatically  shifts 
its  difltance  from  the  fixed  coil  to  meet  the  demands  of  the  load. 

FiQ.  11, — Conatant-Current  TranBtormer, 

In  a  transformer  imder  load  there  is  a  repulsive  force  between 
the  two  coils.  In  ordinary  constant-potential  transformers  this 
force  is  held  in  check  by  the  close-fitting  iron  of  the  magnetic 
ciTcnit.  In  the  constant-current  transformer  this  force  is  allowed 
to  act,  free  motion  of  the  secondary  away  from  the  primary  being 
allowed  by  providing  a  greater  opening  in  the  magnetic  circuit 
than  is  required  by  the  cross-section  of  the  coils.  But  a  separation 
of  the  two  coils,  due  to  a  rise  in  current,  is  accompanied  by  a  fall 
in  the  secondary  voltage,  since  a  portion  of  the  magnetic  field  set 
np  by  the  primary  leaks  across  the  gap  between  the  coils  and  eo 

>y  Google 


Illuhinating  Enqinebrinq 

does  not  pass  through  the  Becondarj.  The  teodency  to  a  rise  in 
current  is  thua  checked  by  a  fall  in  voltage.  By  means  of  suitable 
counter-balancing  of  the  weight  of  the  movable  coil,  and  by  other 
auxiliaiy  devicee,  the  transformer  regulates  very  closely  for  con- 
stant current,  and  arc  circuits  may  be  taken  directly  from  their 

Fig.  11  sho^  a  picture  of  this  transformer.  More  satisfactory, 
however,  is  the  series  continuous  current  arc  circuit,  vhich  may 
be  had  by  combining  with  the  constant-current  tranaformer  a 
mercury-arc  rectifier.     The  combination  gives  excellent  constant 

Pro.  12. — Direct-Current  Serle*  Arc  Rectlfler. 

cuntinnouB  current  regulation.  The  method  of  connection  is  il- 
lustrated in  Fig.  13,  and  the  apparatus  provides  a  very  reliable 
means  of  transformation  between  constant  altsmating-potential 
and  constant  continuous  current.  These  equipments  are  available 
for  any  voltage  between  SSO  and  13,000,  and  for  any  standard  of 
frequency.  They  may  be  had  in  sizes  supplying  as  many  as  75 

(b)  Systems  of  DiitribDtio&.  Arc  lamps  may  be  operated  from 
any  available  source.  Their  operation  on  low-voltage,  constant- 
potential  circuits  is  less  satisfactory  than  on  a  constant-curnmt 
circuit.  The  alternating-current  multiple  tamp  is  the  moat  na- 
satififactory  of  alt,  but  its  use  is  often  justifiable  in  outiying  dis- 

>y  Google 

Qkneratios  and  DieTHiBCTioN  OP  Electricity        257 

tricts  with  widely  scattered  lights.  The  conBtaDt-cnrrent  series 
method  of  connecting  arc  lamps  is  the  moat  common  of  all,  and 
the  series  circnite  may  be  either  alternating  or  continuous  current 
with  preference  for  the  latter.  These  circuits  cover  wide  expames 
of  territory,  and  since  the  connection  from  lamp  to  lamp  is  by  one 
conductor  only  the  distributing  system  is  simple  and  may  be 
looped  in  various  directions. 

Esterior  lighting  by  incandescent  lamps  may  also  be  adapted 
to  any  class  of  service.  For  condensed  loads,  such  as  signs,  it  is 
possible  to  use  the  multiple  connection  from  low-volt^e  circuits 
using  standard  lamps.  Low-power,  low-volt^e  lamps  must  be  con- 
nected two  or  more  in  series  on  continuous-current  circuits;  on 
alternating  circuits  the  use  of  small  transformers  or  economy  coils 
will  avoid  the  undesirable  series  connection.  Street  lighting  is 
sometimes  accomplished  by  the  series-parallel  connection  of  in- 
candescent lamps  OQ  railway  circuits  or  other  available  constant- 
potential  lines,  but  the  most  acceptable  system  for  distributed 
and  uniform  lighting  by  incandescent  lamps  is  the  series  connection 
of  a  number  of  these  lamps  across  constant-potential  high-voltage 
alternating  circuits,  or  in  constant-current  circuits,  either  alter- 
nating or  continuous.  On  constant-potential  high-voltage  service 
the  lamps  are  shunted  by  small  reactance  coils,  so  that  the  circuit 
is  continuous,  and  the  voltage  consumed  by  reactance  if  an  indi- 
vidual lamp  should  fail.  Tungsten  lamps  for  this  system  are  to 
be  had  with  ratings  of  1.75  to  4  amperes,  and  from  8  to  40  volts, 
so  that  it  is  possible  to  operate  360  such  lamps  in  series  on  2800- 
Tolt  circuits.  It  is  claimed  for  the  ^tem  that  operation  is  still 
satisfactory  with  20  per  cent  of  the  lamps  broken  or  out.  The 
series  connection  of  tungsten  lamps  may  also  be  operated  at 
constant-alternating  current  by  use  of  the  constant-current  trans- 
former already  described.  By  this  method  upwards  of  seven  hun- 
dred 30-watt,  3.5  or  6.6  amperes,  lamps  may  be  operated  in  series. 
In  this  system  each  lamp  is  equipped  with  an  insulating  film  which 
withstands  the  lamp  voltage  but  breaks  down  if  the  lamp  fails,  thus 
preserving  the  continuity  of  the  circuit.  The  transformer  ad- 
justs automatically  for  the  lowered  voltage. 

(c)  Des^  of  the  Electric  Syttera.  Constant-potential  regula- 
tion is  obviously  much  less  important  in  outside  than  in  inside 
lighting.  Incandescent  lamps  employed  in  this  class  of  service  are 
comparatively  few,  and  the  objections  to  voltage  fluctuations  are 

>y  Google 

858  Illcmikatinq  Enqikeerinq 

practically  limited  to  the  effect  on  the  life  of  the  lamps.  The  arc 
lamp  is  essentially  a  eonstant-culrent  device  and  is  not  Beriously 
affected  by  slight  voltage  variations.  For  the  short  connections 
usual  in  the  use  of  constant-potential  arc  lamps  no  special  calcula- 
tion is  necessary  for  the  wiring  beyond  providing  ample  current- 
carrying  capacity.  Series  arc  lamps  take  from  4  to  9^  amperes, 
according  to  the  type.  The  commonest  types  are  the  4-ampere 
continuous  luminous  tamp  and  the  alternating  or  con- 
tinuous enclosed  lamps.  The  regulation  of  the  Brusli  generator, 
of  the  constant-current  transformer,  and  of  the  mercury  rectifier 
are  all  extremely  close;  it  follows,  therefore,  that  the  design  of 
the  distributing  system  for  exterior  illumination  will  very  rarely 
involve  any  serious  problems  of  voltage  regulation.  The  series 
circuits  themselves  and  the  resistance  in  lamps  consume  a  large 
part  of  the  applied  voltage,  and  the  constant-current  regulating 
devices  adjust  automatically  to  a  wide  range  of  resistance.  The 
series  circuits  of  large  cities  carry  50,  75  and  100  lamps  at  volt- 
ages from  4000  to  8000,  The  distance  of  the  separation  of  lamps 
varies,  but  averages  from  300  to  300  feet.  The  voltage  drop  in 
the  conductor  itself  is  usually  between  5  and  10  per  cent,  and 
wires  in  the  neighborhood  of  Jfo,  8  B.  &  S.  are  used.  The  low- 
tensile  strength  of  smaller  wires  renders  their  use  inadvisable.  It 
is  not  uncommon  to  find  a  circuit  of  this  kind  comprising  10  miles 
of  single  No,  8  wire  and  seventy-five  4-arapere  lamps. 

Series  incandescent  lighting  from  constant-potential  high-volt- 
age circuits  is  accomplished  with  lamps  taking  from  1.5  to  4 
amperes.  For  voltages  above  550,  the  circuit  should  be  insulated 
from  the  main  line  by  a  transformer.  In  series  incandescent  cir- 
cuits fed  from  constant-current  transformers  the  lamps  may  be 
had  for  currents  between  1.75  and  10  amperes.  The  common  size 
of  wire  for  this  class  of  service  is  from  No,  10  B.  &  S.  up.  The 
regulators  are  rated  in  terms  of  the  aggregate  kilowatt  capacity 
of  total  connected  lamps.  This  rating  includes  an  allowance  of 
5  per  cent  ohmic  and  10  per  cent  reactive  drop  in  the  series 

The  insulation  of  overhead  conductors  should  be  of  the  beet 
rubber  core  and  braided  class  of  manufactured  product.  The 
underground  conductor  may  be  either  fiber-,  paper-  or  rubber-in- 
sulated stranded  conductor,  and  in  every  case  is  surrounded  by 
lead.     These  cables  should  withstand  the  test  prescribed  for  all 

>y  Google 

Generation  and  Disthibution  op  Electhicitt        359 

high-voltage  apparatus,  namely,  they  should  be  subjected  to  double 
the  maxiinuiQ  voltage  for  a  period  of  1  minute. 

S.  Metering 

The  subject  of  metering  is  a  highly  important  one  "in  the  com- 
plete discussion  of  the  entire  electric-lighting  Bvatem.  Meters  are 
asually  owned,  inspected,  tested  and  read  by  the  company  supplying 
power.  'The  illuminating  engineer  will  rarely  be  called  upon  to 
do  more  than  provide  proper  spacing  and  accommodation  for 

Almost  invariably  the  present-day  meter  measures  watt  hours. 
For  continuous-current  service  the  best  meters  are  essentially  the 
same  as  the  original  Thomson  watt-hour  meter.  This  consists  of 
a  continuous-current  shunt  motor  containing  no  iron.  The  line 
current  flows  in  the  field  circuit  and  the  line  voltage  is  applied 
to  the  rotating  armature  with  the  insertion  of  a  very  high  r&- 
sistance.  This  permanent  shunt  connection  across  the  circuit, 
therefore,  takes  current  at  all  times.  While  the  current  of  an 
individual  meter  is  extremely  small,  nevertheless,  the  aggregate  of 
the  meters  of  a  large  system  results  in  quite  an  appreciable  fraction 
of  the  total  load  on  the  station.  The  retarding  force  on  the  arma- 
ture of  this  meter  is  a  copper  disc  rotating  between  the  poles  of 
several  permanent  magnets.  The  shaft  of  the  armature  is  equipped 
with  a  small  pinion  which  engages  a  train  of  gears  counected  with 
dials  constituting  the  recording  mechanism.  For  alternating  cur- 
rents the  induction-watt  meter  has  many  advantages  over  the 
Thomson  type,  although  the  latter  may  be  adapted  to  alternating- 
current  service.  Induction  meters  operate  on  the  induction -motor 
principle,  series  and  shunt  coils  with  different  phase  character- 
istics, giving  the  two  components  of  the  rotating  magnetic  field. 
A  light  aluminum  disc  constitutes  the  rotating  element  or  arma- 
ture. By  their  principle  they  read  trae  power,  and  are  independent 
of  phase  difference  between  current  and  electromotive  force. 

Alternating-current  meters  are,  in  general,  more  permanent  and 
reliable  than  those  for  continuous  currents,  in  that  they  have  no 
commutator  nor  brushes.  The  present-day  meter,  as  furnished  by 
the  best  manufacturers,  has  been  brought  to  a  high  degree  of 
perfection,  and  may  be  relied  on  to  a  very  close  figure  of  accuracy. 
No  meter,  however,  will  maintain  its  calibration  indefinitely,  and 
those  in  service  should  be  tested  and  inspected  regularly.     This 

>y  Google 

860  Illdminatinq  Enoinebrino 

is  generally  carried  out  by  the  eapply  company,  which  in  moat 
cases  OVQE  the  meter.  Id  some  instances  a  meter  rental  is  charged 
for  the  purpose  of  covering  not  only  the  original  cost  of  the  meter 
but  for  defraying  this  regular  charge  for  inspection  and  repair. 
Questions  sometimes  arise  between  customers  and  supply  com- 
panies as  to  the  accuracy  of  meter  readings,  and  public  service  com- 
miseions  have  in  many  places  provided  regulations  by  which  a 
consumer  may  demand  a  test  and  calibration  of  his  meter  at  any 
time  by  the  payment  of  a  small  fee.  The  usual  method  of  testing 
meters  is  that  of  comparing  them  with  portable  standard  meters. 
It  is,  of  course,  necessary  that  these  portable  meters  should  be 
compared  with  permanent  standard  instruments  in  the  laboratory 
at  sufficiently  frequent  intervals.  The  methods  of  charging  for 
power  for  lighting  as  based  on  meter  readings  will  be  referred  to 
later  in  these  lectures. 

The  general  subject  of  meters  baa  been  exhaustively  covered  by 
the  reports  of  the  committee  on  meters  of  the  National  Electric 
Light  Association  for  1909  and  1910. 

Tee  Ikstaliation  of  Electric  LiaHTiHO  Systems 
1.  Interior  Illumination 
(a)  Type  of  Installation,  The  engineering  questions  arising 
in  connection  with  the  installation  of  a  system  of  electrical  con- 
ductors for  distributing  electric  power  for  lighting  are  compara- 
tively simple.  Such  distribution  is  accomplished  at  moderate  volt- 
ages for  which  the  space  requirements  are  not  great.  The  usual 
problem  is  that  of  running  a  more  or  less  elaborate  system  of  two- 
or  three-wire  circuits  inside  a  building.  The  objects  which  must 
be  had  prominently  in  view  are  those  of  safety,  reliability,  per- 
manence and  unobtrusive  appearance.  The  system  must  operate 
without  danger  of  fire  or  to  life.  The  possibility  of  fire  arises  in 
the  results  following  short-circuits  and  grounds  in  the  system. 
The  danger  to  life  is  not  generally  present  in  continuous-current 
service  but  arises  in  alternating-current  distributing  systems  fed 
from  transformers  supplied  by  high-potential  primary  circuits.  It 
is  obvious  that  satisfactory  operation  will  require  that  at  all  times 
the  system  will  perform  its  functions  of  not  only  distributing 
power,  but  in  permitting  its  ready  control  and  the  prompt  elimi- 
nation  of  all  abnormal  conditions  likely  to  cause  interruptions. 

>y  Google 


The  life  of  the  installation  depends  largely  on  the  materiala  and 
quality  of  labor  entering  into  ita  oonetruction.  In  thia  regard 
poeeible  exceptions  may  enter  in  the  installation  of  systems  which 
are  to  have  intentionally  a  short  existence.  Generally  speaking, 
however,  the  material  and  workmanship  of  electric-lighting  in- 
stallations should  be  of  the  best  obtainable,  and  in  accordance  with 
the  latest  recommendationB  of  engineering  bodies.  The  distrib- 
uting system  for  residences,  hotels  and  dwellings,  generally,  as 
well  as  in  all  buildings  where  agreeable  and  attractive  appearance 
is  required,  should  be  as  unobtmsiTe  as  possible.  This  considera- 
tion in  the  instances  mentioned  leads  to  the  entire  concealment 
of  electric  wiring.  In  factories  and  other  buildings  where  no 
particular  attention  is  required  as  to  appearance,  the  conductors 
and  BOpporis  are  often  installed  exposed.  This  method  is  a  per- 
fectly satisfactory  one,  if  due  attention  is  paid  to  the  location  of 
the  conductors  in  such  places  as  wiU  render  them  free  from 
mechanical  injury.  Exposed  wiring  presents  the  general  advan- 
tage of  accessibility  and  convenience  of  inspection.  Concealed 
wiring,  on  the  other  hand,  is  almost  invariably  free  from  the 
danger  of  mechanical  injury.  Dedsion  as  to  which  general  method 
should  be  followed  will  depend  on  the  particular  conditions  of  the 

The  methods  of  installing  electric  wiring  are  rigidly  controlled 
by  the  National  Board  of  Fire  Underwriters,^  and  the  regulations 
governing  this  class  of  work  are  published  by  that  body  in  a  pam- 
phlet known  as  "  The  National  Electrical  Code."  In  addition  to 
these  rules  there  is  published  a  list  of  manufactured  material  which 
has  been  subjected  to  laboratory  test,  and  which  is  known  briefly 
as  "  approved  "  material.  In  many  cities  there  is  a  further  list  of 
requirements  which  apply  to  particular  local  conditions. 

Four  classes  of  interior  wiring  are  usually  permitted.  They  are 
known  as  "  open-work,"  "  moulding,"  "  concealed-knob-and-tube  " 
and  "metal-conduit"  installations.  In  open  work  the  wires  are 
run  entirely  exposed  and  supported  on  porcelain  insulators  and 
knobs;  they  pass  through  all  walls,  joist,  partitions,  etc.,  in  por- 
celain tubes.  The  space  requirements  in  the  way  of  separation  of 
wires  from  each  other,  and  from  walls  and  their  relation  to  other 
circuits,  etc.,  are  rigidly  specified.  This  type  of  installation  is 
entirely  satisfactory  where  its  appearance  can  be  tolerated,  and 
is  the  simplest  and  cheapest  to  install.    The  principal  precaution 

>y  Google 

868  Illuminating  Ekginebbino 

to  be  taken  is  against  mechanical  injury.  MonMiog  and  knob  and 
tube  work  have  been  developed  as  methods  for  inBtalling  wiring  in 
buildings  originally  constructed  without  any  idea  of  future  elec- 
tric service.  They  represent  the  moat  unreliable  and  unsatisfactoTy 
types  of  wiring  installation.  In  the  case  of  moulding  the  wires 
are  run  behind  either  wood  or  metal  strips  which  are  laid  on  the 
ceilings  and  walla  of  interiors.  In  knoh  and  tube  work  the  wires 
are  concealed  by  "  fishing "  them  from  point  to  point  behind  the 
plastering  and  under  the  floors  of  buildings  without  disturbance 
to  these  surfaces.  This  method  is  highly  undesirable,  and  even 
when  most  carefully  installed  during  the  progress  of  building  in- 
troduces great  danger  of  fire.  Both  moulding  and  knob  and  tube 
work  are  make-shifts,  and  should  never  be  installed  by  a  carefol 
engineer  unless  absolutely  unavoidable. 

The  complete  enclosure  of  the  entire  wiring  system  up  to  the 
lamp  or  fixture  outlet  in  metal  conduit  represents  the  best  present- 
day  method,  and  one  which  bids  fair  to  form  the  ultimate  standard 
of  construction.  In  this  system  the  entire  wiring  is  completely 
surrounded  by  metal.  The  materials  are  to  be  had  in  the  form 
of  rigid  or  flexible  metal  conduit.  The  rigid  conduit  consists  of 
iron  pipe  of  various  sizes,  and  usually  in  10-foot  lengths.  Elbows, 
bushings  and  other  fittings  are  also  supplied  for  each  size.  This 
conduit  is  usually  made  as  soft  as  possible  to  permit  easy  bonding 
for  adaptation  to  building  peculiarities.  It  is  either  galvanized 
or  covered  inside  and  out  with  some  protective  enamel  which  is 
valuable  in  protecting  the  metal  of  the  conduit  rather  than  as 
insulation  to  the  conductors  enclosed.  Flexible  conduit  com- 
prises the  several  varieties  of  the  familiar  tubing  made  in  spiral 
form  from  cut  steel.  This  conduit  is  best  adapted  to  locations 
where  straight  runs  are  few,  and  where  there  is  difficulty  of  access 
to  wiring  compartments.  With  either  system  of  conduit  construc- 
tion iron  boxes  are  used  for  all  classes  of  outlets.  The  conduit 
leads  to  these  boxes  and  is  mechanically  and  electrically  connected 
to  them  by  means  of  washers  and  nuts.  These  boxes  form  con- 
venient points  for  pulling  the  wires  into  the  conduit  after  the 
latter  is  installed,  and  also  for  making  connections  for  branch 
circuits.  This  type  of  installation  is  readily  installed  in  new 
buildings,  whether  they  be  frame,  brick,  concrete  or  other  class 
of  construction.  Old  buildings  may  generally  be  equipped  with 
electric  wiring  in  flexible-steel  conduit  with  permanent  damage 

>y  Google 

Generation  and  DieTBiBOTios  or  Electhicitt        263 

to  plastering  only.  In  the  cuBe  of  concrete  buildings  the  outlet 
boxes  for  lamps,  switches,  ping  cut-outs,  etc.,  must  be  located  and 
firmly  attached  to  the  forme  with  complete  conduit  interconnection 
before  the  concrete  is  poured.  The  entire  conduit  system  should 
form  a  complete  metallic  system  which  should  be  grounded.  In 
this  condition  the  installation  provides  practically  absolute  safety 
from  mechanical  injury,  and  when  supplemented  by  proper  cut- 
outs and  fuse  apparatus,  from  fires  originating  in  shori;-circuits  or 
grounded  wires.  The  only  objection  to  thia  type  of  installation 
which  has  arisen  is  the  condensation  of  moisture  inside  of  the 
conduits.  This  has  been  koown  to  take  place  to  such  an  extent 
as  to  result  in  the  rotting  of  the  insulation  of  the  wires  due  to 
their  permanent  immersion  in  water.  This  objection  may  be 
largely  obviated  by  running  the  conduit  so  that  there  are  no 
pockets  in  the  system,  and  so  that  they  have  a  pitch  or  slope  towards 
some  outlet.  It  is  customary  to  run  three-wire  mains,  feeders  and 
duplex  branches  in  one  pipe.  It  is  not  permitted,  however,  to  run 
more  than  one  set  in  a  single  pipe.  Reliance,  therefore,  is  placed 
entirely  on  the  insulating  covering  of  the  wires  without  space  sepa- 
ration, and  on  the  suppression  of  any  arc  or  spark  between  con- 
ductors or  between  conductors  and  ground  by  the  walls  of  the 
conduit.  Two-wire  service  is  now  limited  to  the  smallest  instal- 
lations, the  maximum  number  of  outlets  permitted  by  supply  com- 
panies on  two-wire  service  varying  somewhat,  but  generally  not 
exceeding  25.  It  is  permissible  to  run  the  wires  of  either  two-  or 
three-wire  service  in  a  single  pipe.  The  magnetic  influence  of  the 
iron-protective  covering  in  the  case  of  alternating-current  circuits 
has  never  arisen  as  a  prohibitive  factor.  The  running  of  a  single 
wire  carrying  alternating  current  in  an  iron  pipe  is  prohibited  by 
the  large  increase  of  the  impedance  of  the  circuit  and  by  the  heat- 
ing of  the  conduit  due  to  hysteresis  and  eddy  currents.  A  series 
of  tests  have  been  made  by  the  author  to  determine  whether  two- 
and  three-wire  circuits  in  an  iron  pipe  could  result  in  any  appre- 
ciable increase  of  tlie  impedance  of  the  circuit.  Two  So.  6  B.  &  S. 
wires  were  separated  the  maximum  distance  permitted  by  the  inte- 
rior diameter  of  a  li/^-inch  conduit,  being  rigidly  held  in  position 
by  strapping  to  opposite  sides  of  a  strip  of  wood.  At  60  cycles,  and 
for  currents  between  40  and  80  amperes,  there  was  an  average 
increase  in  the  impedance  of  the  circuit  over  the  value  when  the 
circuit  was  in  air  and  not  surrounded  by  conduit  of  2^  per  cent. 

>y  Google 


It  is  obvious,  therefore,  that  in  the  moderate  lengths  aBoally  met 
with  in  interior  illumination,  this  introduceB  no  disturUng  factor. 

The  lighting  of  interiors  by  are  lamps  fed  from  series  circuits 
is  to  be  avoided.  As  already  mentioned,  these  circuits  operate  at 
high  voltage,  and  special  precautions  must  be  taken  in  insulating 
any  such  circuit  within  a  building.  In  most  localities  the  intro- 
duction of  such  circuits  into  buildings  is  prohibited. 

Multiple-connected  arc  lamps  are  frequently  used  for  the  lighting 
of  stores,  factories,  sheds,  etc.,  and  they  are  supplied  by  low- 
voltage  distributing  mains.  In  such  circumstances  due  considers- 
tion  must  be  given  to  the  regulation  of  these  circuits  if  incan- 
descent lamps  are  also  to  be  operated  from  them.  The  arc  lamp 
takes  from  4  to  9  amperes,  and  when  this  is  the  only  type  of  lamp 
on  the  circnit  the  carrying  capacity  is  often  the  determining  factor 
rather  than  any  question  of  regulation.  The  National  Electrical 
Code  prescribes  the  maximimi  values  of  current  which  it  is  per- 
mitted to  cany  on  various  sizes  of  wire.  Each  lamp  or  series  of 
lamps,  in  case  several  are  operated  in  series,  must  be  provided 
with  a  fused  cut-out.  The  general  description  and  rules  covering 
incandescent  wiring,  as  already  described,  apply  also  to  multiple 
arc  circuits,  but  the  underwriters'  requirements  prescribe  certain 
additional  regulations,  which  are  duly  set  forth  in  the  publications 
mentioned  above, 

(b)  Coatrol.  It  is  obvious  that  the  entire  system  of  an  interior 
installation  should  be  under  control.  We  may  define  "  control " 
as  the  possibility  of  individual  and  separate  operation  of  all  lamps, 
and  the  prompt  cutting  out  of  any  portion  of  the  system  which 
may  develop  trouble.  Thus  every  lamp  or  group  of  lamps  should 
be  operated  by  an  accessible  switch,  and  every  branch  circuit  should 
also  be  equipped  with  apparatus  permitting  its  easy  separation 
from  the  remainder  of  the  system.  Indiridual  distributing  centers 
or  the  feeders  supplying  them  should  be  equipped  with  switches. 
In  addition  to  these  essentials  for  manual  operation,  the  whole 
system  must  be  protected  by  fuses  or  automatic  circuit-interrupting 
devices.  It  is  highly  essential  that  the  main  distributing  center, 
the  service  connectionF,  and  all  subsidiary  centers  should  be  in 
well-illuminated  and  readily  accessible  locations. 

In  the  denser  sections  of  a  distributing  system,  the  service  wires 
vrill  usually  be  brought  in  from  underground.  Connection  to  a 
residence  is  usually  made  from  a  manhole  permitting  access  to 

>y  Google 

Genekation  and  Distribution  op  Electricity        2i)5 

the  underground  network.  The  manhole  is  an  essential  part  of  a 
system  of  underground  ducts.  The  building  connection  ia  usually 
made  from  these  manholes  by  small  conduit  connection,  this  con- 
duit being  made  either  of  fiber,  treated  wood,  terra  cotta  or  any  of 
the  many  types  offered  by  the  market.  These  conduits  are  brought 
through  the  building  line  underground,  and  the  service  wires 
brought  above  the  surface  by  a  continuation  of  the  conduit  or  in 
iron  pipe.  These  conduits  should  drain  back  to  the  manhole,  that 
is,  away  from  the  house,  and  after  the  wires  are  drawn  in  the 
conduit  opening  should  be  stopped  so  as  to  prevent  gases  from 
flowing  into  the  building. 

In  the  outlying  districts  where  the  distribution  is  overhead 
various  methods  are  used  for  bringing  the  service  wires  inside 
buildings.  In  many  instances  this  is  done  by  putting  suitable 
bushings  through  the  walls  near  the  roof  of  the  house.  The  beet 
practice,  however,  takes  the  service  wires  from  the  transformer  into 
an  iron  pipe  some  distance  above  ground  level,  the  pipe  leading 
below  ground  into  the  basement  as  already  described.  This  pipe 
connection  should  be  provided  at  the  top  with  a  rain-proof  bushing, 
and  is  particularly  desirable  in  localities  where  there  is  a  possi- 
bility of  future  underground  service.  The  report  of  the  committee 
on  overhead  construction  of  the  National  Electric  Light  Associa- 
tion, 1910,  describes  in  detail  various  methods  of  making  service 

Interior-lighting  systems,  whether  supplied  from  isolated  plants 
or  from  public-service  companies,  should  be  equipped  with  a  main 
switch  controlling  the  entire  system.  Also  each  feeder  should  be 
equipped  with  a  switch.  The  next  subdivision  at  the  distributing 
centers  should  provide  either  a  switch  or  enclosed  fuse  for  each 
two-wire  branch.  The  main  switch  of  the  system,  and  the  indi- 
vidual feeder  switches,  should  «ach  be  equipped  with  fuses  or  sup- 
plemented by  some  form  of  automatic  circuit-interrupting  device. 
It  is  sometimes  desirable  to  have  a  switch  at  the  distributing 
center,  although  this  is  not  necessary  if  the  feeder  furnishing  this 
center  is  so  equipped.  Branch  circuits  must  be  equipped  with 
fuses,  but  not  necessarily  with  switches.  The  underwriters'  re- 
quirements limit  the  capacity  of  a  single  circuit  from  a  distributing 
center  to  660  watts.  This  figure  was  probably  originally  based  on 
the  demand  of  twelve  55-watt  carbon  lamps.  And,  in  general, 
branch  circuits  in  the  past  have  been  limited  to  10  or  12  ontlets. 

>y  Google 

S66  Illdicinatihb  Ehgineebinq 

It  is  Dov  possible  to  run  many  more  outlets  to  a  branch  circuit 
by  the  use  of  loT-power  tungsten  lamps.  The  branch  circuits  for 
incandescent  lighting  are  usually  protected  by  fusee  of  10-ampere 
capacity.  These  fuses  are  either  of  Edison  "screw-plug"  or  of 
"  cartridge  "  type,  with  present  tendency  to  a  return  to  the  former. 
As  already  stated,  the  feeders  must  be  protected  by  fusee,  and  for 
this  purpose  the  "  cartridge  "  fuse  is  best.  In  many  large  instal- 
lations the  feeders  are  protected  by  circuit  breakers  located  on 
switchboards  of  more  or  less  elaborate  design.  The  requirements 
of  theaters  lead  to  especially  detailed  switching  and  regulating  de- 
vicee.  Fusee  are  manufactured  up  to  500-  and  600-ampere  ca- 
pacity, but  circuit  breakers  are  preferable  above  the  former  figure 
on  account  of  the  cost  of  the  fuses  and  of  the  time  required  for 
their  operation.  Flexible  cable  must  be  used  in  all  conduit  instal- 
lations, and  may  be  had  to  accommodate  practically  any  current. 
In  the  larger  installations  feeders  frequently  have  a  cross-section 
of  500,000  circular  mils,  and  in  extreme  instances  are  even  of 
greater  size.  The  subdivision  in  these  cases  of  the  total  capacity 
required  is  highly  advisable  on  the  score  of  convenience  of  instal- 
lation. The  installation  of  conduit  of  diameter  larger  than  2 
inches  will  usually  involve  difficulties  unless  special  provision  is 
made  in  the  design  of  the  building.  Two-inch  conduit  will  accom- 
modate three  No.  00  wires;  3-  and  4-inch  conduit  has  been  used, 
but  2  inches  marks  the  limit  for  convenient  installation.  The 
neutral  wire  ie  made  of  full  size  in  all  interior  wiring,  so  that 
when  for  any  reason  one  side  of  the  circuit  is  interrupted  the 
neutral  will  provide  full  carrying  capacity  for  the  return  current, 
(c)  Cost  of  lEterior  Wirii^.  Since  the  prices  of  labor  and  ma- 
terial differ  in  different  localities  and  at  different  times,  it  is 
difficult  to  state  even  approximately  what  the  cost  of  distributing 
systems  for  lighting  should  be.  In  large  cities,  however,  these 
variations  are  not  very  wide,  and  it  is  possible  to  state  the  limits 
within  which  the  cost,  expressed  in  terms  of  the  usual  contractor's 
price  per  outlet,  should  lie.  The  figures  given  below  apply  to 
interior  wiring  of  all  classes,  from  the  small  residence  np  to  the 
large  hotel  or  office  building.  They  cover  the  portion  of  the  work 
from  the  main  source  of  supply,  assumed  to  be  at  the  building 
line.  In  case  the  building  is  lighted  from  its  own  plant  these 
figures  will  apply  to  the  portion  of  the  installation  lying  between 

>y  Google 

Gbnehation  and  Disteibution  op  Electbicitt        267 

the  lamp  and  the  plant  switchboard.  No  lamps,  fixtures  or  re- 
fiectore  are  inclnded  in  these  pricee : 

Exposed  wiring,  $1.50  to  $1.60  per  oatlet. 

Wire  in  wooden  moulding,  $2.00  to  $2.50  per  outlet. 

Concealed  knob  and  tube  wiring,  $3.50  to  $3.00  per  outlet,  with 
$1.00  added  per  switch  outlet. 

Wiring  in  iron  conduit  and  in  new  bnildings,  $4,50  to  $5,00  per 

Wiring  in  iron  conduits  in  concrete  buildings,  $5.00  to  $6.00 
per  outlet. 

In  the  above,  switches  and  base-board  plugs  are  considered  as 
outlets  when  the  iron  box  is  inclnded.  If  the  switch  and  plate  is 
also  to  be  furnished,  approximately  $1,00  per  outlet  of  this  nature 
should  be  added.  For  the  larger  InatallationB  in  modern  buildings 
the  price  of  $7.00  per  outlet,  including  all  wiring  and  feeders 
up  to  the  lighting  fixture,  has  been  found  to  be  a  fairly  dose 

For  that  portion  of  the  wiring  which  may  be  necessary  beyond 
the  building  line,  as,  for  instance,  the  service  connection  and 
transformers,  in  those  regions  where  alternating  service  is  sup- 
plied, it  is  hardly  possible  to  state  even  approximate  figures  of 
what  the  prices  will  be.  The  cost  of  wire  follows  that  of  copper 
more  or  leas  closely,  and  transformers  vary  somewhat  in  price. 
Lighting  transformers  suitable  for  erection  on  poles  and  for  60- 
cyele  operation  may  be  had  in  any  capacity  between  6/10  kw,  and 
50  kw.  As  adapted  to  1100  or  2300  primary  circuits,  and  trane- 
forming  to  110  or  380  two-  or  three-wire  secondary,  their  price 
varies  from  $27.00  per  kilowatt  for  the  1-kw,  size  to  between  $7.00 
and  $8.00  per  kilowatt  for  sizes  in  the  neighborhood  of  40  kw.  and 
50  kw.  The  prices  are  somewhat  higher  for  higher  primary  volt- 
ages, and  transformers  adapted  to  location  in  subways  are  from 
10  to  12  per  cent  more  expensive  than  the  usual  out-of-door  type. 
Transformers  for  25  cycles  cost  from  40  to  50  per  cent  more  than 
those  for  60  cycles. 

(d)  Fire  and  Lunranoe  Control.  The  N^atiooat  Board  of  Fire 
Underwriters,  and  in  most  places  municipal  regulations,  require 
strict  supervision  of  the  installation  of  electric  wiring.  It  is  usually 
required  that  the  electrical  contractor  shall  secure  a  permit  for 
any  new  work  or  repairs  to  electric  wiring  in  buildings.  In  many 
cases  the  fire  underwriters  are  satisfied  with  the  municipal  super- 

>y  Google 


TiGiou  and  make  no  independent  demands  of  their  own.  This  ie 
especially  the  eaae  where  the  city  adopts  the  National  Electrical 
Code  for  its  own  regulations.  Presumably  this  permit  for  wiring 
is  followed  up  by  an  inspection  of  the  work  after  completion,  by 
a  city  official.  Too  often,  however,  this  inspection  is  of  the  most 
perfunctory  character.  The  inspector  will  almost  invariably  be 
content  with  a  visual  inspection  of  the  installation.  From  the 
nature  of  the  troubles  and  imperfections  that  are  likely  to  arise 
from  a  system  of  wiring,  electrical  teats  are  the  only  ones  which 
can  yield  complete  evidence  as  to  the  state  and  the  character  of 
the  work.  Insurance  and  city  authorities  therefore  wonld  do  well 
to  require  a  thorough  testing  of  every  installation  before  approval 
and  acceptance.  Since  there  is  at  present  no  municipal  regulation 
which  ensures  tests  of  this  nature,  the  designing  engineer  should 
be  careful  to  incorporate  in  his  specification  clauses  requiring  the 
complete  testing  by  the  contractor.  This  method  of  accomplishing 
the  testing  should  be  easily  available  to  the  city,  which  in  yielding 
a  permit  could  stipulate  that  before  acceptance  proper  tests  should 
be  made  in  the  presence  of  the  city  official. 

It  has  been  already  mentioned  that  the  entire  system  of  metal 
conduit  of  an  interior  installation  should  be  grounded.  Grounding 
means  connecting  as  definitely  and  permanently  as  possible  to  the 
earth,  thus  maintaining  the  grounded  portion  at  the  potential  of 
the  earth.  The  neutral  of  underground  direct-current  systems  is 
almost  invariably  grounded.  Interior-wiring  systems  should,  in 
the  writer's  opinion,  be  always  grounded.  Ground  connections  may 
be  readily  made  by  connecting  between  the  grounding  point  of  the 
circuit  and  the  metal  pipes  of  the  city  water  supply.  Such  connec- 
tions should  he  soldered  and  of  fairly  large  size  of  wire.  To  en- 
sure a  ground  independent  of  water  or  gas  pipes  an  iron  pipe  may 
be  driven  5  or  6  feet  into  solid  soil,  the  damper  the  soil  the  better, 
and  the  ground  connection  soldered  to  this  pipe.  The  conditions 
will  be  improved  by  using  several  pipes  and  by  removing  the  earth 
from  around  the  top  of  the  pipe  to  a  depth  and  diameter  of  about 
1  foot  each,  and  then  filling  this  hole  with  salt. 

There  has  been  a  wide  discussion  as  to  the  advisability  of  ground- 
ing alternating-current  secondary  circuits.  These  circuits  are 
usually  three-wire,  and  the  ground  connection  should  obviously 
be  taken  from  the  neutral.  The  great  advantage  of  grounding  the 
neutral  is  in  the  fact  that  should  the  primary  voltage  reach  the 

>y  Google 

Generation  and  Distbibction  op  Electricity        269 

Becondary  wiring  by  the  failure  of  a  transformeT  or  by  the  croeaing 
of  the  respective  lines,  the  high-voltage  circnit  thus  brought  into 
connection  with  the  low-voltage  wiring  would  be  grounded  and 
thuB  prevent  arcing  and  danger  to  life.  In  many  instances,  also, 
electrostatic  charges  may  be  induced  in  the  secondary  wiring  by 
disturbances  in  the  primary  circnit.  This  may  result  in  serious 
shock  to  persons  handling  the  secondary  circuits  if  these  circuits 
are  not  grounded. 

The  supposed  objection  to  grounding  such  circuits  is  that  it 
places  the  potential  of  one  side  of  the  three-wire  system  between 
the  bare  contacts  on  lamps  and  other  devices  and  the  ground,  thus 
offering  the  possibility  that  persons  receive  shocks.  The  National 
Electric  Light  Association  recommends  that  the  grounding  of  sec- 
ondary circuits  be  limited  to  those  on  which  the  voltage  of  one  aide 
does  not  e:(ceed  150.  This  means  that  no  shock  of  a  higher  value 
than  that  stated  could  be  received  by  anyone  touching  an  unin- 
sulated portion  of  the  circuit.  The  reasons  for  not  grounding  cir- 
cuits of  higher  potential  do  not  appear  to  be  good.  There  can  be 
no  question  that  the  grounding  of  the  circuit  offers  great  pro- 
tection from  any  trouble  that  may  arise  from  the  primary  circuit. 
This  is  undoubtedly  the  most  likely  and  the  most  serious  source 
from  which  trouble  may  come.  The  danger  of  shock  to  persons  is 
hardly  greater  when  the  system  is  grounded  than  when  it  is  not, 
and  in  those  systems  in  which  the  voltage  is  carried  to  vRlnee  dan- 
gerous to  life  it  would  appear  desirable  to  provide  the  safeguards 
in  other  ways,  such  as  complete  insulation  of  all  live  contacts,  or 
by  other  methods  usual  in  high-voltage  circuits. 

(e)  Specificatioiu  and  Contraots.  In  preparing  speciflcatioDa 
and  making  contracts  for  an  installation  it  is  highly  desirable  that 
each  should  be  as  complete  and  explicit  as  it  is  possible  to  make 
them.  The  specifications  should  always  be  accompanied  by  draw- 
ings. Of  the  numerous  clauses  for  the  protection  of  the  client 
which  should  be  inserted,  none  perhaps  is  more  important  than 
that  applying  to  the  charges  for  alterations  or  extensions  of  the 
work,  as  set  down  in  the  specification.  In  competitive  bidding 
on  work  of  this  nature  a  contractor  will  often  look  to  his  charges 
for  extras  and  alterations  for  the  best  part  of  his  profit.  The 
engineer  should  therefore  endeavor  to  describe  on  the  drawings  or 
by  explicit  statement  every  outlet  of  installation.  Qeneral  clauses 
should  be  inserted  which  shall  protect  the  client  during  the  process 

>y  Google 


of  the  work  from  dama^  to  peiBons  and  property,  and  relieve 
him  from  all  reBponsibility  until  the  installation  is  ready  to  be 
turned  over  complete.  In  large  inetallatiotiB  the  contractor  should 
be  required  to  place  insurance  on  completed  portions  of  the  vork 
and  to  give  bond  for  its  completion  within  the  date  stipulated  in 
the  contract.  The  specifications  should  cover  carefully  the  si2M 
of  all  mains,  feeders  and  branches,  together  with  the  conduit  io 
which  they  are  placed.  Full  details  should  be  gives  of  all  switchea, 
distributing  boards,  panels,  etc.  The  trade  names  of  manufactured 
articles  which  will  be  accepted  should  also  be  given,  and  the  gen- 
eral statement  made  that  no  material  not  approved  by  the  Board 
of  Eire  Underwriters  may  be  used.  The  drawings  shoidd  show  the 
accurate  location  of  all  outlets,  service  connections,  distributing 
centers  and  the  run  of  all  feeders.  It  is  highly  desirable  that  the 
engineer  and  architect  should  have  early  consultation  so  that  the 
latter  may  know  what  space  will  be  required  by  the  engineer.  Too 
often  the  architect's  plans  are  completed  before  the  engineer  sees 
them.  The  architect,  as  a  general  thing,  has  a  very  limited  knowl- 
edge of  the  requirements  of  an  electric-wiriog  installation,  and  it 
ie  usually  assumed  that  the  illuminating  engineer  requires  no  space 
at  all  for  his  circuits.  This  consultetion  is  especially  advisable  for 
buildings  of  reinforced  concrete  where  it  is  inadvisable  te  pass 
conduit  through  reinforced  beama. 

The  drawings  should  also  indicate  the  type  of  fixture,  lamp, 
reflector,  mounting  height,  etc.  The  National  Electrical  Con- 
tractors' Association  has  published  a  set  of  symbols  which  are  in 
general  use  for  indicating  the  nature  and  location  of  distributing 
centers  and  the  various  types  of  outlet,  ete.  A  standard  set  of 
symbols  of  this  nature  applying  to  the  difiFerent  methods  of  mount- 
ing lighting  units  and  describing  their  character  would  be  very 

The  wiring  for  lighting  systems  is  often  installed  on  what  is 
known  as  the  time  and  material  basis.  This  means  that  the  con- 
tractor charges  the  cost  of  material  used  and  the  hours  of  labor 
required  to  the  owner,  with  a  certain  percentage  added.  This  is 
rarely,  if  ever,  a  satisfactory  method  to  the  owner.  To  ensure  a 
reasonable  charge  it  requires  a  constant  inspection  of  material  and 
labor-time.  It  will  usually  be  possible  to  secure  competitive  bids, 
and  then  require  the  contractor  to  give  the  owner  the  benefit  of 
any  saving  under  the  contracted  figure  which  results  from  keeping 

>y  Google 

Generation  and  Distribdtion  of  Electhicitt        271 

a  record  on  the  time  and  material  basis.  In  euch  a  case  the  con- 
tractor furnishes  the  engineer  with  a  statement  of  material  and 
labor  time  at  regnlar  intervale. 

(f )  Teati.  Satisfactory  performance  in  wiring  installations  de- 
pends primarily  on  regulation  and  on  the  nature  of  the  material 
and  workmanahip.  The  regnlatton  will  depend  largely  on  the 
sizes  of  conductor  specified  by  the  engineer,  and  a  test  of  r^ulation 
will  only  check  up  the  methods  which  have  been  employed  in 
making  joints  and  contacts.  A  full-load  test,  however,  ahould  be 
invariably  applied  to  the  system  before  its  acceptance.  Every 
switch  should  be  operated  and  each  lamp  socket  and  base-board 
plug  tested.  Insulation  tests  are  rarely  applied  to  interior-wiring 
syetems.  It  is  advisable,  however,  to  apply  at  least  double  the 
normal  operating  voltage  to  the  completed  system.  A  stipulation 
to  this  effect  should  be  included  in  the  specification.  The  con- 
tract should  contain  a  clause  requiring  the  contractor  to  cany  out 
the  teste  in  the  presence  of  the  engineer  and  the  details  of  this 
test  should  be  given. 

S.  Exterior  lUumination 
(a)  The  commonest  form  of  outside-lighting  circuit  is  that  of 
the  series  incande^ent  or  arc  system.  These  circuits  are  usually 
run  overhead,  except  in  the  more  densely  populated  portions  of  the 
city.  N^o  special  comment,  therefore,  seems  needed  as  to  the  instal- 
lation beyond  the  regulations  set  down  by  the  National  Electrical 
Code.  These  circuits  are  of  moderate  voltage  (from  2000  to  8000), 
and  may  therefore  be  handled  by  a  variety  of  approved  grades  of 
manufactured  wire,  insulators,  etc.  Series  circuits  are  controlled, 
as  a  whole,  from  a  generating  station  or  substation,  the  entire 
protective  apparatus  being  installed  there.  Special  precautions 
may  be  necessary  in  some  places  for  the  protection  of  low-voltage 
lighting  circuits  and  of  telephone  and  telegraph  wires.  This  clasa 
of  service  is  more  satisfactory  when  run  in  underground  conduits, 
and  this  is  usually  required  by  the  authorities  in  the  centers  of 
large  cities.  The  cities  usually  own  the  conduit  system  and  rent 
space  to  the  supply  companies.  The  single  conductor  of  the  arc 
or  incandescent  circuit  ie  insulated  with  rubber  or  paper  and  the 
whole  covered  with  lead.  The  manholes  of  the  duct  system  are 
usually  from  400  to  600  feet  apart,  and  individual  lamps  are  fed 
through  branch  conduits  between  the  manhole  and  the  base  of  tha 

>y  Google 


pole.  The  cables  then  rise  inside  the  iron  pole  to  the  lamp.  Since 
there  is  little  or  no  difFerence  in  potential  between  the  two  sides 
of  such  a  loop  from  a  manhole  to  a  lamp,  a  duplex  conductor  may 
safely  be  used  for  this  portion  of  the  circuit.  The  lamp  itself, 
however,  should  be  insulated  from  its  support,  since  it  may  receive 
the  full  potential  of  the  circuit.  Grounds  on  this  class  of  circuit 
are  very  dangerous.  The  lead  sheathing  of  underground  cables 
usually  affords  sufficient  protection  between  mains  of  different 
classes  of  service;  thus  arc  circuits  are  frequently  run  in  the  same 
duct  with  the  low-potential  multiple-distribution  mains.  Instances 
have  been'  known  in  which  trouble  has  arisen  by  reason  of  this 
proximity,  but  a  rental  charge  on  the  part  of  a  city  of  5  cents 
per  duct  foot  per  annum  is  usually  sufficient  to  cause  the  supply 
company  to  pat  as  many  conductors  as  possible  in  one  duct.  Ex- 
cellent data  as  to  the  construction  of  conduits,  their  coat,  etc.,  may 
be  found  in  the  Standard  Hand-Book  (or  Electrical  Engineers. 

S.  Cost  of  Operation 
There  is  probably  no  phase  of  the  general  problem  of  electric 
lighting  which  attracta  more  public  discussion  than  that  of  its 
cost.  Public-service  corporations,  particularly  if  they  have  a 
monopoly  of  the  consuming  market,  are  naturally  the  objects  of 
public  suspicion.  This  is  especially  true  of  companies  selling  elec- 
tricity for  lighting,  and  the  explanation  is  to  be  found  in  the  great 
discrepancy  always  existing  between  the  admitted  cost  of  electrical 
energy  at  the  station  bus-bars  and  the  price  at  which  it  is  sold  to 
the  consumer.  The  latter  figure  is  often  ten  or  more  times  as 
great  as  the  former,  and  consequently  is  often  the  object  of  unin- 
formed public  clamor.  The  reasons  for  the  difference  will  be  better 
understood  after  a  discussion  of  some  of  the  factors  entering  into 
the  actual  cost  of  generating  and  delivering  electric  power. 

(a)  Cost  of  Electric  Power.  The  commonest  basis  of  estimating 
the  cost  of  electric  power  is  the  summation  of  all  expenditure  nec- 
essary to  deliver  the  power  at  the  station  feeder  bus-bars  ready  for 
distribution.  This  total  cost  divided  by  the  total  energy  generated 
gives  the  unit  cost,  i.  e.,  the  coat  per  kilowatt  hour.  This  apparently 
simple  method,  however,  will  rarely  yield  the  same  figure  for  two 
different  months,  or  weeks,  or  even  days  in  the  year,  for  the  total 
cost  of  electric  power  is  not  directly  proportional  to  the  amount 

>y  Google 

Geneeation  and  Distbibdtion  of  Electricity        273 

The  total  cost  may  be  divided  into  two  claBBee:  (1)  fixed  charges 
and  (2)  operating  expenses.  In  the  item  fixed  charges  are  in- 
cluded all  expenditures  necesaary  whether  or  not  the  plant  gen- 
erates power.  Thus  in  this  class  fall  the  items  of  interest,  tases, 
insurance,  depreciation  and  obsolescence.  They  represent  the  ag- 
gregate cost  of  having  an  up-to-date  power  station  ready  to  deliver 
power.  By  depreciation  is  meant  the  outlay  necessary  to  keep  all 
generating  equipment  in  repair,  and  to  replace  efficient  apparatus 
worn  out  in  service.  By  obsolescence  is  meant  the  cost  of  pur- 
chasing apparatus  and  equipment  to  replace  that  which  has  been 
rendered  obsolete  and  inelHcient  by  improvements  and  increased 
knowledge  of  the  art.  Interest  and  taxes  expressed  in  per  cent  of 
the  cost  of  the  plant  will  not  vary  with  the  type  of  plant;  insur- 
ance is  often  eliminated  entirely  in  modem  plants  of  fire-proof 
construction ;  depreciation  and  obsolescence  vary  widely  with  the 
type  and  size  of  plant,  being  greatest  for  reciprocating  steam  plants 
and  least  for  water-power  plants.  The  aggregate  of  fixed  charges, 
in  per  cent  of  the  cost  of  the  plant,  varies  from  9  to  17  per  cent  in 
modern  plants  of  size  required  to  furnish  city  lighting  service. 
The  lower  figure  is  reached  only  in  the  best  type  of  water-power 
plant,  and  the  upper  refers  to  reciprocating  steam  engines  oper- 
ating under  poor  conditions.  The  cost  of  the  power  plant  varies 
from  S80  per  kilowatt  of  installed  capacity,  in  the  case  of  steam 
turbines,  to  $100  or  $125  for  reciprocating  steam  engines,  and  to 
$200  or  more  for  water-power  plants.  Large  gas-engine  plants 
cost  about  $135  per  kilowatt  of  installed  capacity. 

The  second  class  of  expense  in  the  production  of  power  is  called 
the  operating  expense,  and  it  includes  all  items,  such  as  fuel,  oil, 
attendance,  etc.,  which  are  approximately  proportional  to  the 
amount  of  power  generated.  The  proportionality  between  total 
operating  expenses  and  amount  of  power  generated  is  not  exact, 
since  the  efficiency  of  steam  and  electrical  apparatus  is  not  the 
same  for  all  values  of  the  load  upon  them.  With  proper  sub- 
division of  the  total  capacity  into  smaller  units,  however,  it  is 
usually  possible  to  operate  with  machines  loaded  to  more  than  50 
per  cent  of  their  rated  capacity,  and  in  such  conditions  the  oper- 
ating expenses  per  kilowatt  hour  are  approximately  uniform  at 
all  times.  Average  values  of  operating  expenses  in  large  stations 
are  .3  cent  per  kilowatt  hour  for  gas-engine  plants,  A  to  .5  cent  for 
steam-turbine,  and  .6  cent  for  reciprocating-engine  plants. 

>y  Google 


It  18  obvious  that  since  the  fixed  charges  are  constant  and  the 
operating  expenses  proportional  to  the  amount  of  power  geiierated, 
the  cost  per  Idlo^ratt  hour  will  be  least  when  the  station  is  gen- 
erating its  greatest  output.  The  minimum  possible  cost  would  be 
reached  if  the  station  could  operate  continuously  at  its  maximum 
capacity.  In  Buch  a  case,  at  12  per  cent  fixed  charges,  an  up-to-date 
steam-turbine  plant  coilld  generate  poWer  at  the  feeder  terminals 
at  approximately  .5  cent  per  kilowatt  hour.  Unfortunately,  how- 
ever, the  maximum  load  on  the  usual  central  station  lasts  a  very 
short  time,  the  load  curve  having  a  sharp  peak  in  the  late  afternoon 
and  early  evening  hours.  The  value  of  the  maximum  power  output 
ehown  by  this  peak  determines  the  capacity  required  at  the  central 
station.  Consequently,  at  periods  of  light  load,  as  for  instance, 
during  the  morning  hours,  fixed  charges  must  he  paid  on  more 
generating  equipment  than  are  required  to  handle  the  load.  This 
variation  of  thp  load  throughout  the  day,  in  its  effect  on  the  cost 
of  power,  is  described  in  terms  of  a  quantity  known  as  the  "  load 
factor,"  which  is  the  ratio  of  the  average  daily,  monthly  or  yearly 
load  to  the  maximum  loads  occurring  in  the  corresponding  inter- 
vals. The  daily  load  factor  then  is  a  quantity  less  than  1,  and 
represents  the  proportion  of  the  maximum  daily  power  output 
which  may  be  multiplied  by  34  in  order  to  arrive  at  the  total  num- 
ber of  kilowatt  hours  generated  through  the  day.  It  is  therefore 
highly  desirable  to  increase  the  average  daily  load,  and  so  render 
the  load  factor  as  near  to  the  value  1  as  possible.  The  load  factor 
corresponding  to  lighting  service  only  is  very  low,  and  lighting 
companies  make  greet  efforts  to  develop  a  day  load  comprising 
motors  of  all  kinds,  and  heating,  cooking  and  other  domestic  ap- 
pliances. The  daily  load  factor  of  a  large  central  etation,  which 
supplements  its  lighting  load  in  every  way  possible,  is  about  .50; 
the  yearly  load  factor  about  .30.  At  load  factor  .50  the  average 
total  cost  of  generation  in  a  gas-engine  plant  is  .65  cent,  in  a 
steam-turbine  plant  .7  cent,  and  in  a  steam-engine  plant  about 
.9  cent  per  kilowatt  hour. 

(b)  Syatemi  of  Bates  for  Sale  of  Power.  In  the  early  days  of 
electric  lighting  it  was  customary  to  charge  a  consumer  simply  in 
terms  of  the  number  of  lamps  installed  without  reference  to  the 
number  of  hours  they  were  used.  This  method  of  charging,  known 
as  the  fiat-rate  system,  was  ohriously  unfair  to  the  economical  user, 
and  meters  for  reading  the  total  number  of  kilowatt  hours  were 

>y  Google 

Generation  and  Distribution  of  Electhicity        375 

developed  ae  a  basis  for  charging.  This  method  aloBe,  however, 
is  obviously  not  equitable,  since  it  costs  the  supply  company  more 
to  supply  a  consumer  during  the  time  of  peak  load  than  at  other 
times.  Consequently,  consumers  are  often  classified  on  some  basis 
representing  the  times  of  the  day  during  which  they  take  their 
maximum  power,  and  different  rates  apply  to  the  several  classes. 
Such  a  classification  might  wparate,  for  instance,  the  services  to 
residences,  to  stores  or  factories,  and  to  day  motors.  A  further 
refinement  in  the  methods  of  charging  is  found  in  the  so-called 
two-rate  systems,  which  aim  to  charge  a  consumer  a  higher  rat« 
for  the  power  he  uses  during  his  peak  hours  and  a  lower  rate 'for 
the  remainder.  This  method  evidently  aims  to  charge  each  con- 
sumer his  proportionate  share  of  the  fixed  and  operating  charges, 
respectively.  The  obvious  difiBculty  is  that  of  ascertaining  the 
maximum  load  of  each  consumer.  For  residence  lighting  it  is 
usually  assumed  that  some  proportion  of  the  total  number  of  lamps 
connected  will  be  burned  together  for  a  definite  number  of  hours 
each  day.  This  number  of  kilowatt  hours  will  then  be  charged 
for  at  the  higher  rate,  and  all  power  in  excess  at  some  lower  rate. 
Maximum-demand  meters,  which  indicate  the  highest  value  of 
power  taken  during  any  chosen  interval,  have  also  been  used  as  a 
means  of  arriving  at  the  value  of  a  consumer's  peak.  This,  how- 
ever, constitutes  a  separate  measuring  instrument  for  each  con- 
sumer, and  on  account  of  the  expense  involved  the  plan  has  not 
as  yet  been  widely  adopted. 

The  actual  price  at  which  power  for  lighting  is  sold  varies  widely 
in  different  places.  In  the  larger  cities  the  primary  rate  is  rarely 
less  than  10  cents  per  kilowatt  hour,  which  may  be  charged,  for 
instance,  for  all  power  up  to  the  amount  consumed  bj  one-half 
the  connected  load  if  burned  for  30  hours.  All  power  in  excess  of 
this  during  the  month  would  then  be  charged  for  at  a  less  rate, 
say  7  or  5  cents  per  kilowatt  hour.  One  reading  per  month  of  a 
meter  indicating  kilowatt  hours,  therefore,  serves  to  fix  the  amount 
of  the  consumer's  bill. 

The  wide  discrepancy  between  the  prices  at  which  power  is  sold 
and  the  cost  of  its  generation  have  led  to  frequent  agitation  by  the 
public  of  the  question  of  regulating  the  rates  for  the  sale  of  power 
by  law.  This  type  of  discussion  arising  as  well  in  connectiozi  with 
other  dasees  of  public-service  corporations  has  led  to  the  forma- 
tion !n  many  states  of  public-utilities  commissions,  which  have 

>y  Google 

S?6  Illuminatikg  Ekoimeehixo 

the  power  to  investigate  and  regulate  the  conditions  of  manufac- 
ture and  sale  of  the  respective  public  commodities.  The  figures 
of  cost  of  generating  power  which  have  been  given  apply  at  the 
station  bus-bars.  The  discrepancy  alluded  to  above  includes  the 
cost  to  the  supply  company  of  distributing  the  power  to  the  con- 
sumers, the  cost  of  meters  and  their  regular  inspection,  and  the 
general  office  expenses.  While  the  cost  of  distribution,  which  in- 
cludes the- capital  charges  on  all  the  distributing  system,  as  well 
as  its  inspection  and  maintenance,  duct  rentals,  etc.,  is  usually  a 
much  larger  figure  than  at  first  apparent,  the  several  items  men- 
tioned do  not  bring  the  actual  cost  of  delivering  the  power  to  the 
consumer  very  near  to  the  figure  at  which  it  is  sold.  The  remaining 
difl^erence  is  not  all  profit  to  the  company,  however,  but  ie  in  part 
applied  to  paying  the  obligations  of  early  lighting  companies, 
bought  up  by  the  present  one,  and  defunct  through  obsolescence  or 
other  cause.  It  is  worth  noting  that  a  recent  careful  investigation 
by  a  public-utilities  commission  of  the  rates  charged  by  a  lighting 
company  in  a  large  city  in  the  middle  West  resulted  in  a  decision 
that  14  cents  and  8  cents  per  kilowatt  hour  were  equitable  primary 
and  secondary  rates, 


F.  Koeater:     Steam  Electric  Power  Plants. 

Prankhn  and  Esty:     Elements  of  ElectrlcaJ  Engineering. 

C.  W.  Stone;     Modem  Llgbtlng  Systems.    Proc.  A.  I.  B.  E.,  June,  1910. 

Sheldon  and  Hausmann:     Dynamo  Electric  Machinery. 

C.  P.  Stelnmetz:     General  Lectures  on  Electrical  Engineering. 

H.  G.  Stott:     Coat  of  Power.    Trans.  A.  I.  E.  E..  XXVIII.  p.  1479,  1909. 

H.  B.  Gear:    Diversity  Factor.    Proc  A.  I.  B.  E.,  Aug..  1910. 

Reports  to  National  electric  Light  Association. 

H.  Foster:     Electrical  Engineers  Pocket-Book. 

Standard  Hand-Booli  tor  Electrical  Engineers. 

>y  Google 

vn  (1) 




By    E.    a.    COWDBRT 


General  cliaractertiitlcs  of  coal  gaa  and  water  gas. 

Effect  of  different  conetituenta  on  the  caloriDc  value  and  lllumlnBt- 

Ing  power  of  coal  gas. 
Illumlnants,  their  characterletica. 
Manufacture  of  coal  gas. 

Open  furnace  heating  of  bencbee. 
Regenerative  furnace  beating  of  benches. 

Development  In  the  retort  under  varying  beats  and  conclltlons. 
Brief  references  to  through  retorts. 
Brief  reference  to  vertical  retorts. 
Brief  reference  to  Inclined  retorts. 
Brief  reference  to  by-product  coke  oven  process. 
Purification  of  coal  gas. 
Tar  extraction. 

Ammonia  extraction. 
Sulphur  extraction. 
Carburetted  water  gas. 
Oeneral  statements. 

As  made  from  fixed  carbon,  steam  and  oil. 

Harris  process. 
Teasle  du  Motay. 
Lowe  process. 
Treatment  of  different  oils. 
ParafliD  base  oil. 
Seml'paraflln  base  oil. 
Asphalt  base  oil. 
Basic  claim  liowe  patent. 
Efficiency  ol  X^we  apparatus. 
Purification  of  water  gas. 
Carburetted  water  gas  made  from  oil  and  Bteam  only. 
Producer  gas. 

Metering  gas  at  the  manufacturing  station. 
Oas  holders. 

>y  Google 

SI'S  Illcminatino  Enoineebino 

Low  pressure. 
District  holders, 
Relntorclng  pressure  mafni. 

HIgb  pressure  for  Buburb&n  or  long  distance  distribution. 
Seml-blgh  pressure  or  "  Booster  "  srstem. 
Pormula  for  flow  of  gas  tbrougb  pipes. 

Low  pressure. 

High  pressure. 

Bxcesslvelj'  high  pressure. 
LocstlOD  at  gas  works. 
Ststlou  governors. 
Design  of  a  distribution  system. 
Drainage  of  mains. 
Pipe  Joints. 

Brief  mention:  services,  gas  meters,  bouse  piping  and  pbotometry. 

There  are  three  characteristic  ways  in  which  manufactured  gas 
is  used,  each  at  wliich,  in  its  owd  sphere,  results  in  its  ext«Dsive 
employment  as  an  agent  for  the  production  of  artificial  light.  When 
burned  without  previous  mixture  with  air,  it  produces  a  flame  of 
considerable  intrinsic  brilliancy ;  when  burned  after  previous  mix- 
ture with  air,  it  produces  a  non-luminous  flame  of  high  tempera- 
ture; and,  thirdly,  the  application  of  its  explosive,  action,  when 
mixed  with  air  and  ignited  in  the  cylinders  of  gas  engines,  places 
certain  grades  of  artificial  gas  among  the  most  economical  agents 
for  the  production  of  power. 

I  shall  devote  myself  mainly  to  a  description  of  the  principles 
involved  in  tlie  manufacture  and  distribution  of  the  various  gases 
delivered  by  the  artificial-gas  companies  of  the  United  States,  bnt 
attention  is  purposely  called  to  the  use'of  producer  gas  for  the 
production  of  power,  as  being  an  important  modem  moans  towards 
conservation  of  energy,  and  this  phase  of  the  subject  will  be  briefly 

Kinds  of  Gaaei.  The  gases  we  will  consider  are  generally  classi- 
fied as  follows: 

Illuminating  gas  is  divided  into  two  great  classes,  coal  gas  and 
carburetted  water  gas. 

Coal  gas  in  turn  is  divided  into  two  sub-clasaes,  viz.,  that  pro- 
duced by  the  distillation  of  gas  coal  in  comparatively  small  re- 
torts, and  that  produced  by  the  distillation  of  coking  coal  in  larger 

>y  Google 


Carburetted  water  gas,  on  the  other  hand,  ia  dividetl  into  that 
made  from  fixed  carbou,  eteani  and  oil,  and  that  made  from  oil  and 
eteam  only. 

Producer  gas  ie  made  by  the  action  of  stesm  or  air,  or  both, 
upon  fixed  carbon. 

Oenenl  Contiderations.  From  this  clasBiflcation  it  becomes  evi- 
dent that  coal  gas  iB  produced  analytically,  distilled  from  certain 
kinds  of  coal,  while  water  gas  and  producer  gas  are  synthetically 
made,  that  ia,  built  up  from  tlie  action  of  several  conatituenta  upon 
each  other  in  a  manner  to  be  described  later. 

The  reeulte  in  each  case  do  not  widely  differ,  as  is  illastrated 
in  the  table  shown  as  Slide  1.  In  this  connection  it  is  to  be  under- 
stood that  the  analyses  shown  are  representative  only,  and  not  abso- 
lute, under  all  conditions. 

Producer  gas  has  been  omitted  from  this  table,  but  will  be  con- 
sidered later  on. 

It  ie  to  be  noted  that  the  use  of  water  gas  made  from  steam  and 
oil,  owing  to  local  conditions  of  supply  of  the  raw  materials,  is  at 
the  present  time  practically  confined  to  the  Pacific  slope,  where 
it  ia  extendvely  used.  It  is  particularly  interesting  to  note  how 
closely  this  gas  compares  in  composition  with  coal  gas,  although 
produced  from  very  different  materials. 

In  this  country,  generally,  and  in  Europe  and  Great  Britain, 
carburetted  water  gas  is  understood  to  be  the  gas  produced  from 
coal  or  coke,  steam  and  hydrocarbon  oils,  as  shown  in  the  third 
column  of  the  table. 

Table  of  Gehdial  Chabactebistics  op  Coal  anu  Watib  Oas 

Ckwl^uDude  Water  su  made         ^  ^       gj 




Per  cent  by  volume. 

Illumlnants    4.TG 

Ethane    "^ 

Methane  ^^-^^ 

Hydrogeii    47.04 

Carbon  monoxide  . .       S.04 

Carbon  dioxide  1.60 

Oxygen  0.39 

Nitrogen    2.16 

Total  100.00 






4.S      12.8 





3o{  i: 








4S.7       33.9 





4.1      30.9 





1.3        2.8 





0.7        0.6 





8.4        2.8 





100.0    100.0 



Illuminating  Enqikeehinq 

CoBl  Bss  made  Water  gmi  made 

Coke       Steam  died       Staam 
oven*.       C.  and  oil.        and  oil. 


Specific  gravity 42G  .391  .683  .482 

B,  t.  u 678.  67B.  682.  680. 

Candle-power 16.0  15.8  22.  19.69 

Cu.  (t.  air  req'd  for  combnstlOD 

one  cubic  foot 5.65  5.63  6.7*  5.81 

Note. — The  B,  t.  n.  per  cubic  foot  ot  "  lllumlnants  "  Tarlea  consider- 
ably in  different  gasea.  tn  computing  the  amount  ot  air  required  for 
'combustion  the  lllumlnantfi  were  aBBumed  to  have  a  composite  formula 

In  discussing  this  table  it  is  to  be  noted  that  each  gaa  differs 
from  the  other  only  in  the  relative  proportion  of  the  same  con- 
stituents. To  indicate  this  more  clearly,  it  is  seen  that  coal  ga& 
contains  less  iltuminauts,  usually  more  hydrogen,  considerably  lesa 
carbon  monoxide,  and  less  carbon  dioxide  than  water  gas. 

These  characteristic  features  exercise  a  considerable  effect  upon 
the  candle-power,  calorific  value  and  specific  gravity  of  the  gases. 
For  instance,  a  smaller  amount  of  illuminants  means  lower  caudle- 
power,  usually  lower  heat  value  and  higher  specific  gravity.  More 
methane  means  lower  candle-power,  usually  higher  calorific  value, 
but  it  is  of  lesser  specific  gravity  than  the  illuminants. 

An  increased  quantity  of  hydrogen  means  lower  candle-power, 
lower  heating  value  and  very  much  lower  specific  gravity. 

Carbon  monoxide  bums  with  a  blue  flame,  and  in  itself  has  only 
a  relatively  low  calorific  value.  Carbon  dioxide,  being  the  product 
of  the  combustion  of  carbon,  when  present  in  gas,  decreases  the 
candle-power  and  heating  value,  but  increases  the  specific  gravity. 

These  comparisons  are  general  only,  and  give  the  result  of  the 
effect  of  any  one  of  these  constituent  gases,  considered  from  the 
point  of  view  of  such  gas  only,  without  consideration  of  the  effect 
of  other  constituents  at  the  same  time.  For  instance,  it  might 
conceivably  happen  that  an  increase  in  the  proportion  of  methane 
would  be  accompanied  by  such  a  large  decrease  in  the  percentage 
of  carbon  dioxide  and  nitrogen  that  the  candle-power  might  actu- 
ally be  raised.  In  other  words,  it  is  necessary  to  look  at  the  com- 
position of  a  gas  as  a  whole,  in  order  to  arrive  at  a  satisfactory  ide& 
of  the  various  enumerated  properties. 

Dr.  William  B.  Davidson,  of  Birmingham,  England,  in  his  re- 
cent paper  entitled,  "  Experiments  in  Carbonization  on  the  Bir- 

>y  Google 



mingham  Coal-Test  Plant,"  read  before  the  British  Institutiou  of 
Gas  Engineers  in  1910,  gives  some  interesting  data  on  the  effect 
of  these  various  constituents  on  coal  gas.  An  extract  from  the 
same  appears  as  follows: 

"  In  this  connection  it  is  interesting  to  consider  the  effect  of 
each  of  the  main  constituents  of  coal  gas  on  both  the  illuminating 
power  and  the  calorific  value.  On  this  subject,  the  information 
available  in  technical  literature  is  both  incomplete  and  incorrect, 
and  I  have  therefore  undertaken  a  series  of  laboratory  experiments 
with  the  object  of  ascertaining  the  effect  on  candle-power  of  ad- 
mixtures of  small  quantitieg  of  different  gaseous  constituents. 

The  effect  on  calorific  value  is  already  known.  The  approximate 
results  are  given  in  the  following  table,  and  apply  alike  to  No.  8 
and  No.  1  argand  burners  used  with  full  flame. 

Efrct  of  Diffebbnt  ConBTiruBNTs  ok  the  Calobifip  Vjujjb  and  Illuhi— 

NATiNO  PowEE  OF  Coal  Gas  as  a  Basis  or  540  B.  t.  d. 

AND  16  Caudles. 


Per  CeDt. 

—  3.5 

1  to  3.5  decrease 

—  3.0 

1  to  3.0 

—  Z.6 

1  to  2.6 

—  2.7 

1  to  2.7 

—  0.5 

1  to  1.0 

—  0.6 

1  to  1.0 

Increase  In  calorific  value 

-0.6  J 

=twice  the  decrease  In 

Illuminating  power. 

+  10.9 

+  18.0 

1  to  3.0 


1  to  12 

C»H.   +1.9 

C.H.   +6.0 

CoH.    +10,6 

Note.— Qas  saturated  wltb  napbtbalene  vapor  at  60°  F.  contaloa  only 
0.0086  per  cent  by  volume  of  this  constituent.  The  Increase  In  candle- 
power,  due  to  tbU  unall  amount,  is  only  0.16  or  1  per  cent.  It  should 
be  understood  the  per  cent  of  llluminatins  power  given  Is  theoretical 
and  true  only  wltbln  narrow  limits. 

The  figures  for  carbon  dioxide,  oxygen  and  nitrogen  have  been 
confirmed  by  experiments  with  the  large  test  plant.  It  calls  for 
remark,  however,  that  in  short  trials  the  effect  of  the  admission 
of  air  was  not  nearly  so  drastic  as  was  indicated  by  laboratory 
tests.  This  was  doubtless  due  mainly  to  the  fact  that  the  iron 
oxide  underwent  a  large  rise  in  temperature  and  threw  off  certain 

>y  Google 

288  lLi.UMiMi.TiNa  Engikeebing 

hydrocarbonB — chiefly  beDzene — with  which  the  water  Id  the  mate- 
rial had  become  saturated.  In  one  inatanoe,  the  admission  of  3 
per  cent  of  air  appeared  to  efFect  do  reduction  at  all  on  the  multiple. 
In  experimenting  with  air  it  is,  therefore,  necessary  to  allow  the 
plant  to  attain  equilibrium  before  Btarting  the  test,  and  to  prolong 
the  trial. 

It  will  be  observed  that  the  efFect  of  an  admixture  of  1  per  cent 
of  nitrogen  reduces  the  candle-power  by  about  2.6  per  cent.  As 
it  is  this  ingredient  that  Tsries  most  of  all  in  the  composition  of 
coal  gas  as  manufactured  in  this  country,  and  seeing  that  the  effects 
of  carbon  dioxide,  oxygen,  carbon  monoxide  and  benzene  have  all 
nearly  the  same  ratio,  it  follows  from  theoretical  considerations 
that  8  per  cent  reduction  of  illuminating  power  for  1  per  cent 
reduction  of  calorific  value  the  result  previously  indicated  is  ap- 
proximately what  we  should  expect  to  find." 

For  purposes  of  gas-engine  use  a  gae  should  be  able  to  withstand 
a  relatively  high  compression  without  undue  loss  or  premature 
explosion.    Methane  Hithstands  high  compression  without  change. 

However,  it  may  be  stated  that  in  general,  for  illuminating  gas, 
the  iUuminants,  ethane,  methane,  hydrogen  and  carbon  monoxide 
are  all  desirable  constituents,  because  they  all  add  candle-power 
or  heating  value,  but  carbon  dioxide,  oxygen  and  nitrogen  are'  un- 
desirable because  of  the  lack  of  these  properties. 

UlnmiiiBnts.  The  illuminants  play  a  large  part  in  the  charac- 
teristics of  candle-power  and  calorific  value  of  botii  coal  and  water 
gaa.  Some  of  the  more  important  of  these  compounds,  with  their 
special  characteristics,  are  given  in  the  following  table: 


Bp«o.  at»v. 


B.  T.  V. 

Co.  Ft  Air 




GMOr Vapor  Talueper 
































11.963                        : 




































,  Naphthalene  0»H, 






Manufactdre  and  Distribution  op  Gas 

>y  Google 

2Si  Illuminating  £NQiNEEBiNa 

Note. — All  volumes  of  gaBes  and  vapors  are  given  at  60°  F.  and  30" 
pressure.  Benzene,  being  a  liquid  under  ordinary  condltiona,  was  tested 
for  candle-power  b;  mixing  its  vapor  with  bydrogen,  and  a  silt  burner 

Naphthalene,  being  ordinarily  a  solid,  waa  similarly  mixed  with  coal 

Coal-Gas  Manufacture — As  Produced  in  Retorts 
The  art  of  coal-gas  manufacture  is  over  a  century  old.  William 
Murdoch,  in  England,  between  the  years  1792  and  1798,  was  en- 
gaged in  experimenting  with  different  coals,  and  in  devising  appa- 
ratus for  their  distillation.  In  1797-1?98  lighting  by  coal  gas 
was  actually  accomplished,  for  Murdoch,  by  means  of  his  experi- 
mental plant,  first  lighted  up  his  dwelling  house,  and  a  short  time 
later  a  much  larger  building  at  Birmingham. 

Prom  these  first  attempts,  coal-gas  manufacture  has  been  de- 
veloped to  the  present  state  of  the  art. 

Fio.  2. — Simple  Retort  Setting. 

Principles  of  Coal-Oas  Manufacture 
The  generation  of  coal  gas  from  gas  coal  is  a  process  of  destruc- 
tive distillation.    The  solid  coal  is  charged  into  the  retort,  which 
in  laboratory  parlance  would  be  called  a  miifile,  and  the  retort  is 
heated  externally. 

Figure  3  shows  a  setting,  which,  though  too  primitive  for 
modem  use,  exemplifies  the  primary  principles.  It  consists  of 
a  retort  set  upon  parallel  fire-brick  piers  having  openings  through 
them  for  the  passage  of  the  heated  products  from  the  furnace, 
a  furnace  for  heating,  an  open  apace  around  the  retort  to  per- 

>y  Google 

Manufactuee  a>id  D18TKIBUT10N  OF  Gas  285 

mit  its  eoTelopmeiit  by  tlie  heated  products  of  the  fire,  and  a 
flue  for  the  escape  of  the  products.  The  retort  of  burnt  fire-clay, 
3  laches  thick,  croea-section  oval,  D  shape  or  circular,  being  open 
at  the  front  end  only,  has  bolted  to  that  end  a  caat-iron  extension 
called  a  mouthpiece,  which,  projecting  from  the  front  wall  of  the 
setting,  is  fitted  with  a  gas-tight  door,  through  which  opening  the 
coal  is  introduced  and  the  coke  withdrawn.  At  the  top  or  side  of 
the  mouth[»ece  is  an  opening  to  which  is  connected  a  cast-iron  pipe 
rising  vertically,  the  upper  end  dipping  info  a  seal  of  water.  When 
the  charge  of  coal  is  placed  into  the  heated  retort  distillation  im- 
mediately begins,  vapor  and  gases,  air  and  steam  being  given  oB 
until  the  pressure  is  sufficient  to  overcome  the  eeal  in  ihe  dip-pipe, 
when  the  gas  begins  to  bubble  through  and  continues  until  car- 
bonization (by  which  is  meant  destructive  distillation  of  the  coal) 
ceases.  The  door  is  then  opened  for  the  withdrawal  of  the  coke 
lemaining  in  the  retort  and  reintrodnction  of  fresh  coal.  As  soon 
as  the  door  is  opened  there  is  a  return  of  pressure  in  the  retort 
to  normal  atmospheric,  the  water  riBes  in  the  dip-pipe  thereby 
preventing  gas,  from  the  collecting  main  from  all  the  retorts, 
escaping  through  the  open  door. 

The  practical  extravagance  of  such  &  setting  is  at  once  apparent. 
Cold  air  enters  through  a  shallow  fire,  bums  to  carbonic  acid  and 
eteam,  and  the  heated  products  pass  around  the  retorts,  and  while 
still  highly  heated  escape  to  the  chimney.  When  the  door  is  open 
for  charging  fresh  fuel,  which  is  usually  hot  coke  withdrawn  from 
the  retorts,  and  when  clinkering  the  fire,  cold  air  sweeps  over  the 
fire  directly  around  the  retort,  chilling  it.  Again,  the  combustion 
process  is  the  one  least  suitable  for  surrounding,  with  combustible 
gases,  retorts  set  some  distance  sway,  averaging  4  to  5  feet.  Hav- 
ing but  a  short  distance  to  travel  through  the  fire,  the  conversion 
of  the  oxygen  of  the  air  into  CO,  is  almost  instaotaneous,  and  the 
total  heat  of  the  chemical  combination  is  confined  to  the  fire,  with 
the  result  that  the  fuel  becomes  heated  to  a  temperature  well  above 
the  fusing  temperature  of  the  ash.  This  rapidly  seals  oS  the  fire, 
reducing  the  draft  through  it,  and  the  corabostion  rate  diminishes, 
cooling  the  setting,  while  the  retorts  are  surrounded  only  by  the 
products  of  combustion,  and,  except  for  the  bottoms,  immediately 
over  the  fire,  get  only  the  sensible  heat  of  the  products.  Water 
is  placed  in  the  ash-pan  bo  that  a  small  quantity  of  eteam  rising 
therefrom  may  pass  through  the  fire  to  assist  in  keeping  down  the 

>y  Google 

286  iLLL'UTNAriNO  Enoineebiko 

temperature  of  the  fuel  bed.  Thie,  while  ueeessary  to  protect  the 
grates,  to  a  certain  degree  increaaes  the  difficulty,  since  the  hydro- 
gen thus  formed  burns  at  once  to  water  at  the  top  of  the  fire, 
further  localizing  the  intensity  of  combustion  immediately  above 
the  surface  of  the  fire.  The  result  is,  that  uniform  heating  of  the 
retorts  is  difficult  and  uneconomical.  Thirty  years  ago  this  style 
of  setting  was  in  wide-spread  use.  By  having  a  large  mass  of 
fire-tile  and  small  retorts,  however,  good  results,  as  far  as  the 
quality  of  the  gas  was  concerned,  were  obtainable. 

The  difference  between  heating  a  setting  of  retorts  and  a  boiler 
fire,  for  instance,  is  readily  understood.  In  the  latter  case  com- 
bustion must  have  progressed  to  near  completion  before  the  com- 
bustible products  impinge  on  the  comparatively  cold  tubes  or  shell 
and  combustion  is  arrested.  In  a  setting  of  retorts,  where  all 
parts  are  kept  at  a  temperature  well  above  the  ignition  point  of 
the  most  dilute  gaseous  combustibles,  it  is  desired  that  the  fuel 
bed  should  be  kept  at  a  temperature  just  sufficient  to  carry  on  the 
chemical  reaction  for  the  conversion  of  the  atmospheric  oxygen 
into  carbon  monoxide,  and  the  final  combustion  of  that  gas  occurs 
around  the  retorts  situated  at  a  comparatively  remote  distance 
above  it. 

The  solution  of  these  difficulties  led  to  the  adoption  of  the  re- 
cuperative— sometimes  called  regenerative — method.  Here  there  is 
a  furnace  below  an  arched  chamber  containing  nine  retorts  exposed, 
except  where  supported,  to  the  envelopment  of  heated  products. 
This  arched  chamber  and  its  contents  of  retorts  is  called  a  bench. 
Continuous  arches  so  filled  are  called  a  stack  of  benches.  The 
heated  products  of  combustion  on  their  way  to  the  stack  are  led 
through  passages  made  by  thin  fire-clay  tiles ;  the  primary  air  in 
its  passage  to  the  ash-pit,  and  the  secondary  air  in  its  passage  to 
the  nostrils  above  the  fire,  pass  around  these  tile  flues,  absorbing 
heat  that  was  wasted  in  the  former  setting.  Again,  the  fuel  bed 
was  deepened  so  that  the  oxygen  on  entering  the  fire,  being  first 
converted  into  CO,,  passes  up  through  more  fuel  and  becomes  re- 
duced to  CO.  We  have  now  gaseous  firing.  There  will  be  stored 
in  the  fuel  bed  only  the  heat  developed  by  the  combustion  of  car- 
bon to  carbonic  acid,  and  there  will  be  abstracted  from  the  fuel 
bed  the  heat  absorbed  by  the  separation  of  the  hydrogen  from 
o.iygen  of  the  steam,  the  reduction  of  the  carbonic  acid  to  carbon 
monoxide,  and  the  increase  in  the  sensible  heat  of  the  escaping 

>y  Google 

Manufactubb  and  Distribution  of  Gab  287 

atmoepheric  Ditrogen.  The  top  of  the  furnace  is  covered  with 
A  heavy  covering  of  fire-tile  built  with  an  opening  for  the  paesage 
<if  the  combustible  CO  diluted  vrith  N  to  the  setting  above.  At 
this  point,  the  highly  heated  secondary  air  combines  with  the  gases 
from  the  fire  and  combustion  at  liigh  temperature  ensues.  The 
fire,  meanwhile,  being  deeper,  has  an  arrangement  by  which  false 
grate  bars  can  be  driven  in  at  elinkering  time,  some  distance  above 

B«Bcb  oTNiM  Rtnni,  whh  PuD  Depth  Rccupcnion. 
Fio.  3. 

the  fixed  grates,  holding  up  the  fire  while  the  clinker  is  bojng 
removed  from  between  the  false  and  fixed  bars.  There  is  also  an 
arrangement  by  which  small  streams  of  water  drip  through  the 
bottom  of  the  fire,  reducing  the  temperature  further. 

This  method  with  care,  gives  satisfactory  results,  and  is  in 
extended  use  to-day.  A  further  improvement  in  conditions  is 
now  obtained  by  returning  a  small  quantity  of  the  products  of 
combustion  through  the  fire,  diluting  the  oxygen  of  the  air  and 
prolonging  the  period  of  combustion  until  all  the  retorts  are  bathed 
in  flame. 

>y  Google 

288  iLLUHIHATINa  ENQiNaKsmo 

Modem  practice  requires  careful  attention  to  bench  heating. 
The  per  cent  of  combustible  in  the  stack  gases,  the  quantity  of 
primary  and  secondary  air  are  accurately  read  by  meter  and  pro- 
portioned, and  the  heats  of  the  combustion  chamber  and  through- 
out the  setting  are  noted  at  frequent  intervals,  with  the  result 
that  uniform  heats  in  the  retorts  at  from  1700"  F.  to  1900°  P.  are 
maintained  with  an  expenditure  of  fuel  per  ton  of  coal  carbonized 
much  less  than  formerly  obtained.  A  retort,  when  discharged  of 
its  coke,  should  show  a  uniformly  heated  interior  surface  through- 
out— bright  red  in  color. 

The  angle  at  which  the  retorte  are  inclined  to  the  horizon  is  a 
question  of  much  importance.  But  before  we  have  sui&cient  data 
to  take  up  a  discussion  of  this  question,  we  must  look  into  what 
goes  on  inside  the  retort. 

After  a  century  of  close  study,  it  may  be  said,  with  regret,  that 
not  all  of  the  details  of  the  chemical  reaction  attending  the  con- 
version of  coal  into  gas  by  the  retort  process  are  known  at  this  time. 

The  process  is  considered  as  occupying  three  stages : 

In  the  first  one,  a  quantity  (about  350  pounds)  of  cold,  damp 
coal  is  charged  into  a  retort  9  feet  long  and  approximately  14 
inches  by  26  inches  in  cross-section;  then  there  is  a  rapid  cooling 
taking  place  on  the  inner  surface  of  the  retort.  There  is  an  ab- 
sorption of  heat  by  the  coal,  due  to  the  high  therro^  head  existent, 
and  the  heat  rendered  latent  by  the  immediate  evaporation  of  the 
water  and  the  more  volatile  vapors  in  the  coal.  We  know  that  the 
distillation  begins  on  that  portion  of  fhe  coal  in  contact  with  the 
sides  of  the  retort. ' 

When  coal  is  heated  in  a  closed  vessel  at  a  heat  of  about  660° 
to  IQO'  F.,  fusion  of  the  coal  commences  and  hydrocarbon  vapors 
begin  to  come  off.  If  these  vapors  were  condensed,  they  would  be 
found  to  be  mainly  paraffins  and  olefins. 

The  second  stage  ensues  in  which  these  hydrocarbons,  meeting 
with  the  higher  temperatures,  begin  to  be  affected.  It  is  believed 
that  there  is  a  rearrangement  and  loosening  of  the  C-H  and  the 
C-C  bonds,  and  other  compounds  are  formed.  In  the  third  stage 
the  heat  of  the  interior  of  the  retort  rises  still  higher,  the  reactions, 
almost  instantaneoUB  in  many  instances,  are  most  complex,  aiM  so 
far  have  resisted  entire  elucidation.  The  aliphatic  hydrocarbons, 
that  is,  the  open-chain  series,  paraffins  and  olefins,  as  found  in 
gas  coal   and  petroleum,  are,  on  the  one  hand,  loosening  their 

>y  Google 

MaNUFAOTDBE  and   DiBTRIBDTION   OP   Gas  289 

carbon  bonda  and  splittiBg  off  the  initial  or  simplest  members  of 
their  series,  while  the  residues  unite  into  more  complex  closed- 
chain  or  aromatic  compounds,  such  os  benzene,  toluene,  xylene, 
etc.  These  benzol  compounds,  under  the  influence  of  heat,  in  time 
are  decomposed  with  the  liberation  of  hydrogen,  carbon  and  the 
formation  of  still  higher  ring  compounds.  On  the  other  hand, 
the  free  hydrogen  present  reacts  on  the  aliphatic  hydrocarbons. 
In  the,  meanwhile,  the  oxygen  and  the  nitrogen  in  the  coal  are 
forming  other  combinationB,  some  of  the  nitrogen  going  into  am- 
monia and  some  of  the  oxygen  uniting  to  form  phenols. 

A  West  Virginia  coal  would  have  a  hydrocarbon  component  that 
is  expressed  as  approximately  Csi,H4„0g,. 

This  third  stage  is  the  one  which  does  most  to  detennine  the 
candle-power  and  heating  value  of  the  gas  obtained.  The  retort 
ie  filled  to  about  40  per  cent  of  its  volume  with  coal.  After  the 
water  and  first  vapors  are  driven  oS,  the  coal  continues  to  fuse 
and  the  evolution  of  gas  becomes  more  rapid,  and,  passing  above 
the  coal,  is  exposed  to  the  highly  heated  sides  and  top  of  the  retort. 
The  hydrocarbons  and  other  vapors  pass  off  in  gradually  decreas- 
ing proportion  during  the  distillation  period  of  the  charge,  which 
we  are  now  considering  as  being  of  about  4  hours'  duration,  and 
as  the  volume  becomes  less  the  energy  expended  on  them  dimin- 
ishes and  the  retort  gradually  increaeee  in  temperature  toward  the 
end  of  the  period,  at  which  time,  the  temperature  being  higher, 
the  flow  of  gas  slower,  the  effect  produced  upon  the  gas  is  con- 
stantly changing. 

I  have  so  far  asked  your  attention  to  the  consideration  of  that 
form  of  modem  retort  setting  in  which  the  charge  of  coal  is  dis- 
tilled in  the  ehortest  time,  generally  4  hours.  This  design,  being 
only  a  mechanical  improvement  upon  the  centnry-old  chemical 
distillation  of  coal — an  improvement  looking  toward  economy  in 
heating  the  bench  and  procuring  more  "even"  heats,  except  so 
far  as  those  bettered  conditions  could — did  nothing  to  improve 
the  chemical  reactions.  The  distillation  of  coal  is  still  conducted 
under  very  different  conditions  at  the  beginning  than  at  the  end  of 
the  period,  and  the  gas  emanating  from  the  coal  in  the  back  of 
the  retort  is  exposed  to  different  heating  conditions  than  that  in 
the  front. 

Avoiding  a  too  technical  and  voluminous  discussion  of  these 
changes,  it  will  still  be  well  for  me  to  make  a  simple  statement 

>y  Google 

290  Illuminatimq  Enoinebbing 

of  the  most  important  changes  occurring  in  this  third  stage  of 
how  vapors  of  the  paraffin  and  olefin  series,  which  are  those  coming 
from  the  second  stage,  are  affected  by  the  temperatures  of  the 

It  is  doubly  important  to  the  gas  engineer,  because  the  same 
reactions  occur  in  a  water-gas  apparatus  or  in  an  oil-gae  plant, 
where  paraffin  base  oils  are  subject  to  the  "  cracldng-up  "  process. 
It  constitutes  one  of  the  chief  sources  of  interest  and  study  of 
the  gas  engineer,  aod  a  better  knowledge  is  sure  to  be  rewarded 
by  more  economical  operation  of  a  gas  plant,  a  better  profit  and 
improved  product. 

The  higher  members  of  the  paraffin  series  and  olefin  series  break 
down  even  at  temperatures  below  their  boiling  points,  under  normal 
pressure,  to  lower  hydrocarbons  of  the  same  series,  and  the  paraffins 
to  some  extent  are  converted  into  the  olefin  series.  Under  con- 
tinued exposure  to  these  high  temperatures,  the  lower  paraffins  and 
olefins  are  converted  into  members  of  the  benzene  series  with  de- 
posits of  free  carbon ;  if  the  heat  still  continues  there  is  a  produc- 
tion of  acetylene,  followed  at  once  by  a  breaking  down  into  marsh 
gas  and  a  large  deposit  of  free  carbon.  Benzene  (CjH,),  the  lowest 
member  of  the  benzene  series,  at  ordinary  temperatures  exists  as  a 
vapor.  It  has  a  high  illuminating  value,  and,  in  water  gas  made 
from  some  oils,  contributes  largely  to  the  illuminating  power, 
though  not  so  much  to  its  heating  value. 

It  is  clear  that  the  "  cracking-up  "  process  cannot  go  beyond  the 
benzene-forming  period  without  disastrous  effect  on  the  value  of 
the  gas,  and  it  is  true,  further,  that  the  formation  of  the  benzenes 
are  a  loss  in  candle-power  value  over  what  would  have  occurred 
if  the  olefin  gases,  such  as  ethylene,  had  not  been  broken  up.  In 
other  words,  if  we  could  convert  the  paraffins  and  olefins  all  into 
members  of  the  olefin  series,  gaseous  at  ordinary  temperatures, 
the  highest  efficiency  would  be  realized,  but  in  the-  rush  of  gas 
through  the  retort  all  the  reactions  are  taking  place  at  once.  While 
some  of  the  heavy  paraffins  and  olefins  are  breaking  down  into 
lighter  members  of  the  same  series  others  are  being  converted  into 
benzols,  while  some  of  the  benzols  are  going  into  hydrogen  and 
free  carbon.  We  must,  therefore,  use  a  heat  .which  will  crack  up 
all  the  heavy  paraffins  and  olefins,  remove  the  gas  before  the  final 
general  breaking  down  occurs,  and  expect  some  losses  in  the  process. 

The  difficulties  that  the  coal-gas  engineer  has  to  meet  are  now 

>y  Google 

Manufactube  and  Distribution  of  Gas  291 

evident.  He  must  maintain  a  heat  in  his  retorts  that  will  secure 
the  proper  "cracking  up"  of  the  heavier  rush  of  gas  in  the  first 
hour,  and  he  must  expect,  in  the  form  of  retort  under  discussion, 
that  there  will  be,  toward  the  end  of  the  carbonization  period,  too 
great  an  exposure  to  the  heat  of  the  smaller  volume  of  gas,  break- 
ing down  into  free  carbon  and  methane — a  non-illuminating  gas. 
Other  designs  of  retort  settings  suggest  themselvea  as  better  than 
the  one  we  have  so  far  discussed.  Instead  of  withdrawing  the  coke 
from  the  same  door  through  which  the  coal  was  charged,  retorts  are 
used  in  many  installations  which  open  at  both  ends.  The  coal  is 
charged  into  the  retort  until  it  is  nearly  full,  and  the  coke  is 

Fig.  4. — Retorts. 

pushed  out  through  the  other  end,  the  operation  of  pushing  the 
coke  out  and  recharging  the  retort  being  done  by  machinery  in  one 
motion.  By  this  means  the  gas  from  the  coal  flows  through  the 
retort  more  rapidly;  by  reason  of  the  smaller  area  existing  between 
the  coal  and  the  top  and  sides  of  the  retort,  the  temperature  of  the 
retort  is  reduced  and  a  longer  time  is  given  for  the  carbonization 
period.  There  is  still,  however,  direct  exposure  of  the  gas  to  the 
radiant  heat  from  some  portions  of  the  retort. 

Another  development  is  in  the  vertical  retort.  Here  the  coal 
is  charged  into  the  top  and  the  coke  taken  from  the  bottom  of  a 
vertical  retort,  which  usually  tapers  to  somewhat  larger  at  the  bot- 
tom. Here  the  coal  is  fused,  tilling  the  retort ;  there  is  no  appre- 
ciable amount  of  space  between  the  coal  and  the  sides  of  the  retort 
for  the  gas  to  be  highly  heated,  and  the  gases  must  flow,  in  a  large 

>y  Google 

292  Illuminating  Enoineebinq 

part  at  least,  up  through  the  central  unfuaed  core  of  the  coal  itaelf, 
thereby  escaping  the  difficulty  under  discuBsion.  That  there  is  less 
breaking  down  into  free  carbon,  marsh  gas  and  hydrogen  in  the 
vertical-retort  process  than  in  the  horizontal  is  apparent  by  the 
smaller  percentage  of  free  carbon  extracted  from  the  gas  by  the  tar 
in  the  after  processes.  The  coal  is  raised  to  a  greater  altitude 
than  in  the  horizontal  retort,  and  when  charged  into  the  mouth 
at  the  top,  which  by  machinery  can  be  done  with  little  labor,  falls 
of  its  own  weight  out  of  the  bottom  as  coke.    Vertical  retorts  are 

Fia.  6. — Vertical  Retorts. 

in  wide  use  in  Europe,  having  superseded  inclined  retorts,  which  do 
not  appear  to  suit  the  theoretical  conditions  as  well  ae  the  verticals. 

The  coke  oven  is  another  attempt  at  the  solution  of  tbe  problem 
of  getting  uniform,  moderate  heat  throughout  the  body  of  the  coal 
and  throughout  the  carbonization  period.  It  is,  in  effect,  a  large 
"double-end"  horizontal  retort  in  which  large  quantities  (6  to  8, 
sometimes  10  tons)  of  coal  are  exposed  to  carefully  graduated  but 
moderate  heats  for  from  20  to  36  hours. 

What  effect  upon  the  cracking  up  of  hydrocarbons,  of  tempera- 
ture versus  heat  has,  can  hardly  be  discussed  by  me  now.  What  is 
the  relative  effect  of  long-continued  exposure  to  moderate  tempera- 
ture, to  quick  exposure  to  high  temperatures  ?    Some  of  our  leading 

>y  Google 

Manufactuhb  and  DiSTBiBUTiON  OF  Gas  993 

developers  of  the  chemistry  of  gas  manufacture,  notably  the  veteran 
Young,  probably  the  most  original,  as  he  was  the  pioneer  in  this 
field,  maintain  that  radiant  heat  has  a  very  different  effect  on 
"  cracldQg  "  than  conducted  or  convected  heat. 

no.  8. — Otto-Hllgenetock  Coke  Oven.    Rageiieratlve  Type, 

Pio.  7. — Sketch  ot  Purifier, 

Purification  of  Coal  Qas 

The  principal  impuritiee  in  coal  gas,  which  must  be  extracted 
before  the  gas  ie  fit  for  commercial  use,  are  tar,  ammonia,  snlphar 
and  sometimes  cyanogen.  In  connection  with  parificatioQ  the  sub- 
ject of  condensation  will  be  treated. 

The  fundamental  principle  of  condensation  is  to  reduce  the  gas, 
dnring  its  pass^e  through  the  works  to  a  proper  temperatnre,  so 
that  in  its  distribution  through  the  gas  mains  to  the  consumers' 

>y  Google 


Illuminating  Enoinb^iino 

appliances  no  vapors  will  condense  out  of  it.  In  other  words,  after 
proper  condensation  at  the  works,  the  gas  is,  generally  speakiiig, 
in  a  permanent  fixed  form  for  the  ordinary  conditions  of  dis- 

The  principles  employed  in  condensing  coal  gas  are  as  follows : 

First,  gradual  reduction  in  temperature  down  to  about  llO"  F. 

Above  this  point,  as  much  of  the  tar  as  collects  in  the  hydraulic 

main  and  foul  mains  is  allowed  to  pass  o£E  into  the  tar  well.     If 

coal  gas  at  or  beiow  110°  F.  is  allowed  to  remain  in  contact  with. 

coal-tar  a  great  amount  of  the  heavy  hydrocarbons  in  the  ga» 
are  absorbed  by  it.  By  draining  the  tar  off  at  proper  points  in  the- 
process,  the  benzol  and  other  heavy  vapors  are  retained  in  the  gas. 

Some  tar  is  always  carried  with  the  gas  through  the  various 
works  pipes,  and  serves  to  absorb  excess  naphthalene  vapors. 

After  the  primary  condensation  down  to  about  110°  F.,  a  further 
extraction  of  tar  takes  place.  This  is  accomplished  in  variona  ways, 
such  as  hot  washing,  or  scrubbing,  by  centrifugal  force,  or  mechan- 
ically, as  in  a  P.  &  A.  tar  extractor,  where  the  particles  of  tar  are- 
projected  by  high  velocity  against  metal  surfaces,  where  they  ate 
deposited  and  run  off. 

>y  Google 

Manufactdbe  and  Dibthibution  of  Gas  295 

The  condensatioii  principle  of  gradually  cooling  the  gas  is  im- 
portant, as  this  prevents  the  sudden  shocks  to  the  gas,  with  at- 
tendant losses  of  valuable  hydrocarbon  vapors.  Certain  hydro- 
carbon vapors  possess  the  property  of  apparently  carrying  other 
hydrocarbon  vapors  in  a  so-called  state  of  suspension,  up  to  the 
saturation  point,  which  varies  with  the  temperature. 

Naphthalene,  Coal  Oas.  The  subject  of  condensation  would  be 
incomplete  without  brief  reference  to  naphthalene.  Its  formation 
is  believed  to  be  principally  due  to  the  latter-day  high  heats  of 

Fio.  9. 

carbonization,  and  where  it  occurs  in  quantities  it  becomes  ex- 
ceedingly troublesome.  Recently,  washing  the  gas  with  certain 
oils  has  proved  very  successful.  In  mixed  coal-  and  water-gas 
plants  naphthalene  is  very  readily  handled,  owing  to  the  fact  that 
the  rich  hydrocarbons  in  water  gas  absorb  and  carry  it  along. 

The  mechanical  principle  employed  in  condensers  is  simply  the 
transmission  of  the  heat,  either  sensible  or  that  freed  by  reason 
of  the  latent  heat  of  condensation  of  vapors,  through  steel,  usually 
tubes,  to  air  or  water  which  ere  used  as  the  mediums  for  absorption. 

>y  Google 

896  Illchinatin'O  Ehqinsebinq 

Ammonia  is  extracted  from  coal  gas  by  the  well-known  principle 
of  the  poH-er  of  water  to  absorb  it.  The  mechanical  methodB  of 
doing  this  are  by  so-called  washing  and  scrubbing.  In  the  earlier 
stages  of  the  process  it  is  advisable  to  wash  or  scrub  the  gas  with 
crude  ammoniacal  liquor,  which  assists  in  remoTing  tar,  C0„  HjS 
and  CS,  from  the  gas.  The  crude  liquor  also  extracts  ammonia. 
Of  course,  the  final  traces  of  ammonia  are  eliminated  by  the  use 
of  fresh  water. 

Sulphur  exists  in  crude  gas  as  H,S,  and  also  o^anic  compounda, 
the  latter  being  largely  CS,.    Washing  or  scrubbing  the  gas  with 

Flo.  10.— Water-Cooked  Condenser. 

crude  ammoniacal  liquor  extracts  a  portion  of  these  compounds, 
which  form  various  chemical  combinations  with  NH,.  A  recent 
system  of  treating  the  gas,  called  the  Feld  system,  eliminates  usu- 
ally by  far  the  greater  portion  of  H,8,  also  some  organic  sulphur 
is  removed  in  the  purifiers. 

In  the  United  States  iron  oxide  is  used  in  the  usual  system  of  pur- 
ification. The  HjS  in  the  gas  combines  with  the  iron  oxide  to  form 
iron  sulphide.  The  "  fouled  "  material,  by  exposure  to  air,  revivi- 
fies, the  oxygen  of  the  air  combining  with  the  iron  sulphide  to  form 
iron  oxide,  leaving  the  sulphur  in  the  material  in  the  free  state. 

>y  Google 

MaNUPACTCBE  and   DiBTBlBUTION   OF  Gab  297 

The  free  sulphur  probably  does  extract  a  certain  amount  of  CSj 
from  the  gas,  as  CS,  disBolves  sulphur. 

In  England  the  hydrated  form  of  quicklime  is  employed.  This 
process  removes  CSj  as  well  as  H,S,  but  is  not  much  used  in  this 
country  on  account  of  the  expense. 

Carhuretted  Water  Gas  as  Made  from  Fixed  Carbon,  Steam  and  Oil 
It  is  not  the  intention  to  present  for  your  consideration  mere 
history,  but  a  brief  reference  to  the  development  of  carburetted 
water  gae  may  not  be  out  of  place,  and  will  probably  assist  in  the 
clearer  understanding  of  the  principles  underlying  this  process. 

Pre.  11.— Plate  No.  2 — Rotary  Scrubber. 

The  fundamental  chemical  principles  underlying  the  process  of 
making  this  gas  from  fixed  carbon,  steam  and  oil  are  compara- 
tively simple.  In  the  first  place,  there  is  a  bed  of  fuel,  brought  up  to 
high  temperature,  which  we  may  call  incandescent  carbon  for  the 
purposes  of  this  lecture.  Steam  is  admitted  and'  passed  through 
this  fuel,  and,  as  is  well  known,  decomposes  into  its  elements  hydro- 
gen and  oxygen  in  the  presence  of  incandescent  carbon.  To  give 
you  an  idea  of  such  reactions,  and  the  approximate  minimum  tem- 
peratures at  which  such  decomposition  takes  place,  whether  in  the 
presence  of  incandescent  carbon  or  not,  the  following  table  is 
shown : 

H,0-<-»-H  +  0.     Min.  temp,  about  1000°  Cent.-1200°  Cent. 

H,0  +  C  =  CO-)-H.    Min.  temp,  about  600°  Cent. 
Prom  this  you  .will  note  the  comparatively  low  temperature  re- 
quired to  decompose  H^O  in  the  presence  of  incandescent  carbon. 

>y  Google 

298  Illuminating  EKaiNEBsiNa 

The  result  of  this  reaction,  which  takes  place  in  a  fire-brick-lined 
vessel  called  a  generator,  is  the  formation  of  so-called  blue  or 
uncarburetted  water  gas,  which  consists  principally  of  carbon  mon- 
oxide and  hydrogen,  and  burns  with  a  blue  practically  non-lumi- 
nous flame,  and  has  a  calorific  value  of  about  3S0  B.  t.  u.  per  cubic 

This  blue  gas  then  passes  into  a  fire-brick-lined  v«ssel  filled  with 
a  checker-work  of  fire-brick,  which  has  been  heated  to  incandes- 
cence. A  spray  of  hydrocarbon  oil  is  admitted  above  this  checker- 
brick,  is  vaporized  and  gasified  by  the  heat,  and  mixes  with  the 
blue  gas  previously  described-  The  oil  furnishes  the  illuminants 
necesBary  for  candle-power,  and  from  the  analysis  submitted  in  the 
early  part  of  this  lecture,  it  will  he  seen  that  a  good  calorific  value 
is  also  obtained.  The  candle-power  and  calorific  value  depend  very 
largely  upon  the  relative  quantities  of  blue  gas  and  the  gas  re- 
sulting from  the  decomposition  of  the  oil. 

The  mixture  of  blue  and  oil  gas  is  Bubeequently  subjected  to  a 
so-called  "  fixing  "  process,  by  being  passed  through  an  additional 
amount  of  heated  checker-brick,  the  effect  of  which  is  merely  to 
render  the  various  hydrocarbon  gases  more  permanent  under  ordi- 
nary temperatures,  probably  by  the  reason  of  the  decomposition  or 
partial  decomposition  of  some  of  the  richer  hydrocarbons  into  the 
simpler  and  more  stable  forms. 

Development  of  Water  Gas 

The  production  of  water  gas  has  been  attempted  in  three  ways : 

First.    In  the  earlier  forms  it  was  attempted  to  produce  water 

gas  by  contact  of  steam  with  heated  coal  or  coke  contained  in  a 

retort  externally  heated,  as  is  illustrated  by  the  Harris  patent. 


A  bench  of  three  clay  retorts,  shown  in  Figure  1,  was  used.  Re- 
tort A,  or  the  decomposing  retort,  was  provided  with  a  perforated 
tile  (Fig.  3).  The  retort  was  filled  above  the  tile  with  anthracite 
coal  broken  to  the  size  of  an  egg. 

Retorts  B  and  C  were  filled  with  rich  cannel  coal.  Figure  8 
shows  a  cross-section  of  the  decomposing  retort.  Figure  4  il- 
lustrates the  steam  drier  which  was  placed  near  the  base  of  the 
furnace.     Figure  5  represents  the  steam  superheater. 

Steam  supplied  from  a  boiler  heated  by  the  waate  gases  from 
the  bench  was  first  passed  through  superheater  p  into  retort  A, 

>y  Google 


>y  Google 

300  Illdminatinq  Enoikeebino 

passing  through  the  distributing  tile  D  into  the  highly  heated 
anthracite  coal.  Leaving  this  retort  the  gas  was  conducted  to  the 
rear  end  of  either  of  the  lower  retorts,  B  and  C,  through  pipes  H, 
and  from  this  retort  through  stand-pipes  K  to  the  hydraulic  main. 

The  gas  from  the  decomposing  retort  was  supplied  to  one  bitumi- 
nous retort  until  the  rich  hydrocarbon  vapors  of  the  charge  in 
this  retort  were  exhausted. 

This  retort  was  then  closed  oft  by  means  of  cock  3  and  the  gases 
from  retort  A  were  then  passed  to  the  other,  etc. 

Retorts  B  and  C  were  charged  at  intervals  of  about  3  hours- 
These  attempts  were  unsuccessful,  but  your  attention  ia  directed 
to  them  to  illustrate  the  basic  principles.  Various  patent  appli- 
cations, from  time  to  time,  show  the  recurrence  of  this  idea  in 
different  men's  minds.  The  reason  of  the  failure  of  this  process 
is  because  the  chemical  reaction  of  steam  upon  the  fixed  carbon 
of  the  incandescent  coal  or  coke  is  an  endothermic  one,  in  other 
words,  one  which  absorbs  energy  in  the  form  of  heat,  and  requires 
much  more  heat  to  maintain  it,  aud  more  intimate  association  of 
the  steam  and  coai  or  coke  than  can  be  obtained  in  this  way. 

Second.  The  next  step  in  the  process  is  embodied  in  the  ideas 
formulated  by  Tessie  du  Motay.  In  general,  this  process  consists 
in  making  blue-water  gas  intermittently  in  a  generator  and  storing 
same  in  a  holder. 

The  apparatus  consists  of  generator  A,  gas-relief  holder,  bench  of 
retorts  with  furnace  C,  retorts  D,  hydraulic  main  E,  and  naphtha 
vaporizer  B. 

The  generator  is  filled  with  anthracite  coal  or  coke,  through 
which  steam  is  passed,  after  this  bed  of  fuel  has  been  brought  to 
incandescence;  the  resulting  gas  being  a  blue-water  gas,  largely 
CO  and  hydrogen,  this  gas  being  passed  alon^to  the  relief  holder 
for  storage.  The  bench  of  retorts  having  been  brought  to  the 
proper  heat  for  vaporizing,  the  oil  gas  is  admitted  to  the  front  end 
of  retorts  at  point  "H,"  and  at  the  same  point  naphtha  vapor  is 
admitted,  the  naphtha  having  been  vaporized  in  vaporizer  "  B  "  by 
means  of  steam  coils  or  otherwise;  the  naphtha  vapor  and  blue- 
water  gas  are  each  regulated  at  this  point,  "  H,"  to  produce  the 
proper  candle-power  of  gas,  and  passing  through  the  retorts  "  D," 
coming  in  contact  with  the  heated  surface  is  sufficiently  heated 
to  be  largely  converted  into  a  fixed  gas,  passing  off  at  the  opposite 
end  of  the  retorts  to  the  hydraulic  main,  afterwards  treated  in  a 

>y  Google 

Manufacture  and  Distbibution  of  Gas 

>y  Google 

302  Illumihatinq  Enqinesbino 

similar  manner  to  other  gases.  In  operating,  the  generator  was 
first  bronght  up  to  heat  by  blowing  sufficient  air  through  a  bed  of 
fuel  to  raise  this  bed  of  fuel  to  a  high  temperature.  When  the 
fuel  was  hot  enough,  the  blaat  was  cut  off,  a  valve  closed  and  steam 
admitted,  which,  on  passing  through  the  fuel,  resulted  in  the  pro- 
duction of  blue-water  gas.  The  endothermic  action  of  decomposi- 
tion of  steam  in  the  fuel  bed  resulted  in  a  rapid  cooling  of  the 
fire.  When  the  fire  temperature  became  so  low  that  the  steam  was 
no  longer  readily  decomposed,  the  admission  of  the  steam  was  dis- 
continued, and  the  blast  turned  on  again,  as  before,  and  the  cycle 
of  operations  repeated. 

In  the  meantime  hydrocarbon  oils  were  being  vaporized  in  a  sepa- 
rate apparatus,  and  these  vapors,  mixed  with  the  bine-water  gas, 
were  passed  through  an  apparatus  externally  heated,  wherein  the 
gas  was  "  filed  "  or  rendered  permanent. 

The  limitations  of  this  system  of  gas  manufacture  were  thftt 
the  oils  which  could  be  vaporized  were  the  refined  fractions  of 
crude  oil,  called  naphtha,  and  as  these  oils  rapidly  advanced  in 
price  the  limit  of  economical  operation  on  a  commercial  scale  was 
soon  passed. 

The  Lowe  Process 

Third.  The  Lowe  procese.  This  method,  or  modification  of  it, 
is  the  one  in  use  to-day.  To  describe  its  essential  principles  it  is 
advisable  to  insert  a  short  description  of  the  apparatos  used.  A 
Lowe  water-gas  set,  or  its  equivalent,  eonsiste  of — 

First.  A  generator,  or  vessel  built  of  an  iron  shell  with  a  fire- 
brick lining,  and  containing  a  deep  bed  of  fuel. 

Second.  A  carbureter,  or  vessel  consisting  of  an  iron  shell  lined 
with  fire-brick,  and  filled  with  a  checker-work  of  fire-brick.  This 
vessel  has  an  open  chamber  at  the  top  into  which  the  oil  is  sprayed. 

Third.  A  superheater,  or  vessel  built-  and  checkered  similar  to 
a  carbureter. 

To  explain  the  operation  of  such  a  set  we  will  first  assume  it 
cold,  but  with  a  coke  or  anthracite  fire  started  in  the  generator. 
By  means  of  a  blower  an  air  blast  is  turned  under  this  fire,  and  the 
carbon  in  the  fuel  bed  burns  partiy  to  COj,  partly  to  CO.  The  C0„ 
on  passing  through  the  incandescent  fuel  bed,  is  practically  wholly 
decomposed  to  CO,  the  amount  depending  on  blast  velocity,  tem- 
perature, etc. 

>y  Google 

Manufaotdre  Aim  Distbibotion  of  Gas  303 

When  the  producer  gas  (for  such  it  is)  reaches  the  top  of  the 
generator  above  the  fire  it  conaistB  principally  of  N,  CO  and  a  email 
percentage  of  COi-  By  means  of  a  targe  fire-brick-lined  connection 
this  producer  gas  is  conducted  to  the  top  of  the  carbureter.  Here 
an  additional  blast  opening  introduces  fresh  air,  and  a  portion  or 
all  of  the  CO  in  the  producer  gas  burns  to  COj  in  the  carbureter. 
The  resulting  mixture  passes  out  of  the  carbureter,  and  into  the 
bottom  of  the  superheater,  where  still  another  blast  admits  enough 
air  to  bum  the  remaining  CO  to  CO;,  in  case  it  is  desired  to  heat 
the  superheater  higher,  but  if  not,  no  farther  air  is  admitted  here. 
The  final  waste  gases  then  pass  out  of  the  stack  valve  at  the  top 
of  the  superheater  and  escape  into  the  atmosphere,  or  are  first 
passed  through  some  apparatus  to  abstract  as  much  of  the  remain- 
ing sensible  heat  as  possible.  This  process  of  blasting  or  blowing 
is  continued  until  the  entire  fuel  bed  is  highly  incandescent,  the 
checker-work  in  the  carbureter  at  a  high  heat,  and  at  a  reduced 
temperature  in  the  superheater.    The  set  is  then  ready  to-make  gae. 

The  blast  is  first  shut  ofE  from  all  of  the  vessels,  and  the  stack 
valve  on  the  superheater  closed,  live  steam  is  then  turned  into  the 
generator  below  the  fire.  The  resulting  reactions  are  very  instruc- 
tive. The  H3O  vapor  is  first  decomposed  by  the  incandescent  car- 
bon to  hydrogen  and  oxygen.  This  reaction  is  endothermic,  that  is, 
heat  is  absorbed  in  doing  this  work.  The  hydrogen  passes  through 
the  fire  unchanged. 

The  oxygen,  on  the  other  hand,  immediately  combines  with  car- 
bon to  form  CO  and  COj,  and  every  pound  of  carbon  thus  burning 
to  CO,  gives  off  about  14,544  B.  t.  u.,  the  reaction  being  exothermic. 
The  CO  passes  on  through  the  fire,  but  the  CO2,  in  the  presence  of 
the  incandescent  carbon,  decomposes  to  CO,  the  reaction  being 

The  gas  appearing  on  the  top  of  the  fire,  then,  is  a  mixture  of 
hydrogen  and  carbon  monoxide,  in  practically  equal  proportions, 
together  with  a  small  percentage  of  00^  and  some  impurities. 
This  mixture  is  the  so-called  blue  or  nncarburetted  water  gas,  and 
is  merely  one  form  of  producer  gas,  having  a  calorific  value  of 
about  320  B.  t.  u. 

It  will  be  noticed  that  the  reactions  in  the  generator  are  mostly 
endothermic,  and,  in  fact,  the  fire  is  cooled  very  rapidly  during  the 
admission  of  steam,  a  run  being  generally  from  5  to  10  minutes, 
at  the  end  of  which  it  is  necessary  to  blast  again. 

>y  Google 

304  Illcminatisg  Enoinbehing 

Coming  back  to  the  blue-water  gas,  so-called  because  it  bums 
with  a  blue  flame  in  air,  we  find  upon  leaving  the  top  of  the  gen- 
erator that  it  paeses  into  the  top  of  the  carbureter.  Here  it  meets 
with  a  Bpray  of  oil.  This  is  sometimes  the  crude  oil,  bnt  more 
often  a  gas  distillate,  which  is  the  fraction  obtained  from  crude  oil 
after  distilling  ofT  tho  gasolines  and  kerosenes,  and  stopping  before 
the  heavier  lubricating  oils  appear. 

This  oil,  coming  into  the  top  chamber  of  the  carbureter,  vapor- 
izes under  the  intense  heat  and,  mixing  with  the  blue  gae,  starts 
through  the  carbureter.  The  lower  portion  of  the  carbureter  and 
the  superheater  are  merely  heated  checker-work  for  rendering  the 
gases  permanent  under  ordinary  conditions,  or  "  fixing  "  it,  as  it 
is  called  in  operative  parlance. 

Crude  petroleum  consists  of  a  mixture  of  a  great  number  of 
definite  hydrocarbons,  that  is,  hydrocarbons  that  may  be  designated 
by  exact  chemical  formulae,  but  which  are  so  almost  inextricably 
mixed  in  the  oil  that  the  separation  of  any  one  of  the  hydrocarbons 
in  considerable  quantities  requires  repeated  distillations  under  fa- 
vorable conditions  and  chemical  treatment. 

Crude  oils  are  designated  as  paraffin  base,  semi-paratHn  base  and 
asphalt  base,  according  to  the  general  character  and  composition 
of  the  oil. 

Paraifin-baGe  oil,  ae  I  have  stated  in  discussing  coal-gas  mann- 
facture,  is  one  made  up  almost  entirely  of  members  of  the  paraffin 
and  olefin  series.  Paraffins  from  simple  CH,  methane  to  penta- 
tricontane  CjsH,,  have  been  isolated ;  methane  CH„  the  simplest 
member  existing  as  a  goe;  pentatricontane  (CgH;,),  as  a  solid, 
melting  at  76°  F.  This  oil  is  found  in  the  northern  oil  districts, 
such  as  Pennsylvania  and  Ohio. 

Semi -paraffin -base  oil  contains,  in  addition  to  paraffins  and  ole- 
fins, naphthenes  (CnHj,).  These  compounds  have  the  same  chem- 
ical formulae  as  the  olefins,  but  have  markedly  different  character- 
istics. The  explanation  for  this  is  in  the  way  that  the  C  and  H 
atoms  are  united,  differing  in  the  two  series,  the  carbon  particles  in 
the  olefins  existing  as  a  simple  chain,  whereas  the  naphthene  car- 
bon atoms  are  considered  as  being  grouped  as  a  closed  ring.  This 
class  of  oil  is  found  in  southern  districts,  like  Louisiana  and 

Asphalt-  or  naphthalene-base  oils  are  made  up  largely  of  naph- 
thene and  olefins,  paraffins  tioing  almost  entirely  absent.    Examples 

>y  Google 

Mandpaotuhb  and  Distribution  of  Gas  305 

of  this  kind  are  found  in  Texas  and  California.  The  naphthene 
series  are  much  more  stable  than  the  pai-aSin;  they  do  not  yield 
paraffins  or  olefins  in  cracking  under  heat,  hut  pass  at  once  into 
members  of  the  benzene  series,  such  as  benzene,  toluene,  xylene  and 
higher  members.  These  benzenes  exist  in  the  gas  only  as  vapors, 
and  are  subject  to  the  laws  of  vapors  regarding  saturation  and 
precipitation ;  consequently,  gaa  made  from  naphthene  oils  must 
be  very  carefully  handled  in  the  procfsses  subsequent  to  generation 
to  secure  to  the  consumer  equal  candle-power  at  all  seasons  of  the 
year.  For  additional  information  on  the  treatment  of  this  gas  I 
would  refer  you  to  a  paper  presented  to  the  American  Gas  Insti- 
tute by  W.  H.  Gartley  in  1907. 

The  results  of  gasifying  the  oil  show  that  the  various  hydro- 
carbons evolved  depend,  as  to  nature  and  relative  quantities,  on 
time,  temperature,  relative  quantities  of  oil  injected,  and  amount 
of  heat  available  from  the  fire-brick.  The  richer  illuminants  pre- 
dominate, of  course,  and  this  rich  oil  gas,  mixing  with  the  blue 
water  gas,  results  in  carburetted  water  gas,  and  which  has  a  high 
candle-power  and  calorific  value.  By  varying  the  relative  quan- 
tities of  blue  gas  and  oil  gas,  and  the  heats,  time  of  run,  etc.,  the 
candle-power  and  heating  value  may  be  made  high  or  low,  as  de- 
sired. The  maximum  and  miuimum  limits  would  be  about  as 
follows :  With  no  blue  gas  and  all  oil  gas,  the  candle-power  would 
be  about  85,  and  the  calorific  value  about  1300  B,  t.  u.,  or,  with 
all  blue  gas,  and  no  oil  gaa  whatever,  the  candle-power  would  be 
practically  zero,  and  the  calorific  value  about  320  B.  t.  u.  Any 
intermediate  condition  could  be  attained,  but  in  practice  it  is  found 
that  a  gaa  exceeding  2G  to  30  candle-power,  burns  with  a  smoky 
flame  in  ordinary  burners,  under  usual  conditions,  and,  further- 
more, the  tendency  of  the  present  time  seems  to  be  towards  a 
standard  gas  of  an  average  of  about  600  B.  t.  u.  calorific  value. 

When  a  water-gas  set  has  been  making  gas  for  a  certain  number 
of  minutes  it  becomes  too  cool  for  economical  operation.  The  oil 
is  then  shut  off,  next  the  steam,  and  then  the  stack  valve  is  opened. 
Thereupon  the  blast  is  turned  under  the  fire  and  the  whole  cycle 
of  operations  is  repeated. 

On  account  of  the  steam  striking  the  under  side  of  the  fire  and 
cooling  it  too  rapidly,  it  is  now  customary  to  make  a  so-called 
"  down  run  "  every  third  or  fourth  time.  This  simply  means  that 
the  direction  of  flow  of  the  steam  through  tbe  fire  is  reversed,  now 

>y  Google 

306  iLLuinNATiNG  Enginseking 

paBsing  downvards  instead  of  up,  the  connectioiiB  od  the  machines 
being  so  arranged  as  to  permit  of  this  being  done.  As  often  as  the 
fire  requires  it,  fresh  coke  or  coal  is  put  into  the  generator,  the 
ashes  and  clinkers  being  taken  out  at  the  bottom. 

From  the  description  given  of  the  principles  involved  in  the 
Love  procesB  it  will  be  seen  that  it  is  essentially  an  intermittent 
one.  In  the  first  place  the  deep  fuel  bed  is  brought  up  to  a  high 
temperature  in  the  most  economical  way,  namely,  by  internal  com- 
bostion  in  the  generator,  and  in  this  way  differs  from  the  early 
processes  first  mentioned. 

Secondly,  it  differs  from  the  second  type,  or  that  promulgated 
by  Tesaie  dn  Motay,  in  making  use  of  the  heat  from  the  generator 
gases  to  vaporize  and  fiz  the  oil. 

These  differences  may  be  seen  from  the  basic  claim  of  the  Love 
patents,  which,  in  brief,  are  as  follows : 

Basic  Claim  Lowe  Patent 

The  apparatus  consistB  of  the  primary  gas  generator  A,  super- 
heater D,  heat-restoring  stack  I,  boiler  B,  the  usual  washer  T,  and 
scrubber  T. 

The  gas  generator  A  is  filled  with  anthracite  or  bituminous  coal, 
air  is  forced  by  a  blower  through  the  heat-restoring  stack  I  and 
pipe  L  into  generator  A  below  the  grate  bars,  having  been  pre- 
heated in  passing  through  stack  I. 

The  products  of  combustion  are  conducted  from  the  top  of 
generator  A  through  pipe  F,  through  the  superheater,  which  is 
filled  with  loose  fire-brick  above  the  arch,  to  the  atmosphere  through 
stack  I,  Valves  E'  and  H  having  previously  been  opened. 

The  heat  from  the  out-going  gas  is  partially  transferred  to  the 
air  from  the  blower,  which  is  forced  around  the  stack  tubes  into 
pipe  L.  After  the  fuel  in  the  generator  is  thoroughly  incandescent 
and  the  superheater  is  heated,  the  air  ie  cut  off  and  the  valves  E 
and  H  are  closed. 

Steam  is  now  admitted  into  the  top  of  the  superheater  through 
E*  from  boiler  R. 

The  steam  in  passing  through  superheater  becomes  intensely  hot, 
and  is  admitted  to  the  generator  below  the  grate  bars  through 
pipe  H'.  The  steam  in  passing  through  the  heated  carbon  is  de- 
composed, liberating  hydrogen  and  producing  a  proportionate  quan- 
tity of  GO,.    The  COj  in  passing  through  the  heated  carbon  is,  for 

>y  Google 

Manufagtdbe  and  Disteibutioit  of  Gas 

>y  Google 

308  Illuminatixo  Enoineebino 

the  most  part,  changed  to  CO,  and  the  gas  at  the  top  of  the  fuel 
bed  is  H,  CO  and  a  small  part  of  CO,. 

At  the  same  time  steam  is  admitted  to  the  superheater,  petroleum 
or  other  hydrocarbon  oils  are  introduced  in  regulated  quantities 
from  tank  M  on  to  the  top  of  the  hot  coals  in  the  generator,  where 
it  is  volatilized  and  mixes  thoroughly  with  the  gas  coming  through 
the  fuel  bed.  These  gases  arc  then  fixed  by  the  heat  before  leaving 
the  generator  from  which  they  pass  to  the  top  of  the  boiler  E 
through  numerous  tubes,  transferring  some  of  their  sensible  beat 
to  the  water.  All  of  the  steam  used  for  the  gas-making  process  is 
furnished  by  this  boiler,  and  the  heat  of  the  gas  is  the  only  energy 
used  for  generating  the  steam. 

Passing  through  the  boiler  the  gas  enters  the  washer  V,  thence 
through  the  scrubber  Y  into  the  purifiers,  and  finally  into  the 
holder  A'. 

It  should  be  stated  that  this  apparatus  never  worked  satiefac- 
toriiy  for  the  reason  that  the  oil  gas  was  not  subjected  to  suffi- 
cient heat  to  fix  it  into  a  permanent  gaa.  Mr.  Lowe  later  changed 
hia  method,  although  conforming  to  the  original  patent,  and  sub- 
stituted in  place  of  the  superheater  for  drying  and  superheating 
the  steam,  a  superheater  filled  with  checker-brick  properly  heated 
by  internal  combustion  in  the  superheater  of  the  producer  gases 
formed  in  the  generator  at  the  time  of  blasting  up  the  heats.  When 
making  gas  the  blue  water  gas  from  the  generator,  with  the  oil 
vapors  generated  at  the  top  of  the  generator,  pass  through  the 
superheater  for  the  purpose  of  fi.\ing  the  oil  vapors;  this  principle 
being  the  same  as  that  employed  in  all  water-gas -making  appa- 
ratus up  to  the  present  time. 

Returning  for  a  moment  to  the  original  table  giring  composition 
of  gases,  it  may  be  stated  that  the  illuminants  methane  and  ethane, 
result  from  gasifying  the  oil,  while  the  carbon  monoxide  and  hydro- 
gen result  from  the  action  of  steam  upon  the  incandescent  fixed 
carbon  in  tlie  generator  fuel.  The  balance  of  the  constituents  re- 
sult from  both  sources,  but  to  a  varying  extent. 

The  subject  of  the  efficiency  of  a  Lowe  water-gas  set  as  a  heat 
machine  may  be  stated  practically  about  60  per  cent.  The  aubject 
is  too  lengthy  to  be  discussed  here,  but  anyone  interested  is  referred 
to  a  paper  by  Mr,  A.  G.  Glasgow,  Proceedings  American  Gas  Light 
Association,  1890,  or  to  an  abstract  thereof  which  appears  in  the 
"  Mechanical  Engineers'  Pocket-Book,"  by  William  Kent,  under 
the  general  subject  of  illuminating  gas. 

>y  Google 


>y  Google 

310  IiiLUMTNATiNa  Ehginbbbino 

In  general,  we  may  use  the  following  average  figures  to  illustrate 

the  e£5ciency  of  a  Lowe  water-gae  set,  all  pei  thoueand  cubic  feet 

gaa  made  reduced  to  60°  F.: 

Pounds  anthracite  generator  fuel 30 

Founds  oil  admitted  to  carbureter 33 

Founds  steam  used  during  run 30 

Founds  resulting  gas  produced 46 

This  serves  to  introduce  the  principles  of  water-gas  manufacture, 

and  we  will  now  discuss  the  subject  of  treatment  of  this  gas  after 

leaving  the  generating  apparatus. 

Purification  of  Water  Oaa 

Coming  to  the  subject  of  impurities,  we  find  that  tar  and  sulphur 
are  the  predominating  ones  that  must  be  abstracted.  Before  treat- 
ing these,  however,  it  is  to  be  remarked  that  water  gas  is  to  be 
condensed,  in  a  measure,  similar  to  coal  gas.  Water  gas,  however, 
in  modem  practice,  is  not  reduced  to  as  low  a.  temperature  in  the 
works  as  is  coal  gas. 

The  principle  of  water-gas  condensation,  however,  is  the  same 
as  for  coal  gas.  The  heat  to  be  abstracted  consists  of  the  sensible 
heat  plus  the  latent  heat  of  vaporization  of  the  various  gases  and 
vapors  which  compose  the  gas.  This  results  in  deposition  of  some 
of  the  heavier  hydrocarbons,  forming  the  so-called  vater-gas  tar. 

In  modem  practice  water  gas  is  seldom  condensed  below  90°  F., 
because  its  purification  is  most  economical  at  this  or  somewhat 
higher  temperatures,  and  also  because  more  of  the  richer  illumi- 
nants  remain  in  the  gas  at  the  higher  temperatures.  A  large  amount 
of  condensation  takes  place  in  the  relief  holder. 

fTaphthalene  is  easily  avoidable  in  water-gas  practice  by  proper 
regulation  of  the  heats. 

Tar  is  extracted  from  water  gas  by  condensation,  washing  and 
scrubbing,  and  also  by  mechanical  means,  such  aa  a  F.  &  A  tar 
eitraetor.  With  the  oily  water-gas  tar,  however,  the  P.  &  A,  must 
be  operated  between  rather  narrow  limits  of  temperature,  say  be- 
tween 105°  and  110°  F.,  and  under  great  differential  preesure. 

Usually,  after  all  the  washing  and  scrubbing,  there  remains  a 
mist  of  light  tarry  vapors  which  are  exceedingly  difficult  to  ex- 
tract. This  is  perhaps  best  accomplished  by  means  of  shaving 
scrubbers,  in  which  light  wood  shavings  simply  absorb  the  mist 
as  the  gas  slowly  passes. 

>y  Google 

Mandtaotdee  and  Dibtbibution  op  Gas  311 

Sulphur  exists  agaia  as  H,S  and  organic  sulphur,  and  is  usually 
removed  bj  means  of  iron  oxide  as  described  under  coal  gas.  In 
coal  gas  the  purification  is  natially  carried  on  under  lower  tempera- 
tures than  in  water  gas,  because  in  coal  gas  the  gas  ia  previously 
reduced  to  a  low  enough  temperature  to  permit  the  extraction  of 
the  i 

Carburetted  Water  Gas  as  Made  from  Oil  and  Steam  Only 

Lowe  Oil  Oas.     There  is  time  here  only  for  a  brief  mention  of 

carburetted  water  gas  as  made  from  oil  and  steam  only.     This 

process  is  more  largely  used  on  the  Pacific  elope  on  account  of 

the  low  cost  of  oil  and  the  high  cost  of  coal  and  coke. 

Fio.  16. 

The  development  of  this  oil-gas  process  is  due  to  the  efforts  of 
Mr,  Lowe,  as  well  as  largely  to  Mr.  E,  C,  Jones,  Chief  Engineer 
of  The  San  Francisco  Gas  &  Electric  Company. 

The  principles  underlying  the  maniifacture  of  gas  by  this  method 
are  unique  in  a  way.  No  standard  type  of  apparatus  has  been  de- 
veloped, but  there  are  various  forms  of  one-shell  and  two-shell 
types  in  use  on  the  Pacific  coast  to-day. 

These  shells  are  of  iron,  lined  with  fire-brick  and  checkered  with 
fire-brick.  To  heat  up  the  set  oil  is  introduced,  or  sprayed  in  with 
a  steam  spray,  and  burns  by  means  of  an  air  blast,  the  products  of 
combustion  passing  off  through  a  stack  valve  in  the  usual  manner. 

>y  Google 

312  iLLUMiNATtya  Enqinberino 

When  the  set  is  up  to  heat  the  air  blast  is  cut  off,  and  the  oil  and 
steam  admitted  alone.  An  accurate  adjuBtment  of  the  quantity 
of  oil  to  the  heat  is  necessary  for  best  results. 

The  oil  gasifies  under  the  heat  of  the  fire-brick,  and  the  steam 
is  partially  deeompoRed  into  its  elements.  Some  of  the  heavier 
illuminaiits  arc  decomposed,  and  considerable  free  carbon  or  lamp- 
black results.  The  gas  produced,  as  will  be  seen  from  the  early 
tables,  resembles  coal  gas  very  much  in  its  analysis. 

The  impurities  to  be  removed  from  this  oil-f;as  process  are  lamp- 
black, tar  and  sulphur.  The  lamp-black  removal,  handling  and 
treatment  is  a  problem  in  itself,  but  it  is  removed  from  the  gas 
by  washing  with  copious  quantities  of  water,  and  by  scrubbing,  and 
is  subsequently  fired  under  the  boiler  in  a  wot  state,  or  it  can  be 
used  as  generator  fuel  in  an  ordinary  water-gas  set. 

The  tar  and  sulphur  are  removed  in  the  customarv  ways.  Oil 
gas,  as  made  abo\e,  is  trtattd  much  like  ordinary  water  gas,  except 
it  is  never  passed  through  condensers,  but  is  subjected  to  much 
washing  and  scrubbing  This  process  of  treatment  at  once  appeals 
to  anyone  as  being  logical,  on  account  of  the  large  quantities  of 
lamp-black  made  during  ita  generation 

Under  conditions  of  best  practice  to-day,  this  process  of  gas 
manufacture  requires  about  a  total  of  7  to  S  gallons  of  oil  per 
1000  cubic  feet  made,  and  there  is  every  likelihood  that  this  quan- 
tity will  be  materially  reduced.  From  general  figures  it  would  seem 
that  only  about  2  gallons  of  oil  should  be  necessary  to  supply  the 
required  amount  of  heat,  and  if  we  figure  an  average  of  4^/^  gal- 
lons for  making  the  gas.  it  would  seem  as  though  from  6  to  6^ 
gallons  will  ultimately  be  all  that  is  required  for  this  process.  Re- 
cent results  indicate  that  these  figures  may  he  attained. 

Producer  Gas 

Producer  Gaa.  Producer  gas  is  usually  made  by  one  or  both 
processes  already  e.vplainod  under  coal-  and  water-gas  manufacture. 
In  some  forms  it  consists  of  CO  and  X,  produced  by  air  being 
blown  through  a  bed  of  incandescent  fuel,  the  resultant  gas  having 
a  calorific  value  of  about  120  to  i;tO  B.  t,  u.  per  cubic  foot.  If,  in 
addition  to  air,  we  add  steam,  the  resultant  gas  will  contain  H, 
CO  and  N.  If  steam  alone  is  used  the  gas  will  consist  of  H  and  . 
CO,  and  will  have  a  heating  value  of  about  320  B.  t.  u. 

Gas,  as  an  agent  for  the  production  of  light  and  heat,  must  not 
be  understood  to  be  restricted  to  artificial  gas,  as  before  outlined. 

>y  Google 

Manufactl'ke  and  D18THIBUT10N  OF  Gas  313 

but  many  other  forms  besides  these  mentioDed  are  used,  such  as 
retorted  oil  gas,  blast-furnace  gas,  acetylene,  gaaolene  air  gas, 
resin  gas,  wood  gas,  hydrogen -methane  gas,  garbage  gas,  etc. 

Producer  gas  is  only  mentioned  at  this  time  on  account  of  its 
adaptation  to  gas-engine  practice. 

Station  Mlter  Drum. 
Pm.  17. 

Metering  (Jus  at  Works — The  Station  Meier 

Station  Meter.     The  gas  after  passing  through  the  purifiers  is 

ready  to  sell,  except  that  the  amount  made  must  be  determined 

in  order  to  keep  the  several  parts  of  the  works  under  control.    This 

>y  Google 

314  Illuminating  Engineeeino 

meaenhng  is  nsnally  done  bj  meanB  of  a  large  fonr-compaitment 
drum  which  revolves  in  a  east-iron  ease  filled  about  two-thirds 
full  of  water. 

The  inlets  and  outlets  of  the  drum  compartments  are  so  ar- 
ranged that  when  the  outlet  is  below  water  the  inlet  is  above,  and 
the  compartment  fills  with  gaa.  The  drum  revolves  something  like 
a  squirrel  cage,  and  shortly  after  the  inlet  dips  below  the  water  the 
outlet  comes  above  and  the  compartment  discharges  its  contained 
gas.  The  cubical  contents  of  the  compartments  being  ao^rately 
known,  the  motion  of  the  drum  is  communicated  by  gearing  to 
the  dial,  and  thus  we  have  an  apparatus  which  accurately  measures 
the  gae  made.  It  is  customary  to  make  proper  corrections  for  tem- 
perature and  barometric  pressure,  and  in  practice  we  reduce  t^e 
gas  manufactured  to  a  basis  of  60°  F.,  and  30  inches  barometric 

On  account  of  the  large  size  of  station  meters  various  forms  of 
proportional  meters  have  been  tried.  These  measure  only  a  small 
fraction,  usually  1  per  cent,  of  the  make,  and  are  also  arranged  to 
register  the  total,  but  so  far  there  is  really  no  reliable  proportional 
meter  on  the  market  for  measuring  artificial  gases. 

Kecently  various  other  methods  of  measuring  gas  have  been 
tried.  Drums  have  been  made  of  the  rectangular  screw-thread  type, 
rotary  meters  have  been  introduced,  and  the  most  recent  is  the 
electric  gas  meter.  The  time  is  too  limited  to  attempt  to  explain 
these  in  detail. 

Oas  Solders 

Gas  holders  are  simply  inverted  cups  placed  in  water,  and  so 
arranged  that  the  gas  enters  or  leaves  the  holder  above  the  water 
through  pipes  arranged  for  the  purpose.  They  act  as  storage 
reservoirs  for  gas,  and  thus  allow  the  plants  to  manufacture  uni- 
formly during  the  34  hours,  taking  care  of  constantly  varying  con- 
sumption. The  only  principles  involved  are  as  given,  and  the 
great  questions  involved  in  connection  with  gas  holders,  outside 
of  their  design  and  construction,  are  "  How  much  gas-holder  ca- 
pacity is  required  as  related  to  the  maximum  manufacturing  ca- 
pacity ?  "  and  "  Where  shall  these  holders  be  located — at  the  gas 
works  or  in  other  localities?" 

The  latter  question  is  largely  a  matter  of  distribution  methods, 
and  the  former  the  question  of  the  minimum  permissible  holder 

>y  Google 

Mandfaotukb  and  Disteibution  of  Ga8  315 

capacity  any  plant  may  have  and  be  safe.    This  is  a  very  important 
engineering  and  commercial  detail. 


In  the  distribution  we  have  a  vast  subject,  and  one  in  which 
many  problems  remain  to  be  solved. 

Formerly  gas  was  sent  out  from  the  works  at  not  to  exceed  the 
maiimum  preseure  thrown  by  the  works  holders.  In  such  systems 
the  delivery  obtainable  from  a  given  size  and  length  of  pipe  was 

that  due  to  the  difiEerential  head  oi  pressure  between  the  lowest 
permissible  pressure  at  the  extreme  end  of  the  pipe,  say  2  inches 
water  column,  and  the  maximum  holder  pressure,  which  rarely  ex- 
ceeded 5  or  6  inches.  Thus  the  actuating  pressure,  or  differential 
head,  was  very  little,  possibly  less  than  one-tenth  of  1  pound  per 
square  inch. 

Subsequently,  as  cities  spread  out,  gas  holders  were  erected  in 
outlying  localities  and  called  district  holders,  and  these  supplied 
the  district  surrounding  them,  as  before,  by  low  pressure.  These 
district  holders  were  filled  with  gas  through  separate  pumping 
mains  from  the  works. 

In  the  course  of  time  these  systems  were  found  to  be  inadequate, 
and  if  re-enforced  under  the  low-pressure  ideas  would  have  entailed 

>y  Google 

316  Illcminatisg  Engimbeking 

vast  construction  eKpenditures  to  remedy  conditions  sufficiently 
to  produce  good  service. 

To  overcome  bad  distribution  systems,  especially  in  the  larger 
cities,  the  next  stpp  in  the  progress  of  distribution  methods  was  to 
erect  pumping  plants  at  the  works  and  holders,  run  separate  pres- 
sure re-cnforcing  pipe  lines  to  the  heavy  points  of  consumption, 
and  there  install  some  device  to  automatically  reduce  the  pressure 
to  the  required  regular  distribution  pressure.  In  some  cases  pres- 
sure-indicating instruments  were  located  at  such  points  of  hea^y 
consumption,  and  no  regulators  used,  the  pressure  and  amount  of 
gas  pumped  being  controlled  at  the  works  and  holders  so  that  a 
given  pressure  was  maintained  at  this  point  where  the  indicator 
was  located.  The  instruments  transmitted  the  amount  of  pressure 
back  to  the  pumping  plant,  or  a  small  separate-pressure  tell-tale 
line  was  used.  These  systems  used  various  pumping  pressure, 
usually  not  over  5  pounds  per  square  inch. 

In  the  meantime  still  another  development  was  taking  place. 
Communities  were  growing  and  spreading  out  in  all  directions,  and 
especially  around  the  larger  cities  where  suburban  communities 
were  being  formed  at  some  distance  from  the  cities,  and  in  which 
the  houses  were  far  apart.  It  was  not  possible  to  profitably  supply 
such  places  with  gas  with  the  great  investment  required  in  low- 
pressure  mains  under  the  old  system,  so  high  pressure  was  devel- 
oped to  meet  this  requirement. 

Pressures  up  to  50,  CO  and  even  80  pounds  per  square  inch  are 
now  being  used,  as  compared  to  the  old  system  of  low  pressure, 
with  a  maximum  of  about  Vj  pound.  Such  high  pressure  requires 
reduction  to  say  4  or  6  inches  water  column  before  entering  the 
piping  in  the  consumer's  building,  and  this  is  accomplished  by 
automatic  gas  regulators,  a  number  of  different  types  of  which 
are  now  on  the  market. 

Reaiona  for  High  Freuiire.  In  the  meantime  other  forces  were 
at  work  tending  to  hasten  the  advent  of  high  pressure.  The  uses 
to  which  gas  was  applicable  were  increasing  in  number,  it  was 
also  used  more  freely  in  lighting,  heating  and  power  work,  and 
this  resulted  finally  in  very  much  larger  sales  of  gas  per  capita 
per  annum  than  prevailed  formerly.  Add  to  this  the  fact  that 
the  price  of  gas  was  gradually  being  reduced  and  another  stimulus 
is  seen.  Thus  vast  quantities  of  gas  were  being  consumed  as  com- 
pared to  former  years. 

>y  Google 

Manufacture  and  Distribution  of  Gas  317 

The  principles  underlying  the  development  of  high  pressure  then 
resolve  themselves  into  the  fact  that  it  was  necessary  to  provide 
for  vastly  increased  demand,  and  also  that  the  distances  throufrh 
which  it  was  necessary  to  supply  gas  in  lar^e  quantities  were  greatly 

This  development  was  not  rapid  in  the  early  days  of  the  gas 
business,  and,  in  fact,  it  may  be  said  to  have  developed  with  the 
advent  of  fuel  gas,  the  posBibilitJes  of  which  have  only  been  realized 
within  the  most  recent  years.  In  fact,  it  may  be  said  that  the 
application  of  high  pressure  to  artificial  gas  is  a  development  of 
the  last  decade. 

Higher  pressure  permits  of  small  pipes  to  transmit  large  quan- 
tities of  gas.  The  reasons  for  this  are  that  the  differential  head 
is'  very  much  greater  than  under  low  pressure,  and  also  a  given 
mass  of  gas  occupies  a  much  smaller  space  when  compressed. 
■  The  flow  of  gas,  or  any  liquid  through  pipes,  is  governed  by  the 
differential  head  or  effective  driving  pressure,  the  length  of  the 
pipes,  its  diameter,  the  condition  of  its  interior  surface,  whether 
the  line  is  straight  or  full  of  turns,  the  density  of  tlie  traversing 
gas  or  fluid,  and  the  questions  of  pulsations,  obstructions,  etc. 

Fonanlae  for  Pipe  Condnotivity.  Various  formulae  have  been 
devised  to  determine  the  flow  of  gas  in  pipes,  but  the  one  com- 
monly used  for  low  pressure  is  Dr.  Pole's  formula. 


Q  =  quantity  of  gas  in  cubic  feet  per  hour  at  atmospheric  pressure. 

c  =  a  factor,  which  may  vary  from  1000  to  1400,  but  a  fair  average 
value  for  which  is  1350.  This  factor  is  inserted  for  the 
purpose  of  allowing  for  condition  of  the  interior  pipe  sur- 
face, obstructions,  such  as  tar,  etc. 

d^diameter  of  pipe  in  inches. 

h=differential  head,  or  pressure,  in  inches  of  water. 

8=specific  gravity  of  gas,  air  being  1. 
1  =  length  of  pipe  in  yards. 

From  this  formula  it  appears  that  the  capacity  of  a  pipe  to  trans- 
mit gas  under  low-pressure  conditions,  among  other  factors,  varies 
as  the  square  root  of  the  fifth  power  of  the  diameter.  As  a  result 
of  this  it  may  be  stated  that  when  a  pipe  is  doubled  in  diameter 
its  capacitj'  under  low  pressure  is  multiplied  about  5.6  times. 

>y  Google 

318  Illdhinating  Enginbbring 

Pop  Hi^  Prennre.    The  following  formula  covers  the  range  of 
high-pressure  artificial  gas: 
Q-^aa.t    /d'(p.'-p.') 

y      Ls 

Q=qiiantitj  of  gas  in  cubic  feet  per  honr  at  atmospheric  pressare. 

d=diameter  of  pipe  in  inches. 

Pi=ah8olute  initial  presBure  in  pounds  per  square  inch. 
Tp,=dbsotiite  terminal  pressure  in  pounds  per  square  inch. 

L=length  of  pipe  in  miles. 

s  =  specific  gravity  of  gas,  air  being  1. 

Tor  Tery  Hi^-Piessuxe  and  long  Pipe  Lines.  The  formula  for 
ordinary  high-pressure  work,  previously  given  for  use  with  artificial 
gas,  is  found  to  give  results  that  are  too  small  when  applied  to  a 
higher  range  of  pressure  and  long  pipe  lines.  In  particular,  for 
natural-gas  work,  where  pipe  lines  many  miles  in  length  are  in 
use,  it  is  found  in  practice  that  more  satisfactory  results  are  se- 
cured from  the  following  formula : 

Q=42  a  J'Pi*TPl 

Q=:quantity  of  gas  in  cubic  feet  per  hour  at  atmospheric  pressure. 
a=a  factor,  which  in  practice  is  found  to  vary  with  the  diameter 
of  the  pipe,  and  for  which  fairly  satisfactory  amounts  have 
been  determined.     For  instance,  a=95  for  a  6-inch  pipe, 
556  for  a  12-inch,  etc.    See  Ohio  Geological  Survey  report, 
Pi=db3oltite  initial  pressure  in  pounds  per  square  inch. 
f>j=ahsoluie  terminal  pressure  in  pounds  per  square  inch. 
L=length  of  pipe  in  miles. 

This  last  formula  is  based  upon  a  gas  of  0.6  specific  gravity. 
Where  the  gravity  of  the  gas  varies  the  quantity  found  is  multi- 
plied by  the  square  root  of  0.6  divided  by  the  gravity  determined. 
Temperature  corrections  are  usually  neglected  in  nataral-gas 

ElcTstion.  In  the  olden  days  the  question  of  elevation  was 
pertinent.  Gas,  being  lighter  than  air,  in  a  confined  pipe  tends 
to  exercise  greater  pressure  at  higher  elevation,  as  compared  to 
the  atmosphere,  because  it  weighs  less  than  the  equivalent  column 
of  air  under  the  condition  of  being  exposed  to  atmospheric  pres- 
sure at  the  initial  low  point,  as,  for  instance,  through  a  gas  holder. 

>y  Google 


When  gas  was  distributed  entirely  nndet  low  pressures  some  points 
of  a  given  city  lying  much  below  the  level  of  the  works  received 
ineufficient  pressure,  and  other  points,  mnch  above  the  works,  re- 
ceived excessive  pressure. 

Recently,  however,  where  high  pressure  is  used,  the  question  of 
elevation  causes  no  concern  because  of  its  comparatively  slight  ef- 

Fio.  19.~Station  Governor. 

feet  under  such  conditions.  Under  low-pressure  conditions,  and 
with  gas  of  ahont  six-tenths  specific  gravity,  the  difference  in  pres- 
sure due  to  100  feet  elevation  is  about  six-tenths  inches  water 

Station  QoTernor.  A  station  governor  is  an  apparatus  which 
automatically  maintains  a  given  outlet  pressure,  which  must  be 
less  than  the  inlet  pressure.    This  is  simply  produced  by  the  effect 

>y  Google 

320  Illumikatixo  En'oinbebing 

of  the  outlet  pressure  on  a  float  or  a  diaphragm.  Some  governors 
have  been  devised  which  increase  or  decrease  the  pressure  auto- 
maticallj  according  to  the  drmand. 

Principles  of  Design  of  a  Dislribution  System 

We  will  first  consider  the  principles  underlying  the  design  of 
a  low-preBflure  distribution  system. 

Under  this  kind  of  a  system  we  are  limited  to  the  maximum 
pressure  allowable  on  consumers  near  the  plant  or  holders,  and  by 
the  minimum  pressure  allowable  on  the  outlying  consumers.  For 
'  purposes  of  illustration,  assume  this  maximum  and  minimum  to 
be  6  and  2  inches,  respectively.  Then  the  maximum  differential 
head  is  3.8  inches,  allowing  0.2  inches  drop  in  services. 

Neit,  having  a  complete  map  of  the  city,  it  is  necessary  to  deter- 
mine the  maximum  demand  per  unit  of  area,  which  for  purposes 
of  illustration  we  may  assume  as  1  square  mile,  and  having  selected 
the  center  of  each  square  mile,  we  proceed  to  run  low-pressure 
feeders  from  the  works,  in  several  directions  if  necessary,  and  large 
enough  to  furnish  all  the  gas  required  at  peak  load  to  each  unit 
of  area  reached  by  such  main,  and  under  the  limitations  of  pres- 
sure assumed.  If  we  determine  that  the  loss  of  pressure  from 
the  center  of  each  unit  of  area,  to  the  outside  limits  thereof  at 
peak  load,  shall  not  exceed  1  inch,  then  the  maximum  drop  in  pres- 
sure in  the  feeders  from  the  holder  outlets  must  not  exceed  2.8 
inches  to  come  within  our  assumed  limits. 

On  the  basis  of  this  assumption  we  are  thereupon  obliged  to 
design  the  distribution  system  in  each  unit  of  area  so  that  at  peak 
load  the  maximum  drop  in  pressure  from  the  center,  or  point  of 
supply  from  the  feeder  mains,  to  the  farthest  outlying  point  in 
each  area  shall  not  exceed  1  inch  at  peak  load  to  come  within  our 
required  assumed  conditions. 

To  do  all  this  requires  the  knowledge  of  maximum  demand  per 
consumer,  the  probable  maximum  number  of  consumers  per  block 
and  per  unit  of  area,  the  length  of  blocks,  and  certain  other  prac- 
tical considerations,  such  as  presence  of  electric  surface-car  line 
tracks,  etc. 

Naturally,  smaller  and  simpler  systems  for  smaller  cities  are 
easier  to  design,  but  the  principle  of  maximum  permissible  drop 
in  pressure  is  the  same. 

>y  Google 

Manufacture  and  Disthibction  of  Gas 


Bes^  for  High  PrcMlire.  When  we  eorae  to  consider  high- 
pressure  systems  the  same  general  principles  hold  true.  We  may 
run  high-pressure  feeders  to  the  centers  of  the  units  of  area,  or 
we  may  design  them  to  carry  only  moderate  pressure,  say  up  to 
5,  6  or  S  pounds  per  square  inch.  If  such  a  system  is  adopted, 
it  becomes  necessary  to  install  pressure- reducing  devices  at  the 
points  where  the  high-pressure  feeders  deliver  gas  into  the  low- 
pressure  system.    Such  devices  are  called  district  regulators. 

Fra.  20. — Section  of  Manhole  on  G-1b.  Hlgti-PreBSure  Line. 

Another  entirely  different  system  is  to  carry  moderato  or  high 
pressure  on  the  entire  system  of  mains.  In  such  cases  it  is  neces- 
sary to  install  pressure-reducing  devices  on  each  pipe  entering  each 
and  every  consumer's  premises  to  reduce  the  main-pipe  pressure, 
whatever  it  may  be,  to  the  pressure  required  hy  .the  consumer. 
Such  devices  are  called  individual  gas-pressure  regulators  or  gov- 

The  advantage  of  the  use  of  high  pressure  lies  in  the  fact  that 
much  smaller  distributing  pipes  can  be  used,  thus  saving  great 
investment  charges.  The  cost  of  compressing  gas  is  generally  a 
small  item. 

>y  Google 



Drainage  of  Kaiai.  Artificial  gas^  as  it  leaves  the  vorke,  always 
contains  vater  vapor  and  various  hydrocarbon  vapors,  which  con- 
dense out  of  it  as  it  passes  through  the  distributing  pipes,  owing 
to  changes  of  temperature  and  other  causes.  These  vapors  con- 
dense and  liquefy,  forming  the  so-called  drip  water.  For  thie 
reason  it  is  necessary  to  lay  artificial  gas  pipes  on  a  slight  grade, 
and  at  the  low  points  devices  for  collecting  this  drip  water  are 
installed  so  that  it  may  be  pumped  out. 

Fia.  21. — High-Pressure  Main, 

A.  9i"  saddle  with  5/16'  main  top 


B.  %"  corporation  cock  with  Vt" 


C.  %'  street  tee  (galvanized). 

D.  \"  street  ejl   (galvanized). 

B.  ^'    curb    coclc   wltb    full    gas 

F.  %'  street  ell  (galvanized). 

G.  %'  tee  (galvanized). 

H.  ^i"  meter  cock  with  5/16"  gas 

J.   Hlgb-preBsure     governor     (see 

K.  Cross. 

Meter  and  Drip  Installation. 

U    1'  e!l. 

M.  1'  vent  trom  saletjr  seal  (end 

protected  with  No.  16   wli% 

N.  %"  ell. 
0.  Reducing  ell. 
P.  Reducing  ell. 
Q.  1"  ell. 
R.  To  riser. 

S.    Mercury  seal   (see  schedule). 
T.  1"  X  %-  tee. 
U.  %"  long  screw. 
V.  %-  ell. 

W.  %"  vent  from  regulator. 
X.  1-  ell. 

>y  Google 

MaNCPACTURE   and   DiSTBlBUTION   OF   Gas  323 

Haterials  and  Jointa.  For  low-presBure  distribution,  and  in  the 
larger  cities  it  is  customary  to  use  cast-iron  for  the  main  pipes 
on  account  of  its  long  life,  lesigtance  to  corrosion  and  to  electrolytic 
action.  The  joints  are  almost  universally  of  the  bell  and  spigot 
type,  in  cast-iron  mains,  and  the  jointing  material  is  either  lead 
or  cement,  caulked  or  placed  into  the  joint  against  an  inside  roll 
of  jute  packing  to  prevent  it  from  entering  the  pipes. 

Such  joints  are  not  conceded  to  be  safe  at  high  pressure,  and 
when  wrought-iron  pipe  is  used  screw  or  threaded  joints  are  used. 

On  account  of  the  mechanical  strength  of  cast-iron,  it  is  to-day 
the  general  practice  to  use  but  little  pipe  smaller  than  i-inch  cast- 
iron  pipe  for  gas  distribution,  so  that  under  that  size  wrought  pipe 
is  employed.  Under  6-inch  pipe  the  wrought  is  usually  cheaper 
in  Srst  cost  than  cast-iron  pipe.  Pipes  are  usually  much  stronger 
than  required  to  merely  resist  the  internal  pressure.  External  con- 
ditions, such  as  pressure  of  soil,  loads,  settlement,  corrosion,  etc., 
are  the  factors  which  determine  the  minimum  permissible  thickness 
of  pipes. 

Special  types  of  pipes  and  joints  have  at  various  times  been 
brought  forth,  such  as  Universal,  vitrified  clay,  and  even  wood  has 
been  used,  but  to  a  very  small  extent. 

Gas  mains  are  usually  laid  deep  enough  to  be  under  the  frost 
line,  and  are  kept  away  from  car  tracks  and  underground  obstruc- 
stmctions  as  much  as  possible.  Services,  or  the  pipes  leading  from 
the  mains  to  the  consumers'  premises  wherever  possible,  are  graded 
into  the  mains. 

Electric  surface-car  lines  have  proved  a  bug-bear  to  underground 
piping  systems  on  account  of  electrical  current  leakage  setting  up 
an  electrolytic  action.  A  portion  of  the  return  current  from  such 
car-line  systems  finds  its  way  into  the  piping  and  leaves  it  again 
usually  at  some  point  near  the  generating  or  substations,  or  where 
it  jumps  to  some  other  conductor.  The  troubles  occur  where  the 
current  leaves  the  pipes. 

Various  remedies  have  been  suggested  and  tried,  such  as  double 
systems  of  piping,  one  on  each  side  of  the  car  tracks,  also  various 
forms  of  insulated  pipe  covering  and  joints,  also  bonding  the  pipes 
to  the  rails  or  to  the  return  conductors.  All,  so  far,  have  proven 
to  be  more  or  less  in  the  nature  of  palliatives  and  not  complete 
remedies.     The  subject  of  electrolysis  is  one  of  great  importance. 

As  I  have  used  more  than  the  time  allotted  me,  I  shall  not  take 

>y  Google 

384  Illumixatixg  Engineering 

up  the  subject  of  the  gas  meter,  the  instrument  employed  for 
measuring  the  amount  of  gas  used  by  the  consumer,  or  house  piping 
or  photometry,  as  I  understand  some  of  the  subsequent  lectures 
will  incorporate  about  all  there  is  to  be  said  upon  these  subjects. 
I  would  like  to  say  a  few  words  regarding  calorimetry.  Owing 
to  the  fact  that  by  far  tlie  greater  proportion  of  gas  sold  to-day 
is  sold  as  a  heating  agent,  either  through  fuel  appliances  or  through 
mantle  burners,  it  seems  necessary  to  change  our  system  of  meas- 
uring quality  to  one  that  will  define  the  calorific  value.    This  may 

be  determined  in  two  ways,  first,  from  the  chemical  analysis  gas, 
as  the  heating  value  of  its  coustitueiits  are  pretty  well  known. 
There  has  been,  however,  adopted  for  quite  general  use  an  instru- 
ment whose  essential  principle  of  operation  is,  that  the  products 
of  combustion  of  a  gas  shall  be  passed  through  a  vessel  which  is 
water-jacketed,  and  in  which  the  radiated  heat  from  the  flame  and 
the  sensible  heat  from  these  products  of  combustion  are  absorbed 
by  water  in  tlie  jacket.  The  quantities  of  gas  and  water  being 
known,  the  rise  in  temperature  furnishes  a  measure  of  the  amount 
of  heat  liberated  by  the  combustion  of  that  amount  of  gas. 

>y  Google 

VII  (2) 
By  Walter  R.  Addicks 
The  subject  of  this  lecture,  "  The  Monufacture  and  Distribution 
of  Artificial  Gas,  witli  special  reference  to  Lighting",  is  so  com- 
prehensive that  it  is  difficult  to  outline  the  field  without  slighting 
essential  features  of  the  gas  business   covered  by   the   assigned 
The  following  sub-divisions  are  made  to  facilitate  rcfei'ence. 

(A)  Quality  of  Artificial  Gas. 

(B)  Purity  of  Artificial  Gas. 

(C)  Uses  of  Artificial  Gas. 

(D)  Kinds  of  Artificial  Gas  (including  Xatural  Gas  for  corn- 
pa  riaon). 

(E)  How  Artificial  Gas  is  manufactured. 

(F)  The  handling,  within  the  gas  plant,  of  raw  materials,  of  by- 
products, and  of  the  finished  product.  Artificial  Gas.  The  Retort 
Coal  Gas  Process  described  for  illustration,  with  some  reference  to 
an  auJriliarv  carburetted  water  gas  plant  useful  for  enriching  coal 
gas,  for  utilizing  the  colte  by-product  of  the  Retort  Coal  Gas  Plant, 
and  caring  for  variation  in  the  daily  demand  for  gas. 

(G)  Distribution  of  gas  from  Storage  Holder  at  Plant  through 
transfer  mains  to  the  City  Distribution  Holder. 

(H)  tliatribution  of  gas  from  Distribution  Holder  through 
Street  Main  System  to  the  gas  service  pipof  leading  to  the  houses. 

(I)  Distribution  of  the  gas  from  the  Street  Mains  through  gas 
aerrice  pipes,  house  service  pipes,  meters  and  governors  to  appli- 
ances for  utilizing  the  gas. 

(K)  Observations  relating  to  the  piping  of  modem  buildings  and 
its  relation  to  other  utilities  in  use. 

(L)  Observations  relating  to  the  appliances  used  in  burning  gas. 

(M)  Influences  that  govern,  in  the  selection  of  a  particular  type 
of  gas,  in  a  given  geographic  location. 

(X)  The  future  of  the  Artificial  Gaa  business. 

>y  Google 

326  Illuminating  Enoineebino 

A.  Quality  of  Artificial  Qat 
Gas  should  no  longer  be  manufactured  with  special  reference 
to  lighting  alone;  it  must  still  be  designated  by  its  candle  power, 
where  State  laws,  special  and  general,  define  quality  ag  the  candle 
power  given  by  a  specified  quantity  of  gas  burned  throagh  a  flat 
flame  or  argand  bnmer.  The  eame  quantity  of  gas  bnmed  by 
means  of  a  bunsen  burner  as  a  heating  fiame  in  contact  with  the 
Welsbach  mantle  will  give  four  times  the  light.  It  is  quite  common, 
in  describing  an  artificial  gas,  to  say  that  it  is  a  IG,  18,  or  ?0 
candle  power  gae,  meaning  that  when  a  specified  quantity  of  gae  is 
burned  in  a  specified  burner  tliat  it  will  give  16,  18,  or  SO  units  of 
light  when  compared  with  the  original  unit  of  light,  the  candle. 

B.  Purity  of  Artificial  Qas 
It  is  required  in  many  States  that  manufactured  gas  shall  be 
free  from  sulphuretted  hydrogen,  and  contain  but  limited  quan- 
tities of  ammonia  and  fixed  sulphur.  Such  laws  are  quite  unnec- 
essary for  the  reason  that  the  extending  use  of  electricity  will  com- 
pel commercial  purity  in  gas. 

C.  Uses  of  A  rtificial  Qas 

Artificial  gas  is  used  for : — 

(la)  lAgbitBg  by  means  of  the  flat  flame  or  argand  burner. 

(lb)  L^htin;  by  means  of  heat  generated  by  the  gas  when 
burned  in  a  Bunsen  burner  to  a  blue  flame  and  making  incandes- 
cent the  fabric  of  the  gas  mantle. 

(S)  Heating  through  the  use  of  the  Bunsen  fiame  in  gas  ranges 
for  cooking,  in  a  multitude  of  iudustrial  appliances  increasing  day 
by  day,  and  in  steam  boilers. 

(3)  Power  by  means  of  the  internal  combustion  engine,  made 
familiar  to  all  by  the  introduction  of  the  automobile. 

D.  Kinds  of  Artificial  Oaa  {Including  Natural  Qas  for 
Artificial  Oases  are  known  as:- — 

(1)  Water  Oas,  an  odorless  gas,  containing  Hydrogen  and  Car- 
bonic Oxide,  giving  a  non-luminous  fiame  when  ignited;  is  no 
longer  distributed.    It  must  not  be  confused  with 

(2)  Carburetted  Water  Oat  which  is  a  mixture  of  water  gas  and 
oil  gas  having  a  distinct  and  pungent  gas  odor,  and  when  burned 
gives  a  brilliant  white  fiame. 

>y  Google 

Manufacture  and  Distribction  of  Gas  327 

(3)  Retort  Coal  Ou,  a  gas  of  lower  speciJic  gravity  and  less 
brilliaDt  flame  than  Carburetted  Water  Gas. 

(4)  Coke  Oren  Qu,  similar  in  all  respects  to  Retort  Coal 
Gas;  only  35%i  to  50%  of  the  gae  made  is  distributed,  the  portion 
distributed  is  usually  of  equal  candle  power  to  Retort  Coal  Gas. 
The  remainder  of  the  gas  is  burned  under  the  ovens  in  place  of  coke. 

,  (fi)  Oil  flaa,  a  heavy  petroleum  gas  which  when  burned  in  prop- 
erly constructed  burners  gives  a  bright  light.  The  California  Oil 
Gaa  distributed  on  a  large  scale  in  California  is  a  type  of  this  gas. 
The  familiar  Pintsch  Gas  med  in  railroad  passenger  cars  is  a  type 
of  this  gas:  Blau  Gas  is  another.  Garburetted  Water  Gas  contains 
from  twenty-five  to  forty  per  cent  of  oil  gas. 

(6)  Acetylene  Gaa  gives  a  brilliant  white  light  when  properly 
burned.  It  is  prepared  as  required  by  adding  water  to  calcium 
carbide:  the  lamps  of  automobiles  are  a  familiar  example  of  its 
use.  In  country  districts,  hamlets,  villages  and  small  towns  are 
supplied  from  a  central  plant  with  this  gas. 

(7)  Caibnretted  Air  Gas.  This  gas  is  the  familiar  type  iised  in 
country  houses  and  hotels;  it  ie  simply  air  saturated  with  vapors 
of  gasolene. 

(8)  Frodnoer  Gas  contains  Nitrogen  and  abont  S5  per  cent  com- 
bustible gases;  when  cold  usually  requires  beating  to  make  it 
ignite;  is  seldom  distributed  beyond  the  boundaries  of  a  manu- 
facturing establishment. 

(9)  Hatoral  Oas,  one  coming  from  the  earth  usually  in  a  dis- 
trict where  petroleum  oil  is  also  present,  and  frequently  under 
pressure  of  many  atmospheres ;  it  is  usually  sold  at  much  less  cost 
Qian  artiJicial  gae  so  long  as  the  natural  gas  supply  remains 

E.  How  Artificial  Oaa  is  Manufactured 
(1)  Water  Gat,  sometimes  called  blue  gas,  is  made  by  raising 
the  temperature  of  a  fuel  bed,  by  means  of  a  forced  blast  of  air,  to 
incandescence  (the  Producer  Gas  made  usually  being  wasted),  whrai, 
the  air  being  shut  ofT,  steam  (H,0)  is  passed  through  the  fuel  bed 
{G,),  which,  on  decomposing  yields  Hydrogen  and  Carbonic  Oxide 
{CO),  the  Carbon  being  supplied  by  the  fuel.  TJsnally  hard  fuel 
is  used,  either  anthracite  coal  or  coke,  though  bituminous  coal  has 
been  used.  The  fuel  bed  is  usually  contained  in  a  cylindrical  shaped 
fire  brick  furnace  {Fig.  1  illustratoe  a  twin  generator)  which  in 
turn  is  surrotinded  by  a  gas  pressure  tight  cylindrical  ateel  rivetted 

>y  Google 

328  Illlminatixg  Exoixeebisg 

shell,  supplied  with  gaa  tight  stack  valve,  coaliog  and  cleaning 
doors  and  proper  air,  steam  aud  gas  connections  governed  by  valves, 
all  manipulated  by  the  gas  maker.  The  cylinder  containing  this 
fuel  bed  is  commonly  called  a  Generator;  when  single  it  is  eight  to 
twelve  feet  in  outside  diameter  and  twelve  to  twenty  feet  in  height. 
Wat*r  gas  is  colorless,  odorless,  specific  gravity  .330,  yields  ou 
analysis  (Stillman)  CO,  0.14,  Oj  0.13,  illuminants  0.0,  CH,  7.65, 
CO  37.97.  H,  49.32,  S^  4.79;  on  burning  yields  only  a  blue,  non- 
luminous  flame  and  ;t8.">  B.  t.  u.  per  cubic  foot. 


(S)  Carburetted  Water  Gas:  The  water  gas  constituent  of  this 
gas  is  made  in  an  identical  manner  as  above  outlined  for  water 
gaa.  The  oil  gas  constituent  may  be  made  by  heating  oil  in  ex- 
ternally heated  retorts,  but  is  now  usually  made  as  follows:  (2)  The 
cylinder  described  for  making  water  gas  is  connected  with  sim- 
ilar cylinders  in  duplicate  or  triplicate,  though  the  diameter 
may  be  slightly  varied  and  the  height  is  frequently  increased  by 
fifty  per  cent.  The  additional  cylinders  arc  not  used  for  containing 
fuel  but  are  filled  witli  many  hundreds  of  standard  fire  brick  placed 
in  checker  work  fashion  thus  providing  interstices  between  bricks 
for  the  passage  of  gases.    The  checker-brick  work  is  raised  to  in- 

>y  Google 

Manufacture  axd  Distribitiox  of  Gas 


vT.  m  ,.1, 

J   Tl  ■•'    "J  "1   " 






ta--:gg^L 1^ J .  a.4il I 




330  Illdminatinq  Enoineerino 

candeBceot  lieat  by  burning  up  the  Producer  Gaa  (wasted  In  the 
manufacture  of  water  gas)  made  from  the  fuel  bed  of  the  Water 
Gae  generator  when  it  is  being  brought  to  incandescence  by  a  blast 
of  air  preparatory  to  making  water  gas;  ail  additional  air  required 
for  this  secondary  combustion  comes  from  the  same  source  as  for 
the  first.  When  two  additional  cylinders  are  used  the  second  ia 
called  the  carbureter,  because  petroleum  is  dropped  on  the  hot 
bricks  in  this  cylinder  and  on  vaporizing  gives  light-giving  proper- 
ties to  the  water  or  blue  gas  flowing  over  hot  from  the  Generator, 

while  the  third  cylinder,  usually  taller  than  the  carbureter,  is 
called  the  Superheater  or  Fixing  Chamber;  the  function  of  the 
hot  fire  brick  in  this  cylinder  being  the  further  heating  of  the 
water  gas-oil  gas  mixture  and  the  "  fixing  "  of  the  oil  vapor  prod- 
ucts into  fixed  gas.  This  term  fixed  gas  is  used  in  a  limited  sense  to 
include  only  usual  atmospheric  temperatures  and  pressures.  (3) 
Carburefted  Water  Gas  has  the  familiar  pungent  "gas"  odor;  it 
has  a  specific  gravity  of  about  .G60,  contains  normally  as  sold,  no 
sulphuretted  hydrogen,  no  ammonia  and  but  seven  grains  of  sulphur 
compounds  in  100  cubic  feet.     It  yields  on  analysis  approximately 

>y  Google 

Mandfactuhb  and  Distribution  of  Gas 

Trcns verse  Section  Through  Retorts. 

Longitudinal  Section  Tbrough  Retorts. 
Fio.  6. — Bench  ot  Oae  Retorts. 

>y  Google 

332  Illuminating  Ekoineerino 

(Mass.  State  Inapector  1897)  COj  2.91,  O,  0,  Illuminants  14.98, 
Marah  Gaa  S5.90,  CO  25.30,  Hj  27.87,  N,  3.0i.  Ten  thousand 
cubic  feet  of  gas  would  require  in  its  manufacture  about  400  lbs. 
of  fuel  (where  T*aste  heat  boilers  are  not  placed  after  the  car- 
bureters) and  4H  gallons  of  oil:  the  gas  produced  would  be  about 
35  candle  power  and  as  a  by-product  may  yield  from  14  ^'^  9/10 
gallons  of  water  gas  tar. 

(3)  Retort  Coal  Gas  is  made  by  distilling  at  about  3200°-2600'' 
Fahrenheit  as  much  as  1000  lbs.  of  bituminous  gas  coal  in  a  (4) 
clay  retort  having  a  "  D  "  cross  section  typically  16  inches  by  2G 
inches  and  9  feet  to  20  feet  long,  either  vertically,  inclined  or  liori- 
zontally  placed.  The  dimensions  as  well  as  the  position  may  var>' 
and  the  weight  of  charge  is  graduated  to  the  retort  capacity;  in- 
variably the  retorts  are  externally  heated,  (5)  usually  in  groups  of 
six  to  nine,  by  a  single  furnace  but  when  retorts  are  20  feet  long, 
usually  by  two  furnaces.  Fumacea  are  usually  fired  without  forced 
bla.=<t :  The  coke  fuel  is  obtained  hot  from  one  of  the  group  of 
retorts  at  the  end  of  the  distillation  period,  which  varies  from  four 
to  nine  hours.  About  10,000  cubic  feet  of  gas  of  16  to  18  c.  p. 
is  obtained  from  one  gross  ton  of  coal  and  there  remains  as  by- 
products of  manufacture,  about  1000  lbs.  of  coke,  about  13  gallons 
of  tar,  and  ammonia  sufficient  to  produce  20  to  22  pounds  of  sul- 
phate of  ammonia.  Retort  Coal  Gas  in  all  essentials  has  the 
odor  of  carburetted  water  gas,  though  the  manufacturer  may  dis- 
tinguish in  the  odor;  it  has  a  specific  gravity  of  .400  to  .450, 
and  as  distributed  contains  no  sulphuretted  hydrogen,  though  often 
12  or  more  grains  of  sulphur  compounds,  0.3  grains  of  ammonia, 
and  on  analysis  yields  approximately  COj  1.75,  Oj  0,  Illuminants 
488,  Marsh  Gas  33.90,  CO  6.82,  H,  46.15,  N,  may  at  times  be 
found  as  high  as  6.50  though  1.5%  may  be  considered  a  fairer  per- 
centage.   The  heat  units  approximate  600. 

(4)  (6)  Coke  Oven  Coal  Oas  is  mannfactured  by  charging 
several  tons  of  bituminous  gas  coal  in  the  top  of  an  elongated 
"  D  "  oven  26"  wide,  72"  deep,  and  30  feet  long,  and  distilling  it 
normally  at  a  lower  temperature  than  in  the  case  of  Ketort  Coal 
Gas  but  for  periods  varying  from  24  hours  to  36  hours.  The  heat 
for  distillation  is  obtained  by  burning  the  poorer  quality  of  gas 
which  comes  off  after  the  first  10  to  12  hours  and,  after  removing 
the  ammonia  and  tar,  is  supplied  to  tlfe  exterior  of  the  coke  oven 
through  pipes  at  low  pressures ;  air  for  combustion  is  in  some  sys- 
tems heated  in  regenerators  by  the  waste  combustion  gaaes  from  the 


Manofactube  A^fD  Distbibution  of  Gas  333 

ovens  and  is  supplied  to  the  ovens  under  moderate  fan  pressures  or 
by  the  natural  draft  of  tall  stacks.  This  process  is  really  not  a  gaa 
making  process  but  a  coke  making  process,  in  which  gas  is  but  a  by- 
product. 31/2  gross  tons  of  coal  produces  .5200  lbs.  of  Coke  similar 
to  Bee  Hive  Coke  and  as  by-products,  10.000  cubic  feet  of  gas 

Longitudinal  Sections. 

SactlonThmch  K*c*utrUar.     Elvntlon.  TnainrH  flacHoai. 

Fio.  6.— Early  Type  of  Otto-HotTman  By-Product  Coke-Oven. 

of  17  to  19  candle  power,  30  gallons  of  tar,  and  ammonia  suffi- 
cient to  manufacture  73  lbs.  of  sulphate  of  ammonia.  The  by- 
product gas  available  for  distribution  has  a  specific  gravity  and  heat 
unit  value  quite  similar  to  retort  oven  coal  gaa  and  yields  on  analysis 
approximately  COj  0.1,  0^  0.1,  Illuminants  5.55,  Marsh  Gaa  38.90, 
CO  6.57  (may  reach  8.00)  H^  42.1,  Nj  6.65,  and  when  made  with 

>y  Google 

334  Illuuinatinq  Engineesino 

BulphuroUB  coals,  contains  as  impurities,  after  purification  with 
lime  for  carbonic  acid,  foul  lime  for  fixed  sulphur  compounds  aad 
oside  of  iron  for  sulphuretted  hydrogen,  normally  18  or  more  grains 
of  sulphur  compounds  and  0.2  grains  of  ammonia. 

(5)  Oil  Gas  may  be  made  in  iron  retorts  of  similar  pattern  to 
the  clay  retorta  used  in  Setort  Coal  Gas  but  they  are  smaller  in 
cross  section  and  usually  not  exceeding  9  feet  in  length;  latterly 
clay  retorts  have  been  used.  The  externa!  heating  is  effected  by 
means  of  the  best  available  fuel.  As  in  Carburetted  Water  Gas  the 
usual  by-product  is  tar.  Oil  Gas  burned  in  a  special  burner  has  a 
candle  power  of  60  to  100,  specific  gravity  about  that  of  air,  and 
heat  units  of  1200  or  over.  (7)  In  California  {see  paper  by 
E.  C.  Jones,  American  Gas  Institute  1909,  p.  410)  Oil  Gas  is  manu- 
factured on  a  large  scale  and  by  the  use  of  specially  designed  appa- 
ratus, in  which  oil  is  used  for  fuel  to  heat  up  checker  brick  work  in 
chambers  quite  similar  to  the  carbureter  and  superheater  of  the 
carburetted  water  gas  apparatus,  as  well  as  to  make  the  gas.  The 
character  of  the  oil  gas  here  made  is  distinct  from  oil  gas  made 
in  retorts  and  for  a  comprehensive  description  the  student  is  re- 
ferred to  the  able  paper  above  referred  to.  The  only  residual, 
lamp  black,  is  used  in  place  of  coal  to  manufacture  water  gaa 
which  is  mixed  wiUi  the  oil  gas.  The  low  labor  charge  per  thousand 
cubic  feet  made  is  an  argument  for  the  use  of  this  type  of  gas 
where  crude  oil  is  very  cheap.  The  analysis  of  the  distributed  gas 
is  given  as  COj  3.63,  Illuminants  9.10,  0,  0.34,  CO  10.34,  Hj 
36.54,  CH^  33.16,  Nj  6.39,  Candle  Power  21.88,  B.  t.  u.  710.7  and 
specific  gravity  .523, 

(6)  Acetylene  is  made  by  adding  water  to  Calcium  Carbide 
(which  has  previously  been  made  in  the  Willson  (8)  or  similar 
Electric  Furnace  from  lime  and  coke).  When  burned  in  special 
burners  the  resulting  gas  gives  an  intensely  brilliant  white  light 
of  about  250  candle  power,  has  a  specific  gravity  of  .910  and  a 
heat  unit  value  of  756  B.  t.  u. 

(7)  Carburetted  Air  Gas  (9)  is  made  by  forcing  air  through  a 
carbureter  in  such  a  manner  that  it  will  pick  up  10  to  17  per  cent 
of  gasolene  vapor.  It  must  be  burned  in  special  argand  or  mantle 
burners.    Its  heat  unit  value  has  been  given  as  815  B.  t.  u. 

(8)  Producer  Gas  is  made  in  the  Generator  in  the  fuel  heating 
period  of  Water  Gas  and  Carburetted  Water  Gas  manufacture,  (10) 
It  is  likewise  made  by  exhausting  air  or  air  end  steam  through 
any  incandescent  bed  of  fuel,  or  by  means  of  a  jet  of  steam  below 


Mandfactdre  and  Distribution  of  Gas  335 

'Jutis'Oiv  Cab  Sfer. 

>y  Google 

Illuminatixg  Enqinkerino 

>y  Google 

Manufactdhe  and  Distribution  of  Gas  337 

the  ash  pit  iojecting  air  and  steam  Taper  through  the  fuel  bed. 
This  gas  has  a  specific  gravity  of  about  .812  and  heat  units  of 
135-165  B.  t.  u..  Containing  about  75^0  N,  for  maximum  efBciency 
it  should  alvajs  be  burned  in  its  hot  state  as  it  emerges  from  the 
generator,  as  in  the  case  of  Carburetted  Water  Gas  manufacture. 

(9)  Natural  Oas  has  a  specific  gravity  of  .620  and  heat  units  1124. 
Analysis:  CO,  0.0,  0,  1.3,  Illuminants  0.5,  CH«  95.2,  CO  1.0,  H^ 
2.0,  Nj  0.0.  It  issues  from  the  earth  in  some  localities  at  a  pressure 
of  from  300  to  400  lbs.  per  square  inch.  Its  high  heat  units  and 
low  price  jwr  thousand  cubic  feet  make  it,  for  any  purpose,  a  com- 
petitor that  artificial  gas  cannot  compete  with  on  equal  terms. 
Natural  Gas  is  here  mentioned  to  accentuate  this  fact  and  to  give 
a  fairly  complete  view  of  the  commercial  field  occupied  by  gas. 

F.  Handling,  Etc.,  Within  the  Oas  Plant 
The  only  raw  material  required  to  manufacture  simple  coal 
gas  by  the  Retort  Process  (or  the  Coke  Oven  Process)  is  a  coking 
bituminous  gas  coal.  In  general,  a  first-class  bituminous  gas  coal 
should  contain  at  least  36%  volatile  matter,  no  more  than  %  of 
one  per  cent  of  sulphur,  and  should  be  received  in  the  condition  that 
a  %"  mesh  screen  at  the  mines  would  leave  the  larger  portion 
of  the  run  of  mine  coal.  Natural  conditions,  handling,  and  cost  of 
such  a  coal  largely  modify  these  general  specifications. 

Gas  Coal  to-day  is  mined  (Ua)  by  Electric  Mining  Machines, 
is  transported  (lib)  by  pit  wagons  to  the  mine  coal  tipples  (lie) 
and  dumped  (lid)  into  hoppers  with  screens,  and  if  a  gas  plant 
is  favorably  located  cars  loaded  (lie)  with  screened  coal  at  the 
mines  may  be  run  into  the  Gas  Works  coal  storage  shed,  or  even 
into  the  retort  houses  and  there  unloaded  into  the  retort  house  coal 
bins.  In  other  cases  coal  cars  may  go  to  tide  water  and  the  coal 
be  discharged  into  (13)  large  capacity  ocean  going  steamers  that 
will  deliver  the  coal  alongside  the  gas  plant  wharf  several  hundred 
miles  distant,  or  the  coal  cars  may  be  carried  hundreds  of  miles 
and  then  discharge  into  harbor  lighters  of  about  1000  or  more 
tons  capacity  which  deliver  the  coal  to  gas  plants  five  to  forty 
miles  distant.  In  the  latter  case  the  coal  contained  in  the  rail- 
road cars  from  the  mines  may  be  dumped  bodily  or  through  chutes 
into  the  lighters  without  any  hand  labor.  (13a-13b;)_  _  The  Astoria 
(14a)  coal  gas  plant  will  serve  aa  an  illustra^h|*a{  feV^  cotL^)^ 
manufacture:   On  arrival  at  this  gas  plaQt'"(Mbj)  autopajiti^ ^jib  . 

>y  Google 


>y  Google 

Manufacture  and  Distribution  op  Gas 

>y  Google 

340  Illuminatixg  EyaixEEiuso 

buckets  picking  up  two  gross  tons  of  coal  at  a  trip  deliver  the  coal, 
through  the  instrumpntality  of  ao  electrically  driven  traveling 
crane,  to  either  50  ton  railroad  cars,  that  may  be  sent  direct  to  the 
retort  houses,  or  to  temporarj'  coal  storage  pile  at  a  rate  as  great 
as  250  tons  per  hour  per  crane. 

An  electrically  driven  storage  bridge  600  feet  long  (15)  with  a 
7  gross  ton  automatic  bucket  transfers  at  the  rate  of  300  tons  per 
hour,  the  coal  from  the  temporary  storage  to  the  storage  yard,  or  the 
unloading  crane  first  mentioned  may  reclaim  the  coal  from  the 
temporary  storage  and  place  it  in  50  ton  cars  for  its  journey  to  the 

retort  house.  The  storage  bridge  at  appropriate  times  transfers  coal 
in  storage  to  the  same  50  ton  cars,  or  when  conveniently  and  hap- 
pily located,  may  deliver  the  coal  directly  to  the  track  hoppers  in 
front  of  tile  retort  house. 

Ordinarily  a  40  ton  steam  locomotive  places  two  50  ton  cars  con- 
taining different  grades  of  coal  side  by  side  on  two  parallel  sur- 
face railroad  tracks  at  one  end  of  the  retort  house;  beneath  the 
tracks  is  a  hopper  into  which  the  cars  are  unloaded  simultaneously 
at  varj-ing  speeds.  Usually  one  car  contains  a  very  sulphurous  while 
the  other  contains  a  less  sulphurous  coal,  and  thug  a  uniform  mix- 
ture of  coals  is  obtained.  The  track  hopper  contains  a  chain  scraper 
conveyor  which.  ijiQves  tl)e^pql_jit  t^ie  rate  of  135  gross  tons  per  hour 

>y  Google 

Manufactdbe  and  Distribi;tion  of  Gas  341 


348  Illuminating  Escinbbeiso 

up  an  inclined  plane  dropping  it  at  the  end  into  coal  cruaherB,  which 
discharge  the  coal  uniformlv  enislied  into  vertical  elevators  which 
raise  the  coal  to  the  roof  of  the  retort  house  where  the  coal  falls  on 
to  longitudinal  conveyors,  which  in  turn  distribute  the  coal  into 
longitudinal  coal  bins  in  the  inclined  retort  house,  and,  with  the 
aid  of  cross  conveyors  store  it  in  large  bins  convenient  for  charging 
machines  in  the  case  of  the  horizontal  retort  house. 

In  the  inclined  house  (16)  the  coal  drops  by  gravity  into  measur- 
ing hoppers  which  are  manipulated  bv  hand  and  the  coal  thus 
directed  to  its  final  resting  place  by  gravity  into  a  "  D "  retort 

normally  16"  x  26"  s  20  feet  long.  In  the  horizontal  house  the 
coal  drops  into  a  charging  machine  electrically  controlled  (17) 
which  is  run  on  rails  opposite  to  nnd  below  the  level  of  all  retort 
lids;  this  machine  measures  all  charges  and  charges  the  retorts  by 
means  of  large  scoops  driven  into  the  gas  retort,  on  releasing  its 
charge  of  coal  uniformly  distributed  in  the  retort  the  scoop  is 

While  the  charge  remains  in  the  retort  which  is  heated  on  its 
exterior  by  the  combustion  of  hot  coke  drawn  from  a  previous 
charge,  gas  is  being  continually  driven  off  through  a  seven  inch  as- 
cension pipe  as  will  be  presently  outlined.  After  a  distillation  period 
that  may  van'_  from  4  to  even. 8. or  9  hours,  the  mouthpieces  are 

>y  Google 

Mandpactube  and  Distribution  ( 

>y  Google 


>y  Google 

Maxcfactl're  and  Distribution  of  Gas  345 

>y  Google 


>y  Google 


>y  Google 

Illominatino  Enoineehino 




>y  Google 

350  Illuiiixatin'g  Exgixeering 

slacked  off,  liro<1  and  opened,  and  the  residue  of  the  coal  heing 
deprived  of  its  volatile  matter  and  now  called  coke,  is  either,  in  the 
case  of  an  inelined  retort,  allowed,  by  gravity,  to  fall  into  n  con- 
veyor in  front,  or  in  the  case  of  a  horizontal  retort,  is  pushed 
by  an  electrically  controlled  machine  (18)  somewhat  similar  in  ap- 
pearance to  the  charger,  on  to  an  electrically  driven  longitudinal 
hot  coke  conveyor  running  below  the  mouths  of  all  retorts.  Some 
of  this  hot  coke  may  be  deflected  into  the  bench  furnaces  for  heat- 

ing the  retorts  externally,  but  the  larger  proportion  proceeds  along 
the  house,  meeting  sprays  of  quenching  water  in  its  travel,  and 
drops  at  the  end  of  its  journey  into  transfer  bucket  conveyors 
which  convey  it  horizontally  and  vertically  into  a  coke  storage  bin. 
(19)  By  gravity,  coke  is  delivered  from  the  storage  bin  (20)  to 
30  ton  ears  on  a  grade  railroad  and  by  means  of  40  ton  locomotives 
delivered  over  a  track  hopper  (21)  at  the  wharf.  The  coke  falls 
to  an  electrically  driven  belt  which  delivers  the  coke  (23)  at  a 
speed  of  200  net  tons  per  hour  to  a  barge  alongside.  Up  to  this 
time  all  material  has  been  handled  bv  electricallv  driven  mechan- 

>y  Google 

SlAi'UFACTURE   AND    DlSTRIBlTIOX    OF    OaS  351 

ism,  except  in  the  case  of  the  steam  locomotive  and  the  hand 
manipulated  measuring  hoppers  in  the  inclined  retort  house. 

The  barge  is  towed  to  the  coke  distrihutinn  points  and  by  mcflns 
of  a  belt  conveyor  within  the  barge  located  jnst  over  the  keel,  of  a 
bucket  elevator  (33),  and  athwartship  or  cross  belt  conveyor  in  the 
bow  (all  driven  by  a  single  kerosene  internal  combustion  engine) 
is  delivered  at  the  rate  of  60  tons  an  hour  on  to  an  electrically 
driven  inclined  belt  conveyor  located  on  the  wharf  whicli  (24)  de- 
posits the  coke  on  a  coke  platform.  From  the  platform  it  is  deliv- 
ered by  gravitj'  (25)  over  coke  screens  to  teams  or  motor  trucks 
which  in  turn  deliver  it  into  the  sidewalk  chutes  of  commercial 
buildings,  apartment  house  steam  plants  or  other  users  of  coke. 
Only  here,  beyond  the  jurisdiction  of  the  manufacturer  of  gas,  is 
any  hand  labor  applied  since  the  coal  left  the  pick  of  the  miners  in 
the  mine  from  which  it  came.  An  exception  to  this  Htatement 
would  exist  where  coke  has  to  be  carried  in  baskets  from  the  team 
on  the  street  to  the  storage  bins  of  the  user. 

A  convenient  au.xiliary  to  Retort  Goal  (Jas  mannfatturr  is  a  Car- 
buretted  Water  Gas  I'lant.  The  fuel  used  to  make  the  water  gas  is 
coke,  and  this  should  be  delivered  hot  direct  from  the  coal  gas 
retorts  but  when  necessary  quenched  coke  is  taken  from  the  coke 
storage  bins  for  use  in  the  water  gas  generators  in  adjoining  gen- 
erator house.  ■  Detail  explanation  is  omitted  for  want  of  time,  the 
explanation  of  the  manufacture  of  carburetted  Water  Gas  hereto- 
fore made  El  and  2  being  deemed  sufficient. 

As  previously  pointed  out  1000  lbs.  in  coke  from  each  2'JHI 
lbs,  of  bituniiuons  gas  coal  received  at  the  coat  wharf  must  be 
disposed  of  as  a  by-product  in  Retort  Coal  Gas  manufacture;  the 
great  value  of  this  primary  by-product  must  be  at  once  apparent. 

While  the  coal  lay  in  its  hot  bed  in  the  retort  the  36%  volatile 
matter  was  seeking  an  outlet  (26)  via  the  ascension  pipe  before 
spoken  of.  The  length  of  time  the  charge  remains  under  distilla- 
tion depends  upon  the  degree  and  uniformity  of  heat,  the  character 
of  the  coal  and  the  distribution  of  the  charge  in  the  retort,  as  well 
as  the  conductive  qualities  of  the  retort,  for  all  have  their  influence 
on  the  resulting  products.  The  heating  of  a  charge  of  coal  distills 
the  volatile  matter  and  causes  a  gas  pressure  within  the  retort:  it 
is  not  desirable  to  have  too  great  a  pressure  accumulate  boi'aiise  of 
loss  of  gas  through  the  porous  sides  of  the  clay  retorts;  the  gas 
outlet  of  a  retort  or  asci'nsion  pipe  terminates  in  a  metal  chamber, 

>y  Google 

Illcminatino  Enqixsbkinq 

>y  Google 

Manufacture  and  Distbibution  of  Gas  353 

called  a  hydraulic  main  located  above  the  retort  bench;  the  ascen- 
sion pipes  are  sealed  in  water  originally  but  later  by  accumulations 
of  hot  tar  and  ammouiacal  liquor,  products  formed  from  a  portion 
of  the  36%  volatile  matter  in  the  coal  that  become  liquid  at  the 
temperature  of  the  gas  on  passing  the  water  seal.  To  prevent  ex- 
cessive pressure  within  the  retorts  a  gas  pump  called  an  exhauster 

is  installed  in  a  house  {'i7)  beyond  the  retort  house  and  the  ex- 
hauster is  run  at  a  variable  speed,  by  the  aid  of  automatic  governors, 
so  that  all  the  gas  as  Jriven  off  in  the  retorts  is  at  once  drawn  away 
from  the  hydraulic  main  under  a  partial  vacuum  sufficient  to  over- 
come the  water  seal  in  the  hydraulic  main  and  to  prevent  but  a  very 
slight  pressure  in  the  retort. 

Forty  years  ago  cast  iron  retorts  were  in  use  in  coal  giis  benches 

>y  Google 

Illcminatiso  Enoineebing 

>y  Google 

Manufacture  and  Distribution  of  Gas 

>y  Google 

356  Illuminating  Esqineebinq 

and  then  no  exhauster  was  thought  necessary  aa  gaa  (.-ould  not 
penetrate  in  excessive  quantities  the  cast  iron  retorts.  In  the 
Pintech  Oil  Gas  process  east  iron  retorts  may  still  be  used,  but  so 
far  as  1  am  aware,  no  coal  gas  works  to-day  employs  their  use. 
Frequently  gas  in  traveling  from  the  retort  house  to  the  exhauster 
house  is  cooled  by  atmospheric  influences  either  in  the  pipes  lead- 
ing to  the  exhauster  house  or  by  specially  designed  apparatus.    The 

gas  temperature  at  the  inlet  of  the  exhausters  closely  approximates 
120°  Fahrenheit. 

On  passing  through  the  exhauster  outlet  the  gas  immediately  it 
under  pressure,  for  the  exhauster  is  then  forcing  the  gas  to  the 
storage  holder  against  the  storage  holder  pressure,  which  varies, 
depending  upon  its  height,  as  will  be  explained  later;  additional 
pressure  is  pro(liic«d  by  backpressure  in  overcoming  the  resistance 
of  the  gas  travel  through  apparatus  on  the  way  to  the  holder  as 

>y  Google 

Manofactore  and  Disthibdtion  of  Gas  357 

veil  08  by  the  Bkin  friction  of  the  main  gas  pipes  of  the  Bystem. 
Should  the  exhausters  all  stop  simultaneously  a  safety  gas  blow  out 
governor  would  allow  the  gae  egress  to  the  open  air  above  the  Ex- 
hauster house  roof,  thus  preventing  excessive  accumulation  of  gas 
pressure  between  the  exhausters  and  the  retorts. 

Baw  gas  from  the  exhausters  passes  first  through  (28)  mechan- 
ical tar  extractors  having  plates  with  small  openings  that  break  up 
the  gae  volume  in  small  streams  and  by  friction  disengages  tar 
from  these  gas  streams ;  the  raw  gas  next  passee  through  horizontal 
rotary  acrubbere  (29)  where  hydrocarbon  liquids  extract  the  naph- 
thalene in  the  gas.  It  is  now  believed  that  these  washers  should 
be  placed  next  in  order  to  the  condensers  later  spoken  of;  from 
the  naphthalene  scrubbers  the  gas  passes  through  a  liquid  solution 
of  sulphate  of  iron  in  the  (30)  cyanogen  washers  which  deprive 
the  gas  of  any  cyanogen  contained.  The  liquid  on  saturation  is 
sent  to  settling  tanks,  then  to  filter  presses  which  form  in  a  pressed 
cake  a  raw  product  called  cyanogen  sludge  (31)  which  is  shipped 
to  the  chemical  factories ;  residue  liquor  from  the  filter  presses  is 
put  through  drying  processes  and  converted  into  dry  sulphate  of 
ammonia,  which  is  bagged  and  placed  on  the  market  for  sate. 
After  this  purifying  process  the  gas  passes  through  the  cast  iron 
tubes  of  surface  condensers:  (32)  the  tubes  are  surrounded  by  salt 
water,  where  available,  and  the  water  current  is  arranged  so  that 
the  stream  of  warm  water  leaving  the  condenser  meets  the  warm 
gas  entering  the  apparatus. 

On  leaving  the  condensers  the  gas  passes  through  water  in  am- 
monia washers  (33)  quite  similar  in  their  design  to  the  Cyanogen 
and  Naphthalene  Washers,  and  the  gas  having  been  freed  of  all  tar, 
cyanogen  and  naphthalene,  now  surrenders  its  last  trace  of  am- 

The  tar  from  the  hydraulic  mains,  the  main  pipe  connections, 
exhausters,  tar  extractors  and  condensers  is  led  into  underground 
tar  wells  from  whence  it  is  pumped  into  shipping  tanks  near 
the  wharf,  from  whence  the  chemical  contractors  take  it  to  make 
the  tar  into  pitch,  dead  oil,  and  various  coal  tar  compounds. 

The  ammonia  from  the  ammonia  washers  is  sent  to  underground 
ammonia  tanks,  and  together  with  ammoniacal  liquor  recovered 
from  the  hydraulic  mains  and  other  connections  and  apparatus 
where  tar  is  present,  is  ail  transferred  to  ammoniacal  liquor  tanks 
near  the  wharf,  from  which  the  chemical  contractors  remove  it  and 

>y  Google 


Digitized  .yCoO^t^lC 

Mandfacture  and  Distribution  of  Gas  359 

obtain  therefrom  anhydrous  ammonia,  sulphate  of  ammonia  and 
other  ammonia  pioducts. 

The  gas  leaving  the  ammonia  scrubbers  next  passes  into  the  puri- 
fying boxes  now  ueually  a  dry  process  of  purification.  Boxes 
(34)  40  feet  square  and  8  feet  deep  are  uniformly  spread  with 
oxide  of  iron,  usually  deposited  on  white  pine  shavings  supple- 
mented by  iron  borings.  The  layers  vary  in  thickness  in  practice, 
being  in  some  cases  upwards  of  42  inches  thick.  The  gas  is  here 
deprived  of  sulphur  which  is  in  the  form  of  sulphuretted  hydrogen. 
Some  fixed  forms  of  sulphur  are  undoubtedly  taken  up  from  time 
to  time  in  the  journey  of  the  gas  from  the  hyilraulic  main  to  the 
outlet  of  the  purifjing  houf^e  and  unless  a  very  sulphurous  coal  must 
be  used,  no  lime  purification  is  found  necessary  to  meet  a  20  grain 
legal  provision.  Where  it  is  found  necessary  to  use  the  latter  it  is  not 
an  extravagant  statement  to  make  that  the  increased  cost  of  purifi- 
cation (more  particularly  in  the  case  of  Coke  Oven  Gas)  may 
be  ten  times  the  cost  of  ordinary  oxide  purification.  It  has 
been  found  that  in  order  to  eliminate  fixed  sulphur  compounds 
that  all  carbonic  acid  must  first  be  removed  from  gas  in  one  set  of 
boxes  then  lime  fouled  from  sulphuretted  hydrogen  will  attack  the 
fixed  sulphur  compounds  in  a  second  set  of  boxes,  and  finally  a  third 
set  <rf  oxide  boxes  must  be  used  to  remove  any  snlphuretted  hydro- 
gen that  may  be  present.  These  three  independent  processes  must 
he  carried  on  in  the  purifying  house,  where  but  one  is  required 
when  using  fairly  low  sulphur  coal.  One  of  the  three  processes  (the 
second  in  order)  is  exceedingly  disagreeable  to  the  employees  of 
the  gas  company  and  to  the  neighbors  as  well.  So  little  value  is 
now  attached  to  the  requirements  for  fixed  sulphur  compounds  that 
England,  having  passed  through  the  regulating  by  taw  stage,  now 
no  longer  demands  any  specified  freedom  of  fixed  sulphur  in 
gas;  New  York  only  very  lately  passed  laws  respecting  sulphur, 
merely  imitating  the  laws  of  other  places  without  reference  to  any 
necessity.  Massachusetts  in  this  respect  is  also  moderating  its 
position  as  regards  sulphur  in  illuminating  gas. 

Having  passed  the  purifying  house  the  gas  now  goes  through 
IG  foot  station  meters,  (33a-35b)  in  which  the  measuring  drums 
run  in  water;  gas  measurements  are  made  as  near  GO"  Fahrenheit 
as  atmospheric  conditions  permit  but  the  measurements  are  cor- 
rected to  60°  Fahrenheit  and  30  inches  barometer  in  any  event. 
Here  I  might  state  that  unaccounted  for  gas,  not  leakage  as  fre- 

>y  Google 

Illuminatiko  Enqineerino 

>y  Google 


>y  Google 

363  Illuminatixq  Enoineeriko 

quently  ass^rtetl,  ia  agpfi-tain«d  by  taking  the  sum  of  the  readings 
of  the  station  meter?  and  subtracting  therefrom  the  sum  of  the 
readings  of  all  the  consiimers"  meters,  the  rpniaiitder  is  unaccounted 
for  gas,  which  includes  actual  loss  of  gas  by  leakage,  loss  nf  volume 
represented  by  difference  of  temperature  at  which  the  meters  in 

Fm.  3Sa. 

the  cellars  of  houses  measure  their  gas  (which  cannot  l>e  corrected) 
the  slowness  of  the  house  meters,  and  condensation  of  hydrocarbons 
and  aqueous  vapor  present  in  the  gas  itself. 

Passing  from  the  station  meter  the  commercial  gas  now  goes  into 
the  storage  holder.  {36)  The  largest  liolder  in  use  holds  15,000,000 
cubic  feet;  it  is  300  feet  in  diameter  and  233  feet  high,  having  48" 
inlet  and  outlet  pipes.  The  illustration  shows  an  empty  tank  pre- 
pared to  receive  a  second  holder. 

>y  Google 

Mandfactlre  and  DisTRiBrTioK  OF  Gas 


G.  Distrihuiion  of  Gas  from  Storage  Holder  to  City  Disirihution 
In  small  plants  the  works'  holder  distributes  the  gas  direct  to 
consumers  while  in  larfie  plants  Exhausters  (in  this  service  some- 
times termed  boosters  or  Gas  Pushers)  (37)  piiiiip  the  gas  from  the 

works'  storage  holder  into  transfer  mains,  as  large  as  60  inches  in 
diameter,  and  some  times  through  tunnels  under  rivers  into  district 
Distributing  Oas  Holders. 

It  might  be  well  to  here  call  your  attention  to  the  method  of 
obtaining  the  initial  gas  pressure  used  in  gas  distribution.  It  has 
been  stated  that  the  gas  exhauster  withdraws  the  gas  from  the 
retorts  as  rapidly  as  made  and  that  after  the  gas  passes  through 
the  exhauster  it  is  under  pressure  due  to  the  skin  friction  of  pipt-i;, 

>y  Google 


the  resistance  to  the  passage  of  gae  tbroiigb  tbe  apparatus  described 
and  the  preeenre  of  the  gas  storage  holder  due  to  its  weight.  la  the 
case  of  the  district  holder,  identical  in  all  respects  to  the  vorlra' 
storage  holder,  the  motive  power  (steam,  gas  or  electric)  driving  a 
gas  exhauster  fills  the  holder  and  when  gae  has  been  forced  into 
the  holder  the  holder  itself  maintains  the  initial  gae  preesnre  on 
the  supply  mains  in  the  following  manner. 




^-   1 


The  gas  holder  is  free  to  move  vertically  and  when  being  filled 
is  held  in  its  vertical  poBition  by  guide  wheels  rolling  on  guide 
framing  supported  and  bolted  to  the  tank  wall.  The  top  of  the 
holder  prevents  the  escape  of  gas  upward,  the  cylindrical  barrel 
or  sides  of  the  holder  prevent  any  escape  from  the  sides  and  the 
water  in  the  holder  tank  prevents  escape  downward ;  the  only  man- 

>y  Google 

Manufacture  and  Disthibution  of  Gas  365 

ner  that  gas  reaches  the  holder  or  escapes  from  it  is  by  way  of 
pipes  passing  through  the  water  of  the  tank;  gas  flow  is  governed 
by  valves  on  tliese  pipes  in  the  valve  house.  The  gas  is  in  fact  aup- 
porting  the  gas  holder  and  the  weight  and  area  of  the  gas  holder 
determines  the  initial  gas  pressure  obtainable.  , 

It  would  be  quite  impossible  to  have  a  water  tank  as  deep  as  the 
gas  holders  in  our  large  cities  are  in  height,  which  would  be  neces- 
sary if  the  gas  holders  were  not  made  telescopic.  The  photograph 
shows  (a)  a  (38a)  eroes  section  of  a  gas  holder;  grouiiiled,  in  siich 

position  giving  no  presMure  whatever;  (b)  (.tSb)  the  holder  proper 
just  engaging  an  additional  telescopic  section;  (c)  {;19)  the  gas 
holder  entirely  inflated  with  the  first  section  or  holder  proper  and 
its  4  additional  telescopic  sections  filied.  In  the  case  illustrated  the 
holder  is  190'  10"  in  diameter,  230'  high  and  weighs  2,170,203  lbs. 
The  holder  proper  or  first  section  will  give  an  initial  gaa  pressure 
measured  at  the  crown  of  the  holder,  of  4.9  inches  of  water  (you 
remember  that  37.68  inches  of  water  is  equal  to  1  lb.  pressure  per 
square  inch)  the  second  section  when  added  increases  the  pressure 
by  2.8  inches  and  gives  a  total  of  7.7  inches — the  third  a  total  of 
10.4  inches — the  fourth  a  total  of  12.5  inches — and   the  holder 

>y  Google 

366  Illuminating  Enqinksrinq 

fully  inflated  a  total  of  15.0  inches  {if  the  holder  were  filled  with 
air  the  air  pressure  would  be  16.3  inches,  the  difference  1.3  inches 
is  equivalent  to  a  weight  of  173,200  Iba,  When  tiie  holder  is 
grounded  the  water  level  in  the  tank  ie  level,  when  the  holder 
proper  or  first  section  is  raised  the  water  outside  the  holder  is 
4.9  inches  out  of  level  with  the  interior  and  when  the  holder  is 
entirely  inflated  (40)  the  difference  in  level  is  proportionately  in- 
creased. The  inlet  and  outlet  pipes  should  be  higher  in  level  than 
the  tank  walls  and  the  tank  overflow  always  above  the  level  suffi- 
cient to  permit  the  maximum  alteration  in  water  level  without 
wasting  water.  While  the  water  visible  within  the  tank  rises  very 
perceptibly  the  original  water  level  within  the  holder  is  but  slightly 
altered  because  of  the  great  difference  of  the  water  areas  involved. 
The  gas  cannot  escape  from  the  telescopic  joints  because  of  water 
seals  between  the  sections.  When  it  is  necessary  to  ground  a 
holder  and  cut  it  off  from  the  pipe  system  by  valves,  changes  in 
barometric  pressure  or  temperature  would  make  the  crown  of  the 
holder  collapse,  unless  an  opening  to  the  atmosphere  were  made. 

If  the  maximum  initial  gas  pressure  required  does  not  exceed 
the  gas  pressure  given  by  the  holder  proper  or  first  section,  then 
all  the  gas  can  be  sent  out  of  the  holder.  If  a  higher  pressure  than 
this  is  demanded  then  the  gas  exhauster  must  be  called  upon  to 
supply  the  deficiency,  for  the  holder  cannot  always  be  kept  fully 
inflated,  or  its  value  would  be  lost. 

I  have  taken  some  time  to  describe  the  gas  holder,  but  it  is  the 
one  feature  of  the  gas  business  that  is  the  envy  of  our  electric 
brothers.  What  royalty  would  not  an  electric  company  pay  for 
an  equivalent  electric  device  that  would  at  an  equal  annual  cost 
store,  without  loss,  the  latent  energy  of  all  their  boiler  and  engine 
plant  run  for  24  hours,  always  ready  at  any  second  for  maximum 
or  unusual  demand,  and  needing  but  the  turn  of  a  wheel  or  the 
automatic  adjustment  of  a  switch  to  bring  its  latent  energy  into 
action.  The  gas  industry  could  not  be  what  it  is  without  the 
gaa  holder  and  without  the  aid  of  the  hydraulic  seal,  ae  both  are 
essential  to  gas  manufacture  and  distribution. 

H.  Distribution  Holders  to  Services 

The  gas  in  the  District  Distributing  Holders  passes  into  a  Valve 

House  (41)  where  valve  men  may  control  the  initial  pressure  and 

hence  the  rate  of  gas  flow  through  the  various  large  street  mains 

>y  Google 

Manupaotdbe  and  Dibthibution  of  Gab  367 

20"  to  30"  or  more  in  diameter  which  are  connected  later  with  a 
multitude  of  smaller  street  mains  ranging  from  4"  to  16"  in  diam- 
eter. In  special  cases  Gaa  Pushers  are  used  to  foree  the  gas  through 
the  large  mains  for  long  distances  before  the  gas  is  permitted  to 
find  its  way  into  the  smaller  mains.  The  gas  pressure  to  a  district 
is  reflated  either  by  a  gas  regulator  or  by  a  valve  man  adjusting  an 
ordinary  gate  valve.  The  valve  man  regulates  the  gas  flow  by 
watching  the  pressure  of  the  gas  leaving  the  valve  house  but  in 

Fio.  41. 

some  cases  is  assisted  by  the  use  of  an  ingenious  electrical  device 
which,  on  pressing  a  contact  key,  rings  a  belt  whose  strokes  indi- 
cate the  pressure  in  the  mains  a  mile  or  more  away;  in  such  a 
case  the  valve  man  maintains  a  given  pressure  at  that  distant 
point  quite  independent  of  what  pressure  the  gas  is  under  on 
leaving  the  Valve  House.  The  best  the  gas  manager  can  do  is 
to  strive  to  furnish  a  given  locality  with  uniform  pressure — 
the  exact  amount,  within  reasonable  limits,  is  not  so  important. 
It  is  impossible  to  give  a  uniform  pressure  of  gas  everywhere  in 

>y  Google 

366  Illdminatinq  EtraiNBEBiNa 

a  gas  system,  for  in  order  to  distribute  gas  at  all,  difference  in 
pressure  must  be  established  before  the  gas  will  flow  in  the  pipes. 
Water  distribution  requires  a  pressure  obtained  by  the  use  of  reser- 
voirs, stand  pipes,  or  pumps  in  order  to  make  it  flow  tbrougb  a 
city  system.  The  water  pressure  of  a  city  system  is  not  uniform 
any  more  than  the  gae  pressure  in  a  gas  system. 

Id  suburban  districts  the  Oas  Pusher  is  used  to  send  gas  many 
miles  through  small  mains  under  so  called  high  pressure;  in  this 
case  we  speak  of  the  Qas  Pusher  as  a  Gaa  Compressor  (it  is  usually 
of  the  piston  type).  At  appropriate  points  in  the  system  a  branch 
pipe  supplies  a  district  through  a  reducing  valve  (usually  in  dupli- 
cate) into  the  local  district  system;  the  gas  pressure  maintained 
in  such  a  district  is  the  pressure  that  the  district  gas  holder  would 
furnish  in  ordinary  circumstances;  in  fact  a  gas  holder  is  fre- 
quently primarily  supplied  by  a  high  pressure  main  in  which  case 
the  high  pressure  main  is  really  a  transfer  main  similar  to  that 
heretofore  spoken  of,  only  the  diameter  is  small  and  the  gas  pres- 
sure carried  is  greater.  In  some  cases  gas  is  supplied  directly  from 
a  high  pressure  main  to  a  house  en  route ;  in  such  a  case  a  gas  pres- 
sure reducing  valve  is  placed  in  the  house,  but  a  safety  seal  is  also 
provided,  so  that  if  the  reducing  valve  mechanism  fails  to  perform 
its  work,  excess  gas  pressure  cannot  come  on  the  house  meter  or 
house  fixtures,  but  the  gas  seeks  a  safe  course  to  the  atmosphere 
above  the  top  of  the  house  by  means  of  an  escape  pipe. 

The  pressure  carried  in  gas  pipes  in  the  street  is  quite  inde- 
pendent of  the  strength  of  the  gas  pipe  itself  with  respect  to  burst- 
ing by  internal  pressure.  Any  gas  pipe  would  stand  several  hun- 
dred times  the  pressure  it  is  subjected  to  in  the  ordinary  district 
distribution  of  gas. 

The  danger  of  fracture  of  gas  pipes  comes  entirely  from  the  char- 
acter of  the  soil  and  the  use  of  the  streets  by  other  public  utilities — 
electrical  conduits,  sewers,  water  pipes,  steam  pipes,  conduits  for 
telephone  and  telegraph  and  the  like.  During  the  installation  and 
repair  of  these  public  utilities  and  following  these  processes  the 
conditions  underground  are  indefinite  and  complex.  When  a  fault 
does  develop  in  a  gas  main,  the  gas  manager  must  have  at  hand 
emergency  forces  for  instant  service  and  it  is  not  wise  to  give  these 
men  the  task  of  coping  with  this  useful  servant  under  too  high 
pressure,  less  it  become  a  dangerous  foe.  Opinions  may  differ 
on  the  subject  of  gas  pressures  appropriate  to  use  in  public  streets, 

>y  Google 

Mancfactuee  and  DisTEiiBDTiON  OF  Gas  369 

and  fixed  opinions  are  sometimes  modified  by  extended  experience — 
so  that  on  this  subject  we  can  fairly  say  that  "  circumstances  alter 
cases  ", 

It  is  often  suggested  that  the  proper  method  of  distributing  gas, 
water,  electricity,  telephone  service,  steam  and  provide  for  sewage 
should  be  by  installing  all  the  conduits  furnishing  these  in  sub-sur- 
face chambers  called  pipe  galleries  (43)  constructed  the  length  and 
breadtii  of  a  town  or  city.  The  failure  of  many  an  untried  but 
promising  process  is  due  to  not  taking  into  consideration  all  natural 

influences  and  forces.  Nature  never  fails  to  supply  them  all  though 
even  thoughtful  and  experienced  men  sometimes  forget  to  take 
them  all  into  consideration. 

Time  does  not  permit  discussion  of  the  pros  and  cons  of  pipe 
galleries,  but  the  writer  is  opposed  to  their  use,  and  believes  that 
the  pipe  gallery  cure  is  worse  than  the  pavement  disturbing  disease. 
Corporations  using  the  public  thoroughfares  should  be  required  to 
restore  the  street  surface  disturbed  to  as  good  a  physical  condition 
as  they  found  it.  Under  present  conditions  the  maintenance  and 
repair  of  the  conduits  of  public  utilities  is  conducted  with  d  mini- 
mum of  danger  to  the  public ;  before  advocating  pipe  galleries  the 
possibilities  of  wide  spread  disaster  should  be  first  overtome. 

>y  Google 

370  Illuminating  E: 

/.  Distribulion  of  (lax  from  Street  Mains  to  Appliances  for 
Burning  Gas 

Usually  gas  mains  aie  drilled  and  tupped  with  standard  pipe 
thread  and  jnis  services,  nsually  the  smaller  sizes  of  wruught  iron 
or  steel  pipes,  by  the  aid  of  proper  fittings  conduct  the  gas  to  within 
the  foundation  wall  of  the  consumers'  premises.  It  is  good  jirac- 
tieo  to  install  a  gas  service  in  a  straight  line  from  the  street  main 
witli  a  constantly  rising  grade  to  the  housi",  provide  at  the  street 

main  for  reasonable  settlement  of  street  main,  service  or  the  soil 
supporting  them,  and  provide  a  small  sized  opening  within  the 
house  for  aeeess  lo  the  interior  of  the  service  pipe.  Should  the 
service  pipe  be  exposed  to  atmospheric  influences,  ns  in  the  case 
of  areaways  in  the  front  of  the  building,  special  precautions  are 
advisable  where  the  winter  temperature  may  be  expected  to  be  verv- 
low.  Artificial  gas  haw  in  suspension  aqueous  vapors  and  vapors  of 
hydrocarbon,  which  may  be  liquefied  at  low  temperature.  It  is 
advantageous  to  have  tbese  liquids  flow  hack  to  the  street  mains 

>y  Google 

SIaxl-facture  and  Distribotion  of  Gas 


and  collect  into  drips,  which  are  chambers  left  in  the  street  mains 
below  the  lower  level  of  the  mains.  ■  All  gas  mains  in  the  street 
are  very  carefully  laid  on  ascending  and  descending  grades  with 
drips  installed  at  the  low  points.  The  drips  are  pumped  dry  from 
the  street  surfaces  (43)  by  the  use  of  a  pump  and  receiving  tank 
usually  drawn  through  the  streets  by  horses. 

Any  condensation  of  the  lijtlrocarbon  vapors  deprive  the  gas  of 
both  its  illuminating  power  and  its  heat  unit  value  and  is  alwars 

Pre.  44. 

provided  against  so  far  as  possible.  The  aqueous  vapor  in  severely 
cold  weather  may  be  congealed  and  thereby  gradually  reduce  the 
capacity  of  the  service  pipe  below  the  requirements  of  service. 

Naphthalene,  usually  in  the  fall  season,  sometimes  gives  trouble 
by  obstructing  the  service  pipes  in  a  gas  system.  When  the  service 
pipe  passes  through  an  open  area  way  the  size  is  often  increased, 
and  the  pipe  is  sometimes  covered  in  some  way,  more  particularly 
where  entering  the  wails;  this   mitigates  and  sometimes  eradi- 

>y  Google 

372  Illdminatino  Exoineerimo 

cates  trouble  vith  stoppages.  It  is  wise  to  provide  ample  sizes 
of  pipes  for  services  dependent  upon  the  gaB  demand.  Some 
services  are  installed  a^  large  as  the  smaller  sizes  of  street  mains;  in 
such  cases  the  service  pipes  are  connected  with  the  street  main  in  the 
same  manner  in  which  street  mains  are  connected  with  each  other 
(44)  at  intersections  of  streets,  usually  with  lead  or  cement  joints. 
Service  pipes  are  fre<Ju€ntly  provided  with  cut-off  valves  located 
between  the  street  main  and  the  building  supplied;  when  so  placed 
they  should  be  of  a  type  not  apt  to  be  inoperative  through  very  in- 
frequent use.  The  pipe  leading  from  the  street  service  to  the  gas 
meter  is  frequently  spoken  of  as  the  inside  or  house  service  and  its 
location  and  installation  is  dependent  upon  the  needs  of  the  service 
and  the  will  of  the  architect  or  owner. 

The  gas  meter  in  reality  im  a  motor  operated  by  the  gae  pressure 
originating  at  the  gas  holder  but  operating  only  when  any  gas 
&utlet  beyond  the  meter  is  opened  to  the  atmosphere.  Meters  with 
drums  revolving  in  water  have  gone  out  of  use  because  of  the 
difficulty  of  keeping  the  water  level  intact  and  from  freezing 
troubles.    Dry  meters,  bo  called,  are  almost  universally  used. 

The  essential  parts  of  a  typical  dry  meter  are  two  main  cham- 
Iwra,  one  on  each  side  of  a  central  gas-tight  partition ;  each  cham- 
ber is  fitted  with  a  hollow  collapsible  bellows  or  piston  formed 
by  a  tlat  disc  connected  witJi  the  central  partition  by  a  gas  tight 
cylindrical  leather  diaphragm,  each  piston  being  inflated  and  de- 
flated within  its  own  chamber  and  in  rythm  with  its  twin  in  the 
adjoining  chamber  beyond  the  central  gas  tight  partition.  The 
space  within  each  piston  and  the  chamber  surrounding  it  are  filled 
with  gae  and  are  independently  coiinwted  by  valve  passages  to  a 
slide  valve,  usually  of  the  familiar  "  T> "  pattern  met  with  in  simple 
steam  engine  practice,  operating  in  a  gas  tight  chamber  at  the  top  of 
the  meter.  The  mechanism  is  so  connected  that  in  reality  the 
meter  is  a  double  acting  motor.  The  gas  is  measured  by  count- 
ing on  a  dial  located  in  the  front  of  the  meter  the  volumes 
corresponding  with  the  number  of  times  the  collapsible  pistons  are 
inflated  and  deflated.  If  a  piston  is  deflated  the  disc  is  close  to 
the  central  gas  tight  partition  and  a?  it  expands  it  dispels  its 
external  volume  in  gas  from  the  chamber  surrounding  it.  When 
it  is  deflated  the  volume  of  gas  within  it  is  dispelled.  Before 
installing  a  meter  the  volume  dispelled  by  these  pistons  is  care- 
fully calibrated  by  means  (43)  of  a  small  gas  holder,  whose  capacity 

>y  Google 

Manufacture  akd  Distbibution  of  Gas  373 

is  known  by  comparing  its  volume  with  that  of  a  cubic  foot  bottle 
the  accuracy  of  which  is  certified  originally  by  the  Federal  authori- 
ties in  Washington,  If  on  test  the  dial  mechanism  of  a  gae  meter 
indicates  a  measurement  not  within  one  per  cent  of  the  true  volume, 
an  adjusting  device  within  the  meter  provides  for  correction  that 
will  produce  a  final  result  within  that  accuracy. 

By  custom  and  sometimes  by  law  a  gas  meter  {when  tested  on 
complaint  for  inaccuracy)  is  said  to  be  correct  or  accurate  if  it 
measures  within  3%  of  absolute  accuracy,  and  it  is  a  safe  statement 
to  make  that  the  public  buys  no  commodity,  wet  or  dry,  that  so 
closely  meets  the  requirements  of  absolute  accuracy  as  in  the  pur- 
chase of  gas. 

Meters  in  use  are  subject  to  derangement  but,  fortunately  for 
the  consumer,  they  are  more  apt  to  become  slow  (that  is,  the  dial 

>y  Google 

374  Illcmikatinq  Enqixeering 

does  not  indicate  the  total  volume  of  gae  passing  through  the 
meter)  than  fast.  Gas  managers  who  do  not  consistently  maintain 
the  accuracy  of  the  consumers'  meters  will  find  that  their  unac- 
counted for  gas  will  become  larger.  The  certificate  of  a  public 
authority  as  to  the  accuracy  of  the  meter  set  by  a  gas  company  ie 
a  satisfaction  to  the  consumer,  but  no  well  managed  gas  company 
requires  such  inspection  as  a  spur  to  the  maintenance  of  their 
accuracy,  for  the  interests  of  the  stockholders,  whether  private  or 
municipal  has  always  demanded  careful  supervision  of  this  im- 
portant department.  Gas  meters  are  read  when  possible  on  the 
same  date  of  each  succeeding  month,  and  bills  are  thus  rendered 
for  similar  periods,  which  furnishes  a  basis  for  careful  comparison. 
All  consumers  should  learn  to  read  their  meters  themselves;  by  so 
doing  unnecessary  waste  in  their  use  of  gas  is  stopped.  In  a  com- 
munity where  one  apartment  may  be  occupied  by  several  tenants 
during  a  period  of  one  year,  or  where  the  consumer  cannot  pay 
the  necessary  depijsit  required  by  gas  companies,  it  has  been 
found  desirable  to  install  prepayment  gas  meters.  These  meters 
are  the  same  as  other  meters  so  far  as  gas  measurements  are  con- 
cerned. There  is  a  mechanical  contrivance  added  that  permits 
only  that  quantity  of  gas  to  pass  the  meter  that  is  in  value  equal 
to  a  coin  that  is  passed  into  the  coin  bos  by  the  consumer.  As  the 
last  of  this  quantity  of  gas  is  passing  the  meter  the  valve  mechan- 
ism reduces  the  outlet  area  which  reducing  the  size  of  the  gas 
flame  warns  the  consumer  that  an  additional  coin  is  required  in 
the  money  box.  The  coin  attachment  added  to  the  ordinary  meter 
largely  increases  the  cost  of  the  gas  meter,  and  the  coin  collector 
must  be  added  to  the  staff  of  the  gas  company.  The  readings  and 
accounts  of  the  prepayment  meters  must  be  kept  in  the  same  man- 
ner as  the  ordinary-  meters.  Prepayment  meters  should  only  be 
installed  in  locations  convenient  for  immediate  access  and  where 
the  consumer  only  has  control. 

Beyond  the  gas  meter  some  consumers  install  a  small  gas  govern- 
or ;  gas  governors  have  their  place  but  in  practice  when  injudiciously 
installed  they  are  of  little  value.  As  stated  before  the  gas  man- 
ager endeavors  to  furnish  unifonu  pressure  in  a  given  locality'. 
Where  this  is  impossible  a  gas  governor  may  be  profitably  used 
by  the  consumer.  For  photometrical  or  other  delicate  scientific 
work  where  absolutely  uniform  pressure  is  imperative  their  use 
is  necessary.     Gas  pressure  increases  as  the  height  of  a  building 

>y  Google 

Mancfactl're  and  Distribution  of  Gas  375 

increases,  due  to  the  difference  in  specific  gravity  of  gas  when 
compared  with  air;  in  tall  buildings  a  gas  governor  in  the  base- 
ment cannot  serve  uniform  pressure  throughout  the  whole  build- 
ing. The  increase  in  the  pressure  of  coal  gas  may  be  roughly 
stated  for  example  as  one  inch  in  a  difference  in  height  of  one 
hundred  feet;  because  of  this  fact  it  is  customary  to  locate  the  gas 
holder  at  the  lowest  point  in  a  district  to  be  supplied. 

For  lighting  (40)  the  most  common  method  of  burning  gas  is 
by  the  use  of  the  ordinarj'  metal  or  lava  tip  from  which  the  gas  on 
issuing  and  igniting  first  heats  thf  carbon  particles  therein  to  in- 
candescence before  they  are  totally  consumed.  These  burners  are 
the  batswing  burner  in  which  the  gas  issuing  from  a  narrow  slot 

forms  a  thin  sheet  of  flame;  the  fishtail  in  which  two  circular 
streams  of  p;as  coming  in  contact  with  each  other  spread  out  on 
ignition  in  a  similar  sheet  of  flame,  and  the  argaml  burner  where 
a  series  of  small  cylindrical  jets  issue  from  a  multiplicity  of  open- 
ings arranged  in  a  circle  forming  a  cylinder  of  flame  enclosed  by 
a  glass  chimney,  and  air  for  combustion  is  supplied  from  tht'  bottom 
of  the  burner.  These  burners  may  be  expected  to  give  from  three 
to  five  candle  power  per  cubic  foot  of  gas  burned  depending  upon 
the  quality  of  the  gas  used. 

The  gas  mantle  {4'?)  burner  is  displacing  all  other  method" 
of  burning  gas  for  illumination.  In  this  burner  a  partial  mixture 
of  air  and  gas  is  effected  !>efore  the  gas  issues  from  the  burner, 
and  together  with  the  air  present  at  the  burner  outlet  immediately 
affects  complete  combustion  of  the  gas  in  a  bhie  flame  cone,  which 

>y  Google 

376  Illuminating  Enoineerino 

coming  in  contact  with  the  gas  mantle  renders  it  incandescent. 
Such  mantles  are  either  upright  or  inverted  as  the  case  may  be; 
the  candle  power  per  foot  of  gas  is  five  times  that  of  a  flat  flarae 
burner,  but  the  mantles  must  be  of  good  quality  and  properly  ad- 
justed and  maintained  to  give  these  results.  It  is  for  this  reason 
that  many  gas  managers  are  endeavoring  to  maintain  mantle  burn- 
ers for  as  low  a  price  as  possible  to  cover  the  cost  of  this  work. 
In  nearly  all  other  appliances  using  gas  it  is  burned  to  a  blue 
flame  as  in  the  case  of  the  mantle  burner,  and  it  is  this  fact  that 
is  tending  to  make  a  candle  power  requirement  for  gas  obsolete, 

PiQ.  47. 

for  where  a  blue  flame  is  required  only  the  heat  unite  in  the  gas 
are  of  importance.  A  dual  standard  is  illogical  and  imprac- 
ticable for  it  is  a  fact  that  candle  power  and  heat  units  do  not 
rise  and  fall  in  direct  proportion  even  in  the  same  kind  of  gas, 
and  there  is  a  wide  difference  in  the  heat  units  in  different  gases  of 
the  same  candle  power  as  measured  by  the  flat  flarae  or  the  argand 
burner.  The  kitchen  gas  range  burners  all  use  blue  flame,  many  gas 
heating  appliances  likewise,  though  some  radiator  types  of  heaters 
use  a  flat  flame  burner.  Gas  logs  usually  use  no  primary  air  for 
the  reason  that  some  light  is  desirable  to  simulate  a  genuine  wood 
fire.  Gas  to-day  is  used  in  many  more  ways  for  heating  than  for 
lighting,  and  it  has  been  asserted  that  more  gas  is  used  to-day  for 
heating  and  manufacturing  than  for  lighting;  it  is  a  fact  that  the 
proportion  of  gas  used  in  daylight  hours  is  on  the  increase. 

>y  Google 

Manupactcee  and  Distbibdtion  of  Gas  377 

K.  Piping  of  Buildings,  Etc. 

Before  the  era  of  high  buildings,  and  the  advent  of  a  cheap  and 
certain  supply  of  electricity  it  never  occurred  to  the  architect  to 
leave  gas  piping  out  of  buildings,  but  it  is  not  uncommon  to-day — 
and  the  tenedency  is  acquiesced  in  by  our  electric  brethren.  The 
piping  and  wiring  of  a  ten  or  twenty  story  building  for  water,  toilet, 
sprinkler  system,  steam  heat,  refrigeration,  electricity,  telephone 
and  gas  is  a  work  of  no  small  proportions  and  a  great  expense. 
The  natural  impulse  is  to  consider  what  can  be  dispensed  with, 
and  two  methods  of  illumination  being  now  seemingly  unnecessary, 
gas  pipes  are  being  omitted  where  possible.  In  commercial  build- 
ings main  risers  are  installed  for  the  use  of  gas  emergency  hall 
lighting  and  for  manufacturing  purposes  and  in  large  apartment 
buildings  for  cooking  and  heating  and  emergency  hall  lighting. 
Many  thousand  dollars  are  thus  saved  the  builder  and  owner.  The 
consumer  is  then  cut  oif  from  a  choice  of  central  station  illumi- 
nant  in  so  far  as  gas  and  electricity  is  concerned,  though  strange  to 
say  oil  lamps  are  still  in  use  e\en  where  both  gas  and  electricity  are 
available,  and  candles  are  also  used  in  great  numbers.  It  is  a 
mistake  to  leave  gas  piping  out  of  buildings,  but  it  is  unnecessary 
to  pipe  the  large  buildings  so  thoroughly  as  would  needs  be  if  elec- 
tricity were  not  available.  Buildings  which  have  a  private  electric 
plant  should  always  be  piped  for  gas  for  lighting  purposes,  for  it 
is  frequently  needed. 

The  modern  builder  to  save  the  last  dollar  of  construction  coat 
desires  to  save  all  the  steel  possible  by  having  the  dead  floor  weight 
as  small  as  possible  and  hence  cinders  with  and  without  concrete 
are  used  above  tile  or  brick  arches.  The  pipes  are  sometimes  em- 
bedded in  the  cinders  and  concrete.  In  course  of  time  the  sulphur 
in  the  cinders  attacks  the  iron  of  the  pipes  laid  in  the  cinders,  and 
cases  are  many  where  entire  piping  systems  have  had  to  be  aban- 
doned because  of  the  pipes  becoming  unserviceable.  This  deteriora- 
tion in  pipes  is  not  confined  to  those  used  to  conduct  gas.  When 
this  system  is  used  all  pipes  should  be  exposed  or  protected  from 
possible  corrosion. 

In  a  building  in  Xew  York  there  has  been  adopted  a  method  of 
making  the  gas  lighting  and  gas  heating  interdependent  The 
building  is  lighted  by  gas,  its  heat  is  utilized  for  the  heating  and 
when  the  weather  is  cold  thermostats  installed  on  each  fioor  open 
valves  on  radiators  which  thereby  diminishing  by  radiation  the 

>y  Google 


Eteam  pressure  coming  from  a  gas  heating  boiler  in  the  basement 
causes  a  valve  actuated  by  steam  pressure  to  admit  more  gas 
to  the  gas  burners  furnishing  heat  to  the  heating  surfaces  of  the 
boiler,  thus  supplying  to  the  floor  requiring  it  additional  heat. 
Shutting  off  all  light  would  still  further  increase  the  gas  burned 
under  the  boiler,  while  increasing  the  gas  light  would  automatically 
decrease  the  gas  used  under  the  boiler.  This  building  is  ventilated 
thoroughly  with  the  exhaust  fans  used  in  the  modern  systems  of 
building  ventilation.  The  results  arc  satisfactory  at  this  writing. 
The  gradual  extended  use  of  mantle  lighting  tends  to  a  return  of 
the  former  general  practice  of  piping  buildings  for  gas  whether 
electricity  is  used  or  not.  It  is  well  that  this  is  so.  The  writer 
believes  that  all  buildings  should  be  thoroughly  piped  for  the  use 
of  gas. 

L.  Appliances  Used  for  Burning  Gas 

This  subject  is  rather  beyond  the  limits  of  this  paper  and  I  will 
content  myself  with  the  enumeration  of  the  uses  of  gas  as  follows; 

For  Ir^hting^ — By  flat  flame,  argand  and  gas  mantle  burners, 
and  through  the  agency  of  the  gas  engine,  electric  lighting  is 

For  HonBehold  tTse  {48a)  Cooking,  heating,  gas  ironing,  hot  water 
heating  and  heat  for  many  small  appliances,  such  as  coffee  pots, 
chafing  dishes,  hot  water  kettles,  curling  irons,  etc. 

For  Commercial  TTsei  Hot  water,  instantaneous  and  automatic 
(48b),  hotel  ranges,  broilers,  caldrons,  engines,  smelters,  melting 
furnaces,  singeing  furnaces,  china  firing,  smoke  houses,  biscuit 
baking,  steam  boilers  for  feather  and  hat  manufacturers,  as  well 
as  for  heating,  gas  irons  and  mangles,  washing  machines  and  a 
multiplicity  of  other  uses  which  are  increasing  daily. 

For  commercial  uses  it  is  common  both  for  light  and  heat  to 
have  the  gas  under  increased  pressure,  or  the  air  supply  under  pres- 
sure, or  both  and  in  most  appliances  primary  air  is  mixed  witli 
gas  but  not  in  sufficient  amount  to  make  a  mixture  that  would  be 
explosive  as  in  the  case  of  the  gas  engine.  When  increased  pres- 
sure is  used  for  lighting  the  efficiency  of  the  gas  mantle  light  in  the 
use  of  gas  is  doubled.  Any  method  that  uses  the  minimum  amount 
of  air  for  complete  combustion  will  give  the  maximum  temperature 
and  hence  increased  efficiency  per  unit  of  gas  used.  The  steam 
boiler  is  vastly  different  in  its  efficiency  depending  upon  how  many 
pounds  of  air  is  used  per  pound  of  fuel,  and  the  same  principle 
applies  to  the  use  of  gas  as  fuel. 

>y  Google 

Manufacture  and  Distribution  of  Gas 

>y  Google 

380  Illuuinatinq  Enoineerino 

M.  Influences  Qoveming  the  Selection  of  a  Particular  Type  of 
Oas  for  Adoption 

First:  The  first  consideration  should  be  the  laws  exieting  or 
that  reasonably  may  be  expected  to  be  enacted ;  if  the  public  de- 
mand is  for  a  high  candle  power,  to  be  obtained  by  a  flat  flame 
burner,  carburetted  water  gas  best  fills  the  requirements  but  in  a 
very  small  community  acetylene  gas  is  the  choice  if  calcium  carbide 
is  readily  obtainable. 

Second:  A  second  consideration  is  an  adequate  certain  and 
cheap  supply  of  raw  materials.  In  connection  with  this  your  atten- 
tion is  directed  to  a  map  issued  by  the  U.  S.  Geological  Survey, 
entitled  "  Known  productive  Oil  and  Gas  Fields  of  the  United 
States  in  1908  "  and  a  second,  entitled  "  Coal  Fields  of  the  United 
States"  probably  also  1908. 

In  the  manufacture  of  carburetted  water  gas  unless  a  hard  fuel, 
either  anthracite  coal  or  oven  coke,  as  well  as  oil  is  available  then 
coal  gas  would  have  to  form  part  of  the  gas  plant  if  for  no  other 
reason  than  to  supply  retort  coke  for  the  water  gas  generators. 

If  oil  is  available  in  great  quantities  at  a  very  reasonable  price 
as  in  California  then  the  oil  gas  referred  to  under  head  1)5  and 
Eo  may  be  chosen. 

If  gas  coal  is  very  cheap  then  coal  gas  might  be  the  choice,  ex- 
cluding carburetted  water  gas  or  oil  gas,  provided  the  candle  power 
provisions  do  not  make  it  imperative  to  use  these  gases. 

Third:  A  third  consideration  is  the  variation  in  the  demand 
which  may  compel  the  use  of  carburetted  water  gas,  even  where 
coal  gas  is  the  natural  choice,  for  the  ease  of  supplying  sudden 
large  demands,  and  the  small  standby  cost  for  materials  combined 
with  the  smaller  capital  cost  per  unit  capacity  may  make  this  gas 
cheaper  to  use  in  part,  even  if  its  cost  per  unit  made  seems  actually 
greater  than  the  cost  of  the  same  unit  of  coal  gas.  The  coal  gas 
plants  of  large  cities  in  all  parts  of  the  country,  even  in  the  bi- 
tuminous coal  regions,  are  for  this  reason  supplied  with  carburetted 
water  gas  plants. 

Fourth:  The  capital  charge  may  oftentimes  determine  choice. 
Where  the  difference  in  cost  of  production  is  in  favor  of  coal  gas 
remember  that  the  capital  cost  of  equivalent  capacity  of  water  gas 
is  materially  less  than  coal  gas  and  when  the  productive  cost  and 
capital  cost  are  combined  the  choice  may  compel  the  use  of  car- 
buretted water  gas. 

>y  Google 

MaNCFACTDRS  and   DiSTHIBDnON   OF   GA8  381 

Fifth:  The  land  area  available  is  of  importance  for  the  space 
occupied  by  a  coke  oven  plant  is  much  larger  than  a  retort  coal 
gas  plant  of  equal  capacity  and  a  carburetted  water  gas  plant  very 
much  smaller  than  either.  Where  a  small  site  area  is  the  governing 
feature  then  carburetted  water  gas  may  be  the  type  of  gas  best 
adapted  for  the  conditions. 

Sixth:  When  labor  is  Bcarce  and  remuneration  high  that  process 
demanding  least  maniial  labor  may  be  the  type  to  be  chosen,  Car- 
buretted water  gas  frequently  best  fills  this  requirement. 

Fio.  49.— The  First  House  Lighted  by  Gas  In  England. 

Seventh:  Wliere  the  demand  for  metallurgical  coke  i?  great  and 
imperative,  the  coke  oven  procesi'  will  be  installed  in  any  event, 
in  which  case  one  of  its  by-products,  gas,  may  be  utilized  for  the 
supply  of  gas  in  its  neighborhood.  A  carburetted  water  gas  plant 
is  in  this  case  a  useful  auxiliary,  for  in  dull  biisinei>s  seasons  when 
coke  is  a  drug  on  the  market,  the  coke  ovens  may  necessarily  be 
shut  down  and  the  gas  demand  may  be  obtained  for  the  time  being 
from  the  water  gas  plant.    Be  sure  that  the  coke  demand  is  reason- 

>y  Google 

382  Illumixatisq  Exqineerinq 

ably  certain  before  building  a  coke  oven  plant  for  fnrniphing  a 
commercial  gas  supply,  or  financial  disaster  may  result. 

Eighth:  The  only  available  site  for  the  gas  works  may  be  in 
H  district  where  the  choice  may  be  the  plant  producing  the  minimum 
amount  of  dust  and  dirt.  Here  acetylene  gas  or  oil  gas  or  car- 
burctted  water  gas  may  be  the  choice  to  be  made. 

Ninth:  In  addition  to  the  choice  of  the  type  of  gas  liest  adapted 
to  the  situation  there  may  be  here  included  a  consideration  of  the 
system  of  distribution  to  be  adopted.     For  siiburban  districts  with 

Flu.  50 —The  First  House  Lighted  by  Gas  In  Baltimore. 

villages  quite  widely  separated,  each  too  small  to  support  a  gas 
plant,  a  high  pressure  system  may  be  installed  and  the  manufac- 
turing plant  located  in  the  town  best  situated  commercially  for 
economical   production  costs. 

Tenth:  Some  states  have  seen  fit  lo  ])a8s  laws  as  to  maxiniuni 
price  at  which  gas  may  be  sold.  In  such  cases  it  would  be  well 
to  consider  fully  this  rather  unusual  condition,  for  it  might  well  be 
that  while  the  community  might  be  glad  to  have  a  supply  of  gas 
even  at  greater  charge  than  the  law  permits,  it  might  be  many 

>y  Google 


years  before  a  plant  would  pay  a  return  at  the  maximum  per- 
missible price. 

Eleventh:  A  word  of  warning  with  respect  to  a  new  plant;  <lo 
not  make  the  mistake  of  not  looking  many,  many  years  into  the 
future.  Lay  the  plant  out  on  paper  for  the  future  extensions  that 
Rre  in  prospect,  for  if  they  are  not  to  be  expected,  consider  care- 
fully whether  the  plant  should  be  built  at  all.  It  does  not  cost 
money  to  look  ahead  and  design  a  plant  showing  what  is  to  be  done 
in  the  future — but  it  costs  money  to  build  a  plant  to-day  and  within 

Fio.  51.— The  First  House  Lighted  by  Gas  In  New  York  City. 

ten  years  tear  it  down  to  build  a  second  and  then  repeat  the  process. 
Depreciation  by  inadequacy  is  a  cost  which  is  present  in  the  ma- 
jority of  undertakings  in  a  growing  country.  Minimize  this  so 
far  as  possible.  Do  not  actually  build  for  business  so  far  in  the 
future  that  interest  and  depreciation  e-\cce<ls  the  saving  made  by 
erecting  the  structures  in  a  single  operation.  Depreciation  by 
obsolescence  cannot  he  foreseen;  be  careful  to  install  tried  up-to- 
date  apparatus. 

>y  Google 


a  a 

■<  S 


i|  IS  11 







11  g  §  §11  §1  H  IS 
:|  i  s  s  ass  g  s|  es 


I  IS  I  I  I  gli  |i  ii  ii  I 

g  3|  I  S  S  s-'S  g|  gS  Sg  ^ 

"S  33  ••"  33  "8  S  S  S  a"S  S5  St:   28  ' 


s  s  s  isl  as 


Manupaotuhe  and  Distbibdtion  of  Gas  385 


■■■■        ■■     ■■        ■■     ■     ■■        ;i        I        !l     i;     i:     :    1   " 






_  _  :  J  i 






a  "s  s-j  33  e  s  -J  -'s  "a  =■•  b*  s»  -  1 1 

N  in  IN  Mi  :H  H  H  H  H 






386  Illumikatino  Enoiseshino 

N.  The  Future  of  the  Artificial  Oas  Businese 
Judging  by  the  past  the  gas  busineee  is  destined  to  grow  in  the 
future  as  in  the  past.  Gas  stockholdeis  received  a  severe  fright 
when  electric  tight  was  introduced,  but  there  is  little  to  fear  for 
the  growth  of  the  gas  businesB  because  of  electricity  until  someone 
invents  an  economical  and  successful  process  to  manufacture  elec- 
tricity direct  from  coal.  The  fact  that  gas  continues  to  hold  its 
own  even  where  electricity  is  manufactured  on  a  large  scale  with 
water  power  makes  it  unlikely  that  even  such  an  invention  will 
seriously  retard  the  growth  of  the  gas  business.  Gas  for  all  pur- 
poses where  heat  units  are  essential  is  more  economical  and  is  more 
than  holding  its  own  with  electricity  up  to  date.  It  is  quite  clear 
that  the  candle  power  provisions  for  gas  are  bound  to  be  eliminated 
from  legal  requirements  if  for  no  other  reason  than  that  the  burner 
now  specified  in  many  statutes  will  sooner  or  later  go  out  of  com- 
mon use  as  the  argand  burner  has  disappeared.  Should  the  supply 
of  available  oil  be  removed,  as  naphtha  within  twenty  years,  then 
high  candle  power  requirements  must  go.  The  cornerstone  of  the 
gas  business  of  the  future,  as  it  was  before  the  discovery  of  oil,  ia 
gas  made  from  bituminous  coal  and  that  reason  has  influenced 
the  selection  of  a  coal  gas  plant  primarily  for  illustration,  but  not 
only  for  that  reason  but  because  its  auxiliary  machinery  and  manu- 
facture is  the  more  complex. 

While  the  use  of  coal  gas  may  increase  as  pointed  out  heretofore 
it  is  necessary  at  all  times,  though  more  particularly  in  large 
cities,  to  have  a  water  gas  plant  in  combination  with  coal  gas  to 
meet  sudden  demands  for  large  quantities  of  gas  due  to  atmospheric 
variations  from  day  to  day,  and  further  the  use  of  gas  has  displaced 
in  so  many  ways  the  use  of  gas  coke  that  the  utiliiiation  of  this  by- 
product of  coal  gas  by  its  manufacture  into  water  gas  seems  im- 

At  this  writing  the  supply  of  oil  shows  no  prospect  of  failing  for 
gas  making  purposes  and  the  use  of  carburetted  water  gas  ia  in- 
creasing, but  in  a  paper  of  this  kind  a  reference  to  such  a  remote 
possibility,  in  view  of  the  constantly  increased  use  of  the  oil  produc- 
tion for  other  purposes.iban  gas  making,  seems  not  out  of  place. 

>y  Google 



Bt  Edwaed  B.  Hosa. 


I.  I>R0TOHETBic  Unira  and  Nohchclatcbe 

The  luminoat  fiux  (F)  flovlng  awiv  from  a  light  source  falls  upon  and 

lUumlnatea  other  bodies,  the  illumlTtation  (E)  being  the  flux  per 

unit  of  area.'  The  flux  per  unit  BOlld  angle  Is  the  Intensity  (I>. 

F  la  measured  In  lumens.  I  In  candles.    The  flux  per  unit  of  area 

from  a  surface  Is  the  radiation  (B')- 
Tbe  luminous  Hux  Is  the  radiant  power  multiplied  br  the  stlmulna  co- 

efflclent,  vhlch  Is  a  function  of  wave  lengUi. 
The  mean  spberlcal  Intensity  Ii  Is  the  total  flux  F  (In  lumens)  divided 

by  4r;  the  Intensity  I  In  a  particular  direction  Is  proportional  to 

the  rate  of  flux  in  that  direction. 
By  analogy  with  an  electrically  charged  body,  the  total  quanttty  of 

light  Q  on  a  body  is  the  surface  Integral  of  the  speclflc  Intensity  e. 

Total  flux  F  equals  vQ. 
Case  of  extended  sources.     Disk,  plane,  cylinder,  sphere.    Law  of  in- 
verse squares  for  case  of  sphere  and  circular  dlak. 
Reciprocal  relations  between  radiating  bodies.     LnmtiMiis  flux  within 

an  enclosure.    Equations  of  definition  of  photometric  magnitudes. 


A  photometric  standard  is  a  standard  of  light  flux,  either  Its  total  flux 
or  (more  commonly)  Its  rate  of  flux  In  a  particular  direction  being 

The  International  candle  is  a  unit  and  not  a  standard. 

The  two  kinds  of  primary  standards  employed  In  physical  Kicasure- 
ments;  (1)  those  which  are  verified  or  reproduced  from  standard 
specifications,  and  (2)  those  which  are  arbitrary  and  cannot  be  m 
reproduced.    Examples  of  such  standards. 

Flame  standards  are  primary  photometric  standards  of  the  first  kind, 
although  aa  used  are  often  considered  as  of  the  second  kind.  In- 
candescent electric  lamps  are  generally  employed  as  secondary 
standards,  but  are  sometlmet  used  as  primary  standards  of  the 
second  kind.  Other  primary  standards. 

>y  Google 


The  most  Important  flame  sttmdards  are  the  Harcourt  pentane  lamii 
and  the  Hefner  amyl  acetate  lamp.  Advantages  and  dlBadvaiitBgee 
of  each  dlscutwed.  Dlfflcultfea  of  both  11©  (1)  partly  in  the  lamp, 
(2)  partly  tn  the  fuel,  and  (3)  partly  In  the  atmosphere  In  which 
combustion  takes  place.  Bach  of  these  three  queetions  discussed 
for  each  lamii.  The  pentane  l&mp  gives  more  consistent  results  for 
a  single  lamp,  but  different  lamps  disagree  more  than  Is  found  tor 
the  Hefner.  The  Important  difference  between  a  primary  standard 
and  a  working  standard. 
Preparation  and  calibration  of  electric  lamps  as  photometric  standards. 
Their  performance  as  precision  standards,  primary  or  secondary. 
Method  ot  measurement.  Direction  of  improvement  In  primary 


A  disciiBsion  of  photometric  unite  and  standards  may  be  divided 

iato  two  separate  parts,  the  first  including  photometric  units  and 

nomenclature,  and  the  second  primary  and  secondary  photometric 


The  development  of  the  subject  of  photometric  units  and  nomen- 
clature received  a  notable  impulse  through  the  paper  of  Professoi 
Blondel,  presented  to  the  Geneva  Congress  of  1896.  Since  that 
time  various  modifications  of  the  proposals  then  made  have  been 
pnt  forward,  but  no  authoritative  action  on  the  subject  has  ever 
been  taken  by  any  national  or  international  body.  The  nomen- 
clature as  approved  by  the  Geneva  Congress  has,  in  part,  come  into 
general  use.  There  has,  however,  been  a  tendency  to  recognize  as 
few  separate  photometric  quantities  as  possible,  and  some  of  them 
have  been  employed  rather  loosely  in  more  than  one  sense.  This 
is  partly  at  least  due  to  a  lack  of  clearness  in  the  perception  of 
the  physical  relations  of  the  various  photometric  quantities. 

I.  Units  aiid  ITomenclatdee 
1.  Case  of  Point  Source 
We  start  with  the  idea  of  light  as  a  luminovs  fltix  radiating  or 
flowing  away  from  the  source,  and  illuminating  bodies  as  it  falls 
upon  them.  In  the  simple  case  of  a  symmetrical  point  source  the 
fiuz  is  equal  in  all  directions,  and  since  the  entire  flux  falls  uni- 
formly upon  the  interior  surface  of  any  concentric  sphere,  the 
quantity  of  the  luminous  flui  per  unit  of  area  is  inversely  pro- 
portional to  the  square  of  the  distance  from  the  source,  a  law  which 
has  been  verified  by  experiment.     The  quantity  of  the  luminoos 

>y  Google 

Photometric  Units  and  STANDAHre  389 

flux  per  unit  of  area,  or  the  Hux  density  at  the  surface  of  the 
illuminated  body,  is  by  definition  the  specific  illumination  E.  If 
we  represent  the  total  flux  by  F  we  have,  therefore, 

where  r  is  the  distance  from  the  point  source  to  the  body  il- 

Representing  —  by  a  single  letter  I,  we  have 

E=~  m 


P^a.  1. — Hollow  sphere  of  radius  r,  and  surface  irr*,  with  a  ■ymmetrlcal 
point  source  at  center,  has  a  total  tux  F  nnlformlr  distributed  over  It 

I  is  called  the  intensity  of  the  source,  and  is  equal  to  the  flux 
per  unit  of  solid  angle. 

The  illumination  is  equal  to  the  intensity  of  the  source  divided 
by  the  square  of  the  distance  (equation  S),  and  the  total  flux  is 
4ir  times  the  intensity  (3). 

The  intensity  I  is  measured  in  candles*  the  flux  F  in  lumens, 
and  the  distance  r  in  centimeters.  Thus,  from  a  point  source  of 
intensity  I  candles,  there  is  a  luminous  flux  4jrl  lumens. 

■  It  Is  proposed  to  call  the  new  valne  of  the  American  candle,  whicb 
Is  the  same  as  the  BngUsta  candle  and  the  French  bougie  decimals,  and 
which  Ii  also  used  b;  several  other  countries,  the  international  candJs. 

>y  Google 

390  iLLDUItfATING   £)N01NEEBIN0 

The  flux  density  ia  the  luminouB  flux  per  unit  of  area  (normal 
to  the  flux  in  the  case  of  a  point  source),  or  the  total  flux  F  over 
an  area  S  divided  by  the  area.  If  the  flux  denaity  is  variable,  S 
will  be  a  very  small  area,  and  F  the  flux  over  that  small  area. 
Thus  a  pencil  of  flux  Fi  from  a  point  source  falls  on  a  small  area, 
S„  about  the  point  P,  and  the  surface  density  or  illumination  is 

g^-    (Kg.  8.) 

$.  Definiiion  of  Intensity  for  Unsymmetrical  Sources 
For  a  symmetrical  point  source  the  intensity  I  has  been  defined 
as  the  total  flux  F  divided  by  4«-.  If  the  source  is  not  symmetrical, 
but  sends  out  a  total  luminous  flux  F  unequally  in  different  direc- 
tions, then  the  mean  value  of  the  intensity  (F  divided  by  4x)  is 
called  the  mean  spherical  intensity  I,.  We  thus  define  the  mean 
spherical  intensity  with  respect  to  the  total  flux;  and,  similarly, 
the  mean  hemispherical  intensity  is  the  ratio  of  the  flux  through 
a  hemisphere  to  the  solid  angle  Zt,  or  the  average  flux  per  unit 
solid  angle  throughout  a  hemisphere. 

The  intensity  1  in  any  particular  direction  la  the  quotient  of 
the  flux  F»  through  a  small  solid  angle  a  in  that  direction  divided 
by  the  angle.    Thus 

I  =  ^ ,  111  being  a  solid  angle. 

Thus,  if  the  pencil  of  the  flux  (Fig.  2)  occupies  an  angle  »,  the 

intensity  I  of  the  source  A  in  the  direction  of  P  is  — ,     Thus 

the  intensity  I  is  the  angular  density  of  the  flux,  as  the  illumina- 
tion E  is  the  surface  density.  Thus  the  expression  "  the  intensity 
of  a  source  in  a  particular  direction "  means  the  angular  density 

>y  Google 

Photometbic  tTNiiB  ASD  Standabdo  391 

oi  the  flu£  in  that  direction.  Therefore,  both  E  and  I  are  flux 
ratios,  lumens  per  unit  area  and  Inmeua  per  unit  solid  angle,  re- 
spectively.    One  lumen  per  square  meter  is  the  lux,  and  1  Imnen 

per  unit  solid  angle  *  (-r-oi  the  total  angle  about  a  point)  is  a 

If  the  source  is  not  a  point  but  a  small  sphere  of  radius  a,  the 
flux  4irl  passes  out  from  a  radiant  surface  iwa,'.  Thus  the  flux 
density  of  radiation,  or  the  specific  radiation,  is 

8        4^a»       a'      ^  *   ' 

ThuB  we  may  speak  generally  of  the  luminous  flux  at  any  point 
in  space,  and  of  the  flux  density  of  such  radiation.  If  it  falls  on 
a  material  surface  the  incident  flux  density  is  the  specific  illtimi- 
nation  E;  as  it  comes  from  a  luminous  or  other  radiating  or  dif- 
fusing surface,  the  flux  density  is  the  specific  radiaiion  E*.  Al- 
though E  and  E'  are  quantities  of  the  same  nature,  it  is  con- 
venient thus  to  distinguish  them,  and  for  brevity  we  may  often 
omit  the  adjective  "specific." 

The  luminous  flux  density  in  space  is  analogous  to  electric  dis- 
placement in  electrostatics,  which  is  represented  graphically  in 
direction  and  magnitude  by  lines  of  force,  which  start  from  posi- 
tive electricity  and  terminate  upon  negative  electricity.  .  We  think 
of  an  electric  displacement  as  occurring  in  space  between  two  elec- 
tric charges,  but  a  surface  density  of  electricity  occurs  only  where 
there  is  a  material  conducting  body  on  which  the  lines  of  force 
terminate.  In  the  same  way  Hie  terms  luminous  flux  and  flux 
density  apply  generally,  both  at  the  surface  of  the  luminous  and 
the  illuminated  bodies,  and  in  the  space  between.  The  radiation 
is  the  flux  density  at  the  source  of  the  flux,  and  the  Ulamination 
is  the  flux  density  or  flux  per  unit  area  on  the  surface  where  the 
luminous  flux  is  received. 

S.  Distinction  between  Luminous  Flux  and  Energy 

The  total  luminous  flux  F  is  not  to  be  confused  with  the  total 

energy  flowing  from  a  luminous  body.     Luminous  flux,  or  lighi, 

as  we  ordinarily  say,  is  the  physical  stimulus  which  applied  to  the 

■  This  Is  an  aDgle  subtended  by  ^r  of  a  spherical  surface,  and  in  the 
case  where  the  solid  angle  le  a  circular  coae,  Its  section  ttirough  tlie 
apex  is  a  plane  angle  of  66°  32'  28'. 

>y  Google 


retina  prodnces  the  eenBation  of  light.  It  is  equal  to  the  radiant 
power  multiplied  by  the  stimulus  coefBcient.  This  stimulus  co- 
efficient is  different  for  every  different  wave  frequency  or  wave 
length,  and  is,  of  course,  zero  for  all  frequencies  outside  of  the 
visible  spectrum.  Hence,  if  \V»  is  the  power  (expressed  in  watts) 
for  unit  of  wave  length  of  the  specfrum,  and  Ka  is  the  stimulus  co- 
efficient or  luminous  efficiency  whose  value  varies  with  the  wave 
length  A,  we  have  for  the  total  power  radiated  from  a  body 

W=)W,dA,  (5) 

the  integration  being  carried  through  the  whole  range  of  wave 
lengths,  including  non- luminous  radiation. 

For  the  luminous  flux, 

P=JEj.WxdA,  (6) 

the  integration  being  throughout  the  visible  spectrum,  K  being 
zero  elsewhere. 

As  the  values  of  K*  throughout  the  spectrum  are  not  accurately 
known,  it  is  not  possible  to  calculate  F  in  general.  But  by  meas- 
uring W  in  watts  and  F  in  lumens,  we  can  determine  the  ratio 
of  the  luminous  flux  to  the  radiant  power  in  any  paMicuIar  case. 
One  may  properly  say  that  luminous  flux  is  due  to  and  is  always 
associated  with  radiant  power,  but  luminous  flux  and  radiant  power 
cannot,  in  general,  be  converted  into  one  another  like  feet  and 
inches ;  for,  as  stated  above,  the  conversion  factor,  the  stimulus 
coefficient  or  luminous  efficiency,  is  not  a  constant  like  the  ratio 
of  feet  to  inches,  but  is  variable,  having  a  different  value  for  every 
different  wave  length  in  the  visible  spectrum  and  falling  to  zero 
outside  the  visible  spectrum.  "  Luminous  energy  "  should,  there- 
fore, not  be  used  as  synonymous  with  "  luminous  flux." 

i.  Unit  Disk 
Concerning  a  body  charged  with  electricity,  we  have  the  two 
ideas,  (1)  the  electricity  of  density  <t  and  total  quantity  Q  on  the 
surface  of  the  charged  body,  and  (2)  the  flui  of  force  throughout 
the  surrounding  space,  there  being  4)rQ  lines  of  force  for  a  quan- 
tity Q  of  electricity.  We  do  not  believe  in  the  fluid  theory  of  elec- 
tricity in  the  same  way  that  Franklin  did,  but  we  nevertheless  find 
the  idea  of  a  surface  density  of  electricity  very  useful.  In  the 
corresponding  case  with  light  we  may  have  similarly  two  distinct 
ideas,  (1)  a  surface  distribution  of  light  over  a  luminous  area 
of  density  or  specific  quantity  b,  and  total  quantity  Q,  and  (2)  a 

>y  Google 

Photometbio  Units  and  Standabds 


luminous  flux  filling  the  Burronnding  space  and  producing  an  il- 
lumination E  on  any  body  equal  to  the  flux  per  unit  of  area. 

We  have  so  far  defined  illuminatioQ  and  intensity  in  terms  of 
the  flux.  Let  us  now  obtain  their  values  in  terms  of  the  quantity 
of  light  on  the  surface  of  the  luminous  source. 

The  illumination  from  a  very  small  source  is  inversely  propor- 
tional to  the  square  of  the  distance  from  the  source,  and  directly 
proportional  to  the  brightness  of  the  source.  Hence,  for  a  luminous 
plane  of  unit  area,  we  may  write 

'  whra«  b  is  the  total  quantity  of  light  on  the 
disk  of  unit  area,  which  we  define  as  the 
brightness,  and  the  radiation  to  F,  at  a  dis- 
tance r  is  normal  (Fig,  3).  For  a  point  P,  at 
an  angle  e  from  the  normal,  the  illumination 
would  be  (approximately)  proportional  to  -yf^ 
the  cosine  of  the  angle  e ;  if  the  area  of  the 
disk  is  S  we  should  have 


.  bS  cos  e  _ 

Q  cos  e 


Q  is  the  quantity  of  light  on  the  small  disk  of 
area  S,  and  is  equal  to  bS  (Fig.  3) . 

The  total  flux  *  over  the  hemisphere  illuminated  by  the  disk 
is  ttQ. 

Thus  the  total  luminous  flux  F  from  a  small  plane  disk  is  v 
times  the  quantity  of  light  Q  on  the  disk. 

*TblB  iB  found  by  Integrating  the  expreeelan  tor  E  over  tbc  hem- 
isptiere.    TbOB. 

„      f' .    .       ,        „  f2«T' sine  COB  ede 

P=  1 1  B2xT*  Bin  e  de  =  Q  I 

F  =  rQ  [9ln'e]>=»Q.  iq\ 

In  electrostatlCB  tbere  are  2rQ  lines  of  force  on  each  aide  of  a  disk 
charged  with  Q  units  of  electricity,  or  4irQ  total.  In  the  case  of  lu- 
minous flux,  the  flux  Ib  on  one  side  only,  and  owing  to  the  cosine  factor 
the  total  Is  only  one-half  of  what  It  would  be  otberwiae.  Thua,  the 
total  ts  only  one-fourtb  of  tbe  flux  in  the  electrical  caae. 

>y  Google 


The  average  illnminatioii  over  the  hemisphere  of  radius  r  is 
2irr»-  2    r* 

J-  whereas  the  masimum  illumination  E»  normal  to  the 

disk  is -3-.    Thus  the  mean  is  half  the  maximum.    The  intensity 

I  has  been  defined  as  the  angular  rate  of  flux  in  any  particular 
direction.  It  is,  therefore,  proportional  to  the  illumination  pro- 
duced in  the  given  direction.  Thus,  in  the  case  of  the  luminous 
disk  WR  have 

In = maximum  intensity,  normal  =  Q, 

I«  =  meBn  hemispherical  intensity^  -^  .  (10) 

I(=mean  spherical  intenBity=  -^  , 
Thus  P=rf,  =  4,I,.  (11) 

That  is,  the  intensity  is  numerically  equal  to  the  total  quantity  of 
light  on  the  small  disk  for  oil  points  on  the  normal.  It  decreases 
to  zero  as  we  pass  90°  away  from  the  normal,  having  a  mean  value 
of  half  the  maximum  for  the  whole  hemisphere,  and  is  on  the 
average  only  one-fourth  the  maximum  for  the  whole  sphere.  We 
may,  therefore,  say  that  the  hemiaphericai  reduction  factor  for  the 
disk  is  one-half,  and  the  mean  spherical  reduction  factor  is  one- 
fourth,  the  disk  being  supposed  luminous  on  one  side  only. 

Since  the  total  flux  F  from  an  area  is  irQ,  where  Q  is  the  quan- 
tity of  light  on  the  area,  the  flux  from  a  unit  of  area  is  n-b.  This 
is  the  radiation  E'.    Hence,  in  general, 

E'=Tb.  (12) 

For  a  email  sphere  of  radius  a  the  total  flux  is 
P  =  E'  X  surface. 


-•■I  =  -f-  (13) 

That  is,  for  a  unit  sphere  *  the  intensity  is  one-fourth  the  quantity 
of  light  on  the  sphere.  If  the  distribution  of  light  over  the  sphere 
is  not  uniform,  the  mean  spherical  intensity  is  still  one-fourth  the 
total  quantity  of  light  on  the  sphere,  as  it  is  also  for  a  disk.    In 

*  By  unit  sphere  or  unit  disk,  we  mean  a  disk  or  apfaere.  the  linear 
dimensloas  of  which  are  negligible  Id  comparison  witii  the  distance 
from  source  to  receiver. 

>y  Google 

Photohetrig  Units  and  SiANDAJiDa  396 

other  words,  a  sphere  produces  the  same  illumiuatioii  at  a  given 
point  OB  a  disk  of  the  same  diameter  and  same  brightness  placed 
80  that  the  radiation  from  the  disk  to  the  point  is  normal. 

5.  Extended  Sources 
(a)  Ciroolai  Diik.  I^et  AOB  represent  a  circular  disk,  lumi- 
noQB  on  one  side,  of  diameter  AB,  perpendicular  to  the  paper. 
Each  element  of  the  area  sends  out  luminous  flux  toward  the  right 
in  all  directions  (Fig.  4).  Let  us  consider  how  much  of  this  total 
ftuz  falls  upon  a  surface  of  unit  area  at  P  at  a  distance  r  perpen- 
dicular to  the  center  of  the  disk.     The  intensity  of  the  radiation 

in  any  direction  is  assumed  proportional  to  the  cosine  of  the  angle 
of  emission,  the  radiation  falling  on  the  surface  at  P  is  also  as- 
sumed proportional  to  the  cosine  of  the  angle  of  incidence.  Hence, 
the  flux  falling  on  the  area  at  P  is  less  from  the  outer  portions 
of  the  luminous  disk  AOB  than  from  the  center,  not  only  because 
the  distance  is  greater,  but  also  because  the  two  cosine  factors  are 
less  than  unity.  Summing  up  the  radiation  from  the  whole  disk, 
we  find  that  the  flui  falling  on  unit  area  at  P,  which  is  the  U- 
lumination,  is 

r'-Ha*        d' 
where  Q  is  the  total  quantity  of  light  on  the  disk,  and  d  is  the 
distance  to  the'  edge  of  the  disk." 

*  This  Is  Bliovn  by  Integrating  over  the  disk.    See  paper  lA  Transac- 
tlona  III.  Sng.  Soc.,  June.  1910.  p.  479. 

>y  Google 

896  Illuicinatihg  Enqineebinq 

We  cannot  define  the  intensity  I  of  the  disk  in  the  eame  way 
we  have  for  a  point  or  a  unit  disk,  for  the  radiation  is  not  in  a 
diverging  pencil,  aa  from  a  point.  We  can,  however,  define  the 
equivalent  intensity  !„  as  the  intensity  of  a  small  source  at  the 
center  0,  which  would  give  the  same  illnmination  at  P.  If  all 
the  light  on  the  disk  were  concentrated  near  0  the  illumination  at 
P  would  be  greater  than  that  due  to  the  disk.    But,  if  a  smaller 

quantity,  Qo=Q  -35-  were  concentrated  at  0,  the  illumination  at  P 
wonld  be  the  same.  Hence,  the  equivalent  intensity  !«  of  Uie  disk 
for  the  point  P  is  Q  -3^ .    But  if  the  point  P  be  moved  nearer  the 

disk,  the  equivalent  intensity  of  the  disk  is  leee  than  this,  for-jj 
will  be  smaller,  and  if  the  point  P  be  further  away  Ig  will  be 
greater.  Thus,  the  equivalent  intensity  of  an  extended  luminous 
disk  depends  on  the  place  at  which  the  fiux  is  being  received,  in- 
stead of  being  constant  for  all  distances  as  it  is  for  a  point  or  a 
sphere.  In  general,  the  intensity  I,  or  the  angular  density  of  the 
luminous  flux,  does  not  apply  to  extended  sourcee.  The  quantity 
of  light  Q,  however,  has  a  definite  meaning  in  every  case.  It  is 
the  surface  integral  of  b,  the  brightness,  and  is  not  only  a  very 
useful  quantity  to  employ  in  certain  calculations,  but  tends  to 
fi^  our  ideas  concerning  luminous  sources  and  facilitates  exact 

In  the  case  of  a  luminous  cylinder  of  radius  a  and  length  1,  the 
quantity  of  light  upon  the  convex  surface  is  Q=2)ralb,  b  being  the 
brightness.  The  horizontal  illumination  at  a  distance  r,  large  in 
comparison  with  the  length  of  the  cylinder,  is 

The  total  luminous  flux  P  from  the  cylinder  is  wQ,  and,  therefore, 
the  mean  spherical  illuminatioD  on  the  inner  surfaces  of  a  con- 
centric sphere  of  radius  r  is 

E.=  -!S-  .  -3- 

The  spherical  reduction  factor  f  for  the  cylinder  is  the  ratio  E. 
divided  by  E*.    Therefore) 

f-  ^^  +  -^  -  -|-  =0.7854=78.5^  approximately. 

Thus  an  incandescent  lamp  of  one  or  more  straight  filaments 


Photoubtbio  Units  aud  Stahsasds  397 

Bhoold  have  a  spherical  reduction  factor  of  78.5  per  cent.  Thia 
is  nearly  the  value  for  the  tantalum  and  tungsten  lamps,  the  base 
of  the  lamp  catting  off  some  light,  and  bo  making  it  elightly  leae. 
A  round  disk,  luminous  on  both  surfaces,  has  a  spherical  reduction 
factor  of  50  per  cent.  This,  of  course,  assumes  the  cosine  law  as 
holding  exactly. 

The  Lighting 

The  total  luminons  flux  delivered  in  a  given  time,  that  is,  the 
time  integral  of  the  flux,  may  be  expressed  in  Inmen-seconds  or 
lumen-houTB,  according  to  circumstances.  If  this  is  called  the 
lighting,  and  is  represented  by  L,  we  have 

if  F  is  the  total  flux  in  lumens  and  T  is  the  time  in  seconds  or  in 
hours.  The  flash  of  a  fire-fly  may  be  expressed  in  lumen-seconds ; 
the  total  luminouB  radiation  per  gram  of  an  lUnminant,  or  the 
total  lighting  during  the  life  of  an  incandescent  lamp,  may  be 
expressed  in  lumen-hours. 

The  following  list  of  photometric  quantities  is  substantially  as 
recommended  by  the  committee  on  nomenclature  of  the  Illumi- 
nating Engineering  Society,  and  includes  the  quantities  employed 
in  the  preceding  discuseloo. 

Tablx  I 
Photometrlo  niasaltud«       Srmbol  Unit  Bqaatlon  of  deOnltloD 

J.  Intensltr  of  ll^t  I  Candle  ^=  ^ 


2.  Luminous  flax  F  Lumen  F^I«=^=BS  =  t4 

Lumena  or  f         i 

3.  Illumination  E  mllll-lnmoss  E  =  -g'-  =  -^ 

4.  Radiation  E'  B'  =  -g-  =  wb  =  mK 
6.  Brlghtn«w                       b             ^^  b=_i^,* 

6.  QuaaUty  Q  Candles  Q=bS 

7.  Ughtlng  L  Lnmen-bouTB  L^FT 

I,  b,  Q  are  expressed  In  candles.    F,  E  and  &'  are  expressed  In  lumeoa. 

B'  =^Tb  F  =  rQ 

F(   ^  Incident  flux 

Fa  =  emerssnt  flux 

m    ^coefficient  of  dltFuse  reflection  or  transmUalon 
(1  —  m)  =  coefficient  of  absorption. 

*  8,  Is  a  small  plane  area  visible  from  tbe  point  tor  whicb  the  In- 
tenslt?  I,  Is  taken. 

>y  Google 

398  Illouikatino  Enginbebino  , 

What  is  here  called  the  brightness  b  has  Bometimes  been  called 
the  specific  or  intrinsic  intensity,  and  designated  by  i.  But,  if  i 
is  the  Bpecific  intenaty,  the  integral  of  i  ought  to  be  the  total  in- 
tensity, and  that  is  not  true  except  for  very  small  plane  sources. 
For  spheres,  cylinders  or  extended  sources  of  any  shape,  it  is  not 
true,  and  the  term  specific  intensity  is  therefore  nnsatisfactoiy. 

On  the  other  hand,  the  brightness  b  is  defined  as  the  quantity 
of  light  per  unit  of  area,  and  the  integral  of  b  over  the  surface  of 
a  body,  whether  it  be  a  self-luminous  body  of  high  temperature  or 
a  diffusely  reflecting  body  of  low  temperature,  gives  the  total  quan- 
tity of  light,  Q,  which  multiplied  by  t  gives  the  total  luminous 
flux  from  the  body, 

II.  Primaby  and  Sbcondabt  Photometric  Standards 
The  fundamental  quantity  in  photometry  is  the  flux  of  light 
which  produces  illumination.  We  measure  the  flux  from  a  given 
source  by  comparing  it  with  that  from  a  standard  source.  From 
a  source  of  light  of  mean  spherical  intensity  I  candles,  a  total  fiux 
of  4tI  lumens  occurs.  A  standard  source  might  be  rated  in  terms 
of  its  total  luminous  flux  in  lumens,  but  owing  to  the  fact  that  it 
is  more  convenient  to  compare  accurately  the  angular  rate  of  flux 
in  a  particular  direction,  or  the  mean  horizontal  rate  of  flux  of 
two  given  sources,  than  to  compare  their  total  fluxes,  it  is  better 
to  rate  photometric  Btaodards  in  terms  of  their  intensity  in  a  par- 
ticular direction  in  candles,  or  the  mean  horizontal  intensity  in 
candles  than  in  terms  of  their  total  fluxes.  Bemembering  that  the 
intensity  in  a  particular  direction  is  proportional  to  the  luminous 
flux  in  that  direction,  and  is  equal  to  the  fiux  in  lumens  through 
a  small  solid  angle  a  dirided  by  u,  we  see  that  a  standard  source 
of  any  kind,  though  rated  in  candles,  is  really  a  standard  of  light 
fiux,  and  16  candles  in  a  particular  direction  means  a  flux  at  the 
rate  of  16  lumens  per  unit  of  solid  angle  in  that  direction. 

The  international  candle,  as  the  common  unit  of  intensity  of 
England,  France  and  America  is  generally  and  properly  called,  is 
a  unit  and  not  a  standard.  It  will  be  continued  by  international 
co-operative  effort,  through  frequent  comparisons  of  the  material 
standards  maintained  by  the  national  laboratories  of  these  coun- 
tries, but  the  particular  standards  that  are  employed  by  each  coun- 
try in  maintaining  tliis  unit  have  not  been  specified  and  need  not 
always  be  the  same.    The  comparisons  are  made  by  means  of  care- 

>y  Google 

,        Photometbio  Unitb  AMD  Standahds  399 

fully  prepared  carbon-filsmeat  lamps,  and  such  lamps  are  chiefly 
employed  in  maintaining  the  unit  constant.  But  flame  standards 
may  also  be  employed  if  they  are  found  to  be  sufficiently  reliable, 
and  they  can  in  any  case  be  employed  ae  checks  upon  the  work 
done  through  the  carbon-fllaoient  electric  lairips,  which  are  for 
the  present,  at  least,  more  reliable.  The  latter  are  commonly 
called  secondary  Btandarda,  although  in  reality  they  are  at  present 
employed  as^primary  standards. 

Tieo  Kinds  of  Primary  Standards 
The  primary  standards  employed  in  physical  meaeiirements  are 
of  two  kinds:  (1)  those  which  can  be  described  in  such  terms  that 
they  can  be  accurately  verified  or  reproduced  from  their  specifica- 
tions, and  (8)  those  which  are  more  or  less  arbitrary,  and  which 
cannot  be  accurately  reproduced  except  by  copying  other  standards 
of  the  same  kind.  The  international  ohm  is  a  standard  of  the 
first  kind,  as  it  is  specified  in  terms  of  the  resistance  of  a  definite 
column  of  mercury  at  a  certain  temperature,  and  it  can  be  repro- 
duced without  reference  to  any  other  standard  of  resistance.  The 
meter  waa  originally  intended  to  be  auch  a  standard,  being  defined 
in  terms  of  the  dimensions  of  the  earth.  But  when  it  was  found 
that  the  dimensions  of  the  earth  were  different  from  what  had 
been  supposed,  and  that  the  meter  would  require  a  new  definition, 
the  reference  to  the  earth  was  abandoned  and  the  meter  became  a 
standard  of  the  second  kind,  only  to  he  reproduced  by  reference  to 
other  meter  bars,  copies  of  itself,  of  which  there  were  a  sufficient 
number  in  existence  to  make  it  possible  to  maintain  the  meter  in- 
definitely in  this  way.  More  recently  the  meter  has  been  expressed 
in  terms  of  the  wave  length  of  light  so  exactly  that  it  could  be 
reproduced  accurately  if  all  length  standards  were  lost.  Hence, 
the  meter  has  again  become  a  primary  standard  of  the  first  kind. 
However,  meter  bars  are  so  permanent  that  in  practice  they  are 
verified  and  reproduced  by  comparing  with  one  another,  without 
reference  to  the  absolute  specification  in  terms  of  the  wave  length 
of  light. 

The  kilogram  was  intended  to  be  a  natnral  unit,  so  defined  in 
tenns  of  the  unit  of  length  and  the  density  of  water  as  to  be  a 
standard  of  the  first  kind.  But,  owing  to  the  difficulty  of  deriving 
it  in  this  way,  it  is  more  accurate,  as  well  as  more  convenient,  to, 
r^ard  it  ae  a  standard  of  the  second  kind,  and  to  verify  and  re- 

>y  Google 

400  Illuuihatino  Ekoineebikq  , 

produce  standardB  of  mass  by  reference  to  veil-made  platinum 
GtandardB  without  attempting  to  derive  it  according  to  its  original 

Thermometers  are  Btandards  of  the  first  kind,  inasmuch  as  they 
are  referred  to  the  natural  interval  between  the  freezing  and  boil- 
ing pointe  of  water  under  standard  conditionB,  and  they  can  there- 
fore be  verified  or  reproduced  by  referring  to  the  formal  speci- 

The  unit  quantity  of  electricity,  the  international  coulomb,  is 
defined  in  terms  of  the  quantity  of  silver  it  will  deposit  under 
standard  conditions  when  passed  through  a  solution  of  nitrate  of 
silver.    It  is,  therefore,  a  primary  standard  of  the  first  kind. 

Primary  Pkototnetric  Standards  of  the  First  Kind 
Primary  photometric  standards  may  be  of  the  first  kind  or  of 
the  second  kind.  Although  primary  standards  of  the  first  kind 
are  to  be  preferred,  other  things  being  equal,  obviously  a  reliable 
and  convenient  and  permanent  standard  of  the  second  kind  is  better 
than  an  unreliable,  inconvenient  and  temporary  standard  of  the 
first  kind.  Many  primary  photometric  standards  of  the  first  kind 
have  been  proposed,  and  a  considerable  number  have  been  used. 
The  sperm  candle  is  made  to  carefully  stated  specifications,  and  has 
been  more  widely  used  than  any  other  photometric  standard.  But 
it  is  a  very  crude  standard.  The  Careel  lamp  in  France,  the  Har- 
court  pentane  lamp  in  England,  and  the  Hefner  lamp  in  Germany 
are  accepted  as  primary  photometric  standards  of  the  first  kind  in 
the  respective  countries.  They  are  made  and  used  according  to 
very  elaborate  specifications,  but  as  the  light  is  the  result  of  the 
specified  fuel  burning  in  a  specified  lamp,  surrounded  by  a  speci- 
fied atmosphere,  the  standard  is  not  merely  the  lamp,  but  the  com- 
bination of  lamp,  fuel  and  atanoephere,  the  two  latter  of  which  are 
constantly  changing.  For  use  in  ordinary  gas  photometry  fiame 
standards  are  convenient.  But  for  precision  photometry,  in  gen- 
eral, or  for  detennining  and  maintaining  a  photometric  unit,  it  is 
not  unfair  to  say  that  the  best  of  fiame  standards  is  not  as  con- 
venient or  reliable  as  primary  standards  ought  to  be. 

The  difficulties  in  the  use  of  fiame  standards  are,  therefore, 
partly  in  the  lamp,  which  is  the  more  or  less  permanent  part  of  the 
-combination;  partly  in  the  fuel,  which  is  often  founfl  not  to  con- 
form to  the  specifications,  and  in  some  cases  is  liable  to  change  on 

>y  Google 

Photometric  Units  and  Standabdb  401 

standing  even  if  it  confoims  originally  to  specifications,  and  partly 
to  the  atmosphere,  which  ia  constantly  changing  with  respect  to 
barometric  preesare  and  aqueous  vapor,  while  in  and  about  the  lamp 
it  changes  also  with  respect  to  carbon  dioxide  and  oxygen  content. 
All  these  variations  affect  the  light  as  a  Same  standard,  and  make 
the  errors  of  measurement  many  times  greater  than  those  made 
OQ  carbon-filament  lamps.  B;  making  a  long  series  of  measure- 
ments, the  accidental  errors  are  largely  eliminated,  and  a  mean 
result  may  be  obtained  which  is  surprisingly  good  in  view  of  all 
the  difficulties.  But  there  are  constant  sources  of  error  that  are 
not  so  eliminated,  and  perhaps  the  most  perplexing  are  due  to 
the  lamp  itself.  For  example,  although  Hefner  lamps  are  made  by 
different  makers  very  carefully  from  the  same  specifications,  there 
is  a  range  of  2  per  cent  between  the  highest  and  lowest  values  of 
eight  Hefner  lamps  belonging  to  the  Bureau  of  Standards,  four 
from  one  German  maker  and  four  from  another.  There  are  two 
different  devices  in  use  for  observing  the  height  of  the  fiame,  but 
all  (or  nearly  all)  the  lamps  conform  to  the  apeeificatione.  If  one 
requires  only  that  his  Hefner  lamp  be  correct  within  2  per  cent, 
all  these  lamps  are  satisfactory.  But  as  primary  standards  they 
ought  not  to  differ  eo  much  independently  of  fuel  and  atmosphere. 
In  the  same  way,  standard  Harcourt  pentane  lamps  differ  several 
per  cent  in  candle-power,  using  the  same  fuel  and  operating  them 
under  the  most  favorable  conditions.  At  the  Bureau  of  Standards 
we  have  tested  pentane  lamps  from  two  English  makers  and  one 
American  maker.  The  two  Chanoe  lamps  tested  have  the  highest 
candle-power,  averaging  about  9.9  international  candles  under 
standard  atmospheric  conditions,  namely,  8  litera  of  water  vapor 
per  cubic  meter  of  air,  and  standard  barometric  pressure.  The 
Sugg  lamps  tested  average  less  than  9.7  candles,  about  2.5  per 
cent  less  than  Chance  lamps.  American-made  pentane  lamps  also 
average  about  9.7  candles. 

The  standard  Harcourt  pentane  lamp  was  supposed  originally  to 
give  10  British  parliamentary  candles,  and  there  was  supposed  to 
be  no  appreciable  variation  among  different  lamps.  The  National 
Physical  Laboratory  adopted  a  particular  lamp  of  this  kind  as  its 
primary  standard.  When  the  international  candle  was  fixed  by 
agreement  between  the  national  laboratories  of  England,  France 
and  America  the  Bureau  of  Standards  made  a  change  of  1.6  per 
cent  in  its  photometric  unit,  in  order  to  come  into  agreement  with 

>y  Google 

402  Ii  LUMiNiTiNo  Engineering 

England  and  France,  and  at  the  same  time  to  bring  the  gas  and 
electric  industries  of  America  to  a  common  standard  by  bringing 
the  new  unit  midway  between  the  old  nnit  of  the  Bureau,  which 
was  used  by  the  electrical  industries,  and  the  average  value  of  the 
unit  used  in  the  gas  industries.  Theoretically,  therefore,  the 
standard  pentane  lamp  should  give  10  international  candles.  .Bat 
it  happens  that  the  particular  standard  pentane  lamp  of  the  !Na- 
tional  Physical  Laboratory  apparently  has  a  slightly  higher  value 
than  the  average,  and  the  English  maker  of  the  lamp  has  been 
unable  to  furnish  ua  a  lamp  giving  the  same  candle-power.  The 
Bureau  placed  an  order  for  a  lamp  to  agree  with  the  standard  of 
the  National  Physical  Laboratory,  as  shown  by  direct  compari- 
sons made  at  the  National  Physical  Laboratory.  After  several  at- 
tempts on  the  part  of  the  maker  a  lamp  was  accepted  having  about 
1  per  cent  lower  value.  The  Bureau  has  never  tested  a  pentane 
lamp  of  any  make  having  a  value  as  high  as  the  National  Physical 
Laboratory  standard.  The  values  found  range  from  1  to  5  per 
cent  less.  Hence,  it  is  evident  that  the  pentane  lamp  as  a  primary 
standard  cannot  be  a  complete  success  until  different  makers  fol- 
lowing the  same  specifications  can  produce  lamps  agreeing  better 
in  value,  and  until  the  lamps  produced  by  any  experienced  maker 
agree  better  among  themselves  than  they  now  do.  At  the  Bureau 
of  Standards  we  have  made  some .  progress  in  locating  the  source 
of  the  differences,  and  hope  soon  to  see  a  great  improvement  in 
this  respect. 

The  second  source  of  trouble  with  pentane  lamps  is  the  fuel. 
Pentane  (CjHu)  is  a  very  volatile  hydrocarbon,  distilled  from 
gasoline.  It  is  classed  as  explosive,  and  should  be  shipped  in 
strong  sealed  cans  and  stored  and  handled  with  special  precau- 
tions. It  coats  the  Bureau  of  Standards  $3.50  per  gallon,  and  is 
consumed  in  considerable  quantities.  It  is  distilled  between  S5° 
and  40°  C,  and  in  summer  in  an  open  can  evaporates  rapidly  at 
laboratory  temperatures.  The  flame  is  ordinarily  fed  by  the  mix- 
ture of  air  and  pentane  which  come  over  from  the  saturator.  But 
in  hot  weather  instead  of  air  entering  the  inlet,  pentane  vaporizes 
so  rapidly  that  it  flows  out  through  both  outlet  and  inlet,  the 
vapor  escaping  through  the  air  inlet,  passing  out  into  the  atmos- 
phere of  the  laboratory,  and  so  causing  the  pentane  to  disappear 
at,  perhaps,  double  the  normal  rate.  Hence,  pentane  lamps,  as 
ordinarily  constructed,  cannot  be  used  satisfactorily  in  summer  in 

>y  Google 

Phoxometsic  XJnitb  akd  Standabss  ma 

soatfaern  latitades.  Slight  modifications  in  the  lamp  cao  be  made 
to  ovetcome  this  difficulty. 

Moreover,  ta  pentane  is  oot  a  simple  compound,  but  contains 
homologous  compounds  which  are  not  completely  separated  even 
by  repeated  distillatiojiB,  the  density  changes  as  evaporation  pro- 
ceeds, and  hence  the  reservoir  must  be  emptied  and  refilled  with 
freeh  pentane  from  time  to  time,  in  order  to  keep  the  fuel  within 
the  specifications  and  the  light  of  the  Same  sufficiently  near  to  its 
normal  valne. 

The  light  of  a  pentane  flame,  like  other  gas  flames,  is  very  sensi- 
tive to  impurities  in  the  atmosphere  and  to  drafts  or  air  currents. 
There  most  be  excellent  ventilation  of  the  room  and  plenty  of 
pure  air  supplied  to  the  flame,  but  not  too  much.  The  removal 
of  the  products  of  combustion  and  the  screening  of  the  lamp  from 
air  currents,  as  well  as  the  regulation  of  the  supply  of  pentane  and 
the  detailed  manipulation  of  the  lamp,  all  call  for  experience,  pa- 
tience and  skill  in  high  degree,  in  order  to  get  condstent  and  re- 
liable results  from  a  pentane  standard. 

Of  course,  the  atmospheric  humidity  must  be  carefully  deter- 
mined every  time  a  set  of  measurements  is  made,  and  the  barom- 
eter must  be  read  in  order  that  humidity  and  pressure  corrections 
may  be  made.  These  corrections  are  considerable,  the  humidity 
correction,  which  is  the  larger  of  the  two,  sometimes  amounting 
to  10  per  cent. 

These  remarks  apply  only  to  pentane  lamps  which  are  used  for 
the  purpose  of  relatively  accurate  measurements.  As  working- 
flame  standards  they  may  be  used  with  fewer  precautions,  if  ap- 
proximate results  are  sufficient. 

When  a  flame  standard  is  employed  for  testing  illaminating  gas, 
the  humidity  and  barometric  corrections  are  not  applied,  as  the  gas 
Same  is  affected  practically  by  the  same  amount,  and  the  test  is  in- 
tended to  demonstrate  the  quality  of  the  gas  and  not  the  amount  of 
light  given  by  the  given  test  burner  at  that  particular  time.  In 
other  words,  SO-candle-power  gas  is  not  gas  that  always  gives  20 
candle-power  in  a  particular  burner  when  consumed  at  a  stated 
rate,  but  gas  of  standard  light-giving  properties,  that  is  to  say, 
it  gives  20  candle-power  when  burned  at  a  given  rate  in  a  particular 
burner  in  a  standard  atmosphere,  which  is  a  pure  atmosphere  con- 
taining 6  liters  of  water  vapor  per  cubic  meter  and  at  normal 
barometric  pressure.     In  winter,  when  the  humidity  averages  less 

>y  Google 

404  Illohinatihq  Enqineesinq 

than  normal,  the  light  will  be  greater  than  the  average.  In  sum- 
mer, vben  the  htimidity  averageB  greater  than  normal,  the  light 
will  be  less  than  the  average,  and  may  be  10  per  cent  less.  Thua, 
flame  standards  are  foi  this  reason  well  adapted  to  serve  as  work- 
ing standards  for  testing  the  light-giving  properties  of  gasand  oil. 
But  for  primary  standards,  intended  to  maintain  a  photometric 
unit,  they  are  not  as  well  adapted  as  they  would  he  if  unaifected 
by  the  atmosphere. 

Hefner  lamps  have  some  important  advantages  over  pentanes, 
and  some  marked  disadvantages.  Whereas  the  standard  Harcourt 
pentane  lamp  is  bulky,  complicated  in  construction,  relatively  la- 
borious to  manipulate,  and  expensive  both  in  first  cost  and  in  fuel, 
the  Hefner  amylacetate  lamp  is  small  and  very  portable,  simple  in 
construction,  easy  to  assemble  and  make  ready  for  use,  and  lees 
expensive  in  first  cost  and  in  fuel.  The  latter  costs  the  Bureau 
$3.00  per  pound,  but  so  much  less  is  employed  that  it  costs  leas 
per  hour  than  pentane  at  $3.50  per  gallon. 

Its  disadvantages  in  comparison  with  the  pentane  standard  are 
(1)  its  small  candle-power,  (2)  the  redder  color  of  its  flame,  (3) 
its  more  unsteady  flame,  and  (4)  the  greater  difficulty  of  main- 
taining the  correct  flame  height. 

The  Hefner  flame  has  a  horizontal  intensity  of  0.9  candles  when 
the  flame  is  40  mm.  high,  as  officially  prescribed  in  Qermany.  We 
find  at  the  Bureau  that  the  flame  bums  about  aa  steadily  and  is 
nearly  as  easy  to  manipulate  when  maintained  at  45  mm.,  at  which 
height  it  gives  1  international  candle,  or  0.1  candle  more  than  at 
40  mm.  This  change  in  a  standard  lamp  is  made  by  placing  a 
ring  5  mm.  thick  under  the  support  of  the  sight  which  is  used  to 
regulate  the  flame  height.  To  obtain  a  suitable  illumination  on  the 
test  screen  of  the  photometer,  a  standard  of  1  candle-power  must 
be  placed  quite  near,  and  errors  due  to  slight  variations  in  distance 
are  much  greater  than  for  a  lO-candle-power  standard.  In  prac- 
tice, both  a  shorter  distance  and  a  weaker  illumination  are  em- 
ployed with  the  Hefner  standard. 

The  color  difference  between  standards,  or  between  a  standard 
and  a  light  source,  is  necessarily  a  source  of  uncertainty,  and  with 
modern  electric  and  gas  lamps  the  demand  is  for  whiter  standards. 
The  Hefner  is  the  reddest  standard  in  use,  and  its  color  is  one  of 
its  most  serious  objections.  However,  color  screens  are  necessary 
to  pass  from  one  color  to  another,  and  the  difference  between  the 

>y  Google 

Photohbtbio  Units  aitd  Standabds  405 

pentane  color  and  the  Hefner  color  is  not  eBoagh  to  make  this  a 
decidii^  consideration,  as  between  the  two  lamps.  The  voltage 
on  a  carbon-filament  lamp  necessary  to  give  a  color  match  witii 
aereral  different  flame  standards  U  as  follows: 

To  give  4  watte  per  candle   =110  voltB, 

To  match  the  kerosene  lamp  =102-108  volta. 

To  match  the  Carcel  lamp      =98  volts. 

To  match  the  pentane  lamp  =91  volts. 

To  match  the  Hefner  lamp  =8G  volts. 
The  flame  of  a  Hefner  lamp  is  very  easily  disturbed  by  air  car- 
rents,  and  the  tip  is  in  almost  constant  motion  vertically  and 
laterally,  so  that  the  flame  must  be  screened  very  carefully,  and 
then  mnst  be  watched  constantly  by  an  assistant,  and  readings  made 
only  when  it  is  at  the  right  height  and  in  correct  position.  The 
tip  is  only  slightly  luminous,  and  yet  the  height  must  be  main- 
tained coDBtant  to  a  fraction  of  a  millimeter.  Different  observers 
may  differ  sensibly  in  their  judgment  as  to  when  it  is  right,  al- 
though this  source  of  error  is  smaller  than  would  be  supposed. 

The  amylacetate  is  so  volatile  that  the  top  of  the  wick  is  below 
the  top  of  the  wick  tube.  As  the  room  temperature  rises,  the 
wick  must  be  lowered  to  keep  the  height  of  flame  constant,  and 
this  makes  the  flame  more  unsteady.  At  summer  temperature, 
such  as  35°  to  30"  C,  the  flame  is  much  more  unsteady  than  at 
15°  to  30°  C.  In  this  respect  (less  satisfactory  operation  in  hot 
vreather)  both  the  pentane  and  amylacetate  lamps  and  candles  are 
inferior  to  kerosene-oil  lamps. 

Because  the  Carcel  lamp  is  bo  little  used  in  this  country,  or  any- 
vrhere  outside  of  France,  and  because  our  limited  experience  at 
the  Bureau  has  shown  it  to  be  unsatisfactory,  nothing  will  be  said 
of  it  as  a  standard. 

Candles  have  been  of  enormous  service  in  practical  gas  photom- 
etry, but  they  cannot  be  seriously  considered  at  the  present  day 
as  standards.  Kerosene-oil  lamps  are  much  more  convenient  and 
reliable,  and  we  hope  iu  the  near  future  to  publish  experiments 
made  at  the  Bureau  showing  that  as  secondary  standards  for  prac- 
tical photometry  they  may  be  used  with  excellent  results. 

To  sum  up  in  a  few  words,  it  may  be  said  that  as  primary  photo- 
metric standards  of  the  first  kind,  there  are  the  pentane  and  Hefner 
lamps  about  equally  entitled  to  consideration,  each  possessed  of 
important  merits,  but  also  of  serious  limitations  and  defects.     A 

>y  Google 

406  Illuhinaxiko  Ehqinbebino 

given,  pentane  lamp  is  probably  more  consistent  with  itself  than  an 
average  Hefner,  but  different  pentane  lamps  differ  more  than  Hef- 
ner's do.  No  other  Btandard  of  the  first  kind  equals  them  in  con- 
stancy and  reproducibility,  and.  no  other  is  need  where  accurate 
results  are   attempted. 

The  radiation  from  incandescent  platinum  at  it£  melting  point 
was  long  ago  proposed  by  Violle  as  a  primary  photometric  unit  of 
the  first  kind.  But,  although  enormous  progress  has  been  made  in 
obtaining  and  maintaining  and  measuring  high  temperatures,  and 
several  serious  attempts  have  been  made  to  make  Yiolle'B  proposal 
practicable,  nobody  has  ever  succeeded  in  doing  as  well  with  it 
as  can  be  done  with  flame  standards.  Drs.  Waidner  and  Burgess, 
of  the  Bureau  of  Standards,  have  made  an  interesting  proposal, 
namely,  to  employ  the  radiation  from  a  black  body  at  a  particular 
temperature,  tor  example,  at  the  melting  point  of  platinum,  but 
they  have  not  as  yet  attempted  to  realize  it  in  practice.  Dr.  Stein- 
metz  has  recently  also  made  a  new  proposal  for  a  primary  photo- 
metric standard  of  the  first  kind,  but  the  realization  of  this  pro- 
posal to  the  extent  of  obtaining  a  standard  of  precision  seems  very 
diflicult,  and  so  far  as  I  know  has  not  been  attempt«d. 

Photometric  Standards  of  the  Second  Kind 
The  most  successful  photometric  standards  of  the  second  kind 
are  carbon- filament  incandescent  lamps,  which  have  been  employed 
for  many  years  as  convenient  working  standards,  and  in  recent 
years  have  been  employed  in  making  careful  comparisons  of  the 
photometric  standards  of  different  countries.  Their  use  is  so  im- 
portant and  their  operation  under  the  best  conditions  la  so  admir- 
able that  I  wish  to  present  briefiy  the  method  of  their  preparation 
and  use  and  records  of  their  performance.  Such  lamps  cannot, 
of  course,  be  made  accurately  to  specifications,  but  if  they  are 
sufficiently  permanent  they  may  be  employed  to  maintain  the  unit 
of  light  for  an  indefioit*  period.  Probably  nothing  is  more  per- 
manent than  pure  carbon,  sealed  in  a  vacuum  and  kept  at  ordinary 
(room)  temperatures.  Hence,  if  carbon-filament  lamps  can  be 
prepared  which  will  not  change  appreciably  when  burned,  say  100 
hours,  under  working  conditions,  there  is  reason  to  believe  that 
they  will  remain  constant  for  a  long  period  of  years  (barring  acci- 
dents) and  that  a  group  of  such  lamps  will  afford  a  means  of 
mnintaining  the  unit  of  light  constant  for  a  long  time.    How  long 

>y  Google 

Photombtrio  Units  and  Standabis  40? 

and  how  accurately  caB,  of  cotiree,  only  be  detennined  by  ex- 

CarboD-filament  incandescent  lamps  are  usually  operated  as 
standards  at  a  constant  voltage,  the  current  being  measured  as  a 
check.  Sometimee  they  have  been  measured  at  constant  current, 
the  voltage  being  varied  slightly,  if  necessary.  If  the  lamps  have 
constant  resistance,  of  course  these  two  methods  would  amount  to 
the  same  thing.  But,  as  carbon-filament  lamps  do  not  have  con- 
stant resistance,  but  generally  show  a  decreasing  resistance,  fol- 
lowed after  a  longer  or  shorter  period  by  an  increasing  resistance, 
it  becomes  a  matter  of  prime  importance  whether  the  best  per- 
formance can  be  secured  by  operating  lamps  regularly  during  their 
useful  life  as  standards  at  constant  voltage,  or  at  constant  current, 
or  whether  still  better  results  can  be  obtained  by  operating  them 
at  constant  watte.  Obviously,  if  the  radiation  from  the  surface 
of  the  filament  is  unchanged,  and  the  bulb  does  not  blacken  or 
change  its  absorption,  the  most  constant  candle-power  will  be  se- 
cured by  operating  the  lamps  at  constant  watts;  a  constant  rate  of 
energy  supply  and  a  constant  conversion  factor  giving  a  constant 
Aux  of  light.  But,  whether  the  radiation  from  the  filament  and 
the  absorption  in  the  bulb  will  be  constant  at  constant  watts  could 
only  be  detennined  by  experiment. 

At  the  Bureau  of  Standards  we  have  investigated  this  question 
very  carefully,  and  to  obtain  the  highest  possible  precision  have 
made  use  of  a  double-precision  photometer,  with  special  recording 
cylinders,  having  two  observers  measure  the  same  lamp  simul- 
taneously, and  a  third  observer  measuring  the  current  and  voltage 
of  the  lamp  at  once  by  means  of  two  standard  potentiometers.  A 
single  determination  consists  of  the  mean  of  a  large  number  of 
readings,  each  recorded  without  the  observer  taking  his  eye  away 
from  the  photometer,  and  as  the  observer  does  not  know  any  of  his 
readings  until  they  are  ell  completed,  he  reads  without  prejudice. 
By  this  means  each  observer  is  a  check  upon  the  other,  twice  as 
many  determinations  can  be  made  in  a  given  time  as  by  a  single 
photometer,  and  by  the  use  of  Dr.  Middlekauif's  direct- reading  scale 
all  calculations  of  candle-power  are  eliminated,  the  value  of  each 
determination  in  terms  of  the  mean  of  the  group  of  standards  em- 
ployed being  read  off  directly  from  the  record  sheet. 

In  this  way  it  has  been  found  that  with  lamps  in  which  no 
blackening  occurs  the  best  results  are  obtained  by  keeping  the  watts 

>y  Google 


constant  instead  of  using  them  at  constant  voltage.  Life  curves 
have  been  made  of  a  large  number  of  standards,  and  each  curve 
divided  into  the  period  of  seasoning  and  the  period  of  useful  life  aa 
precision  standards.  If  they  are  burned  at  conHtant  volts,  the  season- 
ing ie  carried  on  until  they  reach  constant  resistance.  This  is  a 
longer  or  shorter  operation,  depending  on  the  temperatare  (or 
watts  per  candle)  at  which  they  are  seasoned,  but  is  not  the  same 
for  different  lamps.  If,  however,  they  are  to  be  burned  at  con- 
stant watte,  it  is  not  necessary  that  the  seasoning  be  continued  to 
minimum  resistance;  when  the  filaments  have  nearly  reached  that 
condition  they  may  be  used  with  perfect  satisfaction,  and  long  after 
the  resistance  has  reached  its  minimum  and  has  increased  appre- 
ciably the  lamp  is  still  a  reliable  standard,  provided  only  that  the 
watts  have  been  kept  constant,  and,  of  course,  provided  that  black- 
ening has  not  occurred. 

Blackening  can  be  detected  by  the  decrease  in  the  light  of  the 
lamp  before  it  can  be  seen  on  the  glass.  To  reduce  it  to  a  minimum 
the  lamps  should  be  made  and  selected  with  great  care,  and  the 
filaments  should  preferably  be  mounted  in  larger  bulbs  than  is 
ordinarily  done.  Dr.  Fleming,  of  London,  many  years  ago  advo- 
cated the  use  of  large  bulbs  for  incandescent-lamp  standards,  bat 
as  the  quality  of  lamps  improved  it  did  not  seem  necessary  to  use 
them,  and  hence  nearly  all  laboratories  used  the  ordinary-sized 
16  candle-power  lamps  for  standards  of  the  best  quality.  We  have 
found  in  our  recent  work  at  the  Bureau  of  Standards,  however, 
that  lamps  in  larger  bulbs  give  better  results. 

We  have  seasoned  and  carefully  measured  nearly  200  standards 
as  above  described,  and  selected  the  beet  for  primary  standards. 
A  few  of  these  have  been  burned  for  200  hours  after  seasoniag 
without  the  candle-power  changing  more  than  a  few  hundredths 
of  a  candle.  Such  lamps  would  serve  as  reference  standards  in  a 
photometric  laboratory  for  many  years,  perhaps  for  a  century,  with- 
out being  burned  as  many  hours  as  they  have  been  burned  in  these 
special  tests.  There  should  be  no  depreciation  while  they  are  not 
burning,  for  what  is  more  permanent  than  pure  carbon,  preserved 
in  a  vacuum  at  ordinary  temperatures? 

As  to  the  precision  of  measurement  of  carbon-filament  electric 
lamps,  on  such  a  precision  photometer  as  described  abov^  the  mean 
error  of  the  determination  of  candle-power  on  any  lamp  at  one 
time  is  about  0.2  per  cent,  whereas  the  mean  error  of  the  average 

>y  Google 

Fhotouetrio  Ukitb  and  Staksabds  40d 

value  of  Bix  lamps  measiired  at  one  time  is  about  0.1  per  cent.  If 
a  group  of  six  lamps  be  measured  by  four  different  experieueed  ob- 
Beirers  (as  is  done  at  tlie  Bureau  in  work  of  the  highest  precision) 
the  mean  of  the  four  will  be  still  less  in  error.  These  figures  are 
the  results  of  a  large  number  of  experiments  with  rotating  stand- 
ards, of  the  same  color,  and  stationary  standards  may  be  measured 
with  substantially  the  same  accuracy. 

With  such  precision  of  measurement  and  a  life  performance  of 
standards  such  as  described  above,  it  would  seem  as  though  the 
unit  of  candle-power  not  only  of  a  commercial  laboratory,  but  also 
of  a  national  standardizing  laboratory,  or  even  of  a  group  of  na- 
tional standardizing  laboratories,  could  be  maintained  for  a  long 
period  of  years  by  carbon-fiiament  incandescent  lamps  more  con- 
stant than  has  been  possible  heretofore  with  flame  standards  or 
any  other  form  of  primary  standard  as  yet  realized. 

However,  there  are  possibilities  of  improvement  in  fiame  stand- 
ards, and,  of  course,  possibilities  of  some  new  primary  standard 
appearing  which  shall  surpass  any  flame  standard  as  yet  proposed. 
What  I  wish  to  emphasize  is,  not  by  any  means  that  incandescent 
lamps  are  the  final  standards  or  that  they  are  satisfactory  as  pri- 
mary standards,  but  that  they  really  are,  as  now  used,  primary 
standards,  and  that  by  their  use  a  photometric  unit  can  be  main- 
tained so  well  that  until  the  difiicultdes  of  heterochrome  photome- 
try are  overcome,  and  until  the  demands  for  precision  in  prac- 
tical photometry  are  considerably  increased,  we  need  not  fear  that 
the  international  candle  will  drift  far  enough  from  its  present 
value  to  be  serious.  The  progress  that  has  been  made  in  photo- 
metrical  measurements  in  the  14  years  since  the  Geneva  Congress 
is  gratifying.  Then  it  was  believed  that  the  Hefner  unit  and  the 
bougie  decimale  were  practically  equivalent.  The  uncertainty  in 
the  relative  values  of  the  standards  of  different  countries  amounted 
to  several  per  cent.  Now  the  corresponding  uncertainty  is  not  or 
need  not  be  more  than  a  few  tenths  of  1  per  cent,  so  long  as 
standards  of  a  single  color  are  employed.  It  remains  to  accomplish 
as  much  for  standards  of  a  whiter  color,  and  to  fix  the  ratios 
in  passing  from  one  color  to  another. 

>y  Google 

410  Illuminating  Gnoinsbring 

ebfbrences  on  units  and  nomenclature 

Blondel,  A.,  Lumlgre  Elect.  53,  pp.  7-15,  1894. 

Bltmdel.  A.,  L'Eclalr.  Elect.  8.  pp.  341-36G,  1S9S. 

Broca,  A.,  L'Bclair.  Elect  6.  pp.  148-167,  1896. 

Herner-Alteneck,  F.  v.,  Elektrotech.  Zeltachrltt,  17,  pp.  7S4-6.  1896. 

Kapp,  Q.,  Elektrotech.  Zeltachrlft,  17,  pp.  631-4,  ISUS. 

Weber,  L.,  Elektrotech.  Zeltachrlft,  18.  pp.  91-94.  1897. 

V  rband   Deutscber  Blektrotechnlker,    Elektrotech.    ZelUchrlft,   18,    p. 

474,  1897. 
MlUar,  P.  8.,  Elect.  Rev.  (New  York).  51,  pp.  426-S,  1907. 
Herlng,  C.  Trans.  111.  Eng.  Soc.  3,  pp.  646-678,  1908. 
Roea,  E.  B.,  Bulletin  Bureau  ol  Standards,  6,  pp.  643-672,  1910. 


Hefner-Alteneck,  F.  v.,  Vorschlag  Zur  Beflchaffung  elner  konstanten 
Uchteinhelt,  ElektrotechnUche  Zeltschrift,  5.  pp.  20-24,  lgS4. 

Phyi.-Tech.  Relchaangtalt,  Die  Beglaubigung  der  Hetnerlampe,  Zeltachrlft 
f.  iDBtrumentenkunde,  13,  pp.  367-267,  18S3. 

Liebenthal,  E.,  Ueber  die  Abblltiglgkett  der  Hefnerl&mpe  und  der  Fentan- 
lampe  von  der  Beachaff^nhelt  der  umgabenden  Luft.  Zeitscfarlfl  t 
InBtrumenlenkunde,  15.  pp.  157-171,  1895. 

Vernon  Harcourt,  A.  G.,  On  a  lO'Candle  Lamp  to  be  used  as  a  Standard 
of  Light,  BrltlBh  Assoc.  Report  1898,  pp.  845-346. 

Fleming,  J.  A.,  The  Photometry  of  Electric  Lampe,  Jour.  Inst,  of  Elect 
Ens-  32.  pp.  119-216,  1902-3. 

Paterson,  C.  C,  Some  Investigations  on  the  10  c.  p.  Harcourt  Pentane 
Lamp,  Electrician  (London),  53,  pp.  751-752,  1904. 

PatersOD,  C.  C,  InvestigatlonB  on  Light  Standards,  etc.  Jour.  Inet  of 
Elect.  Eng.  38.  pp.  271-308,  1906-7:  National  Phys.  Lab.,  Coll.  Re- 
searches, S,  pp.  49-66,  1908. 

Dow.  J.  S.,  The  Sources  of  Error  In  the  Harcourt  10  c.  p.  Pentane  Stand- 
ard, Elect  Rev.   (London),  59,  pp.  491-3,  1906. 

Olazebrook,  R.  T.,  The  Photometric  Standard  of  the  National  Physical 
laboratory,  British  Assoc  Report,  1908,  p.  623;  Electrician  (Lon- 
don). 61.  pp.  922-3,  1908. 

Report  of  Committee  on  Taking  Candle-power  of  Oas.  Proc.  Amer.  Gas 
Institute,  a,  pp.  454-509.  1907. 

Bond,  0.  O..  Working  Standards  of  Light  and  Their  Use  In  the  Pho- 
tometry of  Qaa,  Jour.  Franklin  Inst.  165,  pp.  189-209,  1908. 

Roaa  £  Crittenden,  Report  of  Progress  on  Flame  Standards,  Trans.  IlL 
Eng.  Soc.  5,  pp.  763-778,  1910. 

>y  Google 



By  Clayton  H.  Sharp 



Definition  and  scope. 

Quantltleo  to  be  meaaured. 

Are  rel&tlve  to  a  standard. 

Made  by  zero  method,  using  tbe  eye  as  Instrument  for  determining 

DtOculty  due  to  color  difference. 
Apparatus,  general. 

1.  Sfght-box,  photometer  bead,  or,  for  abort,  photometer,  for  pro- 

ducing contiguous  Illuminated  fields. 

2.  Apparatus  vbereby  Intensity  of  one  or  botb  fields  may  be  varied 

according  to  known  law. 

3.  Standard  source  of  Ilgbt. 
Varying  tbe  intensity. 


Effect  of  area  of  sources. 

Apparent  candle-power. 
Sector  disc.     Talbot's  I<aw,  Napoll,  Brodbun,  Hyde. 
Diaphragm.     Comu'a  cat's-eye.     Lens.     Diffusing  plate. 
Inclined  plate. 

Varying  source — belgbt  of  flame,  voltage. 

Fields  muet  be  contiguous  or  adjacent. 

Equality  principle.    Contrast  principle. 

Lambert  or  Rumford. 



BlBter-JoIr  block. 


Oreas»apot    Disappearance.    Contrast—Rildorff  mlrron.    Le» 
son  built-up  dlec.     Theory.     Construction.     Errors  in   use. 
Limitations.    Accurscy. 
Lummer-Brodhun.  * 

Plain.    Contrast.    Sensibility. 

>y  Google 

412  Illdminating  Enqineerino 

PractlcAl  appantuB. 

Precision  bar  photometer.     Scales,  equal  part,  proportional,  direct 

Industrial:  gas.  electric. 

Pori»ble  and  Illumination  photometere. 
Weber.  Martens,  Blonde].  Stiarp-Mlllar. 
Auxiliary  apparatus. 

Lamp  rotators. 

Dletrlbutlon,  elevating  lamp,  three  mirror. 

Arc-lamp  apparatus.    Long  arm. 
Integrating  and  summation  apparatus. 

Blondel.  Matthews,  sphere. 
Heterochrome  photometry. 

Equality  of  contrasts — Leeeon  disc. 

Vlenai  acuity. 

Flicker  photometer. 

Rood.  Slmmance-Abady,  Whitman,  Schmidt  ft  Haensch. 

8  pectro-photo  m  eters. 

Vierordt,  Lummer-Brodbun,  Nichols,  Brace. 

Three-color  apparatus.    Ives  colorimeter, 

Lectcrb  I 

Definition  and  Scope 
Photometry  is  broadly  defined  as  the  science  of  tlie  measurement 
of  light.  Ordinarily  the  name  has  been  used  to  refer  to  the  meas- 
urement of  the  intensity  of  sources  of  light,  since  this  has  been 
the -measurement  most  commonly  made.  The  measurement  of  il- 
lumination as  distinct  from  intensity  of  a  source  has  come  into 
mucli  greater  prominence  in  recent  years,  and  the  term  "illurai- 
nometry  "  has  been  used  for  this  cla^s  of  measurements.  Essen- 
tially, there  is  no  difference  between  illuminometry  and  photometry, 
all  photometric  measurementB  being  essentially  measurements  of 
illumination  or  brightness;  hence,  we  may  say  that  the  term  il- 
luminometry includes  the  term  photometry.  The  term  photometry, 
however,  is  very  much  preferable,  and  is  properly  used  to  include 
all  the  branches  of  the  measurement  of  light  and  illumination. 

Quantities  Measured 
The  fundamental  quantity  with  which  photometry  has  to  deal  is 
luminous  ilux.  The  intensity  of  a  source  is  its  flux  per  unit  solid 
angle.  The  illumination  is  flux  per  unit  area.  These  three  quan- 
tities, flux,  intensity  of  a  source  and  illumination,  are  the  chief 
ones  with  which  photometry  has  to  do;  while  specific  intensity — 

>y  Google 

Thus  Ueasubeheht  of  Light  413 

specific  flux,  etc. — are  also  quastitiee  indnded  in  the  ordinary 
Bcope  of  photometry. 


Tb»  Eye  u  a  Photometric  Ziutr&meiit.  The  normal  human  eye 
being  the  only  instrument  which  is  sensitive  to  light,  in  as  far  as 
light  concerns  the  normal  human  being,  it  is  the  eye  which  must 
constitute  the  fundamental  photometric  inBtrument.  The  eye  by 
itself  is  incapable  of  determining  with  any  accuracy  the  intensify 
of  a  source  of  light  or  the  intensity  of  illumination.  Moreover, 
the  eye  is  incapable  of  forming  any  correct  estimate  of  how  many 
times  one  light  is  brighter  than  another.  It  is  only  by  the  use 
of  special  methods  that  the  eye  is  adapted  to  photometric  work. 
These  melhoda  depend  upon  the  following  properties  of  the  eye: 

First.  The  eye  is  capable  of  determining  with  a  considerable 
degree  of  nicety  the  equality  of  the  brightness  of  two  contiguous 
illuminatt?d  fields.  With  special  devices  the  difference  in  brightness 
which  can  be  detected  by  the  eye  is  quite  small,  therefore  photo- 
metric measurements  may  be  made  by  a  zero  method  relative  to 
a  standard  of  luminous  intensity  or  of  illumination. 

Second.  Any  given  eye  under  given  conditions  is  capable  of  de- 
tecting a  certain  degree  of  contrast  with  a  certain  illumination,  or 
of  just  distinguishing  certain  objects  with  a  certain  illumination; 
for  instance,  a  certain  minimum  illumination  is  required  with  a 
given  eye  in  a  given  condition  to  enable  a  certain  print  to  be  read. 
This  point  is  not  very  well  defined,  but  is  sufficiently  well  defined 
to  enable  photometric  measurements  of  a  certain  class  to  he  made 
in  accordance  with  the  principle  involved.  This  is  called  the 
"  visual-acuity  "  method.  There  is  also  a  zero  method  dependent 
on  the  disappearance  of  flicker.  This  will  he  discussed  in  its  proper 

By  the  zero  method  where  the  eye  is  comparing  the  brightness 
of  one  field  with  that  of  another,  and  deciding  when  they  are  equal, 
difficulty  is  encountered  whenever  the  illumination  of  the  two  fields 
difFers  in  color.  Color  differences  represent  differences  in  quality, 
and,  from  a  theoretical  point  of  view,  substances  which  differ  in 
quality  cannot  directly  be  compared  quantitatively.  The  practical 
effect  of  color  diiference  in  photometry  by  the  zero  method  is  to 
make  the  error  of  measurement  considerably  greater,  and  to  give 
rise  to  personal  differences  between  different  individuals  who  ap- 

>y  Google 

414  Illcmimatino  Enoinsbbino 

praise  the  different  colors  according  to  different  personal  standardB. 
Thus  color  differences  constitute  one  of  the  greatest  inherent  diffi- 
culties in  ordinary  photometry.  In  using  the  second  or  the  liminal 
method  referred  to  ahove,  color  differences  are  eliminated,  since 
only  one  color  is  obserred  at  a  time.  The  illumination  observed 
by  this  method  is  given  a  value  proportionate  to  its  usefulness  in 
enabling  objects  to  be  distinguished.  This  value  may  differ  con- 
siderably from  that  obtained  by  the  zero  method. 

In  the  flicker  method  color  differences  are  eliminated. 

Methods — Direct  Comparison  and  Substitaiion 
There  are  two  general  methods  employed  in  the  use  of  photo- 
metric apparatus.  In  the  direct-comparison  method  the  appa- 
ratus is  set  up  in  such  a  way  that  the  source  of  light  to  be  measured 
is  compared  directly  with  the  standard  source,  one  being  placed  on 
one  side  of  the  photometric  apparatus  and  the  other  on  the  other, 
and  the  balance  secured.  In  making  measurements  after  this 
method,  many  precautions  are  required  to  eliminate  the  errors  due 
to  one-sidedness  of  the  apparatus,  or  to  a  tendency  of  the  observer 
to  favor  one  side  rather  than  the  other.  In  working  by  the  sub- 
stitution method,  the  comparison  between  the  source  of  light  to  be 
measured  and  the  standard  is  indirect.  The  procedure  is,  first,  to 
set  up  the  standard  source  of  light  and  compare  with  it  a  constant 
source  of  light  of  convenient  intensity.  Then  the  standard  source 
of  light  is  removed  and  the  unknown  source  is  substituted  for  it. 
The  unknown  source  is  then  compared  with  the  constant  inter- 
mediate source  of  light,  and  its  value  in  terms  of  the  standard  is 
computed  from  the  two  sets  of  measurements.  This  method  of 
procedure  has  the  advantage  over  the  direct-comparison  method 
that  all  errors  due  to  lack  of  symmetry  in  apparatus,  etc.,  are 
eliminated.  The  substitution  method  is  to  be  preferred  to  the 
direct-comparison  method  in  the  great  majority  of  all  cases  arising 
in  photometry. 

Apparatus  for  Zero  Method 
Any  apparatus  for  making  photometric  measurements  according 
to  the  zero  method,  that  is,  by  balancing  the  brightness  of  two 
adjacent  or  contiguous  fields,  consists  essentially  of  the  following 
dements:  First,  an  arrangement  by  which  the  two  adjacent  or 
adjoining  fields  are  obtained,  one  of  the  fields  being  illnminated 
by  the  standard  light  and  the  other  by  the  light  to  be  measured. 

>y  Google 

The  Meiscbeubkt  of  Liqht  415 

Second,  in  an  arrangement  by  which  the  intensity  of  the  illumina- 
tion of  one  or  both  the  fields  may  be  changed  according  to  some 
known  law  until  equality  is  secared.  Third,  a  standard  source  of 

The  apparatus  hy  which  the  contiguous  fields  are  obtained  is 
called  the  eight-box  or  photometer  head,  or,  for  short,  the  photome- 
ter. Properly  speaking,  the  photometer  includes  the  whole  appa- 
ratus, but  the  distinction  here  noted  is  in  many  cases  a.  coQTenient 
one  to  make,  and  no  confusion  should  arise  because  of  the  use  of  a 
term  proper  to  the  whole  apparatiis  for  a  part  of  the  same. 

The  question  of  a  standard  source  of  light  is  a  separate  one  which 
has  been  treated  by  another  lecturer. 

Apparatus  for  Tarying  the  Illumination 
Variable  Distance.  The  simplest  and  most  common  way  to  vary 
the  illumination  on  the  photometer  disc  in  a  known  manner  is  to 
vary  the  distance  between  the  photometer  disc  and  the  source  to 
be  measured.  For  point  sources  the  illumination  produced  is  in- 
versely proportional  to  the  square  of  the  distance  between  the 
source  and  the  illuminated  surface,  the  illuminated  surface  being 
placed  at  right  angles  to  the  rays.  Hence,  by  varying  the  distance 
of  either  of  the  sources  of  light,  or  by  moving  the  photometer  into 
some  position  along  the  straight  line  adjoining  the  sources,  the 
desired  equality  of  illumination  may  be  obtained.  The  mathe- 
matical relations  are  as  follows: 

If  E  is  illumination  on  the  two  fields  of  the  photometer,  I  is 
the  candle-power  of  the  unknown  source,  I'  the  candle-power  of 
the  comparison  lamp,  r  the  distance  between  unknown  lamp  and 
the  photometer  disc,  and  r*  the  corresponding  distance  for  the  eom- 
parieon  lamp, 

IT-    I    -    I' 

The  most  common  arrangement  is  to  set  the  lamps  to  be  measnted 
at  the  extremities  of  a  straight  horizontal  track  or  bar.  On  this 
bar  is  a  carriage  to  which  the  photometer  head  is  attached.  The 
carriage  is  moved  along  between  the'lights  until  the  desired  equality 
is  obtained.    The  results  are  computed  according  to  the  formula 

>y  Google 

416  Illuminatino  Enoineebino 

in  which  I  and  I'  are  the  intensities  of  the  two  sources  of  light,  and 
r  and  r'  the  dietances  between  the  respective  sources  and  the  pho- 
tometer disc. 

It  is  not  infrequently  desirable  to  alter  the  distance  between  only 
one  lamp  and  the  photometer  disc.    For  instance,  the  photometer 
maj  be  stationary,  and  the  distance  of  the  comparison  lamp  may 
be  adjusted  to  give  equal  illuminations.    In  this  case  the  product 
I'r"  does  not  change,  and  the  formula  for  the  photometer  is 
j_  constant 
~        r** 
Or  the  comparison  lamp  and  the  photometer  carriage  may  be  fas- 
tened rigidly  to  each  other  so  that  the  distance  r'  is  constant  and 
the  distance  r  varied.     In  this  case  the  illumination  on  the  pho- 
tometer disc  is  constant  at  all  times,  a  feature  which  has  some 
advantages.    The  formula  in  this  case  becomes 
I  =  constant  xr^ 

A  further  modification  of  the  variable  distance  method  is  an 
arrangement  wherein  the  length  of  the  path  of  light  is  varied  by 
moving  a  miri'or  or  pair  of  mirrors  set  at  right  angles  to  each  other. 
The  photometric  law  remains  the  same. 

Limitations  of  Variable-Distance  Method.  In  employing  the  in- 
verse square  law  it  is  necessary  to  remember  that  it  applies  in  all 
strictness  only  to  point  sources  of  light,  and  that  for  sources  of 
linear  dimensions  large  in  comparison  with  the  distance  at  which 
they  are  measured,  the  law  doea  not  apply.  This  is  due  to  the 
fact  that  an  element  of  the  luminous  body  which  is  not  situated  in 
the  line  normal  to  the  photometer  disc  sends  rffys  to  the  disc  which 
impinge  upon  it  at  an  an^^e  other  than  90°,  and  consequently 
produce  a  smaller  illumination  than  if  they  fell  normally.  More- 
over, the  angle  of  emission  of  these  rays  from  the  luminous  surface, 
supposing  the  luminous  surface  to  be  parallel  with  the  photometer 
disc,  is  not  90°.  These  two  effects,  according  with  Lambert's 
cosine  law,  produce  a  diminution  in  the  illumination.  It  is  there- 
fore necessary  that  the  angle  at  the  photometer  disc  subtended  by 
the  source  of  light  should  be  below  a  certain  limit.  The  rule  has 
been  given  that  the  linear  dimensions  of  the  source  of  light  should 
not  be  over  five  times  the  distance  between  the  source  of  light  and 
the  photometer  disc.  It  is  safe,  and  usually  entirely  convenient, 
to  keep  far  within  the  limitations  of  this  rule. 

>y  Google 

The  Measurement  of  Light  417 

It  is  necesBarj  also  to  see  that  the  angle  of  IncideDce  of  the  light 
upon  the  two  fields  of  the  photometer  is  the  same.  If  the  angle  is 
different  on  ooe  side  from  what  it  is  on  the  other,  an  error  will  be 
introduced  according  to  Lambert's  cosine  law.  Any  such  error  as 
this,  however,  may  be  eliminated  in  the  substitution  method. 

Apparent  Candle-Power.  It  is  frequently  convenient  to  express 
the  photometric  properties  of  a  combination,  such  rb  a  lamp  with 
a  reflector,  or  a  very  extended  source  of  light,  in  terras  of  the 
candle-power  of  a  point  source  which  would  produce  at  a  giveo 
distance  the  same  illumination  as  the  arrangement  to  be  measured 
produces.  For  example,  the  law  of  inverse  squares  cannot  be  as- 
sumed to  hold  for  a  lamp  with  a  concentrating  reflector  within  rela- 
tively short  distances  from  the  lamp.  However,  for  purposes  of 
illumination  computation,  it  is  important  to  know  what  the  equiva- 
lent candle-power  of  the  combination  is  at  some  practical  distance. 
To  this  quantity  the  term  "  apparent  candle-power  "  is  applied,  the 
distance  at  which  this  apparent  candle-power  is  measured  being 
also  specified.  In  reflector  measurements  the  apparent  candle- 
power  at  a  distance  of  10  feet  is  commonly  given.  This  means 
merely  that  when  the  lamp  and  reflector  are  measured  with  the 
photometer  10  feet  away  the  illumination  which  is-  produced  ia 
equivalent  to  that  of  a  lamp  alone  having  the  candle-power  given. 

Kotatingr  Sector  Disc.  If  an  opaque  disc  from  which  equally 
spaced  sectors  of  definite  angular  dimensions  are  cut  is  placed  in 
the  path  of  a  beam  of  light  and  rotated  rapidly,  the  amount  of  radi- 
ation passing  through  the  open  sectors  hears  to  the  total  radiation 
the  same  ratio  that  the  angular  aperture  of  the  open  sectors  does  to 
360°;  that  ia,  if  the  open  sectors  abrogate  36°  in  aperture,  10 
per  cent  of  the  radiation  will  pass  through.  If  the  light  ao  dimin- 
ished falls  upon  a  screen,  and  the  rotation  is  sufBciently  rapid,  the 
eye  will  observe  the  screen  uniformly  illuminated,  and  the  impres- 
sion made  upon  the  eye  will  be,  in  accordance  with  Talbofa  law, 
the  same  as  if  the  same  flax  of  light  fell  upon  the  screen  in  a 
steady  stream  as  actually  falls  on  the  screen  in  the  intermittent 
stream  transmitted  by  the  disc.  Therefore,  physiologically,  as 
well  as  physically,  the  beam  transmitted  by  the  rotating  disc  varies 
as  the  ratio  of  the  angular  aperture  of  the  open  sectors  to  the 
total  periphery  of  the  disc. 

>y  Google 


Venfioation  of  the  Law  of  the  Duo.*  By  a  series  of  careful  ex- 
periments, Lummer  and  Kurlbaum  have  shown  that  for  lights  of 
the  same  color  Talbot's  law  held  for  the  disc  within  the  errors  of 

As  a  result  of  experiments  by  Ferry  f  doubt  had  been  cast  on 
the  validity  of  Talbot's  law  when  Uglits  of  different  color  are  com- 
pared by  means  of  the  sector  disc.    This  question  has  been  investi- 

FiG.  1.— Sector  DlBc  with  Fixed  Apertures. 

gated  by  Hyde,$  whose  careful  experiments  have  shown  that  Tal- 
bot's law  applies  to  the  rotating-disc  method  within  the  errors  of 
observation,  both  when  lights  of  the  same  and  different  colors  are 
compared  with  all  apertures  of  the  disc  from  15°  to  240°. 

*  Zeltschrlft  ffir  Instrumentenkunde.     Blektrotechntsche  Zeitschritt, 
Aug..  1896. 
t  PhyB.  Rev.,  Vol.  1. 
j:Bul1.  Bureau  ot  Standards,  Vol.  II,  p.  1. 

>y  Google 

The  Mbasoeeuent  of  Light  419 

Pnotioal  Pomu  of  Swtor  Diso.  The  eectoi  disc,  which  is  a  most 
important  adjuoct  in  photometric  work,  can  be  made  either  with 
fixed  openings  or  with  variable  openings.  With  fixed  openinge, 
it  is  conveiiient  as  a  means  tor  reducing  the  intensity  ot  a  beam 
of  light  in  a  known  ratio,  an  operation  which  Ib  often  desired  in 
order  to  bring  a  given  measurement  within  the  range  of  a  given 
photometer  bar.  A  fixed  disc  of  this  sort,  as  used  by  the  Bureau 
of  Standards,  is  illustrated  by  Fig.  1.    Evidently  one  motor  may 

Pio.  2. — Sector  Disc. 

be  supplied  with  a  series  of  discs,  so  that  a  variety  of  ratios  are 
obtainable,  but  in  any  case  the  fine  variations  of  photometric  set- 
tings must  be  made  by  some  other  means.  A  disc  may  be  con- 
structed to  produce  any  required  diminution  from  50  per  cent 
downward  by  taking  two  equal  metal  discs,  out  of  which  equally 
spaced  sectors  are  cut,  of  such  dimensions  that  the  open  sectors 
occupy  one-half  of  the  disc.  These  are  mounted  face  to  face  on  a 
shaft,  and  are  provided  with  a  clamp  to  hold  them  together  in  any 
position.  By  sliding  the  digce  over  each  other,  the  amounts  of 
the  open  sectors  of  the  combined  disc  may  be  varied  at  will,  and 
the  ratio  mav  be  read  from  a  graduated  scale  on  one  of  the  discs. 

>y  Google 



In  the  construction  shown  in  Fig.  2,  the  area  of  the  open  sectors 
may  be  varied  while  the  discs  are  in  full  rotation,  thereby  consti- 
tuting  a  device  bj  which  complete  photometric  Bettings  can  be 
made.  The  sector  disc  D  ia  mounted  on  the  axis  A,  while  a  sim- 
ilar disc  D'  is  fastened  to  the  hollow  sleeve  A',  fitting  over  the  ozia 
of  the  first  disc  and  rotating  with  it.  A'  is  pierced  with  the  spiral 
slot  S,  while  A  has  a  longitudinal  groove  of  the  same  width.  A 
hollow  Bleeve  V  fits  over  the  sleeve  A'  and  carries  a  pin  whi<di 
passes  through  the  spiral  slot  and  terminates  in  the  longitudinal 
groove.  A  longitudinal  movement  of  V,  which  can  be  effected  by 
means  of  a  lever  or  a  micrometer  screw  when  the  discs  are  rotating, 
displaces  the  one  disc  with  respect  to  the  other,  and  varies  the 

Fio.  3. — Brodhun's  Sector. 

effective  aperture  of  the  combination.  The  lever  or  micrometer 
screw  is  calibrated  to  show  the  ratio  of  the  disc  and  can  be  read 
without  stopping  the  disc. 

Brodhun's  Variable  Seotor.  Brodhun  has  not  only  constructed 
a  variable  rotating-aector  disc,  in  which  by  special  optical  arrange- 
ment the  actual  angle  between  the  sectors  can  be  read  from  the 
disc  while  rotating,  but  he  has  also  produced  another  and  much 
simpler  apparatus  for  changing  the  intensity  according  to  Talbot's 
law.  In  the  latter  apparatus  the  variable  sector  remains  fixed, 
while  the  beam  of  light  is  caused  to  rotate  about  it.  The  arrange- 
ment is  shown  in  Fig.  3.  The  beam  of  light  striking  the  Fresnel 
prism  P  is  twice  refiected  to  the  Fresnel  prism  P*  on  the  opposite 
aide  of  the  sector  D,  by  which  it  is  returned  to  its  original  axial 
direction.  The  prisms  are  rotated  rapidly,  and  the  photometric 
setting  made  by  the  aid  of  the  adjustable  sector  disc  D,  the  position 
of  which,  since  it  is  stationary,  can  be  read  at  once  from  an  af&xed 
scale.     In  the  form  in  which  it  is  constructed  by  Schmidt  & 

>y  Google 

The  Me-vscremekt  of  Ijght  421 

Haenach,  this  apparatus  is  adapted  to  the  measurement  of  light 
in  rather  small  beams.  There  is  no  reason,  however,  why  the 
principle  should  not  be  applied  to  a  larger  apparatus  made  with 
mirrors  instead  of  prisms. 

Hyde's  Variable-Sector  Disc.  For  the  special  purpose  of  spectro- 
photometry, in  which  the  beam  to  be  photometered  enters  the  nar- 
row slit  of  a  collimator,  Hyde  has  produced  a  very  simple  form  of 

Pic.  4.— Hyde's  Sector. 

variable-sector  disc.  In  this  form  (Fig.  4)  the  apertures  of  the 
disc  are  not  straight  and  radial,  but  are  curved  in  such  a  way  that 
near  the  center  of  the  disc  the  apertures  are  nearly  100  per  cent, 
and  the  aperture  varies  from  that  to  zero  at  a  point  near  the  circum- 
ference of  the  disc.  It  is  evident  that  if  a  slit  which  is  to  re- 
ceive light  is  placed  bo  that  it  lies  at  right  angles  to  the  axis  of 
the  disc,  the  amount  of  light  which  it  will  receive  will  vary  with  the 
position  of  the  disc  with  respect  to  it.  That  is,  when  the  beam 
of  light  reaching  the  slit  passes  through  the  openings  near  the 

>y  Google 

488  Illumimatino  Enoineerinu 

center,  the  diminution  introduced  by  the  rotating  disc  will  be  small. 
This  diminution  can  be  increased  steadily  by  a  lateral  motion  of  the 
disc.  If  the  relation  of  the  disc  carrier,  with  reepect  t^i  the  slit,  is 
fixed  in  the  apparatus,  the  poBition  of  the  disc,  as  indicated  on  a 
scale  with  vefnier,  will  give,  by  a  previous  calibration,  the  per- 
centage of  light  which  the  disc  is  transmitting. 

Uie  of  BiaphragfrnB.  A  diaphragm  may  be  used  in  several  ways 
as  a  means  for  diminishing  the  intensity  of  a  beam  of  light.  If, 
for  example,  the  source  of  light  is  a  uniformly  illuminated  dif- 
fusing surface,  the  amount  of  light  which  it  emits  varies  directly 
with  its  area,  so  that  if  a  diaphragm  is  placed  before  it  the  light 
emitted  will  vary  directly  as  the  area  of  the  opening  of  the  dia- 
phragm: or,  if  a  converging  lens  is  so  placed  that  an  image  of 
the  bright  surface  which  is  the  source  of  light  is  thrown  by  it  on 
to  the  photometer  screen,  the  flux  of  the  beam  may  be  diminished 
by  stopping  down  the  lens,  and  the  intensity  will  vary  very  nearly 
proportionally  to  the  aperture  of  the  diaphragm.  The  greater 
thickness  of  the  lens  toward  the  center,  as  compared  with  the 
sides  and  the  possible  aberration  of  the  lens,  will  cause  this  law 
to  be  not  quite  rigorous,  and  any  such  arrangement  as  this  needs 
to  be  calibrated  by  experimentation.  With  either  of  these  arrange- 
ments the  diaphragm  may  be  one  which  can  be  adjusted  contin- 
uously, whereby  a  convenient  and  effective  device  is  constituted. 
Ordinarily,  the  bright  surface  which  constitutes  the  source  of  light 
will  be  a  piece  of  translucent  glass.  Ground  glass  should  not  be 
used  for  this  purpose,  since  it  is  a  very  poor  diffuser.  Some  of  the 
other  forms  of  glass,  such  as  alabaster  glass,  etc.,  should  be  used, 
and  it  is  preferable  that  the  surface  of  such  glass  shall  be  ground 
as  an  additional  precaution. 

The  diaphragm  principle  may  also  be  used  in  connection  with 
straight-filament  incandescent  lamps.  If  the  image  of  the  filament 
is  thrown  by  means  of  a  lens  on  to  an  adjustable  slit,  with  the 
image  crossing  the  jaws  of  the  slit  at  right  angles,  the  light  trans- 
mitted by  the  arrangement  will  vary  directly  as  the  width  of  the  slit. 

Any  good  form  of  adjustable  diaphragm  can  be  used  for  photo- 
metric work.  The  one  most  commonly  employed  is  Comu'e  "  caf  s 
eye,"  which  consists  of  two  metal  strips  pierced  with  rectangular 
openings,  and  arranged  to  slide  one  upon  the  other  in  the  direction 
of  the  diagonal  line  of  the  openings.    The  movement  may  be  pro- 

>y  Google 

The  Measureubnt  of  Lioet 


duced  by  a  rack  and  pinion,  or  by  a  micrometer  ecrev,  and  the 
position  may  be  read  from  an  attached  vernier  and  scale.  The 
metal  strips  are  illustrated  in  Fig.  5. 

Another  available  form  of  diaphragm  is  the  iris  diaphragm,  which 
is  very  commonly  used  with  photographic  lenses.  The  calibration 
of  such  a  diaphragm  is  made  empirically. 

Polarisatioii.*  If  the  light  from  one  source  is  polarized  by  pass- 
ing through  &  Nicol  or  other  polarizing  prism,  or  by  reflection  from 
a  pile  of  glass  plates  at  the  angle  of  polarization  (about  56°  2(/ 
for  light  crown  glass),  it  loses  more  than  one-half  of  its  intensity 


Fio.  6. — Comu'B  Cat'B  Bye. 

in  tlie  process,  and  the  intensity  of  the  polarized  beam  may  be  still 
further  cut  down  to  any  extent  by  means  of  an  analyzer.  This 
analyzer  may  be  a  duplicate  of  the  polarizer,  or  it  may  be  any  form 
of  totally  polarizing  device.  When  the  polarizer  and  the  analyzer 
are  "  parallel,"  the  polarized  light  emerges  from  the  analyzer  but 
little  decreased  in  intensity.  When  they  are  "  crossed  "  the  beam 
is  entirely  extinguished.  The  intensity  of  the  beam  varies  as  the 
square  of  the  cosine  of  the  angle  of  rotation  of  the  analyzer  with 

•  The  reader  Is  referred  to  the  [ 
tazt-book  of  pbTslca. 

ibject  of  polarization  In  any  good 

>y  Google 

424  Illuminatinq  Enqinbebinq 

respect  to  the  polarizeT.  The  method  must  be  used  vith  precau- 
tion where  there  ia  any  possibility  thfit  the  light  which  ie  to  be 
measured  is  already  polarized  partially,  as  for  example,  light  from 
the  sky.  The  Nicol  prism  Buffers  from  a  farther  disadvaDtage  of 
being  very  expensive  in  large  sizes  and  absorbing  a  very  consid- 
erable percentage  of  the  incident  light,  thereby  producing  a  dark 
field.  Moreover,  its  absorption  is  selective,  being  very  great  in  the 
blue  and  violet  end  of  the  spectrum.  On  account  of  these  dis- 
advantages, and  since  the  required  end  can  usually  be  attained  by 
simpler  means,  the  polarization  method  is  not  very  extensively  used 
in  photometry. 

Absorbing  Kedia.  The  intensity  of  a  beam  of  light  may  be  cut 
down  in  known  ratio  by  passing  it  through  an  absorbing  medium. 
A  prime  necessity  in  the  case  of  such  media  is  that  they  shall  be 
imcolored;  that  ia,  that  they  shall  transmit  all  colors  of  light 
equally.  This  is  a  condition  which  is  scarcely  fulfilled  to  an  exact 
degree  by  any  medium,  but  various  media  arc  available  which  are 
sufficiently  colorless  for  practical  purposes.  For  diminishing  the 
light  to  a  slight  degree,  a  plate  of  clear  glass  may  be  used,  or 
several  plates  may  be  piled  one  on  the  other.  By  inclining  these 
plates  to  the  axis  of  the  beam,  the  amount  of  diminution  may  be 
changed.  The  diminution  in  this  case  ia  produced  chiefly  by  re- 
flection from  the  surfaceH  of  the  plate.  For  a  glass  of  known 
index  of  refraction,  the  light  reflected  on  one  surface,  the  incidence 
being  normal,  the  coefficient  of  reflection  may  be  computed  from 
Fresnel's  equation, 

where  n  is  the  index  of  refraction.  For  example,  with  light  crown 
glass  having  an  index  of  refraction  of  1.5,  the  value  of  the  beam 
transmitted  from  the  air  into  the  glass  normally  is  96  per  cent  of 
the  incident  beam.  A  further  reflection  of  the  same  percentage 
of  the  beam  which  remains  takes  place  on  emerging  from  the  glass 
into  the  air,  so  that  the  total  light  transmitted  is  96  per  cent  by 
96  per  cent,  or  92.Z  per  cent  plus  such  light  as  is  regained  by 
secondary  reflection. 

Absorbing  media  may  be  divided  into  two  Important  classes. 
First,  those  media  which  permit  the  beam  to  pass  unaltered,  except 
in  intensity;  second,  those  which  diffuse  the  light  as  well  as  ab- 
sorbing it.     An  example  of  the  first  class  of  absorbing  media  ia 

>y  Google 

The  Meascbement  of  Light 


ordinary  smoked  glaae.  An  object  can  be  seen  through  a  piece 
of  smoked  glosB  without  an;  distortion,  only  with  a  diminution  of 
the  brightness.  An  example  of  the  second  ig  a  piece  of  alabaster 
glass  OT  of  thin  paper,  that  is,  media  which  diffuse  as  well  as  absorb 
the  light,  and  which  are  commonly  called  translucent.  The  action 
of  the  media  of  the  two  classes  in  photometric  apparatus  is  quite 
different.  A  piece  of  smoked  glass  can  be  interposed  between  a 
lamp  and  a  photometer  at  any  point  in  the  beam,  and  will  cut 
down  the  light  incident  upon  the  photometer  by  a  definite  amount. 
If  a  diffusing  glass  is  used  for  this  purpose,  it  becomes  a  secondary 


Fio.  I 

-Photometric  Wedge. 

source  of  light,  and  the  amount  of  diminution  found  on  the  pho- 
tometer disc  differs  greatly  with  the  position  of  the  diffusing  glass 
with  respect  to  the  disc,  and  if  the  glass  is  stationary  on  the  pho- 
tometer bar  while  the  photometer  is  moTed,  photometric  distances 
must  be  measured  from  it  rather  than  from  the  lamp  which  is  the 
actual  source  of  light. 

The  Photometric  Wet^.  By  using  a  wedge-shaped  piece  of 
smoked  glass,  the  intensity  of  a  beam  of  light  can  be  diminished 
continuously.  This  was  done  formerly  by  Pickering.  An  im- 
provement in  the  photometric  wedge  was  introduced  by  Spitta,* 
who  used  two  wedges  to  slide  over  each  other  instead  of  a  single 
wedge,  as  illustrated  in  Fig.  (i.  With  this  arrangement)  the  thick- 
ness of  the  absorbing  medium  through  which  the  light  must  pass 

•  Proceedings  of  the  Royal  Society  of  London,  Vol.  47,  p.  16,  1889. 

>y  Google 

426  Illuminating  Enoineering 

can  be  varied  from  a  lower  limit,  which  depends  upon  the  acuteness 
of  the  angle  of  the  wedge  and  upon  the  width  of  the  opening  in 
the  diaphragm  through  which  the  light  is  allowed  to  pass,  to  an 
upper  limit,  which  is  nearly  twice  the  thicknesB  of  one  wedge.  The 
loss  of  light  in  passing  through  an  arrangement  of  this  sort  is  due 
to  two  effects:  tirst,  absorption  in  the  wedges;  second,  reflection 
from  the  surfaces  of  the  wedges.  The  second  loss  enters  in  as  a 
constant  quantity,  superimposed  upon  the  absorption  loss,  which  is 
proportional  to  the  thickness  of  the  wedge.  On  this  account,  and 
because  of  inequalities  in  the  glass,  it  is  necessary  that  wedges 
should  be  calibrated  throughout  their  ejitire  range  before  being 
used  in  photometric  measurements.  The  relative  position  of  the 
wedges  may  be  read  from  a  vernier  and  scale  attached  to  them. 

Another  form  of  graduated  absorbing  medium  has  been  employed 
in  a  portable  photometer  by  Dr.  "Williams,"  who  used  a  photographic 
film  which  had  been  exposed  and  developed  so  that  it  showed  a 
gradually  increasing  density.  Evidently  this  plan  is  capable  of 
considerable  development.  An  arrangement  of  this  kind  must  also 
be  calibrated  empirically  throughout  its  length. 

Inclined  Plate.  If  the  source  of  light  is  a  diffusely  reflecting 
or  transmitting  surface,  as,  for  instance,  an  illuminated  s\irface  of 
plaster  of  Paris,  or  a  window  of  diffusing  glass,  the  light  which 
it  sends  in  a  given  direction  may  be  altered  by  changing  the  angle 
between  the  normal  to  the  plate  and  the  direction  in  question.  If 
the  diffusing  surface  is  a  good  one,  the  diminution  of  light  as 
the  plate  is  turned  from  the  normal  will  vary  proportionately  to 
the  cosine  of  the  angle  for  considerable  angles  from  the  normal. 
In  any  case,  it  is  advisable  that  a  plate  used  in  this  way  should  be 
calibrated  empirically. 

Varying  the  Source  of  Light.  In  ceri:ain  apparatus  the  photo- 
metric setting  has  been  made  by  varying  the  total  amount  of  light 
given  by  the  comparison  source.  With  a  flame  source  this  may 
be  done  by  raising  and  lowering  the  flame.  Then  a  measurement 
of  the  flame  height  indicates  the  photometric  setting.  When  an 
incandescent  lamp  is  used  as  the  comparison  source,  its  intensity 
may  be  varied  over  considerable  limits  by  varying  the  impressed 
voltage.  This  has  the  disadvantage  that  the  color  of  the  light  varies 
at  tlie  same  time.  The  first  of  the  above  methods  is  adapted  to 
only  the  roughest  kind  of  work. 

*  Transactions  IllumlnatiOK  Engineering  Society,  p.  G40,  1907. 

>y  Google 

The  AIeasdkemekt  op  Light  427 


In  tlio  fields  of  photometric  sight  hoxes  two  principles  are  made 
use  of.  First,  is  the  equality  principle.  In  photometers  intended 
to  employ  this  principle  the  fields  are  so  constructed  tliat  the  eye 
compares  their  brightness  directly,  and  endeavors  to  tell  when  they 
are  equally  bright.  It  is  very  difficult  to  do  this  with  any  degree 
of  accuracy  unless  the  fields  are  so  arranged  that  the  dividing  line 
between  them  disappears  when  the  equality  point  has  been  reached. 
Hence,  such  photometers  are  called  also  "  disappearance  "  photome- 
ters. Then  the  eye,  by  observing  the  merging  of  one  field  into 
another,  can  determine  the  point  of  equality  with  considerable 
accuracy.  Second,  is  the  contrast  principle.  In  accordance  with 
this  principle,  each  field  consists  of  two  parts:  first,  a  part  il- 
luminated by  its  own  proper  source;  second,  a  part  illuminated 
by  the  other  source,  but  to  a  different  degree.  With  this  arrange- 
ment equality  exists  when  the  contrast  in  the  right-hand  field  be- 
tween the  portions  of  the  field  illuminated  by  source  A  and  source 
B  is  the  same  as  the  contrast  in  the  left-hand  field  between  tlie 
portions  illuminated  by  source  B  and  source  A.  If  the  degree  of 
contrast  is  the  correct  one,  the  eye  is  able  to  determine  this  equality 
of  contrasts  with  great  precision. 

In  photometers  using  the  contrast  principle,  the  equality  prin- 
ciple may  also  he  employed,  since  when  the  equality  of  contrast 
is  established  equality  of  illumination  is  also  observed.  If,  when 
equality  of  illumination  is  observed,  exact  equality  of  contrast 
ie  not  observed  the  construction  of  the  apparatus  is  faulty.  The 
degree  of  contrast  which  gives  the  most  sensitive  arrangement  de- 
pends to  some  extent  upon