Skip to main content

Full text of "Engineering chemistry : a manual of quantitative chemical analysis .."

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 

'-'-  .'  ■   '.'■:  -"■:■■  '  '  ■    \    \  r> 











RA8TON,  PA.: 



■         •        ■ 

« J  •     -    , 

*  «  V  •  «  '  m 

*     • 

•  «  •    •      ■     ■  *      • 


•  •  •  •  • 

•  •  • 

•       •        * 

.  V 

'  •  •    •  • 

«    « 
•  ♦ 



The  preparation  of  this  manual  has  resulted  from  many  years 
of  experience  in  the  chemical  laboratory,  the  work  of  which  has 
been  closely  connected  with  engineering,  and  with  the  teaching 
of  these  subjects  to  students. 

A  treatise  ot  this  character  cannot  be  too  comprehensive  in 
the  treatment  of  a  subject,  nevertheless  better  results  are  ob- 
tained, from  students,  by  arranging  the  matter  in  such  a  way 
that  the  principles  and  methods  of  work  are  indicated  and  then 
references  given  for  further  study  and  research. 

Students  commencing  quantitative  chemical  analysis  can  with 

profit  perform  the  first  eleven  exercises  given  in  this  work,  with 

proper  supervision,  and  then  select  a  course  of  study  suitable  to 

their  advancement :  either  for  iron  and  steel  chemistry  ;  railroad 

I  laboratory  practice  ;  the  technical  application  ot  water  supply  ; 

I  the  chemical  technology  of  fuels,  etc.,  etc. 

It  will  be  found  conducive  to  thorough  work,  that  each  stu- 
I  dent  before  finishing  any  investigation,  be  required  to  write  not 
[  only  the  analytical  data,  but  also  the  references  to  the  literature 
}  bearing  upon  the  subject  examined,  following  the  plan  outlined 
I  in  the  manual. 

The  articles  upon  gas  analysis  and  valuation,    blast  furnace 
I  practice,    the  heating  value  of  fuels,  the  purification  of  water  for 
technical  purposes,    lubrication,    car  illumination,    and  the  ex- 
amination of  Portland  cement,  have  received  especial  attention, 
since  these  topics,  at  the  present  time,  form  a  considerable  por- 
I  tion  of  the  work  and  investigations  of  engineers. 

The  following  articles  have  been  contributed  by  experts  in 
'  each  line  of  study  : 

'Blast  Furnace  Practice,"  by  Edward  A.  Uehling,  M.E. 
"  Determination  of  the  Heat  Balance  in   Boiler  Tests,"  and 
contributions  of  portion  of  the  article  upon  "Pyromctry,"  by 
Wm.  Kent,  M.E. 


** Carbon  Compounds  of  Iron,**  by  G.  C.  Henning,  M.E. 
"Practical  Photometry,**  by  Alten  S.  Miller,  M.E. 
••Electrical  Units,**  by  Albert  F.  Ganz,  M.E. 
••Energy  Equivalents,**  by  E.  J.  Willis,  M.E. 

The  author  has  endeavored  to  acknowledge  every  excerpt 
made  by  him,  with  the  proper  reference  thereto,  and  his  thanks 
are  due  to  those  chemists  from  whose  experiences  valuable 
methods  of  analysis  have  been  incorporated  in  the  manual. 

Thomas  B.  Stillman. 

Stbvbns  iNSxrruTE  op  Technology, 
HoBOKKN,  N.  J.,  Dec.  31,  1896. 



^   Determination  of  Iron  in  Iron  Wire i 

Determination  of  Alumina  in  Potash  Alum 2 

Determination  of  Copper  in  Copper  Sulphate 3 

As  copper  oxide,  by  precipitation  with  sodium  hydroxide,  3  ; 
Volu metrically  with  potassium  cyanide  solution,  4 ;  As  metal- 
lic copper  by  electrolysis,  5  ;  Giilcher's  thermo-electric  pile,  7. 

Determination  of  Sulphur  Trioxide  in  Crystallized  Magne- 
sium Sulphate 8 

Determination  of  Lead  in  Galena 9 

Determination  of  Iron  by  Titration  with  Solution  of  Potas- 
sium Bichromate 10 

a.  Where  the  iron  solution  is  in  the  ferrous  condition,  b. 
Where  the  iron  solution  is  in  the  ferric  condition,  11 ;  The 
Bunsen  valve,  12. 

Determination  of  Phosphoric  Acid  in  Calcium  Phosphate-  12 
Determination  of  Chromium  Trioxide  in  Potassium  Bichro- 
mate   14 

w   Analysis  of  Limestone 15 

Scheme  for  analysis,  16  ;  Determination  of  the  carbon  dioxide 
17  ;  Calculation  of  the  analysis,  18;  Phosphoric  acid  in  lime- 
stone, 18. 

y  Coal  and  Coke  Analysis 19 

Determination  of  moisture,  volatile  and  combustible  matter, 
fixed  carbon  and  ash,  19 ;  Sulphur  by  fusion  with  sodium  car- 
bonate and  potassium  nitrate,  20 ;  Sulphur  by  the  Eschka- 
Fresenius  method,  21 ;  Determination  of  CaS04,  21 ;  Determi- 
nation of  phosphorus.  22;  Analysis  of  *'Bog  Head  Cannel  *' 
coal,  22;  "Pittsburg  Bituminous "  coal,  23;  Increase  of  ash 
with  decrease  of  size  of  coal,  23 ;  Analysis  of  ash  of  coal  or 
coke,  23  ;  Sample  analysis,  24  ;  Valuation  of  coke,  24  ;  Thor- 
ner  compression  machine  for  coke,  24 ;  Standard  of  strength 
of  coke,  24  ;  Apparent  specific  gravity  of  coke,  25  ;  The  true 
specific  gravity,  25 ;  The  volume  of  pores  in  100  volumes  of 
material,  25  ;  Method  of  making  complete  report  upon  a  coke, 
27;  Weight  per  cubic  foot  of  coke,  27;  Fulton's  standard 
table  for  coke,  28 ;  References  on  the  literature  of  coke,  28. 

Scheme  for  the  Analysis  of  Hematite,  Limonite,  Magnetite 

and  Spathic  Iron  Ores 29 

Determination  of  silica,  30 ;  of  phosphorus  pentoxide,  30 ;  of 
iron,  30 ;  of  sulphur,  30 ;  of  alumina,  30 ;  of  manganese,  31 ; 
of  lime,  31 ;  of  magnesia,  31 ;  of  water  of  hydration,  31 ;  De- 



termination  of  lerroas  oxide  in  PeO.Fc^O„  3a ;  Allen's  metliod, 
33 ;  Method  of  fusion  of  iron  ores  insoluble  in  scids,  13 ;  De- 
termination of  chromium,  33 ;  Genth's  method,  33 ;  Table  of 
analyses  of  yarions  chrome  iron  ores,  34 ;  Determination  of 
titanium,  35 ;  Method  of  Bettel,  35 ;  References  on  literature 
of  iron  ore  analyses,  35 ;  Table  ofthe  composition  of  various 
iron  ores,  36. 

Scheme  for  the  Analysis  of  Blast  Pomace  Slag 37 

Form  of  blank  used  for  reporting  blast  furnace  slag  analyses, 
38 ;  Examples  of  blast  furnace  s&ig  analyses,  39 ;  Analysis  of 
open-hearth  sl«gs,  refinery  slags,  tap-cylinder,  mill-cinder  and 
converter  slags.  39 ;  Calculation  of  the  amount  of  material 
required  for  a  furnace  producing  300  tons  of  pig  iron  per  day, 
39-42 ;  Heat  energy  developed,  42 ;  The  stopping  of  furnaces 
for  repairs,  43 ;  The  charging  of  blast  furnaces,  43-45 ;  De- 
scription of  the  three  different  methods  of  reduction  in  the 
blast  furnace,  46-48;  Calculation  of  blast  furnace  sliur,  48; 
Analyses  of  the  iron  ore,  limestone  and  coal  used,  49 ;  Trans- 
formation ofthe  three  analyses  into  lime,  49-51 ;  Examples  of 
close  coincidence  between  slags  actually  run  from  known  cal- 
culated charges  and  the  slag  determined  a  priori,  52,  53 ; 
Table  of  types  of  slazs,  including  acid,  s^quiacid,  neutral, 
sesqnibasic,  oibasic  and  tribasic,  54 ;  Graphic  method  of  cal- 
culating blast  furnace  charges,  55-57 ;  References  on  the  litera- 
ture of  blast  furnace  slags,  57. 

The  Analysis  of  Water  to  Determine  the  Scale- Forming  In- 
gredients       58 

The  usual  components  of  boiler  scale,  58 ;  The  importance  of 
the  determination  of  the  alkalies  in  water  for  boiler  use,  58 ; 
Example  of  a  boiler  scale  containing  72  per  cent,  of  sodium 
chloride,  58;  Scheme  for  water  analysis  for  scale- forming  in- 
f^redients,  60;  Determination  of  silica,  alumina,  oxide  of 
iron,  calcium  oxide,  ma^esinm  oxide,  sodium  oxide,  potas- 
sium oxide,  carbon  dioxide,  sulphur  trioxide,  and  chlorine  in 
a  sample  of  water  with  the  quantitative  results  of  each  given, 
61-^5  ;  Table  showing  the  number  of  erains  per  United  States 

f;allon  and  Imperial  gallon  corresponding  to  milligrams  per 
iter,  63,  64 ;  Method  of  stating  results  of  an  analysis,  65 ; 
Analysis  of  a  water  not  containing  calcium  sulphate,  66; 
Action  of  magnesium  chloride  as  a  corrosive  a^ent  in 
boilers,  66,  67 ;  Statements  in  grains  per  gallon  ana  not  in 
parts  per  100,000  or  1,000,000,  67 ;  Example  of  a  very  corrosive 
water,  67  ;  Determination  of  free  acid,  68 ;  Determination  of  the 
hardness  of  water,  by  standard  sulphuric  acid,  69 ;  Determi- 
nation of  the  hardness  by  the  soap  test,  70,  71 ;  The  French 
standard  of  hardness,  the  German,  the  English,  and  the 
the  American,  72  ;  Table  showing  the  hardness  of  water  sup- 
plied to  cities,  73. 

The  Sanitary  Analysis  of  Water 73 

Determination  of  chlorine,  73  ;  Amount  of  chlorine  allowable 
in  potable  water,  74 ;  Determination  of  the  free  and  albuminoid 
ammonia,  74-78;  Wolff's  colorimeter,  76;  Preparation  of  the 
standard  Nessler  solution,  74 ;  Standard  ammonium  chloride. 


VI 1 

741  Standard  alkaline  permaDganate,  74;  Amounts  of  free 
and  albuminoid  ammonia  allowable  in  potable  water,  79;  Ap- 

Saratus  used  by  the  New  York  City  Health  Department  for 
etenuination  ol  free  and  albuminoid  ammonia,  80 ;  Determi- 
nation of  nitrates  by  the  phenol  method.  80 ;  Preparation  of 
standard  potassium  nitrate  solution.  80 ;  of  phenol sul phonic 
acid,  80;  Determination  of  nitrites,  81  ;  Griess's  methoa  modi- 
fied by  Glosway,  81  ;  Preparation  of  sodium  nitrite  solution, 
81 ;  of^a-amido-iiapbthelene  acetate  solution,  Si ;  Process  of 
determination  ol  nitrites,  81  ;  Oiyeen  required  to  oxidize 
organic  matter,  81;  Conversion  table,  paris  per  1,000,000  to 
KTBins  pet  gallon,  etc.,  83;  Table  of  analyses  of  thirty-nine 
American  well  and  river  waters,  S4  ;  Table  of  composition  of 
various  European  Inke  and  river  waters,  Ss  ;  Table  of  compo- 
sition of  various  ocean  waters,  85  ;  Description  of  the  filter 
beds  of  city  of  Dublin.  86;  Descriptionof  the  Warren  filters  for 
rapid  filtration  of  water,  88-91  '•  References  on  the  bacterio- 
logical examination  of  water,  92. 

The  Composition  of  Boiler  Scale qi 

Analysis  of  boiler  scale  from  boilers  at  Birmingham,  Ala., 
showing  that  calcium  and  magnesium  hydroxide  may  exist, 
replacing  a  portion  of  the  calcium  carbonate  and  magnesium 
carbonate,  92-94;  Examination  of  boiler  scale,  layerby  layer, 
shows  where  scale  is  in  contact  with  red  hot  iron,  carbon  diox- 
ide is  absent,  94  ;  Change  in  method  of  analysis  of  boiler 
scale,  if  oil  is  present,  94 ;  Atnount  of  water  evaporated  by  a 
too  horse-power  boiler  per  month.  95 ;  Amount  of  loss  of  heat, 
fuel,  etc.,  by  scale  in  boilers,  95;  Estimate  of  the  Railway 
Master  Mechanics  Association  of  the  United  States  of  the  loss 
of  fuel,  repairs,  etc.,  for  locomotive  boilers  due  to  scale,  95. 

[  Water  for  Locomotive  Use g6 

Results  of  experiments  made  by  the  Chicago,  Milwaukee  & 
St.  Paul  Railway,  96;  For  practical  purposes  water  is  classi- 
fied as  incrusting  and  nou-incrusting,  96;  Foaming  of  alkali 
water  in  boilers,  97;  Maximum  residue  allowable,  97  ;  One  to 
ten  grains  solid  per  gallon  is  classed  as  soft  water,  ten  to 
twenty  grains  moderately  hard  water,  above  twenty -five  grains 
very  nard  water,  97 ;  Use  of  boiler  compounds  to  prevent 
scale,  97  ;  Forniulafor  compound  used  by  Chicago,  Milwaukee 
Si.  St.  Paul  Railway  to  prevent  scale,  97;  Washing  out  of 
locomotive  boilers  frequently  with  hot  water  necessary,  99: 
Reasons  why  good  boiler  compounds  to  prevent  scale  can  be 
nsed  profitably,  99. 

|.Feed  Water  Heaters 99 

Feed  water  heaters  as  scale  eliminators  in  boiler  waters,  99 ; 
Boiler  economizers,  99  ;  Principle  upon  wbicb  feed  water  heat- 
ers operate,  100;  Temperature  required  to  precipitate  calcium 
carbonate,  100;  Temperature  required  to  precipitate  calcium 
sulphate,  100 ;  The  Goubert  upright  feed  water  heater,  too: 
BsEiauBt  steam  and  superheated  steam  in  feed  water  beaters.  loi; 
Exhaust  steam  precipitates  calcium  carbonate  but  not  cal- 
cium sulphate,  a  temperature  of  140° 'F.  being  required  for 
the  latter,  100 ;  The  Hoppes  feed  water  purifier,  loi ;  Example 
of  composition  of   boiler   water  before   treatment    with    the 


Hoppes  purifier  and  after  treatment,  102 ;  Table  showing  the 
yearly  saving  effected  by  the  use  of  the  feed  water  heaters  for 
various  horse-powers  at  different  prices  of  coal,  103 ;  Table 
showing  percenta^s  of  fuel  saved  by  heating  feed  water 
(steam  pressure  sixty  pounds),  104;  "Blowing  off"  as  a 
means  of  prevention  of  scale  in  boilers,  105. 

Use  of  Chemicals  and  Filtration  for  Purification  of  Boiler 

Waters f 105 

The  Dervaux  water  purifier,  105,  106;  The  Archbutt  water 
purifier,  107-109 ;  Table  showing  the  cost  of  purification  of 
boiler  waters  from  the  analysis  of  the  same,  by  the  Archbutt 
process,  no;  Use  of  sodium  carbonate,  no. 

Filter  Presses  for  Rapid  Filtration  of  Water.  • iii 

Description  of  the  two  varieties  of,  in,  112  ;  Chamber  presses 
and  frame  presses,  112  ;  The  Porter-Clarke  process  for  soften- 
ing water,  112;  Use  of  fibers  of  cellulose  in  filter  presses  to 
collect  finely  divided  precipitates,  112  ;  Description  of  a  com- 
plete plant  for  water  purification,  using  superheater,  chemi- 
cal precipitation  with  sodium  carbonate,  and  filter  presses, 
113  ;  References  on  water  analysis,  boiler  scale,  purification  of 
water,  etc.,  114. 

Determination  of  the  Heating  Power  of  Coal  and  Coke  •  •  •  •    114 

Ignition  of  coal  in  a  crucible  with  litharge,  115;  Method  of 
calculation  of  results,  115;  Three  methods  available  for  the 
determination  of  the  heating  power  of  coal  and  coke  :  (i)  Cal- 
culation of  the  heating  power  from  an  elementary  analysis  of 
the  coal,  (2)  The  useofcalorimeters,(3)The  combustion  of  large 
amounts  of  coal  in  specially  designed  apparatus  therefor,  115; 
Determination  of  carbon  and  hydrogen  in  coal,  11 5-1 17;  De- 
termination of  nitrogen,  117,  118  ;  Data  for  calculation  of  the 
heating  power  from  the  analysis,  120 ;  Definition  of  a  calorie, 
120;  Definition  of  a  British  thermal  unit  (B.  T.  U.),  120;  For- 
mula for  calculation  of  heating  power  when  products  of  com- 
bustion are  condensed,  121  ;  When  products  of  combustion  es- 
cape in  steam,  121 ;  Calculation  of  the  amount  of  air  required 
for  combustion  of  one  kilo  of  coal,  122;  Calculation  of  the 
amount  of  air  required  for  combustion  of  one  kilo  of  coke,  123 ; 
Calculation  of  the  evaporation  value  of  coal  and  coke,  123 ; 
Table  showing  the  air  required,  the  total  heat  of  combustion, 
evaporative  power,  etc.,  of  one  kilo  of  carbon  and  one  pound 
of  carbon  burning  to  carbon  dioxide,  to  carbon  monoxide,  of 
hydrogen  and  of  sulphur,  124 ;  Cause  of  loss  in  actual 
evaporation  in  boiler  practice,  125  ;  Results  of  boiler  evapora- 
tive tests  made  by  J.  E.  Denton,  125  ;  Ordinary  boiler  evapo- 
ration in  less  than  eighty  per  cent,  of  the  theoretical  value,  125. 

Calorimetry 125 

The  Mahler  calorimeter,  description  of,  125-127;  Determina- 
tion of  the  water  equivalent  of  the  Mahler  calorimeter,  128, 
129  ;  Detail  of  process  of  determination  of  the  heating  power 
of  coal  with  the  Mahler  calorimeter,  129,  130 ;  Example  show- 
ing method  of  calculation,  130,  131 ;  Results  of  tests  upon  five 
samples  of  coal,  made  in  the  laboratories  of  the  Stevens  In- 


stitule,  131;  Refereuces  upon  the  Mahler  calorimeter,  131; 
The  Thompson  calorimeter,  133;  Method  of  determi nation  of 
the  water  equivalent,  132  ;  Determi  nation  of  the  heating  power 
of  a  coal  with  the  Thompson  calorimeter,  1J3  ;  Comparison  of 
the  theoretical  heating  value  of  a  coal,  as  determined  by 
aualysis,  and  of  the  determination  as  made  by  the  Thompson 
calorimeter,  134,  135;  The  Barrus  coal  calorimeter,  135-137; 
Resalts  of  tests,  upon  several  coalsi  with  the  Barrus  coal  cal- 
orimeter, 138;  The  Fischer  calorimeter,  139;  Carpenter's  coal 
calorimeter,  139,  140;  References,  141;  Description  of  the 
Kent  apparatus  for  determining  the  beating  power  of  fuels  in 
large  quantities,  i4i-[43 ;  Boiler  tests  of  coal.  144  ;  Table  of 
the  approximate  beating  value  of  coals,  145  ;  Determination  of 
the  efficiency  of  a  boiler,  147:  Delermioation  of  the  several 
losses  of  heat  in  boiler  practice.  147;  Method  of  making  a 
"heat  balance"  in  boiler  tests,  :47-l49  ;  Sources  of  error  in 
making  a  "  beat  balance,"  150 ;  Estimations  of  radiations  of 
heat  by  difference,  150. 

iDetemiinatioa  of  Sulphur  in  Steel  and  Cast-Iron 140 

Bromine  method.  151;  Aqua-regia  method,  152;  The  potas- 
sium permanganate  method,  iji-154  ;  The  ioiline  method,  154, 
15s ;  References  on  the  determination  of  sulphur  in  steel  and 
cast-iron,  156. 

I  The  Determination  of  Silicon  in  Iron  and  Steel 156 

References,  157. 

[  The  Determination  of  Carbon  in  Iron  and  Steel 157 

Report  of  the  English,  Swedish,  and  American  committees 
upon  the  methods  for  determination  of  carbon,  157  ;  Method 
ofBerzelius,  158:  Method  of  Regnault,  Dcville,  and  Wohler, 
158;  Ullgren's  method,  Eggertz,  Langley,  Richtcr,  Weyl  and 
Binks,  Parry.  McCreath,  Boussingault,  Wiborg,  1591  Selec- 
tion of  the  best  method,  160 ;  Method  as  used  by  author,  160 ; 
Description  of  apparatus  used  in  chromic  acid  process,  160-162; 
Method  of  Langley  modified  as  used  by  author,  with  descrip- 
tion of  apparatus,  163-163;  Wiborg's  method,  165-167;  De- 
scription of  Eggertz's  method  for  combined  carbon  in  steel, 
16S,  169;  Stead's  modification,  169. 

Y  Carbon  Compounds  of  Iron 170 

Microscopical  examination  of  iron,  170;  Marien's  and  Os- 
mond's latest  investigations,  170;  Composition  of  unbardened 
steel,  Fe,C ;  Composition  of  ferrite  and  reactions  of,  171; 
Cementile,  172;  Perlite,  17a;  Martensite,  17a;  Sorbite,  172; 
Troostite.  173;  Systematic  microscopical  examination,  174; 
Distinction  between  martensite  and  perlite,  174  ;  Differences 
in  reai^tions  between  ferrite,  cemenlite,  and  troostite.  174; 
References  upon  carbon  in  iron,  174  ;  References  upon  deter- 
mination of  carbon  in  iron  and  steel,  175. 
The  Determination  of  Phosphorus  in  Cast-iron  and  Steel. .  176 
Tbe  molybdate  method,  176;  Preparation  of  the  standard 
solutions.  177,  178;  Determination  of  phosphoric  acid  in  the 
ammonto-molybdic  phosphate,  by  direct  weighing  of  tbe  yel- 
low precipitate,  178  ;  The  agitation  apparatus  of  Spiegelberg's 


for  precipitation  of  phosphoric  acid,  179;  Volumetric  deter- 
mination of  phosphorus  in  iron  and  steel,  179;  Apparatus  and 
reagents  requirea,  180;  Calculations  of  analyses  made  by 
volumetric  method,  182  ;  References  on  the  determination  of 
phosphoric  acid  in  iron  and  steel,  183. 

The  Classification  of  Steel 183 

Classification  as  made  by  the  Mid  vale  Steel  Co.,  183  ;  Class  O, 
carbon,  o. I  to  0.2  per  cent.;  Class  I,  carbon,  0.2  to  0.3  per 
cent. ;  Class  II,  carbon,  0.3  to  0.4  per  cent. ;  Class  III,  carbon, 
0.4  to  0.5  per  cent. ;  Class  IV,  0.5  to  0.6  per  cent. ;  Class  V, 
carbon,  0.6  to  0.7  per  cent.;  Class  VI,  carbon,  0.7  to  0.8  per 
cent. ;  Class  VII,  carbon,  0.8  to  0.9  per  cent.  ;  Class  VIII,  car- 
bon, 0.9  to  i.o  per  cent. ;  Class  IX,  carbon,  i.o  to  percent.  ; 
Class  X,  carbon,  to  1.20  per  cent.,  183-184  ;  Effect  of  other 
ingredients  besides  carbon  or  tensile  strength.  184 ;  Purp>oses 
for  which  the  different  classes  of  steel  are  recommended,  184  ; 
Phosphorus  limit  in  machinery  steel  must  be  below  0.06  per 
cent.,  185;  Phosphorus  limit  in  gun  forgings,  tool  steel,  and 
spring  steel  must  be  below  0.03  per  cent.,  185  ;  Magnetic  prop- 
erties of  steel,  185 ;  Effect  of  nickel  on  magnetic  properties, 
185  ;  Experiments  made  by  the  Bethlehem  Iron  Co.  on  nickel 
steel,  185  ;  Reauirements  of  carbon,  phosphorus,  manganese, 
silicon,  and  sulphur  for  locomotive  steel  plates,  186 ;  Kent's 
classification  of  iron  and  steel,  187  ;  **  Mitis"  steel,  187. 

Determination  of  Aluminum  in  Iron  and  Steel 188 

Drown*s  method,  188 ;  Table  of  results  of  experiments  on 
quantitative  determinations,  189;  Method  of  Carnot,  190; 
References,  190. 

Determination  of  Sulphuric  Acid  and  Free  Sulphur  Triox- 

ide  in  Fuming  Nordhausen  Oil  of  Vitriol 190 

Determination  of  Manganese  in  Iron  and  Steel 192 

Initial  treatment  of  the  manganese  for  determination  either 
gravimetrically,  volumetrically,  or  colorimetrically,  192 ; 
Gravimetric  method,  193 ;  Preparation  of  standard  solutions 
of  ferrous  sulphate  and  potassium  bichromate  for  the  volu- 
metric process,  193;  Colorimetric  method,  as  modified  by  J. 
J.  Morgan,  194  ;  Textor's  method  for  the  rapid  determination 
of  manganese  in  steel,  194,  195 ;  References,  195. 

Technical  Determination  of  Zinc  in  Ores 195 

Preparation  of  standard  solution  of  potassium  ferrocyanide, 
195  ;  Of  pK>tassium  chlorate  and  ammonium  chloride,  196 ; 
Precautions  to  be  observed  in  the  process,  197. 

Sodium  Cyanide  as  a  Component  of  Potassium  Cyanide.  •  ••    197 

The  valuation  of  p>otassium  cyanide  for  commercial  purposes, 
197;  Composition  of  "  ninety -eight  per  cent."  cyanide,  197; 
Method  to  be  used  for  analysis  of  the  mixed  cyanides,  198 ; 
Determination  of  method  of  manufacture  from  the  analysis, 
199  ;  Comparison  of  the  cost  of  manufacture  of  sodium  cyanide 
and  of  potassium  cyanide,  199,  200. 


The  Chemical  and  Physical  Examination  of  Portland  Ce- 
ment   200 

Limit  of  variation  in  the  composition  of  Portland  cements, 
200 ;  CompK>sition  of,  200 ;  Effect  of  magnesia,  201  ;  Injurious 
effect  of  calcium  carbonate,  201 ;  Scheme  of  analysis  of  Port- 
land cement,  202  ;  Determination,  quantitatively,  of  the  con- 
stituents with  an  example,  203, 204  ;  Analyses  of  *'  Burham's," 
**Dyc.kerho£f's,*'  and  "  Saylor's"  Portland  cements  by  the 
author,  205  ;  List  of  analyses  of  German  cements,  205  ;  The 
mechanical  testing,  205 ;  Rules  of  the  American  Society  of 
Civil  Engineers  for  testing  cements,  206,  207 ;  Description  of 
the  "  Fairbank*s"  cement  testing  machine,  208;  Of  the 
"Riehl^,"  209;  Directions  for  testing  cements  according  to 
the  official  German  rules,  209,  210 ;  Standard  sand,  210 ;  Pre- 
paration of  Briquettes  of  neat  cement,  210 ;  Briquettes  of  a 
mixture  of  Portland  cement  and  standard  sand,  211  ;  Descrip- 
tion of  the  "  Michaelis'*  cement  testing  machine,  212 ;  Of  the 
*'  Reid  and  Bailey"  machine,  213;  The  *'Faija*'  and  *' Grant" 
machines,  213 ;  Causes  of  variations  in  tensile  stren^h  in 
cements,  214;  Experiments  of  Dr.  Bohme,  214;  The  Bohme- 
Hammer  apparatus,  216;  Description  of  Jameson's  automatic 
briquette  molder,  217,  218  ;  Table  showing  results  of  tensile 
tests  on  the  same  samples  of  cement,  by  nine  different  ex- 
perts, 218;  Conditions  required  in  France  for  a  good  cement, 
219  ;  Description  of  the  Buignet  cement  machine^  220,  221  ;  M. 
Durand-Claye's  experiments  on  briquettes  of  Portland 
cement,  221 ;  The  crushing  test  of  cements,  222  ;  Ratio  of 
tensile  strength  to  crushing  strength,  222  ;  The  "  Suchier," 
the  "  Tetmajer,"  and  the  **  Arasler-Laffon"  machines  for  de- 
termination of  crushing  strength,  222-224 1  Variation  in  vol- 
ume of  cements,  224 ;  Hot  water  tests,  224 ;  Porter's  auto- 
matic cement  testing  machine*  225,  226 ;  Resum^  of  tests  re- 
quired for  Portland  cements,  227  ;  References  :  The  Journal 
American  Chemical  Society y  16,  161,  283,  323,  374,  contains  an 
index,  arranged  by  the  writer,  of  the  literature  relating  to 
Portland  cement  from  1870  to  1893. 

The  Determination  of  Nickel  in  Nickel-Steel 227 

Principles  of  the  process,  227,  228;  The  electrolytic  method, 
229,230;  Volumetric  method,  230;  Special  apparatus,  230; 
Preparation  of  the  standard  solutions  of  sodium  phosphate, 
230 ;  of  sodium  acetate,  231  ;  of  potassium  cyanide,  231  ; 
of  nickel  solution,  231  ;  of  cupric  ferrocyanide  solution,  232  ; 
Experiments  show  that  the  volumetric  method  gives  results 
within  0.0003  gram  of  true  nickel  content  in  2.222  grams  of 
nickel-steel,  232. 

Analysis  of  Chimney  Gases  for  Oxygen,  Carbon  Dioxide, 

Carbon  Monoxide,  and  Nitrogen 233 

Description  of  the  Elliott  apparatus,  233,  234  ;  Method  of  col- 
lecting the  gas,  234;  Strength  of  potash  solution  for  absorp- 
tion of  carbon  dioxide,  234 ;  Preparation  of  alkaline  pyrogal- 
late  solution  for  absorption  of  oxygen,  235  ;  Solubility  of  car- 
bon dioxide  and  carbon  monoxide  in  distilled  water,  235  ;  Pre- 
cautions to  be  observed  in  the  determination  of  carbon  mon- 


oxide,  235 ;  Preparation  of  cuprous  chloride  solution  for  ab- 
sorption of  carbon  monoxide,  236 ;  Data  for  converting  percen- 
tages by  volume  to  percentages  by  weight,  237  ;  Example  of  an 
analysis  of  a  chimney  gas,  including  all  the  requisite  calcula- 
tions, 237. 

Analysis  of  Flue  Gases  with  the  Orsat-Miiencke  Apparatus  237 

Advantages  of  this  apparatus  for  rapid  analyses,  238 ;  De- 
scription of  the  apparatus,  238,  239  ;  Method  of  filling  the  ab- 
sorbing tubes  with  the  different  solutions,  239,  240  ;  CompK>si- 
tion  of  chimney  gases  as  an  index  of  the  fuel  consumption 
under  the  boilers,  241  ;  Method  for  determination  of  excess  of 
air  in  furnace  gases,  241  ;  Table  of  ratio  of  carbon  dioxide  and 
air  in  furnace  gases,  241  ;  Amount  of  carbon  dioxide  as  indi- 
cating heating  efficiency,  241 ;  The  dasymeter  of  Messrs.  Sie- 
gert  and  Durr,  as  described  by  W.  C.  Unwin,  242  ;  Automatic 
indication  of  percentage  of  carbon  dioxide  in  the  flue  gases, 
by  the  dasymeter,  242  ;  Loss  of  heat  in  flue  gases,  as  deter- 
mined by  dasymeter  and  Siegert's  formula,  243  ;  Experiments 
on  Ten-Brink  furnaces,  to  determine  the  percentage  of  carbon 
dioxide  as  an  index  of  maximum  combustion,  244  ;  Uehling 
and  Steinbart*s  instruments  for  indicating  automatically  and 
continuously  the  percentages  of  carbon  dioxide  and  carbon 
monoxide  in  furnace  gases,  244. 

Analysis  of  Coal  Gas,  Water  Gas,  Producer  Gas,  Etc.,  by 

Means  of  the  Hempel  Apparatus 245 

Description  of  the  Hempel  apparatus,  245-246;  The  "Wink- 
ler" burette,  247 ;  Method  of  collecting  the  gas  for  analysis, 
248 ;  Example  of  an  analysis  of  a  gas  containing  carbon 
dioxide,  oxygen,  carbon  monoxide,  ethylene,  methane,  hydro- 
gen, and  nitrogen,  251-256  ;  Calculation  of  the  percentages  by 
volume  into  percentages  by  weight,  256 ;  Determination  of 
methane  by  explosion,  257. 

Heating  Value  of  Combustible  Gases 258 

Calculation  of  calories  per  kilo  to  B.  T.  U.  per  p>ouud,  258; 
Data  required,  258;  Method  of  H.  L.  Payne,  258;  Liter 
weights  of  the  gases,  hydrogen,  oxygen,  nitrogen,  air,  carbon 
monoxide,  carbon  dioxide,  methane,  and  ethylene,  258;  Table 
of  the  heating  p>ower  of  combustible  gases  expressed  in  calo- 
ries per  kilo,  B.  T.  U.  per  pound,  and  B.  T.  U.  per  cubic  foot, 
259 ;  Determination  of  heat  units  from  analysis  of  the  gas, 
260 ;  Illuminants,  value  of,  260  ;  Standard  temperature  for  gas 
measurements,  260,  261  ;  Si>ecific  heats  of  the  various  gases, 
261 ;  Volumetric  specific  heats,  261 ;  Calculation  of  heat  car- 
ried away  by  the  products  of  combustion  of  hydrogen  at  328'^ 
F.,  262  ;  of  carbon  monoxide,  263 ;  of  marsh  gas,  363  ;  Table 
of  B.  T.  TJ.  per  cubic  foot,  products  of  combustion  condensed, 
and  products  of  combustion  at  328°  F.,  of  hydrogen,  carbon 
monoxide,  marsh  gas,  and  illuminants,  263 ;  Heating  value  of 
natural  gas  in  B.  T.  U.  per  cubic  foot,  263 ;  Example,  with 
calculations,  of  the  heating  value  of  a  gas  composed  of  carbon 
monoxide,  carbon  dioxide,  illuminants,  hydrogen,  and  marsh 
gas,  stated  in  B.  T.  U.  per  cubic  foot  of  each  constituent,  prod- 
ucts of  combustion  condensed,  and  products  of  combustion 



escapinn  at  328°  F.,  264  ;  Determination  of  calories  per  kilo, 
or  B.  T.  U.  per  pound,  fnun  analysiBof  a  gas  staled  in  volume. 

[.Manufacture   of  Water   Gas   and   Calculation   of  Heating 

Power  of  Variou5  Illuminating  Gases 265 

Description  of  plant,  366,  167;  Operation,  367;  Composition 
of  uncarburetted  water  gas,  267:  Composition  of  c^arburetted 
water  gan,  168  ;  Calculation  of  the  heating  power  of  the  ancar- 
buretted  K^s  in  B.  T.  U.  per  cubic  foot,  from  an  analyses,  268; 
Of  the  carburetted  water  gas.  16S  ;  Analysis  of  a  sample  of 
London  coal  ^as.  36S  :  Calculation  of  its  lieating  power  in  B. 
T.  U.  per  cubic  foot,  products  of  combustion  condensed.  269 ; 
The  same,  products  of  combustion  in  a  state  of  vapor  at  328"  F., 
369^  Analysis  of  Heidelberg  gas,  189;  Konigsberg  gas  and 
Hannover  gas,  289;  Analysis  of  Wilkinson  carburett^  water 
gas.  with  determination  of  its  heating  power  in  B.  T.  U. 
from  the  analysis,  270;  Analysis  of  "Tessie  dn  Motay"  gas, 
with  calculation  of  B.   T.  U.  per  cubic  foot,  270. 

I' Producer  Gas 270 

Constituents,  170:  Analysis  of  Siemen'a  producer  gasi  with  B. 
T.  U.  per  cubic   foot,   270;    Analysis  of  anthracite  producer 

fas,  with  B,  T.  U.  per  cubic  foot,  270;  Analysis  of  soft  coal  pro- 
ucer  gas,  with  B.  T.  U.  per  cubic  foot,  270. 

liOil  Gas 371 

Method  of  manufacture.  271  ;  Keith's  oil  gas,  171  ;  "  Pintsch" 
oil  gas,  171  ;  "  Mineral  Seal"  oil,  271  ;  Compositioa  oF 
"  Pintsch"  oil  gas  as  determined  from  several  analyses,  by  the 
writer,  271  ;  Heating  power  per  cubic  foot,  calculated  from 
the  analysis,  271  ;  Tests  of  the  production  of  oil  gas,  by  the 
"  Keith  "  process,  and  the  "  Pintsch"  process,  by  W.  Ivison 
Macadam.  271.  172 ;  Oil  gas  compressed  to  atmospheres  in  iron 
cylinders  as  an  illuminant  for  cars,  271 ;  Loss  in  illuminating 
power  of  the  gas  hy  excessive  compression,  17a;  References 
upon  gas  analysis  and  oil  gas.  271. 

(Natural  Gas 272 

Composition  of  the  gas  not  uniform.  272;  Chemists  not  in 
agreement  as  to  constituents,  173  :  Analysis  of  Pennsylvania 
natural  gas,  by  Dr.  G.  Hay.  with  calculation  of  the  heating 
power  per  cubic  foot,  272  :  Analyses  of  six  samples  of  natural 
gas,  by  S.  A.  Ford,  373;  Analysis  of  New  Lisbon,  Ohio,  natural 
gas,  by  W.  A.  Noyes,  with  B.  T.  U.  per  cubic  foot.  27^  ;  Inves- 
tigations upon  the  composition  of  natural  gas  by  F.  C.  Phil- 
lips for  the  geological  survey  of  Pennsylvania,  374;  Analvsis 
of  Fredonia  natural  gas,  hy  F.  C.  Phillips,  274  ;  Also  of 'the 
Sheffield  natural  gas,  Wilcox  natural  gas,  and  the  Kane 
natural  gas.  274  ;  Test  of  the  fuel  value  of  natural  gas,  under 
boilers,  by  the  Westinghouse  Air-brake  Co,,  of  Pittsburg, 
Pa.,  274  ;  References  to  literature  on  natural  gas.  374. 

Practical  Photometrj' 275 

How  the  illuminating  value  of  a  gas  is  measured,  175  :  Stand- 
ard sperm  candles,  275  ;  Description  of  the  standard  Bunsen 


photometer,  275-278 ;  Manner  of  using  the  photometer,  279- 
281  ;  Use  of  formula  for  correction  for  pressure  and  tempera- 
ture, 282  ;  Table  to  facilitate  the  correction  of  the  volume  of 
gas  at  different  temperatures  and  under  different  atmospheric 
pressures,  283. 

Hartley's  Calorimeter  for  Combustible  Gases 284 

Description  of  the  apparatus,  284 ;  Method  of  use,  285 ;  Re- 
sults of  tests,  with  this  instrumentt  upon  the  municipal  gas 
of  New  York  City,  by  E.  G.  Love,  285  ;  Determination  of  the 
heating  power  of  the  London  coal  gas,  286  ;  Averacre  value  in 
terms  of  B.  T.  U.  per  cubic  foot,  of  the  water  gas  of  New  York 
City,  286;  Number  of  B.  T.  U.  for  |i.oo,  gas  costing  I1.25  *• 
per  cubic  foot. 

Junker's  Gas  Calorimeter 287 

Description  of  the  instrument,  287  ;  Method  of  operation,  288, 
289  ;  Table  of  resum^  of  tests  upon  London  coal  gas,  291 ;  Ex- 
periments made  at  the  Stevens  Institute  with  Junker  calorim- 
eter upK>n  Lowe  process  water  gas,  291  ;  Analysis  of  Lowe 
process  water  gas,  291  ;  Heating  value  from  calculation  of 
analysis  of  the  gas  equalled  662  B.  T.  U.  per  cubic  foot,  deter^ 
mination  by  calorimeter  668  B.  T.  U.  per  cubic  foot,  292. 

Liquid  Fuel 292 

Evaporative  power  of  petroleum  as  determined  by  Storer, 
292  ;  Heating  power  of  various  petroleums  as  determined  by 
Deville,  292 ;  Evaporative  power  of  liquid  hydrocarbons  as 
determined  by  Dr.  Paul,  292 ;  Table  of  results,  showing  the 
evai>orative  power  in  pounds  of  water  at  212°  F.,  of  CeH^O, 
CjHfiO,  C.oHs,  CuH,o,  CgHjo.  C,H,„  C.oH,,,  292  ;  Calculation  of 
effective  heat,  293  ;  The  determination  of  the  theoretical  evap- 
orative efficiency  of  different  combustibles,  as  given  by  Ran- 
kine,  294  ;  Table  of  evaporation  efficiency  due  to  carbon,  hy- 
drogen, etc.,  of  charcoal,  coke,  petroleum,  etc.,  294  ;  Formula 
representing  the  number  of  times  its  own  weight  of  water  a 
fuel  will  evaporate,  295 ;  Formula  for  the  loss  of  units  of 
evaporation  (Rankine),  295;  The  theoretical  evaporative  power 
of  hydrogen  and  carbon,  295  ;  Relative  heating  value  of  coal, 
gas,  and  petroleum,  as  determined  by  tests  made  by  the  En- 
gineer's club  of  Philadelphia,  296  ;  Tests  of  the  heating  value 
of  petroleum  and  block  coal  under  the  same  boilers,  at  Chi- 
cago, 111.,  296;  Relative  cost  of  oil  I1.93,  coal  $2.15  for  same 
evaporation  performed,  296. 

Valuation  of  Coal  for  the  Production  of  Gas 296 

Method  of  T.  Richardson,  296  ;  Description  of  the  apparatus, 
296,  297 ;  Determination  of  the  amount  of  coke,  tar,  ammo- 
niacal  water,  carbon  dioxide,  hydrogen  sulphide,  and  the  gas 
produced,  297 ;  Newbigging's  experimental  plant  for  the  de- 
termination of  the  gas-producing  qualities  of  coal,  297,  298; 
Method  of  using  the  apparatus,  298,  299  ;  Average  production 
of  gas  from  New  Castle  coal,  299 ;  Amount  of  gas  that  should 
be  produced  by  a  good  variety  of  gas  coal,  299. 

Analysis  of  Clay,  Kaolin,  Fire  Sand,  Building  Stones,  Etc.  299 
Constituents  to  be  determined,  299;  Determination  of  total 


silica.  199;  The  determination  of  conihincJ  silica,  hyrlralert 
silicic  acid,  and  of  quartz  aaod,  300;  Scheme  for  determina- 
tion of  aluntiaa,  ferric  oxide,  manganese  dioxide,  lime,  and 
magnesia,  301  ;  Determination  of  potash  and  soda,  sulpbnr 
trioxide,  and  titanic  oiide,  30a;  Water  of  hydration,  303; 
Composition  of  various  representative  clays,  303, 

■.Physical  Tests  of  Building  Stones ; 

I.  Crashing  stren^h,  how  determined,  304  :  The  RiehlfU.  S. 
standard  automatic  and  autographic  testing  machine,  304; 
Crushing  strength  of  granite,  trap'rock,  marble,  limestone, 
sandstone,  and  red  brick,  304. 

3.  Absorptive  power,  304;  Method  of  determinatioiJ,  304; 
Absorptive  power  of  granite,  marble,  limestone,  sandstone, 
brick,  and  mortar,  jot;  Freciing  test,  306;  The  Tagliabue 
freezing  apparatus,  306;  Preeiing  test  as  required  in  "  Uni- 
form methods  of  Procedure  in  Testing  Building  and  Struc- 
tural Materials"  by  J.  Bauschinger.  ( Mechanisch-lechniaehen 
Laboratorium,  Munchen),  307:  The  testing  of  bricks,  308:  De- 
termination of  soluble  salts  in  bricks,  309:  Examination  of 
unhurnt  clay  for  calcium  carbonate,  iron,  or  copper  pyrites, 
mica,  etc,  309;  Use  of  Papin's  digester  with  steam  at  one  and 
one-quarter  atmospheres.  310;  Microscopical  eiamination, 
310 :  Determination  of  the  character  and  structure  of  the 
stone,  310 ;  Difference  between  sandstones  and  quartzites,  310; 
Metho<l  of  H.  Lynwood  Garrison  tor  microscopical  examina- 
tion, 310;  References  to  literature  upon  testing  of  bulldiug 
atones,  etc.,  311. 

Hoys ■ 

Classification  of  alloys  into  three  classes,  311  ;  First  class 
comprise,  brass,  bronze,  bell  metal,  gun  metal.  Muntz's 
metal,  speculum  metal.  Delta  metal,  311  ;  Scheme  for  analysis 
of  alloys  of  first  class,  31  r  ;  Example  of  analysis  with  weights 
and  calculations,  31 1-313  ;  Alloys  of  the  second  class.  Babbitt 
metal,  Britannia  metal,  type  metal,  solder,  white  metal, 
camelia  metal,  Tobio  bronze,  ajax  metal,  cur-box  metal,  mag- 
nolia metal,  pewter,  "  Argentine,"  Ashbury  metal,  anti-fric- 
tion metal,  phosphor  bronxe.  deoxidized  bronze,  rose  metal, 
Parson's  white  metal,  "B"  alloy,  P.  R.  R.,  313;  Method  for 
analysis  of  Babbitt  metal,  313,  314:  Preparation  of  sodium  sul- 
phide solution.  314;  Mengin's  method  for  separation  of  tin 
and  antimonv  in  alloys,  314:  Scheme  for  analysis  of  white 
metal  containing  Sb,  Sn,  Pb,  Cu,  Hi,  Fe,  Al.  Zn.  315';  Volumetric 
determination  of  antimony  in  presence  of  tin,  315;  Table  of 
the  composition  of  allovs  of  the  second  class,  316;  Alloys  of 
the  third  class  comprising  alnminum  bronie,  ferro-aluminum, 
terro-tungsten.  German  silver,  roaine.  metalline,  aluminum 
bourhounz,  silicon  bronze,  Gutrie's  "entectic,"  arsenic  bronze, 
and  manganese  bronze,  316,  317;  Method  for  analysis  of  alu- 
minum bronze,  317 ;  Determination  of  manganese  in  manga- 
nese bronze,  317:  Method  for  the  analysis  of  ferro-aluminum, 
318,319;  The  determination  of  phosphorus  in  phosphor-bronze, 
319;  Qualitative  tests  for  lead,  copper,  tin,  and  antimony  in 
alloys,  3t9  ;  Thompson's  method  for  determination  of  co]>per, 
tin,  leaa,  and  antimony,  330-322;  References  on  analysis  of 
alloys,  323. 


Analysis  of  Tin  Plate 323 

Method  of  analysis  with  use  of  dry  chlorine  gas,  324 ;  Deter- 
mination in  tin  plate  of  tin,  lead,  iron,  and  manganese,  325  ; 
Table  of  analyses  of  nine  different  samples  of  tin  plate,  con- 
taining tin,  lead,  iron,  manganese,  carbon,  sulphur,  phos- 
phorus, silicon,  325 ;  Iodine  method  for  determination  of  tin, 

Chrdhie  Steel 326 

Classification  of  the  products  of  manufacture,  326;  Determi- 
nation of  chromium,  327 ;  Table  of  mechanical  tests  of  chrome 
steel,  including  limit  of  elasticity,  modulus  of  elasticity,  and 
breaking  strength,  etc.,  328 ;  Determination  of  manganese, 
329 ;  Silicon,  tungsten,  339 ;  Table  of  analyses  of  chrome 
steels  made  at  the  Stevens  Institute,  comprising  "  No.  i 
steel,"  *'No.  3  steel,**  '•  Magnet  steel,**  and  "Rock  Drill 
steel,**  330,  331. 

The  Chemical  and  Physical  Examination  of  Paper 331 

Determination  of  the  nature  of  the  fiber,  331  ;  Use  of  chemical 
solutions  to  detect  fibers  of  pine,  poplar,  and  spruce,  332  ;  De- 
termination of  the  amount  of  mechanical  fiber  in  a  mixture 
of  chemical  fiber,  linen  fiber,  cotton  fiber,  and  mechanical 
fiher,  332  ;  Action  on  wood  pulp  of  solution  of  gold  chloride, 
332 ;  Detailed  instruction  01  procedure  for  examination  of  a 
paper,  333 ;  Microscopical  examination,  334 ;  Description  of 
poplar  wood  fibers  under  the  microscope,  335  ;  Description  of 
spruce  wood  fibers  under  the  microscope,  336;  Description  of 
linen  fibers  under  the  microscope,  335  ;  Difference  in  appear- 
ance of  fibers  before  and  after  manufacture  into  paper,  337; 
Quantitative  determination  of  different  fibers  in  a  paper  by 
means  of  the  microscope,  337 ;  Official  German  directions  for 
the  detection  and  estimation  of  the  various  fibers  in  paper,  337; 
Color  reactions  of  the  different  fibers  with  solutions  of  iodine, 
337»  338»  Determination  of  the  free  acids  in  pai>er,  338;  Process 
for  the  determination  of  chlorides,  338:  for  sulphates,  339; 
Use  of  aluminum  sulphate  instead  of  alum  in  paper,  339;  De- 
termination of  the  nature  and  amount  of  sizing  used,  339,  340; 
Schumann*s  method  for  rosin,  340 ;  Determination  of  the 
amount  of  starch,  340,  341  ;  Tollen*s  formula  for  Fehling's 
solution,  341  ;  Determination  of  the  ash  of  paper,  341-343 ; 
Detection  of  Venetian  red,  Prussian  blue,  ochre,  agalite,  and 
clay,  342  ;  Percentage  of  ash  in  commercial  pulps,  342 ;  Ash 
in  the  various  fibers,  344 ;  Determination  of  the  weight  per 
square  meter,  344 ;  of  the  thickness,  344 ;  of  the  breaking 
stren^h,  345,  346 ;  Description  of  the  Wendler  paper  testing 
machine,  347  ;  Method  of  using  the  instrument,  347,  348 ;  The 
Schopper  apparatus,  348 ;  References  to  literature  upon  paper- 
making  ana  paper-testing,  348,  349. 

Soap  Analysis 349 

Classification  of  soaps  into  toilet  soaps,  laundry  soaps,  com- 
mercial soaps,  and  medicated  soaps,  349 ;  List  of  adulterants 
used  in  soaps,  349  ;  Manufacture  of  the  common  yellow  soap, 
349 ;  Use  of  recovered  grease,  349 ;  List  of  oils  used  in  the 
manufacture  of  soaps,  349  ;  Scheme  for  the  analysis  of  soap, 



350 ;  Scheme  for  the  • 
■oap,  351;  Determinat 
waxes,  isi  :  DeterniiDHt 
Caprjlie  anhydride,  353 ; 

ilysis  of  unaaponifiable  i 

1  of  water.  352  ;    Determination  of 

□  of  total  alkali  and  fatt^  acids,  353  ; 

. ;   Determination  of  glycerine,  354:  of 

silicates  of  the  alkalies,  354;  Factor  lo  convert  weight  of  fatty 
hydrates  to  anhydrides,  354;  Free  alkali,  353;  Determination 
of  resin,  355;  Gottlieb's  method,  355  ;  Hiibl's  method  for  resin 
in  soap,  356  ;  Twitchell's  method  for  determination  of  resin, 
in  (atty  acids,  356  ;  Table  for  the  physical  and  chemical  inves- 
tigations of  fats  and  fatty  acids,  35S:  Determination  of  Klyc 
enne  by  titration.  359:  Table  of  analyses  of  various  kinds  of 
soaps,  361 ;  Washing  powders.  361 ;   References,  361. 

1  Technical  Examination  of  Petroleum 362 

Division  into  three  classes  by  fractional  distillation,  361 ;  The 
method  of  Engler,  36] ;  Variations  in  methods  used  by 
chemists,  36a;  Composition  of  cmde  petroleum  as  deter- 
mined by  fractional  distillation,  363;  Composition  of  two  ftam- 
ples  of  crude  Mexican  petroleum  as  determined  by  the 
writer,  364;  Technical  divisions  of  the  dislillaleB  of  petro- 
leum,  364;  First  class,  cymogene.  rhigolene.  petroleum  ether, 
gasolene,  naphtha,  Hgroiu,  bcnzene^econd  class,  the  vari- 
ous varieties  of  kerosene — and  third  class,  residium,  boiling- 
point  300°  C.  and  above,  364.  365  ;  Average  percentage  compo- 
sition of  the  products  obtained  from  petroleum,  365  ;  Classifi- 
cation of  the  products  from  petroleum  by  the  oil  trade,  365  ; 
Composition  of  valve  oils,  car  oils,  engine  oils,  spindle  oils, 
dynamo  oils,  loom  oils.  36J  ;  Formula  of  engine  oil  as  used  by 
the  Pennsylvania  Railroad,  365  ;  Formula  for  cylinder  oils. 

J  The  Examination  of  Lubricating  Oils 366 

The  generally  accepted  conditions  of  a  good  lubricant,  366 ; 
Determination  of  the  nature  of  the  oil  by  saponification,  etc., 
367;  Description  of  the  process  of  saponification,  367;  De- 
termination  of  fatty  acids  m  vegetable  and  animal  oils,  368; 
Method  of  determining  the  melting-point  of  fatty  acids,  369; 
Table  of  melting-points  and  congealing  points  of  tatty  acids 
of  the  various  animal  and  vegetable  oils  used  in  lubrication, 
370;  Specific  gravity,  371  ;  Baum^  hydrometers,  371  ;  Taglia- 
bue'a  hydrometer,  371  ;  Table  for  converting  Banm^  degrees, 
liquids  lighter  than  water,  into  specific  gravities,  371 ;  Table 
of  Baumf  degrees  with  correction  for  temperature.  371,  373; 
Formula  for  quantitative  determination  of  two  oils  in  a  mix- 
ture, from  the  gravity,  374;  Graphical  method,  375;  The 
Wesiphal  balance,  376 ;  The  Araeo-picnometer.  376 ;  Table  tor 
conversion  of  various  hj^dromeier  degrees  into  specific  gravi- 
ties, 377  L  Table  of  specific  gravity  of  oils,  377 ;  The  coldtest, 
377-379  ;  Description  of  cold  test  apparatus  for  oils,  as  used 
by  chemists  of  Chicago,  Burlington,  and  Quincy  Railroad,  379, 
380  :  Table  giving  the  cold  lest  of  the  principal  oils,  3S0,  381 ; 
Specifications  for  oils,  with  requirements  of  cold  lest  stated, 
381;  Tagiiabuc's  standard  freezing  apparatus,  381;  Viscosity 
of  oils.  383 ;  Pennsylvania  Railroad  viscosity  tests,  383 ; 
Engler's  viscosimeier,  384,  385  ;  Redwood's  viscosimeter,  38s ; 
The  septometer  of  Mr.   Lepenan,   386;    Davidson's  "'—"—"- 


ter,  387 ;  Tagliabae*8  viscosi meter,  389 ;  Gibb*9  viscosimeter, 
390,  391  ;  Table  of  viscosities  of  valve  oils  and  stocks,  392 ; 
Viscosities  of  car  and  engine  oils,  392  ;  Perkin's  viscosimeter, 
393;  Stillman's  viscosimeter,  394-394S ;  Table  of  viscosities  of 
fortv-three  of  the  principle  oils  used  in  lubrication,  at  68°  F., 
122^  F.,  212°  F.,  302°  F.,  392°  F.,  397;  Chart  of  above  tests, 
398;  Conclusions  deduced  from  viscosity  determinations,  3^; 
The  Doolittle  viscosimeter,  400,  401  ;  Iodine  absorption  of  oils, 
401 ;  of  fatty  acids,  402 ;  Table  of  determinations  of  iodine  ab- 
sorption of  various  oils,  403 ;  Flash  and  fire  test  of  oils,  403  ; 
The  *'  Cleveland  Cup**  oil  tester,  404 ;  Tagliabue*s  open  tester, 
405 ;  The  Saybolt  electric  oil  tester,  405 ;  The  Abel  closed 
tester,  405,  406 ;  The  Pensky-Marten's  closed  tester, 
407 ;  Trauniann's  open  tester,  408 ;  Requirements  for  the  flash 
and  fire  test,  408;  Acidity  of  oils,  408,  409;  Method  for  de- 
termining the  acidity  of  oils  as  performed  in  railroad  labora- 
tories, 409,  410 ;  Maumene*s  test  for  oils,  410,  411  ;  Table  giv- 
ing the  rise  of  temperature  of  oils,  by  Maumene's  test,  412  ; 
Color  reactions  of  oils  with  nitric  and  sulphuric  acids,  412,  413 ; 
Heidenreich*s  test,  413 ;  Massie*s  test,  413 ;  Table  of  reac- 
tions of  various  oils  with  nitric  and  sulphuric  acids,  414 ; 
Classification  of  oils  used  in  lubrication  into  two  classes, 
saponifiable  and  unsaponifiable,  414 ;  Detection  of  fatty  oils 
in  mineral  oils  by  method  of  l/ux,  414 ;  Detection  of  rosin  oil 
by  the  method  of  Holde  or  Valenta,  414 ;  Scheme  for  the 
analysis  of  a  lubricating  oil  containing  mineral  oil,  lard  oil, 
and  cotton-seed  oil,  415;  Method  of  Salkowski  for  the  de- 
termination of  the  amounts  of  animal  and  vegetable  oils  when 
mixed  together,  415,  416;  Wool  grease,  416;  Degras  or  sod 
oil,  416;  Bone  fat,  416;  Coefficient  of  friction,  417;  Descrip- 
tion of  the  Thurston,  and  the  Henderson-Westhoven  friction 
machines,  417-419  ;  Description  of  the  friction  apparatus  used 
by  the  officials  of  the  Paris-Lyon  Railway,  419-421 ;  Descrip- 
tion of  the  Riehl^  lubricant  tester,  as  used  in  many  of  the 
railroad  laboratories  in  the  United  States,  422  ;  Record  blank 
used  by  engineers  on  Baltimore  and  Ohio  Railroad  for  testing 
oils  upK>n  locomotives,  423  ;  Detailed  specifications  for  engine 
and  passenger  car  oils,  cylinder,  and  freight  car  oils,  Balti- 
more and  Ohio  Railroad,  423,  424;  Specifications  for  black 
engine  oils  and  cylinder  stock,  Chicago,  Burlington,  and 
Quincy  Railroad,  425,  426;  References  to  literature  of  lubri- 
cation, 426. 

Oils  Used  for  Illumination 426 

Classification  of  illuminating  oils  into  two  groups:  a,  refined 
products  of  petroleum ;  b,  certain  refined  oils  of  animal  and 
vegetable  origin,  426;  Kerosene,  426;  Headlight  oil,  426; 
ST>ecifications  for  petroleum  burning  oils  for  railroad  use,  427; 
150^  fire  test  oil,  427;  300°  fire  test  oil,  427;  Method  of  ma- 
king tests  on  150^^  oil  and  300°  oil,  427, 428 ;  The  cloud  test,  428  ; 
The  "  Wisconsin**  tester  for  the  flash  and  fire  points  of  illu- 
minating oils,  429.  430;  Rules  and  regulations  for  making 
the  tests,  430 ;  Law  regulating  the  standard  of  illuminating 
oils  and  fluids,  state  of  New  York,  430-432  ;  The  grades  of 
colors  in  classifying  kerosenes,  432 ;  The  Stammer  colorimeter 
for  oils,  432  ;  The  Wilson  colorimeter,  433 ;  Colza  and  lard  oil 
for  illumination,  433  ;  Different  methods  of  car  illumination, 
434 ;  Pintsch  oil  gas,  method  of  manufacture  and  use  for  car 


iltamination,  454;  The  Foster  system,  435  ;  Tht  Frost  system, 
435 :  The  electric  system  of  car  lighting.  436-438 ;  Results  of 
experiments  made  upon  different  railroads,  43S;  Relative  ad- 
vantages aud  disadvantages  of  the  Tarioua  systems.  43S,  439; 
Table  showing  the  comparative  cost  o(  car  lighting  systems, 

he  Analysis  of  Lubricating  Oils  Containing  Blown  Rape- 
seed  and  Blown  Cotton-seed  Oils 441 

Rape-seed  oil  as  the  standard  lubricant  in  Europe,  441  ;  Pro- 
portion of  rape-seed  oil  added  to  mineral  oils,  441  ;  Method  of 
duplicating  an  oil  from  the  analysis,  443.  443 ;  Comparison  of 
the  chemical  reactions  of  blown  rape-seed  and  normal  rape- 
seed  oil,  444 ;  Recognition  in  a  mixture  of  the  amounts  of  cot- 
ton-seed and  rape-seed  oil  from  the  difference  in  the  melting- 
point  of  the  fatty  acids,  445 ;  Synthetical  work,  445. 

lie  Analysis  of  Cylinder  Deposits 445 

Classification  of  deposits,  445  :  Composition  of  deposit  taken 
from  a  locomotive  cylinder,  446;  Composition  of  a  deposit 
containing  scale-forming  matter  carried  over  by  the  steam, 
446  :  Corrosive  action  of  fatty  acids  on  iron,  copper,  brass,  etc., 
449:  Action  of  castor  oil  as  a  lubricant,  447;  Method  of  pro- 
cedure in  analysis  of  cylinder  deposits,  449  ;  Scheme  for  the 
analysis,  450:  Composition  of  a  deposit  formed  from  mica 
grease,  453  ;  References.  452. 

^sint  Analysis 452 

What  should  constitute  a  paint,  45I ;  Qualities  essential  in  a 
paint.  4S3;  List  of  red  pigments  with  their  chemical  for- 
mula, 453 ;  brown  pigments,  453 ;  white,  yellow,  and  orange, 
433  <  green,  black,  and  blue  pigments.  454;  Scheme  for  tbe 
aualyeis  of  white  paint  ground  in  oil,  455 ;  Analysis  of  several 
representative  paints,  456;  Scheme  For  the  analysis  of  lemon 
chrome  paint,  457;  Determination  of  water,  volatile  matter, 
and  water  extract  in  chrome  paints,  458  :  Scheme  for  the 
analysis  of  chrome  green,  459;  Spec  i  Beat  ions  for  cabin  car 
color,  Peiinsylvonia  Railroad,  460;  Use  of  gypsum  and  cal- 
cium carbonate  in  red  paints,  46a;  Specifications  for  freight 
car  color,  461  ;  Competition  of  paints  used  for  iron  work. 
Elevated  Railroad,  New  York  City,  i6i,  463;  Asphalt  paint. 
463;  Fire-proof  paints,  silicate  paints,  asbestos  paints,  etc., 
463  ;  Composition  of  the  fire-proof  paint  used  by  the  munici- 
pality of  Paris,  494  ;  Composition  of  ultramarine,  commercial 
Prussian  blue,  and  smalts,  464  ;  Examination  of  the  oil  after 
extraction  from  the  paint,  461;;  Detection  of  turpentine  in 
presence  of  rosin  spirit,  465  ;  Petroleum,  naphtha,  and  turpen- 
■'-■;,  465;  References  on  the  literature  of  paints,  465. 

rometry 466 

Practical  useof  pyrometers,  466 ;  Classification  of  pyrometers, 
466 :  Principles  upon  which  their  operation  depends.  466.  467 ; 
Air  thermometers,  467 ;  Air  pyrometer  of  Siegert  and  Duerr, 
467,468;  Wiborgh's  air  pyrometer,  468;  Hoi&sou's  hot-blast 
pyrometer,  468;  Bristol's  recording  thermometer  for  tempera- 
s  up  to  600°  p.,  469;    Brown's  metallic  pyrometer,  469; 


The  copper-ball  or  platinum -ball  pyrometer,  469  ;  The  Wein- 
hold  pyrometer,  470,  471 ;  The  Saintignon  pyrometer,  472 ; 
Braun's  electric  pyrometer,  473 ;  LeChatelier's  thermo-elec- 
tric pyrometer,  473,  474  ;  Uehling*8  and  Steinbart*s  pyrometer 
for  blast  furnaces,  475-478  ;  List  of  boiling  and  melting-points 
of  metals  as  determined  with  pyrometers,  479 ;  References  to 
the  literature  of  pyrometry,  479. 

The  Electrical  Units 480 

The  electrostatic  and  the  electromagnetic  systems,  480 ;  The 
C.  G.  S.  units,  480;  Unit  magnetic  pole,  480;  Unit  current, 
480;  Practical  units,  480;  Ampere,  480;  The  ohm,  the  volt, 
the  coulomb,  the  Farad,  the  Joule,  the  Watt,  the  Henry,  481 ; 
Kilo-WattSf  482 ;  Relations  between  the  international  units  of 
resistance  and  electromotive  force  to  those  of  the  older  units, 
482 ;  Ohm's  law,  482  ;  Joule's  law,  482;  Measurement  of  elec- 
tric energy,  482;  The  Watt-meter,  483;  Electro-chemical 
equivalents,  483. 

Energy  Equivalents 483 

IVork — in  foot  pounds,  per  second,  per  minute,  per  hour,  483; 
in  B.  T.  U.  per  second,  minute,  hour,  484;  in  pounds  of 
steam,  in  combustion,  in  electricity  and  light,  484 ;  in  rotary 
delivery,  484. 

Heai — B.  T.  U.  to  work,  light  and  electricity,  485 ;  steam  to 
work,  light,  and  electricity,  485 ;  one  pK>und  of  carbon  con- 
sumed in  one  hour,  in  terms  of  combustion,  fuels  to  B.  T.  U., 
steam  work,  486;  one  pound  of  kerosene  consumed  per  hour 
in  terms  of  light  and  electricity,.  486  ;  one  cubic  foot  illumi- 
nating gas  in  terms  of,  487. 

Light— 'One^  candle  power,  in  terms  of  light  to  work,  B.  T.  U., 
electricity,  steam  and  combustibles,  487. 

Electricity— One^  Watt,  in  terms  of  work,  (H.  P.),  B.  T.  U., 
steam,  light,  and  combustibles,  487. 

Tables 488-505 

Index • 506 

List  of  Illustrations. 




Pigrure  t.    Electrolytic  apparatua  for  the  determination  of  copper 6 

*'       2.    Gfilcher'fl  thermo-electric  pile 7 

"       3.    Bunsen  valve la 

4.  Apparatus  for  determination  of  CO,  in  limestone 17 

5.  Lychenheim's  apparatus  for  determination  of  phosphorus  in  coal  and 
coke 22 

**       6.    Thdmer  coke  testing  machine 24 

**       7.    Bunsen  valve 29 

**       8.    Apparatus  for  determination  of  water  of  hydration  in  iron  ores 32 

**       9.    Jenkin's  scale  for  calculation  of  blast  furnace  charges 55 

**      la    Bettendorf's  automatic  water-bath 6z 

11.  Apparatus  for  determination  of  ammonia  in  water 75 

12.  Wolff's  colorimeter 76 

13.  Apparatus  used  by  New  York  City  Health  Board  for  determination  of 
ammonia  in  water 79 

"      14.    Pilter-beds,  water  supply  of  Dublin 86 

"      15-18.    The  Warren  filter 87-92 

"      19,20.    The  Goubert  feed-water  heater 100 

21.    The  Hoppes  feed-water  purifier  and  heater loi 

22,23.    The  Derveaux  water  purifier 106 

24-26.    The  Archbutt  and  Deely  apparatus  for  purification  of  boiler  waters  108 

**      27.    Filter  press iii 

'*      28.    Applicationof  filter  press  for  filtration  of  boiler  waters 112 

**      29.    Apparatus  combining  chemical  precipitation,  feed-water  heater  and  fil- 
ter press  for  purification  of  boiler  waters 1x3 

'*      30,  31-    Apparatus  for  determination  of  carbon  and  hydrogen  in  coal xi6 

32.  Apparatus  for  determination  of  nitrogen  in  coal 117 

33.  Shell  and  connections  of  the  Mahler  calorimeter 126 

34.  35.    The  Mahler  calorimeter 127,  128 

36.  The  Thompson  calorimeter 132 

37.  The  Barms  coal  calorimeter 136 

38-40.    The  Carpenter  coal  calorimeter 139, 141 

41.  Kent's  apparatus  for  determining  the  heating  values  of  fuels 142 

42.  Apparatus  for  determination  of  sulphur  in  iron 151 

43.  Apparatus  for  determination  of  sulphur  in  iron 153 

**      44.    Apparatus  for  determination  of  carbon  in  iron  and  steel ;  chromic  acid 

process 161 

**      45.    Apparatus  for  determination  of  carbon  in  steel  and  iron  ;  oxygen  com- 
bustion process 164 

46.  Wiborg's  apparatus  for  determination  of  carbon  in  iron  and  steel 165 

47.  Bggertz' apparatus  for  determination  of  carbon  insteel 168 

48.  Spiegelberg's  agitation  apparatus  for  phosphoric  acid  determinations. . .  179 
49, 50.    Agitation  apparatus  for  determination  of  phosphoric  acid,  as  used  by 

chemists  of  the  Pennsylvania  Railroad 181 

SI.    Picnometer   191 

'*      52.    Pairbank's  cement  testing  machine 208 

53.    Riehl6's  cement  testing  machine 209 






•  < 






•  I 


Vig,    54.    Briquette  mold 210 

55.    The  MichaelU  ccmeat  tctJog  macfrine xit 

"      56.    The  Paija  testing  machine ati 

'*      S7>  5^*    The  Reid  and  Bailey  testing  marhiwe axa»  213 

59.  Canre  of  breaking  streni^s  of  cememts  (Paijaj 215 

60.  The  B6hme-Hammer  apparatna 216 

6f ,  6a.    Jameson's  briqnette  making  machine- 217 

6$.    Prench  modification  of  the  If  ichaelis  cement  testing  machsoc 2x9 

64.  The  Baignet  cement  testing  machine 220 

65.  The  Sachier  com|>ression  machine 223 

66.  The  B6hme  compression  machine 294 

67.  The  Porter  aotomatic  cement  testing  machine 2a6 

68.  The  Elliott  apparatus  for  analysis  of  chimney  gases,  etc 233 

69.  The  Orsat-MSencke  apparatus  for  analysis  of  fine  gases ^38 

70.  The  dasymeter  of  Siegert  and  Dnerr 242 

71,7a.    Charts  showing  heat  losses  in  boiler  practice 243 

73-®3-    The  Hempcl  gas  apparatus a46-aso,  asa-aS4 

84.  The  Humphrey  water  gas  plant 367 

85.  The  Bunsen  photometer a76 

86.  The  Hartley  calorimeter  for  combustible  gases 384 

87,88.    The  Junker  calorimeter ^T-ago 

89.  Newbigging's  experimental  plant  for  the  determination  of  the  gas-pro- 
ducing qualities  of  coal 297 

90.  The  Kiehl^  United  States  standard  automatic  and  autographic  testing 
machine 305 

91.  TheTagliabue  freezing  apparatus 306 

9a-^.    Microphotographs of  various  fibers 335*336 

100.    Apparatus  for  determination  of  the  thickness  of  paper 345 

loi.    The  Wendler  paper  testing  machine 346 

loa.    The  Westphal  balance  with  Reimann's  plummet 357 

103.  Fractional  distillation  flask  for  petroleum 363 

104.  Beparatory  funnel  for  separation  of  oils 367 

105-108.    Apparatus  for  determination  of  melting-points  of  fatty  acids 369,  370 

109.    Tagliabue's  hydrometer  for  oils 371 

no.    Graphic  method  of  determining  percentages  of  oils  in  mixtures  of  oils..  375 

III.    The  Westphal  balance 375 

I  la.    The  Westphal  balance  modified  for  high  temperatures 376 

113.    Kichhom's  araeo-picnometer 377 

114. 115.    Cold  test  apparatus  foroils 378t  379 

116.  Bectional  view,  Tagliabue's  freezing  apparatus 38a 

117.  Hchubler's  viscosimeter  for  oils 383 

118.  Kngler's  vitcosimeter  for  oils 384 

119.  Redwood's  viscosimeter  for  oils 385 

lao,  tai.    I<epenau's  septometer  for  oils 386 

laa.    Davidnon's  vincosimeter  for  oils 388 

1x3.    Tagliabue'fi  viscosimeter  for  oils 389 

lU*    Oibb'M  vincosimeter  for  oils 390 

tiy    Chart  of  curves  showing  viscosity  of  oils  as  determined  by  the  Gibb's 

viscosimeter 393 

ti6.    Htillman'fi  viscosimeter  for  oils 395 

1*7.    Chart  of  curves  showing  viscosity  of  oils  as  determined  by  the  Stillman 

vincosimeter 398 

laft.    Doolittle's  vincosimeter  foroils 400 

ity.    Apparatus  for  determination  of  the  **  flash  "  and  "fire  "  test  of  lubrica- 
ting oils 404 


Vig.  130.    Ta^liabue' a  open  tester 404 

131.    The  Saybolt  tester 405 

X33.  133-    The  Abel  closed  tester 406 

134.  135'    The  Pensky-Martens  closed  tester 407 

X36.  137-    The  Treumann  open  tester 408 

Z38«  139*    The  Henderson-Westhoven  friction  tester  for  lubricants 417 

140,  X41.    Apparatus  used  by  the  Paris-Lyon  Railway  for  testing  lubricants. 419, 420 

142.  The  Riehl6  machine  for  friction  tests  of  lubricants 421 

143.  The  Wisconsin  tester  for  illuminating  oils 429 

144.  The  Stammer  colorimeter 432 

145.  The  Sozhlet  apparatus 448 

146.  The  air  pyrometer  of  Siegert  and  Duerr 467 

147.  The  Hobson  hot-blast  pyrometer 268 

148.  The  Weinhold  pyrometer 470 

149.  The  Saintignon  pyrometer 472 

i5o>  151*    Prof .  Braun's  electric  pyrometer 473.474 

IS2-IS4-    Uehling  and  Steinbart's  pyrometer 475-477 




Page  80,  line  14,  for  '*  NO,  *'  read  "  NO,.'* 

Page  81,  line  19,  cidd  **  with  dilute  acetic  acid." 

Page  82,  line  22,  for  '*  1000  parts  of  salt  '*  read  '*  1000  parts  of  water.'* 

Page  119,  line  16,  for  "hydroscopic"  read  *'  hygroscopic." 

Page  125,  line  10,  for  "hydroscopic"  read  "hygroscopic." 

Page  170,  line  22,  for  "  L^eduber  "  read  "  Ledeber." 

Page  187,  line  12,  for  "  fidelity  "  read  "fluidity." 

Page  256,  line  30,  for  "  for  carbon  dioxide  24.2  per  cent."  read  "  carbon 
monoxide  24.2  per  cent." 

Page  259,  line  2J,  for  "CjH^  =  11900  calories"  read  "C,H4=ii9ii 

Page  263,  line  3,  for  "  (2.39  cubic  foot  of  air)  "  read  "  (2.39  cubic  feet 
of  air)." 

Page  270,  line  21.  for  "  827.62  B.  T.  U."  read  "  754.6  B.  T.  U." 

Page  271,  line  21.  for  "  1582.  B.  T.  U."  read  "  1391.  B.  T.  U." 

Page  273,  line  36,  for  "  1000.52  B.  T.  U."  read  "  1115.  B.  T.  U." 

Page  288,  line  25,  for  "  pressed  "  read  "  passed." 

Page  314,  line  29,  for  "  hydrogen  oxide  gas  "  read  "  hydrogen  sulphide 

Page  373,  line  3,  for  "24°  Baum^  at  60°  F."  read  "24.7*^  Baum^  at 
(xP  F." 

Page  433,  line  19,  for  "  Wilson's  calorimeter"  read  "  Wilson's  colorim- 




Determination  of  Iron  in  Iron  Wire. 

Weigh  two  samples  of  briglit  iron  wire  (eacli  sample  0.500 
gram)  ;  transfer  to  beakers  (No.  3).  add  twenty-five  cc.  hydro- 
chloric acid,  five  cc.  nitric  acid,  cover  the  beakers  with  watch- 
glasses,  and  warm  gently  until  solution  is  complete. 

Proceed  with  each  sample  as  follows  :  Add  100  cc.  water,  then 
ammonium  hydroxide  ^nK^KK/Zv  until  the  solution  is  faintly  alka- 
line; boil,  filter  upon  a  No.  4  ashless  filter,'  and  wash  precipi- 
tate with  hot  water  until  the  washings  no  longer  react  alkaline. 
Dry  at  105"  C. 

Remove  as  much  of  the  dry  precipitate  as  possible  from  the 
filter  paper  to  a  piece  of  glazed  paper  and  ignite  the  filter  paper 
in  a  weighed  porcelain  crucible  (Meissen  No.  6),  uncovered, 
until  all  carbonaceous  matter  is  consumed.  Add  the  precipitate 
from  the  glazed  paper,  cover  the  crucible,  and  ignite  at  a  red 
heat  for  ten  minutes,  cool  in  a  desiccator,  and  weigh.     Heat  the 

iicible  and  contents  once  more  to  a  red  heat  for  three  minutes, 

K>1  as  before,  and  weigh.     Repeat  until  weight  is  constant. 

Aninunt  of  iron  wire  taken  =0.500  gram. 

Crucible  +  Fe^O, 9.432  grams. 

Crucible 8,711 

I-'CiO,    o.7[lgraiii. 


Fe,0,  :  Fe,  :  :  0.711  :  .r. 
X  =  0.4977  weiglil  of  Fe. 

0.4977  X  "»  =  gg.5^  per  ceul.  Fe  in  the  wire. 

IptrynrrJ.— Freseuius'  "Quantitative  Chemical  Analysis"  (l^nilon 
;<Iitiou).  g  70,1.  I,  a.  ;  "Hints  to  Beginners  in  Iron  Analysis,"  by 
David  H.  Browne, y.  jIho/.  CAciii..  5.  JI5. 



Alumina  in  Potash  Alum. 
( K,SO,  +  Al,(SO,),  +  24H,0). 

Press  finel3'  triturated  potash  alum  between  sheets  of  filter 
paper.  Weigh  out  duplicate  samples,  each  of  two  grams ; 
transfer  to  No.  4  beakers,  and  dissolve  in  about  150CC.  of  water. 

Add  ammonium  hydroxide  in  slight  excess,  fifteen  cc.  solu- 
tion of  ammonium  chloride,  and  boil  gently  a  few  minutes,  the 
liquid  remaining  alkaline.  Allow  the  precipitates  to  settle,  then 
decant  the  clear  supernatant  liquid  upon  No.  4  ashless  filters. 
Pour  boiling  water  upon  the  precipitates  in  the  beakers,  allow 
precipitates  to  settle,  decant  the  liquid  as  before,  and  repeat 
the  operation  three  times,  finally  transferring  all  of  the  precipi- 
tates to  the  filter  papers,  and  washing  with  hot  water  until  the 
reaction  is  no  longer  alkaline.  Dry  at  los**  C,  transfer  to 
weighed  porcelain  crucibles,  and  ignite  as  directed  for  ignition 
of  ferric  hydroxide  (I). 

Example  : 

Amount  of  alum  taken,  2.384  grams. 

Crucible  -f  AljO, 17-5  ^3  grams. 

Crucible 17.258      *• 

AljOj 0.255  gram. 

0.255  X  100  ,  ^     K\  ry 


Theoretical  Percentaf^e  : 

K,,SO,  -f  AljCSO^),-!-  24H,0  :  AljO, :  :  100:  ;i'. 
.r=  10.85  per  cent.  Al^O^. 


Copper  in  Copper  Sulphate. 
(CuS0,  +  5H,0). 

About  five  grams  of  the  crystallized  salt  are  pulverized, 
pressed  between  folds  of  filter  paper,  and  transferred  to  a  small 
stoppered  weighing  tube,  and  the  latter  and  contents  accurately 


Pour  out  about  one  gram  of  the  salt  into  a  No.  3  beaker,  and 
reweigh  the  tube.  The  difference  between  the  two  weights 
gives  the  weight  of  the  salt  taken. 

The  salt  is  dissolved  in  about  100  cc.  of  hot  water,  and,  if 
the  solution  is  not  clear,  add  a  few  drops  of  dilute  sulphuric  acid. 
Warm  gently,  and  add  gradually  a  clear  solution  of  sodium 
hydroxide,  with  constant  stirring,  until  the  reaction  of  the  cop- 
per solution  is  alkaline;  boil:  the  copper  is  precipitated  as  dark 
brown  cupric  oxide.     Thus  : 

CuSO.  +  2(NaOH)  =  CuO  +  Na,SO.  +  H,0. 
The  precipitate  is  allowed  to  settle,  when,  if  sufficient  sodium 
hydroxide  has  been  added,  the  supernatant  liquid  will  be  color- 
less. Filter  by  decautation  upon  a  No.  4  ashless  filter,  wash 
with  hot  water  until  reaction  of  washings  is  no  longer  alkaline, 
and  Aiy  at  105°  C. 

Remove  the  precipitate  (as  much  as  possible)  from  the  filter- 
paper,  and  place  it  upon  a  piece  of  glazed  paper. 

The  filter-paper  (which  will  contain  some  cupric  oxide)  is 
transferred  to  a  weighed  porcelain  crucible  (No.  6  Meissen),  and 

A  portion  of  the  cupric  oxide  is  reduced  to  copper  by  the  in- 
candescent carbon  of  the  filter-paper.     Allow  to  cool,  add  two 
or  three  drops  of  nitric  acid,  warm  gently  to  dissolve  the  copper, 
and,  when  solution  is  complete,  evaporate  to  dryness,  and  heat 
to  redness,  converting  all  the  copper  nitrate  to  cupric   oxide. 
Add  the  rest  of  the  cupric  oxide  remaining  upon  the  glazed 
paper  to  the  crucible,  and  heat,  at  red  heat,  to  constant  weight. 
First  weight  of  weighing  tube  and  CuSO,-}- jHjO.       7.0350  grams. 
Secoad  weight   of  weighing   tube   and  CuSO,  + 

5H,0 5.9605      " 

Copper  sulphate  taken ' t.0645      " 

Crucible -|- CuO '5.3744      " 

Crucible 15.0360     " 

0.3384  gram. 
CuO  ;  Cu  ;  ;  wt.  of  CuO  :  x  (=wt.  Cu) 
79-5  :  63-5  :  :  0.3384  :  a" 

■  Air  -  - 


^*  ■  »       .  ^M  ■ 

1  «  . 

Volumetric  Determination  of  Copper  by  Potassium 

Cyanide  Solution. 

r>i%s//lv*-  ^^rn  ^ran.i  ^f  rx>  cvan:*Ie  in  2-0  cc.  of  water 
;in<J  thorMi'^hly  mix. 

'\A/''ijjh  '>iit  t'A'O  ;(rams  ^f  pure  <-.or»p€:r  wire,  transfer  to  a  one- 
fourth  lit^rr  fla*iW.  a''i<l  t-.venty-tive  cc.  nitric  acid,  warm  gently 
until  th^*  ^oj^j^er  is  ?ill  'Jissr>Ivefl  :  rioil  to  expel  oxides  of  nitro- 
'^/■n  .  fti^A,  Hilnte  with  water  to  the  mark,  mix  well.  Take 
fiff/  ff  of  r)»is  f  opp^jr  V/Iution.  transfer  to  a  No.  3  beaker,  add 
arnrnoTiiiini  hydroxide  until  the  precipitate  formed  dissolves  and 
th^'  <^t]\\t'uni  in  alk;iline. 

Till  ;i  fifty  <  f .  Iiiirt-lto  with  potassium  cyanide  solution,  and 
yjut\u;i]]y  rlrop  the  cyanide  sf^lution  into  the  copper  solution 
fintil  flif  liliio  f  olor  disappears  and   the  solution  becomes  color- 

N'/»f''  flM-  tiiftnbcr  of  cc.  of  jiolassium  cyanide  solution  required 
to  r|o  tlti-;.  iiUf]  mark  npf)ii  the  potassium  cyanide  bottle  the 
vhIim-  of  tiw  rt ,  in  terms  of  copjurr.     Thus  : 

Mm  I 'pom-  fiffv  'c.   (»f  I  lie  copper  solution  required  31.3  cc.  of 

potim';inni  r  vatiidc  solution  : 

TImii  .^i..^  cc.  KCN  =0.40      }(ram  Cu. 
Aiiil  I  re.  KCN  =  0.0127  >^rani  Cu. 

1 1  living  Ihii'i  olitfiiiKMJ  llic*  value  of  the  potassium  cyanide 
Miilnlinii,  ii  I  jin  !»••  usrd  lor  <U'lermininj;  pc*rccMita<;es  of  copper  in 

Jlllnv'i.   liinii/rs.  i\v. 

/■i»/  f  XkUtif^h      Ml  ass  : 


Two  g^ams  of  brass  are  weighed  out  and  treated  with  twenty- 
five  cc.  nitric  acid,  and  the  solution  made  up  to  250  cc. 

Fifty  cc.  of  this  solution  is  made  alkaline  with  ammonium 

hydroxide,  filtered,  and  the  filtrate  titrated  with  the  potassium 

cyanide  solution.      Having   determined   the   number  of  cc.  of 

potassium  cyanide  solution  required  to  decolorize  the  fifty  cc.  of 

the  brass  solution,  the  percentage  of  copper  is  calculated   from 

above  data. 

Consult:  Note  on  the  use  of  potassium  cyanide  in   the  estimation  of 
copper,  by  Geo.  E.  H.  Ellis,  F.  C.  S..  J,  Soc.  Chem.  Industry^  8,  686. 

Determination  of  Copper  by  Electrolysis. 

Weigh  out  five  grams  of  crystallized  copper  sulphate,  dis- 
solve in  500  cc.  water  (preferably  in  a  half  liter  flask),  mix  well. 

Take  fifty  cc,  transfer  to  a  No.  2  beaker,  and  arrange  the 
electroh*tic  apparatus  as  shown  in  Figure  i,  connecting  the 
weighed  platinum  cone  N  with  the  negative  element  of  a  Bun- 
sen  cell  and  the  platinum  spiral  P  with  the  pK)sitive. 

Add  a  few  drops  of  dilute  sulphuric  acid  and  water  enough 
so  that  the  solution  in  the  beaker  covers  two-thirds  of  the  plati- 
num cone. 

Copper  is  deposited  upon  the  platinum  cone  and  the  deposi- 
tion is  generally  complete  in  about  four  hours. 

To  determine  when  all  the  copper  is  precipitated,  take  out  one 
drop  of  the  colorless  solution,  in  the  beaker,  by  means  of  a  glass 
rod,  and  place  the  drop  upK)n  a  watch-glass.  Bring  in  contact 
with  this  drop,  one  drop  of  a  dilute  solution  of  p>otassium  ferro- 

If  copper  is  still  unprecipitated.  brown  copp)er  ferrocyanide 
will  be  formed.  If,  however,  it  is  all  precipitated,  no  brown 
coloration  of  the  drops  will  form. 

When  the  copper  is  all  dep>osited  remove  the  platinum  cone 
quickly,  wash  it  several  times  by  dipping  it  in  distilled  water, 
dr>'  at  100"*  C,  and  weigh. 




Amount  of  copper  sulpliate  takeu  =^  5.000  grams. 

Solution  500  cc. 

Fifty  cc.  taken  for  electrolysis. 

Platinum  cone  +  metallic  copper 36.656  grams. 

Platinum  cone 36-Sa9      " 

Copper  deposited 0.137  gram. 


o.ii?X  IQX  too   ^  ^^^  pp^  j.^^^ 

Where  many  determinations  of  copper,  by  this  method,  are  to 
I  be  made,  the  apparatus  described  by  W.  Hale  Herrick,  /.  Anal. 
I  Chem.,  2,  67,  can  be  used, 

A  very  convenient  instrument  for  generating  the  current  of 
I  electricity  is  Giilcher's  Ihermo-electric  pile.  Figure  2. 

It  consists  of  sixty-six  elements  and  is  etiuivalent  to  two  large 
r  freshly  filled  Bunsen  elements  ;  its  electromotive  force  is  equiva- 
lent to  four  volts,  the  inner  resistance  amounting  to  0.65  ohm, 
so  that  with  an  equal  outer  resistance  the  thermo-electric  pile 
gives  a  current  of  three  amperes.     The  gas  consumption  is  about 
170  liters  per  hour  (6.ooi  cubic  feet). 
The  amount  of  current  should  not  be  excessive,  otherwise  the 
Kdeposit  of  copper  upon  the  platinum  cone  will  be  granular  and 
Knon  -  cohesi  ve . 


References:    "  Bibliography  of  the   Electrolytic  Assay  of  Copper,**   by 
Stuart  Croasdale.y.  Anal.  Chem.,  5,  133-84. 
*' Electro-Chemical  Analysis,**  E.  F.  Smith,  p.  48. 
'*  Quantitative  Chemical   Analysis  by   Electrolysis,**   by  Dr.   Alex. 

Classen,  translated  by  W.  Hale  Herrick.     1894. 
**  The  Utilization  of  the  Electric   Light  Current  for    Quantitative 

Chemical  Analysis,'*  by  P.  T.  Austen  and  J.  S.  Stillwell,/.  Anal. 

Chem.,  6,  127. 
**  On  the  Analysis  of  American  Refined  Copper,**   by  H.  F.  Keller, 

y.  Am.  Chem.  Soc,  16,  785. 


Determination  of  Sulphur  Trioxide  in  Crystallized 

Magnesium  Sulphate. 

Weigh  out  one  and  a  half  grams  of  crystallized  magnesium  sul- 
phate. Transfer  to  a  No.  3  beaker.  Add  100  cc.  water,  a  few 
drops  of  hydrochloric  acid,  and  heat  to  boiling. 

Add  a  solution  of  barium  chloride  in  slight  excess.  Stir  well, 
and  set  aside  for  half  an  hour. 

Filter  upon  two  No.  3*  ashless  filters,  testing  the  filtrate  with  a 
few  drops  of  barium  chloride  solution,  to  make  certain  that  all 
the  sulphur  trioxide  is  precipitated. 

MgSO,  +  BaCl,  =  BaSO,  +  MgCl,. 

Wash  the  precipitate  thoroughly  with  hot  water  until  a  drop 

of  the  filtrate  placed  upon  a  watch-glass  and  brought  in  contact 

with  a  drop  of  solution  of  silver  nitrate  shows  no  turbidity.     Dry 

the  precipitate,  and  ignite  in  a  weighed  porcelain  crucible  to 

constant  weight. 

1st  weight  of  tube  -|-  MgS04  -|-  yHjO 7-9040  grains. 

2ud       '•        •'     "  *'  "       6.5435 

« I 

« I 

MgS04  -h  7H2O  taken 1.3605 

Crucible  -f  BaSO^ 23.502  grams. 

Crucible 22.214      " 

BaSO, 1.288 

BaSO^  :  SO3  :  :  1.288  :  x 

,r  =  0.442  gram  SO3. 

o.44_2  X  100  ^  ^2^8         ^^^^   sQ^ 

1  w  cm.  in  diameter. 

•  I 


Theoretical : 

MgSO,  ^  7H-O  :  SO;  :  :  roc  :  x 
x*^  52.52  p«r  cect.  SO;. 

References:  Frcsenins.  "Qcaat.  Chcm.  Analvsis."  I152.  :. 

"The  Volumetric  Estiinaticz  of   SsIpiLites."    by    D.  Sidersky 
J.  A%al.  Ckem,,  a,  117.- 


Determination  of  Lread  in  Galena. 

Transfer  two  grams  o:  the  tnely  powdered  ore  to  a  four-inch 
porcelain  capsule  :  add  twenty-five  cc.  nitric  acid.  warm,  then 
fifteen  cc.  sulphuric  acid,  and  evaporate  carefully  until  red 
fumes  cease  to  be  evolved,  and  the  residue  is  nearlv  drv. 

Allow  to  cool,  add  a  few  drops  o:  dilute  sulphuric  acid  and 
se\-ent>"-five  cc.  water,  bring  to  a  boil,  niter,  and  wash  well. 
Neglect  the  filtrate.  Wash  the  precipitate  from  the  filter  into  a 
No.  3  beaker,  using  not  over  seventy-five  cc.  water  :  add  100 
cc.  of  a  solution  of  sodium  carb«:>nate  in  water.  \  i  to  10  ^ and  boil 
the  contents  of  the  beaker  for  fifteen  or  twenty  minutes.  Solu- 
tion must  be  strongly  alkaline. 

Bv  this  action  the  lead  sulDcate.  formed  bv  the  nitric  and  sul- 
phuric  acids  upon  the  sulphide,  is  converted  into  carbonate. 
Filter,  wash  well  with  boiling  water  until  reaction  •:•:  washings 
is  no  longer  alkaline.     Neglect  the  filtrate. 

Wash  the  precipitate  into  a  No.  3  beaker  *.vith  about  seventy - 
^\^  cc.  of  water,  add  sevent>*-five  cc.  strong  acetic  acid.  warm. 
and  keep  the  contents  of  the  beaker  at  boiling  temperature  for 
ten  minutes,  beaker  covered  with  2  watch- ^lass. 

The  lead  carbonate  is  therebv  decomrosed  and  soluble  lead 
acetate  formed,  while  any  silica  or  gangue  remains  insoluble. 
Filter,  wash  well  with  hot  water  until  the  washin^^s  are  no 
longer  acid.     Neglect  the  residue  on  the  filter. 

To  the  solution  of  lead  in  the  h^raker.  which  should  not  ex- 
ceed 150  cc.  or  ZQcj  cc.  including  the  washings,  dilute  sulphuric 
acid  is  added  in  slight  excess  until  no  more  precipitate  is  formed. 

After  standing  for  half  an  hour  the  lead  sulphate  is  filtered  off 


upon  a  No.  3  ashless  filter,  and  washed  thoroughly  with  hot 

Dry  at  102''  C.  Transfer  the  lead  sulphate  from  the  filter- 
paper  to  glazed  paper,  and  ignite  the  filter-paper  in  a  weighed 
porcelain  crucible.  After  complete  incineration,  allow  to  cool ; 
add  a  few  drops  of  nitric  acid,  and  warm  gently,  (Any  lead  re- 
duced from  lead  sulphate  by  the  burning  paper  will  be  dissolved, 
forming  lead  nitrate.)  Add  three  or  four  drops  of  sulphuric 
acid  and  evaporate  to  dryness  ;  add  the  rest  of  the  lead  sulphate 
that  is  upon  the  glazed  paper,  and  ignite  contents  of  the  crucible 
to  redness ;  cool  in  desiccator,  and  weigh ;  repeat  to  constant 

Example : 

ist  weighing  of  tube  and  Galena 16.670  grams. 

2d         **  **       "  **      14.503 

f  < 


Galena  taken 2.167 

Crucible  -|-  PbSO^ 17*576  grams. 

Crucible 16.564 


1.012      ** 

PbSO^ :  Pb  : :  1.012  :  x 
;r  =  0.6914. 

2— ?i4_^J^  =131.9  per  cent,  lead  in  the  sample  of  Galena. 

VIII.     . 

Determination  of  Iron   by  Titration  with  Solution  of 

Potassium  Bichromate. 

a.    Where  the  Iron  Soliitio7t  is  t7t  the  Ferrous  Condition. 

Take  one  and  a  half  grams  of  crystallized  ammonium  ferrotis 
sulfate  ;  transfer  to  a  No.  3  beaker,  and  dissolve  in  100  cc.  of  cold 
water  ;  add  ten  cc.  hydrochloric  acid. 

Make  a  solution  of  potassium  bichromate  by  dissolving  14.761 
grams  of  the  **  C.  P.."  salt  in  1,000  cc.  water;  mix  well. 

Each  cc.  is  equivalent  to  0.0168  gram  of  iron.  (Consult  Fre- 
senius,  **  Quant.  Analysis,  London  edition,  §112  b.) 

Fill  a  fifty  cc.  burette  with  some  of  this  solution,  and  drop  the 
bichromate  slowly  into  the  beaker  coniaining  the  iron  solution 
until  a  drop  of  the  latter  placed  upon  a  white  porcelain  slab  and 
brought  in  contact  with  a  drop  of  a  very  dilute  solution  of  potas- 
sium ferricyanide  no  longer  produces  a  blue  or  greenish  colora- 
tion, showing  the  ferrous  salt  to  be  all  oxidized  to  ferric  salt. 
Note  the  number  of  cc.  of  the  bichromate  solution  required  to  do 
,   this,  and  calculate  percentage  of  iron  in  the   ammonium  ferrous 
!  sulphate. 

A mmoninm  ferrous  sulpb ate  taken 1.503  gram. 

I3.Z7  cc.  bichromate  solution  required  to  oxidize. 
1  cc.  =0.0168  gram  iron. 
Then,  ii.78  cc.  =o.ai47  gram  iron. 

?il!47  >=_E^  =  ,4.18  per  cent.  iron. 
1 .503 
Theoretical  percentage  .- 

(NH,),S0,.FeS0(-f-6H,0:  Fe  1  :  100  :  j- 
x^  14.28  per  cent. 

b.    ^'here  the  Iron  solution  Exists  in  the  Ferric  State. 

As  the  use  of  bichromate  requires  the  iron  to  be  in  the  fer- 
rous condition  so  as  to  be  oxidized  by  the  bichromate,  the  ferric 
sah  is  reduced  to  ferrous  as  follows  : 

Take  one  and  a  half  grams  of  ferric  sulphate,"  transfer  to  a  200 
cc.  flask,  dissolve  in  fifty  cc,  water,  add  ten  cc.  hydrochloric 
acid,  and  a  few  pieces  of  "feathered  "  zinc.  All  the  zinc  must 
be  dissolved  and  the  solution  colorless  before  it  can  be  titrated 
with  the  bichromate.  It  is  essential  in  this  process,  that  all  the 
ferric  salt  be  reduced  to  ferrous,  otherwise  the  number  of  cc.  of 
the  bichromate  used  would  give  too  low  a  result  for  the  percent- 
ages of  iron. 

To  keep  the  iron  solution  in  the  flask  from  oxidizing  while  it 
I  is  being  reduced  by  the  hydrogen  from  the  reaction  of  zinc  and 
hydrochloric  acid,  several  methods  are  available  : 

ist.  Method  described  by  Fresenius,  in  which  carbon  dioxide 
j  is  pBssed  through  the  flask  during  reduction  (see  S  112). 

2d.  The  stopper  of  the  flask  is  arranged  to  allow  escape  of  the 

'  nu»  ■mtuiimiiiu  tecric  lulphBtciu^lcad  of  rcmc:  lulphatc. 


hydrogen  generated  by  the  dissolving  of  the  zinc  by  the  hydro- 
chloric acid,  but  prevents  inlet  of  air. 

The  stopper  is  of  rubber  (one  perforation),  through 
which  passes  a  glass  tube.  At  the  upper  end  of  the 
glass  tube  a  piece  of  rubber  tube  (closed  at  b  with  a 
glass  rod),  is  adjusted,  and  at  a  an  opening  is  made 
in  the  rubber  tube,  which,  when  the  contents  of  the 
flask  are  heated,  allows  the  exit  of  gas,  but  which 
closes  and  prevents  the  entrance  of  air  when  heat  is 
removed,  the  so-called  Bunsen  valve. 

3d.  The  method  of  Jones  is  the  most  expeditious 
Fig.  3.       where  a  number  of  reductions  are  to  be  made.— y^. 
Anal,  Cheni.,  3,  124. 
Example  : 

Ferric  sulphate  taken r.520  gram. 

18.01  cc.  bichromate  solution  required  to  oxidize. 

Then,     I?^°-l A°.:2I^2<J5^  =  19.90  per  cent,  iron  in  ferric  sulphate. 


Theoretical  Percentage  : 

Fe.^(S04  \^  +  91^-2^  •  ^^1  '  •  'oo  •  -^' 
A'  ^  19.92  per  cent,  iron  in  ferric  sulphate. 


Determination  of  Phosphoric  Anhydride  in  Calcium 


Weigh  out  one  gram  of  finely  pulverized  calcium  phosphate, 
transfer  to  a  six-inch  porcelain  capsule,  add  twenty  cc.  nitric 
acid,  ten  cc.  hydrochloric  acid,  and  evaporate  nearly  to  dryness. 
Allow  to  cool,  add  twenty-five  cc.  nitric  acid,  seventy-five  cc. 
water,  boil,  and  filter  into  a  one-fourth  liter  flask.  Wash  with 
water  until  reaction  is  no  longer  acid,  and  make  solution  and 
washings  up  to  the  containing  mark  by  the  addition  of  more 

The  reading  must  be  taken  with  contents  of  flask  at  a  tem- 
perature of  15.5°  C.  to  be  accurate. 

Mix  well,  and  take  duplicate  samples,  each  of  twenty-five  cc, 
transfer  to  No.  3  beakers,  and  treat  as  follows  : 


Concentrate  by  evaporation  to  about  fifteen  cc.  Cool  some- 
what, and  add  carefully  ammonium  hydroxide  until  the  solution 
is  alkaline,  then  make  reaction  slightly  acid  with  nitric  acid. 

Add  thirty  cc.  of  standard  ammonium  molybdate  solution, 
with  stirring,  and  then  some  more  ammonium  hydroxide,  but  not 
enough  of  the  latter  to  render  the  liquid  alkaline.  Add  twenty 
cc.  ammonium  molybdate  solution,  and  set  aside  two  hours. 

Filter,  lest  fiitrale  with  a  few  drops  of  ammonium  molybdate 
solution,  to  be  certain  all  of  the  phosphoric  acid  is  precipitated, 
and  wash  precipitate  well  on  the  filter  with  water  containing 
one-eighth  its  volume  of  ammonium  molybdate  solution. 
The  filtrate  and  washings  are  neglected. 

Fifteen  cc.  ammonium  hydroxide  are  poured  upon  the  yellow 
precipitate  on  the  fiher,  and  the  solution  formed  caught  in  a  No. 
2  beaker.  The  filter-paper,  free  from  the  yellow  precipitate,  is 
washed  thoroughly  with  hot  water,  and  the  fiUrate  made  acid  , 
with  hj-drochloric  acid.  This  produces  a  precipitation  of  the 
yellow  ammonium  phosphomolybdate.  Ammonium  hydroxide  is 
added  in  quantity  just  sufficient  to  dissolve  this  and  to  form  a  col- 
orless solution  again. 

Thirty  cc.  of  standard  magnesia  mixture  solution  are  now 
added  gradually  with  constant  stirring,  and  the  beaker  with  the 
precipitated  ammonium  magnesium  phosphateset  aside  for  thirty 

Filter  upon  an  ashless  filter,  wash  with  water  containing  one- 
I   eighth  its  volume  of  ammonium  hydroxide,  dry,  ignite  in  porce- 
[  lain  crucible  to  constant  weight,  and  weigh  as  magnesium  pyro- 
I  phospliate. 
Example  .* 

Amount  ol  calcium  phosphate  taken  =^  1.157  grams. 
Solution  ^  35a  cc. 
35  cc.  taken. 

Crucible  -f  Mg,P,0, 15-^37  grams. 

Crucible 15.5210      " 

Mg,P,0, 0.0827      '■ 

Then.  Mg,P,0, :  P,Oi  : ;  0.0837  ■  '' 

jr  =  o.05i9gram. 
I(  the  P,Oj  in  15  cc.  ^  0.0579  gram,  in  250  cc.  or  entire  solu- 
tion =  0.519  gram. 
.-.  ?-5^J1.^-°g  =  4S.7  per  cent,  P^O^  in  Ca,(PO,),. 


References :  A  very  complete  article  on  **  Mineral  Phosphates  and 
Superphosphates  of  Lime ''  will  be  found  in  the  Atnerican  Chemist^  t% 
103-108;  also 

Bulletin,  "So,  Sg  {Oct.  9,  1892),  "New  Jersey  Agricultural  Experi- 
ment Station,  Analysis  and  Valuations  of  Complete  Fertilizers,  Ground 
Bone,  and  Miscellaneous  Samples.** 

/.  Am.  Chem.  Soc,  X5i  382. 

/.  Anal.  Appl.  Chem.^  5,  418. 

For  method  for  complete  Analysis  of  [Phosphates  and  Superphos- 
phates consult  Fres.  Quant.  Anal.y  p.  689.    Also 

Principles  and  Practice  of  Agricultural  Analysis^  H.  W.  Wiley,  a, 


Determination  of  Chromium  Trioxide  in  Potassium 


Weigh  out  one  g^am  of  the  finely  crystallized  salt,  transfer  to 
a  No.  3  beaker ;  add  100  cc.  of  water,  and  warm  until  com- 
plete solution. 

TAe  twenty-five  cc.  dilute  hydrochloric  acid,  fifteen  cc.  alco- 
hol, add  to  the  solution  of  bichromate,  and  heat  the  mixture 
nearly  to  boiling,  until  the  chromium  trioxide  is  entirely  reduced 
to  chromium  sesquioxide,  the  Itolution  becoming  dark  green  in 
color,  then  boil  out  the  alcohol,  and  add  ammonium  hydroxide  to 
faint  alkaline  reaction.  The  mixture  is  exposed  to  a  temperature 
approaching  boiling,  until  the  liquid  above  the  precipitate  is  per- 
fectly colorless,  presenting  no  longer  the  least  shade  of  red. 

Filter,  wash  with  hot  water  until  the  washings  no  longer  react 

Dry,  ignite,  and  weigh  as  chromium  sesquioxide. 

Example  : 

ist  weight  tube  and  salt 10.942  grams. 

2d        *•         *•       *'      "    8.902      ** 

KjCrjO,  taken 2.040      " 

Crucible  and  CrjOj 43.270      ** 

Crucible 42.230      *' 



This  weight  of  Cr,Os  must  now  be  converted  into  CrO,. 

Cr,0, :  (CrOj), : :  1.044 :  ^ 
;r=  1.3705  grams. 

1.3715  X  ioo_5y  3    percent.  CrO,. 
2.040  /    'JA^  ' 

Theoretical  : 

K,Cr,07  :  (CrOj),  :  :  100  :  ;r 
295  201        :  100 :  X 

^  =  68. 13  per  cent. 
References  : 

Presenius,  Quant.  Anal.,  §  106,  i  a. 

Volumetric  Determination  of  Chromic  Acid,y.  AnaL  Chefn,,  5,  297. 


Analysis  of  Limestone. 

Carbonate  of  lime  is  the  principal  flux  used  by  the  iron  smelter, 
and  as  usually  quarried,  is  called  limestone. 

The  composition  of  this  varies  greatly ;  the  pure  crystallized 
variety  may  be  designated  as  marble,  which  usually  contains 
about  ninety-eight  per  cent,  calcium  carbonate,  the  remainder 
being  srlic^  and  iron  oxide. 

Limestone,  as  distinct  from  marble,  often  contains  organic 
matter  (especially  if  very  dark  in  color),  alumina,  ferrous  or  fer- 
ric oxide,  ferrous  sulphide,  calcium  sulphate,  and  magnesium 
carbonate,  with  the  calcium  carbonate. 

A  small  proportion  of  iron  oxide  is  of  advantage  in  the  smelt- 
ing process,  but  an  excessive  amount  of  magnesium  carbonate 
is  objectionable,  as  it  requires  a  higher  heat  for  fusion  than  cal- 
•cium  carbonate,  and  more  fuel  is  necessar>'  in  the  blast  furnace. 

■1  ■■"im      -T  I   I    inr- 

:2ric<i  **-"fi*  iTis  -^-^.iiri.  rittisi  ::  iarzinst  et7  tr^;  '^not  tni 
Tit  I  =L.':«t*  Z  cni  J  — m-i-r  iii.ia.  -^>*  rr^  i'"'  jusirini 
T'at  J'^:':*  _"-  :-nTrT"T^    utJr'iai    :-:Linrjiit  ^  t:st:r:  i;r-  111:11*- 

lae  «ccK-.^m£  n.T«:? 

2'.  ^v>.nt:tat:v£  AyA;.'!'*:*. 

.  'j"'     «T    ^^     .  "T  T."  _  •>  _  .  .         .'.aL-      ^.  a.^.         w   _'^iA   K  >C  «C 

vjrr.'.  - ^•W'-.z.    .»>■-      14 

-^  Arh 0.052 

M'.--::r': :.5;   r^r  c*nt. 

I-   ■  Y4>--'      •  ^  »"  a     —     ^^  «i.  .. .........        .....        ■^~.      ^"  *  ' 

•    .*";•      .4.     ^  J  .m    ^^  .........................       ,■.-.  .■_ 

1.*- t     i^ 

.  0"A i9C.CC 

It  :•:  n':cts-sary.  now.  to  determine  the  p»ercentage  of  the  sul- 
jyhur  pr*:«^nt  in  the  coal  and  [subtract  it  from  the  amounts  of 
volatile  and  combustible  matter  and  fixed  carbon. 

Th'.-  method  is  as  follows  : 

Take  one  gram  of  the  finely  powdered  coal,  mix  it.  upon  a 
piece  of  black  glazed  paper,  with  about  ten  grams  of  sodium' 
carlx^nate  'dry »  and  five  grams  of  sodium  nitrate. 

Place  a  small  portion  in  a  platinum  crucible  of  fifty  cc.  capac- 
ity, find  heat  to  redness.  When  combustion  is  complete  add 
v>me  more  of  the  coal  mixture,  repeating  the  operation  until  all 
has  >H:en  transferred  to  the  crucible  from  the  glazed  paper.  Heat 
at  a  red  heat  for  fifteen  minutes,  making  certain  that  no  parti- 
fles  of  carlxjn  remain  unconsumed. 

Allow  to  cool,  transfer  crucible  and  contents  to  a  No.  3 
Inraker.  adrl  \r/j  cc.  water,  and  wann  carefullv  until  the  mass 

Remove  the  crucible  from  the  beaker,  washing  it  once  with 
hot  water,  allowing  the  washings  to  run  into  the  beaker.  Fil- 
ter the  solution,  acidify  the  filtrate  with  hydrochloric  acid,  boil, 
and  add  solution  of  barium  chlonde  in  slight  excess.  Allow  to 
stand  twelve  hours,  filter,  wash  well,  dry,  ignite,  weigh 
barium  sulphate,  and  calculate  to  sulphur. 


Atiiouiit  of  f:i>al  taken ].oi6  grams. 

CriK  ilih:  -|-  HaSO, 16.553 

C'riK:i!>lf 16.511 

•  • 

>  • 

iJaSO, 0.042 



S  =  OXJO^'  g^^a:. 

0.00^7  X  TOO  _*• 

*^-^"?/  ^  «^"  =0-^  per  cent.  ^. 

1. 016 

Taking  this  amonnt  and  subtracting  one-half  of  i:  from  the 
volatile  and  cxymbastible  matter  of  the  coal,  and  one- half  from  the 
fixed  carbon,  the  coal  analysis  will  be  : 

Moisture o.  53  per  cent. 

Volatile  and  combnstible  matter 35-7* 

Fixed  carbon 59-7^ 

Snlphnr 0.56 

Ash 3-46 

Total 99-99 

In  most  cases  the  snlphnr  in  coal  exists  combined  with  ixt>n 
to  form  ferrous  sulphide :  it  also  occurs  as  calcium  sul- 
phate, or  both  forms  may  be  present  in  the  same  coal. 

To  determine  the  sulphur  trioxide  combined  with  the  lime. 
take  ten  grams  of  the  finely  powdered  coal  and  digest  at  a  gen- 
tle heat,  two  hours,  in  a  solution  of  sodium  carbonate.  It  is  fil- 
tered, washed  with  hot  water,  the  filtrate  made  acid  nnth  hvdro- 
chloric  acid,  and  the  sulphur  trioxide  precipitated  with  barium 
chloride  solution. 

From  the  weight  of  barium  sulphate  obtained,  the  amount  of 
sulphur  trioxide  Ls  calctilated. 

DeterminatwH  of  Sulphur  in  Coal  b\  ihe  Est'hka-Fresenius 


One  gram  of  the  finely  powdered  coal  is  mixed,  in  a  platinum 
crucible,  with  twice  its  \x>lume  of  a  mixture  of  one  part  sodium 
carbonate  and  two  parts  of  calcined  magnesia,  then  heated  in  an 
uncovered  platinum  crucible  until  the  mass  becomes  heated  to  a 
low  red  heat  and  the  grey  color  of  the  mixttire  changes  to  a  yel- 
low or  brownish-yellow  hue.  Allow  to  cool,  treat  with  bromo- 
hydiochloric  acid,  filter,  boil  out  the  excess  of  bromine,  and 
precipitate  the  sulphur  trioxide  with  barium  chloride  solution, 
as  barium  sulphate  and  determine  percentage  of  sulphur. 

Ann,  Cketm.  •  Liebig  .  76.  90. 
Dimg,  PioljUck.J.,  si  a.  403. 
Am^  Ck^misi,  6.  83. 
y.  Aumi.  Ckrm.,  6,  86. 
/.  Aumi.  Okrm..  6,  585. 
y,  Aumi.  Orm.,  6,611. 

**  Ob  the  manner  in  which  Snlphur  in  Coal  and  Coke  is  Combined.*' 
by  Dr.  F.  Mnck./.  .S^.  Chem.  /m/itsiry.  6.  468. 


Delerminalion  of  Phosphorus  in  Coal  and  Cote. 

Five  grams  of  the  powdered  coal  or  coke  are  transferred  to  a~ 
platinum  boat  (Fig.  5).     This  boat  is  two  inches  square,  one- 
half  inch  deep,  and  made  from  0.002  platinum  foil. 

Care  should  be  taken  in  making  the  boat  that  the  comer  flaps 
fit  tightly,  so  that  none  of  the  ash  will  be  lost  by  getting  into 
the  interstices.' 

A  tripod,  Erdnian  chimney,  and  two  pieces  of  platinum  wire 
bent  three-fourths  of  an  inch  below  top  of  the  chimne\  complete 
the  apparatus. 

The  heat  applied  for  the  first 
five  minutes  should  be  a  low 
red,  ill  order  that  none  of  the 
coal  shall  be  lost  in  the  escape  of 
the  volatile  matter.  After  that 
the  gas  should  be  turned  on  full, 
and  a  bright  red  heat  main- 
tained. It  is  not  necessary  that 
the  sample  be  ground  very  finely. 
After  complete  combustion  of 
the  carbon,  the  ashistransferred 
to  a  platinum  crucible  and  fused 
with  five  grams  of  sodium  car- 
bonate and  one  gram  of  po- 
tassium nitrate.  The  fused  j 
mass  is  dissolved  in  forty  cc.  of 
dilute  hydrochloric  acid  in  a  No.  4  beaker  and  evaporated 
therein  nearly  to  dryness ;  thirty  cc.  of  strong  nitric  acid  are 
added  and  evaporated  also  nearly  to  drj'ness.  The  solution  is 
then  diluted  with  water,  filtered  from  the  silica,  and  the  phos- 
phoric acid  precipitated  with  molybdate  solution. 

The  analyses  of  a  few  representative  coals  are  here  givot:  j 
'■  Boo  Hhad  Cannei."  Coal. 

Moisture   0.60  per  ceat. 

Volatile  and  combuatible.  matter 71.30     "       " 

Fixed  carbon 






Pittsburg  Bituminous  **  Coal. 


Volatile  and  combustible  matter 

Fixed  carbon 



1.28  per  cent. 












Total 100.00 

II      II 

"  Pbnn  Anthracite,'*  Wilkes-Barre,  Del.  &  Hudson  Canal  Co.'s 

**  Vein  No.  5." 

Moisture  4.182  per  cent. 

Volatile  and  combustible  matter  4.283 

Fixed  carbon 85.320 

Sulphur 0.794 

Ash  5.521 

Total 100.000 

It  is  found  in  practice  that  coal  from  the  same  vein  or  seam 
varies  in  composition  with  the  size  of  the  coal ;  the  percentage 
of  ash  increasing  as  the  size  of  the  coal  diminishes.  Thus,  sam- 
ples collected  from  the  Hauto  Screen  building  of  lychigh  Coal 
and  Nav.  Co.,  Pa.,  gave  the  following  :* 

size  of  coal. 




.    1.722 

•  1.426 

•  1-732 
.    1.760 



Buckwheat    1.690 





3.518  88.489 
4.156  83.672 
4.046         80.715 

3.894  79<H5 
4.058        76.918 

These  coals  afe  separated  into  different  sizes  according  to  the 
mesh  of  the  screen  over  which  they  pass.  The  sizes  noted  in 
the  above  table  passed  over  and  through  sieve  meshes  of  the  fol- 
lowing dimensions  : 

Broken  or  grate  size through 

Egg  "    






f  I 




4.00  in. 














The  composition  of  the  ash  of  coal  or  coke  is  sometimes  de- 
sired. The  analysis  can  be  made  in  a  manner  similar  to  scheme 

1  7>unsaciioms  Amer.  Inst.  Mining  Engineers,  14,  720. 


Analysis  of  a  sample  of  ash  of  a  Welsh  coal,  bj' J.  A.  Phillips, 
gave : 

Silica    26.87  per  cent. 

Alumina  and  iron  oxide 56*95 

Lime 5.30 

Magnesia 1.19 

Sulphuric  acid    7.23 

Phosphoric  acid   0.74 

rndetermined  1.72 

Total 100.00 



All  analysis,  by  Gaultier,   of  the  ash  of  a  sample  of  English 
coke,  gave  the  following  : 

Silica    42.10  per  cent. 

Alumina    34.40    ** 

Calcium  carbonate 4.80 

Magnesium  carbonate 0.40 

Calcium  sulphate 1 2.55 

Ferric  oxide 5.28 


((      f  I 

(4  44 

44  44 

Total 99.53     ••      ** 

Coke  is  the  best  solid  fuel  for  the  blast  furnace  in  the  manu- 
facture of  pig-iron. 

Charcoal,  while  having  less  ash,  and  producing  combustion 
more  readily,  cannot  be  used  in  furnaces  carrying  large  burdens, 
since  it  easily  crushes  and  pulverizes. 

Anthracite  coal  ignites  and  burns  slowly  in  the  furnace,  and 
though  it  can  withstand  the  burden,  generally,  without  crush- 
ing, its  slow  work  in  the  furnace  has  caused  coke  to  supersede  it. 

The  value  of  a  coke  is  determined  : 

FJrsf,  by  chemical  analysis  ;  a  good  coke  showing  a  low  per- 
centage of  ash,  sulphur,  and  phosphorus,  and  a  high  percentage 
of  fixed  carbon. 

Second,  by  mechanical  tests,  which  comprise  '*  Crushing 
Strength,"  "Porosity,*'  specific  gravity,  etc. 

The  crushing  strength  can  be  determined  by  taking  several 
samples  of  the  coke,  each  one  cm.  high,  and  placing  them  in 
proper  position  in  a  Thorner  compression  machine,  Fig.  6. 

Good  coke  gives  a  compression  strength  of  160  to  175  kilos 
per  cubic  centimeter. 

a  oooEpreasiciE  s 





i2scd  bv 

i«irt,"i>r  gjjtil*  car  ^  5e£e^E=3ed 


»  ts>  f 


c  g:ai:i3    cc 

j-,3  ■ 
od  posts'  ix  2 


xi  <:c 



absorbed  by  the  mass,  gives  the  true  volume  of  water  displaced 
by  its  particles,  and  hence  the  means  of  determining  their  speci- 
fic gravity.  The  division  of  the  amount  of  water  absorbed  by 
the  amount  of  water  displaced,  gives  the  amount  by  volume  of 
the  pores  in  a  unit  of  material,  and  the  division  of  the  weight  of 
the  water  absorbed  by  the  weight  of  the  dry  mass,  gives  the 
volume  of  pores  in  a  unit  of  weight  of  the  material : 

Leta  =  the  weight  of  the  dry  material. 

b  =  the  weight  of  the  water  which  the  material  can  absorb. 

c=i  the  loss  in  weight  in  water,  of  the  saturated  material. 


c  :  a  :  :  looo  :  a  =  the  apparent  specific  gravity,  or  the  specific 

gravity  of  the  mass. 

c —  b'.aw  looo :  a  •=.  true  specific  gravity,  or  specific  gravity 

of  the  particles,    water    being  looo. 

^  :  ^  :  :  ICO  :  a  =  percentage  by  volume  of  the  pores  in  the  ma- 

a\b\\  ICO  :  a  =  volume  of  pores  in  loo  parts  by  weight  of  the 

material,  say  cc.  in  loo  grams. 

In  filling  porous  substances  generally  with  water  two  methods 
are  in  use,  one  to  soak  the  specimens  in  water  for  a  time  and 
then  to  place  them  in  water  under  the  receiver  of  an  air-pump 
and  exhaust  until  no  more  air  is  given  off ;  and  the  other  to 
keep  them  suspended  in  boiling  water  until  the  pores  are  filled 
with  water,  as  is  shown  by  their  ceasing  to  gain  weight  on  tak- 
ing them  out,  cooling,  and  weighing. 

A  combination  of  both  methods  will  be  found  advisable  in  ex- 
perimenting with  coke.* 

A  series  of  nine  specimens  from  the  Bradford  Works  of  Frick 
&  Co.,  yielded  as  follows  : 

True         Apparent  Per  cent  of    Cc.  in  xoo 
Moisture,      sp.  g^r.         sp.  t^r.     cells  by  vol.     grams. 

Maximum 0.096  1.79  1.033  54-37  66.31 

Minimum 0.008  1.73  0.819  42.20  40.83 

Average 0.034  1.76  0.802  49.37  55.73 

Coke.     El  Moro,  Colorado.  Twelve  samples. 

Maximum 0.225  1.85  1.047  54.66  71.36 

Minimum 0.025  1.61  0.766  61.47  41.56 

Average 0.114  169  0.919  45.75  50.39 

I  Fiifls,  Mills  and  Rowan,  pp.  149-150- 



The  following  is  a  report  upon  a  sample  of  Connelsville  coke 
Anai,ysis  op  thb  Coai«  prom  which  thb  Cokb  was  Madb. 

Per  cent. 

Water 1.105 

Volatile  and  combustible  matter 29.885 

Fixed  carbon 57-754 


Ash 9.895 

to    Jh 

1.113^-3  ^ 

<    . 



Analysis  op  thk  Coke. 

Per  cent 

Water 0.030 

Volatile  and  combustible  matter 0.460 

Fixed  carbon 89-576 


Ash 9.1 13 

Total 100.000 

-   9> 
to    u 

0.821  yrs    o 



8  W 

Specipic  Gravity,  Porosity,  Pbr  Cent  op  Cei,i,s,  Weight 
per  Cubic  Foot,  Etc.  op  the  Coke. 

Apparent  specific  gravity 

True  specific  gravity 

Per  cent,  of  cells  by  volume 

Volume  of  cells ;  cc.  in  100  grams 
Weight  per  cubic  foot  (lbs. ) 





ricthotf  of  flamiiactnre. 

Charge    Yield  Time 

in          per  of 

pounds,    cent,  coking. 


7600        63  and 

12'  X  '6'  72 

11X5' 6" 




To  be  used  for. 


Size  of 

Iron  blast.    7o'Xi6' 

John  Fulton,  M.  E.,  gives  the  following  as  the  standard  for 
the  chemical  and  physical  properties  of  coke  : 


lirj-.  I  Wei  ;Dry.  |  Wet.  Cok*  Oclli.  g  ^[£f  *^  « 


>.i.6r  s'-^  ■> 

7M    49.96 

,«>■««       .K" 



-  = 

"    "c^,,^ 


=  -= 




i  ^ ' 

r  i 



MHlHlanl  Coke 

kefnemea  :  "On  tlie  Ilcnsity  of  Coke."by \Vm.  A.Tildeu,  F.  R.S.,y. 
Sm\  Chfm.  /lid..  3i  610. 

"An  InveHtiffation  Kcgardiii);  the  Differences  Between  Cokes,"  bjr  Sir 
1.  t.(>w'thiiiu  Iii-11,y.  h-au  and Sterl  iHstitHte,  1885. 

"  The  I'hvBical  ami  Chemical  Properties  of  Coke,"  by  John  Fulton, 
Transactions  .hiifrifan  InstUule 0/ Mining  Engineers.,  1885, 

<'•  nil)  I  Unwell  <ltT  tCoks-C  hemic,  von  Oscar  Simmersbach,  Berlin.  1S95. 

"  A  MvtUixl  of  Obtaiiiiiit;  tile  Si>ecific  Gravity  and  Porosity  of  Coke," 
ljy  W.  Cnrrick  .Anderson, y.  .Sloe.  Cheni.  Ind..  15,  10. 

"  An  Invfstiftatinn  nf  Coaln  for  Making  Coke  in  Seniet-Solvay  Ovens," 
liy  J.  n.  I'liiiKK-k ,  y.  Ana!.  .Xppl.  Cbetn.,  7.  135. 




Jiff  I 

li  fflii 


°-  aa'gga 


ill!  -ills -I 



Example : 

Ten  grams  of  iron  ore  taken. 

Insoluble  residue  and  crucible 10.551  grams. 

Crucible 10.301      ** 

0.250      ** 

o.25Xioo_.^       per  cent,  insoluble  matter. 

Solution  =  500  cc. 

Phosphorus  pentoxide,  (100  cc.) 

a.  Crucible  +  Mg,P,07 8.923  grams. 

Crucible   8.919      *« 

Mg,PA=  0.004      ** 

h.  Crucible  +  MgjPjO; 7.6140  grams. 

Crucible 7.6105     ** 

Mg,P,0,  =  0.0035     *' 
Mg^PjO; :  PjOj  :  :  0.0038  :  x 
X  =  0.0024 
0.0024  X^5X  _ioo  ^  ^^^  ^^  ^^^^   p^Q^ 

=  0.054         **  P. 

Iron  determination. 

Fifty    cc.    reduced    with    zinc    required    34.65    cc.  standard 

K,Cr,0,  solution.     One  cc.  K,Cr,0,  corresponds  to  0.0168  gram 

iron.     34.65X0.0168  =  0.58212  gram  iron  in  fifty  cc.  of  the 

iron  solution. 

Then  o^5822_22<J^X_Loo  ^  ^g^^  ^^  ^^^^   p^  .^  ^^^  ^^^ 


=  83. 16        *  *  Fe,0,  in  the  ore. 

Sulphur  Trioxide  (50  cc.) 

Crucible  H- BaS04 11.126  grams. 

Crucible 1 1 .01 1       ** 

BaS04=  0.015       " 
BaS04  :  SOj  :  :  0.015  :  x 
X  ^0.0051 
0.0051  X  10  X  log  _  ^5,         ^^^j  so, 


Alumina  (50  cc.  from  250C.C.  =  1-5  of  100  cc.) 

Crucible  -f  AljOj.Fe^Oa •• 12.6614  grams. 

Crucible 12.3160       *' 

Al203,Fe203=    0.3454       ** 



Fifty  cc.  of  the  iron  solution,  by  titration,  gave  0.5S212  gram 
of  iron  or  0.3326  gram  of  ferric  oxide  for  fifty  cc.  of  the  250  cc. 
solution  of  Fe,0,A],O,in  (4)  scheme  XIII.  Subtract  this  weight 
(0.3326)  from  weight  of  alumina  and  ferric  oxide,  (0.3454)  in 
the  fifty  cc.  The  remainder  equals  0,0128  grams  alumina. 
o.o»8X»5Xioo  ^ 3  ,^  ^^  „„^  Al,o.. 

Another  method  of  determination  of  alumina  in  presence  of 
ferric  oxide,  where  the  aluminum  oxide  is  in  small  amount,  is 
to  fuse  the  weighed  oxides  with  potassium  hydroxide  in  a  silver 
capsule,  and  extract  with  water.  The  alumina  forms  a  soluble 
salt  whereas  the  ferric  oxide  remains  undissolved. 

Filter  off  ferric  oxide,  wash,  ignite  and  subtract  weight  from 
former  weight  of  both  oxides.     Difference  is  weight  of  alumina. 

rUnganese  oxide  ( 100  cc. ) 

Crucible  +  MnjO, u,  166  grams. 

Crucible 12.131       " 

0.35  X  5  X  I 

Mu,0,=   0.0 
=  1.75  per  cent,  MnjO,. 

Une  ( 100  cc. ) 

Crucible +  CaO 8.936  grams. 

Crucible B.9»9 

0.007       '■ 
?:???I2il2ii5e  =  o.35  per  cent,  CaO. 
riacneris  ( loo  cc.  I 

Crucible  +  Mg,P,0, 8. 929  grams. 

Crucible S.919 

Mg,P,0,=     0.010      ■• 
Mg,P,0,  :(MgO),  1:0.010  :.r 

A-  =  0.0036 

0.0036  X  5  X  l«>  =  o.i8  per  ceil.  MgO. 
Water  of  Hydration. 

Amount  oi  ore  taken 1.1(17  grains. 

CaCl,  lube  +  H,0 19065 

CaCI,  tube 28.963      " 

Carbou  dioxide  absent. 


8.03  percent.  H,0(hydrateii) 



H  O  (h  d  ated 


Total.     99.8a         ■' 
If  tile  ore  is  a  magnetite,  the  iron  exists  asFeO,Fe,0,.    The] 
are  several  methods  of  determining  the  FeO  in  presence  of  Fe,03 
The  one  recommended  by  Whittlesay  &  Wilbur,'  is  trequentljg 
used,  but  the  method  of  Allen  is  simpler  and  is  to  be  preferr 
It  Is  as  follows  : 

One  gram  of  the  very  finely  powdered  iron  ore  is  heated  | 
a  small  sealed  combustion  tulie,  half  full  of  fuming  hydrocMorf 
acid  (25  cc.  of  the  acid  being  sufficient).     The  heating  is  f 
performed  in  the  water  bath  for  two  or  three  hours,  then  in  ^ 
hot-air  oven  at  150°  C  for  four  hours  more. 

The  ore  is  thus  completely  decomposed  and  after  cooling  t 
tube,    it  is  broken    under  water  in  a   beaker,  and   the  ferroiU 
oxide  immediately  determined  by  titration  with   standard  solltf 
tion  of  potassium  bichromate.     The   amount  of   ferrous  oxict 

r    ORE    ANALYSIS. 



mbtracted  from  the  total  oxides,  determined  in  another  sample 
i  the  ore,  gives  the  amount  of  ferric  oxide. 
Some  iron  ores  resist  solution  in  acids,  in  which  case  the 
iieme  is  modified  as  follows  i 

Two  grams  of  the  finely  pulverized  ore  are  fused  with  fifteen 
-ams  of  fusion  mixture  (Na,CO,  4- K,CO,)  in  a  large  plati- 
[~Bum  crucible  for  one  hour.  After  cooling  the  fused  mass  is 
treated  with  boiling  water,  the  contents  transferred  to  a  four  inch 
porcelain  capsule,  made  acid  with  hydrochloric  acid  (carefully), 
and  evaporated  to  dryness,  twenty-five  cc.  hydrochloric  acid,  five 
cc.  nitric  acid  are  added,  wanned  until  solution  of  iron  is  com- 
plete,then  fifty  cc.  of  water  added,  and  the  solution  filtered  from 
e  silica, etc.    The  analysis  caunow  befinished  byschemeXIII. 

Determination  of  Chromium  in  Chrome  Iron  Ore' 
Take  a  half  gram  of  the  very  finely  divided  mineral  and  inti- 
mately mix  it  with  twelve  grams  of  a  mixture  containing  equal 
parts  of  dry  sodium  carbonate  and  barium  dioxide,  transfer  to  a 
large  platinum  crucible  and  fuse  over  the  Bunsen  burner  for  one 
hour.    At  the  end  of  this  time  a  quiet  fusion  is  obtained  and  the 
decomposition  is  completed.     The  crucible  is  then  placed  in  a 
beaker,  covered  with  water,  and  hydrochloric  acid  added,  a  lit- 
at  a  time,  till  the  mass  is  completely  disintegrated.     The 
cible  is  then  removed,  the  solution   made  strongly  alkaline 
ith  caustic  potash,  and  ten  cubic  centimeters  of  a  five  per  cent, 
ilution  of  hydrogen  dioxide  added  to  oxidize  the  small  amount 
chromium  sesquioxide  that  may  be  present.     The  solution  is 
iw  boiled  for  twenty  minutes  to  remove  any  excess  of  hydrogen 
[oxide,  made  acid  with  hydrochloric  acid,  and  the  amount  of 
imic  acid  determined  by  the  aid  of  a  standardized  solution  of 
'ous  chloride,  one  cubic  centimeter  of  which  corresponds   to 
,015  gram  Cr,0,. 
The  usual   method   for  the   determination  of  chromium   in 
ne  iron   ores,  is   that   of  Genth's'  which   consists  in  the 
n    of   the  finely  divided  ore  with  potassium  bisulphate. 
detail  as  follows  : 

[occBi  of   Donath  modified  by  t,.  P.  ElnolciKI  and  G.  W.  PatteruD.   J.  Amml. 



A  half  gram  of  the  pulverized  ore  is  fused  in  a  platinum  cni-' 
cible  with  ten  grams  of  potassium  bisulphate  (or  one  hour. 
This  is  allowed  to  cool  when  five  grams  of  dry  sodium  carbonate 
and  one  gram  of  potassium  nitrate  are  added  and  the  mass  sub- 
jected to  fusion  for  one  half  hour.  After  cooling  the  crucible  is 
transferred  to  a  No.  4  beaker  and  the  contents  treated  with 
water.  Filter,  wash  well,  and  evaporate  the  filtrate  to  dryness 
in  a  porcelain  capsule  after  acidulating  with  hydrochloric  acid. 
Treat  with  hydrochloric  acid,  filter,  wash  with  hot  water,  and 
reduce  the  chromium  trioxide  to  chromium  sesquioxide  by  the 
addition  of  ten  cc.  of  alcohol  and  boiling  (consult  scheme  X). 
Filter,  dry  and  ignite  the  precipitate,  which  may  contain  some 
alumina,  etc.,  with  a  small  amount  of  sodium  carbonate  and 
potassium  nitrate  in  a  platinum  crucible;  cool,  dissolve  the  fused 
mass  in  water  and  transfer  to  a  platinum  capsule  and  evaporate 
to  a  syrupy  consistency.  Add  gradually  crystals  of  potassium 
nitrate  and  continue  this  until  effervescence  ceases,  add  ammo- 
nia to  alkaline  reaction  and  filter.  This  precipitate  contains  the 
alumina,  etc.,  that  might  have  been  present  in  the  first  precipi- 
tation of  the  chromium  sesquioxide. 

The,chroniium  trioxide  in  the  filtrate  is  reduced  to  the  sesqoj 
oxide  by  the  addition  of  excess  of  solution  of  sulphurous  aci 
Boil,  make  faintly  alkaline  with  ammonia  and  continue  boilii^ 
for  several  minutes.  Filter,  wash  well.  dr\-,  ignite  and  wetd 
OS  Cr,0.. 

The  fflllowing  analyses  indicate  the  varying  amounts  of  chrfl 
mium  sesquioxide  in  chrome  iron  ores : 













■Hi  IS 


t  ass:,?;:.?::::::::;::;:'.:' 

IS-*'''    "   "5R 



,J  iS^'SS''::T^:.'^±: 






IRON    ORB    ANAI.VSIS.  35 

Jieference. — "  New  process  for  the  oxidation  of  cliromijim  ores  and  the 
manufacture  of  chromates,"  by  J.  Massiguoii.  /,  A'lal.'Appl.  Chem., 
5.  46s 

Determination  of  Titanium  in  Iron  Ores. 

The  method  of  Bettel'  is  generally  used. 

Fuse  about  half  a  gram  of  the  finely  powdered  ore  with  six 
grams  of  pure  potassium  bisulphate  iu  a  platinum  crucible  at  a 
gentle  heat,  carefully  increased  to  redness,  and  continued  till 
the  mass  is  in  tranquil  fusion.  Remove  from  the  source  of  heat, 
allow  to  cool,  digest  for  some  hours  in  150  cc.  of  cold  distilled 
water  (not  more  than  500  cc.  are  to  be  used,  as  it  generally 
causes  a  precipitation  of  some  titanic  acid)  ;  filter  off  from  the 
silica,  dilute  to  i20occ.,  add  sulphurous  acid  until  all  the  iron  is 
reduced,  then  boil  six  hours,  replacing  the  water  as  it  evaporates. 

The  titanic  acid  is  precipitated  as  a  white  powder,  which  is 
now  6!tered  off.  washed  by  decantation,  a  little  sulphuric  acid 
being  added  to  the  wash  water  to  prevent  it  carrying  away 
titanic  acid  in  suspension.  Dry,  ignite,  allow  to  cool.  moisten 
with  solution  of  ammonium  carbonate,  re-ignite  and  weigh. 
The  titanic  acid  is  invariably  obtained  as  a  white  powder  with  a 
faint  yellow  tinge,  if  the  process  has  been  properly  carried  out. 

The  tableon  the  next  page  gives  the  composition  of  theprinciple 
varieties  of  iron  ores. 

Re/er(n(es.—{\xo\i.QxK.f,.)  "The  Iron  Ores  of  the  United  States.  /Vo- 
ctedtHgs  0/ Ike  Iron  and  Steel  InsHluU.,  wpcz\a.l  \o\vttae,  1890,  pages  6S-91. 

"  Htnta  (or  Beginners  in  Iron  Analysis,"  by  David  H.  Browu./.  Anal. 
Afpt.  Chem..  %,  125. 

'■  Dctermi nation  of  Iron  by  Stannous  Chloride,"  by  R.  W.  Mahoii, 
Amer.  Ckem.JoHmal.  15,  360. 

" The  Volumetric   Determination  of  Titanic   Acid  and  Iron  ill  Ores," 

r  H.  I,.  Wells  and  W.  L.  Mitchell,  /.  Am.  Chem.  Soc.  17,878. 

"The' Constitution  of  Magnetic  Onide  of  Iron,"  by  W.  G.  Brown.y. 
i/.  Appl.  Chem.  7.  26. 



sp*0*      •    ti     •    O^    •      •      • 
-9JO  3|q)Bds  t^    I    ON    !     I    CO    I    dv    I     I     I 


.paBH^SBia.,  w    d 

•    O   c«   «   « oo 

•         •••••••••  • 

.   o  o  o  ^ t^ 


w    «    O 
_,__.^_   ..^     -  .  __    _    O    «    "^ 

'9jo*aoaf  Xb{5  ioC>d»OMMO'^dd 

•31JOA  >^3N  00    "^^^^    '^OOoO    O 






On   «     »0 












'pOBliClBIV    ^ 

•aao  ^ 

aaxi  auiaxqo  (i^ 



:  t 


CI   00     (>l   00   00 

•Xasj^f  M3M  lo  "-J   N-    «    «o 


lO  1-4 

o  o 

•  • 

o  o 







.    fo    .    <^   *!?  2J 

lO    .    »o 

x^iomuax  :  ^  :  1 "? '°. 

•     ,      • 

a^iuomn    •  JT    •    '^  "^    '^ 

fo  :  o 



'jBinoads  pan 
aouadns  s^ib^ 

.    «4^0nOwqn.vo»oQ,.     • 

o   g  M  o  b  »o 

ON    4-1 

o  o 


*••  I  "^  ^     *^ 

?  /^"^  d*  «  ^2  o  qdS.2  ^•oa 



sp*0*      •    ti     •     O^    *      •      • 
'9JO  oiq^Bds  t^    I     9^    I      *     ro    *    dv    I      I     ! 


.paBa^^Bia.,  «  o 


«   w   2. 9S. 

•         •••••••  ■ 

o  o  ^ t^ 



•31JOA  ^'N  w   ^  ^  d5    fO  6b' 
'9JO  aoji|  Xbi3  io  d    O    io  •-«    m 

t^  u^  00    w    •-•    O 

o5   P   «   •^ 

c>  •-•  o 




ON    « 

')n3p9aaao3  «9  ^    . 

























pOBliClBIV    J? 


:  Z 


CI    00     C4    ^    00 

M    vO 

to  M 

o  o 

•  • 

o  o 









«    fO  «    lo     .     »o 
"-•     »0  VO     'tf     .     w 

•  •  •  «  • 



.WO^QnOmon.^JPQL"      • 

UBinD^ds  pan    •*?w'?^1"*     '^iO.^:     I 

aouddns  aaiBl    IdS'^OOioIOO'^     .     . 


:••::::::::::«         * 
:::.:.:::.::.  g«{     | 

^  /^  d*  -  ^2  o  q.o  S-S  JO'S 

Cntciblc  +  SiO, 

Cnrablc  +  Mg^^,  =  1  i.«Ms  BiMi 



'                       ■*"*■■ 

-£aiE?E5g     g q  I » J  - 

t  Snbtiacl  Pe^,      I .  "^-(""SrSSS 

HooDdttyliUm^Diit  —  ".mft  ,3*"6    5-    5^B    5 

**  \^  '""  =  ^.46  per ctDl.  CaO.         =^  =f        5  E?  altv         ?  !  3  I  " 

acftile  +  liBiPiO,  =  ii.Ma5]Cianu.  ?5_5i-.  S    ^ta*?,  JS^tS         =■  =  ""« 

—         =11. S:^  =^3-2  a     =~."5  "-"e-J'S         ^mK'm 

,P,0,:  IMgfl},::  0.01353    J-  ==-2-  ~!.tt,»>  53= -2  "23«" 


per  «Dt  MgO.         •  "  E„  S;^  J.       I      -a  jE.! 

Crvcible  +  BaSO,  =  ii.9i3j6Bniiu.  El!£"r»  E  ?  ■  B. 

Si!??''  "3*3 

ii»so.=  O.OMJ6  ,=.  J,;.-*  "Is* 

S  =  iS3P««nL  „Iii,_^  S3  5  B 




Lime  (CaO) 36.46  per  cent. 

Magnesia  (MgO) 2.12 

Silica(SiO,) 4325 

Alumina  ( A1,0,) 15-94 

Ferrous  oxide  ( FeO ) 0.31 

Sulphur  (S) 1.53 

Manganese  Oxide  ( MnO, ) 0.09 

Phosphoric  Acid  (P,Oji) 0.09 

Undetermined 0.21 

Total,  100.00 


f  • 

h  i 
•  1 


Form  op  Blank  usbd  for  Reporting  Blast  Furnacr 

Slag  Analyses. 



Furnace. . 
Ores  used 

No.  of  Iron 

Lime  (CaO) 

Magnesia  ( MgO) 

Silica  (SiO,; 

Alumina  ( Al^Oj) 

Oxide  of  Iron  ( FeO) 

Calcic  Sulphide  (CaS) 

Manganese  Oxide  (MuOJ 
Phosphoric  Acid  (P,O.0  •  •  •  • 

Chemist  or  laboratory 


Examples  0/ Blast  FmriuueSags  Analyses. 

No.  1.'  No 

■  -  —  ■-- •■• 0.370  per  cent.  0.456  p 

45-460        ■'  35.000 

■aIjO, 16.590        ■'  14.36* 

C«0 32-805         "  45-370 

MgO 1.080        ••  1.398 

MnO,   0.083        "  trace 

Sniphar  |  Sulphide  of  1    1.571         "  1,873 

Calcium.'     Calcium     '    I-963         "  l-S<>o 

Phosphoric  Acid  ( P,Os) 0.008       -  0.059 

Undeienntned  Loss 0,070        "  

'  Some  varieties  of  slag  are  soluble  in  hydrochloric  acid,  in 
which  case  the  analysis  can  be  made  by  scheme  XIII.  This 
applies  also  to  open-hearth  slags,  refinery  slag,  tap-cinder,  mill- 
ciuder  and  converter  slag. 

Basic  slags,  from  the  Thomas- Bessemer  Process,  often  con- 
tain as  high  as  thirty  per  cent,  of  phosphoric  acid  and  require 
a  somewhat  different  process  of  analysis.     Thus  : 

One  gram  of  the  finely  pulverized  slag  is  fused  with  excess  of 
sodium  carbonate  in  a  platinum  crucible.  Extract  with  water, 
acidify  solution  with  nitric  acid  and  evaporate  in  porcelain 
capsule  to  drj'ness.  Take  up  with  hydrochloric,  dilute  to  half  a 
liter  and  precipitate  the  phosphoric  acid  by  the  Acetate  process.' 

The  precipitate  is  filtered,   dissolved  in    hydrochloric  acid, 

^cess  of  nitric  acid  added,  and  the  solution  concentrated  until 

hydrochloric  acid  and  acetic  acids  are  expelled.     The  nitric 

id  solution  is  diluted  to  half  a  liter  and  two  portions  are 
taken  {each  250  cc. )  and  the  phosphoric  acid  determined  in 
these  by  the  niolybdate  method  :  see  scheme  IX. 

Blast  furnaces  capable  of  producing  300  tons  of  pig  iron  per 
day  are  becoming  the  rule  rather  than  the  exception,  while  an 
ontput  of  400  tons  in  twenty-lour  hours  is  often  reached.  To 
show  the  amount  of  material  required  every  twenty-four  hours 
to  keep  such  a  furnace  running,  we  will  assume  as  follows : 

>  Slag  nude  duriDg  Itae  ruD  of  KWct  furnace,  on  mixture  conuiniag  EaterpiiK  oie 
1  Slajc  made  (I  Uic  Slou  tumace  in  June,  iSM,  on  Mo.  i  fouodry  iron.    (C«uall 
"Truuactiooiot  Amciit:*!!  InsUlnie  of  Miolog  Ka(iJi«ef«,"  Vol.  XVI,  p.  14S). 




Height  of  furnace,  eighty  feet;  internal  diameters  at  the 
hearth,  bosh  and  stockline  respectively  fourteen,  twenty  and 
seventeen  feet ;  cubical  contents,  about  22,000  cubic  feet.'  To 
produce  one  ton  of  iron,  would  require,  approximately,  160,000 
cubic  feet  of  air,  engine  measurement,  which  would  be  at  the 

rate  of  33»333  cubic  feet  per  minute (i^ — !29?^_3^  =  33,333.  ) 

\       24  X  60  / 

To  deliver  this  quantity  of  air  a  blowing  power  of  not  less 
than  2,000  horse-power  should  be  available  and  200  horse-power 
more  is  required  to  hoist  the  stock  and  pump  the  water  needed 
for  cooling,  etc.  The  blast  should  be  heated  from  1200°  to 
1400**  F,  and  for  this  purpose  four  regenerative  stoves  twenty 
feet  in  diameter  and  seventy  feet  in  height  are  employed. 
These  stoves  contain  about  48,000  cubic  feet  of  fire-brick,  and 
are  kept  at  such  a  temperature  as  will  heat  the  blast  to  the  de- 
sired degree,  by  burning  in  them  the  waste  gases  of  the  fur- 
nace. If  we  assume  the  ore  smelted  to  contain  sixty  per  cent, 
of  iron  and  twelve  per  cent,  of  silica,  it  will  require  one  and  six- 
tenths  tons  of  ore  and  0.4  tons  of  flux  to  make  a  ton  of  iron, 
assuming  that  two  per  cent,  of  the  silica  be  reduced  and  alloyed 
with  the  pig  iron.  It  will  further  require  one  ton  of  fuel*  to 
make  a  ton  of  iron*  which,  containing  ten  per  cent,  of  ash,  will 
require  an  additional  amount  of  0.15  ton  of  flux.  Thus,  for  one 
ton  of  iron  is  required   1.6  +  0.40  + 0.15  +  i.  =  3.15  tons  of 

solid  material  and '- =5.81  tons  of  air.     In  one  day 


therefore,  there  would  be  300  X  8.96=2688  tons  of  material 
passing  through  such  a  furnace.  Supposing  the  flux  to  be  car- 
bonate of  lime,  and  to  contain  two  per  cent,  of  silica  and  one  per 
cent,  of  alumina,  the  furnace  would  produce  (0.55 — 0.03  X 
0-55)  0-56  +  0.03  X  0.55  +  1 .6  X  o.oi  +  I  X  o.oi  =0.57526  ton 
of  slag  per  ton  of  pig  iron,  or  0.57526X300=172.5  tons  per 

1  '•  The  Modern  Blast  Furnace,"  K.  A.  Uehling:.  Sleven*s  Indicator,  8,  p.  17. 

>  Well  equipped  and  well  managed  furnaces  using  "  lake  ores"  are  making  a  ton  kA 
iron  (2,240  Ihs.)  with  i.Hoo  lbs.  of  coke,  and  in  some  instances  the  fuel  consumption  has 
been  as  low  as  1.600  lbs.  of  coke. 


Summing  up : 

Material  charged  into  the  blast  furnaces  per  day  : 

Ore 480  tons. 

Coke 300 

Flux 165 


*  • 

t  • 


945  tons. 
BUst 1743     ** 

Toul a688    - 

Tapped  from  bottom  of  furnace  in  molten  state  : 

Pig  Iron 300  tons. 

Slag  172.5 

Total  molten  product 472.5 

Gaseous  product  passing  out  at  top  of  furnace  : 

Total  blast 1743.00  tons. 

Oxygen  from  ore 144.00    •• 

Gasified  Carbon,  as  CO  and  CO,.   246.00    ** 

Carbon  dioxide  from  flux    70.42    ** 

Volatile  matter  in  ore  and  fuel 12.08    ** 

Total  gaseous  product 2215.50  tons. 

Thus  it  is  shown  that  of  the  material  charged  into  a  blast  fur> 
nace  somewhat  less  than  sixty >five  per  cent,  is  gaseous,  while 
over  eighty  per  cent,  passes  off  in  the  form  of  gas. 

In  addition  to  the  2215.5  ^<>°s  of  gas  there  must  be  added  an 
equal  weight  of  air,  or  nearly  so,  since  considerable  excess  is 
required  for  combustion. 

Thus  the  chimney  of  a  300-ton  blast  furnace,  when  in  full 
operation,  discharges  into  the  atmosphere  ever>'  twenty-four 
hours  about  4,450  tons  of  gaseous  material,  which  is  at  the  rate 
of  over  three  tons  per  minute.  The  heat  energy  developed  is 
enormous.  In  twent>'-four  hours  fully  7,500,000,000  heat  units 
are  generated,  which,  if  utilized  in  a  first-class  steam  plant, 
would  develop  over  13,000  horse  power.  The  average  amount  of 
solid  and  molten  material  contained  in  a  300-ton  furnace  is  prob- 
ably not  far  from  900  tons.  The  temperature  varies  from 
3000**  F,  in  the  hearth,  to  300"*  at  the  stockline.  If  the  heat 
varied  regularly  the  average  temperature  would  be  1650'  F  ;  but 
since  the  stock  becomes  denser  as  it  gets  lower  in  the  furnace, 
and  also  since  a  red  heat  reaches  quite  high  up  in  the  furnace, 


2,000'  is  probably  nearer  the  average  temperature  of  the  whole. 
The  specific  heat  of  such  a  conglomerate  is  not  definitely  known, 
but  it  will  be  between  two-tenths  and  three-tenths  :  assume  it  to 
be  0.25.  Hence,  is  obtained,  for  the  heat  stored  away  in  the 
incandescent  furnace  stock.  900  X  2000  X  2000  X  0.25  =  908,- 
000,000  heat  units. 

The  lining  of  the  furnace  will  weigh  800,000  lbs ;  its  average 
temperature  will  not  be  less  than  800  degrees :  the  specific  heat 
of  fire-brick,  at  that  temperature,  is  about  0.18  :  therefore  the 
amount  of  heat  stored  away  in  the  lining  is  800,000  X  800  X 
0.18  =  115,000,000  heat  units. 

The  regenerative  stoves  contain  something  like  48,000  cubic 
feet  of  fire-brick,  which,  at  150  lbs.  per  cubic  foot,  would  make 
48,000  X  150  =  7,200,000  lbs.  The  average  temperature  of  the 
brick-work  in  these  stoves,  when  the  temperature  of  blast  is  car- 
ried at  1400''  may  be  taken  at'  looo""  and  the  specific  heat  of 
the  brick- work  at  that  temperature,  at  0.20.  Upon  this  basis, 
the  heat  stored  away  in  the  regenerative  stoves  amounts  to 
7,200,000  X  1000  X  0.2=  1,440,000,000  heat  units.  Thus,  in 
a  blast  furnace  of  300  tons  daily  capacity,  there  are  the  follow- 
ing quantities  of  materials  consumed  and  heat  units  developed  : 

Charged  into  the  furnace  : 

Solid  material  at  the  top 945  tons. 

Gaseous  material  (blast)  at  tuyeres 1743     ** 

ToUl  charged 2688     " 

Discharged  from  furnace  : 

Molten  material  from  hearth   472.5  tons. 

Gaseous  materials,  dust  and  fume 2215.5 


Total  discharge 2688.0    " 

Heat  energy  developed  : 

I'rom  fuel  consumed  in  twenty-four  hours,  7,500,000,000  heat  units. 

j  Stored  in  the  incandescent  material  )  «  ,  . 

,            ^  '      i  '    r  c     908,000,000  heat  units. 

y     contained  in  furnace >      ^ 

j  Stored  in  regenerative  stoves 1.440,000,000 

•  Total  heat  energy  stored *  2,348,000,000 

Thus  the  stored  energy  is  equal    to  ^^348,000,000  ^  778  = 


913,272,000  foot-tons  of  mechanical  energy. 




The    mechanical     energy    developed     during     twenty -four 

urs     in    the  process  of   smelting   is   7,500,000.000X778^:^ 

^(835. 000,000, 000      foot      pounds,      or      at       the       rate      of 


-  383  niile-lons  per  minute. 

t  X  60X  2000  X  5.2S0 

When  working  well,  a  blast  furnace  gives  but  little  evidence 
i  the  immensity  of  the  force  it  contains  :  it  is  only  when  ' '  run- 
ning off"  that  one  realizes,  in  a  measure,  what  a  monster  it  is. 
It  is  furthermore  quite  evident  that  the  process  must  be  continu- 
ous, twenty-four  hours  a  day  and  three  hundred  and  sixty-five 
days  in  a  year,  from  the  beginning  to  the  end  of  the  blast, 
which  may  last  from  six  weeks  to  as  many  years. 

When  in  good  condition  a  furnace  may  be  stopped-  for  twenty- 
four  or  even  forty-eight  hours,  without  serious  consequences, 
and  when  properly  prepared  may  be  "banked"  for  mouths  and 
started  up  again. 

■■  The  Charging  of  Blast  Furnaces. 

K  The  process  of  smelting  in  a  blast  furnace  is,  of  necessity,  a 
contiuuousoperalion.     The  raw  materials,  ore,  fuelandflux,  are 
charged  in  at  the  top,  keeping  the  furnace  practically  lull,  and 
the  molten  metal  and  slag  are  tapped  out  at  the  bottom  at  inter- 
vals as  required.  The  lime  necessary  for  a  charge  to  pass  through 
le  furnace  varies  from  ten  to  forty  hours  according  to  the  cubic 
intents  of  the  furnace,  the  character  of  the  ore,  and  the  relative 
[Uantity  of  air  driven  through  the  furnace   in  a  unit  of  time, 
.iiy  reducible  ores  require  less  time  than  tliose  of  a  refractory 
iture.     The  average  time  in  modern  furnaces  may  be  taken  at 
twenty  hours.     In  view  of  this  fact,  and  of  the  further  fact  that 
the  effects  of  bad  fillings  do  not  become  positively  manifest  until 
the  badly  proportioned  or  irregularly  distributed  charges  have 
entered  the  zone  of  fusion,  and  also  that  the  correction   for  such 
irregularity  can  only  become  effective  in  the  same  zone,  it  be- 
comes very  evident  that  serious  consequences  might  result  from 
bad  filling  before  the  remedy  could  have  had  lime  to  act. 

The  proper  charging  of  a  blast  furnace  is,  therefore,  of  the 
utmost    importance.     This  fact  has  long  ago  been  acquired  by 
ictical  experience,  and  the  success  of  blast   furnace  manage- 



ment  ver>'  largely    depends   on   the  proper  proportioning  and 
distribution  of  the  fuel,  ore  and  flux,  in  charging  the  furnace. 

Since  the  successful  running  of  a  blast  furnace  depends  more 
directly  upon  proper  charging  than  upon  any  other  one  thing, 
it  may  be  profitable  to  inquire  how  a  furnace  should  be  charged 
to  obtain  the  best  results.  To  do  this  we  must  study  the  chemi- 
cal reactions  as  well  as  the  physical  changes  which  take  place 
within  a  blast  furnace. 

The  first  requirement  is  heat,  which  must  not  only  be  sufficient 
in  quanity  and  intensity,  but  it  must  also  be  properly  distributed. 
a.  The  temperature  must  be  a  maximum  at  the  tuyere-line  and 
a  minimum  at  the  stoclc-line.  The  former  temperature  must  be 
higher  than  the  fusing  point  of  the  iron  and  slag,  and  the  latter 
should  be  below  the  point  at  which  carbon  dioxide  is  reduced  to 
carbon  monoxide  by  the  fuel.  b.  Each  horizontal  layer  of  the 
contents  should  have  practically  the  same  temperature  through- 
out its  whole  area.  c.  The  temperature  of  these  horizontal 
layers  should  be  fixed  at  fixed  heights. 

The  second  requirement  is  an  abundant  supply  of  an  efficient 
reducing  agent.  Since  all  the  sensible  heat  in  a  blast  furnace 
is  due  to  the  combustion  of  carbon  to  carbon  monoxide  at  the 
tuyeres,  except  that  brought  in  by  the  blast,  and  carbon  monox- 
ide, as  we  shall  presently  show,  being  the  most  desirable  reducing 
agent,  it  follows  that  if  the  first  requirement  is  fulfilled  the 
second  must  be  also. 

The  third  requirement  is  that  the  ore  and  flux  shall  be  so  pro- 
portioned and  mixed  that  the  impurities  of  the  former  will 
assimilate  with  the  latter  and  with  the  ash  of  the  fuel  and  form 
a  fusible  slag. 

Of  the  solid  material  charged  at  the  top,  over  fifty  per  cent. 
passes  off  in  the  form  of  gas — first,  by  the  evaporation  of  the 
hygroscopic  and  combined  water ;  second,  by  the  volatilization 
of  the  hydrocarbons  of  the  fuel  and  the  carbon  dioxide  of  the 
flux  and  air  ;  third,  by  the  reduction  of  the  ore,  the  oxygen 
combining  with  carbon,  forming  carbon  monoxide,  or  with  car- 
bon monoxide  forming  carbon  dioxide  ;  and  lastly,  by  the  oxi- 
dation of  the  carbon  of  the  fuel,  which  unites  with  the  oxygen 
of  the  blast,  forming  carbon  monoxide  at  the  tuyeres,  part  of 




,icli  in  its  upward  course  reduces  the  ore.  as  already  stated, 
fonniug  carbon  dioxide,  the  remainder  passing  off  as  carbon 
monoxide  with  the  other  gaseous  products. 

The  principle  object  in  view  in  operating  a  blast  furnace  is  to 
secure  the  best  possible  conditions  for  reducing,  carbonizing  and 
melting  the  iron  contained  in  the  ores  to  be  smelted,  and  to  do 
this  with  the  greatest  regularity  and  the  least  expenditure  of 
fuel  for  the  quality  of  iron  desired. 

The  chemical  phenomena  which  take  place  in  a  blast  furnace 
are  manifold  and  complicated,  and  not  altogether  understood; 
but  for  our  present  purpose  it  is  necessary  only  to  consider  the 
two  principal  reactions,  vis.  ■  "  Reduction"  and  "  Oxidation  ;" 
the  latter  always  generating,  and  the  fonner  absorbing  heat. 

One  pound  of  iron  in  being  reduced  from  its  ore  (Fe,0,) 
irbs  3.396  heat  units.'     Carbon   in  being  oxidized  to  car- 

in  monoxide  generates  4,466  heat  units,  and  when  another 
atom  of  oxygen  is  added,  forming  carbon  dioxide.  10,078  heat 
units  more  are  developed. 

The  reduction  of  ore  to  the  metallic  state  may  and  does  take 
place  in  three  different  ways ;'  First,  by  oxidizing  the  carbon 
at  the  tuyeres  to  carbon  monoxide  by  the  oxygen  of  the  entering 
blast.  The  carbon  monoxide  thus  formed,  taking  auotheratom 
of  oxygen  from  the  ore,  forming  carbon  dioxide,  which  passes 
oflt  as  such;  second,  the  carbon,  taking  direct  from  the  ore  two 
atoms  of  oxygen,  forming  carbon  dioxide  as  which  it  escapes  ; 
third,  the  carbon,  taking  directly  from  the  ore  only  one  atom  of 
oxygen,  passing  off  as  carbon  monoxide.  Although  it  is  true 
that  the  conditions  in  a  blast  furnace  do  not  permit  the  full  reali- 
zation of  the  first  two  modes  of  reduction,  it  is  none  the  less  a 
fact  that  all  three  modes  take  place  side  by  side  in  the  process  of 
smelting ;  and  it  now  remains  to  be  seen  which  one  is  the 
it  economical,  and  what  can  be  done  in  the  way  of  charging 
[furnace  to  realize  that  one  to  the  fullest  degree. 

One  pound  of  iron  (Fe)  in  the  form  of  hematite  ore  {Fe,0,) 
holds  -in  combination  three-sevenths  pound  of  oxygen.     One 

'  3'196  TcprcacDla  ihc  mcBDO<  Uulong's  Apdrcwii'  and  Pavrond  Silbernun'sespcri- 




pound  of  carbon  to  become  oxidized  to  carbon  monoxide  requir 
one  and  one-third  pounds  of  oxygen,  forming  two  and  one-third 
pounds  of  carbon  monoxide,  which  is  capable  of  combining  with 
one  and  one-third  pounds  more  of  oxygen,  resulting  in  three  and 
two-thirds  pounds  of  carbon  dioxide. 

The  heat  absorbed  in  reducing  one  pound  of  iron  from  its  ore 
is  not  afiected  by  the  mode  of  reduction,  being  in  each  case  3,396 
heat  units.     The  heat  generated  by  the  oxidation  of  the  cartx 
by  the  first  mode  is 

4466  :=  1435.5  ''eat  units  generated  at  the  tuyeres. 
10078  =  3339.5         "  "         inprocessof  reduclioi 

Total.     4675 

absorbed  " 

leaviug     1379         "  surphis. 

By  the  second  mode  of  reduction  we  have 

f  X  i  X  :4544  =  ^337  lieal  units  generated. 
3396  "  absorbed, 

leaviug       1059  "  deficiency. 

By  the  third  mode  of  reduction  we  have 

•I  X  j  X  4466  ^  1436  heat  units  generated. 
and,  as  before.  3396  "  absorbed. 




Thus  we  see  that  in  the  first  case  we  have  a  surplus  of  1279 
heat  units,  in  the  second  case  a  deficiency  of  1059  heat  units, 
and  in  the  third  case  a  deficiency  reaching  i960  heat  units,  mak- 
ing a  difference  between  the  first  and  third  cases,  for  the  : 
consumption  of  carbon,  of  3239  heat  units  in  reducing  one  pouti 
of  iron. 

Time  and  space  will  not  permit  at  this  time  to  point  out  s 
formulate,  quantitatively,  the  heat  requirements  in  addition  Ij 
that  absorbed  in  the  process  of  reduction,  nor  is  it  necessary  I 
our  present  purpose. 

The  total  carbon  requirement  varies  with  the  nature  1 
amount  of  impurities  contained  in  the  ore,  the  temperature  i 
hygroscopic  state  of  the  blast,  the  size,  shape  and  construct 



of  the  blast  furnaces,  etc.,  etc.;  for  our  purpose  it  will  suffice  to 
denote  the  heat  required  for  smelting,  in  addition  to  that 
absorbed  in  the  process  of  reduction,  by  the  symbol  X.  Sub- 
tracting from  this,  our  +  or  ^  surplus,  we  get  tor  the  addi- 
tional heat  required  to  complete  the  process  of  smelting  : 
t  First    case,  X  —  1179  heat  uuits. 

Second  "      X  +  1059  " 

Third  ■■  X -t- i960 
■  Now,  since  all  the  heat  required  to  satisfy  X  (excepting  what 
is  carried  in  by  the  healed  blast )  must  be  generated  by  burning 
carbon  at  the  tuyeres,  where  a  higher  oxidation  than  carbon 
luonoxide  cannot  result,'  we  have  only  4466  heal  units  available 
per  pound  of  carbon  consumed  by  the  blast. 

The  weight  of  carbon  consumed  in  reducing  one  pound  of 

iron,  according  to  the  first  mode  ol  reduction,  is  5  -|-  j  ^  ^j 

pounds  of  carbon,  resulting,  as  already  shown,   in  a  surplus  of 

1 279  heat  units,  available  to  meet  the  requirement  of  X,  which 

surplus,  expressed  in  weight  of  carbon,  is  equal  to      '     =  0.284 

pounds  burnt  at  the  tuyeres, 

^B     By  the  second  mode  of  reduction  we  have  S  X  »  —  i'e  pounds 

^^H  carbon  oxidized,  resulting  in  a  deficiency  of  1059  heat  units. 

^■K^e  have,  however,  in  this  case  consumed  only  one-half  as  much 

^Hitrbonasincase first:  raakingthisthesame,  wegetj^  X4466:ir7i8 

^^■eat  units  additional,  which  reduces  the  deficiency  to   1059 — 

^fpiS  =^  341  heat  units.     Expressed  as  above,  we  have  available 

for  X,  —  — f,r^^ — "-077  ptu'iid  of  carbon. 

Similarly  we  get  by  the  third  mode  of  reduction:  |X|=A 

lUnds  carbon  oxidized,  resulting  in  a  deficiency  of    196a  heat 

Its,  which,  expressed  in  pounds  of  carbon  available  for  X, 



carlioD  moooiidc  r 
lonned  first,  and  th 

socUdoDj.  carhon  dioidde  is  immrdla 
dioxide  is  fiix  formed,  it  can  only  be  of 
lunuce,  and  does  not  allrr  Ibe  effect. 

-0,443  pound  of  carbon. 


To  simplify  the  comparison,  let  us  suppose,  for  the  moment, 
that  X  is  completely  satisfied  by  the  first  mode  of  reduction,  in 


which  case  we  would  have  — —  —  0.284  =  o-oo  pound  of  carbon 

required  to  complete  the  process  of  smelting  one  pound  of  iron. 


By  the  second  mode  of  reduction  we  would  have   — 7^  — 


( — 0.077)  =  0.363  pound  of  carbon  required  to  complete   the 

process  of  smelting,  and,  according  to  the  third  mode  of  reduc- 


tion  we  would  have  —  - — ( — 0.443)  =0.727  pound  of  carbon 

required  to  complete  the  process  of  smelting  one  pound  of  iron. 

From  these  few  calculations  it  becomes  evident  that  it  is  a 
matter  of  vital  importance  which  mode  of  reduction  prevails  in 
the  furnace,  and  that  the  first  mode  should  be  sought  to  be  real- 
ized to  the  fullest  extent,  and  the  third  mode  should  be  avoided 
entirely  if  possible. 

The  first  mode  of  reduction  implies  that  no  direct  reduction 
shall  take  place,  /.  e.,  no  oxygen  shall  be  abstracted  from  the 
ferric  oxide  by  carbon,  which  necessitates  that  all  the  carbon 
shall  reach  the  zone  of  combustion  and  be  there  burned  to  car- 
bon monoxide,  which,  on  its  upward  course,  reduces  the  ore 
(Fe,0,)  and  becomes  itself  oxidized  to  carbon  dioxide.' 

Since  the  direct  combination  of  the  carbon  of  the  fuel  with  the 
oxygen  of  the  ore  can  only  take  place  by  actual  contact,  it  does 
not  require  much  meditation  to  arrive  at  the  conclusion  that 
such  contact  should  be  avoided  as  far  as  possible.  This  can  be 
done  to  a  limited  extent  in  charging  the  furnace. 

Calculation  of  Blast  Furnace  Slag. 

Assume  that  we  have  a  mixture  of  ores,  of  which  the  average 
composition  is :' 

1  This  mode  of  reduction,  the  desirability  of  which  was  first  scientifically  estebllshed 
by  Prof.  Gruner  in  his  Etude  sur  Us  Haute  Foumeaux,  cannot  be  fully  realised  in  Uie 
present  form  of  blast  furnace,  because  contact  between  fuel  and  ore  cannot  be  entirely 

'<<  Rossi  ij.  Ant.  Chem.  Soc.,  la,  321. 


Iron  Orb. 

SiO, 20.00  per  cent. 

A1,0, 3«> 

CaO 3.10 

MgO 2.60 

Fe,0, 70.00 

MnjOf 0.20 

PA 105 

SO, o.  I  o 

i«  «< 
<<  «« 
«<      <i 

i<      << 

Metallic  Iron  =  50.  per  cent. 


SiO, 6.00  per  cent. 

A1,0, 1. 15 

CaO 30.00 

MgO 19.00 

CO, 44.20 

a  ti 

<<  II 


11  II 

<i         i< 



Anthracite  coal  containing  6.28  per  cent  ash.    Of  which  ash,  the  com- 
position, in  the  coal,  in  per  cent,  is  : 

SiO, 3.35  per  cent. 

A1,0, 2.73    "      '* 

CaO    **      " 

MgO 0.10 

<i      (I 


<i      <• 

We  have  decided  to  obtain  a  slag  of  such  a  character  that  the 
fusibility  will  be  about  that  of  a  sesquibasic  slag,  that  is,  if  pre- 
ferred, of  one  in  which  oxygen  ratio  is  4:3.  Looking  at  the 
table'  we  see  that,  for  such  a  type,  one  of  lime  saturates  0.7 14  of 
silica,  or  one  of  silica  takes  up  i  .400  of  lime.  Assuming  any 
proper  amount  of  coal  per  ton  of  ore  smelted  and,  in  most  cases, 
0.75  ton  is  all  that  is  required,  we  have  all  the  data  necessary 
for  our  calculations.     Transform  all  Ihe  analyses  into  lime  : 

1  Sesaui-  Sesqui-  Tri-     Quad- 

CompoAiUon.  Acid.       acid.    Neutral,  basic.  Bibasic.  basic,  nbasic 

SiCVpc^  cent 68.19       61.65       51.72       41.66       14.88       26.30       21.13 

CaOpcrcent 31.81       38.35       48.28       58.34       65.12       73.70       78.87 

O  of  SiO,  :  O  of  bases 4:1       3:1       2:1       4:3       1:1       2:3       1:2 

iCaO  saturates  5<0« 2.143       1.6071      1.0713     0.714       0.538       0.357       0.268 

i&iO,        "         CaO 0.466       0.622       0.932       1.400       1.858       2.821)       3.73a 




SiO,= 20.00  per  cent. 

AljOs  =  3.20X  1.631 5.22 

CaO  =  3.10 3-IO 

MgO  =  2.60  X  1.40 3.64 

MdjO^  =  0.20 0.15 

«i      <( 

«4  I* 

It  i( 

CaO=i2.ii     '*      " 

The  ore  is  equivalent  per  ton  to 

SiO, 20.00  per  cent. 

CaO 12.11 



SiO, 6.00  per  cent. 

Al,08=  1. 15  X  1.631 1.87 

CaO  =  30 30.00 

MgO  =  19  X  1.40 26.60 

(<  (4 

<<  <C 

«<  l< 

(<  (( 

n  n 


The  limestone  is  equivalent  per  ton  to 

SiO, 6.00  per  cent. 

CaO CaO=58.47    **       ' 


SiOj 3.35  per  cent. 

Al203  =  2.73X  1.631 4.45     **      " 

CaO 0.10    **      •' 

MgO 0.14 


The  coal  is  equivalent  per  ton  to 

Si0.j 3.35  per  cent. 

CaO 4.69     '*      ** 

Hence,  as  we  use  only  three-fourths  of  a  ton  of  coal  per  ton  of 
ore,  the  coal  used  is  equiv^alent  to  three-fourths  of  the  above 
analysis,  or  : 

SiOj 2.52  per  cent. 

CaO 3.52 

SiO.  =  20 -f  2.52=3102 22.52 

per  ton  of  ore  ;  the  coal  and  ores  are  equivalent  to  CaO  =  12. 11 
+  3-52—  I5-63-  Since,  to  make  the  proper  silicate,  one 
of  lime  takes  up  0.714  of  silica,  the  15.63  of  lime  in  coal  and 
ores  will  take  up:  0.714  X  15.63=  11. 16  per  cent,  of  silica, 

<(     (» 

{(     <( 


leaving  of  free  silica  in  the  ore  and  coal  22.52  —  11.16  =  11.36 
silica  to  saturate  with  limestone.  The  six  per  cent,  of  silica  of 
the  stone  will  require,  at  the  rate  of  1.400  pounds  lime  per  pound 
of  silica,  6  X  1.40=  8.40  lime,  leaving  of  free  lime  or  the  equiva- 
lent in  the  limestone,  58.47  —  8.40  or  50.07 /r^^  lime.  We  have 
to  saturate  in  coal  and  ores,  11.36  free  silica.  At  the  rate  of 
saturation  adopted,  it  will  take:  11.36  X  1.40  lime  =  15.91 
lime  ;  as  50.07  free  lime  in  one  ton  of  limestone  requires  only 
15.91  of  lime  to  saturate  the  silica  in  coal  and  ores,  we  need 
only  per  ton  of  ore  and  three-fourths  ton  coal, 


==0.317  ton  of  limestone. 


The  charges  are  thus  :  one  ton  of  ore,  0.75  ton  of  coal,  0.317 
ton  of  limestone  and,  as  the  ore  contains  fifty  per  cent,  of  iron ^ 
wc  require  two  tons  of  ore,  one-and-a-half  tons  coal  and  0.634 
ton  limestone  per  ton  of  pig  made. 

The  composition  of  the  slag  is  : 

Silica  in  ore  and  coal  per  ton  of  or«  and  per  \  ton  of 

coal 22.526  per  cent. 

In  stone  6  X  0.317  ton 1.902 

((      t< 

Total  SiOj 24.428    "      '* 

Lime  in  J  ton  coal  and  i  ton  ore  (per  ton  ore)  15.63  '*  '* 
In  stone  0.318  ton  X  58.47  per  cent  =  18.59 

Total  lime 34-22 

equivalent  to 

SiO,  =  24.428  or  reducing  to SiOj  =  41.66  per  cent. 

CaO  =  34.220     percentage: CaO  =  58.34 

((      (t 

« (      <  • 

58.648  Total,  100.00 

exactly  a  sesquibasic  silicate. 

Using  the  preceding  charges  of  ores,  stone  and  coal,  we  should 
have  every  reason  to  expect  a  slag  of  the  above  composition  or 
of  one  very  close  to  it. 

We  have  adopted  one  and  one- half  tons  coal  per  ton  of  pig.  If 
greater  accuracy  were  necessary  the  preceding  calculations  could 
be  made  over  again  with  the  new  charges  in  coal ;  but,  practi- 
cally, it  is  absolutely  useless,  the  ash  of  coal  entering,  as  it  may 





Per  cent. 

Per  cent 









be  seen,  as  a  small  percentage  into  the  general  composition. 
With  inferior  cokes  or  anthracite  it  becomes  an  important  factor 
not  to  be  neglected  but  too  often  ignored.  Cokes  with  fifteefn 
per  cent,  of  ash  are  not  uncommon  in  certain  localities. 

As  an  example  of  the  close  coincidence  between  slags  actually 
run  from  known  calculated  charges  and  the  slag  determined,  a 
priori,  we  quote  the  following  slag  run  in  a  furnace  sixty  feet 
high,  sixteen  feet  bosh,  running  on  hot  blast  850**  to  900*  F. 
Pressure  of  blast,  seven  and  one-half  pounds,  American  furnace,, 
anthracite  coal.     The  analyses  of  materials  were  as  follows  : 

Per  cent. 

SiO, 23.31 

AljOj 4.51 

CaO 1.61 

MgO 3.41 

Alkalies 2.67 

Mn,04 traces 

PA  031 

S 0.08 

Making  the  calculations  proportionally  to  the  quantity  of  the 
different  materials  charged,  we  have  : 

Chargbs  : 

Iron  ore,    Dolomite,     Coal, 
924  lbs.       378  lbs.      588  lbs.         Total. 

Containing : 


Silica 215.38 

Alumina    41.67 

Lime i4-90 

Magnesia  3i'50 

Alkalies 24.67 

Mang.  oxide  Traces 

Amount  of  slag  ingredients 

Nine  hundred  and  twenty-four  pounds  of  ore  gave  in  iron  425 
pounds,  the  ores  having  46.60  per  cent.  iron.  With  such  slag, 
of  which  the  character  was  sesquibasic,  a  light  grade  of  iron 
was  to  be  expected,  such  pig  as  contains  in  an  average  1.50  per 
cent,  silicon  or  3.20  per  cent,  silica  corresponding,  in  425  pounds 
of  pig  iron,  to  13.60  pounds  of  silica,  which,  subtracted  from  the 
















■  •  •  • 

•  •  •  • 


•  •  •  « 

•  •  •  « 

.  • .  • 

. .    K.nK.tA 


total  silica  which  went  to  form  slag  and  pig,  leaves  a  balance  of 
256.84  pounds  silica  to  be  expected  in  slag.  The  composition 
of  the  slag  was  then  : 


I*b8.  Per  cent. 

SiO,    256.84  45-44 

Al,Ot 69.85  12.36 

CaO 121.33  21.40 

*  MgO 92.45  16.36 

Alkalies 24.67  4.40 

565.14  9996 

The  analysis  gave : 

Per  cent. 

SiO, 44.27 

AljOj    12.91 

CaO 20.00 

MgO 16.50 

Alkalies  3.98 

FeO 2.47 

MnO  Traces 

S 0.56 


This  quantity  of  iron  oxide,  2.47  per  cent.,  is  not  abnormal,  but 
occurs  in  many  slags.  If  we  take  it  into  consideration  in  calcu- 
lating the  slag  we  have  99.96  +  2.40=  102.36.  Reducing  to  a 
percentage  we  find  : 

Calculated  Slag  (iron  added).  Actual  Analysis. 

Per  cent.  Per  cent. 

SiO, 44.34 44.27. 

A1,0, 12.06 12.91. 

CaO 20.88 19.81. 

MgO 16.00 16.50. 

Alkalies 4.22 3.98. 

FeO 2.47 2.47. 

The  iron  was  found  to  be  No.  3  light  gray,  containing  1.53 
per  cent,  silicon.  Transformed  into  lime,  this  slag  corresponds 

Per  cent. 

SiO, 40.66 

CaO   59.34 










*,  R. 

*5      —  « 

i,  o 




5    1    = 

;S  I-  ? 

:;  '^  5 





r     a  cs  « 
c  ''• 



.■5     S 





•*    tea 

be  o  K 



Sf  CI 








a  ..  - 

i'    '    II 


..    es 
O  ^ 

fl    •• 

V    C8 



C8     O 

o  ^ 



..    es 
O  - 

35  O 

c   •• 


O  ^ 

es    O 

«   O 

be  ^ 

o  '-^ 



5  2 

^.  <; 



'^    r    V 

«  « 

-  •x; 

T    IT. 


X  «' 

t>.  PI 


c   : 

o     • 
X  .2 


o    • 

U    es    4, 



!7v    - 

r.  • 

O  • 

&  : 

a  • 

•"  • 

o  • 

X  « 


—  X 

a  ;/ 


a  ^3 


a  t/ 
v  a 

•^   •• 
»^  '^^ 


-.  ..  n 


•*  •  *      I 

a  •   •         I 

a  V 
y  a 


W    Ml    O 

«  •   • 

u    .    • 

3  :  • 

a  '   ' 


r"  ••• 


'^        *« 

.-i  •  a 
»r>  •  •»< 
v<  ,   » 

••   ■*  *  vO 

I"   ►»   1;  vC 

a  i:  a 

(<4     •<    4p' 

5ti  a 


>t  a  s 

£  3  s^ 

1  e 
a  « 




TJ  O  O    «J3 

^  ;; 

"x    3 

t:  be 

-.  V 

^  1?  **a 
■;  a  a  o 

a  '*-  e 

£  >^ 

„  >»•-  ^  a  a 

§  a  be  «  a  o 

'-  a      z  X  3 

SI   t^x 

a  a.  « 

5  2"-.  "S  >. 
C  a  be  a  a 
—  c  cs  a  «- 

a—  3   o* 

"—     X 

U.    ft; 

-  «8 

o  u  »  ^ 







•  «« 












—  mm  bi 

O  fl  "    «; 

C  8  bA  -S 

-  o  a  * 

O  u  »  ^ 

4;  a 

"250  =^ 

X     3 

a  u. 



fl  Ou 

.M  o  a 

o  o  « 

a*-  3 
I-  >»« 



-iqiluc  Method  for  Calculating  Blast  Pomace  Charges. 

be  xnle  consists  of  two  eqnal  scales  at  right  angles.  Fig.  9, 
o<  which  («)  is  fixed  to  a  small  board,  while  the  other  {h) 
ced  at  right  angles  to  «  npon  a  block,  c,  which  is  capable  of 
ng  motion  in  a  groove  parallel  to  a.' 


•'  i 


3     4     5     6 

7    8     9 

1           ;  , 

12    13    14    IS     16    17    18 


Fig.  9- 

he  point  A,  given  by  the  intersection  of  the  zeros  of  the 

es,  is  marked  upon  the  board,  and  from  it  a  line  AB  parallel 

he  groove  is  drawn.     With  A  as  a  centre,  lines  AC,  AD,  AE. 

also  drawn,  making  with  AB,  angles  whose  tangents  are 

H.  C.  Jenkins,  Iron  and  Sietl  Institute^  1S91. 


equal  to  the  ratios  between  the  weight  of  the  silica  to  weight  of 
base  in  the  respective  silicates  which  it  is  desirable  to  produce 
in  order  to  form  the  typical  fusible  slags  ordinarily  met  with  in 
blast  furnace  practice. 

The  lines  AC,  AD,  AE  are  marked  with  the  names  of  the 
bases  for  which  they  have  been  calculated. 

Thus  AC  makes  an  angle  of  28**  10'  with  AB — ^this  angle  having 
a  tangent  whose  value  is  0.5357,  which  is  the  ratio  of  the  atomic 
weight  of  silica  to  twice  the  atomic  weight  of  lime,  and  corres- 
ponds to  calcium  silicate  :  this  line,  therefore,  is  marked  **Lime.** 

Similarly  the  line  AD  makes  an  angle  of  36®  52'  with  AB,  the 
value  of  whose  tangent  is  0.75,  or  the  ratio  of  the  atomic  weight 
of  silica  to  the  atomic  weight  of  2  MgO  :  hence  it  is  marked 
**  Magnesia.'* 

Also  the  line  AE  is  at  an  angle  of  41**  25',  and  this  having  a 
tangent  corresponding  to  the  ratio  of  the  atomic  weight  of  3  SiO„ 
to  that  of  2A1,0„  makes  the  line  correspond  to  the  value  of  the 
component  parts  of  silica  and  of  alumina  in  aluminum  silicate, 
and  so  it  is  marked  **  Alumina.'* 

With  such  a  scale  it  is  a  ver>^  simple  matter  to  at  once  read  ofF 
either  the  excess  of  silica  in  any  ore,  or  else  the  amount  required 
to  properly  flux  off  the  earthy  bases  present. 

As  an  example,  take  a  spathic  ore  containing  : 


FeO  50  per  cent per  cent. 

MgO  3        ''         ...     2.25 

CaO     5        "         2.68  *« 

Al,03  3        "         265 

SiO,    3 
CO,    36 

Then  setting  the  movable  scale  b  against  3  on  the  fixed  scale 
a  and  looking  along  b  until  the  line  marked  ''Magnesia"  cuts 
it,  we  find  the  value  2.25  as  being  the  amount  of  silica  required 
to  satisfy  the  magnesia.  In  like  manner  is  found  the  amount 
(2.68)  of  silica  required  for  the  lime,  and  the  amount  (2.65)  for 
the  alumina  respectively  :  adding  all  these  together  we  find  a 
total  of  7.58  parts  of  silica  required  for  every  hundred  of  the  orie. 

But  as  there  are  already  three  parts  present,  every  hundred 


I  ^arts  of  the  ore  require  7.58 — 3.  =:  4.58  parts  of  silica  added  to 
flux  it. 

Due  allowance  is  also  made  for  the  ash  of  the  coke,  and  any 
small  quantity  of  sulphur  in  the  mixture.  In  the  treatment  of 
several  kinds  of  ores  to  be  smelted  together  they  should  be  mixed 
and  divided  into  three  classes,  one  having  /ess  and  another  more 
iron  than  is  required  in  the  final  charge,  and  one  should  be  add 
and  another  6asic  after  the  correction  for  the  ash  of  the  coke  is 
made,  or  one  of  these  three  maybe  a  limestone  or  a  siliceous  flux: 
it  need  not  necessarily  contain  iron. 

Then  let  it  be  required  to  have  «  parts  of  iron  per  hundred  of 
the  charge,  and  let  a„a„  a",  be  the  percentages  of  iron  in  the  ores, 

i  and  A,,  i„  6,  percentages  of  deficiency  (or  excess)  of  silica  in  the 
,same,  and  x.j/,  z,  the  number  of  parts  required  of  the  component 

.  -ores  per  hundred  of  the  charge 

X  (a,' +5,) 

e         («,  ±  *,) 

(I)  x+y  +  ^  = 


'■•  +  >".  +  ^a, . 

(3)  X*.  — ^-i,  zt  ?*.  —  O 
By  solving  these  simple  equations  there  is  obtained,  at  once. 
the  number  of  parts  of  each  component  required  to  satisfy  the 
conditions  ol  the  charge. 

If  it  is  desired  to  produce  a  more  acid  or  a  more  basic  slag, 
it  only  requires  that  the  scale  d  he  replaced  by  one  having  a 
['length  of  one-half  (.for  bi-siJicate  slag),  or  twice   (for  bi-basic 
Lslag)  that  of  the  normal  scale. 
Ktferences  : 

"  Noteon  the  DetermJitatioii  of  Silica  in  Blast  Furnace  Slag,"  by  P.  W. 

Shimer./.  ylm.  Chem.  Sac.  16,  501. 
"  Estimation  of  Metallic  Iron  in  Slag"  by  G.  Neumann,  Ztschr.  anal. 

Chem..  6,  680. 
"Bstimation  of  Phosphoric   Acid   in   Basic  Slags'   by   V.OHveriy. 

Anal.  Chem.  5.  415. 
"  The  Determiaation  of  Phosphoric  Aciil  in   Basic  Slags"  by  Adolph 
P.  JolUB.y.  Anal.  Chem..  6,  625. 



The  Analysis  of  Water  to  Determine  Scale-Forming 


The  scale-forming  ingredients  of  a  water  are  usually  composed 
of  calcium  and  magnesium  carbonates  and  calcium  sulphate,  and 
though  an  analysis  of  a  water  for  boiler  purposes  usually  states 
the  number  of  grains  per  gallon  of  the  above  constituents,  the 
analysis  should  also  comprise  the  determination  of  other  ingre- 
dients, not  scale  forming,  that  are  necessary  to  a  proper  estima- 
tion of  the  former.  This  is  especially  true  of  the  alkalies,  which 
are  not  always  determined  in  a  commercial  analysis,  of  water, 
for  boiler  purposes ;  the  amounts  of  lime,  magnesia,  chlorine, 
carbon  dioxide  and  sulphuric  acids,  being  considered  a  sufficient 
index  of  the  character  of  the  water. 

The  alkalies  and  their  salts  rarely  form  scale  in  boilers'  and 
so  cannot  be  classed  as  scale-forming,  yet  they  play  fully  as  im- 
portant a  part  in  the  relation  they  sustain  to  the  sulphuric  acid 
and  chlorine. 

If  all  the  sulphuric  acid  in  a  water  were  combined  with  the 
alkalies,  there  would  be  no  sulphate  of  lime  present,  and  the  lat- 
ter would  be  eliminated  as  a  part  of  the  scale  ingredients.  This 
is  a  condition  rarely  occurring,  however,  since  in  most  waters  a 
portion  of  the  sulphuric  acid  is  united  with  the  alkaline  earths 
and  the  alkalies.  The  indirect  estimation  of  the  carbon  dioxide 
would  be  changed  also.  That  is  to  say,  where  the  carbon  diox- 
ide is  estimated  by  uniting  all  the  lime  and  magnesia  (left  un- 
combined  with  sulphur  trioxide  and  chlorine),  with  carbon 
dioxide,  it  is  evident  that  if  all  the  sulphur  trioxide  is  united 

1  A  sample  of  Boiler  scale,  from  Charlestown  S.  C,  analyzed  by  the  author  in  1887 
had  the  followiug  composition  : 

Carbon i .01  per  cent, 

SiOa 1.52 

AlaOj 0.43 

NaCl 72.12 

CaCl  a 10.32    " 

KCl i.oi     •• 

MkCIo 1.71 

CaS04 11.20 

rndctermined 0.6S 

It  r« 

ti  I. 

14  it 

k  t 

li  (< 

it  t> 


<•         ti 


with  lime,  when  a  large  portion  belonged  to  the  alkalies,  the 

amount  of  calcium  carbonate  would  be  too  small,  and  also  that 

the  proportion  of  the  carbon  dioxide  would  be  deficient  by  the 

amount  required  to  saturate  the  lime  incorrectly  united  with  the 

sulphur  trioxide.     There  is  nothing  in  the  usual  commercial 

analysis  to  indicate  whether  the  sulphuric  acid,  as  determined  in 

the  water,  is  all  united  with  the  lime  to  form  calcium  sulphate  or 

not :  but  the  custom  has  been  so  to  unite  it,  with  the  result  that 

calcium  sulphate  may  be  represented  as  a  large  component  of 

the  scale-forming  material,  when,  in  reality,  none  whatever  may 

be  present. 

In  a  report  of  a  partial  analysis  of  the  Monongahela  River 

water,  {Transactions  Atner,  hist,  Minijig  Engineers^  Vol.  XVII, 

P-  353)  >  the  amount  of  objectionable  substances,  for  boiler  use, 

are  given  as  follows : 

Total  lime 161  parts  per  100,000  =   94  grains  per  gallon. 

••      magnesia    33      **       "        "       =    19      "        " 

Sulphuric  acid.  210      **       *'        "       =122 
Chlorine 38      '*       "        "       =   22 

((         i(         <( 

t (  <4  <  < 

It  further  states  the  amounts  of  carbonates  of  lime  and  mag- 

nesia  precipitated  upon  boiling  to  be  : 

Carbonate  of  lime 130  parts  per  100,000  =  76.    grains  per  gallon. 

"  magnesia.  21      "       "         "       =12.2       " 

The  alkalies  not  having  been  determined,  the  proportion  of 
sulphuric  acid  combined  with  them  becomes  problematical :  in 
fact,  the  inference  is  that  there  are  not  any  present,  when  in  all 
probability,  they  may  amount  to  a  large  percentage. 

For  this  reason  it  is  essential  that  the  alkalies  be  included 
in  the  analysis,  and  the  following  scheme  is  so  arranged  as  to 
include  them : 


1  Mil 



■8  lis' 

W     ►  ;  £  i* 


B  C""  —S^C'::?  ■3 B  iS £ " E^ii B'i'35 



5  Is; 


t  ■=■■->« 

!l  ii-ifeoi 

II  -I 

8  3= 






To  show,  in  detail,  the  method  of  using  the  scheme,  the  fol- 
wing   water  analysis  is  given.      (Preliminary   tests    having 

shown  the  water  to  contain  but  little  residue,  'eight  liters  of  it 

were  evaporated.) 

Platinum  capsule  and  residue  (8  liters) 147.460  grams. 

I"  ■■  without  residue 146.610  " 
Total  residue 0.840  " 
Before  ignition,  capsule  and  residne 147.460  " 
After  "  ........  147.197  " 
Organic,  volatile,  CO,,  etc.... o.l6j  " 
Crncible  +  SiO, 15.970  •■ 
Cmcible 15*904      " 

SiO, 0.066      ■' 

Solution  made  to  100  cc. — seveuty-five  cc.  for  bases,  twenty- 
6ve  cc.  for  SO,. 

Twenty-five  cc.  (SO,). 

Cmcible  and  BaSO, i6,03j  grama. 

Crucible 15-903      ' ' 


Seventy-five  cc. 

Crucible  +  FejOjf AljOj) 159338  grams. 

Crucible 15*903        ** 

Fe,0,  (AljO,) 0.0308 

Crucible  +  CaO 16.0197 

Crucible 15*903 


CaO 0.1167 

Crucible  -f  MgO 15*928 

Crucible 15*903        ** 


MgO 0.025 

Platinum  dish  +  alkaline  sulphates-hMgSO^    53*370 
Platinum  dish 53.197        ** 



Sulphates 0.173 

Dissolved  in  water,  made  solution  up  to  fifty  cc. :  twenty-five 
cc,  for  magnesia  determination,  and  twenty-five  cc.  for  potash 

Crucible  -f-  'M.g^Ffi^ 15*942  grams. 

Crucible 15*904      ** 

MgjPjO, 0.038      ** 

0.038   X    2  =  0.076    MgjPjOy. 

MgjPaO,  :  (MgS04)j  •  •  0.076  :  x 
MgS04  =0.082  gram. 

Potassium  platinic-chloride  on  tared  filters  =0.023  grani,  cor- 
responds to  0.0046  potassium  sulphate  in  the  fifty  cc. 

Having  determined  the  amounts  of  magnesium  and  potassium 

sulphates,  the  residue  remaining  is  sodium  sulphate,  as  follows  : 

Total  sulphates 0.173  grams. 

Magnesium  sulphate 0.082       ** 

Sodium  and  potassium  sulphates 0.091 

Potassium  sulphate 0.0046 


Sodium  sulphate 0.0864 

and  calculated  to  their  oxides  would  be  as  follows : 

MgO  =  0.027  grams  for  75  cc.  =  0.036 gram  for  100  cc.  or  the  8  liters. 
Na^O  =0.0377     ''        ••    75  "    =0.0502     "         •'       *'  ••  ** 

K,0    =0.0025     "         ••    75  "    =0.0034     "         "       **  **  ** 

The  chlorine  found  by  titration  amounted  to  0.0055  grani  per 
liter  or  0.32  grain  per  gallon. 



The  weights  thus  obtained  are  in  terms  of  the  total  residue, 

eight  liters,  and  are  converted  into  values  corresponding  to  one 

liter,  the  result  being  as  follows  : 

SiOa 0.0082  gram  per  liter. 

SO, 0.0206 








Organic,  etc< 


Oxygen  in  excess  of  CI 0.0016 

Total  residue 0.1046 

It  is  necessary  now  to  convert  these  values,  grams  per  liter, 
into  grains  per  gallon,  in  doing  which  the  following  table  is 

TABLE    SHOWING    THE    NUMBER    OF    GRAINS    PER    U.    S.    GALLON    (58,318 
grains)    and    imperial    GALU^N    (70,000    GRAINS)   CORRES- 

per  liter. 







Grains  per  Imperial 

Grains  per 
U.  S.  grallon. 


o.  1400 










12 0.8400 

13 0.9100 

14 0.9800 

15 • 1.0500 

16 1. 1 200 

17 1. 1900 

18 1.2600 

19 I  3300 

30 1 .4000 




1. 1080 
1. 1663 



Milli.  Grains 
grams  per 

per  Imperial 

liter.  gallon. 

21 1.4700 

22 1.5400 

23 I.6100 

24 1.6800 

25 1. 7500 

26 1.8200 



29 2.0300 





















40 2.8000 

41 2.8700 



























3- 7800 


56 39200 

57 3.9900 

58 4.0600 

59 4-1300 

60 4.2000 


U.  S. 









2.041 1 




















•  4.2700 

>  4.4800 

66 4*6200 




















86 6.0200 



























98 6.8600 

99 6.9300 

*  7.0000 
































The  amounts  obtained  being  : 


CI  .. 


CaO 1.1313 

CO, 0.7987 

Fe,0„  A1,0, 0.2973 

Organic 1-1254 

Oxygen  in  excess  of  CI  •  0.0932 

Total 6.0980 


0.4771  grains  per  gallon. 












« « 



« I 

1 1 




Having  determined  the  component  parts  of  the  water  residue 
in  grains  per  gallon,  it  becomes  necessary  to  unite  these  in  chem- 
ical union,  as  near  as  possible,  as  they  exist  in  the  water. 

The  general  rule  may  be  stated  as  follows :  The  chlorine  is 
combined  with  the  sodium,  if  still  in  excess,  then  with  the  po- 
tassium, magnesium,  and  finally  calcium.  The  sulphuric  acid 
to  the  alkalies,  provided  there  is  not  enough  chlorine  to  saturate 
them,  then  to  the  calcium,  and  finally  to  the  magnesium. 

The  carbon  dioxide  is  united  with  the  calcium  and  magne- 
sium after  the  other  combinations  are  made.  There  are  excep- 
tions to  this  rule,  mineral  waters  and  many  artesian  well  waters 
forming  notable  examples. 

Carrying  out  the  above,  the  following  is  obtained  : 

per  gallon. 




per  liter. 

NaCl 0.0091 

Na,S04 0.0033 

KjSO^ 0.0009 

CaS04 •  o  031 1 

CaCO, o.oi  18 

MgCO, 0.0162 

Fe,0„  A1,0, 0.0051 

SiO, 0.0082 

Organic,  etc 00193 

Total o.  1050 

1. 8136 
I. 1254 


This  analysis  shows  that  the  principal  scale-forming  ingre- 


dient  is  calcium  sulphate,  being  more  than  equal  to  the  calcium 
and  magnesium  carbonates. 

The  following  analysis  is  of  a  water  containing  sulphuric  acid, 
but  the  alkalies  being  present  in  sufficient  amount  to  combine 
with  all  of  it,  as  well  as  the  chlorine,  no  calcium  sulphate  is 
present : 

Gram  Grains 

per  liter.  per  gallon. 

SiO, 0.0038  0.2215 

SO, o.oiio  0.6414 

CI 0.0062  0.3615 

K,0 0.0033  0.1923 

Na,0 0.0185  1.0788 

MgO 0.0165  0.9624 

CaO 0.0466  2.7175 

AljOj.FejOj 0.0020  0.1166 

COj 0.0530  3.0908 

Organic 0.0246  i*4345 

0.1855  10.8173 

Oxygen  in  excess  of  CI 0.0021  0.1224 

Total o.  1834  10.6949 

Combined  as  follows : 

Gram  Grains 

per  liter.  per  gallon. 

NaCl 0.0154  0.8980 

Na,S04 0.0141  0.8223 

KjSOi 0.0061  0-3557 

CaCOa 0.0833  4.8577 

MgCO, 0.0338  1.9710 

AljOj.FejOs 0.0020  0.1166 

SiO, 0.0038  0.2215 

Organic 0.0246  i*4345 

Total o.  1831  10.6773 

Where  all  the  chlorine  is  not  in  combination  with  the  sodium 
and  potassium,  magnesium  chloride  is  usually  present. 

The  latter  compound,  while  not  scale-forming,  is  considered 
as  an  active,  corrosive  agent — upon  the  supposition  that  at  the 
temperature  of  100°  C,  and  higher,  it  is  decomposed,  and  hydro- 
chloric acid  formed  and  liberated.     Consult  Journal  Society  of 


Chemical  Industry,   Vol.   IX,  p.  472  ;    also    Treatise  on  Steam 
Boilers,  by  Wilson,  p.   168. 

The  report,  given  below,  is  of  a  water  from  a  driven  well  in 
Florida.  Complaint  having  been  made  that  not  only  was  the 
scale  excessive  in  amount,  but  that  corrosive  action  was  also  very 
marked,  an  analysis  was  made  ;  reference  to  which  readily  ex- 
plains the  diJBScult>'  encountered  in  the  boilers. 

Gram  Grains 

per  liter.  per  gallon. 

NaCl 0.323  18.83 

KCl 0.067  3.90 

MgCl, 0.104  6.06 

CaS04 0.197  11.48 

CaCO, 0.293  17.08 

MgCOa 0.144  8.40 

SiO, o.oii  0.64 

Al,0„Pe,Os  . . .' 0.007  0.40 

Organic 0.138  8.04 

Totel 1.284  74.83 

In  all  of  the  above  analyses  the  constituents  have  been  stated 
in  grains  per  gallon,  rather  than  in  parts  per  100,000,  the  former 
being  in  general  use  by  the  mechanical  profession  as  the  proper 
method  by  which  to  express  the  weights  of  the  component  parts 
of  the  residue  of  a  water. 

The  following  is  an  analysis  of  boiler  water,  in  which  no  scale 
-was  present,  but  where  corrosion  was  rapid.  Sample  was  marked 
**  Stand  Pipe  in  Boiler.'* 

NaCl 33'70  grains  per  gallon. 

KCl 2.26      •• 

Na,S04 16.33 

MgSOf 1926 

Pe,0,( suspended  particles)    5.64 
Fe,(NO,), 6.12 

Cu(NO,), 3.18 

Ca(NO,), 12.11 

Mg(NO,),-- 14.08 

Silica 14.16 

HNO,(free) 12.27 

Organic  matter 24. 12 

Undetermined 2. 15 

<  {  ( t  { I 
1  <  t  (  t  ( 
« i  <  {  <  t 
t  i  i  (  i  • 
It  1 1  1 1 

<  •  it  1 1 

<  1  <  t  1 1 
1 (  1 1  it 

t  { <  t  > 


it  1 1  it 

Total 164.38 

(I        i  <        It 


The  water  supplied  to  this  boiler,  also  acid,  was  composed  as 
follows : 

NaCl 0.73  grains  per  gallon. 

MgCl,(KCl) 0.87      '* 

MgSO* 1. 71       •* 

CaSO* 1.53      **        •* 

CaCNO,), 0.38      *• 

SiOj 0.52 

Fe,(NO,), 0.44 

HNOjCFree) 0.90 

Organic  matter 0.64 

«<  <(  (t 

•<  («  (i 

«(  *<  «( 

it  14  (( 

<«  t(  11 

Total  solids 7.72 

This  increase  of  total  solids  from  7.72  grains  per  gallon  in  the 
supply  water  to  164  grains  total  solids,  per  gallon  of  water  in  the 
boiler  shows  neglect  in  the  management  of  the  boilers. 

By  neutralizing  the  free  acid,  in  the  supply  water,  with  sodium 
carbonate,  corrosive  action  in  the  boiler  was  prevented. 

In  coal-bearing  districts  the  boiler  waters,  while  usually 
selected  with  care  regarding  the  total  solids,  often  contain  free 
sulphuric  acid,  derived  from  the  oxidation  of  the  iron  pyrites  in 
the  coal  beds  and  which  enter  into  the  water  supply. 

The  free  acid  in  water  can  be  determined  quantitatively  as 
follows : 

250  cc.  of  the  water  are  transferred  to  a  six-inch  porcelain 
evaporating  dish,  a  few  drops  of  litmus  solution  added,  and  the 
water  boiled  five  minutes,  then  titrated  with  a  very  dilute  stand- 
ard solution  of  caustic  soda. 


250  cc.  of  water  taken,  which  required  one  and  two-tenths  cc. 
of  the  caustic  soda  solution. 

The  caustic  soda  solution  is  of  such  a  strength  that  31.  i  cc.  of 
it  neutralizes  five  cc.  of  normal  sulphuric  acid  solution. 

One  cc.  of  the  normal  sulphuric  acid  solution  contains  0.049 
gram  sulphuric  acid. 

Then  one  cc.  of  the  dilute  caustic  soda  solution  corresponds 
to  0.00787  gram  sulphuric  acid. 

If  250  cc.  of  the  water  required  one  and  two-tenths  cc.  of  the 
caustic  soda  solution,  one  liter  will  require  four  and  eight-tenths 


<:c.  caustic  soda  solution  =  0.0377  Rram  sulphuric  acid,  corres- 
ponding to  3. 30  grains  of  free  sulphuric  acid  per  gallon  of  water. 
Determination  of  the  Hardness  of  Water. 

The  hardness  of  w 

r  both. 



water  may  be  temporarj',  permanent, 

The  former  is  caused  by  calcium  and  magnesium  carbonates 
and  which  are  held  in  solution  by  the  excess  of  carbon  dioxide  in 
the  water. 

Boiling  the  water  drives  out  the  excess  of  the  carbon  dioxide 
and  the  calcium  and  magnesium  carbonates  are  thereby  pre- 

The  permanent  hardness  is  usually  caused  by  calcium  sul- 
phate, which  is  not  precipitated  by  boiling,  or  by  magnesium 
chloride.  The  former,  however,  can  be  separated  out  of  boiler- 
feed  water  by  heating  to  240°  F.  since  at  230*  F,  it  is  insoluble. 
Temporary  Hardness. 

The  temporary-  hardness  is  determined  as  follows  :  Standard 
sulphuric  acid  solution  and  standard  sodium  carbonate  solution 
are  required  and  are  prepared  as  follows  : 

1 .96  grams  of  pure  ignited  sodium  carbonate  are  dissolved  in 
oneliter  of  distilled  water.  One  cc.=o.ooio6gram  corresponding 
too.ooi  calcium  carbonate.  The  standardsulphuricacidsolution 
is  made  of  such  strength  that  one  cc.  of  it  exactly  neutralizes 
one  cc.  of  the  standard  sodium  carbonate  solution. 

100  cc.  of  the  water,  to  which  a  few  drops  of  lacmoid  solution' 
have  been  added,  are  heated  to  boiling,  and  the  sulphuric  acid 
gradually  added  until  the  color  changes. 

Each  cc.  used  represents  one  part  of  calcium  carbonate  per 
looooo  parts  of  water,  or  il  it  be  desired  to  express  it  in  grains 
per  gallon,  the  result  in  parts  per  rooooo  is  multiplied  by  0.583. 
Permanent  Hardness. 

One  hundred  cc.  of  the  water  are  taken  and  an  excess  of  the 
sodium  carbonate  is  added  thereto,  generally  speaking  the  same 
volume  will  be  sufficient.  This  is  evaporated  to  dr>-ness  in  a 
platinum  dish,  and  the  soluble  portions  are  extracted  with  dis- 
tilled water  through  a  small  filter  and  the  filtrate  is  titrated  with 


the  Standard  acid  for  the  excess  of  sodium  carbonate ;  the  differ- 
ence represents  the  permanent  hardness. 

Reference. — Consult  Sutton's  volumetric  analysis,  p.  67. 

Determination  of  Hardness  by  the  Soap-Test. — (Phillips.) 

The  degree  of  hardness  of  a  water  is  determined  by  ascertain- 
ing the  amount  of  standard  soap  solution  necessary  to  form  a 
permanent  lather  with  a  definite  volume  of  the  sample  ;  the 
**  harder*'  the  water  the  more  soap  it  will  consume,  owing  to 
the  formation  of  insoluble  calcium,  magnesium,  etc.,  soaps 
(**  curd"),  brought  about  by  the  decomposition  of  the  soda  or 
potash  soap  added,  by  the  salts  of  the  alkaline  earths  present  in 
the  water. 

The  hardness  of  water  is  usually  expressed  in  terms  of  calcium 

Preparation  of  the  standard  solutions : 

First,  Solution  of  '*  Hard  Water."  Dissolve  i.ii  grams  of 
pure  fused  calcium  chloride  in  a  little  water,  and  dilute  to  one 
liter  at  is'^C,  or  dissolve  one  gram  of  pure  calcium  carbonate  in 
fifty  cc.  of  dilute  hydrochloric  acid,  evaporate  to  dryness,  dis- 
solve in  fifty  cc.  of  water,  and  dilute  to  one  liter.  In  either  case 
each  cubic  centimeter  of  the  solution  will  correspond  to  0.00 1 
gram  calcium  carbonate. 

Second,  Solution  of  Soap.  Castile  soap,  which  is  supposed 
to  be  made  with  soda  and  olive  oil,  is  much  used  for  standard 
soap  solutions,  but  it  has  been  found  liable  to  considerable  de- 
terioration on  keeping,  especially  in  cold  weather,  owing  to  the 
deposition  of  sodium  palmitate. 

Sodium  oleate  makes  a  standard  soap  solution  which  suffers 
very  little  change  on  keeping,  and  can  be  generally  recom- 
mended for  the  purpose. 

Thirteen  grams  of  it  are  dissolved  in  a  mixture  of  500  cc.  of 
alcohol  and  500  cc.  of  water,  and  filtered  if  necessary.  It  now 
becomes  necessary  to  standardize  it,  so  that  one  cc.  will  be 
equivalent  to  o.ooi  gram  of  calcium  carbonate.  In  order  to 
effect  this,  twelve  cc.  of  the  standard  hard  water  are  run  into  a 
250  cc.  stoppered  bottle  from  a  burette  and  diluted  to  58.3  cc. 
A  burette  is  now  filled  with  the  soap  solution  which  is  run  into 



the  bottle  one  cc.  at  3  time,  and  the  bottle  vigorously  shaken 
aher  each  addition,  until  a  point  is  reached  where  a  persistent 
lather,  lasting  for  at  least  five  minutes,  is  obtained.  Note  the 
volume  required.  Twelve  cc.  of  hard  water  should  require  thir- 
teen cc.  of  soap  solution  (distilled  water  itself  requiring  one  cc. 
to  form  a  lather),  but  it  will  be  a  figure  less  than  this,  and 
therefore  the  soap  solution  is  too  strong  and  will  require  diluting, 
so  that  twelve  cc.  of  standard  "  hard"  water  wnll  require  thir- 
teen cc.  of  the  soap  solution.  An  example  of  an  actual  prepara- 
tion of  a  standard  soap  solution  will  explain  this. 

Thirteen  grams  of  sodium  oleate  were  dissolved  in  a  mixture  of 
500  cc.  of  alcohol  and  500  cc.  of  water,  and  filtered.  On  testing 
in  the  manner  described,  twelve  cc.  of  the  standard  "hard" 
water  diluted  to  58.3  cc.  required  1 1.4  cc.  of  the  soap  solution  to 
(onn  a  persistent  lather. 

Now,  since  thirteen  cc.  should  have  been  required,  every  11,4 
cc.  of  the  soap  solution  left,  requires  diluting  by  13 — 11.4  = 
1.6  cc. 

There  were  ^60  cc.  of  the  solution  left,  therefore  "^ —  =  84.2. 


and  84.2  X  1.6  =^  134-7  '^^-  more  of  the  mixture  of  alcohol  and 
water  to  be  added.  On  adding  this  quantity,  thoroughly  mix- 
ing, aud  testing  as  before,  twelve  cc.  of  the  standard  hard  water 
required  exactly  thirteen  cc.  of  the  soap  .solution. 
Delermination  of  Total  Hardnas. 
58.3  cc'  of  the  clear  sample,  of  the  water  to  be  examined,  are 
run  into  a  250  cc.  flask,  and  the  standard  soap  solution  added  in 
the  manner  described  above,  until  a  lather  capable  of  persisting 
for  five  minutes  is  produced.  The  number  of  cubic  centimeters 
required,  minus  one  cc.  for  the  water  itself,  will  give  the  degrees 
of  hardness  in  terms  of  calcium  carbonate  in  grains  per  gallon. 
If  the  water  contains  a  fair  proportion  of  magnesia  salts,  there 
will  be  some  difficulty  in  obtaining  the  right  point,  owing  to  the 
slowness  with  which  magnesia  salts  decompose  soap  ;  an  appar- 

■I  gallon  conUln*  70.00 


ent  persistent  lather  is  formed,  which  on  being  allowed  to  stand 
a  little  while  and  again  shaken  up,  will  disappear ;  a  little  ex- 
perience with  magnesian  hard  waters  will  familiarize  the  operator 
with  this  peculiarity. 

The  Permanent  Hardness, 

250  cc.  of  the  water  are  poured  into  a  500  cc.  flask,  and  boiled 
for  one-half  hour,  the  original  volume  being  kept  up  by  frequent 
additions  of  boiling  distilled  water,  free  from  carbon  dioxide. 
After  cooling,  it  is  quickly  poured  into  a  250  cc.  graduated  stop- 
pered flask,  diluted  if  necessary  to  exactly  250  cc.  at  15"  C.  with 
distilled  water,  well  mixed  and  filtered.  58.3  cc.  of  the  solution 
are  now  poured  into  the  bottle  and  the  permanent  hardness  de- 
termined as  described. 

The  Temporary  Hardness, 

The  temporary  hardness,  or  that  hardness  removed  by  boiling, 
is  obtained  by  deducting  the  degree  of  permanent  hardness  from 
that  of  the  total. 

Standards  of  Hardness, 

The  French  standard  of  hardness  of  water  is  stated  in  terms  of 
milligrams  of  calcium  carbonate  in  100  grams  of  water,  or,  parts 
calcium  carbonate  per  100,000  parts  of  water. 

The  German  standard  represents  milligrams  of  lime  in  100 
grams  of  water,  or  parts  lime  per  100,000  parts  of  water. 

The  English  standard  represents  grains  of  calcium  carbonate 
per  gallon  of  70,000  grains. 

The  American  standard  represents  grains  of  calcium  carbon- 
ate per  gallon  of  58,381  grains.  The  French  standard  is  to  be 

Table  Showing  the  Relative  Hardness  of  the  Water  Suppi^ied  to 
Cities.    Determinations  made  by  A.  R.  Leeds. 

































Parts  per  100,000 









Grains  per  U.  S.  gallon 










The  Sanitary  Analysis  of  ^Vatcr. 
This  comprises  the  determination  of 

1.  Chlorine. 

2.  Free  and  albuminoid  ammonia. 

3.  Nitrates. 

4.  Nitrites. 

5.  Total  solids. 

6.  Organic  and  volatile  matter  by  ignition  of  residue. 

7.  Oxygen  required  to  oxidize  organic  matter. 

/.  DeUrminatioH  of  Chlorine.  Standard  Silver  Solution. 
Dissolve  five  grams  of  pure  crystallized  silver  nitrate  in  1000 
cc.  of  distilled  water.  One  cc.  of  the  solution  is  equivalent  to 
0.001  gram  chlorine.  If  the  water  to  be  tested  shows  by  qual- 
itative analysis  a  small  amount  of  chloride  present,  350  cc.  of 
the  water  should  be  evaporated  to  about  fifty  cc,  allowed  to 
cool,  three  drops  of  saturated  solution  of  potassium  chromate 
added,  and  the  silvernitrate  solutionHropped carefully  fromabu- 
rette  until  a  faint  permanent  red  color  is  produced  in  the  water. 
This  point  indicates  that  all  the  chlorine  has  combined  with  the 
silver,  and  that  any  additional  silver  solution  added  forms  sil- 
ver chromate.     Thus : 

ajo  cc,  of  the  water  used  for  exam iaal ion. 

"         "     "        "'     required  t. 3  cc.  silver  nitrate  solution. 
Iocx>cc.  ■■     ■■        ■■  ■■         5-^cc. 

Bqaivalent  to  0.0051  grams  of  chlonae  per  liter. 

■■  0.52  parts  chlorine  in  lou, 000  parts  of  the  water. 

It  is  customary  to  state  the  amount  of  chlorine  as  "  chlori 
as  chlorids" — as  NaCl.     Thus  : 

0.0053  gram  chlorine  per  liter  :=  0.0085  gram  sodium  chloride  per  liti 
0.51  parts  chlorine  per  too, 000  =  0,85  parts  soilium  chloride  per  :oo,ooi 

The  amount  of  chlorine  allowable  in  good  drinking  water  can- 
not be  stated  positively,  since  the  source  from  which  it  is 
derived  must  be  taken  into  account. 


Results  from  a  great  many  analyses  of  various  waters  would 
indicate  the  amount  allowed  as  follows  : 

Rain  water Traces  to  one  part  per  1,000,000. 

Surface  water One  to  ten  parts  per  1,000,000. 

Subsoil Two  to  twelve  parts  per  1,000,000. 

Deep  well  water Traces  to  large  quantity. 

2,  Free  and  Albuminoid  Ammo7iia. 

Solutions  required  are  : 

a.  Standard  solution  of  ammonium  chloride,  made  by  dis- 
sohnng  0.382  gram  dry  ammonium  chloride  in  100  cc.  of  ammo- 
nia-free distilled  water.  One  cc.  of  this  solution  is  diluted  to 
100  cc.  with  distilled  water,  each  cc.  of  the  latter  solution  cor- 
responding to  0.000012  gram  ammonia. 

b.  Standard  Nessler  Reagent. — Dissolve  seventeen  grams  of 
mercuric  chloride  (pulverized)  in  300  cc.  of  water,  and  thirty- 
five  grams  of  potassium  iodide  in  100  cc.  of  water.  Pour  the 
mercuric  chloride  solution  into  the  potassium  iodide  until  a  p>er- 
manent  red  precipitate  is  formed.  Add  a  twenty  per  cent,  solu- 
tion of  sodium  hydroxide  until  the  volume  of  the  mixed  solution 
amounts  to  one  liter.  Add  some  more  mercuric  chloride  solution 
until  a  permanent  red  precipitate  forms  and  allow  to  settle. 

c.  Alkaline  potassium  permanganate,  formed  by  dissolving 
eight  grams  of  potassium  permanganate  and  200  grams  of  potas- 
sium hydroxide  in  a  liter  of  distilled  water. 

This  solution  is  concentrated  by  boiling  to  about  750 cc,  then 
250  cc.  of  ammonia-free  water  is  added.  When  properly  pre- 
pared this  solution  gives  but  traces  of  ammonia  by  distillation. 
In  any  event,  however,  it  must  be  tested,  and  if  an  appreciable 
amount  is  found,  it  must  be  deducted  from  the  determination  of 
albuminoid  ammonia  in  any  sample  of  water  under  examination. 

Ammonia-free  water  is  made  by  distilling  water  acidulated 
with  sulphuric  acid. 


The  apparatus  shown  in  Fig.  1 1  is  well  adapted  for  this  pur- 

Place  250  cc.  of  the  water  to  be  tested  in  a  flask,  capacity 


one  liter,  add  one  cc.  saturated  solution  sodium  carbonate,  con-^ 
nect  with  the  coudeiiser  and  distil  until  no  reaction  for  ammo- 
nia is  shown  in  the  distillate,  (caught  in  one  of  the  comparison 
tubes)'  when  two  cc.  of  the  Nessler  solution  is  added  thereto, 
a  yellowish  brown  color  being  indicative  of  ammonia.  The 
apparatus  being  free  froni  ammonia,  500  cc.  of  the  water  are 
now  added  to  the  water  remaining  in  the  flask  and  one  cc.  of  the 
saturated  sodium  carbonate  solution  (free  from  ammonia)  added. 
Distillation  proceeds  until  three  distillates,  each  of  fifty  cc,  have 
been  received  in  the  comparison  tubes,  when  the  distillation  is 
stopped  and  the  heat  removed  until  the  distillates  can  be  exam- 
ined. The  comparison  tubes  are  protected  by  being  enclosed  in 
a  glass  vessel,  with  a  movable  top,  as  shown  in  Fig.  11,  at  the 
base  of  which  is  an  opening  filled  with  cotton  wool. 



ix'bei]  the  distillate  reaches  that  mark,  the  handle  of  the  stand 
containing  the  comparitor  tubes  is  turned  and  another  compari- 
tar  tube  placed  under  the  outlet  of  the  condenser.  The  revolv- 
ing stand  contains  seven  comparitor  tubes,  sufficient  for  both  the 
free  and  albuminoid  ammonia  determinations.  C.  H.  Wolff's 
colorimeter,  Fig.  12,  has  an  extended  use  in  water  analysis  for  the 
purpose  of  comparing  tints  of  color  of  the  water,  also  in  the  de- 
termination of  the  difference  in  color  in  the  estimation  of  free 
and  albuminoid  ammonia. 

One  of  the  tubes  contains  the  nesslerized  standard  ammonium 
chloride  solution,  the  other  tube  a  portion  of  the  water  distillate, 
I  nesslerized.  to  compare  with  the  former.  The  contents  of  the 
tube  containing  the  darker  liquid  are  partially  drawn  off  by 
means  of  the  glass  stop-cock  at  the  base,  and  the  remaining 
liquid  dUuted  with  distilled  water  until  a  uniform  tintof  color  is 
obtained  in  both  glasses.  As  these  tubes  are  graduated,  the 
calculations  are  simplified  and  rendered  more  expeditious. 
Ammonia  Determinations. 

The  first  fifty  cc.  of  distillate  is  now  tested  for  ammonia,  as 
follows  : 

The  tube  is  removed  and  placed  in  a  comparitor  and  two  cc.  of 
the  Nessler  solution  added.  Thecolorproduced  mustbematched 
taking  another  tul>e  and  filling  to  the  fifty  cc.  mark  with 
-free  distilled  water,  adding  two  cc.  Nessler  solution 
and  one  cc.  of  the  standard  ammonium  chloride  solution.  Allow 
to  stand  five  minutes  for  full  development  of  color,  then  compare 
the  color  of  the  liquids  in  the  tubes. 

If  the  solution  containing  the  ammonium  chloride  is  too 
Strong,  divide  it  and  add  distilled  ammonia-free  water  to  fifty  cc. 
mark  and  compare  again,  and  repeat  until  the  tints  are  identical. 

tf.  however,  the  solution  containing  the  ammonium  chloride 
is  not  deep  enough  in  color,  add  one  cc.  more  of  the  standard 
ammonium  chloride  solution  and  compare  as  before. 

The  second  and  third  distillate  are  treated  in  a  similar  man- 
ner, but  if  the  third  distillate  shows  over  a  trace  of  ammonia,  a 
fourth  distillate  must  he  taken,  or  until  no  appreciable  amount 
of  free  ammonia  can  be  obtained. 


Free  Ammonia. 

500  cc.  of  the  water  taken. 

Pint  distillate  '50  cc.^  repaired  1.5  cc.  ammonxam  chloride  solation. 

Second     "  50CC.,  0.3  cc. 

Third  50  cc. ,  none 

•  * 

Total  for  500  cc.  1.8  cc. 
••  1000  cc.  3.6  cc. 
Onecc.  ammoninm  chloride  solution  is  equivalent  to  0.00001  gram  nitro- 
gen, or  0.000012  gram  ammonia. 
Then  one  liter  of  the  water  contains  0.000043  gram  free  ammonia. 
Equivalent  to  0.0043  part  ammonia  per    100.000. 
••  0.0430    **  ••  •*  1.000.000. 

Fifty  cc.  of  the  alkaline  solution  potassium  permanganate  are 
added  to  the  contents  of  the  flask,  after  the  determination  of  the 
free  ammonia.  The  contents  of  the  flask  must  be  cooled  some- 
what  before  the  addition  of  the  alkaline  permanganate  solution. 
The  latter  is  placed  in  the  flask  by  means  of  the  glass  delivery- 
tutie,  which  passes  through  and  is  fused  to  the  glass  stopper  of 
the  flask.  By  this  arrangement  any  solution  can  be  added  to 
the  contents  of  the  flask  without  remo\nng  the  stopper. 

The  distillation  and  comparison  of  distillates  by  known 
amounts  of  ammonium  chloride  solution  are  made  in*  the  same 
manner  as  for  the  determination  of  free  ammonia. 

Albuminoid  Ammonia. 

Iffi  re.  of  the  water  taken. 

Fir»t  distillate  required  3.2  cc.  ammonium  chloride  solution. 

Second     '*  **        0.7  cc. 

(»  4t 

ToUl       "        3-9  cc. 
1000  cc.       ••        5.2  cc. 
Kquivalent  to  0.000063  gram  ammonia  per  liter. 

"  0.0052  part  ammonia  per    100,000  parts. 
*•  0.0520     ••  ••  •*  1.000,000      *• 

It  must  be  remembered  that  the  free  ammonia  was  determined 
in  the  50c^cc.  of  water  after  the  free  ammonia  was  expelled  from 
the  250  cc.  of  water  first  placed  in  the  flask. 

As  the  albuminoid  ammonia  is  not  developed  until  the  addi- 
tion  of  the  alkaline  permanganate  solution,  the  determination  of 
the  albuminoid  would  be  upon  750  cc.  of  water,  as  above  stated. 



Tbe  atoMnits  of  free  and  albuminoid  ammonia  allowable  in 
%ood  drinking  water  are  thus  stated  by  Wanklyn  :  "  If  a  water 
yield  o.oo  part  of  albuminoid  ammonia  per  million,  it  may  be 
passed  as  organically  pure,  despite  of  much  (ree  ammonia  and 
chlorides;  and  indeed  if  the  albuminoid  ammonia  amounts  too. 02, 
or  to  less  thano.05  parts  per  million,  tbe  water  belongs  to  theclass 
of  very  pure  water.  When  the  albuminoid  ammonia  amounts 
to  0.05 ,  then  the  proportion  of  free  ammonia  becomes  an  element 
in  the  calculation,  and  I  should  be  inclined  to  regard  with  some 

suspicion  a  water  yielding  a  considerable  quantity  of  free  ammo- 
nia along  with  more  than  0.05  part  of  albuminoid  ammonia  per 
million.  Free  ammonia,  however,  being  absent,  or  very  small, 
a.  water  should  not  be  condemned  unless  the  albuminoid  ammo- 
nia reaches  something  like  part  per  million.  Albuminoid 
ammonia  above  0,10  per  million  begins  to  be  a  very  suspicious 


sign ;  and  over  0.15  ought  to  condemn  a  water  absolutely.  The 
absence  of  chlorine  or  the  absence  of  more  than  one  grain  of 
chlorine  per  gallon,  is  a  sign  that  the  organic  impurity  is  of 
vegetable  rather  than  of  animal  origin,  but  it  would  be  a  great 
mistake  to  allow  water  highly  contaminated  with  vegetable  mat- 
ter to  be  taken  for  domestic  use.** 

The  apparatus  for  the  determination  of  free  and  albuminoid 
ammonia,  used  by  the  New  York  City  Health  Department,  is 
shown  in  Fig.  13,  a  description  of  which  will  be  found  in  the 
Jotimal  of  the  American  Chemical  Society^  16,  871. 

J.  Determination  of  Nitrates  by  the  Phenol  Method, 

a.  Standard  potassium  nitrate  solution,  formed  by  dissolving 
0.722  gram  potassium  nitrate,  C.  P.,  in  a  liter  of  water.  One 
cc.  of  this  solution  is  equivalent  to  0.00044  NO,. 

b.  Phenolsulphonic  acid,  formed  by  adding  three  cc.  of  water, 
six  grams  pure  phenol  and  thirty-seven  cc.  of  concentrafted  sul- 
phuric acid  together. 

The  operation  of  determining  the  nitrate  is  as  follows  : 

Twenty-five  cc.  of  the  water  are  evaporated  to  dryness  in  a 
No.  2  porcelain  capsule,  on  a  water  bath.  One  cc.  of  the  phenol- 
sulphonic acid  is  added  and  incorporated  thoroughly  with  the 

Add  one  cc.  water,  three  drops  of  concentrated  sulphuric  acid 
and  warm.  Dilute  with  twenty-five  cc.  water,  make  alkaline 
with  ammonium  hydroxide  and  make  solution  up  to  100  cc.  with 
water.  If  an  appreciable  amount  of  nitrate  is  present,  it  forms 
picric  acid  with  the  phenolsulphonic  acid,  imparting  a  yellow 
color  to  the  solution,  when  the  ammonia  is  added  by  the  forma- 
tion of  ammonium  picrate.  The  intensity  of  the  color  is  pro- 
portional to  the  amount  of  ammonium  picrate  present. 

One  cc.  of  the  standard  potassium  nitrate  solution  is  evapo- 
rated in  a  porcelain  capsule,  treated  as  above,  and  the  solution 
made  up  to  100  cc.  The  two  solutions  are  placed  in  comparitor 
glass  tubes  and  distilled  water  added  to  one  or  the  other  until 
the  colors  agree  in  tint.  Suppose  twenty-five  cc.  of  the  original 
water  after  treatment  and  subsequent  dilution  to  100  cc.  corres- 
ponded in  color  to  the  standard  solution  of  one  cc,  which  after 


trealment  and  dilution  to  too  cc,  was  diluted  to  200  cc.  Then 
twenty-five  cc.  of  the  original  water  contained  0.00005  gram 
nitrogen,  or  1000  cc.  contained  0.0020  gram  nitrogen  or  0,009 
gram  XO,  per  liter,  corresponding  to  0.52  grains  per  gallon,  or 
0,9  part  per  100,000,  or  9,0  parts  per  1 ,000,000. 
4.  Nitrites. 

Glosway's  modification  of  Griess's  method  is  to  be  recom- 
mended for  simplicity  and  accuracy. 

Th«  solutions  required  are  : 

1.  Sulpha uilic  acid.  Dissolve  one  gram  in  300  cc.  of  acetic 
acid  (sp.  gr.  1.04). 

2.  Sodium  nitrite.  Formed  by  dissolving  0.272  gram  silver 
nitrite  in  100  cc.  water,  adding  a  dilute  solution  of  sodium 
chloride  in  slight  excess,  and  diluting  to  250  cc. 

Take  100  cc.  of  this  solution,  dilute  to  one  liter  for  use.     One 
.  —  o.ooooi  gram  nitrogen. 

3.  a-amido-naphtha!ene  acetate. 
Two-tenths  gram  of  naphtbylamJne  is  boiled  with  forty  cc.  of 

water,  filtered  and  diluted  to  400  cc. 

Process  0/ Determination. 

Twenty-five  cc.  of  water  are  taken  and  placed  in  one  of  the 
color  comparitors,  two  cc.  of  the  sulphanilic  acid  and  two  cc.  of 
the  a mido- naphthalene  acetate  are  added.  If  nitrites  are  pres- 
ent, a  pink  color  is  produced,  which  must  be  compared  with  the 
colorproducedby  onecc.  of  the  standard  nitrite  solution,  to  which 
two  cc.  of  the  sulphanilic  acid,  two  cc.  of  the  ami  do-naphthalene 
acetate  and  twenty-five  cc.  of  pure  distilled  water  (free  from 
nitrites)  are  added. 

Suppose  twenty-five  cc.  of  the  water  required,  six-tenths  cc. 
of  the  standard  nitrite  solution,  or  0.000006  gram  nitrogen,  or 

10.00002  gram  NO,. 
Corresponding  to  o.oooS  gram  NO,  per  liter, 
■'  "  0.0466  grain  per  gallon, 

"  "  0.0800  part  per  100.000. 

"  0.8000  part  per  1,000,000. 
5.  The  total  solids  are  determined  by  evaporating  500  cc.  of 



the  water  in  a  platinum  disb  and  drying  the  residue  at  105**  C. 
to  constant  weight.  The  amount  obtained  multiplied  by  2 
equals  the  weight  per  liter. 

6,  The  organic  and  volatile  matter  is  approximately  deter- 
mined by  igniting  the  weighed  residue  until  all  carbonaceous 
matter  is  consumed,  and  weighing ;  the  difference  between  the 
weight  of  the  total  solids  and  the  weight  after  ignition  is  the 
volatile  and  combustible  matter. 

7.  Oxygen  required  to  oxidize  the  organic  matter  in  the  water. 

Solutions  required  : 

Standard  potassium  permanganate  ^  formed  by  dissolving  0.395 
gram  potassium  permanganate  in  1000  cc.  water.  Each  cc.  con- 
tains o.oooi  gram  available  oxygen. 

Potassium  Iodide  Solutio7i.  100  grams  of  the  pure  salt  dis- 
solved in  1000  cc.  water. 

Dilute  Sulphuric  Acid  Solution,  One  part  by  volume  of  pure 
sulphuric  acid  is  mixed  with  three  parts  by  volume  of  distilled 
water  and  solution  of  potassium  permanganate  dropped  in  until 
the  whole  retains  a  very  faint  pink  tint,  after  warming  to  80**  F. 
for  four  hours. 

Sodium  Thiosulphate.  One  part  of  the  pure  crystallized  salt  in 
1000  parts  of  salt. 

Starch  hidicator  is  made  by  mixing  six  grams  of  starch  with 
100  cc.  pure  glycerine,  heating  for  one  hour  to  100**  C,  pouring 
it  into  200  cc.  of  water,  then  adding  sufficient  strong  alcohol  to 
precipitate  the  soluble  starch,  which  is  filtered  o£f  and  preserved 
in  a  moist  pasty  state.  When  required,  a  minute  quantity  is 
taken  with  a  glass  rod. 

Determination  of  the  Oxygen  Absorbed, 

Two  separate  determinations  are  required,  viz,,  the  amount  of 
oxygen  absorbed  during  fifteen  minutes  and  that  absorbed  dur- 
ing four  hours.  Both  are  to  be  made  at  a  temperature  of  27**  C. 
It  is  most  convenient  to  make  these  determinations  in  twelve 
ounce  stoppered  flasks,  which  have  been  rinsed  with  sulphuric 
acid  and  then  with  water.  Put  250  cc.  of  the  water  to  be  tested 
into  each  flask,  which  must  be  immersed  in  a  water  bath  or  suit- 
able air  bath  until  the  temperature  rises  to  27**  C.     Now  add  to 


ich  flask  ten  cc.  ol  the  dilute  sulphuric  acid,  and  then  ten  cc. 
'o(  the  standard  permanganate  solution.  Fifteen  minutes  after 
the  addition  of  the  permanganate,  one  of  the  flasks  must  be 
.Ttmoved  from  the  bath  and  two  or  three  drops  of  the  solution  of 
^potassium  iodide  added  to  remove  the  pink  color.  After  thor- 
ough admixture,  run  from  a  burette  the  standard  solution  of 
ithiosulpliate  until  the  yellow  color  is  nearly  destroyed,  add  some 
of  the  starch  solution  and  continue  the  addition  of  the  thiosul- 
phate  until  the  blue  color  is  just  discharged.  It  the  titration 
bas  been  properly  conducted,  the  addition  of  one  drop  of  per- 
manganate solution  will  restore  the  blue  color.  At  the  end  of 
four  hours  remove  the  other  flask,  add  potassium  iodide  and 
titrate  with  thiosulphate,  as  just  described.  Should  the  pink 
■color  of  the  water  in  the  flask  diminish  rapidly  during  the  four 
tours,  further  measured  quantities  of  the  standard  solution  of 
permauganate  must  be  added  from  time  to  time  so  as  to  keep  it 
markedly  pink.  The  thiosulphate  solution  must  be  standard- 
ized, not  only  at  first,  but  (since  it  is  liable  to  change),  from 
time  to  time  in  the  following  way  ;  To  250  cc.  of  pure  distilled  water 
add  two  or  three  drops  of  the  solution  of  potassium  iodide,  and 
ten  cc.  of  standardized  solution  of  permanganate.  Titrate 
with  thiosulphate  solution  as  above  described.  The  quantity 
nsed  will  be  the  amount  of  thiosulphate  solution  corresponding 
ten  cc.  of  the  standardized  permanganate,  and  the  factor  so 
'found  must  be  used  in  calculating  the  results  of  the  thiosulphate 
titrations  to  show  the  amount  of  standard  permanganate  solu- 
tion used,  and  thence  the  amount  of  oxygen  absorbed.  The 
amount  of  thiosulphate  solution  thus  found  to  be  required  to 
combine  with  the  iodine  liberated  by  the  permanganate. 
CosvBBSios  Tablh. 
Parts  per  X  0.7  ^=  Grains  per  UnperUl  Galloa. 
..  ..  X  0.07  = 
■•       •■        100.000     X  0.583      =         •■        '■         V.  S. 

■■       "    1,000,000    X  0.058     =        

'•       "     1,000.000     X  0,00833  =  Avoir,  pounds  per  1000  U.S.  Gal. 
Grains  "    Imp.  ^al.    -i-  0.7  —  Paris  per     100.000 

■■        ■•    U.  S.  ■■      +O.S8J      =       00.000 








1:51  L 






















su»|iq]iidon  i.iiapa 









5  ^^ 





S'S  a  3.  S  RS  5  *  K.ff  5  R  S.'S.'R  S  S  S 




!?■?■,?£  fti-a^B&ffaiEyS^KSffiSSRSif&SK^SRS'SaiS^i?' 

1  ?d 

aoninai  ua«so^ 






,¥sjpa^a8  8 

•3  .™i  m  npitos  IBIOJ, 
























If  ii 






"I""""  ".I'U'K 


=  ssa 


























hm ! ! 


ItHrl  i  i 





«??;?  ?  ? 



a:  :  :  : 







b:  :  :  ; 





nm  i  1 




fill  1 1 


:.$:;    S   S 






;i!Mr  ;  i 







Si  52 






in  ■,o«,ooo.      i| 



s      j 








:i::  : 

:t;  : 



-.9:  : 

:  ;^: 


■  t'i'- 


■■%■■  ■■ 


















:  : ;: ;J::-  : 




5i  ; 


Water  supplied  to  large  cities  is  usually  filtered  through  sam 
filter  beds. 

Fig.  14  shows  the  section  of  a  well-arranged  filter  bed  buiS 
for  the  city  of  Dublin.  The  bottom  of  the  filter  i 
posed  of  puddled  clay  three-fifth  meter  in  thickness  A.  built  in 
with  stones  one-fifth  meter  in  thickness.  The  next  layer,  thn 
quarters  meter  thick,  consists  of  coarse  angular  stones  I, 
then  fifteen  centimeters  of  smaller  stones  C,  followed  by  a  layer 
fifteen  centimeters  in  depth  of  coarse  gravel  0,  then  the  same 

depth  of  hue  gra\el  E   and  finallj  three  fourth  meter  of  sand  1 
To  collect  the  v.  aler  there  are  tw  o  channels  B  situated  half  in  the  " 
bed  of  claj  ind  half  in  the  stratum  of  large  stones      Each  clian 
nel  is  se^ent\  fi\e  centimeters  lu  width  and  sixty  centimeters  iit 
depth.     The  surface  of  sand  in  each  meter  is  sixty-one  by  thirty-' 
one  meters  ;  the  depth  of  this  water  is  sixty  centimeters. 

The   speed  of  filtration  varies  in  'the   existing  sand   filte 
from    one    and    four-tenths    to    fifteen    meters    per    twenty-fotO 
hours.     Each  water  requires,  if  it  is  to  be  well  filtered  by  a  givcafl 
sand,  a  determined  speed  of  filtration.     Thus,   under  otherwiaej 
similar  conditions,  three  and  five-tenths  cubic  meters  of  Tharae!^  ■ 


water  may  be  filtered  in  twenty-four  hours  per  square  meter  of 
filtering  surface,  but  only  one  and  seven-tenths  cubic  meters  of 
Elbe  water,  as  the  latter  contains  much  more  finely  divided  dirt. 
In  a  well  managed  filtration  the  turbid  water  passes  so  slowly 
through  the  sand  that  each  of  the  fine  particles  of  dirt,  though 
far  smaller  than  the  intervals  between  the  grains  of  sand,  has 
opportunity  to  attach  itself  to  one  of  the  grains.  Therefore,  the 
finer  and  more  numerous  the  particles  of  dirt  are,  the  finer  must 
be  the  sand  and  the  slower  must  be  the  rate  of  filtration.    If  this 

rate  is  too  great,  the  suspended  particles  flow  simply  through 
between  the  grains  of  sand.  But  if  the  sand  is  too  fine,  the  fil- 
ter bed  may  easily  become  water  light,  but  if  the  sand  is  too 
coarse  slower  filtration  is  to  some  extent  a  remedy.  The  best 
size  of  the  sand  grains  is  from  one-half  to  one  millimeter,  and  the 
sand  is  the  better  the  more  uniform  the  grains  are.  A  sand  con- 
taining much  finer  grains  cannot  be  used,  as  it  is  easily  ren- 
dered too  compact  by  the  pressure  of  the  water.  iVagttfr's 
Ckcm.   Tifh.,  p.   236. 


A  very  complete  description  of  the  sand  filter  beds,  constructed 
for  the  Massachusetts  Water  Works,  will  be  found  in  The  Engi- 
neering Record,  1895.  These  works  represent  the  latest  ad- 
vancement in  this  lineof  engineering. 

A  very  complete  article  on  "  PuriEcation  of  Sewage  and  of 
Water  by  Filtration,"  by  H,  F.  Mills.  C.E.,  will  be  found  in  the 
Transaclions  0/ the  American  Society  of  Civil  Engineers.  1894. 

To  show  the  methods  in  use  for  quick  filtration  as  well  as  the 
general  arrangement  of  the  apparatus,  the  Warren  filter  is  taken 
as  an  example.  The  filter  plant  usually  consists  of  a  settling 
basin,  one  or  more  filters,  and  a  weir  for  controlling  the  head, 
together  with  the  necessary  pipe  connections.  Each  filter  contains 
(see  Figs.    15   and   16)   a   bed   of  fine   sand,'  C   two   feet   in 

depth,    supported   by  perforated    copper    bottom,    B,    and   lor  J 
cleaning   this    bed  an  agitator,    D,    is   provided.      This  con- 
sists   of    a  heavy    rake    containing    thirteen     teeth     twenty- 


five  inches  long,  rotated  by  a  system  of  gearing,  A',  and  capa- 
ble of  being  driven  into  the  bed  by  means  of  suitable 
screw  mechanism,  L,  M,  whereby  the  entire  bed  is  thoroughly 
scoured.  The  process  of  filtration  is  as  follows :  The  water 
enters  the  settling  basin  through  a  valve  operated  by  a  float,  by 
which  a  constant  level  is  maintained  in  the  entire  filter  system. 
The  water  entering  through  this  valve  passes  through  an  eight- 
bladed  propeller  of  brass,  from  ten  to  sixteen  inches  in  diameter, 
so  arranged  as  to  revolve  freely  with  the  passage  of  the  water. 
This,  by  means  of  two  small  bevel  gears  and  an  upright  shaft, 
operates  an  alum  pump  of  unique  design,  consisting  of  six  hol- 
low arms  radiating  from  a  chambered  hub,  and  bent  in  the  direc- 
tion of  rotation.  This  pump  revolves  in  a  small  tank  containing 
a  dilute  solution  of  aluminum  sulphate,  or  other  coagulant,  and 
by  its  revolution  each  arm  takes  up  its  modicum  of  alum  water, 
passes  it  into  the  hub  and  to  the  deflector,  which  sends  it  down 
to  the  incoming  water. 

The  latter,  having  received  its  proper  amount  of  coagulant,  is 
then  allowed  to  remain  in  the  settling  basin  from  thirty  to  forty 
minutes,  to  enable  the  chemical  reaction  between  the  coagulant 
and  the  bases  and  organic  matter  in  the  water  to  take  place,  and 
to  permit  of  the  heavier  sediment,  together  with  a  portion  of  the 
coagulated  matter,  to  settle  by  subsidence  to  the  bottom  of  the 
tank,  where  it  can  be  drawn  off  at  inter\*als  into  the  sewer.  The 
water,  with  all  the  suspended  matter,  as  well  as  practically  all 
the  bacteria  present  in  the  water,  bound  and  held  together  by 
the  insoluble  hydrate  of  alumina  resulting  from  the  addition  of 
the  coagulant,  passes  on  through  suitable  piping  and  valves  to 
the  filter  A,  and,  filling  the  tank,  passes  down  through  the  fine 
zinc  sand  bed,  leaving  all  the  coagulated  matter  upon  it,  and 
makes  its  exit  from  the  filter  through  the  main  /,  bright  and 
clear  and  perfectly  adapted  in  every  way  for  domestic  purposes. 

The  main,  collecting  the  filtered  water  from  the  various  filters, 
passes  along  between  them  to  the  head  box,  or  weir,  over  which 
the  water  is  compelled  to  pass  and  which  controls  the  operation 
of  the  filters.  The  top  of  this  weir  is  twenty  inches  below  the 
water  level  maintained  in  the  filter  system,  and  this  head  of 
twenty  inches  (equivalent  to  a  pressure  of  three-quarters  of  a 


pound  to  a  square  iuch.)  is  the  extreme  pressure  that  can  be 
brought  to  bear  upon  the  filters,  and  it  is  evident  that  they  can 
at  no  time  be  pushed  beyond  the  rate  which  experience  has  shown 
to  yield  the  best  results. 

When  the  bed  o(  a  filter  becomes  clogged,  and  it  seems  best  to 

clean  it,  the  inlet  and  outlet  valves  EF.  are  closed,  and  the 
washout  G,  opened,  allowing  the  contents  of  the  tanks  to  escape 
to  the  sewer.  Fig.  16.  The  agitator,  D,  is  then  set  in  motion  by 
I  means  of  the  friction  clutch  with  which  it  is  equipped,  and  as  the 
I  teeth  on  therakebeginlo  plough  up  the  surface  of  the  bed  a  slight 
aniouot  of  filtered  water  is  allowed  to  flow  back  up  through  the 
bed.  in  order  to  rinse  off  the  dirt  loosened  by  the  rake.  This  is 
kept  up  until  the  rake  penetrates  to  the  bottom  of  the  bed,  and 
thoroughly  agitates  everj'  particle  of  material  therein. 

As  soon  as  the  water  following  to  the  sewer  is  clear,  the  motion 
of  the  rake  is  reversed  and  it  is  slowly  withdrawn  from  the  bed. 
Whenthe  teeth  are  raisedabove  the  bed,  the  water  pipe  is  closed 
:  inlet  valve   E  opened,  and   the  filter  tank  allowed   to  fill. 



After  waiting  a  few  minutes  the  outlet  valve.  F,  is  slowly  opened  ] 
andfiltrationisresumed.     A  filterten  feetsix  inches  in  diameter, 
net  area,  eighty-four  square  feet,  will  61ter  375000.  gallons  of 
water  per  twenty-four  hours. 

Bacteriological  Examination. 

The  bacteriological  examination  of  water  is  dependent  more 
upon  the  Microscopic  than  the  Engineering  Chemist. 

The  following  references,  however,  are  inserted  ; 

"  Micro-organisms  in  water"  by  Percy  and  G.  C,  Frankland,  1894. 

'■  Manual  of  Bacteriology,"  by  Dr.  George  M.  Sternberg,  1893. 

■■  A  Bacterial  Study  of  Drinking  Water."  by  Dr.  V.  C.  Vaughn,  1891. 

"  Bacteriological  Diagnosis,"  by  Dr.  James  Bisenberg.  Vienna,  1S87. 

"  Report  of  the  Masaachusetts  State  Board  of  Health  for  1891. 

"Bacteria  and  other  organisms  in  water"  by  John  W.Hill.  Transae,  I 
Amer.  Soc.  Civil  Engineers,  Vol.  ixiiii  pp.  423-467. 

"  Practical  Bacteriology"  by  Dr.  W.  Migula,  London,  1893. 
The  Composition  of  Boiler  Scale.' 

The  results  of  an  analysis  of  boiler  scale  usually  represent  the 
lime  and  magnesia  as  carbonates  with  a  portion  of  the  former  as 
sulphate — on  the  general  principle  thatthe  scale  made  continues 
to  exist  in  the  form  in  which  it  was  precipitated.  In  those  por- 
tions of  the  boiler  where  the  direct  heat  does  not  come  in  contact 
with  it,  the  scale  remains  unchanged  after  formation,  but  the 
conditions  are  altered  where  the  scale  is  subjected  to  intense  heat. 
In  the  latter  case,  while  the  deposition  of  the  scale- forming 
material  at  first  occurs  as  carbonate  and  sulphate,  the  gradual 
heating  expels  some  of  the  carbonic  acid,  and  the  oxides  of  cal- 
cium and  niaguesium  are  formed. 

That  portion  of  the  scale  nearest  the  iron  and  to  the  heat  loses  J 
more  of  its  carbonic  acid,  and  becomes  caustic  so  long  as  the  fin  I 

As  soon,  however,  as  the  fires  are  drawn,  the  oxides  of  calcium  \ 
and  tnagnesium  become  hydrated  by  absorption  of  water. 

If  now  a  sample  of  the  scale  were  taken  for  analysis,  the  water  I 
of  hydration  becomes  an  important  factor  in  the  analysis. 

A  sample  of  scale  from  some  boilers  at  Birmingham,  Ala., 
gave  the  following  result: 



Silica  aod   clay 11.70  per  cent. 

Fe,0„  A1,0. a.8i  ■■  '■ 

CaO i3-6a  "  - 

MgO 41-3'  "  " 

CO, 6.93  ■■  ■■ 

SO, 0.96  "  •■ 

H,0  (of  hydration) 21.78  -  " 

H,0  (inoi8tureat3ia°F.) 0.69  "  " 

Undetermined  " 

Total,   100.00    ■■ 

An  examination  of  this  analysis  shows  an  unusually  small 
amount  of  carbonic  and  sulphuric  acids,  a  large  amount  of  water 
and  ot  magnesia. 

The  great  excess  of  the  latter  over  the  lime  indicates  that  the 
water  from  which  thescale  was  formed  isa  magnesia  water,  but  its 
presence  in  this  amount  does  not  in  any  way  alter  the  conditions 
of  the  problem. 

With  less  than  one  per  cent,  of  sulphuric  acid  and  less  than 
seven  per  cent,  of  carbonic  acid,  the  oxides  of  calcium  and  mag- 
nesium could  not  exist  in  their  entirety  as  carbonates  or  sulphates, 
for,  combining  the  above  acids  to  form  carbonates  and  sulphates 
the  result  indicated  over  twenty  per  cent,  lacking  in  the  roo 

The  large  percentage  of  the  oxides  of  calcium  and  magnesium 
left  after  combination  with  the  acids  suggested  water  of  hydration. 

A  sample  of  the  scale  (dried  at  100°  C.)  was  transferred  to  a 
platinum  crucible  and  heated  over  the  blast  lamp  to  a  constant 
weight.  The  loss  of  weight  was  over  tweuty-eight  percent,  and, 
of  course,  included  the  carbonic,  but  not  the  sulphuric  acid. 

To  check  this  result,  a  sample  of  the  dried  scale  was  ignited  in 
a  combustion  tube  and  the  water  collected  in  a  weighed  calcium 
chloride  tube.  The  result  was  21.78  percent,  of  water  of  hydra- 

This  satisfied  the  conditions  existing,  and  the  combinations 
gave  as  follows : 


Silica  and  clay 1 1.70  per  cent. 

FCjOs,  AljOs 2.81 

CaSO^ 1.69 

CaCOj 5.45 

MgCOs 736 

Ca(OH)2 13.70 

Mg(OH), 56.37 

H,0  ( Moisture  at  212^  F.) 0.69 


Undetermined 0.20    ** 

Total,  99.97   ** 

A  section  of  the  scale  was  subjected  to  examination,  layer  by 
layer,  and  the  following  results  confirm  the  above. 

That  portion  of  the  scale  next  the  iron  and  nearest  the  fire 
contained  but  traces  of  carbon  dioxide,  and  was  principally  the 
hydrated  oxides.  The  middle  portion  of  the  scale  was  a  mix- 
ture of  carbon  dioxide  and  the  h)'drated  oxides,  while  the  upper 
portion  of  the  scale  contained  carbonates,  but  no  hydrated 
oxides.  In  other  words,  the  composition  of  the  scale  will  de- 
pend, in  a  great  measure,  upon  what  portion  of  the  boiler  the 
deposit  is  made.  That  deposited  on  the  iron  or  shell  not  in  con- 
tact with  the  flame  or  not  subjected  to  extreme  heat,  will  remain 
as  deposited — as  carbonates  and  sulphates,  while  the  scale  de- 
posited upon  the  iron  subject  to  the  flame  or  heat  sufficient  to 
drive  out  any  carbonic  acid  from  the  scale,  will  vary  in  the 
amounts  of  carbon  dioxide  and  water  of  hydration  as  indicated. 

Scale  formed  in  which  the  lime  all  exists  as  calcium  sulphate 
and  in  which  no  magnesium  carbonate  is  present  will  be  subject 
to  but  little  variation. 

When  oil  has  been  indicated,  by  qualitative  analysis,  as  pres- 
ent, the  method  of  analysis  requires  the  following  modification : 

The  sample  of  pulverized  scale  is  dried  at  gS""  C.  to  constant 
weight,  and  a  portion  of  this,  one  and  one-half  gram,  is  trans- 
ferred, to  a  Soxhlet  tube  and  the  oil  extracted  with  ether.  The 
ether  solution  evaporated  carefully  in  a  platinum  capsule  and 
the  amount  of  oil  determined. 

The  residue  in  the  Soxhlet  tube  is  dried  again  and  the  analysis 
made  in  the  regular  way. 

The  following  is  an  analysis  of  a  boiler  scale  containing  some 
lubricating  oil : 


SiO, 7.36  per  ceut, 

Al,0,.Fe,0, 1.91     " 

CaCO, 62.71     " 

MgCO, 18.15     "  " 

MglOH), 4J1     '■  " 

H.O.atllQ'-C 3.51     ■•  " 

Oil  Clubricating) 3.53     "  " 

Undetermined   0.63     "  " 

Total,  - — - — ■ 

Nearly  all  waters  contain  foreign  substances  in  greater  or  less 
degree,  and  though  this  may  be  a  small  amount  iu  each  gallon, 
it  becomes  of  importance  where  large  quantities  are  evaporated.' 

For  instance,  a  100  H.P.  boiler  evaporates 30, 000  lbs.  of  water 
in  ten  hours  or  390  tons  per  month  ;  in  the  comparatively  pure 
Croton  water  there  would  be  88  lbs.  of  solid  matter  in  that  quan- 
tity, and  in  many  kinds  of  spring  water  as  much  as  2000  lbs. 

The  nature  and  hardness  of  the  scale  formed  of  this  matter 
will  depend  upon  the  kind  of  snbstances  held  in  solution  and 
suspension.  Analysis  ot  a  great  variety  of  incrustations  show 
that  calcinm  carbonate  and  sulphate  form  the  larger  part  of  all 
scale,  that  from  carbonate  being  soft  and  granular,  and  that 
from  sulphate  hard  and  crystalline.  Organic  substances,  in 
connection  with  calcium  carbonate  will  also  make  a  hard  and 
troublesome  scale. 

The  presence  of  scale  or  sediment  in  a  boiler  results  in  loss  of 
fuel,  burning  and  cracking  of  the  boiler,  predisposes  to  explo- 
sion and  leads  to  extensive  repairs.  It  is  estimated  that  the 
presence  of  one-sixteenth  inch  of  scale  causes  a  loss  ot  thirteen 
per  cent,  of  fuel,  one-fourth  inch  thirty-eight  percent,,  and  one- 
half  inch  sixty  per  cent. 

The  Railway  Master  Mechanics'  Association  of  the  U.  S.,  es- 
timates that  the  loss  ol  fuel,  extra  repairs,  etc,  due  to  tncms- 
I  tation.  amount  to  an  average  of  $750  per  annum  for  every  loco- 
V  BJOtive  in  the  Middle  and  Western  States,  and  it  must  be  nearly 
Pthesame  for  the  same  power  in  stationary  boilers.  When  boil- 
ers are  coated  with  a  hard  scale  difficult  to  remove,  it  will  be 
found  that  the  addition  of  one-fourth  lb,  of  sodium  hydroxide 
per  horse  jwiver  and  steaming  for  some  hours,  just  before  clean- 


ing,  will  greatly   facilitate  that  operation  often  rendering  the 
scale  soft  and  loose. 

Water  for  Locomotive  Use. 

After  many  years  of  experiment  upon  waters  for  Locomotive 
use,  by  the  chemists  of  the  Chicago,  Milwaukee  &  St.  Paul 
R.  R.,  the  results  obtained  may  be  stated  as  follows : 

Varieties  of  water  may  be  classified  by  either  of  two  methods  : 

1.  By  their  chemical  composition. 

2.  By  their  effect  in  use. 

The  second  is  manifestly  what  is  wanted  by  master  mechanics 
and  superintendents. 

The  following  may  be  placed  in  the  first  class : 

a.  Alkaline  waters. 

b.  Non-alkaline,  bad  and  good. 
In  the  second  class  (2): 

a.  Those  causing  foaming  and  corrosion,  but  non-incrusting. 

b.  Hard,  or  incrusting. 

c.  Soft  non- alkaline  and  good. 

These  two  classes  are  related  as  follows : 

**a*'  of  class  1,  ** alkaline  **  waters,  will  produce  the  trouble 
mentioned  in  '  *  a ' '  of  class  2  :  that  is,  foaming  and  in  certain 
cases  corrosion. 

**^,'*  the  bad  ^^non-alkaline,*'  would  be  classed  as  hard  or 

*V,'*  **soft  waters/*  would  include  all  those  having  little 
mineral  impurities  of  any  kind. 

It  is,  however,  impossible  to  set  hard  and  fast  limits  for  each 
class,  one  generally  shading  into  the  other,  and  what  would 
be  called  good  water  in  the  West,  for  instance,  would  be  thought 
poor  enough  in  the  Hast. 

In  making  an  analysis  all  ingredients  are  grouped  broadly 
under  two  heads,  **  incrusting*'  and  **  non-incrusting."  Under 
the  former  are  put  such  salts  as  are  thrown  out  of  solution  by 
heat,  and  in  the  latter  case  those  which  do  not  precipitate  until 
great  concentration  occurs — a  condition  which  hardly  ever  hap- 
pens with  locomotives. 




In  the  "  non-incrusting"  group  is  found  a  variety  of  actions. 
A  well  known  property-  of  alkali  in  water  is  to  cause  foaming 
and  priming,  when  sudden  reduction  of  pressure  occurs  upon 
opening  the  throttle.  At  just  what  point  this  action  begins  to 
be  apparent  depends  on  a  number  of  circumstances.  With  a 
tjoiler  overworked  and  foul  from  mud,  it  appears  sooner  than  in 
one  having  ample  heating  surface,  with  moderate  train  load, 
uniform  resistance  and  consequent  regular  consumption  of  steam. 
For  a  maximum  allowable  with  good  results  in  service  and  in 
the  West,  where  really  good  water,  as  before  mentioned,  is  un- 
common, fifty  grains  per  gallon  of  alkaline  water  are  taken. 
When  this  figure  is  exceeded  it  certainly  pays  to  institute  a 
regular  search  for  better  water.  With  these  non-incrusting 
salts  are  associated  a  few  that  are  readily  decomposed  in  contact 
with  iron,  and  attack  it,  causing  gradual  corrosion.  These  are 
most  commonly  the  magnesium  chlorides  and  sulphates,  a  very 
small  amount  of  which,  say  ten  grains  per  gallon,  should  con- 
demn the  water.  Organic  matter  is  supposed  also  to  have  this 
action,  but  in  the  presence  of  alkali  the  danger  is  not  great  and 
with  frequent  blowing  out  little  attention  need  be  given  it.  The 
Tvater  may  be  classified  as  follows  : 

I  to  lo  grains  of  solids  per  gallon,  soft  water 

■  ID  to  20     "      "      moderately  hard  water 

Above  25  "      "      very  hard  water. 

On  this  railroad  "boiler  compounds"  are  employed.  Waters 
having  thirty-five  to  forty  grains  of  incrustating  matter  per  gal- 
lon can  be  dealt  with  successfully,  provided  no  alkali  be  present. 
The  above  reser\'ation  is  made  because  the  "compound"  is 
itself  an  alkali ;  so  in  adding  it  to  a  water  care  must  be  taken 
not  to  bring  the /o/fl/ aZ-td/*  al>ove,  say,  fifty  grains  per  gallon, 
or  there  will  be  trouble  from  foaming.  In  the  ' '  Report  of  Analy- 
sis" blanks,  directions  are  given  fixing  the  amount  of  compound 
to  use  in  each  case.'  A  few  examples  of  the  different  kinds  of 
water  used  on  this  road  are  here  given,  illustrating  the  distinc- 
tions above  drawn.     The  best  is  surface  water,   in  the  forest 

■odium  catbonslt,  dii»l«(l  In  one  gallon  ot  wslet.    The  oveiBKe  coi.1  lor  a  mo  of  1.OD6 
ni1c«  brtog  almul  Inrly  cents. 


region  of  Wisconsin  ;  for  example  that  from  Wausau,  as  follows : 

Total  solid  residue 6.78  grains  per  gallon. 

(  Oxide  of  iron. ...  0.23 

Incrusting  matter  i  CaCOs 2.26 

(CaS04 0.56 

((         <t        i< 
(<         («        << 

t(  it  a 

Total 2.95 

(I         <(        (t 

<t         »(        <( 

Non-incrusting  /  Organic  and  volatile  3.15 
matter I  Alkaline  chlorides . .  0.68      " 

it         ((        t( 

Total 3.83 

For  boiler  purposes  this  water  could  not  be  better,  the  in- 
crusting  matter,  about  three  grains,  being  inappreciable. 

For  a  good  example  of  badly  incrusting  water,  but  non-alka- 
line, the  following  from  Lennox  Creek,  Dakota,  may  be  given  : 

Total  solid  residue 109.20  grains  per  gallon. 

Incrusting  matter {^^'^0^; :;:;::;:     ^o-ai 

ti  ii  n 

<t  ((  <( 

Total 47.48 

((         i(        (< 

Non-incrust- f  ^''^***^^  *°^  ^^^*^*^^     ^^'34      *' 
ison-incrust- 1  ^^^q^ ^5  o^      ^ 


^°^°**"^'' I  Alkaline  chlorides..       i.\i      ** 

it         tt        (• 

Total 61.72 

This  water  could  not  be  properly  purified  by  the  addition  of 
caustic  or  carbonated  alkali  without  introducing  an  inadmissible 
amount  of  the  latter,  as  above  noted. 

It  will  be  noticed  that  the  magnesium  sulphate  is  classed  as 
**  non-incrusting'*  matter.  It  is,  however,  much  more  hurtful 
than  the  lime  salts  on  account  of  its  corrosive  properties.  The 
organic  matter  is  also  high,  but  not  more  so  than  is  usual  for  a 
surface  water  in  that  locality. 

For  examples  of  absolutely  worthless  water,  notice  first,  that 
from  an  artesian  well  at  Kimball,  D.  T. 

Total  solid  residue 182.06  grains  per  gallon. 

Incrusting  f  Calcium  carbonate 61.85      **  ** 

matter.  \  Calcium  sulphate 41.44      **  '* 

Total 103.29  **  "  •* 

Non-incrust- f  Alkaline  sulphates..  64.83  *'  **  ** 

ing  matter  \  Alkaline  chlorides..  13.94  **  *'  ** 

Total 78.77  •*  **  •« 

'  FEED   WATEK  HEATERS,    ■  99 

And  again,  Irom  a  130  feet  driven  well  at  Fargo,  D.  T. 

Total  soiiil  residue 416.84  grains  per  gallon.  . 

'°"""V°KJcald"mVaVboD^^^  ito!*     "         "        '" 

'"  iCalciuni  sulphate 33.46      ■■  "         "  I 

Total IJO.46 I 

IMagnesiutn  sulphate    10.90  "         "        "  I 

AlkaliDCSulptaates..   150.91 I 

Alkaline  chlorides..       1. 14  "         "         "  I 

Organic  and  volatile     33.41  ''          "         "  I 

Total 196.38 

It  is  manifestly  useless  to  attempt  the  purification  of  these 
waters  practically. 

All  the  rouud-liouses  are  provided  with  hydrants  and  high 
pressure  steam  connections  for  the  purpose  of  obtaining  a  power- 
ful stream  of  hot  water  for  wash-out  use. 

On  eastern  divisions,  locomotives  having  run  from  1,500  to 
2,000  miles  are  blown  off  at  low  pressure,  cooled,  and  the  stream 
of  hot  water  thrown  iu  at  hand  holes,  front  tube-sheet  and  back 
head,  and  scraper  worked  in  and  out.  The  .sediment  is  found 
mostly  loose  and  in  the  form  of  fine  mud,  to  the  amount  of  ten 
to  fifteen  buckets  full.  After  thorough  cleaning,  the  boiler  is 
again  filled  with  hot  water,  and  is  ready  for  service. 

On  the  western  divisions  the  frequency  of  washing  out  is 
increased,  doing  so  as  often  as  once  every  300  or  400  miles 
run.  As  to  the  economy  of  using  hot  water  ahcays.  there 
can  be  no  question.  Fully  seventy-five  per  cent,  iu  the  number 
of  cracked  fire-box  sheets  are  saved  by  this  practice  alone,  and 
it  materially  reduces  the  force  of  repairers  in  round-houses, 
notwithstanding  a  very  large  increase  of  engine  mileage. 

Many  people  are  opposed  to  the  use  of  chemicals  in  boilers, 
rightly  upon  general  principles  ;  but  when  the  proper  ones  are 
used,  the  experiments  have  failed  to  show  the  slightest  injurj' 
therefrom,  while  the  economy  resulting,  both  in  ser\-ice  and  re- 
pairs, has  amounted  to  an  enormous  sum  on  this  system. 
Feed  Water  Heaters. 

Feed  water  heaters  as  well  as  boiler  economizers  are  often  used 
as  eliminators  of  the  scale-forming  materials  in  water.     This  is 


due  to  the  fact   that  waters  containing  much  calcium  and  mag-si 
nesium  carbonates  when  heated  to  the  usual  temperature  in  feed  ' 
water  heaters  (20o°-2io°F),  give  up  the  excess  of  carbon  dioxide 
that  holds  the  calcium  and  magnesium  carbonates  in  solution, 
and  the  latter  are  precipitated  and  removed  before  the  water  e 
ters  the  boiler. 

Where  calcium  sulphate  is  a  large  constituent  of  the  water,  fe< 
water  heaters  using  exhaust  steam  do  not  precipitate  the  lime  sa] 
but  if  the  feed  water  be  heated  by  live  steam  under  pressure to-J 
temperature  of  240°  F,  then  the  calcium  sulphate  precipitates,  ' 
addition  to  the  water  of  the  necessary  amount  of  sodium  carbonate 
will  precipitate  the  lime  as  carbonate,  at  ordinary  temperatures, 
and  it  will  thus  be  found  more  economical  in  this  case  to  use 
feed  waters  heaters, using  exhaust  steam  with  sodium  carbonate, 
than  feed  water  heaters  using  live  steam  only. 

An  example  ot  an  upright  feed  water  heater  heated  by  exhaust 
steam  is  the  "  Goubert." 

The  exhaust  steam  from  the  engine  is  admitted  to  the  s! 
through  the  nozzle  on  one  side  and  spreading  between  the  h 


'  tubes,  impinges  upon  them  on  its  passage  across  to  the  outlet  on 
the  opposite  side ;  the  water  of  condensation  being  removed  by 
the  drain  pipe.  The  cold  water  may  be  admitted  at  either  top 
or  bottom  of  the  heater,  passing  out  at  the  opposite  end  :  but  for 
bad  waters  the  feeding  should  be  at  the  top.  This  being  a 
closed  heater  and  the  water  being  forced  through  against  boiler 
pressure,  the  flow  along  the  heating  tubes  will  be  the  same  whether 
thewater  movesinan  upward  oradownwarddirection.but  in  the 
latter  case  the  separation  and  settling  of  sediment  will  be  much 
more  thorough,  while  the  heating  will  be  the  same. 


This  purifier  consists  of  a  round  shell  of  best  boiler  steel,  hav- 
ing a  solid  pressed  flange  steel  head  riveted  in  the  back  end,  and 
a  solid  pressed  flange  steel  head  bolted  to  a  heavy  ring  on  the 
front  end,  by  studs  and  nuts.  Within  the  shell  are  a  number 
of  trough-shaped  pans  or  trays,  placed  one  above  another,  and 
supported  on  steel  angle  ways,  fixed  longitudinally  by  means  of 
brackets  to  the  sides  of  the  shell.  These  pans  are  formed  from 
thin  sheet  metal,  the  heads  or  end  pieces  being  malleable  iron, 
whereby  a  verj-  light, strong  and  durable  construction  is  obtained, 
and  a  degree  of  elasticity  secured  to  the  pans,  which  permits  the 
lime  or  other  incrustations  being  easily  removed.  Six  pans  are 
placed  in  a  tier,  and  from  one  to  four  tiers  used,  according  to 
capacitj'  required.     The  purifier  is  connected  to  the  boiler  by  a 


large  steam  pipe  A,  and  the  exit  pipe  D,  A  blow-off  pipe  is 
also  connected  at  C,  The  feed  pipe  from  the  pump  or  boiler 
feed  is  attached  at  B, 

In  operating  the  purifier,  the  water  is  pumped  in  at  B  and 
distributed  into  the  upper  pans  through  the  pipes  leading  into 
each  pan.  While  the  purifier  is  in  operation,  the  pans  remain 
full  of  water,  and  afford  ample  settling  chambers  for  the  heavier 
solids,  such  as  mud,  sand  etc.,  etc.,  while  the  carbonates  and 
sulphates  (scale-forming)  adhere  to  their  under  sides. 

An  analysis  of  a  sample  of  water  before  passing  through  one  of 
these  heaters  at  Rochester,  N.  Y.,  is  as  follows: 

Before  Use. 

Inorganic  solids 128.74  grains  per  gallon. 

Organic  matter 3.38 

((        (<  «( 

((         »i  «« 

Total  solids 132.12 

After  Passing  Through  Heater. 

Inorganic  solids 8.44  g^rains  per  gallon. 

Organic  matter 3.20 

t«  (4  (< 

Total  solids 11.64 

««         i(  (( 

£"3  ffi.o  O  w  i»  JT'O  -a  -J  1- ?  S^  3-ui  * 

fffffi  &  Si's    «  w"o-'&'&  &  ~    MA 

"S*^^  «Su<  »SoD  oo-b  ~j  ^-t^  ^  u  ce  So 

Sa  8.-:i-:3-?J  SS'£  8>£2S53'iS  S 


o-jSq  "S-  S'Sdn  S""^  *" 


-    B4  ^    u  &*t>l  ■'^    aj  OD  00  OO'fi 


n-S£->i  o'Soic  E^u■  KE  -  «l  ^  A  s*-^ 




























u:    6 

ill    *> 











^  O  <O>O00  O  M  ^^00 

•        ••••••••• 

O^t^QO  *ONO   ON  O  O^  «>i  "S" 

^  0\  M  ^^  00  •-•  n  ^>o 

•         ••••••••• 

lO^^eOM  •-•  ^   O   OvOO 



•         ••••••••• 

«o^^*o«  "^  o  o  ONoo 



lO  r^  POOO  M  eo  ^  «o  •-•  r>» 
«  lOX  O  fO  »0  r^  O^  «  « 

»0^e0«0«   '-'   O  ON  CSOO 



OQ  O  NO  O  ^nO  no  >0  rO  on 
On  ^nO   On  m   ro  >0  t>i  On  O 

•  ••••••••• 

^^fOC«  C«  i-i  o  o»«  X 

►i-  «  ONfONOXX  I^^Q 
X  C«  ^I^ONt^  rOiOI^ON 

•  ••••••••« 

^  ^  fO  «    N*    •-•     O     ONX    »^ 



•          ••••• 

"^  ^  rO  C«   ^   i-i 

^  OnX  r^ 

«       •       •       • 

O  O^X  r«> 



•          ••••• 

^  eo  row  ►^  O 

•        •        •        • 

O  ONX  r^ 




eO  «>i  On  «  ^»0  X  O  n  ro 

•  ••••••••• 

^  fOC*    PI    N*    O    ON  ONX    »^  I 

ON  Q  Tf\0  t>.  r^NO  fOX  •- 
»o  O  «  ^nO  X  o  «  «o  »o 

•         ••••••••• 

C«  r>«  iM   O   OnX  X  t^NO   lO 



-  «  ^i  ovo 

«  lO  c^ 

lOfs.  ONl-l    C« 



•       «       •       •       « 

•         •         • 


O  0 

ONX  ts.fs.NO 

lO"^  ro 


M     M 

ON  C>»X    l>»vO   «    C>»  M    fO  ^ 

•-  lO  e*»  ON  »^  ro  ^no  r«*oo 

*  •••••••• 

ONX    t>.NO  NO    •'^  ^  fj  PI    M 

ON^  ip  f«^  I-I  ^x  0  ►^   . 

^X  O  «  ^  ^O  X  ON  . 

^•«       •       •       .       •       •       .  ., 

ts.NO  ^O  >0  ^  ro  C«  •>«  O    . 

-8jdudS  at  pdqjos 
-q«  qt  Jdd  «)ian  iBdH 

t^vO   »0  ^  rO  C«   M   O 



'jd)Vdq  Saud)Ud 

O  5t-  ionO   O.X   On  S  M  S 

USE   or   CHEMICALS   i 



I       ■'  Blowing-off.  ■  ■     The  arrangement  in  a  boiler  for  this  purpose 

I  usually  consists  of  one  or  two  internal  pipes  extending  along  the 

I  "bottom  of  boiler  and  connected  with  the  blow-out  tap.     They 

I  are  placed  about  one  and  one-half  inches  clear  of  the  plates  and 

are  perforated  on  their  under  side.     It  is  usual  to  blow-out  the 

sediment  every  two  or  three  days  just  before  drawing  the  fires 

and  the  sediment  in  the  water  has  had  time  to  settle.      Consult 

also  "  Wilson  on  Boilers,"  page  169-171. 

Use  of  Chemicals  and  Filtration. 
Dervaux  Water  Purifier  for  boiler  use. 
This  apparatus  (Figs.  22,  23)  is  automatic  in  action,  and  is 
thus  described.'  The  purifier  is  intended  to  act  as  an  eliminator 
for  both  calcium  sulphate  and  calcium  and  magnesium  carbon- 
ates. It  not  only  acts  to  precipitate  the  dissolved  impurities,  but 
also  to  collect  those  that  are  in  suspension.  These  last  are  caught 
and  held  in  the  tower-shaped  holder  D.  The  water  enters  at  H, 
passes  down  through  £,  and  is  made  to  rise  through  a  series  of  fun- 
pels  or  inclined  funnel-shaped  walls.'  0\\  these  walls  the  coarsest 
particles  are  caught  and  from  them  they  flow  down  to  the  bottom 
of  the  tower,  where  they  collect :  the  water  then  passes  upwards 
though  the  filters  F,  ivhich  are  made  of  wood  shavings,  and 
flows  off,  freed  from  its  mechanical  impurities  through  the  open- 
ing T.  In  the  mean  time,  by  the  addition  of  lime  and  soda,  the 
water  has  been  chemically  purified  in  the  following  way:  The 
water  first  flows  into  the  reservoir  C.  through  the  pipe  H.  In 
C,  there  is  a  float  for  regulating  the  flow  of  water.  A  portion 
of  the  water  goes  into  E.  through  the  pipe  P,  while  the  rest 
passes  through  the  valve  ('' into  the  lime  satnrator  ^;  5" is  filled 
with  lime  :  the  water  first  meets  the  lime  at  the  bottom  of  the 
aaturator  and  passes  up  through  it;  the  conical  shape  of  S  causes 
the  rise  to  be  slower  and  slower  as  the  water  nears  the  top,  so 
that  the  milk  of  lime,  at  first  formed,  has  plenty  of  time  to 
clarify  itself.  The  lime  water  usually  contains  some  calcium 
i:arbonate  in  suspension :  and  as  this  is  worthless  for  purposes 
of  purification,  it  is  eliminated  by  causing  the  water  to  flow  over 

iPapiir  Zniung.  34.S84 

>  Tlic  Cfacn 




into  the  cone  K,  which  is  closed  at  the  bottom  In  this  cow 
the  carbonate  settles  out,  and  may  be  dran  d  o9  through  i 
The  clear,  saturated  lime-water  containing  i  s  gram  of  lime 
per  liter,  runs  then  directly  into  the  miving  tube  L  A  solution 
of  soda-ash  is  made  by  taking  a  known  weight  of  Ilieash,  which 
is  placed  in  the  tank  Z,  after  which  the  tank  A',  ts  filled  to  a  de- 


finite  mark  with  water.  This  solution  slowly  passes  through 
the  tube  provided  with  strainers :  a  float  in  the  tube  keeps  the 
water  in  £  at  a  constant  level .  The  siphon  A',  one  end  of  which 
dips  to  the  bottom  of  S.  allows  the  alkaline  solution  to  flow  into 
fi.  The  regulation  of  the  flow  in  £  is  performed  as  follows: 
The  siphon  A'  is  joined  by  a  chain  Q,  to  the  float  in  C.  In  case 
the  flow  of  water  through  //  to  C  is  cut  off,  the  float  sinks,  rais- 
ing .Vand  thus  stopping  the  flow  of  the  solution.  At  the  same 
time  the  level  in  C  sinks  so  low  that  the  flow  of  water  through 
P  and  (^ceases  ;  as  soon  as  the  flow  of  water  through  yV  recom- 
mences, the  apparatus  is  again  set  in  operation  automatically. 

The  chemical  operations  may  be  stated  as  follows :  The  addi- 
tion of  the  lime  softens  the  water  by  precipitating  any  bicarbon- 
ate which  may  be  present,  and  the  excess  of  lime  is  thrown 
down  by  the  sodium  carbonate.  This,  by  its  precipitation  throws 
out  much  of  the  finely  divided  organic  impurity.  The  apparatus 
may  be  easily   modified  to   work   with   alum   where  desirable. 

This  Der\'eax  Purifier  is  extensively  used  in  France  and  Ger- 

In  England  the  apparatus  devised  and  patented  by  I*.  Arch- 
buit,  F.  I.  C.  and  R.  M.  Deeley,  M.  E.,  has  an  extensive  use 
for  the  purification  of  boiler  waters. 

The  drawings  ( Figs.  24,  25,  26 )  show  the  construction  and  rep- 
resent a  purifier  suitable  for  the  treatment  of  from  5,000  to  10,000 
gallons  of  water  perhour.  Ilconsists  of  acast-iron  tank,  measur- 
ing 32  feetX  16  feetX  10  feet  deep,  divided  into  two  equal  parts 
by  a  transverse  partition  of  cast  or  wrought  iron.  The  two 
tanks  are  intended  to  be  used  alternately,  so  as  to  maintain  a 
continuous  supply  of  softened  water. 

The  water  to  be  purified  is  admitted  to  either  tank  by  means 
of  the  supply  pipe,  1,  which  is  connected  up  to  a  pump  or  main. 
The  water  fills  up  nearly  to  the  level  of  the  top  of  the  well,  4. 
While  the  tank  is  filling  the  proper  amounts  of  lime  and  sodium 
carbonate  are  weighed  out,  with  the  addition,  in  some  cases,  of 
3  verj-  small  quantity  of  aluminum  sulphate,  or  alumina-ferric 
cake,  and  these  are  boiled  up  with  water  in  the  small  chemical 
tsnk,  2,  by  means  of  steam  Irom  the  steam  pipe.  The  trajector, 
,  is  put  into  action  by  opening  its  steam  valve,  and  then  the 


WnwT  WUtO  ■WTM  PI 



■^^^saical  liquid  is  run  out  of  the  chemical  tank  into  the  well. 

"^  trajector  creates  a  powerful  current  of  water  from  the  well, 
^titoiigh  the  projecting  pipe,  across  the  tank,  and  into  this  cur- 
'^at  the  chemicals  pass.     After  the  chemicals  have  thus  been 
^^ded  and  mixed  with  the  water,  and  the  trajector  shut  off, 
steam  is  admitted  to  the  blower,  5,  which  causes  air  to  be  sucked 
^o^Ti  the  orifice  and  forced  out  of  the  perforations  in   the  pipes 
'aid  close  to  the  bottom  of  the  tank.     After  the  blower  has  been 
in  operation  for  fifteen  minutes,  the  steam  is  turned  off  and  the 
water  is  allowed  to  rest.     The  result  is  that  in  about  thirty  min- 
ules  very  nearly  all  of  the  precipitate  will  have  settled  to  the  bot- 
tom of  the  tank.     The  drawing-off  and  carbon;iting  are  opera- 
tions  that   are   automatically   and   simultaneously  effected  by 
means  of  the  floating  discharge  pipe,  9,  of  rectangular  section. 
Fuel  gas,  from  the  coke  stove,  7,  constructed  so  as  to  produce 
aminimum  of  carbon  monoxideandamaximnm  of  carbon  dioxide 
is  forced  continuously  by  means  of  a  very  small  steam  blower,  8. 
The  gas  and  water  pass  together  through  the  ball  tap  fixed  over 
the  small  supply  tank,  i2,  into  which  the  softened  and  carbona- 
ted water  falls,  and  from  which  it  is  drawn  off  for  use,  whilst  the 
residual  gas  and  nitrogen,  etc.,  escape  into  the  air.     The  mud 
is  removed  by  extending  the  main  blower  pipe  through  the  side 
of  the  tankwhereitterrainatesinavalve,  14,  which  by  opening  for 
a  few  minutes  at  inten-als  the  accumulation  of  mud  is  prevented. 
The  reasons  for  carbonating  the  softened  water  are  fully  ex- 
plained in  a  paper  read  before  the  Society  of  Chemical  Industry 
in  June,  1891.     Uncarbonated  softened  water  often  forms  a  de- 
posit in  pipes  and  especially  in  the  feed  apparatus  of  steam 
boilers,  which  may  become  very  troublesome.    This  is  not  a  pe- 
culiarity of  water  softened  in  this  apparatus. 
The  output  can  be  calculated  as  follows  : 
V  =:  the  number  of  gallons  of  softened  water  supplied  continu- 
ously per  hour. 

jT^the  working  capacity,  in  gallons,  of  each  tank, 

=  the  number  of  minutes  required  to  fill  each  tank. 
I  x^the  number  of  minutes  required  for  settling. 
_        60  .r        _ .,_ 












1    O 



•  •  •  •  • 



«  ro 


•          • 


M^  0 




O  fO 

•         • 


00  c«  d  c«  ri 


.    rooO 



•           *          * 

.  vO  O 




^OOnOn*    to'oOON 

«   to  O 

NO  ^^ooo  . 

O   ON^Or*.  On  • 

•         •         •         •         •  . 

w  O  r^»^  On  . 

*    t^vO 

•  *  * 

.  NO    o 

OvOt^»£)»    ON.    ^OM 

•       ••■«        ••        ••• 

ONt^Cil^,   ro.   OnOO 

On^O  On^vO*     *     (QO*^ 
OOt^ON^ON*     •     •«»*) 

O  «  «  «oo    .    .    .   lOO 

ON  **   O  00 
to  fO  f*5  « 

no"  d  ^  •-•* 

•3  •  V?^ 

2  •  9  "^ 

5  I  rod 


^00  NO 

•  •  •         ^  • 

X  M  eo   .    1^ 



N    fO 

«  o 



S         3   «    3  C   S    ;, 



bcji  be;::  ics  bTji.i 
55  u  S '/.  S  -^  S  •/.  -x 


"o  remove  calcium  carbonate  by  chemical  means  from  water 
costs  very  tittle,  because  lime  alone  is  necessary,  and  is  very 
cheap.  To  remove  calcium  sulphate,  alkali  must  be  used,  which 
greatly  increases  the  cost.  Both  lime  and  alkali  are  necessary 
for  the  removal  of  certain  magnesium  salts,  and  the  alkali  must 
be  used  in  greater  relative  proportion.  Waters  containing  much 
magnesium  salts  are  therefore  the  more  costly  to  treat.  The  table 
on  page  i  to  gives  the  analyses  of  nine  typical  samples  of  water, 
together  with  the  cost  of  chemicals  needed  to  soften  each  by  this 
process,  and  reduce  the  hardness  to  3^  and  not  exceeding  5°. 
Filter  Presses. 
Filterpresses  are  often  used  for  rapid  filtration  ofwater.  These 
presses  consist  of  a  number  of  filter  chambers  with  solid  separat- 
ing walls,  which  are  held  between  two  head  pieces,  one  of  which 
is  fast  and  the  other  movable ;  the  latter  as  well  as  the  filtering 
frames  slide  along  two  strong  iron  rods.     Between  the  chambers 

the  filtering  cloth  is  hung  and  this  also  helps  to  make  the  outer 
edges  fit  closer  together.  The  whole  system  is  pressed  together 
by  a  screw  or  lever  or  by  hydraulic  pressure  and  forms  a  num- 
ber of  hollow  spaces  lying  together  and  communicating  with  one 
another.  Between  these  hollow  spaces  the  liquid  to  be  filtered 
is  pressed  by  a  pump  or  other  means.  During  this  process  the 
separation  of  the  liquid  and  solids  takes  place,  in  that  the  liquid 


is  forced  through  the  cloth  and  runs  out  clear  through  chaii 
to  a  common  outlet,  leaving  the  solids  behind. 
We  distinguish  two  varieties  of  filter  presses. 

1,  Chamber  Presses,  (Fig.  27)  by  which  the  space  fori 
the  cake  i.  e.,  the  solid  matter  remaining,  is  formed  by  the  edg%s:l 
of  each  two  filters  plates,  so  that  the  cake  falls  out  when  the  pres&fl 
is  opened. 

2.  Frame  Presses,  by  which  the  space  for  the  cake  is  formed 
by  frames  that  are  placed  between  each  two  filter  plates,  so  that 
the  cake  can  be  lifted  out  with  the  frames. 

In  order  to  dr>-  the  cake  completely  and  to  wash  it,  when 
necessary,  there  are  in  most  filter  presses  two  extra  canals  inl 

each  chamber,  one  in  which  the  washing  fluid  enters  and  the 
other  by  which  it  is  removed.  There  is  also  an  attachment  by 
which  liquids  can  be  filtered  hot  or  cold. 

The  Porter-Clarke  process  for  softening  hard  water,  largely 
used  in  England,  makes  use  of  filler  presses  to  remove  the  pre- 
cipitated material  in  the  water.  Where  this  latter  precipitate  is 
ver>'  fine  and  small  in  amount,  manufacturing  establishments 
sometimes  arrange  a  system  as  shown  in  Fig  28  in  which  fibers, 
of  cellulose  are  added  to  collect  the  fine  precipitate.     The  ajv, 


Qiaratus  consists  of  a  high  horizontal  reservoir  H  (Fig.  28)  for 
reception  of  the  water  to  be  filtered,  another  reservoir  or  tank 
M,  in  which  the  floating  material  (or  fibers  of  cellulose  or  asbestos) 
is  mixed  with  water,  a  reservoir  R  into  which  the  purified  water 
flows  and  the  filtering  apparatus  proper  F.  T!ie  latter  is  com- 
posed, as  are  the  filter  presses,  of  a  series  of  frames  on  which 
metal  sieves  are  fastened.  The  filtration  takes  place  in  the 
following  manner:  The  thin  mass  of  cellulose  or  asbestos  fibers 
are  caught  by  the  sieves  and  remain  on  them  :  the  water  is  then 
allowed  to  pass  from  the  reservoir  H  through  the  sieves  which 

I  vow  holds  back  all  suspended  matter,  so  that  clear  water  flows' 
Ja  the  reservoir  R. 

Another  method  made  use  of  in  some  large  industrial  plants, 
is  to  combine  the  action  of  a  heater,  chemical  precipitation  and 
filtration  by  filter-presses  as  shown  in  Fig.  29. 

The  water  passes  first  through  the  heater  A  in  which  it  is 
brought  to  the  temperature  of  the  heater,  thence  into  the  pre- 
"pitalion  tank  B  in  which  it  is  mixed  with  the  chemicals  in 
solotions  the  latter  being  pumped  into  B  from  F  by  means  of  the 
pump  D.  The  water  then  passes  into  the  filter  press  C,  in  the 
'^'•aiubers  of  which  the  suspended  matter  is  retained,  and  is  then 
Pttn3]>ed  by  the  pump  E  either  directly  to  the  boiler  or  else  to  a 
enoir.     The  Water  and  chemicals  are  mixed  in  the  propor- 


tions  shown  to  be  necessary  by  analysis.  This  system  of  water 
purification  has  shown  itself  to  be  very  successful,  but  the  filter 
press  must  be  cleaned  every  two  to  eight  days  according  to  the 
composition  of  the  water. 

References :  * 'Boiler  Deposits,'*  Vivian  B.  Lewes,  F.C.S.i  Transactions 
Inst,  of  Naval  Architects,    Vol .  XIV. 

**  Boiler  Incrustation,"  Treatise  on  Steam  Boilers ^  Robert  Wilson  C.E., 
pages  158-187. 

"  The  Purification  of  Water  for  Domestic  and  Manufacturing  purposes, 
{Hyatt  System.)  By  J.  S.  Crone.  Trans,  Am,  Soc,  Mech,  Engineers ^  7, 

"The  use  of  Kerosene  oil  in  Steam  Boilers,  as  a  preventative  of  scale. 
By  Lewis  F.  Lyne,  Trans,  Am,  Soc,  Mech,  Engineers^  8,  247-259 

*'  Corrosion  of  Steam  Boilers.'*  By  David  Phillip,  Proceedings  Institu- 
tion  of  Civil  Engineers,  65,  73. 

"On  the  Results  of  an  examination  of  the  Chemical  Composition  of 
steam-raising  waters  and  of  the  incrustations  formed  from  such,  with 
notes  on  the  action  of  the  more  common  materials  employed  as ''ante- 
incrustators"  and  of  the  various  processes  for  softening  water  for  steam 
purposes."  By  W.  Ivison  Macadam,  F.C.S.,  /.  Soc,  Chem,  Industry^  a, 

"  The  Porter-Clark  Process"  (for  softening  water.)  By  J.  H.  Porter, 
y.  Soc,  Chem.  Industry  3,  51-55. 

"  Suggestions  on  Boiler  Management.'*  By  VeroC.  Driffield.  J,  Soc, 
Chem.  Industry,  6,  178-189. 

"  The  Analytical  Examination  of  Water  for  Technical  Purposes."  By 
Alfred  H.  Allen,  V.  C.  S../.  Soc,  Chem.  Industry,  7,  795-806, 

*•  The  .\ction  of  Water  on  Lead  Pipes."  By  Percy  F.  Frankland,  F.I.C., 
/.  Soc.  Chem.  Industry  8,  240-256. 

••  The  Treatment  of  Hard  Water."  By  L.  Archbutt.  F.I.C.,  and  R.  M. 
Deehiy.    ,/.  ^'tv.  Chem,  Industry  10,  511. 

'*  The  Purification  of  water,  on  the  large  scale,  by  means  of  Iron."  By 
William  Anderson.  Proceedings  of  the  Institution  of  Civil  Engineers, 
81,  27g. 

Determination  of  the  Heating  Power  of  Coal  and  Coke. 

The  simplest  method,  but  which  gives  only  approximate  re- 
sults, is  the  ij^iiilion  of  coal  with  litharge  in  a  crucible.  In  de- 
tail the  process  is  as  follows  :  one  gram  of  the  finely  powdered 
cvKilis  imimalely  mixed  with  thirty  grams  of  litharge  (PbO), 
tiaiisierreil  to  a  No.  ^;  Hessian  crucible,  twent>*  grams  more  of 


litharge  placed  on  top  of  the  charge,  the  crucible  covered  and 
heated  at  a  high  heat  in  the  furnace  for  fifteen  miuutes.  The 
crucible  is  removed,  allowed  to  cool,  broken,  and  the  button  of 
metallic  lead  cleaned  from  the  slag  and  carefully  weighed. 

Duplicate  results  should  not  vary  more  than  0.025  gram.     To 
calculate  the  result: 

One  part  of  carbon  reduces  thirty-four  times  its  weight  of  lead, 
and  if  one  kilo,  of  carbon^: 8 140  calories,  then  each  part  of  lead 
is  equivalent  to  S140  ^^^^  ^^j^^j^^ 

Suppose  the  lead  button  from  one  gram  of  coal  weighed  31  05 

gram,  then X 3 1.05  =  7420.9  calories  per  kilo,  or  13357.7  B. 

T.  U,  per  pound  of  coal,  which  represents  the  heating  power  of 
the  coal. 

The  heating  power  of  coke,  containing  no  volatile  combustible 
matter,  can  be  calculated  from  the  analysis,  thus 

^^^  X  8140  — 7686.6  calories—  13837  B.  T.  U.  per  pound. 

Bituminous  coals  contain  volatile  combustible  matter  as  well 

as  var>'ing  amounts  of  water,  for  which  reasons  both  of  the  above 

methods  give  very  incorrect  detenninations  of  the  heating  power. 

Three  raelliods  are  available  (which  include  all  varieties  of 

I  coals  0 

I        I,  Calculation  of  the  heating  power  ■from  the  results  of  an  ele- 
F  tnentarj'  analysis  of  the  coal.  viz. :  determination  of  the  percent- 
ages of  carbon,  hydrogen,  nitrogen,  oxygen,  sulphur  and  ash. 

2  The  use  of  calorimeters  in  which  a  sample  of  coal  is  burned 
and  its  heating  power  determined  directly  from  the  experiment. 
3.  The  combustion  oi  large  amounts  of  coal  in  specially  de- 
signed apparatus  therefor,  and  also  boiler  tests. 
Ca/cu/afhn  of  the  Heating  Power  from  t/ie  Results  of  an  Elemen- 
tary Analysis  of  the  Coal. 
a.  Determination  of  the  carbon  and  hydrogen.     Select  a  Bo- 
[  bemian  glass  combustion  tube  about  seventy  cm.  long,  two  cm. 
in  diameter,  open  at  both  ends    (Fig.  30).      Place  in  it  at  j 

granulated  cupric  oxide  for  a  dis- 
tance o(  about  thirty  cm.,  and  at  *  a 
plug  of  asbestos;  place  the  tube  in 
a  combustion  furnace  c,  connect  it  at 
b  with  the  drying  apparatus  a.  and  at 
d  with  calcium  chloride  tube  e  filled 
with  CaCl,.  granulated.  The  latter 
is  connected  with  an  aspirator,  and 
air  is  very  slowly  drawn  through 
the  apparatus  ;  at  the  same  time  the 
furnace  is  gradually  lighted  and  the 
heat  increased  until  all  the  cupric 
oxide  has  reached  a  red  heat.  Main- 
tain this  for  fifteen  minutes,  turn  off 
the  gas,  and  continue  the  aspiration 
of  air  until  the  tube  is  nearly  cold. 
;  This  preliminary  heating  is  necessary 
^  to  eliminate  any  moisture  that  may 
•  be  in  the  tube  or  in  the  cupric  oxide. 
Transfer  five-tenths  gram  of  the 
finely  powdered  coal  to  a  weighed 
porcelain  boat  and  place  in  the  tube 
at  // ;  at  ^  is  a  coil  of  platinum  foil. 
The  calcium  chloride  tube  e  (Fig, 
31)  is  now  accurately  weighed,  as 
well  as  the  potash  bulbs  /, '  and  when 
all  the  connections  are  properly  made, 
heat  is  turned  on  in  the  furnace  at 
the  end  d,  and  oxygen  gas  is  very 
slowly  passed  through  the  apparatus. 
At  inlen-als  of  a  few  minutes  the  heat 
is  turned  on  in  the  furnace  until  the 
cupric  oxide  is  at  a  red  heat,  and 
finally  the  entire  tube  from  i  to  g-  is 
also  at  that  temperature. 
After  the  complete  combustion  of 






the  carbon  of  the  coal,  which  is  indicated  by  the  absence  of 
black  particles  in  the  porcelain  boat,  turn  off  the  heat  in  the 
furnace,  but  continue  the  slow  current  of  oxygen  until  the  appa- 
ratus is  nearly  cold.  The  hydrogen  in  the  coal  by  its  combus- 
tion is  converted  into  water  and  absorbed  by  the  calcium  chlo- 
ride tube  e  ;  the  carbon  of  the  coal,  by  its  combustion  with  ex- 
cess of  oxygen  has  produced  carbon  dioxide,  and  is  absorbed  in 
the  potash  bulbs/.  From  the  increase  of  weights  thus  obtained 
the  percentages  of  hydrogen  and  carbon  are  calculated,  thus  : 

Amount  of  coal  talcen  =:  0.500  gram. 

Calcium   chloride  tube  -(-  H,0  =  36.5118  grams. 
■'  ■■  ■'  =  36.4025        ■' 

H,0=    0.1093        ■' 
0.109  gram  H,0  =  o.oiii  gram  H. 

■                     =  2.41  per  cent,  bydrogea. 
The  potash  bulbs  and  COj  ^  34.9554  grams. 
=33^3300       ■' 

1-6354       ■■ 
'■^354  gi'an'S  CO,  s=  0.4460  gram  C. 

5:M^>1_'°?  =  per  cent,  catboa. 


The  ash  is  as  follows : 

Remove  the  porcelain  boat  from  the  combustion  tube  care- 
fully and  weigh  ;  the  increase  of  weight  is  ash.     Thus  : 
Porcelain  tube  +  residue  (aah)  =  S.9693  grams. 

o-°'33  X  1 


Ash  =  0.0333 

b.  The  nitrogen  determination  is  n 
Select  a  combustion  tube  about  si> 

five-tenths  cm.  diameter,  drawn  to  a 

at  the  other  end  (Fig.  32). 

i  follows  : 

.   long  and  two  and 

at  one  end  and  open 



At  a  place  three  grams  of  crystallized  oxalic  acid,  then  a  few 
layers  of  freshly  ignited  soda-lime,  and  at  b  insert  five-tenths 
gram  of  the  powdered  coal  mixed  with  about  twenty  grams  of 
soda-lime,  fill  the  rest  of  the  tube  with  soda-lime  and  finally 
some  asbestos  near  the  open  end  of  the  tube.  Connect  with  a 
bulb  tube  d  containing  fifteen  cc.  of  a  standard  solution  of  sul- 
phuric acid,  each  cc.  of  which  contains  0.049  gram  sulphuric 

The  combustion  tube  is  now  placed  in  the  combustion  furnace 
and  heat  is  gradually  applied  under  the  tube  at  e  and  extended 
slowly  towards  a.  The  soda-lime  between  c  and  the  coal  must 
be  at  a  red  heat  before  heat  is  applied  under  the  coal.  Now 
heat  the  tube  until  the  soda-lime  and  the  coal  are  well  heated 
and  maintain  this  until  no  more  gas  is  generated  or  passes  into 
the  standard  acid  ;  being  careful,  of  course,  that  none  of  the 
oxalic  acid  has  yet  been  heated. 

Gradually  heat  the  oxalic  acid,  which  slowly  vaporizes,  and 
in  passing  over  the  soda-lime  is  converted  into  carbon  dioxide. 
The  nitrogen  in  the  coal,  by  this  ignition  with  soda-lime,  is  con- 
verted into  ammonia  and  forced  out  of  the  tube  into  the  stand- 
ard acid  by  the  excess  of  carbon  dioxide  generated  from  the 
oxalic  acid. 

After  the  operation  is  completed,  disconnect  the  ||-tube  con- 
taining the  standard  acid,  transfer  its  contents  to  a  No.  3  beaker, 
add  u  few  drops  of  litmus  solution  and  titrate  with  normal  soda 
solution  to  determine  the  amount  of  ammonia  united  with  sul- 
phuric acid.     Thus  : 

Ci>iil  taken  =  o.5<x>gram  v^dried) 

H*St\  solution  taken  ^  15  cc. 

Noruuil  siHla  solution  required  to  neu-  )       , ,  _^o 
trali/e  f r.^  acid  j  =  **-7^  ^• 

^i>nc  vv.  Nal>H  solution  neutralizeti  one  cc.  HJSO4)  0.232  cc.  of  H,S04 

solution  nouirali.-od  l»v  the  ammonia. 


It  one  vv.  H;SO,  solution  =0.049  grani   H,S04  •  •  o-23*  cc.  ^  0.0113 
kjiam  H.SO.. 

v^  .^ii^ij:r;un  HjSC\  =  o.v.x\;o2  gram  NHj. 

=  o-^5  wr  oenik  nitrogen. 
o  5vv^ 



The  method  of  Kjeldahl  can  also  be  used  for  the  determina- 
tion of  nitrogen  in  coal.  Consult ''Contribution  a  T etude  des 
combustibles/'  P.  Mahler,  1893,  p.  19. 

The  sulphur  is  determined  as  directed  in  scheme  XII,  and  in 
this  sample  amounted  to  0.19  per  cent. 

Having  determined  all  of  the  constituents  in  the  dried  coal 
but  oxygen,  the  latter  is  estimated  by  subtracting  the  sum  of 
the  other  constituents  from  100.     Thus  : 

Carbon  •  •  • 
Sulphur  • 


Oxygen . . 

89.21  per  cent. 





Total 100.00 

d.  We  will  now  include  in  this  analysis  the  hydroscopic 
water  (the  above  analysis  having  been  made  upon  the  dried 
sample) . 

This  moisture  in  the  coal  is  a  direct  loss  in  the  calorific 
power,  since  it  absorbs  heat. 

Amount  of  coal  taken 2  grams. 

Watch-glass  and   coal   before  drying   twenty 

minutes  at  102^  C 12.162  grams. 

Watch-glass  and  coal  after  drying  twenty  min- 
utes at  102^  C 12.101 

Loss  (moisture)    0061 

t  i 

<  < 

0.061  X  too  _  ^^^  p^j.  ^^jj^   moisture. 


The  complete  analysis  of  the  coal  will  now  be  : 

Moisture 3.05  per  cent. 

Carbon 86.49 

Hydrogen  • 2.36 

Nitrogen 0.63 

Sulphur 0.18 

Oxygen 2.76 

Ash 4.53 

Total 100.00 

The  calorific  power  of  the  coal  is  calculated  from  the  follow- 
ing data : 


A  calorie  is  the  standard  heat  unit,  and  represents  the  heat 
required  to  raise  the  temperature  of  one  kilo  of  water  from  4®  C. 
to  5'  C. 

A  British  thermal  limit  (**  B.  T.  U.")  is  the  heat  required  to 
raise  the  temperature  of  one  pound  of  water  i^F.,  at  its  temper- 
ature of  maximum  density,  (39.1°).'  To  reduce  calories  per 
kilo  to  **B.  T.  U.*'  per  pound,  multiply  by  |;. 

One  kilo  of  carbon  (from  wood  charcoal)  in  burning  to  car- 
bon dioxide  produces  8140  calories. 

These  figures,  8140,  obtained  by  Berthelot  and  Bunte  are 
probably  nearer  correct  than  the  figures  8080  given  by  Fa\'Te 
and  Silbermann. 

One  kilo  of  sulphur  in  burning  to  sulphur  dioxide  produces 
2220  calories. 

One  kilo  of  hydrogen  in  burning  to  water  (condensed)  pn> 
duces  34500  calories. 

If  the  water  produced  by  the  burning  of  the  hydrogen  is  not 
condensed,  but  remains  in  the  form  of  steam,  part  of  the  34500 
calories,  produced  by  the  combustion  of  one  kilo,  appears  as  latent 
heat  and  as  sensible  heat  in  the  steam.  Thus,  suppose  one  kilo 
of  hydrogen  and  eight  kilos  oxygen,  both  at  15°  C.  unite  to  form 
nine  kilos  of  steam  which  escapes  at  loo"*  C. 

The  total  heat  of  one  kilo  of  steam  at  100^  C,  measured  from 
water  at  15**  C.  is  622.1  calories,  and  of  nine  kilos,  9  X  622.1  = 
5599  calories,  which  subtracted  from  the  34500  calories  pro- 
duced by  the  combustion  of  one  kilo  of  hydrogen,  leaves  28901 
calories  as  the  available  heat  of  combustion  of  hydrogen  at  15** 
C.  when  the  product  of  combustion  escapes  as  steam  at  100**  C. 

If  the  steam  escapes  at  some  other  temperature,  or  if  the  ini- 

1  One  French  calorie  =3.968  British  thermal  units  :  one  B.  T.  U.  =.0.352  calorie.  The 
"  pound  calorie  "  is  sometimes  used  by  Engrlish  writers  :  it  is  the  quantity  of  heat  re- 
quired to  raise  the  temperature  of  one  ponnd  of  water  i*C,  one  pound  calorie  =  a.2046  B. 
T.  U.  =  I  calories. 

The  heat  of  combustion  of  carbon,  to  COf.  is  said  to  be  8140  calories.  This  figure  is 
used  either  for  French  calories  or  for  pound  calories  as  it  is  the  number  of  pounds  of 
water  that  cau  be  raised  i*C.  by  the  complete  combustion  of  one  pound  of  cartxMi.  or  the 
number  of  kilof^rams  of  water  that  can  be  raised  i*C.  by  the  combustioA  of  one  kilo- 
((ram  of  carbon.    [Kent]. 

<(  (< 

<(  (( 

il  II 


tial  temperature  of  the  hydrogen  is  other  than  is^'C.  the  avail- 
able heat  units  will  vary  accordingly. 

In  practical  calculations  of  the  heating  value  of  fuel,  it  is  gen- 
erally most  convenient  to  take  the  total  calorific  power  of  the 
hydrogen  it  contains  at  34500  calories  per  kilo,  and  after  ob- 
taining the  total  heating  value  of  the  fuel  on  this  basis  to  make 
the  necessary  corrections  for  the  initial  temperature  of  the  hydro- 
gen and  for  the  latent  and  sensible  heat  of  the  steam  in  the 
products  of  combustion. 

The  heating  value  of  coal  is  thus  calculated  from  the  analysis  : 

Let  C^B  the  percentage  of  carbon  in  the  coal. 
Let//^=    ••  **  ••  hydrogen 

Let  (?»    "  •*  "  oxygen 

Let  5=    *•  **  "  sulphur 


„    ,.                       8140  C  +  345oo(/^—  g  )  4-  2220  S. 
Heating  power  » — 

(8140  X  86.49)4-34500(2.36  —  0.345)  -I- 2220X0.18 

~"  TOO 

^704028  4-69517.5-1-  399.6 

=  7739.4  calories  per  kilo  of  coal. 

Where  the   products  of  combustion  of   hydrogen  escape   as 
steam  at  100°  C,  the  formula  will  be  : 

8140 C4-  28901  (//'—  o)  -I-  222o5—  622W 
Heating  power  =» -^ 

W  =  moisture  of  the  coal. 

8140  X  86.49 -I- 28901  (2.36  —  0.345)4-2220X0.18  —  622X3.05 


^  704028.6  4-  5B335.5  4-  399-6  —  1897.3 
"™  100 

=  7645.6  calories  per  kilo  of  coal. 

To  calculate  the  amount  of  air  required  for  complete  com- 
bustion, the  following  data  are  required  : 



(  kilo  of  carbon  burning  to  carbon  dioxide  requires  2.66  kilos  of  oxygen, 
t     "      "  hydrogen     "        "        water  *'       8.00    •'      **         " 

I     **       "sulphur       '•         "  sulphur  dioxide       *'       i.oo    **      **         ** 

Air  is  composed  of  a  mechanical  mixture  of  oxygen  and 
nitrogen  in  the  proportion  by  weight,  of  26.8  parts  of  nitrogen 
with  eight  parts  oxygen  ;  that  is,  3.35  parts  of  nitrogen  with  one 
part  of  oxygen;  or  in  volumes  3.76  cubic  meters  of  nitrogen 
with  one  cubic  meter  of  oxygen. 

The  volume  of  i  kilo  of  oxygen 

t  < 
t  < 





oxygen               is 



meter  at  16.67"  C. 

nitrogen              ** 


i<        ff      it      it 

hydrogen             " 


il                 4  1               it              «t 

sulphur  dioxide  " 


tf                 it              It              tf 

carbon  dioxide   " 


ft                t f               tt              14 



tf                 tt               tf              ft. 

One  kilo  of  carbon  requires  11.6  kilos  of  air  to  produce  car- 
bon dioxide. 

Thus,  the  oxygen  required  2.66  kilos,  which  combined  with 
8.94  kilos  of  nitrogen  (the  proportion  of  oxygen  and  nitrogen  in 
air)  gives  11.6  kilos  of  air  or  9.5  cubic  meters. 

I. o  kilo  carbon  Carbon    i. 00  kilo 

11.6  kilos  air  (Oxygen   2.66  kilos 

^^^  \  Nitrogen  8.94    " 


3.66  kilos  CO, 
Products  of  ^  J    8.94    "     N 
r  combustion 

Total  12.60    *'      J 

12.60    " 

One  kilo  hydrogen  requires  for  combustion  34.8  kilos  of  air^ 
or  28.58  cubic  meters: 

i.o  kilo  hydrogen         Hydrogen  i.o  kilo  "j 

«  Oxygen     8.0  kilos  j 

14.S  kilos  air 



I.  Nitrogen  26.8   ** 


•  t 

!  Products  of  _ 
'1'  combustion 


9.  kilos  H,0 
26.8  **        N 

35.8  " 

In  a  similar  manner  it  it  found  that  one  kilo  of  sulphur  re- 
quires 4.35  kilos  of  air  to  produce  sulphur  dioxide,  or  3.6 
cubic  meters. 

The  amount  of  air  required  for  the  combustion  of  one  kilo  of 
the  coal  will  be  : 

~ .  '-■. 


Carbon        ^  86.49  per  cent.  =  10. a      kilos  c 

Hydrogen  ^    1.36     "      "     =   0.82       " 

Snlphnr     =    0.18     "      ■'     =   0.007      0.003      "  " 

One  kilo  of  Ihe  coal  requires   11.037      "       "     "     "    8.993       "  " 

or  one  pound  of  the  coal  would  require  11,027  pounds  or  144.9 
cubic  feet  of  air  at  62°  F.  for  its  combustion. 

In  a  similar  manner  the  amount  of  air  required  for  the  com- 
bustion of  one  kilo  of  coke  (partial  analysis  given  on  page  115) 
would  be : 

Carbon  94.43  X  11.6  — 10.95  kilos  of  air,  or  8,97  cubic 
meters,  equivalent  to  144-4  cubic  feet  of  air  per  pound  of  the 

The  evaporative  power  of  a  coal  or  coke  expressed  in  kilos  of 
water  evaporated  per  kilo  of  coal,  is  determined  by  dividing  the 
total  heat  of  combustion  of  one  kilo  of  the  combustible  by  620, 
which  is  Ihe  total  heat  (degrees  C)  of  one  kilo  of  steam  at 
atmospheric  pressure,  raised  from  water  supplied  at  62°  F,  or 
16.67° C,  or  by  536.5  (degrees  C.)  if  the  water  is  supplied  at 

If  the  results  are  stated  in  pounds  of  water  evaporated  per 
pound  of  fuel,  it  is  obtained  by  dividing  the  total  heat  of  com- 
bustion in  "B.  T.  U,"  by  11 16.6'  F.,  which  is  the  total  heat  of 
atmospheric  steam  raised  from  water  supplied  at  62°  F. ,  and  by 
dividing  by  956.7°  F.  when  the  water  is  supplied  at  212°  F. 

The  evaporative  value  of  one  kilo  of  the  coal  will  therefore 
be,  theoretically,  assuming  the  water  to  be  supplied  at  16.67°  C. 
(62°F.) ,  and  the  steam  generated  at  atmospheric  pressure: 

Carbon,  86.49  X  8140  -f-  100  = 
Hydrogen,  (2-36— i  X  34500.)-!- « 
Sulphur,  0.18  X  2220  -f-  100  = 

7040.28  calories. 

7739.45  +  610=1 13.48  kilos  ol 

evaporated  per  kilo  of  coal. 






9  8  «'     -  -« 

o  ^^  f 
-5-8    u  o 


a    ^ 
2§  o 


5    o  a 


!?    *  * 


s  S 

2  8. 


•^  8. 

i  8  *s 



^    2 

NO        ^ 

CD      *0 

•^  8. 


o    a 








S  T  ^ 

S  5  ^. 

?  S  P 

^    O  5«  lo 

^    «p  a  o& 

30      "*  «  5 


CD      -, 

8  P 


'C     • 



♦^     ♦*  tt 

q^      4)  41 

£    6  <S 

u     U  V 



s  8;  8 


*0    »• 

I   5 

^  ^  £ 

s  s  » 

u  o  u 

r«  00  O 

*^  ^  NO 

\ri    IS, 

?  '-2   »•" 

U      (J      > 

NO    r^   ^ 

lO    o 







































4,      ^ 

0      0 

o  o 

8  « 

2  2 

0  .2 
0   ^ 

o    S 

1 1 

o  -a 

o    ® 

§  .2 


4»      4» 

0      0 

o  o 



0    o 
0     0 

^  5 


O  <o 

4»  4»  4> 

0  0  O 

O  O  O 



CO  g 

O  bo 

bo  •- 

0  .S 

o  -. 

s  ^ 

JO  H 

0  ? 


0  ? 

«  6 

S  ^ 

0  43 




II  the  water  be  supplied  at  100°  C,  the  evaporative  value 
will  be  7739-45-^536.5;=  14-42  kilos  of  water  evaporated  per 
kilo  of  coal- 

The  actual  evaporation  is  less,  in  boiler  practice,  than  the 
theoretical  as  computed  above,  for  the  following  reasons: 

1.  There  may  be  a  loss  due  to  incomplete  combustion. 

2.  There  is  necessarily  a  considerable  amount  of  heat  carried 
ofi  by  the  chimney  gases. 

3.  There  is  loss  of  heat  due  to  radiation. 

4-  Heat  is  also  lost,  due  to  the  evaporation  of  the  hydroscopic 
moisture  contained  in  the  coal  and  to  the  heat  in  the  vapor 
formed  by  the  combustion  of  the  hydrogen  in  the  coal. 

For  example,  in  a  test  of  a  standard  type  of  boiler  made  by 
Prof.  J-  E.  Denton,  where  the  fuel,  anthracite  coal,  was  burned 
so  thoroughly  as  to  practically  eliminate  the  loss  due  to  incom- 
plete combustion,  the  remaining  losses  were  as  follows  : 

L09&  of  heat  bv  chimney '3-^3  pc  cent- 

"   radiation 3.64     "     " 

Total 16,55     ■'     " 

These  per  cents  being  in  terms  of  total  heat  per  pound  of 
combustible.     Consult  article  on  Boiler  Tests. 

The  total  heat  as  determined  by  calorinietric  measurements 
being  14302.  "B.  T.  U."  per  pound  of  combustible. 

The  heat  imparted  to  the  steam  was  100 —  16.55:=  83-45  P^f 
cent-  of  the  total  heat. 

This  is  a  high  economical  result.  Ordinarily  the  heat 
imparted  to  the  steam  is  not  over  80  per  cent,  of  the  total  heat, 
so  that  the  available  heat  is  usually  less  than  80  per  cent,  of  the 
theoretical  heat. 


Of  the  many  instruments  in  use  in  calorimetry  for  determining 
the  heating  power  of  coals,  the  Mahler,  the  Thompson,  the 
Barrus,  and  the  Carpenter  are  selected  for  description. 

For  rapidity  and  accuracy,  the  Mahler  is  to  be  recommended. 

This  apparatus  consists  of  a  modified  form  of  Berthelot'sbomb. 

Berthelot's  instrument,  which  was  originally  made  for  the 



combustiou  of  gases  under  pressure,  consisted  of  a  steel  C}*Ui 
der  lined  with  platinum. 

Mahler  uses  porcelain  as  a  lining  to  the  steel  cylinder  in  piao 
of  the  platinum,  thereby  materially  reducing  the  cost  of  the  a^ 

The  accompanying  sketches  represent  a  vertical  section  of  t!it 
calorimeter  itself,  showing  all  of  the  attachments,  and  also  a  ver 
tical  section  of  the  shell  to  a  larger  scale.  The  shell  is  forgtd 
out  of  mild  steel  having  a  tensile  strength  of  thirty-one  Ions  in 
the  square  inch,  and  au  elongation  of  twenty-two  per  cent.  I', 
is  about  eight  millimeters  thick  and  usually  weighs  about  3.500 
grams,  with  a  capacity  of  814.6  cc.  The  capacity  of  the  instru- 
ment was  made  much  greater  than  that  of  M.  Berthelot  for  two 
reasons  :  First,  to  insure  complete  combustion,  and  second,  be- 
cause many  gaseous  fuels  used  for  industrial  purposes  contain 
nitrogen  and  carbon  dioxide.  It  is  necessary  to  take  a  large 
quantity  of  them  in  order  to  obtain  a  measurable  rise  in  tem- 
perature. The  shell  is  coated  on  the  inside  with  porcelain  to 
protect  it  from  corrosion  or  oxidation.  The 
porcelain  being  very  thin,  does  not  inter- 
fere with  the  transmission  of  heat.  The 
cover  is  fitted  with  a  ferro-nickel  cock  R. 
with  a  conical  screw  and  stuffing  box  E, 
for  the  introduction  of  oxygen  under  pres- 
sure. The  cover  is  screwed  down  upon  a 
ring  of  lead  P,  placed  in  a  circular  groove 
cut  in  the  rim  of  the  shell,  making  a  tight 
joint.  Through  the  co\-er  passes  an  iso- 
lated electrode,  to  which  a  platinum  rod  is 
fastened  by  means  of  a  clamp.  Another 
platinum  rod  is  fastened  to  the  cover,  and 
the  pan  which  contains  the  substance  to  b 
burned  is  attached  to  this  by 
other  platinum  rod  and  two  clamps, 
tached  to  the  platinum  rods  and  passing 
;  through  the  substance  to  be  burned  is  a 
small  lielix  of  fine  iron  wire.  Ignition  i& 
produced  by  heating  this  wire  white-hotb 



on  i& 

ins  of  a  battery.     The  calorimeter,  the  outer  vessel,  and  the 
various  details  of  M.  Mahler's  apparatus  differ  from  the  analo- 
gous parts  of  M.  Berthelof  5  instrument.     The  calorimeter  is  of 
*^in  brass  and  contains  about  2.j  kilos  of  water.     The  large 
amount  of  water  practically  eliminates  all  error  due  to  evapora- 
tion. '  The  agitator  S  is  worked  by  the  lever  L,  which  pushes 

wn  the  rod  K.  to  which  the  agitator  is  attached.     This  rod 

s  a  spiral  thread  on  it  and  moves  through  a  nut,  so  that  in 

"pressing  it  down  it  also  receives  a  revolving  motion,   thus  very 

thoroughly  stirring  the  water.     The  thermometer  T  should  read 

to  the  one  hundredth  of  a  degree.     For  igniting  the  substance  a 

■jtwtter}-  capable  of  giving  a  current  of  two  amperes   with  an   E. 



M.  F.  of  ten  volts  is  required.  The  oxygen  is  supplied  in  cj'I- 
indera  of  125  cubic  feet  capacity  at  a  pressure  of  150  atmos- 
pheres. Such  a  cylinder  will  supply  oxygen  enough  for  about 
140  determinations. 

Before  this  instrument  can  be  used  for  determining  calorific 
power,  it  is  necessary  to  find  the  water  equivalent  of  the  shell 
and  its  appendages.  This  must  be  determined  with  the  utmost 
care,  as  upon  it  depends  the  correctness  of  all  the  results  after- 

ward obtained.  It  may  be  calculated  directly  from  the  weights 
and  known  specific  beats  of  the  parts.  It  may  also  be  obtained 
experimentally.  The  method  by  calculation  can  only  be  ap- 
proximate, because  of  the  weight  of  the  porcelain  of  the  shell  is 
not  known  and  can  only  be  estimated.  This  method  gives  the 
following  results :  ^m 

Weight  in  Specific       Water  equivaleat^H 

Miitriil.                          ciamt  lust.  in  jimns.        -'^^ 

Hrass  ot  calorinictcv 703'°7  0.094  66.08S           H 

St«el  of  c4lorii»eter 3.3*3-25  0.1165  3S7.157 

PorceUin 134078  <^iT9  »4-o 

rUttnum ji.i  0.0314  0,68 

LwhI 9-0  0.0314  o.aSi 

GU>*  of  thermometer 11.69  0.17968  3>ti4 

Mercury  of  thennomcter  .--        15.03  0.03331  0-833 

Oxygen I9-1»>5  o-"55  3S"3 



In  determining  Ihe  water  equivalent  the  following  method  is 
employed.  The  shell  is  charged  with  oxygen  at  twenty-five  at- 
mosphtres  pressure. 

A  known  weight  of  water,  about  2000  grams,  is  then  placed 
in  the  calorimeter,  the  shell  immersed  in  it,  and  the  whole  appa- 
ratus placed  under  the  same  conditions  that  would  exist  during 
an  actual  combustion.  The  water  is  then  agitated  until  the 
temperature  becomes  constant,  when  about  300  grams  of  water 
at  a  much  lower  temperature  are  added,  and  the  whole  agitated 
until  the  temperature  again  becomes  constant.  Readings  of  the 
thermometer  are  taken  everj-  half  minute.  From  the  observed 
(all  of  temperature  the  water  equivalent  may  be  calculated  by 
means  of  the  following  formula  : 

Let  A'  —  water  equivalent  of  calorimeter  shell  and  appendages. 
/  =  final  temperature  of  water  in  calorimeter. 
/,  —  initial         "  "       •■      ■•         '• 

',  :=  initial  temperature  of  cold  water  added. 
^=.  weight  of  water  in  calorimeter  at  beginning  of  ex- 
(■  =.  weight  of  cold  water  added. 
■  Then  we  have ; 


Y_{t  —  /,)w—(/,-~i)W 

The  results   of    twenty-five  determinations  gives  a   mean  of 
J. 97.  or  practically  490. 

Method  of  Making  a  Detenmnalion  with  Coal. 
About  ten  grams  of  coal  to  be  tested  is  finely  powdered  and 
Missed  through  a  sieve  having  10,000  meshes  to  the  square  inch. 
"it  is  necessary  that  the  coal  be  very  fine  or  it  will  not  bum  com- 
pletely.    The  powdered  sample  is  placed  in  a  glass  weighing 
tube  and  carefully  weighed.      The  platinum  wires  and  pan  are 
attached  to  the  cover  of  the  shell  and  the  iron  wire  helix  placed 
in  position.     A  sample   of   the   coal   is  now   poured   into   the 
pan    from   the    weighing    tube,    its  weight   determined,    care 
rbeing   taken   to   see   that    none   is   spilled   and   that   the   iron 


wire  helix  passes  throagh  the  coal.     The  cover  is  then  placed 
on  the  shell  and  screwed  down  firmly.     The  shell  is  now  con- 
nected with  the  oxygen  cylinder,  and  the  oxygen  allowed  to 
flow  in  until  the  gauge  shows  a  pressure  of  about  twentv-fire 
atmospheres.     The  stop-cock  is  then  closed  and^ the  shell  placed  in 
the  calorimeter,  which  has  been  previously  partially  filled  with 
about  2,400  grams  of  water.     The  thermometer  and  agitator aie 
adju.sted,  and  the  whole  well  stirred  to  obtain  a  uniform  tem- 
perature.    The  temperature  is  then  observed,  from  minute  to 
minute,  for  four  or  five  minutes,  so  as  to  determine  its  rate  of 
change.     The  charge  is  then  ignited  by  connecting  one  pole  of 
the  batter>'  to  the  electrode  P,  and  touching  the  other  pole  to 
any  part  of  the  shell.     The  temperature  is  observed  each  minute 
until  it  begins  to  fall  regularly,  and  then  each  minute  for  five 
minutes  in  order  to  ascertain  the  law  of  cooling.     The  agitator 
should  be  kept  going  constantly  during  the  whole  period  of  the 
observation.     The  shell  is  now  removed  from  the  water,  the 
stop-cock  R  opened  to  let  out  the  gas,  and  then  the  shell  itself 
is  opened.     The  shell  should  be  rinsed  out  with  distilled  water 
to  collect  the  acid  formed  during  combustion.     The  amount  of 
acid  carried  out  with  the  escaping  gas  is  negligible.     The  calo- 
rific power  of  the  coal  may  now  be  calculated  as  follows  : 
I^et  Q  =  calorific  power  of  the  coal. 

J  =  observed  rise  of  temperature. 

X  =  correction  for  radiation. 

P  =  weight  of  water  taken  in  grams. 

P'=z  water  equivalent  of  shell,  appendages,  and  gas. 

/>  =  weight  of  nitric  acid  found. 

/>'  =:  weight  of  iron  wire  helix. 

0.23  calorie  =  heat  of  formation  of  one  gram  of  nitric  acid. 

1.6  calories  ~  heat  of  combustion  of  one  gram  of  iron. 

iJ^  {J  +  x)  (/>+/>)  _(o.23/+  1.6/'). 
/ixtifNp/c  S/iouhti^  Mv^/tod  of  Ca/cu/aiton. 
1.042  gram  of  coal  was  taken. 

The  calorimeter  contained  2,276.6  grams  of  water. 
The  water  equivalent  of  apparatus  =  490  grams. 



1?lie  pressure  of  oxygen  =  25  atmospheres. 
I^he  law  of  variation  of  temperature  in  the  calorimeter  before 
i^oixi bastion  is  expressed  by  X^  =  O, 

law  of  variation  during  subsequent  period  is 

X^  =  27.46  —  27.395  =  0.065''  c. 

ence,  during  the  period  of  combustion  the  system  lost  0.065 

by  radiation, 
lie  apparent  variation  of  temperature  is 

27.460  —  24.855**  =  2.605'  c. 

.ctual  variation  =  2.605°  +  0.065**  =  2.67**  C. 
'he  nitric  acid  formed  =  0.15  gram.     And  the  weight  of  iron 
=  0.025  gram.     Hence,  heat  of  formation  of  nitric  acid  = 
^-  "K  ^  X  0.23  =  0.0345  calorie,  and  heat  of  combustion  of  wire  = 
•35  X  1.6  =  0.04  calorie.     Heat  of  combustion  of  coal  =  2.67 
(2,276.6  +  490), 

=  7,386.8  calories. 
7,386.6  —  (0.0345  +  0.04)  =  7,386.72 

-J-  1.042  =  7,088.9 

7088.9  calories  per  kilo.  =  12760.  B.  T.  U.  per  pound  of  coal. 

To  show  the  accuracy  with  which  this  calorimeter  works,  five 
^mples  of  willow  charcoal  were  burned,  with  the  following  re- 

Average  of  five  determinations 7973  calories  per  kilo. 

Highest  determination 7975       **  " 

Lowest  determination 7971      "  "      *' 

Five  determinations  of  a  sample  of  bituminous  coal  from  Cole- 
man County,  Texas,  gave  as  follows : 

Average  of  five  determinations 6766.0  calories  per  kilo. 

Highest  determination 6793.6 

Lowest  determination 6720.3 

References. — "  On  the  Berthelot-Mahler  Calorimeter  for  the  Calorific 
Power  of  Fuels."  Prof.  A.  M.  Mayer.  Stevens'  Indicator ^  April,  1895, 
p.  135-148. 

'*  Zur  Werthbestimmung  der  Brennstoffe."  (Verfahren  und  Calorime- 
ter von  Mahler,  Bnnte,  Fischer,  Scheurer-Kestner),  Stahl  und  Eisen,  13, 

Determination  industrielle  du  pouvoir  calorifique  des  combustibles. 
Mahler.    La  Sucrerie  Jndi^^ne,  Aii  443. 

((         (<      It 



This  instrument,  in  general  use  in  England  for  caloriiuetii 
determinations  of  solid  fuels,  is  shown  in  Fig.  36. 

The  water  equivalent  (theoretical)  of  the  calorimeteris  found 
by  weighing  each  part  carefully  and  multiplying  by  its  specific 

Thus  J 

Specific          wxer  ^M 

Weight.                           E«l.        cqnlTBlenl.  ^M 

'^ruS'vf,;',i:t',i; } »"-» «»«■• "  »■■"«  -  *-«  ■ 

Gtan  bell  jar 7S-J*i             X  0.19      =    14.3'3  ^| 

BraMUae 9*353     ■*       X  0.09391  =      9.377  ^1 

I^our  copper  dtoks 65.100              X  0.0951s  ^      6.J94  ^^M 

Brass  ov«r  top  ol  )k1)  \tT-    n-ynj             X  Oi.09ji9i  ^     3.001  ^^M 

Copperiube joAn             X  0.0M"  =     '-9ao  ^| 

Kubbercork 1.578             X  0^331      =     0.552  ^H 

Rubber  tube 1,7*4             x  0.331     =     0.591  ^" 

nitiann  crttciUc is^tit      "      X  0.314     =>     0,490 

Uncnrr  o(  iber«tOMCt«r-      9~sS3  X  4-333     ~     0^319 

Gtw  ot  therwwawWr  -  -  -  -      7.35a  X  olK)      =     '■>96 

T»m »o-478 



This  theoretical  water  equivalent  should  be  checked  by  a  de- 
termination by  direct  experiment,  as  follows  : 

The  calorimeter  is  taken  and  adjusted  under  the  conditions  of 

2000  grams  of  distilled  water  are  weighed  out  and  the  temper- 
ature taken  :  call  this  temperature  /. 

Let  i,  =:  temperature  of  the  apparatus. 

The  2000  grams  of  water  are  poured  into  the  glass  cylinder 
ab.  Fig.  36,  the  other  parts  c.  d,  g.  It.  etc. ,  placed  in  position  in- 
side the  cylinder,  and  the  water  kept  well  stirred  by  means  of 
the  discs  K.  K.  on  the  side  of  the  bell  jar. 

After  agitating  it  about  fifteen  minutes  (about  the  time  re- 
quired for  a  coal  combustion)  the  temperature  is  taken  ;  this 
temperature  call  /„.  To  correct  for  radiation  it  is  necessary  to 
continue  this  operation  for  an  equal  period  of  time,  calling  the 
last  temperature  c,  from  which  we  obtain  the  fall  of  temperature 
to  be  (l^  —  c)  ^  r.     Expressing  this  in  a  formula 

2000  (/—(/,  +  /■)  ,  .     ,     , 

— --; — ; ;^ —  watcr  equivalent 

r  being  the  fall  of  temperature  due  to  radiation. 

Tetnperalttre  of  apparatus  ^  14.6"    C. 
"  water  =  19.5°    C- 

Final     "  "      "      =         18.65"  C. 

Correction  f  18.65  —  18.3)  =0.35. 

3W»(l9-5—  '9'0)    =  221.12. 

ig  —  14.6 
By  calculation  the  water  equivalent  is  210.47. 
■'    eitperimeiil  "        ■'  "  "  1*7. »;. 

The  combustion  with  a  sample  of  coal  is  performed  as  follows : 
An  incandescent  paper  (about  one  mm.  long)  is  dropped  into 
the  crucible  (/)  containing  one  gram  of  the  very  Jindy  pulver- 
ized coal,  the  oxygen  supply  being  slowly  turned  on  and  the  in- 
verted bell  jar  (/)  containing  the  crucible  (/)  is  gently  lowered 
into  the  2000 grams  of  watercontained  in  theglass cylinder  (oi). 
The  combustion  will  be  quite  active  i  the  gaseous  products 
will  bubble  through  the  water  and  give   up  their  sensible  heat. 
After  the  fuel  has  been  consumed  the  supply  of  oxygen  is 
stopped  and  the  glass  tube  c  is  opened,  permitting  the  water  to 


enter  the  bell  jar  and  flow  into  and  submerge  the  crucible  so 
that  the  whole  of  the  apparatus  and  water  is  raised  to  a  uniform 

It  will  be  noted  that  the  coal  bums  gently  at  first.  The  oxy- 
gen introducing  pipe  ig  h)  should  not  be  projected  too  low  into 
the  bell  jar  until  the  volatile  hydrocarbons  are  consumed  ;  the 
residual  fixed  carbon  is  more  difficult  to  bum. 

The  oxygen  supply  tube  should  consequently  be  projected  so 
as  to  deliver  the  oxygen  immediately  over  the  platinum  cruci- 
ble, and  to  more  effectually  burn  the  fuel,  the  tube  may  be 
slightly  rotated. 

Great  care  must  be  taken  in  reading  the  thermometer  before 
and  after  the  gram  of  coal  is  burned :  the  difference  of  these 
two  readings  gives  the  rise  in  temperature  for  the  amount  of 
coal  taken,  which  when  multiplied  by  2000  plus  the  water 
equivalent  of  the  calorimeter,  gives  the  heating  power  of  the 

But  since  heat  is  being  radiated  to  the  air  during  the  experi- 
ment, a  correction  must  be  made.  To  determine  this,  it  is 
necessary  to  note  the  time  required  to  bum  the  coal,  and  then 
agitate  the  apparatus  for  a  corresponding  period.  During  this 
last  agitation  the  temperature  will  fall  somewhat ;  this  fall,  di- 
vided by  two,  will  give  tho  proper  correction. 

The  figure  obtained  is  an  average  of  the  whole  radiation  : 
should  the  fall  be  taken  direct,  it  would  give  the  correction  for 
radiation  when  the  water  is  at  its  maximum  temperature.  The 
following  is  the  analysis  of  a  sample  of  coal,  the  theoretical 
heating  power  calculated  from  the  analysis,  and  calorimetric  de- 
termination of  the  coal  by  means  of  the  Thompson  calorimeter, 
and  a  comparison  of  the  number  of  calories  per  kilo  derived  by 
calculation  and  by  direct  experiment. 

Analysis : 

Carbon 84.80  per  cent. 

Hytlrojjen 2.42 

Sulphur 0.62 

Nilroj^on 0.93 

^[oislu^c 1.03 

Oxygon 3.19        " 

Ash 7.01 


Total 100.00 




the  theoretical  heating  value  being  : 

(8140  X  84.8)  +  (34500  X  2.42  —  0.4)  +  2220  X  0.62 

_ =  75,3 

calories  per  kilo  of  coal. 

The  test  of  the  coal  by  the  Thompson  calorimeter  gave  as  fol- 

Amonnt  of  coal  taken  ^  0.445  gram. 

Temperature  of  water  and  apparatus  (initial) '8.95'^  C. 

Maximum  temperature  "  •'  ^o-Ab"  C. 

Filial  temperature  (used  for  correction  o(  radiation)  10.40°  C. 

Correction  for  radiation  \  ^°45  —  ao.4Q^  _  o.oas'^  C. 



Rise  in  temperature  for  0.445  gram  =  1.525'^'  C. 

■'      ■'  "  ■'      ■■      =  3-43^0. 

1337  X  3.43  =  7638.6  calories  per  kilo  of  coal. 

The  complete  apparatus  is  shown  in  the  accompanying  figure 
(37),  The  calorimeter  itself  consists  of  a  glass  vessel  five 
inches  in  diameter,  nine  and  a  half  inches  high,  which  holds  the 
water  of  the  calorimeter.  Submerged  in  the  interior  is  a  bell- 
shaped  glass  vessel  two  and  a  half  inches  in  diameter,  four  inches 
high,  having  a  long  neck  three-fourth  of  an  inch  in  diameter, 
which  is  closed  at  the  top  with  a  stopper. 

The  upper  end  of  the  neck  stands  five  inches  above  the  top  of 
the  outside  vessel.  The  glass  bell,  or  "combustion  chamber,"  as 
it  may  be  termed,  rests  upon  a  metal  base,  to  which  it  is  held 
by  means  of  spring  clips,  the  bottom  of  the  chamber  being  pro- 
vided with  an  exterior  rib  by  means  of  which  the  clips  are  made 
fast.  The  base  is  perforated,  and  at  the  center  is  mounted  a 
short  lube,  for  the  reception  of  a  crucible  in  which  the  combus- 
tion takes  place.  The  crucible  is  made  of  platinum.  It  is  sur- 
rounded by  a  layer  of  non-conducting  material,  which  is  placed 
between  it  and  the  outer  metal.  A  small  glass  tube  is  inserted 
in  the  stopper  at  the  top  of  the  neck,  and  this  is  carried  down  to 
the  interior  of  the  combustion  chamber.  It  is  fitted  somewhat 
loosely,  so  that  a  slight  pressure  will  move  it  up  or  down,  and 
thereby  adjust  its  lower  end  to  any  height  desired  above  the 
crucible.     The  tube  has  a  slight  lateral  movement  also,  so  that 

1  TnnuctiDDi  American  Sodrty  Mechanical  Enstnecra,  14,  816, 



tay  be  directed,  at  the  will  of  the  operator,  toward  any  part 
of  the  crucible. 

This  tube  is  connected  with  a  tank  containing  oxygen  gas, 
and  through  it  a  current  of  gas  is  passed,  so  as  to  enable  the 
combustion  of  the  coal  to  be  carried  on  under  water. 

The  pressure  of  the  gas  drives  out  the  water  which  would 
otherwise  fill  the  chamber,  and  keeps  its  level  between  the  base. 
The  products  of  combustion  rising  from  the  crucible  pass  down- 
ward through  the  perforations  in  the  base,  escaping  around  Che 
edge  of  the  base,  and  finally  bubbling  up  through  the  water  and 
emerging  at  its  surface.  A  wire  screen  is  secured  to  the  neck 
of  the  combustion  chamber,  extending  to  the  sides  of  the  outer 
%'esM],  thereby  holding  back  the  gas  and  preventing  its  imme- 
diate escape  to  the  surface  of  the  water. 

In  making  the  test  the  quantity  of  water  used  is  2000  grams 
and  the  quantity  of  coal  one  gram.  The  equivalent  colorific 
^-alue  of  the  material  of  the  instrument  is  1S5  milligrams  (0.1S5 

One  degree  rise  of  temperature  o(  the  water  corresponds, 
therefore,  to  a  total  heat  of  combustion  of  2185  B,  T.  U.  The 
number  of  degrees  rise  of  temperature  for  ordinary  coals  varies 
from  5  5°  to  6.5°  F. 

The  thermometer  used  for  determining  the  temperature  of  the 
water  is  graduated  to  twentieths  of  a  degree  :  and  as  the  divi- 
sions are  about  one-thirtieth  of  an  inch  apart,  they  may  be  sub- 
diWded  by  the  eye  so  as  to  readily  obtain  a  reading  to  hun- 
dredths of  a  degree. 

The  scales  shown  at  the  extreme  left  of  the  cut  are  used  for 
weighing  out  the  water,  and  the  chemical  scales  shown  in  the 
center  are  employed  in  weighing  the  coal  and  ash. 

The  process  of  making  a  test  is  as  follows  : 

Having  dried  and  pulverized  the  coal,  and  weighed  out  the 
desired  quantities  of  coal  and  water,  the  combustion  chamber  is 
immersed  in  the  water  for  a  short  time,  so  as  to  make  the  tem- 
perature of  the  whole  instrument  uniform  with  that  of  the  water. 
On  its  removal  the  initial  temperature  of  the  water  is  obser\'ed, 
I  of  the  chamber  lifted,  the  gas  turned  on.  and  the  coal 
kly  lighted,  a  small  paper  fuse  having  been  previously  in- 



serted  in  the  crucible  for  this  purpose.  The  top  of  the  combus- 
tion chamber  is  quickly  replaced,  and  the  whole  returned  to  its 
submerged  position  in  the  water.  The  combustion  is  carefully 
watched  as  the  process  goes  on,  and  the  current  of  oxygen  is 
directed  in  such  a  way  as  to  secure  the  desired  rate  and  condi- 
tions for  satisfactory  combustion.  When  the  coal  is  entirely 
consumed,  the  mterior  chamber  is  moved  up  and  down  in  the 
water  until  the  temperature  of  the  whole  has  become  uniform, 
and  finally  it  is  withdrawn  and  the  crucible  removed.  The  final 
temperature  of  the  water  is  observed,  and  the  weight  of  the  re- 
sulting ash. 

The  initial  temperature  of  the  water  is  so  fixed  by  suitably 
mixing  warm  and  cold  water  that  it  stands  at  the  same  number 
of  degrees  below  the  temperature  of  the  surrounding  atmosphere 
(or  approximately  the  same)  as  it  is  raised  at  the  end  of  the 
process  above  the  temperature  of  the  air.  In  this  way  the  effect 
of  radiation  from  the  apparatus  is  overcome  so  that  no  provision 
in  the  matter  of  insulation  is  required,  and  no  allowance  needs 
to  be  made  for  its  effect. 

Results  op  Tbsts  with  tub  Barrus  Coai«  Caix>rimbtbr. 

Cumberland  Coals, 

Kind  of  coal  :  Mine  or  locality 

•*  JS 

=  s 

o  d 
JS  s 
p  o 

1       7.6 

2        8.2 

;     Gci^rge*s  Creek 6.1 

4  *'  *      6.6 

5  ••  ••      8.6 

6  ••  •'      6.5 

:  *  *     7.0 

8  *•            **      ;  American  Co. '^ 5.0 

Q  "      vMd.  Coal  Co."* 5.1 

K*  *•      v^»-  C.  Coal  and  Iron  Co.>  5.7 

II      6.1 

\2     Oe\'>rj^?'s  Creek 5.1 

^^      ••••    7-5 

14      Kureka 5.1 

15         '     H 

10      S. 

17     v.Vx>i>:e's  Creek 4^ 

13.868  B.  T.  U 

14,217  " 

13.925  " 
12.874  " 
12,921  " 

13.3^  •' 
I3.4«7  " 
13.656  " 

13.424  " 

13.534  " 

13*745  " 
13.617  " 

13.653  " 
13.427  " 
12,973  " 
"3.923  " 

«  <( 

<  II 

C  <( 

(  «< 

<  «( 
.  tc 

<  (< 
(  «< 

•  •• 
«  <• 
«     «l 

<  «« 

<  4C 

carpenter's  coai.  calorimeter. 


Fischer's  calorimeter,  while  somewhat  more  complex  than  the 
Mahler  or  Thompson's,  is  an  accurate  instrument  for  the  deter- 
mination of  the  heating  power  of  fuels.  Consult  Chemische 
TfihnologU  der  Brennstoffe,  von  Dr.  Ferdinand  Fischer,  p.  401- 
carpenter's  coal  calorimeter. 

R.  C,  Carpenter'  has  devised  a  calorimeter  for  the  determina- 
tion of  the  heating  power  of  coals,  which  is  thus  described.  The 
general  appearance  of  the  instrument  is  shown  in  Fig.  38,  a  sec- 
tional \new  of  the  interior  is  shown  in  Fig.   39.  from  which  it  is 

seen  that,  in  principle,  the  instrument  is  a  large  thermometer, 
in  the  bulb  of  which  combustion  takes  place,  the  heat  being  ab- 
sorbed by  the  liquid  which  is  within  the  bulb.  The  rise  in  tem- 
perature is  denoted  by  the  height  to  which  a  column  of  liquid 
rises  in  the  attached  glass  tube. 

In  construction,  Fig.  39,  the  instrument  consists  of  a  chamber, 




No.  15,  which  has  a  removable  bottom,  shown  in  section  in  Fig. 
39  and  in  perspective  in  Fig.  40.  The  chamber  is  supplied 
with  oxygen  for  combustion  through  tube,  23,  24,  25,  the  prod- 
ucts of  combustion  being  discharged  through  a  spiral  tube,  29, 
28.  30. 

Surrounding  the  combustion  chamber  is  a  larger  closed  cham- 
ber, I,  Fig.  38,  filled  with  water,  and  connecting  with  an  open 
glass  tube,  9  and  10.  Above  the  water  chamber,  i,  is  a  dia- 
phragm, 12,  which  can  be  placed  in  position  by  screw,  14,  so  as 
to  adjust  the  zero  level  in  the  open  glass  tube  at  any  desired 
point.  A  glass  for  observing  the  process  of  combustion  is  in- 
serted at  33  in  top  of  the  combustion  chamber,  and  also  at  34  in 
top  of  the  water  chamber,  and  at  36  in  top  of  outer  case. 

This  instrument  readily  slips  into  an  outside  case,  which  is 
nickel  plated  and  polished  on  the  inside,  so  as  to  reduce  radia- 
tion as  much  as  possible.  The  instrument  is  supported  on  strips 
of  felting,  5  and  6,  Fig.  39.  A  funnel  for  filling  is  provided  at 
37,  which  can  also  be  used  for  emptying,  if  desired. 

The  plug  which  stops  up  the  bottom  of  the  combustion  cham- 
ber carries  a  dish,  22,  in  which  the  fuel  for  combustion  is  placed  ; 
also  two  wires  passing  through  tubes  of  vulcanized  fiber,  which 
are  adjustable  in  a  vertical  direction  and  connected  with  a  thin 
platinum  wire  at  the  ends.  These  wires  are  connected  to  an 
electric  current  and  used  for  firing  the  fuel.  On  the  top  part  of 
the  plug  is  placed  a  silver  mirror.  38,  to  deflect  any  radiant  heat. 
Through  the  center  of  this  plug  passes  a  tube,  25,  through 
which  the  oxygen  passes  to  supply  combustion.  The  plug  is 
made  with  alternate  layers  of  rubber  and  asbestos  fiber,  the  out- 
side only  being  of  metal,  which,  being  in  contact  with  the  wall 
of  the  water  chamber,  can  transfer  little  or  no  heat  to  the  out- 

The  discharge  gases  pass  through  a  long  coil  of  copper  pipe, 
and  are  discharged  through  a  very  fine  orifice  in  a  cap  at  30. 

The  instrument  has  been  so  designed  that  the  combustion 
can  take  place  in  oxygen  gas  having  considerable  pressure,  but 
in  pressure  it  has  been  found  that  very  excellent  results  have 
been  obtained  with  pressures  of  two  to  five  pounds  per  square 

cahpenter's  coal  calorimeter. 

inch,  and  these  having  been  commonly  used  in 
the  determinations. 

Two  instruments  have  been  built  at  the  pres- 
ent time,  which  difier  from  each  other  some- 
what in  detail,  but  principally  in  dimensions. 
The  first  instrument  held  about  one  pound  of 
water,  and  was  intended  for  use  with  about  one 
gram  of  coal.  In  thai  instrument  the  entire 
bottom  of  the  water  chamber  was  removable 
and  the  whole  of  the  combustion  chamber. 
This  form,  while  giving  full}'  as  good  results  as 
the  one  described,  was  more  likely  to  leak,  and, 
consequently,  was  difficult  to  keep  in  good  con- 
dition. The  first  form  built  employed  an  ad- 
justing piston  to  regulate  the  initial  heading  of 
the  water  column,  which,  possibly,  may  have 
been  as  good  as  the  diaphragm  used  at  pres- 

The  instrument  described,  which  is  of  later 
design,  holds  about  five  pounds  of  water,  and  is 
large  enough  for  the  consumption  of  two  grams 
of  coal. 

Full  details  for  manipulation  of  the  apparatus 
are  given  in  Trans.  Amer.  Society  Mechanical 
Engineers,  Vol.  XVI,  (1895). 

Jie/erences. — ■'  Uberdie  Bestimmung des  Heizwertlies  der  festen  Brenn- 
materialtea  und  Bericht  iiber  die  wichtigere  neure  Litteratur  dieses 
Gcbietes ;"  von  Koorre,  Die  Chemische  Industrie,  17,  93. 

'■  Etude  sur  les  combustibles  et  la  combustion,"  Vivien.  La  Sucrarie 
indigitu.  44,  ]6i. 

Determination  of  the  heating  power  of  coal  by  the  use  of  large 
^^  amounts  of  coal  either  (a)  in  specially  eonstnicied  apparatus 

^K  for  the  same,  or  {6)  under  boilers  in  actual  practice. 

^^B  Apparatus  (or  determining  the  heating  value  of  Fuel,  by  Wm. 
Hkent.  M.£.,  (Fig.  41). 

^H  Its  principal  feature  is  that  it  is  not  a  steam  boiler  but  a  water 
^^Bieater.  It  consists  of  two  sheet-metal  cylinders,  each  twelve 
^^^feet  long,  the  upper  one  four  feet  in  diameter  and  the  lower  one 
^M  three  feet,  and  connected  by  a  short  neck  at  one  end  only. 



The  Upper  cylinder  is  provided  with  a  fire-box  three  and  a 
half  feet  in  diameter  and  six  feet  long,  and  its  rear  end  is  filled 
with   about   lOO  two-inch  tubes.     The  lower  cylinder  is  com- 

pletely filled  with  two-inch  tubes.  The  fire-box  is  lined  through- 
out with  fire-brick,  and  contains  a  grate  surface  two  feet  wide 
by  two  and  a  half  feet  long.  A  hanging  bridge-wall  of  fire- 
brick is  placeii  in  the  upper  part  of  the  fire-box  in  the  rear  o£fl 


the  bridge-wall  proper,  for  the  double  purpose  of  presenting  a 
hot  fire-brick  surface  to  the  flame  before  allowing  it  to  touch  the 
heating  surfaces  of  the  tubes  and  tube-sheet,  and  of  changing 
iLi  direction  so  as  to  cause  the  gases  to  thoroughly  commingle, 
and  thus  to  iusure  complete  combustion.  In  testing  highly 
bituminous  coals,  it  might  be  advisable  to  have  more  than  one 
of  these  hanging  walls,  and  to  give  the  fire-box  a  greater  length, 
to  more  certainly  insure  complete  combustion  of  the  gases.  The 
gases  of  combustion  pass  through  the  tubes  of  the  upper  heater, 
then  down  through  a  fire-brick  connection  into  the  tubes  in  the 
lower  heater,  after  leaving  which  they  pass  into  the  chimney. 
Air  is  fed  to  the  fire,  under  the  grate-bars,  through  a  pipe  lead- 
ing from  a  tan-blower.  The  air  is  measured  by  recording  the 
revolutions  of  the  blower,  and  the  measurement  is  checked  by 
an  anemometer  in  the  air-pipe.  Its  weight  should  be  calculated 
from  the  barometric  pressure,  and  its  contained  moisture  should 
also  be  determined.  Its  temperature  should  be  taken  before  it 
enters  the  ash-pit. 

The  temperature  of  the  escaping  gases  should  be  taken  by  sev- 
eral thermometers,  the  bulbs  of  which  reach  to  different  por- 
tions of  the  chimney  connection.  Cold  water  is  supplied  to  the 
bo«om  of  the  lower  heater,  at  the  chimney  end,  its  temperature 
being  taken  before  it  enters  by  a  thermometer  inserted  in  the 
pipe.  The  water  supply  pipe  may  be  conveniently  attached  to 
the  city  main.  The  water  passes  through  the  two  heaters  in 
an  opposite  direction  to  that  of  the  gases  of  combustion,  and 
escapes  at  the  outlet  pipe  at  the  top  of  the  upper  heater  by  which 
it  is  taken  to  two  measuring  tanks,  which  are  alternately  filled 
and  emptied.  The  temperature  of  the  outflowing  water  is 
taken  by  a  thermometer  inserted  in  the  overflow  pipe.  The  rate 
of  flow  of  water  through  the  apparatus  is  regulated  so  that  the 
temperature  of  the  outflowing  water  does  not  exceed  200'  F, 
The  measuring  tanks  have  closed  tops,  which  prevent  evapora- 
tion, small  outlet  pipes  being  attached  to  the  top  of  each,  which 
serve  both  as  indicators  when  the  tanks  are  full,  and  to  allow  air 
to  escape  from  the  tank  when  it  is  being  filled  with  water. 

The  grate  surface  being  only  five  square  feet  and  the  heating 


surface  about  looo  square  feet,  the  ratio  of  200  to  i,  or  more 
than  five  times  the  usual  proportion  in  a  steam  boiler,  and  the 
water  being  much  colder  than  that  in  a  steam  boiler,  the  gases 
of  combustion  should  be  cooled  down  to  near  the  temperature  of 
the  air  supplied  to  the  fire,  especially  when,  as  is  usually  the 
case,  the  water  supply  is  colder  than  the  air.  For  extremely 
accurate  tests,  the  water  might  be  cooled  before  entering  by  a 
refrigerating  apparatus  or  by  ice. 

The  whole  apparatus  being  thoroughly  protected  by  felting 
from  radiation,  the  heat  generated  by  the  feel  is  all  measured  in 
the  increase  of  heat  given  to  the  water  which  flows  through  the 
apparatus,  and  in  the  increase  of  temperature  of  the  gases  of 
combustion  as  taken  in  the  chimney,  over  the  temperature  of 
the  air  supplied  to  the  fire.  This  increase,  however,  being  in 
any  case  very  slight,  and  the  quantity  of  air  being  known,  the 
amount  of  heat  from  the  fuel  which  escapes  up  the  chimney  can 
be  calculated  with  but  small  chances  of  error. 

Boiler  Test. 
R^uM^  OF  Tests  Upon  Babcock  &  Wilcox  Boilers.' 

«  S  »E      5V-2  0    1-0  -      3=  S 

..^.Mco...    I    1^    ss   I  pJL=-MI  SI'S  I    L 

£l    ul    ssr    It  sli.£sis,$Ss  ill  si  is 

Anthracite       1 

ton.'pl'.  S^mT  ■"*  ^■°  3"*  ^-^  "■*  •'■3'  "-a"  "■*»  "72  448 
bitum  i.  I 

Jackson,  O.,  nut     8    48.0    3358      9.6    31.1    4. 11      8.93      9.88    262    460 

Castle  Shan'n  ) 

Pa.    i   nut,    ;  \  4il     69.1     4784     10.5     37.9    4.13     10,00     IJ.J7    416    570 

lump.  ) 

CanliEf,  lump  -.  6]    «,»    1564      11.7    at.y    3.69    lojj?    11.40    136    189 

I  Tnni.  Amcr.  Soc.  MmhiLD.  UnKinecr*.  4,  K^. 

■  Thctrrm  ")>cr  pound  ol  cvmbuitihlc"  reprcHBii  otw  pound  of  llw  heatliia;  con- 


Approximatb  Heating  Value  of  Coals.  (Kent.) 




























16. 1 1 
















The  use  of  the  table  may  be  shown  as  follows : 

Given  a  coal  containing  moisture  two  per  cent.,  ash  eight  per 
cent.,  fixed  carbon  sixty-one  percent.,  and  volatile  combustible 
matter  twenty-nine  per  cent.,  what  is  its  probable  heating 
value  ? 

Deducting  moisture  and  ash  we  find  the  fixed  carbon  is  61.90 
or  sixty-eight  percent,  of  the  total  fixed  carbon  and  volatile 
combustible  matter. 

One  pound  of  coal  dry  and  free  from  ash  would,  by  the  table, 
have  a  heating  value  of  15480  thermal  units,  but  as  the  ash  and 
moisture  having  no  heating  value,  are  ten  per  cent,  of  the  total 
weight  of  the  coal,  the  coal  would  have  ninety  per  cent,  of  the 
table  value,  or  13932  thermal  units.  This  divided  by  966,  the 
latent  heat  of  steam  at  213°  F. ,  gives  an  equivalent  evaporation 
per  pound  of  coal  of  14,42  pounds. 

The  heating  value  that  can  be  obtained  in  practice  from  this 
coal  would  depend  upon  the  efficiency  of  the  boiler,  and  this 
largely  upon  the  difficulty  of  thoroughly  burning  the  volatile 
combustible  matter  in  the  boiler  furnace.  If  a  boiler  efficiency 
of  sixty-five  per  cent,  could  be  obtained,  then  the  evaporation 
per  pound  of  coal  from  and  at  212'  F.  would  be  14.42  X  0.65  ^ 
9-37  pounds. 



With  best  anthracite  coal,  in  which  the  combustible  portiot 
is,  say  ninety-seven  per  cent,  fixed  carbon  and  three  per  cent,  ' 
volatile  matter,  the  highest  result  that  can  be  expected  in  a 
boiler  test  with  all  conditions  favorable,  is  12.2  pounds  of  water 
evaporated  from  and  at  212°  F,  per  pound  of  combustible,  which 
is  eighty  per  cent,  of  15. 28  pounds,  the  theoretical  heating 

With  the  best  semi-bituminous  coals,  such  as  Cumberland 
and  Pocahoutas,  in  which  the  fixed  carbon  is  eighty  per  cent,  of 
the  total  combustible,  12.5  pounds,  or  seventy-six  per  cent,  of 
the  theoretical  16.4  pounds  may  be  obtained. 

For  Pittsburgh  coal,  with  fixed  carbon  ratio  of  sixty-eight  per 
cent.,  eleven  pounds,  or  sixty-nine  per  cent,  of  the  theoretical 
16.03  pounds,  is  about  the  best  practically  obtainable  with  the 
best  boilers. 

With  some  good  Ohio  coals,  with  a  fixed  carbon  ratio  of  sixty 
per  cent.,  ten  pounds,  or  sixty-six  per  cent,  of  the  theoretical 
15.9  pounds  has  been  obtained  under  favorable  conditions,  with 
a  fire-brick  arch  over  the  furnace  with  coals  mined  west  of  Ohio; 
with  lower  carbon  ratios,  the  boiler  efficiency  ia  not  apt  to  be  as 
high  as  sixty  per  cent. 

From  these  figures  a  table  of  probable  maximum  boiler  test 
results  from  coals  of  different  fixed  carbon  ratios  may  be  con- 
structed as  follows  : 

Fixed  carbon  ratio 97.0    80,0    68.0    fio.O    54.0    50.0 

Evaporated  from  and  at  Jii^  F.  per 
pound  combustible,  maximum  in  boil- 
er tests 15.1     "5     "i-O     10.0      8,3       7.0 

Boiler  efficiency,  percent 80.0    76.0    69.0    GG.o    6o.a    SS-|> 

Loss,  chimney  radiation,  imperfect  com. 

Imstion,  etc 20.0    24.0    31.0    34.0    40.0    45.0 

The  difference  between  the  loss  of  twenty  per  cent,  with  an- 
thracite and  the  greater  losses  with  the  other  coals  is  chiefiy  due 
to  imperfect  combustion  of  the  bituminous  coals,  the  more  highly 
volatile  coals  sending  up  the  chimney  the  greater  quantity  of 
smoke  and  unbumed  hydrocarbon  gases.  It  is  a  measure  of 
the  inefficiency  of  the  boilerfurnace  andof  the  inefficiency  of  heat- 
ing surface  caused  by  the  deposition  of  soot,  the  latter  being 
primarily  caused  by  the  imperfection  of  the  ordinary  fun 

^■^     ▼  BOILER   TEST. 

and  its  unsuitability  to  the  proper  burning  of  bituminous  coal. 
If  in  a  boiler  test  with  an  ordinary  furnace  lower  results  are  ob- 
tained than  those  in  the  above  table,  it  is  an  indication  of 
unfavorable  conditions,  such  as  bad  firing,  wrong  proportions  of 
boiler,  detective  draft,  and  the  like,  which  are  remediable. 
Higher  results  can  be  expected  only  with  gas  producers,  or 
other  styles  of  fumaceespecially  designed  for  smokeless  combus- 

The  efficiency  of  a  boiler  is  the  percentage  of  the  total  heat 
generated  by  the  combustion  of  the  fuel,  which  is  utilized  in 
heating  the  water  and  in  generating  steam.  With  anthracite 
coal  the  heating  value  of  the  combustible  portion  is  very  nearly 
14500  "  B.  T.  U."  per  pound,  equal  to  an  evaporation  from  and 
at  212°  F.  of  14500  -r-  966  =^  fifteen  pounds  of  water.  A  boiler 
which  when  treated  with  anthracite  coal  shows  an  evaporation 
of  twelve  pounds  of  water  per  pound  of  combustible  has  an  effi- 
ciency of  12  -^  15  =  80  per  cent.,  a  figure  which  is  approximate, 
but  scarcely  ever  quite  reached  in  the  best  practice. 

With  bituminous  coal  it  is  necessary  to  have  a  determination 
of  its  heating  power  made  by  a  coal  calorimeter  before  the  effi- 
ciency of  the  boiler  using  it  can  be  determined,  but  a  close  esti- 
mate may  be  made  from  the  chemical  analysis  of  the  coal. 

The  difference  between  the  efficiency  obtained  by  the  test  and 
100  per  cent,  is  the  sum  of  the  numerous  wastes  of  heat,  the 
chief  of  which  is  the  necessarj'  loss  due  to  the  temperature  of 
the  chimney  gases.  If  we  have  an  analysis  and  a  calorimetric 
determination  of  the  heating  power  of  the  coal,  and  an  average 
analysis  of  the  chimney  gases,  the  amounts  of  the  several  losses 
may  be  determined  with  approximate  accuracy  by  the  method 
described  below.  Data  given  : 
1-  Analysis  OF  THK  Coal.    Cum-       2.  Analvsisofthr  Drv  Chimnby 


Carbon 80.55  per  cent.  C.            O.  n. 

Hydrogen 4.50         "  CO,  13.6  3.71         9.89 

Oxygen    1.70        "  CO  0.2  0.09        o.:i  .... 

Nittogen I.08         "  O  11.3  11,10  .... 

UoiRlurc J.9*         '■  N  75.0  75.0 

Total  100,0       3.80       75.0 



The  gases  being  collected  over  water,  the  moisture  in  them  is 
not  determined. 

Heating  value  by  Dulong's  formula  =  14243  heat  units. 

3.  Ash  and  refuse  as  determined  by  boiler  test  10.25  per  cent, 
or  two  per  cent,  more  than  that  found  by  analysis,  thedifiference 
representing  carbon  in  the  ashes  obtained  in  the  boiler  test. 

4.  Temperature  of  external  atmosphere  60**  F. 

5.  Relative  humidity  of  air,  sixty  per  cent,  corresponding  to 
0.007  pound  of  vapor  in  each  pound  of  air. 

6.  Temperature  of  chimney  gases  =  560**  F. 

Calculated  results : 

The  carbon  in  the  chimney  gases  being  three  and  eight-tenths 
per  cent,  of  their  weight,  the  total  weight  of  dry  gases  per 
pound  of  carbon  burned  is  100  -7-  3.8  =  26.32  pounds.  Since 
the  carbon  burned  is  80.55  —  2.0  =  78.55  per  cent,  of  the 
weight  of  the  coal,  the  weight  of  the  dry  gases  per  pound  of  coal 
is  26.32  X  78.55-7-100  =  20.67  pounds.  Each  pound  of  coal 
furnishes  to  the  dry  chimney  gases  0.7825  pound  C,  0.0108  N, 

and  (  2.70  —     o~  j  "^  100  =  0.0214  pound  O;  a  total  of  0.8177 

or  0.82  pounds.  This  subtracted  from  20.67  pounds  leaves 
19.85  pounds  as  the  quantity  of  dry  air  (not  including  moisture) 
which  enters  the  furnace  per  pound  of  coal,  not  counting  the  air 
required  to  bum  the  available  hydrogen,  that  is,  the  hydrogen 
minus  one-eight  of  the  oxygen  chemically  combined  in  the  coal. 
Each  pound  of  coal  burned  contained  0.045  pound  of  hydro- 
gen, which  requires  0.045  X  S  =  0.36  pound  O  for  its  combus- 
tion. Of  this  0.027  pound  is  furnished  by  the  coal  itself,  leav- 
ing 0.333  pound  to  come  from  the  air.  The  quantity  of  air 
needed  to  supply  this  oxygen  (air  containing  twenty-three  per 
cent,  by  weight  of  O)  is  0.333  -^- 0.23  =  1.45  pounds,  which 
added  to  the  19.85  pounds  already  found  gives  21.30  pounds  as 
the  quantity  of  dry  air  supplied  to  the  furnace  per  pound  of  coal 
burned.  The  air  carried  in  as  vapor,  0.0071  pound  for  each 
pound  of  dry  air,  or  21.3  X  0.0071  =  0.15  pound  for  each  pound 
of  coal.  Each  pound  of  coal  contained  0.029  pound  of  moisture, 
which  was  evaporated  and  carried  into  the  chimney  gases.  The 




0.045  pound  of  hydrogen  per  pound  of  coal  when  burned  formed 
0.045  X  9  =  0.405  pound  of  water. 

From  the  analysis  of  the  chimney  gas  it  appears  that  0.09  -4- 
3.80  =  3. ,17  per  cent,  of  the  carbon  of  the  coal  was  burned  to 
carbon  monoxide  instead  of  carbon  dioxide. 

We  now  have  the  data  for  calculating  the  various  losses  of 
heat,  as  follows,  for  each  pound  of  coal  burned  ; 

ai.^  pounds  dry  air  X  (560^  —  60'^  F.)  X  sp.  beat  0,138  2534.7  17.80 

O.IS  pound  vapor  in  air  X  (560-  —  60')  X  ap.  heat  0.48  36,0  0.15 

0.039  poond  moisture  in  coal  heated  from  60*-  to  1 1  a'-'  F.  4.4  0.03 

"         "       evaporated  from  aud  at  311^  ;  0.019  X  966  iS.o  0.20 

steam  (heated  from  tli'  F,  to  560'-)  X  348 

X  0.48 4-7  003 

o.405poiindsw8ter  from  H  in  coal  x  (560— 60- )  x  0.48  97.*  0,68 
0.0337  pound  carboD  burned  to  carbon  monoxide,  toss 
by   incomplete   combustion,  0.0137  ^  (MS44  ^ 

4450 ^39.3  1.68 

0.03  pound  cobI  lost  in  ashes  ;  0.03  X  14S44 ^9°'9  3. 04 

Radiation  and  unaccounted  for  by  difference 712.1  5.00 

3947.8  37.71 

Utilized   in   making   steam,    equivalent    evaporation 

10.66  pounds  from  and  at  III-  per  pound  of  coal   10395.7  72-I9 

U143.0  100.00 

The  heat  lost  by  radiation  from  the  boiler  and  furnace  is  not 
easily  determined  directly,  especially  if  the  boiler  is  enclosed  in 
brick  work,  or  is  protected  by  non-conducting  covering.  It  is 
customary  to  estimate  the  heat  lost  by  radiation  by  difference, 
that  is.  to  charge  radiation  with  all  the  heat  lost  which  is  not 
othenvise  accounted  for. 

One  method  of  determining  the  loss  by  radiation  is  to  block 
off  a  portion  of  the  grate  surface  and  build  a  small  fire  in  the 
remainder,  and  drive  this  fire  with  just  enough  draught  to  keep 
up  the  steam  pressure  and  supply  the  heat  lost  by  radiation 
without  allowing  any  steam  to  be  discharged,  weighing  the  coal 
consumed  for  this  purpose  during  a  test  of  several  hours  dura- 


Estimates  of  radiation  by  difference  are  apt  to  be  greatly  in 
error,  as  in  this  difference  are  accumulated  all  the  errors  of  the 
analyses  of  the  coal  and  of  the  gases.  An  average  value  of  the 
heat  lost  by  radiation  from  a  boiler  set  in  brick  work  is  about 
four  per  cent. ;  when  several  boilers  are  in  a  battery  and  enclosed 
in  a  boiler  house  the  loss  by  radiation  may  be  very  much  less, 
since  much  of  the  heat  radiated  from  the  boiler  is  returned  to  it 
in  the  air  supplied  to  the  furnace,  which  is  taken  from  the  boiler 

An  important  source  of  error  in  making  a  **heat  balance," 
such  as  the  one  given  above,  especially  when  highly  bituminous 
coal  is  used,  may  be  due  to  the  non-combustion  of  part  of  the 
hydrocarbon  gases  distilled  from  the  cold  immediately  after  fir- 
ing, when  the  temperature  of  the  furnace  may  be  reduced  below 
the  point  of  ignition  of  the  gases.  Each  pound  of  hydrogen 
which  escapes  burning  is  equivalent  to  a  loss  of  heat  in  the  fur- 
nace of  62500  B.  T.  U. 

The  Determination  of  Sulphur  in  Steel  and  Cast  Iron. 

Of  the  various  methods  described  for  this  purpose,  the  follow- 
ing three  are  selected  as  giving  the  best  results  in  general 
practice : 

(a).  Bromine  method. 

{d).  Aqua  Regia  method,  si 

{c).  Potassium  permanganate  method. 

a.  Bromine  Method. 

Dilute  hydrochloric  acid  is  allowed  to  act  upon  the  steel  or 
iron  ;  the  sulphur  is  expelled  as  hydrogen  sulphide  and  is 
oxidized  by  the  bromine  to  sulphuric  acid. 

This  latter  is  precipitated  by  barium  chloride  as  barium  sul- 
phate, filtered,  washed  and  weighed  as  such,  then  calculated  to 

The  apparatus  used  is  as  follows  : 

In  the  fl^isk  A,  capacity  about  400  cc,  is  placed  the  steel  or 
iron  (five  grams  of  steel  or  three  grams  of  cast  iron),  and  con- 
nection made  with  the  absorption  apparatus  D}     In  the  latter, 

I  All  Stoppers  are  of  grlai^s- 


E",  is  placed  five  drops  of  bromine  and  twenty-five  cc.  of  hy- 
drochloric acid  (sp,  gr.  1.18).  Seventy-five  cc.  of  hydrochloric 
acid  (sp.  gr.  1.12),  is  placed  in  tlie  delivery  funnel  B  and  about 
ten  cc.  allowed  to  run  into  the  flask.  The  action  is  quite  often 
violent,  and  care  must  be  exercised  that  small  amounts  of  acid 
only  be  admitted  at  a  time  from  B  until  all  action  of  the  acid 
upon  the  steel  or  iron  ceases. 

I  Heat  is  now  gently  applied,  and  contents  of  the  flask  brought 
to  boiling  :  continue  the  boiling  two  or  three  minutes  ;  remove 

the  beat,  connect  the  delivery  tube  B  with  a  ' '  Bennert  drying 
apparatus."  and  connect  the  absorbent  apparatus  with  an  aspi- 
rator. Gradually  aspirate  about  one  liter  of  the  air  through  the 

Between   the   aspirator   and   the   absorbent  apparatus   there 

should  be  placed  a  wash  bottle  containing  dilute  ammonium  by- 

'  droxide,  (250  cc.  strong  ammonia  to  600  cc.  water),  to  absorb 

^  fumes  of  bromine  that  may  pass  out  of  the  absorbent  appa- 


ratus  during  aspiration.  Transfer  the  liquid  in  the  absorbent 
tube  to  a  No.  3  beaker,  washing  the  tubes  with  water  and  add 
the  washings  to  solution  in  beaker. 

Bring  to  boil,  expel  any  excess  of  bromine,  add  solution  of 

barium  chloride  and  set  aside  twelve  hours.      Filter  upon  two 

No.  3  ashless  filters,   wash  with  hot  water,  dr>-,   ignite,   and 

weigh  as  barium  sulphate  and  calculate  to  sulphur. 

b.  Aqua  Regia  Method. 

Five  grams  of  the  iron  or  steel  (in  fine  turnings)  are  trans- 
ferred to  a  No.  4  beaker  and  the  latter  covered  with  a  watch- 
glass.  Introduce  into  the  beaker  (in  quantities  not  exceeding 
ten  cc.  each  time)  some  nitric  acid,  until  the  iron  or  steel  is  dis- 
solved. Warm  gently  and  evaporate  to  dryness  on  an  iron  plate, 
adding  some  sodium  carbonate  previously,  so  that  no  sulphuric 
acid  may  be  lost  by  vaporization. 

Allow  to  cool_,  treat  with  hydrochloric  acid,  warm  until  solu- 
tion of  iron  is  complete  and  filter  off  the  silica.  Wash  well  and 
to  the  filtrate  containing  the  washings  add  a  few  cc.  of  solution 
of  barium  chloride,  and  set  aside  twelve  hours,  Filter,  wash 
with  hot  dilute  hydrochloric  acid,  then  with  water  thoroughly  ; 
dry,  ignite,  weigh  as  barium  sulphate  and  calculate  to  sulphur. 

This  method  is  preferred  where  the  iron  or  steel  contains  any 
metals  (even  in  minute  amounts)  that  are  precipitated  by  hydro- 
gen sulphide.  Thus  the  presence  of  one-fourth  percent,  of  cop- 
per would  render  the  bromine  or  permanganate  process  unrelia- 
ble since  hydrogen  sulphide  is  generated,  forming  copper  sul- 
phide, and  the  resulting  amount  of  sulphur  would  be  too  low. 

In  the  "  aqua  regia  method  "  the  oxidation  is  performed  at 
once  upon  the  addition  of  the  nitric  acid,  no  hydrogen  sulphide 
being  formed. 

c.    The  Potassium  Permanganate  Method.' 

a  is  a  flask  holding  300  cc. ,  with  pure  rubber  stopper,  through 
the  latter  passing  a  thistle  tube  with  stop-cock  for  the  delivery 
of  the  acid,  as  required.      Fig.  43. 

i  is  a  flask  with  rubber  stopper.  The  glass  tubing  must  not. 
reach  below  the  neck  of  the  flask.     This  flask  should  be  large 




enough  to  hold  the  contents  of  the  bottles  in  case  back-suction 
should  occur. 

The  bottles  f.  d,  and  e  contain  a  solution  of  potassium  per- 

manganate, five  grams  potassium  permanganate  to  1000  cc. 
water,  and  are  filled  to  the  amount  shown  in  the  figure  (about 
twenty-five  cc.  each),  ' 

/  contains  an  ammoniacal  soUition  of  silver,  and  is  used  to 


test  whether  the  hydrogen  sulphide  is  all  oxidized  by  the  perman- 
ganate ;  if  not  oxidized,  the  solution  in  /  becomes  black  from 
the  silver  sulphide  formed. 

The  process  is  as  follows  : 

Three  grams  of  cast  iron  or  six  grams  of  steel  are  placed  in 
flask  a  and  hydrochloric  acid  (sp.  gr.  1.12)  gradually  added 
until  seventy-five  cc.  have  been  used.  Warm  contents  of  the 
flask,  observing  that  the  evolution  of  gas  is  not  too  rapid. 

When  the  iron  or  steel  is  dissolved,  bring  the  liquid  to  boil- 
ing, connect  bottle  /  with  an  aspirator  and  slowly  draw  air 
through  the  apparatus  ten  minutes. 

Transfer  contents  of  bottles  c,  d,  e^  to  a  No.  3  beaker,  dissolv- 
ing any  oxide  of  manganese  that  may  have  deposited  in  bottles 
with  hydrochloric  acid.  Wash  the  bottles  with  hydrochloric 
acid,  then  with  water,  adding  the  washings  to  contents  of  the 

The  solution  in  the  beaker  is  warmed  and  enough  hydro- 
chloric acid  added  that  it  becomes  colorless  or  nearly  so,  and 
barium  chloride  added  in  sufficient  quantity  to  precipitate  the 
sulphuric  acid.  Allow  to  settle  twelve  hours.  Filter  upon  two 
No.  2  ashless  filters,  wash  with  boiling  water,  dry,  ignite, 
weigh  as  barium  sulphate  and  calculate  to  sulphur. 

Great  care  must  be  exercised  in  this  process,  that  the  potas- 
sium permanganate  is  free  from  sulphurous  or  sulphuric  acid 
before  use. 

The  iodine  method  for  the  determination  of  sulphur  in  pig 
iron  and  steel,  as  used  by  the  chemists  of  the  Duquesne  Steel 
Works,  is  as  follows  : 

Five  grams  pig  iron  or  steel  are  weighed  off  into  a  dry  500 
cc.  flask,  provided  with  a  double  perforated  rubber  stopper, 
with  a  long  stem  four  ounce  funnel  tube  with  a  stop-cock,  and 
a  delivery  tube  bent  at  right  angles,  on  which  a  short  piece  of 
one-quarter  inch  rubl^er  tubing  is  placed,  making  connection 
with  a  delivery  tul>e,  also  l>ent  at  right  angles  reaching  to  the 
Ixntoni  of  a  one  inch  by  ten  inch  test  tube,  suitably  supported. 
AK^ut  ton  CO.  of  the  animoniacal  solution  of  cadmium  chloride 
is  inlrvxhioeil  into  the  test  tul>e,  which  is  diluted  with  cold  water, 



until  the  tube  is  about  two-thirds  full.  Eighty  cc.  of  dilute  hy- 
drochloric acid — one  acid  to  two  water — is  poured  into  the  fun- 
nel tube,  a  file  marked  on  the  bulb  indicating  this  amount, 
which  is  allowed  to  run  into  the  flask,  the  stop-cock  is  then 
closed,  and  a  gentle  heat  applied,  till  the  drillings  are  all  in 
solution,  and  finally  to  boiling  by  raising  the  hent,  until  noth- 
ing but  the  steam  escapes  from  the  deliver>-  tube. 

The  apparatus  is  then  disconnected,  and  the  delivery  tube  is 
placed  in  a  No.  4  beaker  iu  which  the  titrations  are  made,  the 
contents  of  the  test  tube  are  then  poured  into  the  beaker,  the 
test  tube  filled  to  the  top  twice  with  cold  water,  the  sides  of  the 
I  tube  rinsed  down  with  about  twenty-five  cc.  dilute  hydrochloric 
I  acid  and  filled  again  with  cold  water.  The  total  volume  of  the 
solution  equaling  about  400  cc,  both  acid  and  water  being  sup- 
plied from  overhead  aspirator  bottles  and  suitable  rubber  con- 
nections with  pinch  cocks  ;  the  delivery  tube  is  now  rinsed  off 
inside  and  out  with  dilute  hydrochloric  acid,  and  about  five  cc, 
starch  solution  added  to  the  beaken 

Without  waiting  for  complete  solution  of  the  cadmium  sul- 
phide, the  iodine  solution  is  run  in  from  a  burette,  stirring  gen- 
tly, till  a  blue  color  is  obtained,  the  solution  is  then  stirred  vig- 
orously, keeping  a  blue  color  by  fresh  additions  of  the  iodine 
solution,  till  the  precipitate  of  cadmium  sulphide  is  all  dissolved, 
and  the  proper  permanent  blue  color  is  obtained.  The  amount 
of  iodine  solution  used  in  cc,  is  hundredths  per  cent,  sulphur. 

Iodine  iolutioH  is  made  by  weighing  off  into  adry  500CC.  flask 

about  thirty-five  grams  potassium  iodide,  and  sixteen  grams 

iodine,  fifty  cc.  water  added  and  shaken  and  diluted  cautiously 

L  until  all  are  in  solution,  and  finally  diluted  to  3500  cc.     This  is 

I  Standardized  with  steels  of  known  sulphur  contents,  so  that  one 

cc.  equals  0.0005  grams  sulphur. 

Cadmium  ehloride  solution  is  made  by  dissolving  100  grams 
cadmium  chloride  in  one  liter  water,  adding  500  cc.  strong  am- 
monia, and  filtering  into  an  eight  liter  bottle;  two  liters  of  water 
are  now  added,  and  the  bottle  filled  to  the  eight  liter  mark  with 
strong  ammonia. 

Starch  solution  is  made  by  adding  to  one-half  gallon  boiling 
L  water,  in  a  gallon  flask,  about  twenty-five  grams  pure  wheat 



Starch,  pre\'iously  stirred  up  into  2  thin  paste  with  cold  water; 
this  is  boiled  ten  minutes  and  about  twenty-five  grams  pure 
granulated  zinc  chloride  dissolved  in  water  added,  and  the  solu- 
tion diluted  with  cold  water  to  the  gallon  mark.  The  solution 
is  mixed  and  set  aside  over  night  to  settle,  the  clear  solution  is 
decanted  into  a  glass  stoppered  bottle  for  use. 
This  solution  will  keep  indefinitely. 

Rejerences  :  "  Volumetric  Method  of  Elliott,"  Chem.  News,  aj,  61. 

'■  Wiborgh's  Colorimetric  Method,"/.  Anal.  Chem.,  6,  301. 

"  Sulphur  Determinations  in  Iron  and  Steel,"  by  different  methi 
By  L.  S.  Clymer.  /.  Anal.  Chem.,  4,  318. 

"Cadmium  Chloride  as  an  Absorbent  of  Hydrogen  Sulphide"(Sulpboi 
in  Irou  and  Steel).     By  Frank  L.  Crobaugb./.  Anal.  Chem.,  7.  280. 

"The  Reduction  of  Barium  Sulphate  to  Sulphide  on  Ignition  with  Fil 
ter  Paper."     By  C.  W.  Marsh,/.  Atiai.  Chem.,  3,  2. 

"  Duplicate  DetermiQations  of  Sulphur  in  Iron  and  Steel  Should  Agrca 
within  0.005  Per  Cent."     By  C.  B.  Dudley,/.  Am.  Chem.  Soc. 

"  The  Determination  of  Snlphur  ia  Iron  and  Steel.  By  L,  Archbutt,i 
F.I.C.,/.  Soc.  Chem.  Industry,  4.  75, 

Determination  of  Silicon  in  Iron  and  Steel 

Five  grams  of  steel  or  three  grams  of  pig  iron  in  fine  borings, 
are  transferred  to  a  No,  3  beaker'  and  fifty  cc.  of  dilute  sulphuric 
acid  added.  When  the  action  of  the  acid  ceases  and  the  iron  is 
dissolved,  twenty-five  cc.  nitric  acid  (sp.  gr.  is  cautiously 
added  until  effer\'escence  ceases. 

Apply  heat  and  evaporate  until  white  fumes  of  sulphur  triox- 
ide  appear  ;  allow  to  cool ;  add  strong  hydrochloric  acid  until 
the  residue  is  thoroughly  saturated  with  it,  then  add  seventy- 
cc.  boiling  water.  * 

Filter,   wash  with   dilute   hydrochloric   acid,  then  with 
water,  dry,  ignite,  weigh  as   silicon  dioxide  and  calculate 

This  method  must  be  used  in  the  determination  of  silicon 
pig  iron,  but  in  wrought  iron  and  steel  the  insoluble  residue 
the  determination  of  phosphorus  may  be  used  for  the  silicon, 

1  Porcelain  bokeriare  W  be  ptcferred  to  giMS  beaker*  for  IhlideteTDUsatiaa. 



CARBON    IN    IRON    AND   STEEI,.  I57 

In  all  determinations  of  this  element,  the  ignited  and  weighed 
-ilicon  dioxide  must  be  nhite  in  color  and  a  fine  non-coherent 

Rtferencei  :  "Iirefrular  Dislribulioci  of  Silicon  iu  Pig  Iron.'  By 
J.    W.   Thomas. /.  ^nn/.  (~A>rfn.,  J,  t4S. 

■■  Silicon  in  Pig  Iron."     By  Clemens  Jones,/.  Anal.  Chetn.,  3,  m. 

--The  Influence  of  Silicon  on  the  Determination  of  Phosphorus  in 
Iron-'"     By  Thomas  M.  Drown,/.  Anal.  Chem..  3.  288. 

"  Notes  on  Silicon  in  Foundry  Pig  Iron.  By  David  H.  ^Tova.J.Anal 
Chem.,  6,  459-467. 


The  Determination  of  Carbon  in  Iron  and  Steel. 

The  determination  of  carbon  in  iron  and  steel  has  probably 
received  more  attention  in  later  years  from  chemists  than  any 
other  subject  in  analytical  chemistry. 

To  secure  a  method  at  once  complete  and  rapid  whereby  car- 
bon varying  in  amounts  from  four  per  ceut  to  o.ooi  per  cent,  in 
di&erent  irons  and  steels  could  be  accurately  determined  has 
been  a  desideratum. 

Processes  that  are  satisfactory  for  special  grades  of  irons  or 
steels  rarely  can  be  relied  upon  in  general  practice. 

So  important  has  this  subject  become  to  the  metallurgical 
world  that  committees  acting  in  union  from  Sweden,  England, 
and  America'  have  been  appointed  to  determine  not  only  the 
best  methods  of  iron  and  steel  analysis,  but  also  to  analyze 
standard  .samples  of  iron  and  steel,  compare  the  results,  and 
select  methods  which  should  be  uniform   lor  the  different  coun- 


The  determination  of  carbon,  , 
pies,  are  thus  reported  : 

5  made  upon  the  standard  sam- 

The  English  and  Swedish  committees  have  not  yet  selected 

the  method  to  be  adopted  as  standard  in  carbon  determination^ 
but  the  American  committee  have  rendered  their  report  suggest- 
ing certain  modifications  in  the  use  of  solvents  for  the  iron  and 
the  separation  of  the  total  carbon. 

The  use  of  the  double  chloride  of  copper  and  potassium,  as  a 
solvent  for  iron,  is  recommended  in  place  of  the  double  salt  of 
chloride  of  copper  and  ammonium,  owing  to  the  great  diflSculty 
in  obtaining  the  latter  salt  free  from  pyridin  and  other  tarry 

Of  the  many  methods  used  to  .obtain  the  amount  of  carbon 
from  the  iron,  the  following  few  are  selected  to  indicate  not  only 
the  variety  of  the  processes,  but  a  gradual  improvement  by  com- 
bination of  different  methods  : 

Berzelius'  first  suggested  that  the  iron  or  steel  be  finely  pi 
verized  and  then  ignited  in  a  current  of  oxygen  and  the  resi 
ing  carbon  dioxide  weighed. 

Regnault'  made  use  of  combustion  of  the  powdered  iron  wU 
chromate  of  lead  and  chlorate  of  potash  and  the  amount  of 
bon  dioxide  weighed. 

Berzelius,  however,  in  1840,  separated  the  carbon  from  the 
iron  by  dissolving  the  latter  in  copper  chloride  and  igniting  the 
carbon  in  oxygen.^ 

From  this  period,  the  methods  for  total  carbon  can  be  included 
in  two  general  classes  : 

First  class. — Combustion  of  the  carbon  in  the  powdered 

Secotid  class. — Separation  of  the  carbon  from  the  iron  by  chei 
ical  means  and  combustion  of  the  carbon. 

Deville  and  Wohler'  describe  processes  by  which  the  iron  can 
be  separated  from  the  carbon,  by  volatilization  of  the  iron  with 
chlorine  or  hydrochloric  acid  gas,  and  combustion  of  the  remain- 
ing carbon. 

With  the  exception  of  a  method  described  by  Gmelin' 
which  the  powdered  iron  is  treated  directly  with  chromium 

1  Ann.fkyi,  chtm..  iSjg, 

*Anm.cllim.p*3ti.,i9yi.  107. 

*J.  frail.  Oiim.,  tifo.  J47. 

•  ZUcAr.  amil.  Ctetn..  I,  tol. 

t  OeilrrUArr  ZeiUchri/l/ir  Brrg  <.nd  ffiltfnmit*.  iSSj,  39J. 



oxide  and  sulphuric  acid  and  the  carbon  oxidized  to  carbon 
dioxide,  the  methods  of  the  first  class,  above  given,  are  no 
longer  used. 

Second  Class. — These  methods  give  better  results  in  general 
practice,  and  nearly  all  the  advances  and  improvements  have 
been  made  in  this  direction. 

^  Ullgren'  dissolved  the  iron  with  solution  of  copper  sulphate 
and  oxidized  the  carbon  to  carbon  dioxide  by  heating  with 
chromium  trioxide  and  sulphuric  acid. 

Eggertz'  dissolved  the  iron  with  bromine  or  iodine,  and  the 
separated  carbon  was  ignited  with  chroniate  of  potash. 
^    Langley'  modified  Ullgren's  method  by  ignition  of  the  carbon 
in  oxygen  after  solution  of  the  iron  by  copper  sulphate. 

Richter  dissolved  the  iron  with  chloride  of  copper  and  potas- 
sitim  and  burned  the  carbon  in  oxygen. 

Weyl  and  Binks  dissolved  the  iron  in  dilute  hydrochloric 
acid  passing  an  electric  current  at  the  same  time  and  ignited  the 
carbon  in  oxygen.* 

Parry  dissolved  the  iron  in  solution  of  copper  sulphate,  the 
carbon  burned,  mixed  with  copper  oxide,  in  vacuo,  and  the 
\T3lume  of  carbon  dioxide  measured.' 

Eggertz's  method  for  combined  carbon',  in  which  the  iron 
was  dissolved  in   nitric  acid  and  the  amount  of  carbon   (com- 
Llrined)  determined  by  color  of  the  solution  formed. 

McCreath  and  Pearae  dissolved  the  iron  with  chloride  of  cop- 
per and  ammonium  and  ignited  the  carbon  in  oxygen.' 

Boussingault'  decomposed  with  mercuric  chloride  and  oxi- 
dized the  carbon  to  carbon  dioxide. 

Wiborgh*  dissolved  the  iron  with  solution  of  copper  sulphate, 
oxidized  the  carbon  by  heating  with  chromium  trioxide  and 
sulphuric  acid,  and  measured  the  volume  of  carbon  dioxide 

XAmm.ibatln.aPkaT..  II4.  » 

•  Dtmtlrr't  PolylrtJuntluiJimntal,  170.  j!o. 

•  AmrrUAn  C^ifmtst.  ft.  165. 
*Am*.^ji.  Otm.,  114,  y>j. 

•  Oem.  JVtw,.  15. 301 

^  Ettfiiutfitig  and  Wiling  JiMtmol,  t\,  151. 

•  OiKgler-i  /Ww«A./.  I9T.  W. 

•  DtmgliT-i  Polylfdt.  / .  iCj.  ya. 



Kxperience  has  shown  that  the  methods  of  Ullgren,  Lang- 
ley,  Richter.  and  Wiborgh  give  the  best  results  for  the  total 
amount  of  carbon  in  iron,  and  that  the  Eggertz  method  for  com- 
bined carbon  in  steel  can  be  relied  upon  as  the  best  for  the  pur- 

The  determination  of  total  carbon,  as  made  in  ray  labbratorj-, 
is  either  by  the  Ullgren  or  Langley  methods,  somewhat  modi- 
fied. The  Ullgren  method  is  thus  performed  :  Six  grams  of  the 
iron,  in  fine  turnings,  are  transferred  to  a  No.  3  beaker  and  100 
cc.  of  a  solution  of  copper  sulphate'  (i  to  5)  added,  the  solution 
being  first  rendered  neutral  by  a  few  drops  of  a  very  dilute  solu- 
tion of  potassium  hydroxide.  Digest  at  a  gentle  heat  until  all 
the  iron  is  dissolved  (no  smell  of  hydrocarbon  given  off),  add 
100  cc.  cuprous  chloride  solution  (i  to  3)  and  seventy-five  cc. 
hydrochloric  acid  (specific  gravity  1.2)  and  warm  until  the 
metallic  copper  is  dissolved.  Filter  upon  an  asbestos  filter, 
washing  first  with  dilute  hydrochloric  acid,  and  finally  with 
water  until  uo  reaction  for  hydrochloric  acid  is  obtainable  with 
a  drop  of  silver  nitrate  solution.  Transfer  the  asbestos  filter 
containing  the  carbon  to  the  flask  .-J,  Fig.  44,  using  not  over 
twenty-five  cc.  water  in  the  operation.  Add  ten  grams  chromium 
trioxide,  and  in  the  deliverj"  flask  place  fifty  cc.  concentrated 
sulphuric  acid,  and  connect  the  flask  with  the  system  of  [J-tubes. 

D  contains  water  sufficient  to  cover  the  neck  of  the  (J-tube, 
and  is  made  slightly  acid  with  sulphuric  acid. 

C  and  D  contain  granulated  calcium  chloride  free  from  litne. 

E  and  /"contain  soda  lime,  medium  granulated,  and  are  care- 
fully weighed  before  use. 

G  contains  granulated  calcium  chloride.  Allow  the  sul- 
phuric acid  in  the  delivery  tube  to  enter  flask  A  and  close  the 
stop-cock,  Warm  the  contents  of  the  flask  gradually  to  boiling, 
and  when  no  more  gas  passes  through  B  open  the  side  stop- 
cock of  flask  A  and  connect  ivith  the  Trauber  drying  apparatus. 
The  aspirator  is  connected  with  G  and  the  air  is  slowlj-  aspir- 
ated through  the  entire  apparatus.  Continue  this  until  about 
five  liters  of  air  have  been  aspirated. 

■  Copper  chloride  slid  hydrochloric  aciil  cmi  I)*  lubsUlaled  ■>  reconimeadEd  br 



After  twenty  minutes  weigh  tubes  E  and  F\  the  increase  of 
weight  represents  the  carbon  dioxide  produced  by  the  oxidation 
of  the  carbon. 

Thus,  six  grains  of  cast  iron  taken : 

Tubes  -ff  and  F-f  CO, 65.700 

Tubes  E  and  F* 65.002 

CO, 0.698 

CO, :  C  : :  0.698  :  x  =  0.1904. 

0.1904X100^  3, 8  per  cent,  carbon. 

The  carbon  in  cast  iron  being  generally  a  mixture  of  com- 
bined and  graphitic  carbon,  it  is  essential  to  determine  the 
graphitic  carbon,  and  this  amount  being  subtracted  from  the 
total  carbon  gives  the  combined  carbon.  In  steels  where  the 
carbon  is  all  combined  the  color  test  of  Eggertz  suffices.  The 
graphite  is  thus  determined  : 

Add  fifty  cc.  hydrochloric  acid  (specific  gravity  i .  i )  to  six  grams 
of  cast  iron  or  ten  grams  of  steel  in  a  Ko.  3  beaker ;  warm 


gently  until  the  iron  is  all  dissolved,  bring  to  boiling  tempera- 
ture for  five  minutes,  allow  the  graphite  to  settle,  and  decant 
the  supernatant  liquid  upon  an  asbestos  filter ;  wash  by  decan- 
tation  four  times  with  hot  water  and  treat  residue  in  beaker 
with  twenty-five  cc.  solution  of  potassium  hydroxide(sp.  gr.  i .  1 2) 
and  boil.  Transfer  to  the  asbestos  filter,  wash  thoroughly  with 
boiling  water,  then  with  alcohol  and  ether,  and  transfer  the  as- 
bestos filter  to  the  flask  A,  Fig.  44,  and  oxidize  the  carbon  to 
carbon  dioxide  with  chromium  trioxide  and  sulphuric  acid,  as 
in  the  process  previously  given  for  total  carbon. 
Thus,  six  grams  of  iron  taken  : 

Tul>cs  E  and  /^-|-  CO, =  66.053  grams. 

Tul>es  E  and  F =  65.621      •* 

CO,  =    0.432      " 

C  ^0.1178  gram. 

Graphitic  carbon =  1.96  per  cent. 

Combined  carbon =«  1.22        ** 

Total  carbon 3.18        " 

^^^^^V  CARBON    IN    IRON   AND    STEEI..  163 

^^^^HP  Method  of  Langley  Modified. 

In  this  process  the  sample  is  treated  in  the  same  manner  for 
^solution  of  the  iron  as  described  for  total  carbon  in  the  UUgren 
:«nethod.  After  the  carbon  has  been  thoroughly  washed  upon 
Hie  asbestos  filter,  it  is  dried  and  transferred  to  a  porcelain  boat 
^lich  is  placed  inside  of  a  combustion  tube  in  the  furnace  C, 

1^%.  45- 
The  tube  D  connected  with  the  combustion  tube  contains 
granulated  calcium  chloride,  and  E  and  F  soda  lime  ;  another 
tube  G  containiug  calcium  chloride  (not  shown  in  the  figure), 
is  also  used.  Oxygen  under  pressure  in  the  tank  A  is  allowed 
to  pass  slowly  through  the  Trauber  drying  apparatus,  which 
nemoves  all  moisture  and  carbon  dioxide,  into  the  combustion 
tube  and  through  the  tubes  D.  E.  /^and  G. 

Heat  is  gradually  turned  on  in  the  furnace  and  increased  until 
tlie  carbon  is  completely  burned  to  carbon  dioxide.  Turn  off  the 
Ixeat,  disconnect  the  oxygen  tank  and  slowly  aspirate  air  through 
Xlie  apparatus  by  means  of  an  aspirator. 

After  cooling  thirty  niinutes  wei^h  the  tubes  E  and  F  and 
^::^alculate  the  result  as  given  previously. 

It  will  be  noticed  that  no  Liebig's  potash  bulbs  are  used.     I 

^llave  obtained  better  results  by  the  use  of  soda  lime  in  y-tubes 
than  by  the  potash  bulbs,  and  in  general  practice  they  will  be 
iound  much  more  convenient  and  less  liable  lo  variation  in 

The  use  of  the  double  chloride  of  copper  and  ammonium  as  a . 
solvent  for  the  iron  has  been  quite  general  in  this  countr>'.  The 
American  Committee  on  Standard  Methods  of  Iron  Analyses 
found  that,  contrary  to  the  usual  practice,  this  solvsnt  must  not 
be  neutral,  but  strongly  acid  with  from  five  to  ten  per  cent,  of  its 
volume  of  strong  hydrochloric  acid. 

T.  M.  Drown,  in  his  report  to  the  committee,  describes  his 
process  as  follows  :  ' '  Three  grams  of  the  steel  were  treated  with 
200  cc.  of  a  solution  of  copper  potassium  chloride  (30c  grams  to 
the  liter)  and  fifteen  cc.  of  hydrochloric  acid  (sp.  gr.  1.3). 
After  complete  solution  of  the  iron  the  carbon  was  filtered  off  on 
an  asbestos  lined  platinum  boat,   thoroughly  washed  with  hy- 


drochloric  acid,  and  then  with  water  until  the  wasliings  gave  no 
reaction  with  silver  nitrate.  After  dr\ing  the  boat  was  put  into 
a  porcelain  tube  and  the  carbon  burned  in  a  current  of  oxygen." 
This  is  a  modification  of  Richter's  process. 

There  does  not  appear  to  be  much  choice  in  the  method  of  the 
combustion  of  carbon.  Some  chemists  prefer  oxidation  with 
chromium  trioxide  and  sulphuric  acid,  and  others  ignition  in  a 
current  of  oxygen  gas. 

Fig.  46. 

For  rapidity  of  execution  and  simplicity  of  apparatus  (Fig. 
44),  Ilpreferthe  former. 

Wiborg's  method,'  in  which  the  carbon  dioxide  is  measured 
instead  of  being  weighed,  consists  as  follows:  The  apparatus 
'required  is  shown  in  Fig.  46.  A  test  tube  A,  140  mm.  long  by 
twenty  mm.  internal  diameter,  is  surrounded  by  a  cage  of  brass 

dt.trt'JW.  /HdHilry,  6.  -iH. 



wire  gauge,  and  fitted  with  a  caoutchouc  cork  with  two  perfor- 
ations. Through  one  perforation  passes  the  narrow  end  of  the 
stop-cock  funnel  B,  which  should  project  for  about  fifteen  to 
twenty  mm.  beneath  the  cork  ;  through  the  other,  but  not  pro- 
jecting beneath  the  stopper,  passes  the  connecting  tube  D.  This 
latter  tube  consists  of  two  portions,  united  by  India  rubber  tub- 
ing ;  the  part  more  remote  from  A  and  carrj'ing  the  stop-cock  E 
is  bent  to  pass  through  one  of  the  perforations  of  another  caout- 
chouc stopper  in  the  graduated  tube  C,  the  other  perforation 
serving  to  connect  the  latter  with  a  stop-cock  funnel  F. 

The  tube  C  should  for  the  distance  of  seventy  mm.  downwards 
have  an  internal  diameter  of  sixteen  mm;  it  should  then  be 
widened  to  a  bulb  G,  of  about  twenty-five  centimeters  capacity, 
and  be  finally  reduced  for  the  remaining  200  mm.  to  about  nine 
mm.,  this  narrow  portion  being  graduated  into  divisions  of  one- 
tenth,  or  preferably,  one-twentieth  of  a  cubic  centimeter,  denot- 
ing in  each  case  the  capacity 'of  the  whole  of  that  portion  of  the 
tube  above  the  respective  graduations.  Beneath  this  tube  is 
the  stop-cock  H,  communicating  by  flexible  tubing  with  the 
movable  water  reser\'oir  /.  The  test  tube  A  is  warmed  by  a  gas 
or  spirit  lamp,  and  the  whole  apparatus  should  be  mounted 
suitable  stand.  The  measuring  tube  is  surrounded  by  a  water 
jacket  A' to  preserve  an  even  temperature. 

To  conduct  an  analysis  two-tenths  gram  of  finely  divided 
wrought  iron  or  steel  or  one-tenth  gram  of  cast  iron  is  intro- 
duced carefully  into  the  test  tube  A.  taking  care  that  none  of 
the  filings  adhere  to  its  sides.  Four  cc.  of  a  saturated  soluttoa 
of  pure  copper  sulphate  are  then  introduced  and  allowed  to  act, 
with  frequent  stirring,  during  ten  minutes,  unless  an  apprecia- 
ble smell  of  hydrocarbon  is  obser\'ed,  when  the  action  must  be 
suspended  after  three  or  four  minutes.  One  and  two-tenths 
grams  of  crystallized  chromic  acid  are  added  to  the  solution. 
Meanwhile  the  tube  C  must  have  been  filled  with  water  by  rais- 
ing the  reservoir  /  until  the  liquid  has  risen  above  the  bulb  tube 
G,  the  remaining  space  up  to  the  cock  being  filled  by  water  in- 
troduced through  F.  The  test  tube  is  now  corked  and  con- 
nected with  the  burette  C 

Eight   cc.   of  sulphuric  acid   (sp.   gr.    1.7)    are   introduced 



)  STEEL.  167 

'  through  B,  the  cock  of  the  latter  is 
*^«i,  that  marked  £"  opened,  and  the  liquid  in  tlie  test  tube 
gra*i\ia!ly  raised  to  boiling,  the  pressure  having  been  diminished 
^V  previously  lowering  the  water  reser\'oir  /.  After  ten  minutes' 
boiling,  during  which  the  reser%'oir  has  been  still  further  low- 
ered, if  necessary,  to  maintain  the  diminished  pressure,  the  tube  is 
cooled  somewhat,  and,  together  with  the  connecting  tube  D,  is 
carefully  filled  with  water  introduced  through  B.  The  cock  E 
is  then  closed  and  the  total  volume  of  air  and  carbon  dioxide 
read  off  after  leveling  with  the  reservoir. 

/  is  then  once  more  lowered  and  the  cock  H  closed  in  order  to 
draw  in  a  quantity  of  a  ten  per  cent,  potassium  hydroxide  solu- 
tion through  F.  After  the  carbon  dioxide  has  been  completely 
absorbed,  H  is  reopened,  the  liquid  leveled  again  and  a  reading 
of  the  residual  air  is  taken. 

The  difference  between  the  two  readings  will  be  the  volume  of 
carbon  dioxide  evolved  from  the  carbon  in  the  iron. 

E\-idently  if  two-tenths  gram  of  substance  were  used,  each 
cc.  of  carbon  dioxide  will  correspond  to  0.253  P^''  cent,  of  car- 
bon, and  the  factor  0.353  multiplied  by  the  number  of  cubic 
centimeters  of  gas  should  give  a  direct  reading  of  the  percentage 
of  carbon. 

But  this  is  not  quite  correct,  since  a  certain  quantity  of  car- 
bon dioxide  (to  be  found  by  experiment)  is  absorbed  by  the 
water  in  the  tube.     By  treating  pure  anhydrous  sodium  carbon- 

»Bte  in  the  apparatus  instead  of  iron  and  comparing  the  actual 
with  the  theoretical  yield  of  carbon  dioxide,  the  factor  may  be 
Thus  the  true  factor  was  found  to  be  0.28,  and  this  was  uni- 
versally correct  tor  cast  irons  :  but  for  wrought  irons  or  steels, 
■nrhich  contain  less  carbon,  it  should  be  0.29, 
When  one-tenth  gram  of  iron  is  used  the  factor  must  of  course 
be  doubled. 
Where  the  temperature  of  the  operation  differs  much  from  the 
normal  eighteen  degrees,  correction  must  be  made  by  multiply- 
ing or  dividing  by  (i  -\-  0.00367  X  /),  where  /is  the  variation  in 
temperature,  according  as  the  solution  is  cooler  or  warmer  than 
i  normal. 


This  process  is  expeditious,  and  a  verj-  delicate  nieasuremenl 
of  the  carbon  dioxide  cau  be  obtained,  thus  : 

One-tweutieth  cc.  of  carbon  dioxide  from  two-tenths  gram  ( 
iron  represents  0.014  P^r  cent,  of  carbon,  but  weighs  onlyo.oood 

G.    Lunge'    gives  this  process  the   preference   where  smal 
quantities  of  carbon  are  to  be  determined  in  cast  irons. 
Determination  of  Combined  Carbon  in  Steel.     Eggeriz'  Method. 

This  method  depends  upon  the  color  given  to  nitric  acid  (sp. 
gr,  1.2)  when  steel  is  dissolved  therein  :  the  carbon  present  pro- 
ducing a  light  brown  or  dark  brown  coloration  to  the  liquid  in 
proportion  as  the  carbon  is  in  small  or  large  amounts.  The  ap- 
paratus, Fig.  47,  is  well  arranged  for  this  test.     It  consists  of  , 

a  series  of  graduated  tubes,  of  glass,  each  27.5  centiraeters 
long,  fifteen  mm.  in  diameter,  and  graduated  to  hold  thirty 
cc.  divided  by  one-fifth  cc.  The  back  plate  of  the  appa- 
ratus is  of  white  porcelain,  25,5  centimeters  wide,  twenty-sevei 
centimeters  high,  and  three  mm.  thick,  and  I  have  found  \ 
much  better  than  the  various  cameras  to  obtain  correct  compai 


CARBON   IN   ISON    AND    STEKI,.  169 

sons  of  colors  of  solutions  in  the  different  tubes.  Three  stand- 
ard steels  are  required,  one  containing  one  per  cent,  combined 
carbon,  for  tool  steels,  etc.,  one  containing  four-tenths  per  cent. 
carbon,  for  tires,  rails,  etc.,  and  two-tenths  per  cent,  carbon,  for 
soft  steels ;  these  percentages  of  carbon  having  been  verj-  accu- 
rately determined  by  combustion. 

The  process  is  as  follows:  Two-tenth  gram  of  the  standard  steel 
is  transferred  to  one  of  the  graduated  tubes,  four  cc.  of  nitric 
acid  (sp.  gr.  1.20)  added,  and  the  tube  placed  in  cold  water  to 
prevent  energetic  action  of  the  acid.  After  a  few  minutes  inter- 
val the  tube  is  placed  in  warm  water,  and  the  latter  gradually 
raised  to  the  boiling-point  and  maintained  at  that  temperature 
about  twenty  minutes.  The  sample  of  steel,  in  which  the 
amount  of  carbon  is  unknown,  is  treated  in  a  similar  manner, 
using  the  same  amount  of  steel  and  acid. 

Suppose  the  standard  steel  contains  0.S4  per  cent,  of  carbon, 
the  solutiou  in  the  tube  is  diluted  with  water  to  16.8  cc.  Each 
cubic  centimeter  therefore  contains  0.0001  gram  of  carbon. 
Suppose  that  upon  dilution  of  the  test  sample  solution  to  four- 
teen cc.  and  placing  the  two  tubes  side  by  side  in  the  frame 
(Fig.  47),  that  the  test  sample  is  somewhat  stronger  in  color 
than  the  standard  sample  ;  upon  diluting  it,  however,  to  fifteen 
cc.  it  is  slightly  lighter  in  color.  This  would  indicate  that  the 
unknown  or  test  sample  contains  more  than  0.70  per  cent. 
(o.i  X  y )  of  carbon,  but  leas  than  0.75  per  cent,  (o.i  X  '/ ). 
The  steel  can  be  thus  assumed  to  contain  0.73  per  cent,  carbon. 

The  use  of  Eggertz'  color  test  for  combined  carbon  requires 
that  steels  should  have  been  subjected  to  a  similar  physical  treat- 
ment to  which  the  standard  steels  had  been  subjected  in 
order  to  secure  accurate  results.  A  steel  shows  less  carbon,  by 
color,  when  hardened  than  when  unhardened,  and  less  unan- 
neated  than  when  annealed.  Several  modifications  of  the  pro- 
cess have  been  submitted  by  various  chemists,  hut  they  offer  no 
special  advantages.  Stead'  renders  the  nitric  acid  solution  of 
the  steel  alkaline  with  sodium  hydroxide,  which  dissolves  the 
carbon,  producing  a  solution  about  two  and  a  half  times  stronger 

color  than  the  solution  in  nitric  acid.     The  precipitated  iron 



oxide  is  filtered  off.  and  a  measured  quantity  of  the  colored  fil-  J 
trate  Is  transferred  to  a  Stead's  chromometer  and  the  color  com* 
pared  with  a  standard  steel  under  similar  conditions.     Except  J 
where  the  carbon  is  present  in  minute  quantity  only  is  this  pro*  J 
cess  of  any  advantage  over  the  Eggertz  method. 

Carbon  Compounds  of  Iron. 

Microscopical  examinations  of  iron  have  led  to  remarkable 
developments  of  our  knowledge  of  its  structure.  Recent  inves- 
tigations by  Osmond,  Martens,  Arnold  and  others  have  shown 
that  eminently  practical  results  are  to  be  obtained  from  this 
microscopical  examination.  < 

These  microscopical  examinations  indicate  that  structure  of 
iron  depends  upon  a  number  of  partially  identified  compounds 
which  have  been  given  the  names  of  pearlite,  cementite,  mar- 
tensite,  etc..  and  which  are  all  compounds  of  carbon  and  iron. 

Marten's  and  Osmond's  latest  investigations,  as  well  as  those 
of  Arnold,  have  demonstrated  that  the  presence  or  absence  of 
one  or  more  of  these  compounds  determines  and  identifies  the 
qualities  and  properties  of  different  kinds  of  iron  and  also  de- 
termines the  methods  of  manufacture  and  heat  treatment 
which  they  were  subjected. 

Messrs.  Abel,  Mueller,  and  Leduber  showed  some  years  ago 
that  carbon  in  unhardened  steel  exists  chiefly  as  the  definite 
carbide,  Fe,C ;     but  microscopical  investigations    have    further 
proven  the  coexistence  of  many  other  carbides,  especially  aft« 
heat  treatment.     Professor  Arnold  claims  to  have  proven 
existence  of : 

{a)   Crystals  of  pure  iron  which  remain  bright  upon  etching. 

(A)  Crystals  of  slightly  impure  iron  which  become  pale  brown 
on  etching,  probably  owing  to  the  presence  of  a  small  quantity 
of  an  intermediate  carbide  of  hypothetical  formula  Fe,„C. 

{c)   Normal  carbide  of  iron.  Fe,C.  which  exists  in  three 
tinct  modifications  ;  each  one  conferring  upon  the  iron  in  whid 
it  is  found  particular  mechanical  properties: 

( I )  Emulsified  carbide  present  in  an  excessively  fine  state 
division  in  tempered  steels. 


^  dis->l 

(a)    Diffused  carbide  of  iron  occurring  in  normal  h 
forms  of  small  ill-defined  stride  and  granules. 

(5)  Crystallized  Fe^C  occurring  as  well  defined  lamiuse  in 
annealed  and  in  some  normal  irons, 

(d)  Subcarbide  of  iron,  a  compound  of  great  hardness  exist- 
ing in  hardened  and  tempered  irons  and  possessing  formula 
Fe„C.  This  substance  is  decomposed  by  the  most  dilute  acids. 
and  at  400°  C.  it  is  decomposed  into  Fe,C  and  free  iron  with 
evolution  of  heat.  One  of  the  most  remarkable  properties  of 
this  compound  is  its  capacity  for  permanent  magnetism, 

(f )   Graphite  or  temper  carbon. 

Chemists  have  heretofore  identified  only  graphitic  carbon. 
combined  carbon,  and  carbide  of  iron.  These  alone  are  not 
sufficient  to  identify  iron,  and  what  must  be  done  is  to  devise 
accurate  methods  (or  determining  all  of  the  above  enumerated 
substances  by  chemical  analysis. 

Professor  Arnold  says:  "  The  existence  of  Fe,.C  is  proved  by 
the  fact  that  iron  containing  o.8g  per  cent,  carbon  presents  sev- 
eral co-relative  critical  points  when  examined  by  different 
methods  of  observation  :  (i )  Well  marked  saturation  points  in 
raicro- structure  of  normal  annealed  and  hardened  steels.  (2) 
A  sharp  maximum  in  a  cur%-e,  the  coordinates  of  which  are 
heat  evolved  or  absorbed  at  Ar  i  (point  of  recalescence }  and 
carbon  percentage.  (3)  A  point  in  the  compression  curve  of 
hardened  steels  at  which  molecular  flow  absolutely  ceases. 
(4)  A  sharp  maximum  in  a  curve,  the  coordinates  of  which  are 
carbon  percentage  and  permanent  magnetism  in  hardened  steels. ' ' 

The  famous  French  micrographist,  F.  Osmond,  defines  and 
describes  five  distinct  carbon  compounds  which  can  only  be 
found  and  identified  by  the  microscope,  as  existing  in  iron  sub- 
jected to  heat  treatment. 

( 1 )  The  first  he  calls,  with  Howe,  ferrite,  because  it  is  almost 
pure  iron  ;  it  at  first  retains  a  dull  polish  (po!i  sp^culaire)  when 
relief  polished :  after  continued  polishing,  especially  with  precipi- 
tatedchalkandwater,  itbecomesmoregranularasitislessmassive, 
but  when  forming  large  masses  it  finally  shows  as  polyhedral 
crystals.  Etch  polishing  with  tincture  of  iodine  produces  no 


(2)  The  second  is  called  cementite,  which  is  distinguished  by 
its  hardness  (felspar,  No.  6,  Mohr's  scale).     This  hardness, 
which  is  greater  than  that  of  all  other  carbon  compounds,  per- 
mits its  identification  even  when  polishing  with  emery  paper,  pro- 
vided it  is  not  so  imbedded  in  softer  particles,  that  the  micro- 
scope is  no  longer  able  to  identify  it,  and  chemical  analysis  alone 
is  able  to  prove  its  presence.     This  substance  corresponds  to 
that  imagined  by  Karsten  and  Caron,  and  isolated  by  Dr.  F.  C. 
G.  Mueller,  Sir  Fred.  Abel,  and  Professor  Ledebur  as  carbide, 
and  of  the  probable  formula  Fe,C,  and  which  Howe  also  calls* 
cementite.     Osmond  believes  that  cementite  of  iron  of  cementa- 
tion (de  racier  poule)  can  now  be  identified  with  the  hard  com- 
ponent of  cast  and  forged  steels. 

(3)  The  third  compound  is  called  sorbite,  after  Dr.   Sorby^ 
which  was  first  described  as  *' pearly  constituent,  * '  ( Howe  called- 
itperlite),  which  could  be  identified  under  a  magnification  o'f 
800  diameters,  as  a  substance  with  the  sheen  of  mother-of-pearl-. 
It  is  possible,  with  oblique  light,  to  separate  this  in  bands  of 
alternating  hard  and  soft  flakes. 

Osmond  questions  the  accuracy  of  the  conclusion  generally 
held  that  this  is  Fe,C,  because  he  points  out  that  etch  polishing 
gradually  changes  the  color  from  yellow  to  brown  and  then  from 
purple  to  blue,  and  at  a  certain  period  there  is  great  difference  be- 
tween the  colors  in  adjacent  ** islets**  (ilots). 

The  uncolored  flakes  (lamelles)  may  appear  elevated  or  de- 
pressed. With  tincture  of  iodine  similar  results  are  obtained. 
It  must  be  remembered  that  neither  ferrite  nor  cementite  take 
such  colors  under  similar  conditions  even  when  extract  of 
liquorice  root  or  tincture  of  iodine  is  used. 

He  can  offer  no  suggestion  in  regard  to  the  chemical  compo- 
sition of  *'  sorbite." 

(4)  A  fourth  compound  always  found  after  quenching  iron, 
which  is  already  well  known,  is  *J  martensite,**  named  after  Pro- 
fessor A.  Martens,  the  famous  micrographer  of  Berlin. 

When  iron  with  0.45  per  cent,  carbon  is  heated  to  825®  C, 
and  then  at  720°  C.  quenched  in  a  freezing  mixture  of  20**  C,  re- 
lief polishing  produces  no  effect ;  but  etch  polishing  shows  the 
structure.    Groups  of  needles  (fascicles)  or  groups  of  rectilinear 


parallel  fibers,  which  are  separated  or  not  by  a  scarry  or  vertni- 
fom  filling,  and  are  shown  in  very  slight  depths.  Three  groups 
of  fibers,  parallel  to  the  sides  of  a  triangle,  are  often  seen  in  one 
spol,  as  crystalline  bodies  of  the  cubical  system.  Etch  polish- 
ing does  not  always  color  martensite,  and  then  only  takes  a  light 
vellow  sheen.  However,  when  applying  tincture  of  iodine  it 
takes  a  yellow,  brown,  or  black  color,  according  to  the  percent- 
sec  of  carbon  present.  Because  of  the  non-uniformity  of  color 
Jt  is  not  quite  certain  whether  martensite  can  be  considered  a 
lundaGienta!  compound.  It  does,  however,  retain  its  forms  even 
in  the  quenched  parts,  in  the  softest  as  well  as  in  the  hardest 
iron,  with  the  single  difference  that  the  fascicles  (needles)  are 
sometimes  longer,  sometimes  more  varied,  in  accordance  whether 
ihe  iron  is  more  or  less  carbonized.  The  shapes  are  character- 
"tic  and  permit  the  determination  of  differences  in  hardness. 
Martensite  is  not  positively  a  definite  compound  of  iron  and  car- 
'mii ;  it  represents  rather  the  crystalline  arrangement  of  an  allo- 
Irapic  modification  oi  iron  under  the  influence  of  carbon. 

15)  A  fifth  well  defined  iuiidiiinental  conipoutid  found  in  me- 
dium iron  quenched  while  undergoing  structural  changes  (into 
'Is  allotropic  modifications  according  to  Osmond),  is  named 
troQstite.  after  the  famous  French  metallurgist  Troost.  When 
iron  with  0.45  per  cent,  carbon  is  heated  to  825°  C.  and  then 
quenched  at  690''  C.  is  relief  polished,  nodules  in  relief,  de- 
pressed tatters  or  tongues  (lanibeaux),  and  between  the  two 
intercalations  of  varying  breadth  and  medium  hardness  are 
"ieveloped.  Etch  polishing  proves  that  the  hard  nodules  are 
martensite.  and  the  soft  tatters  or  tongues  are  ferrite.  The  in- 
tercalated bands  show  temper  colorations,  but  they  harden  less 
rapidly  than  sorbite  under  identical  conditions,  and  these  colors 
produce  an  irregular  marbleized  appearance  :  they  are  almost 
amorphous,  slightly  granular  and  warty.  Tincture  of  iodine 
'"first  and  second  application  produces  quite  similar  effects,  in 
this  filth  fundamental  compound,  troostite. 

"!•  is  noticed  that  it  is  a  transitory  form  between  soft  iron  and 

1  steel.     But  troostite  is  identified   by  the  microscope 

I  just  like  the  sorbite  :  its  composite  character  is  still  to  be 

1 74 


e  precipitaii 


determined.     The  systematic  microscopic  examination  consi«s 
briefly  in  the  application  of  three  methods: 

(i)  Relief  polishing,  (2),  etch  polishing,  and  (i)  etching 
with  tincture  of  iodine. 

In  relief  polishing  it  is  sometimes  advisable  to 
chalk  as  well  as  rouge,  to  preserv-e  the  ferrite. 

Id  etch  polishing  with  precipitated  chalk,  the  fundameiu. 
compounds,  with  the  exception  of  marteusite.  are  divided  )ii;i 
two  groups : 

(a)   Not  colored  :  ferrite,  cementite,  or  niartensite. 

(ft)  Colored  :  martensite,  troostite,  or  sorbite. 

Martensite  takes  only  a  yellowish  color  and  isdistinguishablt 
by  its  crystalline  form.  A  novice  might  take  martensite  ioi  per- 
litc,  especially  by  oblique  light,  for  both  have  irridescent  sheen, 
and  its  structural  elements  may  be  of  equal  dimensions:  but 
they  are  easily  distinguished,  as  the  needles  of  martensite  are 
straight  and  crossed,  while  those  of  perlite  are  curved  and  never 
cross  each  other. 

Ferrite  and  cementite  are  distinguished  by  their  great  dittcr 
ences  in  hardness  ;  the  former  is  low,  the  latter  is  high,  Tro" 
stite  takes  less  color  and  more  slowly  than  sorbite,  but  the  trut 
distinctive  mark  is  that  troostite  accompanies  martensite,  whik 
sorbite  goes  with  cementite  in  perlite. 

By  etching  with  tincture  of  iodine  two  groups  can  be  distin- 
guished, vis.  .■ 

{a)  Uncolored  :  ferrite  and  cementite. 

{d)  Colored  :  sorbite,  troostite  and  martensite. 

In  group  (i)  the  three  compounds  varj-  in  color,  in  kind  and 
depth  in  proportion  to  the  percentage  of  carbon  and  of  the  quan- 
tity of  tincture  of  iodine  used. 

References  :  "  Unification  of  Methods  of  Iron  Analyses."  By  Ptef. 
H.  Wedding.     SiaA/  und  Ehen,  No.  11,  /*aj. 

"Microphotography  of  Iron."  By  F.  Osmond,  Paris,  /.  Irvn  and 
Steel  Inst..  iSgo.  No.  i. 

"  Micro  photography  of  Iron."  By  A.  Martens,  Berlin.  SlaSt  und 
Eisen,  No,  10,  /Spj, 

•■On  the  Influence  of  Carbon  on  Iron."  By  John  Oliver  Arnold,  Bir- 
minghain,  England. 


"  Report  of  the  French  Commission  on  Testing  Materials."  Minisltrg 
des  7HIWJUJ-  Publics.  Paris.  1891  and  1895. 

•'Teslingof  Materials."  By  R.  A.  Hadfield.  /.  Iron  and  Steel  Inst,. 
-•«g./,So,  I. 

'■  The  Microstructure  of  Ingot  Irou  in  Cast  Ingots."  By  A.  Marteus. 
TroKs.  Am.  Inst.  Mining  Engineers,  as,  37-63,  iSgs. 

"  Dttemii nation  of  Combined  Carbon  in  Steel  by  the  Colorimetric 
Method."     By  J.  Blodgel  Britton.     Chem.  Nen's.  a6,  139. 

"  On  the  Estimation  of  Carbon  in  Pig  Iron."  By  Charles  H.  Pierce. 
(^^11.  News.  aS,  199, 

■"Colorimetric  Carbon  Estimation."  By  Fred  P.  Sharpless.  /.  Anal. 
C^em..  a,  55. 

"A  Fnnnel  for  Filtering  Carbon."     By  Thomas  M.  Drown.    /.  Anal. 


I.  33°- 

L.   Blum. 
By  L.  L.  de  Konick. 

Steels  by 

on  of  Carbon  in  the  Irons  of  Commerce. 
f-'^f'n.  Xews.  60,  167. 

"  Determination  of  Carbon  in  Iron  and  Steel.' 
-^'.'fAr.  auat.  duin..  1S8S.  463. 

''A  New  Form  of  Apparatus  for  Determination  of  Carbon 
Color."     By  C.  H.  Risdale.    /.  Soc.  Ckem.Ind.,  5,  583. 

"'  International  Standards  for  the  Analysis  of  Iron  and  Steel."  /,  Anal. 
^*cm..  6, 403. 

"  Nates  fiB  Carbou  in  Experimeutal  SiandardB."  By  P.  W.  Shimer. 
-/-  Wj»o/.  Chem..  6.  ijg. 

"  The  Determination  of  Carbon  in  Steel."     By  A.  A.   Blair.    /.  Anal. 

"  Researches  on  the  Carbon  of  White  Cast  Iron."  By  Isherwood. 
^''Si'ieer,  44.461. 

**  I>e termination  of   Combined   Carbon   in   Cast  Iron  and  Steel."     By 
"o^rnsend.     Proceedings  0/ the  EngiHtei s'  Club,  Philadelphia,  Pa.,  a,  31. 
"  Determination  of  Carbon  in  Iron  and  Steel,"     By  Zabudsky,     Ber.d. 
'**■»«.  Ges.,  16,  3318. 

"  *  The  Colorimetric  Determination  of  Combined  Carbon  in  Steel."  By 
■*Jfred  E.  Hunt.     Trans.  Am.  Inst.  Miti.  Eng..  la,  303. 

"Apparatus  for  the  Determination  of  Carbon  in  Iron  and  Steel  by 
Measurement  of  the  Evolved  Carbon  Dioiide."  By  Reinhsrt.  Stakl 
***ti  Eisrn,  la.  648, 

"  The  Determination  of  Carbon  in  Iron  and  Steel."  By  C.  fl.  Dudley 
»t>d  F.  N.  Pease.     The  American   Engineer  and  Railroad  Journal.  67, 

"A  Method  for  the  Determin 
Juliw,    J.  Anal.  Chem. .5.  i63. 

I   of  Carbon  in   Steel."     By  Franlt 



Determination  of  Phosphorus  in  Cast  Iron  and  Steel. 

The  molybdate  method,  as  described  by  Troilius/  gives 
uniform  and  satisfactory  results.  It  is  as  follows  :  Five  grams 
of  drillings  are  dissolved  in  a  No.  4  Grifl&n's  beaker,  in  nitric 
acid  (sp.  gr.  1.20),  using  about  fifty  cc.  of  the  acid.  The  solu- 
tion is  then  evaporated  with  excess  of  strong  hydrochloric  acid 
by  rapid  boiling  on  a  large  iron  plate  by  one  of  Fletcher's  solid 
flame  burners. 

The  plate  is  so  heated  that  the  heat  gradually  decreases  from 
the  centre  towards  the  edges.  The  hottest  part  ought  to  be 
rather  above  than  below  300**  C.  The  evaporation  is  continued 
on  the  hottest  part  of  the  plate  until  signs  of  spattering  are 
noticed.  The  beaker,  or  beakers,  are  then  moved  to  a  less  hot 
part  of  the  plate.  When  the  tendency  to  spatter  has  ceased  the 
beakers  are  moved  back  to  the  hottest  part  of  the  plate  for  at 
least  half  an  hour.  This  heating  is  necessary  in  order  to  com- 
pletely oxidize  and  decompose  the  last  traces  of  iron  phosphide, 
which  would  otherwise  remain  insoluble  with  the  silica.  The 
presence  of  hydrochloric  acid  lessens  the  tendency  to  spatter, 
which  is  always  less  in  high  carbon  steels  than  in  low  carbon 

The  beakers  are  now  slowly  cooled  and  strong  hydrochloric 
acid  added  in  excess.  This  acid  is  at  once  brought  to  a  boil, 
which  effects  a  solution  of  the  residue,  and  the  boiling  is  con- 
tinued until  only  a  small  bulk  remains. 

This  boiling  serves  two  purposes  : 

1.  To  convert  any  pyrophosphoric  acid  (H^P,0,),  which  may 
have  been  formed  by  the  strong  heating  into  orthophosphoric 
acid  (H,PO,). 

2.  To  concentrate  the  solution  and  remove  the   excess  of  ^ 

hydrochloric  acid  which  would  otherwise  interfere  with  the  pre-^ 

cipitation  of  phosphoric  acid  by  means  of  molybdic  acid. 

Hot  water  is  added  and  the  insoluble  residue  filtered  off  am 
thoroughly  washed  with  dilute  hydrochloric  acid,  and  afterwards 
with  hot  water. 

1  '•  Notes  on  the  Chemistr>'  of  Iron."  by  Magnus  Troilius,  E.  M. 

The  phospboric  acid  in  the  filtrate  is  precipitated  as  the  yel- 
low phospbo-molybdate  of  ammonia. 

For  this  precipitation  is  used  a  solution  of  about  one  part  by 
weight  of  molybdic  acid  in  four  weights  of  ammonia  (0.96  sp. 
gr. ).  and  fiftceu  parts  of  nitric  acid  (r.20  sp,  gr. ).  The  molyb- 
dic acid  is  first  dissolved  in  the  ammonia,  and  this  solution 
slowly  poured  into  the  nitric  acid,  which  must  be  shaken  con- 
stantly in  order  to  prevent  the  separation  of  molybdic  acid, 
which  redissoives  with  difficulty.  After  a  few  days'  standing 
the  solution  may  be  siphoned  off  clear.  Fifty  to  one  hundred  cc. 
of  this  solution  are  used  for  each  phosphorus  determination. 

To  precipitate  the  phosphoric  acid  in  the  filtrate  from  the  in- 
soluble residue  (silica,  etc.)  sufficient  ammonia  is  added  to 
nearly  neutralize  the  solution.  The  fifty  cc.  of  molybdic  acid 
solution  are  then  added  and  the  solution  well  stirred.  If  the 
yellow  precipitate  is  slow  in  coming  down,  a  little  more  ammo- 
nia may  be  added.  If  too  much  ammonia  is  added,  a  little 
strong  nitric  acid  must  be  introduced  to  redissolve  the  iron  pre- 
cipitate. As  a  rule  the  yellow  precipitate  comes  down  very 
quickly.  By  neutralizing  the  solution  before  adding  the  molyb- 
dic acid,  as  described,  the  yellow  precipitate  becomes  granular 
and  easy  to  filter.  When  precipitated  in  any  other  way  it  has 
a  tendency  to  pass  through  and  creep  over  the  edges  of  the  fil- 

The  yellow  precipitate  is  allowed  to  settle  over  night  at  about 
40°  C,  or  during  a  few  hours  at  80°  C. 

After  settling  the  clear  supernatant  liquid  is  siphoned  off  and 
the  precipitate  washed  with  copious  quantities  of  molybdic  acid 
solution  diluted  with  an  equal  volume  of  water.  About  300  cc.  of 
washing  are  not  too  much  to  insure  the  complete  removal  of  the  last 
traces  of  iron.  The  yellow  precipitate  is  then  treated  on  the  fil- 
ter with  six  cc.  hot  ammonia  (0.96  sp.  gr, )  and  the  filtrate  al- 
lowed to  tun  back  into  the  beaker  in  which  the  precipitation  was 
made.  When  all  is  dissolved  theammoniacal  solution  is  thrown  on 
the  same  filter  again,  but  now  allowed  to  run  into  a  loocc.  beaker. 
The  filter  is  then  washed  well  with  small  portions  of  cold  water, 
so  that  the  bulk  of  the  ammoniacal  solution  will  not  exceed  forty 
This  is  now  made  faintly  acid  with  hydrochloric  acid,  then 


alkaline  with  a  few  drops  of  ammonia,  enough  to  dissolve  any 
yellowsalt  that  may  have  separated.  Add  ten  cc.  of  magnesia 
mixture  and  stir  well  until  the  white  crj-stalline  precipitate  ot 
phosphate  of  magnesia  and  ammonia  appears;  about  six  cc.  of 
ammonia  (0,96  sp.  gr. )  are  then  added.  Allow  to  stand  two 
hours,  filter  upon  a  No.  2  ashless  filter  and  wash  with  diluted 
ammonia  (one  part  ammonia,  0.96  sp.  gr. ,  with  three  parts 
water) . 

About  eighty  cc.  of  this  mixture  are  sufficient  for  washing  the 
precipitate.  It  is  advisable  not  to  use  more  than  this  amount, 
as  the  same  has  a  slightly  solvent  action  upon  the  precipitate. 
The  white  precipitate  must  be  rubbed  loose  from  the  sides  of  ths 
beaker  with  rubber  tubing  on  a  glass  rod. 

The  "magnesia  mixture"  is  prepared  by  dissolving  110 
grams  of  crystallized  magnesium  chloride  together  with  280 
grams  of  ammonium  chloride  in  1300  cc.  of  water  and  adding 
700  cc.  oi  ammonia  (0.96  sp.  gr. )  to  the  solution. 

The  filter  with  the  well  washed  precipitate  is  ignited  in  a  small 
weighed  platinum  crucible  and  weighed  as  magnesium  pyro- 
phosphate, care  being  taken  that  the  ignited  precipitate  whea 
weighed  is  white  and  uniform  in  color.  Calculate  the  weight  of 
phosphorus  from  this  magnesium  pyrophosphate. 

If  it  be  desired  to  estimate  the  phosphorus  from  the  yellow 
precipitate  (ammonio-uiolybdic  phosphate)  directly,  proceed  as 
follows:  The  yellow  precipitate,  when  dried  at  95°  to  100'  C, 
contains  1.63  per  cent,  of  phosphorus.  It  must  be  washed  with 
water  containing  one  per  cent.,  by  volume,  of  nitric  acid  (1.2  sp, 
gr. )  instead  of  the  dilute  molybdic  solution.  After  drying  it  is 
transferred  from  the  filter,  by  shaking  and  brushing,  into  a 
weighed  watch-glass,  or  some  other  suitable  vessel  and  weighed. 
When  much  phosphorus  is  present  this  method  can  be  used  with 
great  accuracy,  but  when  little  the  risk  of  loss  is  too  great. 
Weighed  filters  must  then  be  used. 

The  magnesia  method  is,  however,  undoubtedly  the  better  of- 
the  two  in  general  working. 

When  precipitating  phosphoric  acid  with  the  molybdic  acid, 
solution  it  should  be  borne  in  mind  that  100  cc.  of  the  acidsoliif 
tion    are    required  for   the    complete  precipitation  of  one-tenth 





gram  of  phosphorus  pentoxide  containing  0.044  grant  ot  phos- 

Maay  forms  of  agitation  apparatus  have  been  devised  for  the 
thorough  precipitation  of  theammonio-magnesium  phosphate. 

The  apparatus  of  Spiegelbergs  (Fig.  48),  which  is  run  hy 


-water  power,  is  well  adapted  for  the  purpose  of  continued  aud 
\'ioIent  agitation  of  the  liquids. 

Volumetric  Dfterminatioii  0/ Phosphorus  in  Iron  and  Steel.' 

Put  one  gram  of  the  steel  in  a  ten  or  twelve  ounce  Erlen- 

meyer  flask  and  add  seventy-five  cc.  of  nitric  acid  (1.13 

When  solution  is  complete,  boil  one  minute  and  then  add  ten 

H  IMclhodsilapleilby  Motive  Cower  Dcpirtmcntol  Pcno.  R.  K.  Co.,  Uu<]lcyitid  Pease. 



cc.  of  oxidizing  potassium  permanganate  solution.  Boil  until 
the  pink  color  disappears  and  manganese  dioxide  separates,  re- 
move from  the  heat  and  then  add  crjstals  of  ferrous  sulphate, 
free  from  phosphorus,  with  agitation  until  the  solution  clears  up, 
adding  as  little  excess  as  possible.  Heat  the  clear  solution  to 
185'  F.,  and  add  seventy-five  cc.  of  molybdate  solution,  which 
is  at  a  temperature  of  80°  F.,  close  the  flask  with  a  rubber  stop- 
per and  shake  five  minutes,  keeping  the  flask  so  inclosed  during 
the  operation  that  it  will  lose  heat  very  .slowly.  Allow  to  stand 
five  minutes  for  the  precipitation  to  settle,  and  then  filter  through 
a  nine  cm.  filter  and  wash  with  acid  ammonium  sulphate  until 
the  ammonium  sulphide  tested  with  the  washings  shows  no 
change  of  color.  Dissolve  the  yellow  phospho-niolybdate  on  the 
filter  in  five  cc.  of  ammonia  (sp.  gr.  0.90),  mixed  with  twenty- 
five  cc.  of  water,  allowing  the  solution  to  run  back  into  the  same 
flask  and  thus  dissolve  any  yellow  precipitate  adhering  to  it- 
Wash  until  the  washings  and  filtrate  amount  to  150  cc,  then 
add  ten  cc.  strong  C.  P.  sulphuric  acid  and  dilute  to  scx)  cc. 
Now  pass  the  liquid  through  a  Jones  reductor  or  its  equivalent, 
wash  and  dilute  to  400  cc,  and  then  titrate  in  the  reduction 
flask  with  potassium  permanganate  solution. 

Apparatiii  and  Reagents. — The  apparatus  required  needs  no 
especial  comment,  except  perhaps  the  shaking  apparatus  and 
the  modification  of  the  Jones  reductor.  Accompanying  cut* 
represent  these  two.  The  shaking  apparatus  is  arranged  to 
shake  four  flasks  at  a  time,  which  is  about  al3  one  operator  can 
manipulate  without  the  solutions  becoming  too  cold.  The  cut  is 
about  one-twelfth  the  actual  size  of  the  apparatus.  The  flasks 
containing  the  solutions  rest  on  a  sheet  of  India  rubber  about 
one-quarter  inch  thick  and  are  held  in  position  by  the  coiled 
springs  as  shown.  There  is  a  recess  in  the  spring  arrangement 
to  receive  the  cork  of  the  flask.  Of  course  during  use  the 
door  of  the  box  is  closed,  the  cut  showing  it  open  so  that  the 
interior  may  be  seen.  The  niodificatiou  reductor  seems  to  work 
equally  as  well  as  the.  more  elaborate  apparatus.  The  cut 
about  one-fourth  the  actual  size.  As  will  be  seen  the  tube 
fitted  with  two  rubber  corks,  the  top  one  of  which  holds  the  fi 
nel  and  the  bottom  one  a  small  tube  which  also  fits  into  the  nil 



ber  cork  in  the  flask.  Next  to  the  bottom  cork  in  the  tube  is  a 
disk  of  perforated  platinum  ;  then  about  three  fourths  of  an  inch 
of  clean  white  sand,  then  another  perforated  platinum  disk  and 
then  the  tube  is  nearly  filled  with  powdered  zinc.  At  least  half 
the  zinc  may  be  used  out  before  it  is  necessarj-  to  refill. 

The  oxidizing  potassium  permanganate  solution  is  made  as  fol- 
lows: To  two  liters  of  water  add  twenty-five  grams  of  C.  P. 
cr>'stallized  potassium  permanganate  and  allow  to  settle  before 
using.     Keep  in  the  dark. 

The  molybdate  solution  is  made  as  follows :  Dissolve  loo 
grains  of  moiybdic  acid  in  400  cc.  of  ammonia  (sp.    gr.   0.96). 




and  filter.     Add  the  filtrate  to  one  liter  of  nitric  acid   (sp.  gr. 
1. 20).     Allow  to  stand  at  least  twenty-four  hours  before  using. 

The  acid  ammonium  sulphate  solution  is  made  as  follows  : 
To  one-half  liter  of  water  add  27.5  cc.  of  ammonia  <sp.  gr.  0.96) 
aud  then  twenty-lour  cc.  strong  C.  P.  sulphuric  acid  and  make 
solution  up  to  one  liter. 

The  potassium  permanganate  solution  for  titration  is  made  as 
follows :  To  one  liter  of  water  add  two  grams  of  crystallized  po- 
tassium permanganate  and  allow  to  stand  in  the  dark  not  less 
than  a  week  before  using.  Determine  the  value  of  this  solu- 
tion in  terms  of  metallic  iron.  For  this  purpose  0.150  to  0.200 
gram  of  iron  wire  or  mild  steel  are  dissolved  in  dilute  sulphuric 
acid  (ten  cc,  C.  P.  sulphuric  acid  to  forty  cc.  water)  in  a  long- 
necked  flask.  After  solution  is  complete,  boil  five  to  ten  min- 
utes, then  dilute  to  150  cc,  pass  the  liquid  through  a  reductor 
and  wash,  make  the  volume  up  to  300  cc.  Now  titrate  with  the 
permanganate  solution.  Severaldeterminationsshould  be  made. 
The  figures  showing  the  value  of  the  permanganate  solution  in 
terms  of  metallic  iron  should  agree  to  the  hundredth  of  a  milli- 

Calailations. — An  example  of  all  the  calculations  is  given 
herewith.  The  soft  steel  employed  in  standardizing  the  potas- 
sium permanganate  solution  contains  99.27  per  cent,  metallic 
iron.  0.1498  gram  of  this  contains  {0,1498  X  0.9927)  0.14S7064 
gram  Fe.  This  requires  42.99  cc.  permanganate  solution 
cc.^  0,003466  gram  Fe.  But  the  same  amount  of  permanganate 
solution  used  up  in  producing  the  characteristic  reaction  in  this 
amount  of  metallic  iron,  will  be  used  up  in  reaction  with  90,76 
per  cent,  of  the  same  amount  of  niolybdic  acid.  Hence  one  cc. 
of  the  permanganate  solution  is  equivalent  to  (0.003466  X  0.9076) 
0.003145  gram  of  the  molybdic  acid.  But  in  the  yellow  precipi- 
tate obtained  as  above  described,  the  phosphorus  is  1.90  per 
cent,  of  the  molybdic  acid.  Hence  one  cc.  of  permanganate 
solution  is  equivalent  to  (0.003145  X0.0190),  0.0000597  gram  oi 
phosphorus.  If,  therefore,  in  any  sample  of  steel,  tested  as 
above,  the  yellow  precipitate  requires  eight  and  six-tenthscc.  of 
permanganate,  the  amount  of  phosphorus  in  that  steel  is 
(0,0000597  X  8,6)  —  0.051  percent. 



Ten  cc.  of  the  '  *  magnesia  tnixiure  ' '  are  required  for  the  same 
quantity  of  phosphorus  pentoxide. 

RtfrreHtei.  "Volumetric  Estimation  of  Phosphorus  in  Iron  and 
Steel,"     By  Edward  D.  Csmpbell.    /.  Anal.  Chem..  i.  370. 

"  Note  on  Percentage  Composition  with  Table  (or  Phosphorus."  By 
William  St.  G.  Kent.    J.  Anal.  Chfin..  i,  164. 

"The  Elimination  of  Arsenic  in  Phosphorus  Determinations."  By  F. 
D.  Campbell.    /.  Anal.  CheiH.,  a,  370. 

"  Determination  of  Phosphorus  in  Iron  and  Steel."  By  Porter  W.  Shi- 
mer.    J.  Anal.  CAem.,  1,  97. 

"The  Inflnence  of  Silicon  on  the  Determination  of  Phosphorus  in 
Iron."     By  Thomas  M.  Drown.    /.  Anal.  Chtm..  3,  288. 

■'  Phosphorus  in  Pig  Iron,  Steel  and  Iron  Ore."  By  Clemens  Jones, 
/.  Anal.  Appl.  Ckem..  4,  368. 

■'  PhoBphorns  Determination  by  Neutraliiation  of  the  ■  Yellow  Precipi- 

te  ■  with  Alkali."     By  C.  E.  Manby.    /.  Anal.  Appl.  Chem..  6,  141. 

"  Note  on  the  Precipitation  of  Phosphorus  from  Solutions  of  Iron  and 
\  Steel."     By  Robert  Hamilton.    /.  Anal.  Appl.  Chem..  6,  573. 


ClassiBcation  of  Steel.' 

Classification  of  Steel  Made  hy  the  Midvale  Steel  Company. 

CIttssO.  Carbon  o.i  to  0.3  percent. 

Approximate  tensile  strength  from  55.0 

Approximate  tensile  strength  from  65, 

Class  II.  Carbon  o 

Approximate  tensile  strength  from  75, 

Class  in.  Carbon  c 

Approximate  tensile  strength  from  85, 

Class  IV.  Carbon  c 

Approximate  tensile  strength  from  95,0 

Class  V. 
Approiimate  le 

Class  VI. 
Approximate  \t 

Class  VII. 

sile  strength  from  105, 
Carbon  < 

silestrength  from  130, 

I  to  63.000  pounds. 
.3too.3per  cent. 
300  to  75,000  pounds. 
.3too.4  per  cent. 
300  to  85,000  pounds. 

Doo  to  95,000  ponnds. 
Sto  0.6  per  cent. 

00  to  103,000  pounds. 

00  to  120,000  pounds. 
.7100.8  per  cent. 
00  to  135,000  pounds. 
.8(00.9  P*'  cent. 

On  heats  of  this  carbon  and  above,  tensile  strength  is  not  con- 
sidered, as  they  are  generally  used  for  spring  steel  and  tool  steel, 
in  which  the  fitness  of  the  material  for  the  purpose  wanted  can- 
not be  decided  by  the  tensile  strength  ol  a  test  bar. 


Class VIII.  Carbon  0.9  to  i.o  per  cent. 

Class  IX  Carbon  1.00  to  per  cent. 

Class  X  Carbon  to  per  cent. 

It  is  of  course  understood  that  while  this  classification  holds 
good  in  a  general  way,  the  other  chemical  ingredients  besides 
carbon,  as  well  as  treatment,  may  so  effect  the  tensile  strength 
that,  while  the  percentage  ot  carbon  would  place  it  in  one  class, 
other  chemical  ingredients  or  physical  treatment  may  bring  it 
(as  far  as  tensile  strength  goes)  into  one  of  the  other  classes. 
As  a  general  thing,  it  has  been  found  that  a  high  percentage  of 
manganese,  say  above  seven-tenths  per  cent.,  up  to  and  includ- 
ing one  per  cent.,  will  exert  a  much  greater  hardening  influence 
on  steels  of  high  carbon  than  o(  steel  below  five-tenths  per  cent, 
in  carbon  ;  while  the  other  chemical  ingredients  seem  to  exert 
a  uniform  hardening  influence  on  all  grades  o(  steel. 

The  purposes  for  which  the  different  classes  of  steel  are  recom^j 
mended  by  the  Midvale  Steel  Company,  taking  into  considera- 
tion the  many  different  specifications  for  the  same  purposes  that 
are  received,  are  as  follows; 

Classes  I  and  II  are  used  for  propeller  shafting,  axles,  and 
general  machinerj-  work.  Also  used  for  rifle-barrel  steel,  steel 
castings  where  toughness  is  the  principal  requirement,  and 
finally,  in  the  higher  grades,  where  it  approaches  Class  III,  for 
gnn  tubes. 

Class  III  is  principally  used  for  Pennsylvania  Railroad  axlf^ 
and  crank  pins,  and  for  parts  of  machinery  where  a  high  elastic 
limit  is  required.  This  class  is  recommended  for  axles  and 
crank  pins,  and,  where  the  choice  is  left  with  the  makers,  they 
invariably  use  it  for  this  purpose.  It  ts  a  class  which,  in  their 
opinion,  is  best  suited  for  steel  forgings  of  all  descriptions,  with 
the  conditions,  however,  that  the  forgings  should  be  thoroughly 
annealed.  If  this  is  not  done,  the  lower  class  is  preferable,  as 
the  strains  left  in  the  forging  are  not  apt  to  be  injurious  in  the 
lower  carbon  steel.  This  class  is  also  used  for  gun  forgings, 
jackets  and  hoops,  the  high  requirements  as  to  elastic  limit 
making  it  necessary  to  have  a  good  percentage  of  carbon. 

Class  IV  is  used  also  principally  for  gun  forgings  and  fori 
large  locomotive  tires- 






Class  V  is  used  principally  for  tires  for  freight  service  and  car 
wheels,  and  for  forgings  for  air  vessels  lor  torpedoes,  and  also 
for  steel  castings  where  greater  wear  is  desirable,  such  as  ham- 
mer dies,  roll  pinions,  etc. 

Class  VI  is  used  mostly  for  surgical  iustruments  and  grinding 

Class  VII  is  used  for  spring  steel. 

Classes  VIII,  IX  and  X  are  used  for  various  grades  of  spring 
and  tool  steel,  the  highest  grades  being  used  for  cutting  tools 
and  the  lower  grades  for  chisels,  reamers,  etc. 

It  is  necessary  to  remember  in  this  classification,  that  while 
the  carbon  and  tensile  strength  governs  the  classes,  the  chemical 
composition  of  the  different  heats  that  come  under  one  class 
varies  considerably.  In  the  case  of  ordinary  machinery  steel 
and  tires,  the  makers  endeavor  to  keep  the  phosphorus  limit 
below  0.06  per  cent.  This  is  the  case  also  with  their  steel  cast- 
ings. In  gun  (orgiugs,  on  the  other  hand,  their  phosphorus 
limit  is  below  o.oj  per  cent.,  as  well  as  in  tool  steel  and  spring 
steel . 

At  the  present  moment,  the  greatest  interest  is  taken  in  the 
magnetic  qualities  of  steel,  as  compared  with  the  best  Norway 
iron;  and  from  recent  experiments  it  will  be  seen  that  conclu- 
sions can  not  be  drawn  with  safety  from  a  few  experiments,  par- 
ticularly in  regard  to  the  alloys  of  various  metals  with  steel. 
The  statement  has  been  broadly  made  that  a  large  percentage 
of  nickel  introduced  into  steel  castings  destroyed  the  magnetic 
qualities  of  the  steel  to  such  an  extent  as  to  make  this  alloy  par- 
ticularly desirable  or  useful  for  the  bolts  that  damp  the  puuch- 
ings  of  the  armature  in  the  dynamo  together,  the  general  idea 
being  that  these  foreign  substances  were  all  acting  injuriously. 
Some  recent  experiments  have  been  made  by  the  Bethlehem 
Iron  Company,  bearing  upon  the  dynamos  that  are  to  be  made 
for  Niagara,  which  seem  to  show  that  when  a  small  quantity  of 
nickel  only  is  used,  the  magnetic  qualities  are  improved  to  such 
a  degree  as  to  make  its  employment  advisable,  making  nickel 
steel,  properly  prepared,  higher  in  its  capability  of  magnetiza- 
tion than  even  the  best  Norway  iron.  This  statement  does  not 
hold  good  in  all  degrees  of  excitation,  but  is  said  to  be  particu- 


lady  good  at  the  amount  of  excitation  to  which  field  magnets  are 
usually  subjected. 

Mr.  L.  B.  Stillwell,  in  his  examination  of  this  metal,  con- 
cludes a  report  on  the  subject  with  the  words  :  **  I  am  emphati- 
cally of  the  opinion  that  no  better  material  can  be  secured.** 
The  effect  of  the  mixture  of  foreign  substances  with  steel  is  one 
that  is  worthy  of  the  most  careful  attention  of  the  students  of 
technical  colleges,  and  would  form  an  admirable  subject  for  a 
thesis,  as  the  experiments  to  be  reliable  need  not  involve  great 
cost,  and  would  give  opportunity  for  a  considerable  display  of 
ingenuity  in  devising  methods  of  making  the  tests,  and  the  man- 
ner of  showing  the  results  by  graphical  methods. 

Steel  plate  for  locomotive  use  requires  the  carbon  to  be  not 
under  0.15  per  cent,  nor  over  0.20  per  cent;'  phosphorus  0.03 
per  cent,  to  0.04  per  cent. ;  manganese  0.35  per  cent,  to  0.50  per 
cent. ;  silicon  0.025  P^^  cent,  to  0.04  per  cent.  :  sulphur  0.02 
per  cent,  to  0.04  per  cent. ;  copper  (if  any)  not  over  0.04  per 

1  The  Engineer,  March  30,  1895. 


■  11 











5"  D 

1  1  mi 



:    n' 

187         ' 



■  CAST   IRON. 

.  -  -  Ordinary 





.,       r. 



5.5  s 

5  15 





Obtained      Obuined 
by  direct     byindirect 

from  ores,     cast  iron. 
Catalan.        Finery, 
Siemens,    Lancashire 
Chcnot  and       and 
other  pro-     puddled 
cess  irons.       irons. 






2  O 

■  ■  Will  not  harden..                              1 

WROUGHT  STEEL'                              1 

ObUiued  by                                      ■ 

Shear                                            ■ 
Blister  and                                    ^M 

"'   m 




Determination  of  Aluminum  in  Iron  and  Steel. 

The  direct  determination  of  aluminum  in  iron  and  steel  is 
somewhat  difficult,  especially  if  the  amount  of  aluminum  be 

Drown*  describes  a  process  which  gives  good  results,  as  fol- 
lows : 

Dissolve  five  to  ten  grams  of  iron  or  steel  in  sulphuric  acid, 
evaporate  until  white  fumes  of  sulphuric  anhydride  begin  to  come 
off,  add  water,  heat  until  all  the  iron  is  in  solution,  filter  off  the 
silica  and  carbon,  and  wash  with  water  acidulated  with  sulp*huric 
acid.  Make  the  filtrate  nearly  neutral  with  ammonia,  and  add 
to  the  beaker  in  which  the  electrol3'sis  is  to  be  made,  about  loo 
times  as  much  mercury  as  the  weight  of  iron  or  steel  taken. 
The  bulk  of  the  solution  should  be  from  300  to  500  cc.  Con- 
nect with  the  battery  or  dynamo  current  in  such  a  way  that 
about  two  amperes  may  pass  through  the  solution  over  night. 
This  is  generally  accomplished  by  using  three  lamps  (thirty-two 
candle  power)  arranged  in  parallel  on  an  Edison  circuit.  In  the 
morning  the  solution  is  tested  for  iron,  and,  if  necessary,  the 
electrolysis  is  continued  after  adding  enough  ammonia  to  neu- 
tralize the  acid  that  has  been  set  free  by  the  deposition  of  the 
iron.  The  progress  of  the  operation  may  be  obser\'ed  by  the 
changing  color  of  the  solution.  At  first  it  becomes  darker  in 
color  near  the  anode  ;  after  five  or  six  hours  it  is  nearly  color- 
less, and  finally  becomes  pink,  from  the  formation  of  permanga- 

When  the  solution  gives  no  test  for  iron,  it  is  removed  from 
the  beaker  with  a  pipette  while  the  current  is  still  passing. 
When  as  much  has  been  removed  as  possible  without  breaking 
the  current,  water  is  added,  and  the  operation  continued  until 
the  acid  has  been  so  far  diluted  that  there  is  no  danger  of  dis- 
solving iron  from  the  mercury'.  The  anode  is  now  taken  out 
and  the  mercury  washed  with  water  until  the  last  traces  of  the 
solution  have  been  removed  from  it. 

After  filtering,  to  remove  any  fiakes  of  manganese  dioxide 

ly.  Anal.  Appl.  Chem.,  5.  631. 


which  may  be  suspended  in  the  solution,  sodium  phosphate  is 
added  in  excess  and  ten  grams  of  sodium  acetate.     The  solu- 
tion is  now  made  nearly  neutral  with   ammonia  and  boiled  for 
not  less  than  forty  minutes.     The  precipitate  of  aluminum  phos- 
phate is  then  filtered  off.  ignited,  and   weighed.     It  should  be 
white  after  ignition.     If  it  has  more  than   the  faintest   shade  of 
color  it   must  be   dissolved  by  fusing  with  acid  potassium  sul- 
phate, in  a  platinum  crucible,  and  again  electrolyzed  for  two  or 
three   hours.       The   second  precipitate  has  been    found   to    be 
always  white  without  a  trace  of  iron.     The  precipitate  of  alumi- 
.,nuiu  phosphate,  produced  as  above,  does  not  always  have  the 
Kcompositton  A1,0,.P,0,.     It  is  more  nearly  expressed  by  the 
Iformula  jA],0,^,  containing  24.14  per  cent. 

The  following  table  gives  the  results  obtained  in  determining 
I  by  the  above  process  the  aluminum  added  in  known  amounts 
|to  solutions  of  steel: 

«tl  Uken.  Per  cHil.  o(  nlunii-  Per  ginl  of  aluroi- 

UFim*.  DUm  added.  num  lound. 

5  o.W  0-36 

S  0.39  0.38 

S  0-39  0.38 

5  0-39  0-38 

5  0-39  0-37 

5  °o43  0.04S 

5  o  043  0048 

10  0.017  0.015 

to  0.300  0.160 

10  0.04S  0.044 

S  0.085  0088 

A  blank  experiment  with  the  same  steel,  without  the  addition 
\  of  any  aluminum,  gave  a  precipitate  of  aluminum  phosphate 
'  equivalent  to  0.004  P^""  cent,  of  aluminum.     Itmightbe  thought 
that  the  process  would  be  simplified  by  reducing  the  iron  to  the 
state  of  protoxide,  and  then  precipitating  alumina  as  basic  ace- 
tate, subsequently  removing  by  electrolysis  the  small  amount  of 
1  precipitated  with  the  alumina.     A  number  ot  experiments 
I  proved,  however,  that  this  modification  not  only  gave  less  accu- 
I  rate  results,  but  involved  much  more  work  than  the  precipitation 
f  all  of  the  iron  by  electrolysis. 


Method  of  Carnot, 

Treat  ten  grams  of  the  iron  or  steel  in  a  platinum  dish  covered 
with  platinum  foil,  with  hydrochloric  acid,  and  when  solution  is 
complete,  dilute  and  filter  into  a  fiask,  washing  the  carbon, 
silica,  etc.,  on  the  filter,  thoroughly  with  distilled  water.  Neu- 
tralize the  solution  with  ammonia  and  sodium  carbonate,  but 
see  that  no  permanent  precipitate  is  formed ;  then  add  a  little 
sodium  hyposulphite,  and  when  the  liquid,  at  first  violet,  be- 
comes colorless,  two  or  three  cc.  of  a  saturated  solution  of  sodium 
phosphate  and  five  or  six  grams  of  sodium  acetate  dissolved  in  a 
little  water.  Boil  the  solution  for  about  three-quarters  of  an 
hour,  or  until  it  no  longer  smells  of  sulphurous  acid.  Filter  and 
wash  the  precipitate  of  aluminum  phosphate  mixed  with  a  little 
silica  and  ferric  phosphate,  with  boiling  water.  Treat  the  pre- 
cipitate on  the  filter  with  hot  dilute  hydrochloric  acid,  allow  the 
solution  to  run  into  a  platinum  dish,  evaporate  to  drj'uess,  and 
heat  at  lOo"*  C.  for  an  hour  to  render  the  silica  insoluble.  Dis- 
solve in  hot  dilute  hydrochloric  acid,  filter  from  the  silica,  dilute 
to  about  100  cc.  with  cold  water,  neutralize  as  before,  add  a 
little  hyposulphite  in  the  cold,  then  a  mixture  of  two  grams  of 
sodium  phosphate  and  two  grams  of  sodium  acetate,  boil  until 
all  sniell  of  sulphurous  acid  has  disappeared,  filter,  wash,  ignite, 
and  weigh  as  Al,0,.P,Oj,  which  contains  22.18  per  cent,  of 

References  .•  **  A  Rapid  Method  for  the  Determination  of  Aluminum  in 
Iron  and  Steel."   Chem.  News,  61,  313. 

''  On  the  Determination  of  Minute  Quantities  of  Aluminum  in  Iron  and 
Steel."     By  John  E.  Stead,  F.  I.  C.,/.  Sac,  Chem.  Industry,  /8S9,  p.  956. 


Determination  of  Sulphuric  Acid  and  Free  Sulphur  Trioxide 
in  Fuming  Nordhausen  Oil  of  Vitriol. 

As  this  acid  fumes  immediately  upon  exposure  to  the  air,  also 
rapidly  absorbing  moisture,  great  expedition  must  be  exercised 
in  obtaining  the  samples  tor  analysis. 

ly.  Ana/.  CMfm.,  5,  178. 



Select   a   small   picnometer   (Fig.   51),  weight   about   eiglit 
grams,  and  determine  its  weight  with  great  accuracy. 
Insert  a  pipette  into  the  Nordhausen  acid,  and  witli- 
out  suction  allow  about  t^x'o  grams  of  the  acid  to 
run  into  the  pipette.     Remove  the   stopper  of  the 
picnometer,  insert  the  lower  end  of  the  pipette  into 
it.  allow  the  acid  to  flow,  remove  the  pipette  and  in- 
sert the  stopper  of  the  picnometer.     Weigh  the  pic- 
p'v  i'-      nometer  and  acid  carefully  to  the  fourth  decimal ; 
then  drop  it  into  a  tall  beaker  (capacity  800  cc. )   containing 
about  500  cc.  of  distilled  water  and  cover  with  a  watch  glass; 
remove  ihe  stopper  of  the  picnometer  at  the  moment  the  latter 
is  dropped  into  the  water.     Too  much  acid  should  not  be  used, 
three  grams  being  the  maximum  amouut. 

Determine  the  amount  of  acid  present  by  titration  with  a  solu- 
tion of  soda,  which  will  give  the  total  sulphur  trioxide,  but  as 
Nordhausen  acid  is  composed  of  varying  amounts  of  a  mixture 
of  sulphuric  acid  and  sulphur  trioxide,  it  will  be  well  to  explain 
the  niethod  in  detail. 

Picnometer  and  Nordhausen  acid  =  S.yiro  grams. 
I  Picnometer  ^7.6320       " 

Nordhausen  acid^  1.0890 

Amount  of  soda  solution  required  to  neutralize  1.089  grams  of 
the  acid  =  28.7  cc. 

One  cc.  of  the  soda  solution  is  equivalent  to  0.0401  gram  sul- 
phuric acid  or  0.0327  gram  sulphur  trioxide. 

The  acid  therefore  contains  86.2  per  cent,  of  sulphur  trioxide 
atid  13.8  per  cent,  of  water. 

To  determine  the  proportions  of  sulphur  trioxide  and  sul- 
phuric acid  the  following  formulas  are  used  ; 

Let  X  =  H,SO.  in  the  acid. 

:  SO. 

x-\-j'=  100. 

9»jr+g8y  —  9800. 
8o;rH-98j'=  8447.6 


-^  =  75.1  per  cent,  of  H,SO,  in  the  acid. 
100  —  75.1  =  24.9  per  cent,  of  SO,  in  the  acid. 
y  =  24.9  per  cent  of  SO,  in  the  acid. 

75.1  +  24.9^  joo 

Nordhausen  acid  often  contains  small  amounts  of  sulphur 
dioxide.  This  should  be  boiled  out  of  the  water  before  titra- 
tion with  the  soda  solution. 

Determination  of  Manganese  in  Iron  and  Steel. 

Manganese  can  be  determined  accurately  in  iron  and  steel 
colorimetrically,  graviraetrically  or  volumetrically.  The  latter 
method  is  in  general  use  as  being  expeditious. 

For  the  gravimetric  and  volumetric  methods  the  initial  treat- 
ment may  be  the  same,  that  is,  solution  of  the  steel  in  nitric 
acid ;  the  precipitation  of  the  oxide  of  manganese  by  means  of 
the  nitric  acid  and  potassium  chlorate,  and  its  filtration  and 

Five  grams  of  the  steel  are  transferred  to  a  No.  5  beaker  and 
150  cc.  of  nitric  acid  (sp.  gr.  1.2)  added.  After  solution  of  the 
iron  and  concentration  to  about  100  cc,  there  is  added  fifty  cc. 
nitric  acid  (sp.  gr.  1.42)  and  the  boiling  continued  till  the  bulk 
of  the  liquid  amounts  to  about  100  cc.  To  this  is  added  crys- 
tals of  potassium  chlorate  (not  over  three  grams)  gradually^ 
and  the  boiling  continued  until  no  more  fumes  of  chlorous  gas 
are  emitted.  Allow  to  cool,  add  twenty-five  cc.  nitric  acid  (sp. 
gr.  1.42)  and  filter  upon  an  asbestos  filter,  washing  twice  with 
strong  nitric  acid  and  five  times  with  cold  water.  Transfer  the 
filter  and  contents  to  a  beaker  and  treat  a,  for  gravimetric  de- 
termination, or  b  for  volumetric  determination  of  the  manganese. 

a.  Add  seventy-five  cc.  hydrochloric  acid  (strong)  and  boil; 
the  manganese  dioxide  is  dissolved.  The  solution  is  diluted 
with  water  and  the  asbestos  separated  therefrom  by  filtration 
upon  a  No.  4  filter,  and  well  washed.  The  filtrate  is  made 
faintly  alkaline  with  ammonia,  then  to  acid  reaction  with  acetic 






acid,  and  boiled.  Filter  off  any  basic  acetate  of  iron  that  may 
be  present,  and  to  the  filtrate  add  ammonium  hydroxide  to  alka- 
line reaction  and  then  bromine  (not  over  one  cc. );  shake  well, 
set  aside  two  hours,  then  boil,  filter,  dry,  ignite,  and  weigh  as 
Mn.O..     Consult  scheme  XIII. 

6.  Instead  of  dissolving  the  manganese  dioxide  in  hydro- 
chloric acid,  as  in  a.  it  is  dissolved  in  a  measured  amount  of 
standard  acid  solution  of  ferrous  sulphate,  and  the  excess  of  fer- 
rous sulphate  determined  hy  a  standard  solution  of  potassium 
bichromate.  The  ferrous  sulphate  solution  is  made  by  dissolv- 
ing twenty  grams  crj-stallized  ferrous  sulphate  in  1600  cc.  water 
and  adding  thereto  400  cc.  of  sulphuric  acid  (sp.  gr.  1.5). 

The  bichromate  solution  is  made  by  dissolving  ten  grams  of 
potassium  dichromate  in  1000  cc.  water.  One  cc.  of  the  ferrous 
sulphate  solution  corresponds  to  o. 01 1  gram  of  iron,  that  is,  it 
will  oxidize  the  amount  of  ferrous  sulphate  to  ferric  sulphate 
that  corresponds  to  0.011  gram  of  iron.  One  cc.  of  the  bichro- 
tuate  solution  corresponds  to  0.0054  gram  manganese. 

The  manganese  dioxide  precipitate,  obtained  from  the  five 
grams  of  steel,  is  dissolved  in  100  cc.  of  the  acid  ferrous  sul- 
phate solution  ;  it  is  then  titrated  with  bichromate  solution  until 
a  drop  of  the  liquid  placed  on  a  porcelain  slab  and  brought  in 
contact  with  a  drop  of  /res/i  dilute  solution  of  potassium  ferricy- 
anide  shows  no  blue  or  green  color,  but  a  faint  brown  color, 
{Scheme  VIII)  indicating  complete  oxidation. 

The  amount  of  bichromate  that  would  he  required  to  oxidize 
the  total  iron  in  the  100  cc.  would  be  iS.i  cc,  but  in  this  ex- 
periment 15.1  cc.  were  required,  showing  that  the  oxidizing 
action  of  three  cc.  of  the  bichromate  solution  had  been  sup- 
planted by  the  action  of  the  manganese  dioxide.  Since  thre# 
cc.  ofthe  bichromate  corresponds  to  0.0162  gram  manganese 
dioxide,  and  this  amount  is  obtained  from  five  grams  of  the 

,     ,  ,  , ,    o.oi6a  X  100 

steel,  the  per  ceut.  of  manganese  dioxide  will  he — 

0.324  per  cent.      Some  chemists  prefer  the  use  of  a  solution  of 

potassium    permanganate    instead    of    potassium    bichromate. 

(Consult.  Trans.  Ami^rican  /mi.  Mining  Engineers,    10.    100.) 

T-he  color  method  may  he  stated  briefly  as  follows :  In  a  test- 


tube,  similar  to  that  used  for  the  estimation  of  carbon,  place  two- 
tenths  gram  of  the  sample  to  be  tested,  and  in  a  like  tube  the 
same  quantity  of  a  standard  steel,  in  which  the  manganese  has 
been  carefully  determined  by  weight.  To  each  add  five  cc. 
nitric  acid  (sp.  gr.  1.20),  and  boil  in  a  beaker  of  hot  water  until 
solution  is  complete.  Cool  the  tubes,  and  to  each  add  an  equal 
bulk,  about  two  cc.  of  water ;  replace  in  the  beaker,  and,  after 
boiling  for  a  few  minutes,  add  an  excess  of  lead  peroxide,  which 
must  be  free  from  manganese,  and  ten  drops  of  nitric  acid  (sp. 
gr,  1.42.)  After  boiling  for  four  minutes  the  tubes  are  with- 
drawn and  placed  in  a  beaker  of  cold  water.  When  the  per- 
oxide of  lead  has  completely  settled,  transfer  two  cc.  of  the  clear 
supernatant  liquid  of  the  standard  solution  to  the  graduated 
tube  used  in  the  colorimetric  estimation  of  carbon,  dilute  to  five  cc. 
with  cold  water,  mix.  In  a  similar  tube  place  the  same  quan- 
tity of  the  solution  of  the  sample  which  is  being  tested,  diluting 
with  water  until  its  color  is  of  the  same  intensity  as  that  of  the 
standard.  Read  off  the  number  of  cc.  to  which  dilution  is  car- 
ried, from  which,  by  a  simple  calculation,  the  percentage  is 
easily  determined.* 

Textor's  Method  for  the  Rapid  Determination  of  Manga- 
nese in  Steel. 

To  one-tenth  gram  of  steel,  in  a  No.  2  beaker,  add  fifteen  cc.  of 
nitric  acid  (sp.  gr.  1.20)  ;  boil  until  the  brown  oxides  of  nitro- 
gen are  gone;  add  fifteen  cc.  of  hot  water,  and  while  boiling 
introduce  one-half  gram  of  lead  peroxide.  Boil  three  minutes 
after  the  addition  of  the  lead  peroxide,  filter  through  asbestos, 
and  wash  with  water  containing  two  per  cent,  nitric  acid  (sp. 
gr.  1.20).  Titrate  with  a  solution  of  arsenious  acid  till  the  pink 
color  is  gone  ;  each  cubic  centimeter  of  solution  equals  one- 
tenth  per  cent,  of  manganese.  ^ 

Precautiojis, — The  brown  fumes  must  all  be  expelled  before 
adding  water,  otherwise  low  results  may  be  expected.  Before 
filtering,  the  asbestos  must  be  treated  with  nitric  acid.  For 
steels  containing  0.75  per  cent,  of  manganese,  one-half  gram  or 
more  lead  peroxide  should  be  added,  and  the  solution,  after  the 

I  J.  J.  Morf^an  :  Chcm.  Sru's,  56,  82. 


ZINC    IN   ORES.  195 

addition  of  the  lead,  should  be  boiled  not  less  than  three  miuiites, 
otherwise  low  results  may  be  obtaiued.  To  secure  rapid  filtra- 
tion, a  special  filter  is  required.  It  may  be  constructed  as  fol- 
lows :  Fill  a  two  and  one-half  inch  funnel  one-third  to  a  half 
full  with  piecesof  glassrodone-quarterto  one-half  inch  long;  on 
this  place  a  disk  of  platinum  foil  fitting  the  funnel  at  the  point 
where  the  disk  rests  on  the  broken  glass.  The  platinum  disk 
is  perforated  by  means  of  a  pin,  over  its  whole  surface  ;  the 
rough  side  is  turned  down.  Pour  suspended  asbestos  upon  the 
foil  till  a  layer  is  formed  one-half  inch  in  thickness.  When  the 
filter  becomes  clogged  and  works  slowly,  the  thin  layer  of  lead 
peroxide  can  be  removed  by  carefully  scraping  with  a  wire,  a 
fresh  surface  of  asbestos  thereby  becoming  exposed. 

For  the  arsenic  solution,  twenty  grams  of  arsenic  trioxide  in 
powder  and  sixty  grams  of  sodium  carbonate  are  dissolved  in 
750  cc.  of  hot  water,  filtered  and  diluted  to  2000  cc.  An  equiva- 
lent amount  of  sodium  arsenite  may  be  conveniently  taken.  Of 
this  solution,  87.5  cc.  are  diluted  to  2500  cc.  and  tested  with  a 
steel  containing  a  known  percentage  of  manganese.' 

Reference!:  "Colorimetric  Estimation  of  Mangauese  in  Steel."  By 
B.  W-  Cbeever./.  Anal.  CJiem..  i.  88. 

"  Volumetric  Determination  of  Manganese."  By  J.  Pattiaon.y,  CAetx. 
■Smr.,  35.  365. 

"  Method  for  the  Rapid  Determinatiou  o{  Manganese  in  Slags,  Ores, 
Etc."     By  F.  G.  Myhlertz,y.  A»a/.  Chem..  a,  367. 

Technical  Determination  of  Zinc  in  Ores. 
Prepare  a  solution  of  potassium  ferrocyanide  by  dissolving 
iorty-four  grams  of  the  pure  salt  in  distilled  water  and  diluting 
to  one  liter.  Standardize  as  follows :'  Dissolve  200  milli- 
grams oi  pure  zinc  oxide  in  ten  cc.  of  strong  pure  hydro- 
chloric acid.  Add  seven  grams  of  C.  P.  ammonium  chloride, 
and  about  100  cc.  of  boiling  hot  water.  Titrate  the  clear 
liquid  with  the  ferrocyanide  solution  until  a  drop,  tested  on 
a  porcelain  plate  with  a  drop  of  a  strong  aqueous  solution  of 
uranium  acetate,  shows  a  brown  tinge.     About  sixteen  cc.  of 

lEDKidccri'SoiHelyol  WesWrn  pa..  Ttnun.,  1892. 
«  Hclbod  o(  VOQ  Schul.  and  Low. 


ferrocyanide  will  be  required,  and  accordingly  this  amount 
may  be  run  in  rapidly  before  making  a  test,  and  then  the  titra- 
tion finished  carefully  by  testing  after  each  additional  drop 
of  ferrocyanide.  As  soon  as  a  brown  tinge  is  obtained  note  the 
reading  of  the  burette,  and  then  wait  a  minute  or  two  and  ob- 
serve if  one  or  more  of  the  previous  tests  do  not  also  develop  a 
brown  tinge.  Usually  the  end-point  wilt  be  found  to  have  been 
passed  by  a  test  or  two,  and  the  proper  correction  must  then  be 
applied  to  the  burette  reading.  Finally  make  a  further  deduc- 
tion from  the  burette  reading  of  the  amount  of  ferrocyanide  re- 
quired to  produce  a  brown  tinge  under  the  same  conditions 
when  no  zinc  is  present.  This  correction  is  about  two  drops,  or 
0.14  cc.  Two  hundred  milligrams  of  zinc  oxide  contain  160.4 
milligrams  of  zinc,  and  one  cc.  of  the  above  standardized  sol- 
ution will  equal  about  0.01  gram  of  zinc,  or  about  one  per  cent., 
when  one  gram  of  ore  is  taken  for  assay. 

Prepare  the  following  solutions  for  the  assay  of  ores : 

A  saturated  solution  of  potassium  chlorate  in  nitricacid,  made' 
by  shaking  an  excess  of  crystals  with  the  strong  acid  in  a  flaskj 
Keep  the  solution  Jn  an  open  flask. 

A  dilute  solution  of  ammonium  chloride  containing  about  teit'l 
grams  to  the  liter  ;  (or  use  heat  to  boiling  in  a  wash  bottle. 

Take  exactly  one  gram  of  the  ore  and  treat  in  a  three 
one-half  inch  casserole  with  twenty-five  cc,  of  the  above  chlorate 
solution.  Do  not  cover  the  casserole  at  first,  but  warm  gently 
until  any  violent  action  is  over  and  greenish  vapors  have  ceased 
to  come  off.  Then  cover  with  a  watch-glass  and  boil  to  com- 
plete dryness,  but  avoid  overheating  and  baking.  Cool  suffi- 
ciently and  add  seven  grams  of  ammonium  chloride,  fifteen  cc. 
strong  ammonia  water,  and  twenty-five  cc.  hot  water.  Boil  the 
covered  mixture  one  minute  and  then,  with  a  rubber-tipped 
glass  rod,  see  that  all  solid  matter  on  the  cover,  sides  and  bot- 
tom of  casserole  is  either  dissolved  or  disintegrated.  Filter  into 
a  beaker  and  wash  several  times  with  the  hot  ammouium 
chloride  solution.  A  blue  colored  solution  indicates  the  pres- 
ence of  copper.  In  that  case  add  twenty-five  cc.  strong  pure 
hydrochloric  acid  and  about  forty  grams  of  granulated  test-lead. 
Stir  the  lead  about  in  the  beaker  until  the  liquid  has  become 






perfectly  colorless  and  then  a  Httle  longer  to  make  sure  that  all 
the  copper  is  precipitated.  The  solution,  which  should  bequite 
hot,  is  now  ready  for  titration.  In  the  absence  of  copper  the 
lead  is  omitted  and  only  the  acid  added.  'About  one-third  of 
the  solution  is  now  set  aside,  and  the  main  portion  is  titrated 
rapidly  with  the  ferrocyauide  until  the  end-point  is  passed,  us- 
ing the  uranium  indicator  as  in  the  standardization.  The  greater 
part  of  the  reserved  portion  is  now  added,  and  the  titration  con- 
tinued with  more  caution  until  the  end-point  is  again  passed. 
Then  add  the  remainder  of  the  reserved  portion  and  finish  the 
titration  carefully,  ordinarily  by  additions  of  two  drops  of  ferro- 
cyanide  at  a  time.  Make  corrections  of  this  final  reading  of  the 
burette  as  in  the  standardization. 

Gold,  silver,  lead,  copper,  iron,  manganese,  and  the  ordinary 
constituents  of  ores  do  not  interfere  with  the  above  scheme.  Cad- 
mium behaves  like  zinc.  When  known  to  be  present  it  may  be 
removed,  together  with  the  copper,  by  the  proper  treatment  with 
hydrogen  sulphide,  and  the  titration  for  zinc  may  be  made  upon 
the  properly  acidified  filtrate  without  the  removal  of  the  excess 
of  gas. 

Sodium  Cyanide  as  a  Component  of  Potassium  Cyanide. 

The  valuation  of  potassium  cyanide  for  commercial  purposes, 
is  deRendeut  upon  the  amount  of  cyanogen  present,  the  salt  being 
rated  from  "thirty  percent,  cyanide"  to  "ninety-eight  percent. 
cyanide  " — the  former  selling  for  twenty  cents  and  the  latter  for 
sixty  cents  per  pound.  The  determination  of  the  percentage  of 
cyanogen  is  usually  made  by  titration  with  semi-normal  silver 
solution,  and  in  chemical  manufactories  where  potassium  cyan- 
ide is  made,  generally  constitutes  the  entire  analysis.  Potas- 
sium cyanide,  when  pure,  contains  forty  per  cent,  of  cyanogen ; 
"  niuety-eight  per  cent."  would,  therefore,  indicate  39.2  per 
cent,  of  cyanogen,  and  "  thirty  per  cent.,"  twelve  per  cent,  of 
cyanogen.  An  analysis  of  a  sample  of  the  former  gave  by  titra- 
tion 42.33  per  cent,  of  cyanogen,  or  a  rating  of  105.87  per  cent, 
of  potassium  cyanide.  This  result  immediately  showed  that 
another  base  than  potassium  was  present,   and  one  also  whose 


combining  weight  was  less.  Sodium  being  indicated  by  quali- 
tative analysis,  a  quantitative  analysis  of  the  sample  was  neces- 
sary to  determine  the  proportions  of  potassium  and  sodium  com- 
bined with  the  cyanpgen. 

The  method  adopted  was  as  follows:  The  cyanide  was 
weighed,  transferred  to  a  platinum  capsule,  sufficient  water 
added  for  solution,  then  dilute  sulphuric  acid  in  excess  and  con- 
tents evaporated  to  dryness  and  ignition  to  constant  weight. 
This  represented  sulphates  of  potassium  and  sodium,  and  after 
solution  in  water  and  acidifying  with  hydrochloric  acid,  the 
sulphuric  acid  was  precipitated  and  weighed  as  barium  sulphate 
and  calculated  to  SO,. 

These  determinations  gave  a  method  of  obtaining  the  propor- 
tions of  potassium  and  sodium  in  the  weighed  alkaline  sulphate 
as  follows : 

94.2  parts  Kfi  require  80  parts  SO,  for  K^SO^ 
62.0    '*      Na,0     *'        80      **      SO,   "   Na,S04 

Let  G  =  weight  of  sulphates. 
*'    x=        *'       **   K,0. 

80  80 

Or,     G  =  j: -\-y -{- o.Ss  X  +  i,2gy 

__   1.85  SO,  —^85^ 

^"         0.4387 

;r  =  G^— (SO,-|->) 

Having  obtained  the  values  of  potassium  oxide  and  sodium 
oxide,  they  are  calculated  to  potassium  and  sodium.  These 
weights  are  multiplied  by  100  and  divided  by  the  weight  of  cyanide 
taken,  the  results  being  the  percentages  of  potassium  and  sodium 
respectively  in  the  cyanide.  If  to  these  results  is  added  the  per- 
centage of  cyanogen,  as  determined  by  titration  with  semi-nor- 
mal silver  solution,  the  analysis  is  completed. 

A  sample  of  the  cyanide  above  mentioned  as  containing 
sodium  as  well  as  potassium,  gave  the  following  : 


Amount  of  salt  taken  for  analysis,  i. 519 grams. 

Platinum  capsule  and  alkaline  sulphates 46.625  grams. 

"      44.573      " 

K,S04  +  Na,S04 2.052      '* 

Crucible  and  BaSO^ 25.165      '* 

22.306      ** 

BaSOi 2.859      ** 

Equivalent  to  0.981  gram  SOg.    Cyanogen  by  titration  was  42.33  per 

Na,0  =  1-85  (0-981)  -  0-85  (2.05a)  ^  ^^6, 

=3  0.120  gram  sodium,  or  7.90  per  cent. 

K,0  =  2.052  —  (0.981  +  0.161)  =  0.910  gram. 
«  0.755  gram  potassium,  or  49.70  per  cent. 

Resulting : 

Sodium 7.90  per  cent. 

Potassium 49-7© 

Cyanogen 42.33 

Undetermined 0.07 

Total 100.00    **      '* 

Equivalent  to  : 

Sodium  cyanide 16.90  per  cent. 

Potassium  cyanide 82.83    "      ** 

Difference 0.20    **      ** 

Undetermined 0.07 

II      II 

II      II 

Total 100.00    •*      ** 

This  cyanide  of  potassium  and  sodium  (though  marked  *'  pot- 
assium cyanide,  ninety-eight  per  cent.  '*)  is  sold  at  a  lower  rate 
than  the  "  ninety-eight  per  cent,  potassium  cyanide,'*  and  for 
many  purposes  is  superior,  as  it  contains  a  higher  percentage  of 
cyanogen.  An  examination  of  the  formula  for  its  manufacture 
shows  that  it  can  be  made  at  a  less  cost  than  the  potassium 
cyanide  alone.  Potassium  ferrocyanide,  or  sodium  ferrocyanide 
when  heated  in  covered  crucibles  is  converted  into  potassium 
or  sodium  cyanide,  iron  carbide  and  nitrogen : 

2K,Fe(CN),=  8KCN+2FeC,+  N, 
2Na,Fe  (CN).=  8NaCN+  2FeC,  +  N, 


lOO  pounds  of  potassium  ferrocyanide,  at  thirty  cents  per  pound, 
produces  70,63  pounds  of  potassium  cyanide,  ninety-eight  per 
cent.,  at  a  cost  of  forty-two  cents  per  pound  ;  and  loopoundsof 
sodium  ferrocyanide,  at  twenty  cents  per  pound,  produces  64.47 
pounds  of  sodium  cyanide,  ninety-eight  per  cent.,  at  a  cost  ol, 
thirty -one  cents  per  pound. 

If  a  mixture  composed  of  117  pounds  of  potassium  ferrocyan- 
ide and  twenty-six  pounds  of  sodium  ferrocyanide  be  heated  in 
covered  crucibles,  the  resulting  compound,  weighing  loopounds, 
will  closely  approximate,  in  composition,  the  sample  submitted, 


The   Chemical    and    Physical    Examination    of    Portland 


The  enlarged  consumption  of  Portland  cement  in  this  countryi 
during  the  past  few  years  has  caused  the  subject  of  its  chemicaT. 
and  physical  properties  to  receive  increased  consideration.  Not 
only  has  the  consumer  been  directly  interested ,  that  the  cements 
used  should  stand  special  tests,  but  the  attention  of  the  manu- 
facturer has  been  drawn  in  the  same  direction,  resulting  in  im- 
provements in  methods  of  production.  1 

A  number  of  causes  have  prevented  the  use  of  American  Port-' 
land  cements  in  the  home  market,  one  of  the  chief  being  that 
the  imported- German  cements  always  give  higher  ph3-sical 
tests  when  made  by  the  German  methods  of  testing  than  the 
American  cements  under  the  American  system  of  testing. 
There  are  a  number  of  American  Portland  cements  fully  as  good 
as  the  best  German  cements,  and  have  shown  fully  as  high  ten- 
sile strength  when  tested  by  the  same  methods. 

These  differences  in  results  are  not  due  entirely  to  the 
cements,  but  rather  to  the  methods  in  use  in  the  different  coun- 
tries for  testing  them,  for  Portland  cements  cannot  vary  mucb. 
in  their  chemical  composition  without  losing  their  value. 

The  limit  of  variation  is  as  follows  : 

CaO 58.0  to  67.0  per  ccn 

Eindt  pnulique  mr  U  Cim 



After  manufacture  it  is  practically  Ca,SiOj,  and  is  quite  dis- 
tinct from  another  product  made  and  largely  consumed  here 
called  •*  hydraulic  cement/* 

Experience  has  shown  that  Portland  cements  containing  over 
two  per  cent,  of  magnesia  (MgO)  are  inferior  in  lasting  quali- 
ties, and  by  the  gradual  absorption  of  water  produce  cracking 
and  disintegration.* 

Calcium  carbonate  (CaCO,),  formed  by  the  absorption  of 
carbon  dioxide  by  the  lime  in  the  cement  after  manufacture,  is 
another  injurious  compound  found  in  cements  containing  more 
lime  than  sufficient  to  unite  with  the  silica  to  form  tri-silicate  of 
lime.  This  carbonate  of  lime  gradually  produces  seams  and 
fractures  after  the  setting  of  the  cement.  The  '*  Ecole  Nation- 
ale,**  of  Paris,  rejects  all  cements  containing  over  one  and  five- 
tents  per  cent,  of  sulphuric  acid.  Thus,  if  upon  chemical  analy- 
sis, magnesia  is  found  present  in  amounts  over  two  per  cent. 
carbonic  and  sulphuric  acids  in  amounts  over  one  and  one-half 
per  cent.,  the  cement  can  be  condemned  at  once  without  any  mechani- 
cal tests.  Therefore,  it  is  evident  that  a  careful  test  of  a  Port- 
land cement  requires  :  (i)  a  chemical  analysis  to  determine  the 
proportion  of  the  ingredients,  and  (2)  the  mechanical  or  physi- 
cal tests  to  determine  fineness,  tensile  strength,  and  resistance  to 

The  following  scheme  is  arranged  to  show  the  method  of 
making  a  cement  analysis  : 

1  Compt.  rend..  May,  1886. 













E  • 











1  ft  ? !  i^!4^:  stf 

liltl                 i 



iiiiffli  '' 

!#p       ^ 

liii     ^ 



Weight  of  SiO,  X  100  .0/^ 
=  per  cent.  SiO,. 

(1)  Crucible  +  SiO, 11.205  grams. 

Crucible 10.721      ** 

SiO,  ^  0.484 

0.484  X  100  ^  -.^ 

—  =  24.20  per  cent.SiO,. 



Weight  Al,Oa  X  100  *    ai/^        ..t.    •       %  u^  j 
2 i—f ^  per  cent.  Al^Oj  in  the  insoluble  residue. 

( 1 )  Crucible  +  A1,0, 10.743  grams. 

Crucible 10.721      ** 

A1,0,  =    0.022      *' 

0.022  X  100  . ,  ^ 
^  per  cent.  A1,0,. 

Weight  of  Fe,Os,  X  2.5  X  100  .   r^    ^^ 

-2—2 ^ =  per  cent.  Fe,Oj. 

(2)  Crucible  +  Fe,0, 10.745  grams. 

Crucible 10.721 

Fe,0,  ^    0.024 

0.024     X     2.5     X      100  .       T>        /^ 

=  3.00  percent.  FCjO.,. 




Weight  of  AljO,  X  2.5  X  100  , ,  ^ 
5 !1_! 2 —  per  cent.  AljO,. 

(2)  Crucible  +  A1,0, 10.762  grams. 

Crucible 10.721 



AljOj  =    0.041 

0.041  X  2.5  X  100  .     4 ,  /^ 

..-    ._  .y_      —  =  5.12  per  cent.  AljOs. 

5.12  +  =  6.22  per  cent.  AljOj. 

Weight  of  CaO  X  2.5  x  100  .  r^  r^ 

^  per  cent.  CaO. 


(5)  Crucible  +  CaO....* 11.2223  grams. 

Crucible 10.7210      *' 

CaO  =    0.5013 

0.5013  X   2.50  X    100  ,      ,  ^     r^    rx 

—^ — ^ =  62.67  per  cent.  CaO. 

t  i 


(7)  Crucible  +  MgO 10.725  grams. 

Crucible 10.721 

MgO  =    0.004 

<  t 




BaSO*  =>8 
SOj  =    0.0027    " 

0.0027  X  5  X  100         ^r-^,,^„.    on 
=  0.07  per  cent.  SOj. 


SiO, 24.20  per  cent. 

AljO, 6.22 

FejOs 3.00 

CaO 62.67 

MgO 1.22 

KjO  1. 10       ** 

Na,0 0.86 

SO3 0.67 





Platinum  dish ^33'7S50  *' 

=  0.0500  (Total  sulphates)     •  * 

( MgSO^.KjSO^.Na^SO*). 
Crucible  +  Mg^PjOT  =  10.729  grams. 

Crucible  =  10.721       " 

0.008      ** 
Mg2P807=o.oo8gm8.=o.oo8MgS04X  2=    0.0176  (MgS04) 

0.0324  (K,SO44-Na^04) 
KjPtCle  =  0.0232  =  0.0082  KjSO*  X  a  =       0.0164  l^^SOj 

0.0160  (Na^SOi) 
0.0176  MgSOf  =  0.0058  MgO  and  is  added  to  MgO  in  (7) 

MgO  from   (7) 0.004 

MgO  from  (8) 0.0058 


0.0098  X  2.5  X  100  ^^  ^   w  /^ 
=  1.22  per  cent.  MgO. 

0.0164  K,SO,    =  0.0088  K,0  then  9j?^J^..^'S  X  100  ^  j  lopg^cent —  K,0. 

0.0160  Na^SO*  =0.0069  Na,0  then— ?^—  ^-^  ^-i??=  0.86  per  cent.K^«tO. 

(SO3). Crucible  -j-  BaSO^ 10.729  grams. 

Crucible 10.721 

Total    99.94       ** 

The  following  well  known  brands  of  Portland  cements  were 
analyzed  in  my  laboratory  by  above  method. 


Bnrham'8.  Dyckerhoff's.  Saylor's. 

>» 21.70  per  cent.  19.05  per  cent.  21.25  per  cent. 

O, 6.82        **  7.90        "  4.21 

tO, 2.37        **  4.48        **  8.25 

D 62.26        *'  63.62        "  61.25 

lO  1.48        ••  1.87        "  1.50 

3  1.84        "  0.88        •*  i.oi 

aO 0.98         **  0.36         "  0.99 

»     1.20         **  0.94         '*  1.38 

k..        -.Tin  **  **  '< 


•    • 

99.95  **  100.00  '*  99.84  '* 

In  some  cements  quartz  is  a  constituent  in  amounts  varying 

>iii  five-tenths  to  six  per  cent.     It  can  be  separated  from  com- 

tied  silica  by  the  method  of  Fresenius.* 

W'here  carbonic  acid  has  been  indicated  by  the  qualitative 

alysis  the  quantitative  analysis,  for  this  constituent,  should  be 

%d.e  upon  at  least  eight  grams  qi  the  cement. 

The  carbonic  acid  rarely  reaches  one  per  cent.,  and  while  it 

generally  absent  in  well  burned  cements,  it  is  by  no  means  an 

oommou  constituent  to  the  amount  of  0.15-0.30  per  cent.,  as 

-    following  table  of  analyses  of  German  cements  will  show  :  * 


^^ 61.99  62.89    63-71  63.27  65.59  5996  64.51    60.81 

^^ 23.69  22.80    25.37  19.80  22.85  23.70  22.38    22.63 

«^3» 2.71  3.40      3.14  3.22  2.76  3.15  2.24      2.42 

1 8.29  7.70      4.31  6.73  5.51  8.20  9.45      7.06 

0.47  1.20      1.25  2.02  1.24  i.oo  ...       2.89 

'^alics 0.95  1.30      0.84  1.48  0.92  r.05  ...       2.83 

*^  0.69  0.71      0.87  r.o8  1.69  0.88  1.44      0.47 

•^ 0.27  •••         ...  0.23  ...  0.26  ...       0.33 

Soluble 0.44        1.38  ...  0.80       

The  Mechanical  Testing. 

^he  method  recommended  for   use  in  this  countrj*^  by  the 
"^^erican  Society  of  Civil  Engineers  is  as  follows  : 
^i)  Determination  of  fineness. 

(2)  Liability  to  checking  or  cracking. 

(3)  Tensile  strength. 

^  Quant,  Chem,Anai.,  p.  359. 

^Der  Portland-cement  und  seine  A  nwenduMg fit  im  Bauwesen,  Berlin,  1892.  p.  18. 

206         •  QUANTITATIVE   ANALYSIS. 

Fineness, — Tests  should  be  made  upon  cements  that  have  passed 
through  a  No.  loo  sieve  (10,000  meshes  to  the  square  inch), 
made  of  No.  40  wire,  Stubb's  wire  gauge.  The  finer  the 
cement  the  more  sand  it  will  unite  with  and  the  greater  its 

Liability  to  Checking  or  Cracking, — Make  two  cakes  of  neat 
cement  two  or  three  inches  in  diameter,  about  one-half  itizh 
thick,  with  thin  edges.  Note  the  time  in  minutes  that  these 
cakes,  when  mixed  with  water  to  the  consistency  of  a  stiff, 
plastic  mortar,  take  to  set  hard  enough  to  stand  the  wire  test 
recommended  by  General  Gillmore,  one-twelfth  inch  diameter 
wire  loaded  with  one-fourth  pound,  and  one  twenty-fourth  inch 
diameter  wire  loaded  with  one  pound. 

One  of  these  cakes,  when  hard  enough,  should  be  put  in  water 
and  examined  from  day  to  day  to  see  if  it  becomes  contorted  or 
if  cracks  show  themselves  at  the  edges,  such  contortions  or 
cracks  indicating  that  the  cement  is  .unfit  for  use  at  that  time. 
In  some  cases  the  tendency  to  crack,  if  caused  by  too  much 
lime,  will  disappear  with  age.  The  remaining  crack  should  be 
kept  in  the  air  and  its  color  obser\'ed,  which,  for  a  good  cement, 
should  be  uniform  throughout. 

Tensile  Strength. — One  part  of  the  cement  mixed  with  three 
parts  of  sand'  for  the  seven  days  and  upward  test,  in  addition  to 
the  trials  of  the  neat  cement.  The  proportions  of  cement,  sand 
and  water  should  be  carefully  determined  by  weight,  the  sand 
and  cement  mixed  dry,  and  all  the  water  added  at  once.  The 
mixing  must  be  rapid  and  thorough,  and  the  mortar,  which 
should  be  stiff  and  plastic,  should  be  firmly  pressed  into  the 
molds  with  the  trowel  without  ramming  and  struck  off  level, 
the  molds  in  each  instance,  while  being  charged  and  manipu- 
lated, to  be  laid  directly  on  glass,  slate  or  other  non-absorbent 
material.  The  molding  must  be  completed  before  incipient 
setting  begins.  As  soon  as  the  briquettes  are  hard  enough  to 
bear  it,  they  should  be  taken  from  the  molds  and  kept  covered 
with  a  damp  cloth  until  they  are  immersed.  For  the  sake  of 
uniformity,  the  briquettes,  both  of  neat  cement  and  those  con- 

1  White  crushed  quartz,  which  passes  through  a  No.  ao  sieve,  but  remains  upon  a 
No.  30  sieve,  is  standard. 



taiiiing  sand,  should  be  immersed  in  water  at  the  end  of  twenty- 
lour  hours,  except  in  the  case  of  one  daj'  tests.  Ordinary  clean 
water  having  a  temperature  between  60°  F.  and  70°  F.  should 
be  used  for  the  water  of  mixture  and  immersion  of  sample.  The 
proportion  of  water  required  is  approximately  as  follows  : 

For  briquettes  of  neat  cement,  about  twenty-five  per  cent. 

For  briquettes  of  one  part  cement,  one  part  sand,  about  fifteen 
per  cent,  of  total  weight  of  cement  and  sand. 

For  briquettes  one  part  cement,  three  parts  sand,  about  twelve 
per  cent,  of  total  weight  of  cement  and  sand. 

The  object  is  to  produce  the  plasticity  of  plasterer's  stifi 

An  a\-erage  of  five  briquettes  may  be  made  for  each  test,  only 
those  breaking  at  the  smallest  section  to  be  taken.  The  bri- 
quettes should  always  be  put  in  the  testing  machine  and  broken 
immediately  after  being  taken  out  of  the  water,  and  the  tem- 
perature of  the  briquettes  and  of  the  testing  room  should  be  con- 
stant between  60°  F.  and  70°  F. 

The  following  table  shows  the  average  minimum  and  maxi- 
mum tensile  strength  per  square  inch  which  some  good  cements 
have  attained.  Within  the  limits  given  the  value  of  a  cement 
varies  closely  with  the  tensile  strength  when  tested  with  the  full 
dose  of  sand. 

American  and  Foreign  Portland  Ckmsnts.— Neat. 
Oue  day,  ( i  hour,  or  until  set,  in  air,  the  rest  of  the  24  hours  in  water, ) 
from  100  to  140  pounds  per  square  inch. 

One  week,  (1  day  in  air.  6  days  in  water),  from  350  to  550  pounds  per 

One  month,  28  days,  (1  day  in  air,  37  days  in  water),  from  350  to  70a 
pounds  per  square  inch, 

One  year,  (iday  in  air,  the  remainder  in  water),  (rom  450  to  800 pounds 
per  square  inch. 

American  and  Fobkign  Portland  Cements. — i  Part  of  Cement  to 
3  P.4.RTS  OF  Sand. 

One  week,  t  r  day  iu  air,  6  days  iu  water),  from  80  to  125  pounds  per 
'Square  inch. 

One  mouth,  18  days.  (I  day  in  air,  27  days  in  water),  from  100  to  zoo 
pounds  per  square  inch. 

One  year, (I  day  in  air,  the  remainder  in  water),  from  20010350  pounds' 
per  square  iuch. 

7>Biu,  ^mtr.  Set.  of  Civl  Engmrnt.  Augusl,  11*91,  V-  3»5 


The  machines  for  determining  the  tensile  strength  of  Portland 
cements  in  use  in  this  country  are  the  "Fairbanks,"  Fig,  52. 
the  "  Riehle."  Fig.  53  and  the  Olsen. 

The  Fairbanks  machine  is  automatic  and  is  operated  as  follows ; 

Hang  the  cup  on  the  end  of  the  beam  :  see  that  the  poise  is 
at  the  zero  mark  and  balance  the  beam  by  turning  the  ball. 
Place  the  shot  in  the  hopper.  Place  the  briquette  in  the  claraps 
and  adjust  the  hand  wheel  so  that  the  graduated  beam  will  be 

inclined  upward  about  45".  Open  the  automatic  valve  so  as  to 
allow  the  shot  to  run  slowly.  When  the  specimen  breaks  the 
beam  drops  and  closes  the  valve  through  which  the  shot  has 
been  pouring.  Remove  the  cup  with  the  shot  in  it  and  hang 
the  counterpoise  weight  in  its  place.  Hang  the  cup  on  the  hook 
under  the  large  balance  ball  and  proceed  to  weigh  the  shot, 
using  the  poise  on  the  graduated  beam,  and  the  weights  on  the 
counterpoise  weight.  The  result  will  show  the  number  of 
pounds  required  to  break  the  specimen. 


The  'Riehle,"  while  not  automatic,  is  accurate,  and  responds 
Lto  difterences  as  slight  as  one  pound  in  2,000.      The  distinctive 

(a)  The  poise  moves  quietly  and  smoothly  on  the  weighing 

it)  The  weighing  beam  is  long  and  the  marks  not  too  close 
together.  The  slightest  movement  of  the  beam  is  promptly  and 
plainly  observed  by  the  motion  of  the  indicator. 

(c)  The  levers  are  tested  and  sealed  to  U.  S.  standard 

( d)  The  arrangement  of  the  * '  grips ' '  to  hold  the  briquette  is 
such  that  they  are  always  swung  from  pins,  thus  giving  the  test 

"upon  the  cement  when  the  briquette  is  on  a  dead  straight  line. 

Directions  for  Tesling  Portland  Cement  According  to  the  Official 

German  J!ules.' — The  quality  of  a  mortar   made   with   cement 

depends  not  only  on  the  strength  of  the  cement  itself,   but  also 

1  Pertla^d  C/mrHl,  hy  GuslBV  C.  IB  will. 


on  the  degree  of  sub-dh'ision  of  the  same.  It  is  therefore 
sary  to  make  the  tests  both  with  neat  cement  and  with  a  mixture 
of  the  same  with  "standard  sand."  This  tatter  as  used  at  the 
Royal  Testing  Station  at  Berlin,  is  produced  by  washing  and 
drying  quartz  sand,  which  must  be  clean  as  possible,  and  after- 
wards be  silted  through  a  sieve  of  sixty  meshes  per  square  cen- 
timeter (587  meshes  per  square  inch),  by  which  process  the 
coarsest  particles  are  separated.  The  sand  is  again  sifted 
through  a  sieve  having  120  meshes  to  the  square  centimeter 
{774  meshes  per  square  inch).  The  residue  remaining  in  this 
sieve  is  the  standard  sand  for  experiments,  the  coarsest  and 
finest  panicles  having  been  eliminated.  It  is  absolutely  neces- 
sary in  order  to  obtain  uniform  results  to  use  only  the  "standard 
sand,"  as  the  size  of  the  grain  has  a  material  infliience  on  the 
results  of  the  testing.  The  sand  must  be  clean  and  dr>',  and  all 
earthy  and  other  substances  previouslj'  removed  by  washing. 

Preparation  of  Briquettes  of  Neat  Portland  Cement. — Upon  s 
slab  of  metal  or  marble  are  laid  five  sheets  of  filtering  paper, 
which  have  been  previously  saturated  with  water,  and  upon 
se  are  placed  five  brass  molds  ( Fig.  54 )  thoroughly 

cleaned  and  moistened  with  water.  One  thousand 
Lffls  of  cement  and   250  grams  of  water  must  be 

ihoroughly  mixed,  well  worked  up,  and  when  the 
ulting  mass  has  been  rendered  perfectly  homogen- 
is,  it  is  poured  into  the  molds.     The  lattermust  be 

gently  tapped  by  means  of  a  wooden  hammer  with 
Figs*.  equal  force  on  both  sides  during  ten  to  fifteen  minutes 
to  insure  the  escape  of  confined  globules  of  air.  The  molds 
must  be  carefully  filled  up  until  the  mass  becomes  plastic,  the 
.superfluous  mortar  is  then  struck  off,  and  the  mold  carefully 
withdrawn.  The  samples,  after  remaining  twenty-four  hours 
exposed  to  the  air,  at  a  temperature  of  about  60°  F.,  must  be 
immersed  in  water  having  the  same  temperature,  and  care  must 
be  taken  that  tliey  remain  covered  with  water  until  the  time  ar- 
rives for  breaking  them.  In  order  to  obtain  a  proper  average  at 
least  ten  briquettes  should  be  prepared  for  every  examination. 
Preparation  of  Briquettes  from  a  Mixture  of  Portland  Cement 
and  Standard  Sand. — Place  the  molds  on  metal  as  described 




preparation  of  neat  cement  briquettes.  The  quantities  { by  weight) 
specified  of  cement  and  sand  are  thoroughly  mixed  and  to  this 

is  added  the   requisite  quantit\    of  water.     The   whole  mass  is 
then  worked  up  with  a  trowel  or  spatula  uutil  it  becomes  uni- 


form.  In  this  manner  is  obtained  a  very  stiff  mortar. 
molds  are  filled  and  mortar  heaped  up.  The  latter  is  then 
beaten  into  the  molds  with  an  iron  trowel,  at  first  lightly,  and 
afterwards  more  heavily,  until  it  becomes  elastic  and  water  ap- 
pears on  the  surface.  The  superfluous  morter  is  then  scraped 
off  with  a  knife  and  by  means  of  the  same  the  surface  is  leveled. 
The  further  treatment  of  these  briquettes  is  the  same  as  for  neat 

cement  briquettes.     The  average  of  ten  breaking  weights  fui^l 
nishes  the  strength  of  the  mortar  tested. 

The  machine  in  general  use  in  Germany  for  determining  the  J 
tensile  strength  of  cements  is  the  Michaelis  {Fig.  55).  and  from  I 
this  is  derived,   with   modifications,  the  "Reid  and  Baileyj' 



machine  in  use  in  England,  and  the   "Fairbanks"  previously 

No  standard  specifications  for  the  testing  of  Portland  cement 
are  required  in  Great  Britain,  the  determination  of  fineness,  ten- 
sile strength  and  variations  in  volume,  being  considered  suffi- 
cient to  determine  the  value  of  a  cement.  The  machines  forten- 
sile  strength  are  the  "Faija."  (Fig.  56).  the  "Reid  and  Bailey,"' 
(Figs. 57 and 58). or  the  "Grant,"  similar  to  the  Riehle,  and  de- 

scribed in  Proceedings  of  the  Institution  of  Civil  Engineers,  6a, 
113.  The  "  Reid  and  Bailey  "  is  essentially  the  "  Michaelis" 
(Fig-  55},  excepting  that  water  is  used  instead  of  fine  shot  for 
the  breaking  power. 

It  is  readily  seeu  that  the  "  Faija  "  and  "Grant  "  machines, 
not  being  automatic,  require  the  application  of  the  power  at  a 
certain  uniform  speed  to  obtain  comparable  results,  since  a  dif- 
ference of  twenty-five  per  cent,  of  tensile  strength  may  be  ob- 
tained by  applying  the  strain  very  quickly  or  very  slowly.' 


Faija  has  determined  this  variation  with  e»treme  care,  the  re- 
sults being  indicated  in  the  curve  shown  in  Fig.  59.  To  over- 
come these  variations  a  uniform  speed  of  400  pounds  per  min- 
ute has  been  accepted  as  the  standard. 

Not  only  are  comparable  methods  required  in  the  use  of  the 
machines  to  obtain  uniform  results,  but  the  briquettes  must  also 
be  constructed  under  similar  conditions. 

It  is  manifestly  unjust  to  compare  the  tensile  strength  of  two 
cements  (even  when  the  briquettes  are  broken  upon  the  same 
machine)  unless  the  briquettes  have  the  same  weight  of  water 
for  mixing ;  the  same  pressure  with  the  trowel  when  being 
formed  in  the  molds,  and  the  same  length  of  time  of  exposure 
under  water  before  submitting  the  briquettes  to  the  tensile 

For  instance  :  Comparing  tests  made  upon  the  DyckerhofP 
Portland  cement  by  Dr.  Bohme,  Director  of  the  Royal  Commis- 
sion for  testing  building  material,  at  Berlin,  and  by  E.  J. 
DeSmedt,  General  Inspector  Engineering  Department,  District  of 
Columbia,  we  find  that  the  German  method  gives  a  much  higher 
tensile  strength  than  the  method  in  use  in  this  country. 

Dr.  Bohmb. 

Average  tensile  strength 
Age  of  briquettes.  per  square  inch.  Number  of  tests. 

7  days 767  pounds  10 

28     **     895        "  10 

E.  J.  DeSmedt,  C.  E. 

Average  tensile  strength 
Age  of  brifjuettes.  per  square  inch. 

5  days 250  pounds. 

30    ••     700        " 

Showing  : 

109  pounds  increase  per  day  (7  days),  Dr.  Bohme. 
59      "  "  **      "   (5  days),  DeSmedt. 

or  over  100  per  cent,  difference  upon  the  same  cement  on   the 
seven  days  test. 

These  variations  are  undoubtedly  due  principally  to  the  dif- 
ferent pressures  upon  the  cement  during  the  making  of  the  bri- 
quettes, and  to  overcome  difficulties  of  this  nature  the  Vereins 
deutsche  Portland  Cement  Fabrikanten  have  modified  the  rules 



ill  the  coustructioii  of  the  briquettes  so  that  two  methods  are  a 
ceptable  : 

First,  tile  iiomial  method,  above  given,  with  the  trowel,  etd 
(■'  Haiidarbeit  ".) 

Second,  the  use  of  the   Bohme- Hammer  apparatus  or  "  n 
chine  method,"  by  which  the  cement  in  briquette  form   (aft 



mixing  with  proper  amount  of  water) ,  is  submitted  to  a  pres- 
sure of  150  blows  from  a  hammer  weighing  two  kilos  (Fig.  60). 
The  briquette  of  cement  is  then  removed  from  the  mold  and 
treated  for  tensile  strength  as  usual. 

This  subject  is  receiving  considerable  attention  at  the  pres- 
ent moment,  the  evident  purpose  being  to  render  the  tests  of 
tensile  strength  as  uniform  as  possible  by  making  the  working 
of  the  apparatus  automatic  and  the  production  of  cement  bri- 
quettes with  the  least  possible  variation  in  the  pressure  in  the 

In  this  case,  no  matter  how  careful  the  experimenter  may  be, 

the  "personal  equation  "  enters  largely  into  the  results  of  test- 
ing hand-made  briquettes,  for  which  reason  the  manufacture  of 
the  briquettes  should  be  as  automatic  as  possible.  In  no  other 
way  can  results  obtained  by  different  experimenters  be  compared. 

Prof.  Charles  D.  Jameson  describes  an  apparatus  for  this  pur- 
pose (Figs.  6[  and  62).' 

The  method  of  operating  is  as  follows  1  The  lever  being  raised 
so  that  the  lower  end  of  the  piston  or  main  plunger  is  above  the 
hole  in  the  side  of  the  cylinder  communicating  with  the  hopper, 
cement  is  put  in  the  hopper  and  pushed  down  into  the  cylinder. 
The  molding  plate  is  pushed  against  one  of  the  stops,   so  that 

le  American  Soei«ly  of  Civil  Enttinter.H,  IJ,  301. 

2  3?- 

5     «-    ^H 












Is      2      al 

ill  i 

Ij  II 









li  II  II 

II 11  n 






Ill  1 

9     i      i 

1 1 1 





»  i  || 

i  1  I 









:!    ?    s 

?  s  ; 



^        ^        :      }i      :        :        :        :        : 




§ft     5       R 

E     I     S 




SJ    3     S 

4     •     i 





js    5    J 

S    «     S 






!3    »     1 

E     S     S 


J ! ' 



~?    *     ? 

=     8     S 





1     '    1 

3      ^     s 
^     >     > 

.>...:»  .sS  lilt     I 



one  ol  the  openings  is  beneath  the  bore  of  the  cylindeT.  The 
long  lever  is  forced  down,  causing  the  plunger  to  force  the  ce- 
ment into  the  opening  of  the  molding  plate.  After  this,  the 
molding  plate  is  swung  against  the  other  stop,  cutting  off  the 
briquette,  placing  it  over  the  plungers,  throwing  the  other 
opening  in  the  molding  plate  directly  beneath  the  cylinder.  The 
smaller  lever  is  lifted,  raising  the  plunger,  and  forces  the  bri- 
quette out  of  the  mold,  after  which  it  is  removed.  The  plunger 
is  then  pressed  down,  the  main  lever  also,  the  molding  plate 
swung  back  to  the  first  position,  the  other  plunger  lever  lifted, 
and  another  briquette   is  ready  to   be  taken   away,    and  so  on. 

^P After  making  three  hriijuettes,  the  main  lever  is  lifted  and  more 
Cement  placed  in  the  cylinder.  The  machine  is  best  operated 
by  two  men.  one  to  feed  and  operate  the  long  lever,  and  the 
other  to  swing  the  molding  plate,  remove  the  briquettes  and 
lower  the  plungers.  The  pressure  on  the  briquette  is  1 75  pounds 
per  square  inch. 

The  conditions  required  in  France  for  a  good  cement  are :  ' 
First.  Analysis  to  determine  the  chemical  composition. 
Second.  The  determination  of  density. 
Third.  The  determination  of  fineness. 

1  E-  Canilloi,  Ommli  /I  C*a»i  HyJ^aHltqun.  paria.  Baudry  &  Co.    ifi9i. 


Fourth.  The  determtnatioti  of  tensile  strength. 

Fifth.  The  determination  of  crushing  strength. 

Sixth.     The  determination  of  variations  in  volume. 

The  tensile  strength  is  determined  by  the  use  of  a  Mtcbaelis 
machine,  Fig,  63,  or  the  use  of  a  Buignet  apparatus,  Fig.  64, 
this  latter  being  upon  an  entirely  different  principle  than  any 
yet  in  use,  and  is  thus  described  by  the  designer,  M.  Buignet, 
Conductor  des  ponts  et  chauss^es  au  Havre : 

It  is  composed  of  a  basin  A  and  frame  B. 

The  basin  A,  filled  with  mercury  and  water,  closes  up  by  a 

Fig.  64. 

diaphragm  of  rubber  covered  with  a  metallic  disk,  and  is  in 
direct  communication  with ; 

(a)  Manometrique  tube  D. 

{b)  With  a  movable  resen-oir  R,  filled  with  merpury,  by 
means  of  a  thick  rubber  tube  T. 

The  grips  G  G,  in  which  are  to  be  placed  the  briquettes  to  be 
tested,  are  fastened,  one  to  the  frame  B  by  the  support  V,  the 
other  to  the  support  A/,   which  rests  upon  the  center  of  the 




metallic  disk  over  A.  It  is  operated  as  follows :  The  briquettes 
are  placed  in  the  grips  G  G,  and  the  support  V  moved  up  or 
down  until  equipoise  is  established,  and  then  firmly  secured  by 
a  crank  in  irame  B. 

The  support  J/is  adjusted  until  the  point  at  its  lower  end  just 
touches  the  metallic  disk  in  A. 

By  gradually  lowering  the  reser\'oir  /f  an  upward  pressure  is 
given  to  the  metallic  disk  in  A,  which  is  transferred  to  the  sup- 
port .1/,  until  when  sufficient  pressure  is  exerted  the  briquette  is 
broken.  The  moment  rupture  of  the  briquette  takes  place,  the 
pressure  required  to  do  this  is  indicated  by  the  float  /  in  the 
manometer  tube  D. 

By  a  comparison  of  the  various  machines  used  in  Germany, 
England,  France,  and  the  United  States,  we  find  practically 
but  two  in  general  use  :  the  "  Michaelis"  and  the  "  Grant." 
While  nearly  all  engineers  require  cements  to  be  subjected  to 
the  tensile  strength  test,  in  fact  relying  more  upon  this  one  test 
than  any  of  the  others,  it  might  be  well  to  include  here  the 
Opinion  of  H,  I,e  Chatelier,  professor  at  the  5co!e  des  Mines, 
Paris,  France,  as  given  in  a  paper  presented  at  the  last  meeting 
of  the  Americau  In.stitule  of  Mining  Engineers,  August,  1893, 
entitled  "Tests  of  Hydraulic  Materials,"  p,  44. 

"The  method  of  tension  is  at  present  most  widely  used,  but 
the  preference  for  it  is  not  well  founded.  Here,  as  in  rupture 
by  bending,  only  the  surface  o(  the  briquettes  acts  in  a  really 
useful  way,  and  its  ine\ntable  irregularities  and  alterations  so 
greatly  affect  the  precision  of  the  results  that  they  can  in  no 
case  be  trusted  nearer  than  about  twenty  per  cent. 

"This  preponderant  influence  of  the  superficial  parts  was 
first  shown  by  the  fact  that  the  resistance  of  briquettes  of  differ- 
ent sizes  increases,  not  with  the  section,  but,  on  the  contrary-, 
with  the  perimeter.  Finally,  M.  Durand-Claye  has  shown  that 
the  interior  of  a  briquette  may  be  removed  without  notably 
diminishing  its  resistance  to  rupture  by  tension,  and  has  given  a 
complete  theoretical  explanation  of  the  phenomena  which  seemed 
at  first  sight  paradoxical." 


The  Cmshing  Test. 

This  test  is  not  official  in  this  country  and  is  seldom  required 
by  our  engineers,  who,  however,  have  confined  their  experiments 
in  this  direction  mainly  to  crushing  tests  of  concrete,  formed  by 
mixing  Portland  cement,  sand,  and  broken  stone. 

Tests  upon  cubes  of  neat  cement  and  of  mortar  composed  of 
one  part  cement  and  three  of  standard  sand,  are  generally  in- 
cluded in  reports  given  upon  the  examination  of  cements  in 
Europe,  the  ratio  being  that  the  crushing  strength  is  about  ten 
times  greater  than  the  tensile  strength. 

Thus,  a  cement  of  good  quality  should  show  the  following 
resistance  per  square  centimeter  : 

Tensile  Strength. 

7  days.  28  days. 

Neat  cement 25  kilos.  35  kilos. 

I  partcement)    ^^    ..  ^g     ,. 

3  parts  sand    J 


Crushing  Strength. 

7  days.  28  days. 

Neat  cement 250  kilos.  350  kilos, 

ipartceraent) ^^    ..  ^go    " 

3  parts  sand    > 

To  convert  kilos  per  square  centimeter  to  pounds  per  square  inch,  the  equivalents 
used  are  :  one  kilo  =  2.204  pounds  English;  6.451  square  centimeters  =  one  square 
inch,  English. 

The  hydraulic  presses  made  use  of  for  this  purpose,  a  few 
years  since,  gave  very  discordant  results,  as  it  was  impossible  to 
distribute  the  pressure  evenly  over  the  surface  of  the  cubes. 
This  has  been  overcome,  and  there  are  now  several  machines 
upon  the  market  whose  results  are  comparable,  viz,  : 

The  "Suchier,"  Fig.  65,  the  *'B6hme,''  Fig.  66,  the  *'  Tet- 
majer,"  as  improved  by  Prof.  Amsler-Laffon,*  the  **  Brink  and 
Hubner,'"  the  **Riehle,"  the  '*  Fairbanks,*'  the  **01sen"  and 
the  **  Bailey." 

l^ariation  in  volume  (expansion  or  contraction) — The  method 
of  Faija,"  the  one  generally  used  for  this  purpose,  is  as  follows  • 

1  Consult :  Schiceizer  Bauzeit,  January  12.  1889. 

a  Description  of  the  "  Suchier,"  "  Bohme."  and  •*  Brink  and  Huber  "  machines  will 
be  found  in  Det  I*ortland  Cfment  und  seinf  An-vendungen  im  Bauwtien^  Berlin,  x8q2. 

8  The  determination  of  liability  to  *  checking  "  or  "  cracking  "  (variation  in  >'ol- 
ume)  in  Portland  cements  as  recommended  by  Amet-ican  Society  Civil  Engineers^  is  not 
as  complete  as  Faija's  method.    See/.  Am.  Chfm.  Soc.,  15. 184. 


2  23 

Three  pats  should  be  made  on  pieces  of  glass  or  other  non- 
porous  substance,  and  their  behavior  watched  under  the  follow- 
ing conditions : 

Pat  No.  I  may  be  left  in  the  air,  and  No.  2  should  be  put  in 

water  as  soon  as  it  is  set  hard. 

Pat  No.  3  should  be  treated  in  the  apparatus  for  determining 
the  soundness  of  cement.     The  apparatus  consists  of  a  covered 



vessel  in  which  water  is  maintained  at  an  even  temperature  of 
no"  C;  the  space  above  the  water  is  therefore  filled  with  the 
vapor  rising  therefrom,  and  is  at  a  temperature  of  about  ioo°  C. 
Immediately  the  pat  is  gauged,  it  should  be  placed  on  a  rack  in 
the  upper  part  of  the  vessel,  and  in  five  or  six  hours  it  may  be 
placed  in  the  warm  water  and  left  therein  for  nineteen  or  twenty 
hours.     If,  at  the  end  of  that  period,  the  pat  is  still  fast  to  the 

glass  and  shows  no  signs  of  blowing,  the  cement  may  be  consid- 
ered perfectly  sound  ;  should,  however,  any  signs  of  blowing 
appear,  the  cement  should  be  laid  out  in  a  thin  layer  for  a  day 
or  two,  and  a  second  pat  made  and  treated  in  the  same  manner, 
as  the  blowing  tendency  may  only  be  due  to  the  extreme  new- 
ness of  the  cement. 

If  pat  No.  3  shows  the  cement  to  be  unsound,  pats  Nos,  i  and 
2  will  eventually  prove  it,  but  it  may  be  weeks  or  even  months 
before  they  develop  the  characteristics.  If  pat  No.  2  blows,  it 
may  be  .because  it  was  put  in  the  water  before  it  was  set.  A 
cement  is  considered  set  hard  when  it  can  no  longer  be  marked 
by  the  pressure  of  the  thumb  nail. 



An  Automatic  Cement  Testing  Machine. 

To  promote  coiiveuience  and  rapidity  and  secure  uniformity, 
regularity,  and  a  standard  method  of  work  as  free  as  possible 
from  the  irregularities  coming  under  the  head  of  "persona! 
equation,"  Prof.  J.  M.  Porter  has  devised  the  adjustable  auto- 
matic loading  and  balancing  attachments  which  are  illustrated 
by  the  accompanying  elevation  and  details  of  the  special  mechan- 
ism added  to  the  2,000-pound  Olsen  machine  o!  standard  pat- 
tern. Fig.  67.  The  load  is  applied  by  filling  with  water  a 
tank  suspended  from  the  long  arm  of  a  15  to  r  lever,  the  con- 
nection of  which  has  a  pin  bearing  on  a  cylindrical  surface 
which  rests  on  the  adjustment  screw  of  the  lower  grip.  Neither 
the  tank  nor  its  contents  are  weighed,  but  the  exact  rate  of 
loading  per  minute  is  accurately  known  from  previous  tests. 
Water  is  admitted  to  the  tank  from  a  large  reservoir  on  the  roof, 
where  a  practically  constant  height  of  surface  level  is  maintained, 
so  thar  there  is  no  sensible  variation  of  pressure  in  the  stream 
admitted  through  a  carefully  fitted  gate  valve.  The  position 
of  this  valve  at  "open,"  "closed."  and  alHntermediate  points  is 
shown  by  an  index  attached  to  the  stem  and  registering  on  a 
dial  marked  off  with  the  number  of  pounds  per  minute  applied  to 
specimen  as  determined  and  verified  by  previous  experiments. 

When  the  specimen  breaks,  the  load  lever  drops  and  permits 
the  load  tank  to  fall  a  few  inches,  so  that  the  chain  is  brought 
into  tension  and  arrests  the  descent  of  the  valve  before  its  seat 
stops  descending.  Thus  the  bottom  of  the  tank  is  opened  and 
the  contents  quickly  escape  into  the  hopper  of  the  receiving  case, 
and  are  carried  off  through  the  waste  t«  the  sewer.  The  actual 
load  can  be  applied  at  from  zero  to  eighty  pounds  per  minute, 
thus  giving  an  increase  strain  of  zero  to  1.200  pounds  per  min- 
ute on  the  specimen,  A  small  electric  motor  is  belt -connected 
to  the  pulley  that  continuously  drives  a  friction  disk  and  its 
eugaged  wheel.  The  wheel  is  feathered  to  a  sleeve  that 
runs  loose  on  its  shaft,  and  carries  a  coned  clutch  that  is 
nominally  disengaged  from  its  cone,  which  is  feathered  to 
shaft,  and  can  be  moved  slightly  longitudinally  on  the  shaft 
into  contact  with  the  wheel  by  the  action  of  a  lever. 





When  the  scale  beam  rises,  it  makes  a  contact  which  com- 
pletes the  electric  circuit  and  sends  a  current  through  the  elec- 
tromagnet and  causes  it  to  attract  its  armature  (here  shown  not 
in  contact),  which  moves  to  the  right  about  a  pivot  a  sufficient 
distance  to  make  the  friction  clutch  with  the  coned  wheel  and 
drive  shaft.  This  shaft  in  turn  operates  the  sprocket  wheel  and 
chain,  which  draw  the  weight  out  011  the  scale  beam  until  the 
latter  falls,  and  breaking  the  electric  circuit,  releases  the  arma- 
ture and  allows  the  friction  clutch  to  disengage.  By  turning 
the  capstan-head  nut,  the  friction  wheel  is  set  at  a  greater  or  less 
distance  from  the  center  of  the  disk,  and  the  chain  is  overhauled 
faster  or  slower  accordingly.  The  arrangement  was  constructed 
in  the  college  laboratory  and  is  positive  and  simple.  It  does 
not  get  out  of  order  and  is  considered  accurate  and  satisfactory, 
and  to  enable  more  rapid  and  better  comparable  tests  to  be  made 
more  than  twenty  specimens  per  hour  have  been  broken  by  its 

Jiesumf:  The  determination  of  the  value  of  Portland  cement 
therefore  requires  the  following  tests  : 

First.  Chemical  analysis. 

Second.  Determination  of  fineness. 

Third.  Determination  of  tensile  strength,  including  the  use  of 
automatic  briquette  machines  as  well  as  an  apparatus  for  mix- 
ing the  cement  with  water,  as  "  Faija"  mixing  machine. 

Fourth.  Determination  of  crushing  strength. 

Fifth.  Determination  of  variation  of  volume. 

References;  J.  Win.  Chein.  Soc.  16,  381-386,  contains  an  index,  ar. 
xsDged  by  the  writer,  of  the  literature  relating  to  Portland  ceaieut.  from 
J870  to  1893. 


Determination  of  Nickel. 

The  principles  involved  in  the  processes  are  the  following  :' 

Firsi.the   iron   is  precipitated  as  ferric   phosphate  in  cold, 

strong  acetic  acid  solution,  under  which  condition  it  precipitates 

perfectly  free  from  nickel,  although  retaining  a  small  amount  of 


I  E.  D.  CamplieLl  i  /.  Am  Ckrm.  Sat.,  17,  ,,y 



Second.  The  copper  is  separated  from  manganese  and  nickel 
in  hydrochloric  acid  solution  by  means  of  granulated  lead. 

Third.  The  manganese  and  lead,  which  displaced  the  copper, 
are  separated  from  the  nickel  by  means  of  cold  ammoniacal  solu- 
tion of  sodium  phosphate. 

Fourth.  The  nickel  is  determined  in  the  ammoniacal  filtrate 
from  the  phosphate  of  manganese  and  lead,  by  titration  with 
standard  potassium  cyanide,  or  by  electrolytic  deposition. 

In  case  the  nickel  is  accompanied  by  cobalt  the  latter  metal 
remains  wuth  the  nickel  and  may  be  separated  from  it  by  any  of 
the  well-known  methods  after  dissolving  off  the  electrolytically 
deposited  nickel. 

The  two  methods  described  below  are  identical  up  to  the 
point  where  a  portion  of  the  filtrate  from  the  phosphates  of  man- 
ganese and  lead  is  taken.  The  description  of  that  part  of  the 
methods  common  to  both  will  be  first  given,  and  then  the  two 
ways  of  treating  the  above  filtrate  for  the  final  determination  of 
nickel  will  be  added. 

Take  2.2222  grams  of  nickel-steel,  place  in  a  500  cc.  gradu- 
ated flask,  add  twenty  cc.  nitric  acid,  sp.  gr.  1.20,  and  five  cc. 
hydrochloric  acid,  sp.  gr.  1.2 1.  Boil  until  the  solution  is  clear, 
which  will  usually  require  not  more  than  from  five  to  ten  min- 
utes. Remove  from  the  plate  and  add  155  cc.  sodium  phos- 
phate solution.  If  a  slight  precipitate  should  form  which  does 
not  dissolve  upon  shaking,  add  carefully  a  few  drops  of  hydro- 
chloric acid  until  the  solution  clears  up.  Add  twenty-five  cc. 
acetic  acid,  sp.  gr.  1.04,  then  100  cc.  sodium  acetate  solution, 
shake,  dilute  with  water  to  502.5  cc,  shake  again,  and  allow  to 
stand  fifteen  minutes.  Filter  through  a  dry  twenty-five  cc.  fil- 
ter,  catching  the  filtrate  in  a  dry  beaker.  As  soon  as  enough  of 
the  filtrate  has  run  through,  which  requires  about  ten  minutes, 
draw  off  with  a  pipette  250  cc.  of  the  filtrate,  transferring  to  a 
No.  4  beaker.  This  will  give  one-half  of  the  solution,  since  it 
was  found  by  experiment  that  the  ferric  phosphate  from  the 
amount  of  steel  taken  occupies  two  and  a  half  cc.  Bring  the 
solution  to  a  boil  and  add  twenty  grams  potassium  hydroxide 
previously  dissolved  in  forty  cc.  of  water.  Boil  five  minutes, 
then  keep  just  below  boiling-point  until  the  precipitate  has  set- 



tied  and  the  solution  Js  clear.  This  precipitates  copper,  mau- 
ganese,  and  nickel  so  completely  that  the  filtrate  gives  no  color 
with  hydrogen  sulphide.  Filter  through  asbestos,  using  a 
pump,  decanting  as  much  of  the  solution  as  possible  before  al- 
lowing the  precipitate  to  get  upon  the  filter.  Wash  with  water. 
Dissolve  the  precipitate  on  the  filter  in  a  hot  solution  of  six  cc. 
strong  hydrochloric  acid  with  an  equal  volume  of  water.  Wash 
the  filter,  using  only  as  much  water  as  is  necessary.  To  the 
solution  in  the  flask,  which  should  not  exceed  fifty  cc.  and 
should  have  a  temperature  of  40"  to  50°  C,  add  fifteen  grams  of 
granulated  lead  and  agitate  at  frequent  inter\-als  for  five  or  ten 
minntes.  This  will  completely  precipitate  the  copper,  a  small 
amount  of  lead  going  into  solution.  Filter  through  a  small 
glass  wool  filter,  catching  the  filtrate  in  a  No.  z  beaker ; 
wash  the  granulated  lead  with  a  small  amount  of  water  and  boil 
the  solution  down  until  it  does  not  exceed  sixty  cc.  Add  ten 
cc,  of  sodium  phosphate  solution,  then  ammonium  hydroxide 
until  a  precipitate  begins  to  form,  then  hydrochloric  acid  suffi- 
cient to  clear  the  solution,  cool  until  cold,  and  transfer  to  a  cylin- 
der or  flask  graduated  to  1 1 1 .  i  cc.  Add  five  cc.  strong  ammo- 
nium hydroxide,  sp.  gr.  0.90,  dilute  to  the  mark,  shake  well, 
and  allow  to  stand  fifteen  minutes.  Filter  through  a  dry  nine 
cm.  filter,  receiving  the  filtrate  into  a  dry  beaker.  Draw  off,  by 
means  of  a  pipette,  100  cc.  of  the  filtrate,  which  is  equivalent  to 
one  gram  of  the  original  steel,  and  treat  by  one  of  the  two  fol- 
lowing methods : 

Ekdrolylk  Method. 
Transfer  the  100  cc.  of  filtrate,  above  mentioned,  to  a  large 
platinum  dish  having  a  capacity  of  alwut  200  cc.  Add  twenty- 
five  cc  of  strong  ammonium  hydroxide,  sp.  gr.  0.90,  and  dilute 
to  175  cc.  Electrolyze  for  at  least  four  hours,  using  a  current 
)'ielding  four  cc.  of  electrolytic  gas  per  minute.  This  strength 
o(  current  can  be  easily  obtained  by  connecting  three  medium- 
sized  cells.  The  end  of  the  precipitation  of  the  nickel  is  indi- 
cated when  a  drop  of  the  solution  placed  in  contact  with  a  drop 
of  ammonium  sulphide  gives  no  color  due  to  nickel  sulphide. 
Wheu  the  nickel  is  completely  precipitated,  disconnect  the  bat- 


ter\',  wash  the  nickel  thoroughly  with  water,  then  finally  twice 
with  alcohol,  and,  after  draining  off  as  much  as  possible,  heat 
for  a  lew  minutes  in  an  air  bath  at  iio^C.  Cool  and  weigh. 
After  getting  the  combined  weights  of  the  platinum  dish  and 
nickel,  dissolve  off  the  latter  by  warming  with  five  to  six  cc.  of 
nitric  acid  (sp.  gr.,  then  wash  the  platinum  dish  by  means 
of  water  and  alcohol,  and  dry  and  weigh  as  before.  The  differ- 
ence in  the  two  weighings  gives  the  nickel.  It  is  more  satis- 
factory to  weigh  the  empty  dish  after  the  precipitated  nickel  has 
been  dissolved  off  than  before  electrolysis,  since  in  this  way  a 
shorter  time  will  elapse  between  the  two  weighings  and  conse^^^J 
quently  less  error  will  be  introduced  from  variations  in  atmo»-^H 
pheric  conditions. 

Volume Iric  Method. 

Take  lOO  cc.  of  the  filtrate  from  the  phosphate  of  manganese 
and  lead,  add  hydrochloric  acid  very  carefully  until  the  blue 
color  of  the  double  ammonium  nickel  chloride  disappears,  then 
add  ammonium  hydroxide,  drop  by  drop,  until  the  blue  just  re- 
appears, add  an  excess  not  exceeding  one  cc,  Dilute  to  zoo  cc. , 
add  five  cc.  of  cupric  ferrocyanide  indicator,  and  run  in  standard 
potassium  cyanide  until  the  solution  turns  from  the  purple  color 
of  the  indicator  to  a  perfectly  clear  light  straw-yellow.  Sub- 
tract from  the  number  of  cubic  centimeters  of  potassium  cyanide 
used,  the  correction  for  the  indicator.  The  difference  gives  the 
amount  necessary  to  convert  the  nickel  into  the  double  cyanide 
of  potassium  and  nickel.  Multiplying  this  by  the  factor  of  the 
potassium  cyanide,  expressed  in  metallic  nickel,  gives  the 
amount  of  nickel  in  one  gram  of  the  original  sample. 
Special  Apparatus  and  Reagents. 

Five  hundred  cc.  graduated  flask  with  an  additional  mark  1 
502.5  cc. ;  250  cc.  drop  pipette;    100  cc.  drop  pipette;  glassi 
stoppered  cylinder  or  flask  graduated  to  iii.i  cc.     The  gradii«J 
ated  apparatus  should  be  carefully  calibrated  and  compared  be-B 
fore  using. 

Sodium  phosphate  solution,  made  by  dissolving  200  grams  of 
the  ordinary  crystallized  disodium  hydrogen  phosphate  in  1S60 
cc.  of  water.     Ten  cc.  of  the  solution  contain  one   gram  of  the 



f  the  ■ 


cr>'Stallized  salt,  and  it  requires  seventy  cc.  to  precipitate  one 
gram  of  iron  as  ferric  phosphate. 

Sodium  acetate  solution,  made  by  dissolving  250  grams  crys- 
tallized sodium  acetate  in  8zo  cc.  of  water.  100  cc.  of  this  solu- 
tion contain  twenty-five  grams  of  sodium  acetate,  which  is  a 
slight  excess  over  that  which  is  necessary  to  convert  the  nitric 
and  hydrochloric  acids  to  sodium  nitrate  aud  chloride,  with  the 
i.liberatioii  of  the  corresponding  amount  of  acetic  acid. 
I  Gramdaled  lead  is  of  the  same  quality  as  that  used  in  assay- 
ing. In  size  it  should  be  that  which  passes  through  a  sieve 
with  twenty  meshes  to  the  inch,  but  remains  upon  a  sieve  with 
forty  meshes.  Before  using,  the  lead  should  be  washed  with 
dilute  hydrochloric  acid  (one  part  o(  acid  to  two  parts  of  water) 
in  order  to  dissolve  any  oxide  that  may  be  present. 

Standard  nickel  solution.  This  may  be  made  from  chemically 
pure  nickel  by  dissolving  two  and  a  half  grams  nickel  in  fifty 
cc.  nitric  acid,  sp.  gr.  1.20,  adding  an  excess  of  hydrochloric 
acid,  evaporating  on  a  water-bath  nearly  to  dryness,  then  dilu- 
ting to  one  liter.     One  cc  =■  0.0045  gram  of  nickel. 

Standard  potassium  cyanide  solution. — Take  twelve  grams  of 
C.  P,  potassium  cyanide,  dissolve  in  water,  dilute  to  one  liter. 
This  must  be  standardized  against  a  standard  nickel  solution. 
Since  the  presence  of  ammonium  salts  interferes  somewhat  in 

tde,  necessitating  the  use  of  a 
,um  cyanide  than  would  be  re- 
sahs  present,    it  is  better 

the  titration  with  potassium  cyani 
slightly  greater  amount  of  potass! 
quired  if  there 

that  the  potassium  cyanide  be  standardized  under  the  same 
conditions  as  are  met  in  analysis.  To  standardize  the  potassium 
cyanide,  take  fifteen  to  twenty  cc.  of  the  standard  nickel  solu- 
tion, add  six  cc.  of  hydrochloric  acid,  sp.  gr.  r.20,  ten  cc. 
sodium  phosphate  solution,  ammonium  hydroxide  until  the  solu- 
tion turns  blue  and  then  five  cc.  in  excess.  Now  add  hydro- 
chloric acid  until  the  blue  color  of  the  double  nickel  chloride 
disappears,  then  ammonium  hydroxide  until  the  blue  color  just 
reappears,  and  an  excess  not  exceeding  one  cc.  Dilute  to  200 
cc.,  add  five  cc.  cupric  ferrocyanide  indicator  and  run  in  potas- 
sium cyanide  until  the  solution  changes  from  the  purplish  color 

23  = 


imparted  by  the  indicator  to  a  perfectly  clear  light  straw-yellow. 

Divide  the  amount  of  nickel  in  the  standard  nickel  solution 
taken,  by  the  number  of  cubic  centimeters  of  potassium  cyanide 
used,  less  the  correction  for  the  indicator.  The  result  will  gj^'e 
the  strength  of  the  potassium  cyanide  expressed  in  metallic 

Cupric  ferrocyanidf  indicator. — Take  two  and  a  half  grams  of 
crystallized  cupric  sulphate,  dissolve  in  tweuty-five  cc.  of  water. 
add  to  this  a  solution  of  ammonium  oxalate  until  the  precipitate 
first  formed  just  redissolves,  then  dilute  to  500  cc.  Dissolve  two 
and  a  half  grams  of  potassium  ferrocyanide  in  500  cc.  of  water, 
then  slowly  pour  this  solution  into  the  cupric  sulphate  solution, 
stirring  constantly  during  the  operation.  This  will  give  a  deep 
purplish  brown  solution  of  cupric  ferrocyanide  which  may  pre- 
cipitate partially  on  standing  ;  but  the  precipitate  so  formed 
will  be  so  fine  that  it  will  easily  remain  in  suspension  for  a  long 
time,  upon  shaking  the  bottle,  thus  insuring  unifonn  composi- 
tion. To  find  the  correction  for  the  indicator  take  200  cc.  of 
water,  add  six  to  eight  drops  of  ammonium  hydroxide,  then  five 
cc.  of  indicator,  taken  after  shaking  the  bottle  well,  and  then 
run  in  potassium  cyanide  until  the  characteristic  change  of  color 
is  obtained.  Five  cc,  of  cupric  ferrocyanide  of  the  above 
strength  require  from  0.15  to  0.20  of  potassium  cyanide,  one  cc, 
of  which  is  equivalent  to  0,0025  nickel.  If  a  stronger  end  reac- 
tion is  desired,  ten  or  even  fifteen  cc.  of  the  indicator  may  be 
used  and  a  suitable  correction  made. 

Repeated  analyses  of  steel  have  shown  that  the  nickel  may  be 
determined,  by  the  volumetric  method,  within  from  0.0003  to 
0.0005  gram  of  the  true  nickel  content,  duplicate  determinations 
being  made  in  three  hours.  The  electrolytic  method  requires 
three  hours  to  the  time  the  solution  is  ready  for  electrolysis. 




Analysis  of  Chimney  Gases  for  Oxygen,  Carbon  Dioxide,^ 
Carbon  Monoxide,  and  Nitrogen. 

The  determinations   usually  made   are   the   percentages.  byJ 


olume,  of  oxygen,  carbon  dioxide,  carbon  monoxide,  and  nitro- 

The  apparatus  used  (a  modified  form  of  the  Elliott)  is  shown 
I  Fig.  63.  and  consists  of  two  glass  tubes,  ib  and  aA,  the  tube  ib 

taving  a  capacity  of  about  125  cc.  and  is  accurately  graduated 
a  o  cc.  to  100  cc.  iu  oue-tenth  cc.     At  d  and  e  are  three-way 


glass  stopcocks,  connected  by  means  of  rubber  tubing  to  the 
water-supply  bottles,  /and^.*  The  manipulation  of  the  appa- 
ratus is  as  follows  : 

Remove  the  funnel  cap  r,  and  connect  in  its  place  a  glass  tube 
of  small  diameter,  but  of  sufl&cient  length  to  reach  well  into  the 
flue  from  which  the  gases  are  to  be  taken.  Open  the  stop-cocks 
a  and  b  and  slowly  raise  g  and  /  until  both  tubes  are  full  of 
water  including  the  glass  tube  in  the  flue.  It  is  necessary  in 
this  operation  to  be  certain  that  no  air  is  in  the  tubes  and  that 
the  displacement  by  water  is  complete.  Now  gradually  lower 
the  bottle/ whereby  the  gas  is  drawn  into  the  tube  ah.  As  soon 
as  sufficient  gas  has  been  obtained  for  the  analysis,  the  lower 
portion  of  the  tube  containing  water  two  or  three  inches  above 
the  point  A,  the  stop-cock  a  is  closed,  the  small  glass  tube  con- 
necting a  with  the  flue  removed,  and  the  funnel  cap  c  replaced. 
After  allowing  the  gas  to  stand  in  the  tube  ah  fifteen  minutes  to 
secure  it  the  temperature  of  the  room,  and  thus  insure  correct 
measurements,  the  bottle^  is  slowly  lowered  until  the  surface  of 
the  water  therein  is  on  an  exact  level  with  o  on  the  tube  /d,  the 
stop-cock  b  opened  and  the  bottle  /gradually  raised  until  suffi- 
cient gas  from  ah  has  been  transferred  to  bi^  indicated  by  the 
volume  taken  reading  from  the  mark  o  on  the  graduated  tube  ib 
to  the  mark  loo  cc.  immediately  in  contact  with  the  stop-cock  b. 

Having  thus  obtained  loo  cc.  of  the  gas,  the  stop-cock  b  is 
closed  and/ is  raised  until  all  the  remaining  gas  in  ah  and  ab  is 
displaced  by  the  water.  The  first  constituent  of  the  gas  to  be 
determined  is  the  carbon  dioxide  (CO,).  The  gas  is  now 
transferred  to  the  tube  ah  by  raising  g  and  opening  ^,  keep- 
ing a  closed  and/ lowered.  When  the  water  reaches  b  the  latter 
is  closed. 

Fifty  cc.  of  a  solution  of  caustic  potash  are  placed  in  the  funnel 
cap  c,  (The  solution  being  made  by  dissolving  280  grams  of 
potassium  hydrate  in  1000  cc.  of  distilled  water.) 

Open  the  stop-cock  a  only  partially,  so  that  the  solution  of 
caustic  potash  in  c  may  slowly  drop  down  through  the  gas  in  the 
tube  ah  and  absorb  the  carbon  dioxide  in  so  doing. 

I  The  water  used  in  this  apparatus  should  contain  too  grams  sodium  chloride  i& 
each  liter  of  distilled  water. 




When  all  the  caustic  potash  in  c  (with  the  exception  of  two  or 
three  cc.)  has  passed  through  a,  the  latter  is  closed,  thus  pre- 
venting entrance  of  any  air  ;  b  is  opened,  /is  slowly  raised  and 
g  lowered.  Continue  the  raising  of/until  the  water  in  the  tube 
ha  reaches  the  stop-cock  b  and  immediately  close  the  latter. 
Allow  the  gas  to  stand  in  the  tube  ib  five  minutes  before  taking 
the  reading  of  the  volume  on  the  tube,  bearing  in  mind  that  the 
level  of  the  water  in  g  must  be  on  a  level  with  the  water  in  ib  to 
obtain  equal  pressure.  The  difference  between  o  and  the  point 
indicated  by  the  water  in  the  tube  ib  will  give  the  amount  of 
carbon  dioxide  absorbed  from  the  gas  by  the  caustic  potash. 

Original  volume  iBdicated  at 0,0 

After  removal  of  carbon  dioxiite 11. 1 

or  1 1.2  per  cent,  carbon  dioxide  by  volume. 

To  obtain  the  oxygen  the  gas  is  forced  from  ib  into  ah,  as  be- 
fore, and  in  c  is  placed  fifty  cc.  of  an  alkaline  solution  of  pyro- 
galtic  acid. 

This  latter  solution  is  formed  by  dissolving  ten  grams  of  pyro- 
gallic  acid  in  twenty-five  cc.  of  distilled  water,  placing  it  in  c 
and  adding  thirty-five  cc.  of  the  caustic  potash  solution.  This 
is  allowed  to  pass  slowly  through  a  and  gradually  absorbs  the 
oxygen  in  the  gas.  a  is  closed  before  all  the  liquid  passes  out 
of  c.  Repeat  with  the  same  quantity  of  alkaline  pyrogallic 
solution.  Transfer  the  gas  in  the  usual  manner  to  ib,  and  after 
allowing  to  stand  five  minutes,  take  the  measurement  thus  : 

Previous  reading 11.3  cc. 

.After absorbing  oxygen ig.6  " 

Oiygen 8.4  " 

or  8.4  per  cent,  by  volume. 

Before  transferring  the  gas  to  ah  for  the  determination  of  the 
carbon  monoxide,  all  the  water  in/and  ah  must  be  replaced  by 
distilled  water;'  to  do  this,  open  the  three  way  cock  c,  open  a, 
and  all  the  water  can  be  caught  in  a  large  beaker  at  e.     Wash 

I  CarbOQ  dioxide  ii  much  ihqtc  soluble  in  distilled  waier  Ibao  carbon  tuonoilde  or 
tnied  water  can  be  iitcd. 


out /and  ah  three  times  with  the  water,  then  close  e  in  the  prop- 
er manner  so  that  the  water  placed  in/ will  rise  in  the  tube  ha 
to  a,  then  close  a,  lower/,  raise ^,  open  b,  placing  the  gas  in  ah  for 
treatment  with  a  solution  of  cuprous  chloride  to  determine  the 
carbon  monoxide. 

The  cuprous  chloride  solution  is  made  by  dissolving  thirty 
grams  of  cuprous  oxide  in  200  cc.  hydrochloric  acid  (sp.  gr. 
1. 19),  and  using  fifty  cc.  as  soon  as  the  solution  has  reached  the 
temperature  of  the  room. 

Experience  has  shown  that  a  freshly  made  solution  acts  much 
better  as  an  absorbent  of  carbon  monoxide  than  one  that  has 
stood  several  days.  Fifty  cc.  of  this  solution  are  placed  in  c 
and  allowed  to  slowly  drop  through  a  and  absorb  the  carbon 
monoxide  as  it  passes  through  the  gas.  This  absorption  should 
be  repeated  at  least  three  times.  The  heat  generated  during 
this  absorption  often  causes  such  an  increase  in  the  volume  of 
the  gas  that  when  the  latter  is  transferred  to  the  tube  ib  for  meas- 
urement, the  reading  may  prove  minus.  To  insure  accuracy 
proceed  as  follows : 

The  gas,  after  fifteen  minutes,  is  transferred  in  the  usual  way 
to  bi,  and  the  water  in  /and  ah  is  replaced  with  distilled  water. 
The  gas  is  now  returned  to  ah  and  a  solution  of  potassium  hy- 
droxide is  placed  in  c  and  allowed  to  pass  through  the  gas  in 
ah,  absorbing  all  traces  of  hydrochloric  acid  gas.  Repeat  with 
this  once.  Return  the  gas  to  bi,  allow  to  stand  fifteen  minutes, 
then  take  the  reading  : 

Previous  reading 19.6  cc. 

After  using  Cu^Cl,  solution 20.7    *• 



The  nitrogen  is  determined  by  subtracting  the  total  amounts 
of  carbon  dioxide,  oxygen  and  carbon  monoxide  from  100. 
Thus  the  analysis  will  read  : 

Carbon  dioxide 11. 2  per  cent,  by  volume* 

Oxygen 8.4    "       "      "        *' 

Carbon  monoxide i.i    "       **      **        ** 

Nitrogen 79.3    **      **      *«        ** 

Total loo.o    **      **      "        «* 



In  this  analysis  no  corrections  are  required  for  the  tension  of 
the  aqueous  vapor,  since  the  original  gas  is  saturated  with 
moisture,  and  during  the  analysis  all  measurements  are  made 
over  water.* 

To  convert  percentages  by  volume  to  percentages  by  weight 
proceed  as  follows  : 

liter  of  oxygen  gas  weighs  1.430  grams. 

"   hydrogen  "  "       0.0895 

**   nitrogen  "  '*       1.255 









'*   carbon  dioxide 
•*   carbon  monoxide 
"   methane 
*'  acetylene 
Then  11.2  cc.  carbon  dioxide 
8.4   **   oxygen 
I.I   "   carbon  monoxide 
79.3   "   nitrogen 




1 1 




1 1 


<  I 

1 1 



1. 251 



gas  weighs  0.02202  gram. 

"         ••       001201       " 

0.00138      " 



<  I 

f  I 



1 1  t  i 



If  the  100  cc.  of  gas  weighs  0.13467  gram,  then 

0.02202  X  100 

The  carbon  dioxide  = 
The  oxygen  = 

0.1349 I 

0.01200  X   100 


=  16.32  per  cent,  by  weight. 
=  8.97  per  cent,  by  weight. 

^*_          1.                    .J         0.00138  X   100  ^  .  ... 

The  carbon  monoxide  = =  1.02  per  cent,  by  weight. 


^,        •  0.09952  X   100  ,  ^   .  ... 

The  nitrogen  =  — --— =  73.69  per  cent,  by  weight. 

**  0.13491 

Total     100.00  per  cent,  by  weight. 

Analysis  of  Flue  Gases  with  the  Orsat-Muencke 


Where  the  determinations  to  be  made  are  the  percentages  of 
carbon  dioxide,  carbon  monoxide,  oxygen  and  nitrogen,  this 

1  The  solubility  of  these  four  gases,  at  normal  temperature  and  pressure,  are  as 
fonows  : 

1  volume  of  air-free  water  at  15*  C.  absorbs  1.002  volume  of  carbon  dioxide. 
1        "        *•        *'  »•        "    *•    "         "         0.024        "        "   carbon  monoxide. 






» ■  .ft  it 

tl  tt  tt 


*   oxygen. 
'   nitrojfen. 


apparatus  offers  many  advantages  over  any  other.     It  is  shown  I 
in  Fig,  69,  and  is  thus  described: 

The  measuring  burette  .4,  of  100  cc.  capacity,  is  surrounded 
by  a  large  cylinder  filled  with  water,  in  order  10  tree  the  gas 
from    changes  of  temperature,  and   the  first   forty-five   cc. 

divided  into  tenths  cc,  the  remaining  fiity-five  cc.  into  cubic! 
centimeters.  The  thick  capillary  glass  tube  is  fastened  at  both  I 
ends,  at  /  in  a  cut  of  the  dividing  panel,  and  at  0  by  means  of  a  / 
small  brace,  attached  to  the  cover  of  the  case. 

The  capillary  tube  is  bent  at  its  further,  end  and  connected] 
with  the  jj  tube  B,  containing  cotton,  and  at  the  bend  is  filled  J 



with  water  in  order  to  retain  all  dust  and  to  saturate  the  gas 
thoroughly  with  moisture  before  measuring  takes  place. 

The  rear  end  of  the  three  way  cock  c  is  connected  by  means  of 
a  rubber  tube  a  with  the  rubber  aspirator  C.  which  fills  the  tube 
vpith  the  gas  to  be  analyzed. 

The  absorption  takes  place  in  the  "  U  "  formed  vessels  /?,  E, 
and  F,  which  are  connected  with  the  stoppers  b}-  short  rubber 
tubes.  For  the  enlargement  of  the  absorbing  surface,  D.  £'and 
y^are  filled  with  glass  tubes.  Since  the  mark  m  is  above  the 
place  of  connection,  the  latter  is  always  moistened  by  the  re- 
spective liquid  and  therefore  can  easily  be  maintained  air  tight. 
The  other  end  of  the  [)  tube  vessel  is  closed  by  a  rubber  cork, 
■which  contains  the  small  tube  .v  :  the  small  tubes  are  al!  con- 
nected to  one  rubber  bulb  of  about  200  cc.  capacity  in  order  to 
keep  out  the  atmospheric  oxygen.  The  entire  apparatus  is  en- 
closed in  a  wooden  case  fifty  centimeters  high  and  twenty-five 
centimeters  wide.  Its  use  is  indicated  as  follows:  The  glass 
cylinder  surrounding  the  burette  A  as  well  as  the  bottle  L  are 
filled  with  distilled  water.  In  order  to  fill  the  three  absorbing 
cylinders,  the  stoppers  are  removed  as  well  as  glass  tubes  a^  and 
the  rubber  bag  G,  and  110  cc.  potassium  hydroxide  solution 
(sp.  gr.  1.26)  poured  into  the  vessel  D.  so  that  the  latter  is  about 
half  full.  This  is  for  the  absorption  of  the  carbon  dioxide.  E 
contains  a  solution  of  eighteen  grams  of  pyrogallic  acid  in  forty 
cc.  of  hot  water,  which  is  poured  into  E,  and  then  sevent.v  cc. 
of  potassium  hydroxide  solution  (sp.  gr.  1.26)  added,  whereby 
the  oxygen  is  absorbed  in  the  gas  under  examination. 

The  carbon  monoxide  is  absorbed  in  the  cylinder  F.  which 
contains  a  solution  of  cuprous  chloride  made  as  follows;  Thirty- 
five  grams  of  cuprous  chloride  are  dissolved  in  200  cc.  hydro- 
chloric acid  (concentrated),  fifty  grams  of  copper  clippings 
added  and  the  mixture  allowed  to  stand  in  a  glass -stoppered  bot- 
tle lor  twenty-four  hours.  Each  glass  tube  in  F  contains  a 
spiral  of  copper  wire.  100  cc.  of  water  is  added  to  the  solution 
(no  precipitate  forming),  and  enough  is  transferred  to  F  to  fill 
to  the  required  poiut.  The  solutions  in  the  rear  section  of  D, 
E.  and  F  are  transferred  to  the  front  sections,  where  the  absorp- 
tion of  the  gas  takes  place  as  follows  :  The  three  glass  stoppers 


are  closed,  the  stop-cock  c  turned  horizontal  and  the  bottle  /,, 
containing  distilled  water,  raised  so  that  the  water  fills  the  bu- 
rette A ,  give  a  quarter  turn  to  the  left  to  the  stop-cock  c.  so  that 
the  second  passage  leads  to  the  tube  5,  open  the  stop-cock  of 
the  vessel  D,  lower  the  bottle  L  and  carefully  open  the  pinch- 
cock  placed  on  the  tube  s,  so  that  potassium  hydroxide  solution 
rises  to  the  mark  m.  whereupon  the  stop-cock  is  closed.  The 
fluids  of  the  two  other  absorbing  vessels  are  raised  in  the  same 
way  to  the  mark  m.  The  three  stoppers  with  the  glass  tubes 
.V  are  then  attached.  About  one  cc.  of  water  is  placed  lu  the 
tube  B,  loose  cotton  placed  in  both  sides,  the  stopper  reinserted 
and  connected  with  the  tube  >i.  After  filling  the  burette .-/  with 
water  to  the  100  cc.  mark  by  raising  the  bottle  L,  the  stop-cock 
is  turned  so  that  the  connection  of  the  rubber  aspirator  C  with' 
the  chimney,  containing  the  flue  gases,  is  brought  about  throughi 
the  tube  B.  Aspiration  of  the  gas  into  the  apparatus  is  now 
performed  by  compressing  Cten  or  fifteen  times  till  the  whole 
conductor  is  filled  with  gas.  This  is  easily  done  by  compress- 
ing Cwith  the  left  hand,  closing  the  attached  tube  r  with  the 
thumb  of  the  right  hand,  and  then  upon  opening  the  left  hand 
allowing  Cto  expand,  raising  the  thumb  agaiu,  compressing  C, 
etc.,  till  the  object  is  obtained.  To  fill  the  burette  .-I  with  the 
gas,  the  stop-cock  f  is  turned  horizontal,  the  pinch-cock  of  the 
tube  s  opened,  and  the  bottle  L  lowered  until  the  gas  reaches 
the  zero  point  in  A,  whereupon  c  is  closed. 

To  determine  the  carbon  dioxide,  the  stop-cock  of  ZJ  is  opened 
and  L  raised  with  the  left  hand,  so  that  on  opening  the  pinch' 
cock  of  s  with  the  right,  the  gas  enters  the  cylinder  D ;  L  ia 
loweredagain  until  the  potassium  hydroxide  solution  in  iP  reaches 
to  about  the  tube  connection  under  w,and  once  again  drives  the 
gas  into  the  potassium  hydroxide  vessel  by  the  raising  of  i. 
This  is  repeated  two  or  three  times,  and  the  gas  retunied  to  the 
burette  .i  by  opening  the  pinch-cock  of  s  and  raising  L,  and. 
closing  the  glass  stop-cock  of  />.  To  measure  the  amount  of 
absorbed  carbon  dioxide,  the  bottle  /-  is  held  next  to  the  burette 
in  such  a  way  that  the  water  stands  at  the  same  level  in  both 
vessels,  the  pinch-cock  of  s  closed,  and  the  remaining  volume  of 
gas  read  off.     This  amount  subtracted  from   100  cc.  gives  the 







amount  of  carbon  dioxide.  The  gas  is  now  passed  into  the  ves- 
sel Sin  the  same  manner  as  in  /?,  the  oxygen  being  absorbed  by  the 
alkaline  pyrogallate  solution.  This  absorption  must  be  re- 
peated three  or  four  times  or  until  no  diminution  of  volume 
takes  place.  The  gas  is  then  returned  to  the  measuring  burette 
j4  and  the  amount  o(  absorption  measured. 

The  gas  is  then  passed  into  the  vessel  F  for  the  absorption  of 
carbon  monoxide.  After  repealing  for  a  number  of  times  the 
absorption  in  F  the  gas  is  passed  into  D  before  measurement  in 
vJ  of  the  absorbed  carbon  monoxide.  This  is  necessary  on  ac- 
count of  the  vapors  of  hydrochloric  acid  retained  by  the  gas  after 
contact  with  the  cuprous  chloride  solution  in  hydrochloric  acid. 
After  passing  the  gas  into  D  three  or  four  times,  it  is  then  meas- 
ured as  usual  in  .4,  the  remaining  gas  being  nitrogen. 

The  composition  of  the  chimney  gases  is  an  index  of  the 
working  of  the  furnaces  under  the  boilers.  When  the  fuel  is 
properly  consumed,  the  furnace  gases  should  contain  only  nitro- 
gen, oxygen,  steam,  and  carbon  dioxide,  and  to  secure  this  re- 
sult, excess  of  air  is  required,  but  this  excess  must  not  exceed  a 
certain  amount,  otherwise  too  great  a  volume  of  air  is  heated 
and  the  heat  wasted. 

This  excess  of  air  can  be  determined  by  finding  the  amount 
carbon  dioxide  in  the  furnace  gases  ;    thus  the  percentages  ol 
carbon  dioxide,  herewith  given,   show  the  amount  of  air  used 
in  the  furnace.' 

4  ptr  cent,  carbon  dioxide  indicates  4.9  times  the  theoretical 
I  required  was  in  Ihe  gases. 

5  per  cent,  carbon  dioxide  indicates  3.5  times  the  theoretical 
r  required  was  in  the  gases. 

6  per  cent,  carbon  dioxide  indicates  3.0  times  the  theoretical  amount  o< 
■ir  required  was  in  the  gages. 

7  per  cent,  carbon  dloiide  indicates  2.5  times  tbe  Iheorelical 
r  requited  was  in  the  gases. 

8  per  cent.  cBrt>on  dioxide  indicates  2.3  times  the  theoretical 
r  required  was  in  the  gases. 

9  per  cent,  carhou  dioxide  indicates  3.0  times  tbe  theoretical 
r  required  was  in  the  gases. 

10  per  cent,  carbon  dioxide  indicates  1.7  times  the  theoretical 
r  required  was  in  the  gases. 



II  per  cent,  carbon  dioxide  imlicates  1,5  times  the  iheoreticnl  a 
air  required  was  in  the  gases. 

17  per  cent. carbon  dioxide  indicates  l.o  times  the  theoretical 
air  required  was  in  the  gase^. 

It  is  customary  in  boiler  trials  to  make  analyses  of  furnace 
gases  and  calculate  the  amount  of  air  required  for  combustion 
from  the  percentage  of  carbon  dioxide  found  in  the  furnace 

Prof.  W.  C.  Unwin,  F.R.S..'  states  that  this  method  is  accu^  1 
rate  in  principle,  but  the  samples  analyzed  are  a  very  minute  i 
fraction  of  the  total  chimney  discharge,  and  the  samples  may  i 
not  be  the  average  samples.   What  is  wanted  is  an  instrument  i 


easily  read  as  a  pressure  gauge,  and  giving  continuous  indica-  ■ 
tions,  such  as  the  dasymeter  of  Messrs.  Siegert  &  Durr,  of  | 
Munich.  (Fig.  70.)  This  is  a  fine  balance  in  an  enclosed  case, 
through  which  a  current  of  the  furnace  gases  is  drawn.  At  one  ' 
end  of  the  balance  is  a  glass  globe  of  large  displacement,  at  the 
other  a  brass  weight.  Any  change  of  density  of  the  medium  in 
the  chamber  disturbs  the  balance.  A  finger. on  the  balance 
moving  over  a  graduated  scale  gives  the  amount  of  the  altera- 
tion of  density. 
An  air  injector  draws  the  funiace  gas  from  tlie  flues,  and  it  ii 



filtered  belore  entering  the  balance  case.  An  ingenious  mer- 
curial compensator  counterbalances  any  effect  due  to  change  of 
temperature  or  barometric  pressure. 

The  dasymeter  is  usually  combined  with  a  draught  gauge, 
and  an  air  thermometer  or  pyrometer  in  the  flue  is  required  if 
the  amount  of  waste  heal  is  to  be  calculated. 

The  losses  through  sensible  heat  in  the  escape  gases  can  be 
easily  determined  with  the  assistance  of   the  dasymeter   and 
Siegert's  approximate  formula  in  the  following  way  ; 
Lei  the  carboo  dioxide  =  .r  in  per  cent. 

temperature  of  discharged  gases  =  /"( Celsius  scale), 
temperature  of  draught  at  grate  ^  /. 
then  tbe  loss  of  Twith  the  coats  as  fuel  equals  : 
T  —  I 
1'=  0-65-^—  in  per  cent,  of  the  heat  value. 

With  lignite,  peat,  wood,  etc.,  the  coefficient  varies  accord- 
iiig  to  the  contents  of  water  and  the  coefficient  of  heat  of  the  fuel 
aud  is  so  much  the  greater,  the  less  valuable  the  combustible  is. 

With  coal  furnaces  the  loss  of  heat  can  immediately  be  ob- 
tained  from  the  following  diagrams  without  any  calculation  : 


■  M 

"  K^      - 

-  cS 

■■  fix 



.      ^  ^NS, 


..  s^"^:i 
























In  Fig.  71  look  up  the  carbon  dioxide  =  contents  of  carbon 
dioxide  in  the  lower  horizontal  (abscissa  row),  follow  the  ver- 
tical line  appertaining  thereto  till  it  intersects  the  curve  of  the 
irplus  temperature,  draw  from  this  point  of  section  a  horizon- 
1  tal  line  to  the  left  and  it  will  give  the  amount,  per  cent,  of  the 
I  loss  of  heat  as  indicated  by  that  point  on  the  scale  at  which  tt 
I  was  intersected. 


In  Fig.  72  look  up  the  amount,  per  cent,  of  the  surplus  tem- 
perature on  the  bottom  abscissa  line,  raise  a  perpendicular  line 
from  the  point  till  it  intersects  the  line  drawn  diagonally  for  that 
amount  of  carbon  dioxide  indicated  by  the  dasymeter^  The  hori- 
zontal line  through  this  point  of  section  indicates,  on  the  scale  for 
the  loss  of  heat  on  the  left,  the  loss  to  be  determined.  Points 
lying  between  two  given  abscissa  can  easily  be  assumed  on  both 
diagrams  by  eye  measurement. 

Experiments  show  that  when  using  horizontal  and  step  grate 
furnaces,  as  also  the  Ten-Brink  furnaces,  the  most  profitable 
combustion  is  obtained  when  about  ten  to  fourteen  per  cent,  of 
carbon  dioxide  is  contained  in  the  escaped  gases,  and  in  the  use 
of  gas  furnaces  about  seventeen  to  eighteen  per  cent. 

Thedasymeter  requires,  initially,  exceedingly  delicate  adjust- 
ment, and  its  indications  must  be  checked  from  time  to  time  by 
analysis  of  the  gas.  It  is  set  to  read  zero  with  pure  air,  and  then 
any  increase  of  density  due  to  carbon  dioxide  is  read  as  a  percen- 
tage on  the  scale*  When  in  adjustment,  it  is  as  easy  to  read  the 
percentage  of  carbon  dioxide  in  the  furnace  gases  as  to  read  the 
pressure  on  a  pressure  gauge.  When  the  dasymeter  is  fitted  to 
a  boiler,  the  stoker  has  directions  to  adjust  the  supply  of  air,  so 
that  the  furnace  gases  have  about  twelve  per  cent,  of  carbon 

With  practice  he  learns  what  alterations  of  the  damper  or  fire- 
door,  or  thickness  of  fuel  on  the  grate  are  necessar>',  or  whether 
an  alteration  of  grate  area  is  desirable. 

After  a  little  practice  the  percentage  of  carbon  dioxide  can  be 
kept  very  constant. 

Uehling  &  Steinbach  describe  an  instrument  they  make  use 
of  to  determine  the  composition  of  furnace  gases,  and  which  in- 
dicates automatically  and  continually  the  percentages  of  carbon 
dioxide  and  monoxide  present  in  furnace  gases.  It  is  fully  de- 
scribed in  United  States  patent  No.  522746. 




COAL    GASN^WATEK     GAS,    OIL    GAS,     PRODUCER    GAS.     ETC.,    BY 

In  technical  analysis  of  gases  the  most  complete  experiments 
may  be  conducted  with  the  aid  of  the  Hempel  apparatus.  The 
essential  feature  of  this  apparatus  consists  in  the  fact  that  meas- 
urements and  absorptions  may  be  conducted  separately  and  in 
special  apparatus.  Gas  burettes  serve  the  first  purpose  and  for 
the  latter  gas  pipettes.'  The  gas  burette,  as  shown  iu  Fig.  73, 
consists  of  two  parts,  the  calibrating  tube  b  and  the  levelling 
tube  a.  The  first  has  a  constant  diameter  and  ends  above  iu  a 
capillary  tube  about  one-half  millimeter  iu  diameter  and  three 
centimeters  in  length  ;  at  the  ixittom  it  tapers  into  a  small  tube 
bent  at  an  angle  and  passing  through  and  protruding  from  the 
wooden  stem^,  supported  by  an  iron  base.     ^^ 

The  calibrating  tube  is  divided  from  the  capillary  part  down 
to  a  little  above  the  wooden  support  into' two-tenths  cc,  the 
total  graduation  comprising  100  cc.  A  rubber  tube  about  120 
cm.  long,  ha\'ing  a  short  length  of  glass  tubing  inserted  at 
about  the  middle,  as  shown  in  Fig.  73,  serves  to  connect  the 
glass  tube  projecting  at ^  with  the  levelling  lube  a.  which  at 
the  bottom  is  .similarly  fastened  into  the  base  at  e.  The  tube 
a  sXh  widens  into  a  funnel  to  facilitate  pouring  in  the  water. 
Over  the  capillary  tube  r  of  the  measuring  tube  a  short  piece 
of  heavy  rubber  tubing  is  fastened  by  means  of  wire.  A 
strong  "  Mohr"  pinch-cock /enables  one  to  close  the  measuring 
tube  directly  alxive  the  capillary  tube.  The  rubber  tube  at  d 
has  a  -—I  shaped  capillary  glass  lube  leading  from  it  (see  E  Fig. 
75).  to  provide  for  a  connection  with  the  various  gas  pipettes. 

Since  in  this  simple  gas  burette  water  is  used  as  a  sealing 
fiuid.  it  is  not  adopted  lor  the  analysis  of  gases  containing  con- 
stituents easily  soluble  in  water.  In  such  cases  Winkler's  gas 
burette.  Fig.  74,  is  used.     The  capillary  tube  b  is  closed  lielow 



by  a  three-way  cock  c  and  above 
by  means  of  a  simple  stop-cock  rf. 
Similarly  to  the  Hempel  burette 
both  the  measuring  tube  li  and 
levelling  tube  a  are  fastened  into 
iron  stands  and  are  connected  by  a 
rubber  tube.  The  space  between 
the  stop-cocks  c  and  d  is  divided 
into  too  cc. ,  and  each  of  these  into 
fifths  of  cc.  Before  use  the 
"  Winkler"  burette  must  be  thor- 
oughly dried,  by  rinsing  with  alco- 
hol and  ether,  and  thereupon  pass- 
ing a  current  of  dry  air  through  it. 
In  order  to  admit  a  sample  of  gas 
to  be  analyzed.  1:  is  connected  by 
rubber  or  glass  tubing  with  the 
source  of  gas,  and  the  length-bore 
of  the  three-way  cock  f,  which 
communicates  with  the  inside  of  d, 
is  attached  to  an  aspirator  or  rub- 
ber pump.  Gas  is  drawn  through 
till  all  the  air  has  been  displaced, 
thereupon  closing  c  and  d.  In 
order  to  transfer  the  gas  into  the 
pipettes,  the  levelling  tube  a  and 
the  rubber  tube  are  filled  with 
water  till  the  latter  commences  to 
flow  from  the  stop-cock  c.  which  at 
this  moment  communicates  with  a 
through  its  length-bore.  The 
flow  of  water  is  checked  by  closing 
with  a  rubber  tube  and  glass  rod, 
or  a  pinch-cock.  At  Winkler's 
suggestion  the  calibrating  tube  6 
of  Hempel's  simple  gas  burette  is 
jacket  in  order  to  reduce  the  effect  of 
temperature  upon  the  gas  in  the  burette. 


The  larger  glas.i  tube  serving  a.s  a  water  jacket,  Fig  75,  is 
closed  above  ant!  below  by  two  rubber  corks,  through  which  the 
calibrating  tube  passes,  and  lias  also  above  and  below  two  small 
projecting  glass  tubes,  used  for  filling  or  discharging  the  water  ; 
they  are  either  simply  closed  by  rubber  corks  or  have  attached 
to  them  rubber  tubes  to  produce  a  continuous  flow  of  water  in 
the  jacket. 


On  the  working  table  there  rests  a  stand  G  upon  which  the 
pipettes  are  placed  and  whose  height  is  so  adjusted  that  the 
entrance  to  the  pipette  and  the  capillary  tube  of  the  burette  are 
at  one  level.  These  pipettes,  which  must  be  equal  in  number  to 
the  various  absorptions  which  are  to  be  executed  (since  each 
one  remains  charged  with  one  liquid  and  always  serves  for  the 
determination  of  only  one  gas  constituent^  have  according  to 
the  purpose  which  they  serve,  different  attachments. 

The  simple  absorption  pipette,  Fig.  76,  consists  of  two  glass 

Fglobes  a  and  i,  connected  by  means  of  a  bent  glass  tube  d,  and  fast- 
f  ened  to  a  wooden  stand  to  prevent  breakage.  A  capillary  tube  c 
f  passes  from  the  globe  6  before  a  plate  of  milk  glass  w,  which  is 
[  let  into  the  wooden  stand,  in  order  to  easily  trace  the  move- 
t  ments  of  the  liquid  thread  in  the  capillary  tube  c.  The  exit  tube 
_/'of  the  globe  a  and  the  capillarj- tube  c  extend  above  the  wooden 
frame  :  a  small  rubber  tube  e  is  connected  to  the  protruding 
tube  c  and  fastened  by  means  of  wire.  The  reagent  to  be  used 
in  the  pipette  is  poured  in  at  /,  entirely  filling  the  globe  6,  a 
I  only  partially,  and  the  capillar>-  tube  c  to  the  junction  with  the 
[  rubber  tube  near  <•.  When  not  in  use. /is  closed  by  a  cork  and 
f  by  a  glass  rod,  which  during  use  is  displaced  by  a  pinch-cock. 



A  label  designating  the  contents  of  tlie  pipette  is  attached  to  the 
wooden  frame.  The  gas  is  transferred  into  these  pipettes, 
brought  into  intimate  contact  with  the  reagent  by  shaking  and 
thus  freed  from  the  constituent  gas  under  consideration.  The 
simple  burette,  containing  caustic  potash  solution  (i  to  2)  is 
used  for  absorption  of  the  carbon  dioxide.  The  pipette  contain- 
ing fuming  sulphuric  acid,  Fig.  77,  is  so  modified  that  shaking 
is  avoided.  Above  the  globe  *,  filled 
1  with  disulphuric  acid,  the  smaller  globe 
jf  also  filled  with  the  fuming  sulphuric 
acid  and  pieces  of  broken  glass  (the  lat- 
ter placed  there  by  the  glass-blower). 
When  the  gas  passes  into  the  pipette  it 
comes  into  contact  with  large  surfaces 
of  the  broken  glass,  which  are  covered  I 
\\-ith  the  absorbing  liquid.  Passing  the  J 
gas  through  ff  three  or  four  times  s 
fices  for  complete  absorption.  TlW 
Fig.  ^.  heavy  hydrocarbons  in  the  gas  a 

sorbed  in  the  pipette  by  the  disulphuric  acid. 

Fig.  78  shows  the  compound  pipette,  two  of  which  are  used:^ 



one  lor  the  determination  of  oxygen  in  the  gas.  the  other  for  the 
determi nation  of  the  carbon  dioxide. 

This  pipette  is  charged  with  alkaline  pyrogallate  solution  for 
the  former  and  with  cuprous  chloride  solution  for  the  latter  de- 

The  bulb  a  next  the  capillary  tube  is  filled  with  a  solution  of 
alkaline  pyrogallol,  the  bulb  b  partially  filled  with  the  same 
solution,  the  bulb  c  is  empty  or  nearly  so,  and  the  bulb  d  con- 
tains distilled  water.  The  alkaline  pyrogallol  solution  is  made 
by  dissolving  one  part  of  a  twenty-five  per  cent,  pyrogallol  solu- 
tion in  water,  in  six  parts  of  a  sixty  percent,  solution  of  caustic 

In  order  to  illustrate  the  working  of  the  Hempel  apparatus, 
an  analysis  is  here  given  of  a  gas  containing  carbon  dioxide, 
oxygen,  carbon  monoxide,  ethylene,  methane,  hydrogen,  and 
nitrogen.  A  sample  of  this  gas.  loo  cc,  is  collected  and  meas- 
ured in  the  gas  burette.  The  carbon  dioxide  is  first  absorbed 
by  passing  the  gas  into  the  potassium  hydroxide  pipette.  Fig. 
76,  containing  a  solution  of  one  part  potassium  hydroxide'  in 

o  parts  of  water.  Agitate  the  gas  and  potash  solution, 
and  after  waiting  five  minutes  pass  the  gas  back  into  the 
measuring  burette  and  determine  the  carbon  dioxide  absorbed. 
The  contraction  produced  gives  directly  the  percentage  of  car- 
bon dioxide,  since  100  cc.  were  used  at  starting." 

The  oxygen  is  next  absorbed  in  the  compound  pipette,  Fig. 
'8.  The  absorption  is  complete  in  about  five  minutes.  The 
I    amount  of  oxygen  absorbed  is  now  measured.' 

Some  chemists  prefer  to  use  stick  phosphorus  for  the  absorp- 
tion of  oxygen,  The  phosphorus  pipette  is  shown  in  Fig.  79. 
The  bulb  b,  contains  pieces  of  phosphorus  inserted  through  the 
opening  at  k,  which  is  closed  by  a  rubber  stopper.  The  bulb  b, 
and  the  capillary  tube  c,  are  filled  with  water,  likewise  a  portion 
of  a,. 

After  the  absorption  of  the  oxygen  the  next  step  is  to  absorb 
the  acetylene  by  means  of  disulphuric  acid  in  the  pipette.  Fig. 


77.  The  absorption  is  complete  in  a  few  minutes,  but  the  re- 
maining gas,  previous  to  measuring,  should  be  passed  into  the 
potassium  hydroxide  pipette  twice,  in  order  to  free  the  gas  from 
fumesof  sulphur  trioxide.  Allow  the  gas  to  stand  in  the  graduated 
burette  five  minutes  before  taking  measurement  of  volume.     The 

GAS    ANALYSIS.  253 

carbon  monoxide  is  next  absorbed  by  means  of  a  solution  of 
cuprous  chloride  (Cu, CI,)  in  hydrochloric  acid'  in  the  compound 
pipette.  Fig.  80. 

Complete  absorption  of  carbon  monoxide  is  somewhat  slow. 
Fifteen  minutes  should  be  givenforthis  with  frequent  agitation  of 

I  thegaswithcuprouschloridesolution.  The gasisthenpassed into 
the  potassium  hydroxide  pipette  to  absorb  fumes  of  hydrochloric 
acid  before  it  is  transferred  to  the  measuring  burette,  where  after 
wraiting  five  or  ten  minutes  the  volume  can  be  measured.  The 
residual  gas  now  contains  hydrogen,  methane,  and  nitrogen. 

The  hydrogen  is  determined  by  passing  the  gases  over  palla- 
dium sponge  in  the  palladium  tube  £,  Fig.  81. 
Ad  improved  form  is  shown  in  Fig.  83. 

•  prepared  by  di 
hyilrochlDTic  acid,  ■ 
CHTbon  dioxide. 



The  palladium  tube  is  kept  at  a  temperature  not  exceeding 
ioo°  C,  either  by  using  a  minute  flame  as  shown  in  Fig.  8r.  or 
by  immersing  the  tube,  Fig.  82.  in  a  beaker  ol  water  at  the  re- 
quired temperature.  Under  these  conditions  the  combustion  of 
hydrogen  proceeds  without  the  combustion  ol  any  methane. 
The  gas  is  passed  and  repassed  through  the  tube  slov'ly  at  least 
three  times,  the  palladium  tube  during  the  whole  operation 
ing  connected  with  an  ordinary  absorption  pipette  filled  v 
water,  Fig.  81.  , 

Finally  the  gas  is  transferred  to  the  measuring  burette  and  the! 
volume  determined.  As  the  hydrogen  has  burned  to  water  by 
uniting  with  the  occluded  oxygen  in  the  palladium  sponge,  the 
diminution  in  volume  represents  the  amount  of  hydrogen  in  the 
gas  directly.  The  residual  gas  now  contains  methane  and 
nitrogen.  Of  this  gas  ten  cc.  are  now  taken  and  the  rest  allowed 
to  escape,  or  if  necessary  can  be  collected  and  saved  in  a  pipette. 

To  the  ten  cc.  of  the  gas  in  the  burette  ninety  ec.  of  air  arc 
added  and  this  mixture  of  air,   methane,  and  nitrogen  passed] 
into  the  explosion  burette,  Fig.  83. 

The  gases  are  thoroughly  mixed  and  then  exploded.     The  | 


current  required  to  do  this  is  generated  by  a  dip  battery  of  four 
cells  connected  with  a  small  induction  coil,  and  with  the  explo- 
sion pipette  by  the  platinum  wires  kk.  Fig.  83,  which  are  fused 
into  the  pipette.  After  the  explosion  the  gases  are  led  into  the 
potassium  hydroxide  pipette  to  absorb  the  carbon  dioxide  formed 
by  the  combustion  of  the  methane.  One  volume  of  methane  re- 
quires four  volumes  of  oxygen  for  its  combustion,  producing  one 
\'olume  of  carbon  dioxide  and  two  volumes  of  water.  One-third 
of  the  loss  of  volume,  after  the  explosion  and  measurement  in 
the  burette,  gives  the  volume  of  methane  in  the  ten  cc.  of  the 
gas.  This  amount  subtracted  from  ten  cc.  gives  the  amount  of 
nitrogen  in  the  ten  cc,  and  these  amounts  must  be  calculated 
back  into  values  of  the  total  amount  of  gas  left  in  the  burette 
before  mixing  with  air. 

The  following  analysis  of  a  sample  of  carburetted  water  gas 
will  indicate  the  working  of  the  method. 

100  cc.  of  the  gas  taken. 

Before  use  of  KOH loo.o  cc. 

After      *'     *•      •*     96.2 

( < 


COi  =         3.8 

No  oxygen  present. 

Before  use  of  H^S^O, 96.2  cc. 

After     ••    '*        "       81.6    *• 

(Illumiiiants, )  C,H4,  etc.  ^     14.6    " 

Before  use  of  Cu,Cl)  solution 81.6  cc. 

After      "    *'        *•  ''         53.6   " 

CO  =  28.0   " 

Before  use  of  Palladium  tube 53.6  cc. 

After      "     **         "  "     18.0   •• 

H  =  35.6   - 

CH^  +  N  remaining  =  18  cc. 

Ten  cc.  taken  +  ninety  cc.  air.  After  explosion  in  the  explo- 
sion pipette  (Fig.  78)  and  measurement  after  absorption  of  car- 
bon dioxide  formed,  the  volume  was  71.8  cc,  which  corresponds 
to  28.2  cc.  of  absorption  or  nine  and  three-tenths  cc.  of  methane. 
If  ten  cc.  of  the  gas  gave  nine  and  three-tenth  cc.  of  methane. 


eighteen  cc.  of  the  gas  (amounts  remaining  before  mixing  with 
air)  will  give  16.7  cc.  of  methane. 

Calculated  to  x8  cc. 

The  volume  of  gas  before  adding  air 18.0  cc. 

after  explosion  and  absorption(N)     1.3 

<<  •*  <<         <<      affA*- A<VT\1rkcirk«  a«#1  aKa^»•^4^;^•«/Xr^         ▼    'v        << 

CH4=i6.7    " 
Resum6  : 

Carburetted  water  gas.  By  volnrae. 

COj 3.8  per  cent. 

(Illuminants)  CjH^,  etc 14.6    *•       •* 

CO 28.0    " 

H 35.6    - 

CH, 16.7     •* 

N 1.3 

it  44 

Total loo.o     "       ** 

and  by  weight : ' 

COj 9.6  per  cent. 

( Illuminants)  CjH^,  etc  23.7 

CO 45.1 

H 4.1 

CH, 15.4 

4(  (( 

44  44 

(4  4< 

44  44 

N 2.1 

44  44 

Total 100.00    '*       ** 

Some  chemists  prefer  to  determine  the  hydrogen  and  methane 
by  explosion,  instead  of  using  the  palladium  tube  for  the  hydro- 
gen. In  this  case  suppose  a  partial  analysis  of  gas  gave  as  fol- 
lows ( 100  cc.  of  gas  taken)  : 

Carbonic  acid 2.2  per  cent. 

Oxygen 0.0     **       ** 

Illuminants 12.8     **       ** 

Carbon  dioxide 24.2     **•      ** 

39.2    ^*       *• 

The  remaining  constituents  being  (in  the  60.8  cc.  gas  left) 
methane,  hydrogen  and  nitrogen.     These  are  treated  as  follows  : 

Fifteen  cc.  of  this  residual  gas  are  taken  and  mixed  with 
eighty-five  cc.  of  air.     This  is  then  passed  into  the  explosion 

1  For  method  of  calctilatioa  se«  Analysis  of  Chimney  Gases,  p.  237. 


burette  containing  water  previously  saturated  with  the  gas.  It 
is  well  shaken  to  insure  thorough  mixture  of  the  air  and  gas, 
and  then  exploded  by  means  of  a  spark  from  the  induction  coil. 
After  fifteen  minutes  the  reading  is  taken  :  the  latter  being  77.4 
cc.  or  22.6  cc.  contraction.     (loo — 77.4=  22.6.) 

Methane  produces  an  equal  volume  of  carbon  dioxide  by  com- 
bustion ;  therefore,  if  the  carbon  dioxide  produced  be  measured 
by  absorption  with  potassium  hydroxide,  the  amount  represents 
the  methane.  Pass  the  gas  into  the  potassium  hydroxide 
burette  and  determine  carbon  dioxide. 

77.4  cc.  —  73  cc.  =4.4  cc.  methane  in  the  fifteen  cc.  of  the  gas 
mixed  with  the  air,  or  in  per  cent,  of  100  cc.  of  original  gas  : 

15  cc.  :  60.8  : :  4.4  :  x  or  17.83  per  cent,  methane. 

The  hydrogen  is  determined  as  follows  : 

Let  C=  contraction  (15  cc.  gas  +  85  cc.  air)  after  explosion. 
**   Z>  =  CO,  =  CH,  in  15  cc.  of  gas. 

^1.  LT  2C ±D  45.2  17.6  27.6  ., 

Then /r  = -1 —  =  ^^ '- —  =  -^ — =  9.2percent.* 

3  3  3 

hydrogen  in  fifteen  cc.  of  the  residual  gas,  or  32.07  per  cent, 
hydrogen  in  the  100  cc.  of  the  original  gas. 

By  adding  together  all  of  the  constituents  determined,  and 
subtracting  this  amount  from  100,  the  residue  is  nitrogen.  Thus 
the  complete  analysis  will  l;e  : 

CO,     2 .  20  per  cent. 

Illuminants 12.80    "       " 

O 0.00    " 

CO 24.20    "      •• 

CH, 17.83 

H 37-95 

N 5.02 

<<       11 
(I       (I 

Total 100.00    "      " 

iConfult  Gasometriiche  Methoden,  R.  Bunsen,  2nd  edition,  pp.  4^51. 


Heating  Value  of  Combustible  Gases. 

In  the  calculation  of  the  fuel  value  of  gases,  the  method  as 
given  by  H.  L.  Payne*  will  be  found  accurate  and  convenient. 
Since  the  results  of  gas  analyses  are  stated  volumetrically  the 
calculation  of  a  number  of  analyses  is  greatly  facilitated  if  the 
calories  per  kilo  are  converted  into  heat  units  per  volume,  and 
custom  requires  results  to  be  stated  in  **  B.  T.  U.**  per  cubic 
foot  of  gas.  Calories  per  kilo,  multiplied  by  f  (the  ratio  of 
the  Fahrenheit  thermometer  degree  to  the  Centigrade)  will  give 
British  thermal  units  per  pound.  Dividing  this  result  by  the 
number  of  cubic  feet  of  each  gas  per  pound  will  give  **  B.  T.  U.** 
per  cubic  foot.  To  calculate  the  cubic  feet  per  pound  of  gas  the 
following  fundamental  relations  are  used  : 

I  pound  =  453-59  grams. 

I  meter  =  39.37  inches. 

1728  (cubic  inches  to  one  cubic  foot)  divided  by  (3.937)'=: 
28.317  ;  therefore  one  cubic  foot  =  28.317  liters.  From  this  is 
obtained:  **  B.  T.   U."  per  cubic  foot  =  Calories  per  kilo  X 

-^  X    — '.^—L  X  the  liter  weight  in  grams  of  the  gas  in  ques- 

5         453-59 


The  following  table  contains  the  liter  weights  as  determined 
by  actual  weight  of  some  of  the  gases  : 

Gas.  Grams  per  liter. 

H 0.0895 

O 1 .430 

N 1 .  257 

Air 1*293 

CO 1. 251 

CO, 1.966 

CH, 0.7155 

CjH, 1.252 

For  the  other  hydrocarbons  and  gases  not  given  above,  we 
may  substitute  in  the  formula  in  place  of  liter  weight  the  ex- 


This  formula  then 


\  2  1.007       ' 

may  be  reduced  to  the  following  form:    "  B.  T.  U."  per  cubic 

foot=  calories  (per  kilo)  X  ""'""'"tJ^'SJ'  X  0.01003.  which 

is  comparatively  simple  and  can  be  used  in  all  cases  n-ithout 
very  serious  error. 

Applying  this  formula  to  the  different  combustible  gases  we 

Table  of  Heating  Valphs  of  Gases. 

One  kilo  of  H  evolves  upon  complete  combust  ion  34.500  calories,  or 
61100  B.  T.  U.  per  pouud,  or  348  B.  T.  t'.  per  cubic  foot  at  0'=  C.  and  760 
mm.  pressure. 

One  kilo  of  CO  evolves  upou  complete  combustion  1,487  calories,  or 
4.476  B.  T.  r.  per  pound,  or  349  B,  T.  U.  per  cubic  foot  at  o*^  C,  and  760 
tntn.  pressure. 

One  kilo  of  CH,'  (methane)  (marsh  gas)  evolves  upon  complete  com- 
bustion 13,14s  calories,  or  33,851  B.  T.  U.  per  pound,  or  1,065  ^-  T.  t-'.  per 
cubic  foot  at  o-  C.  and  760  mm.  pressure. 

One  hilo  of  C,H,  (acetylene )  evolves  upon  complete  combustion  1 1,935 
calories,  or  21,465  B.  T.  U.  per  pound,  or  1,555  B-  T.  U.  per  cubic  foot  at 
cP  C.  and  760  mm.  pressure. 

One  kilo  of  C,H,'  (ethylene)  (olefiant  gas)  evolves  upon  complete 
combustion  11.900  calories,  or  Ji,44o  B.  T.  U.  per  pound,  or  1,673  B.  T.  U. 
per  cubic  foot  at  o^'  C.  and  760  mm.  pressure. 

One  kilo  ol  C,H,  (ethane)  (ethyl  hydride!  evolves  upon  complete  com- 
plete combustion  1 1,350  calories,  or  21,130  B.  T.  V.  per  pound,  or  1.858 
B.  T.  U.  per  cubic  foot  at  o^'  C.  and  760  mm.  pressure. 

One  kilo  of  C,H,  (propane)  (propyl  hydride)  evolves  upon  complete 
combustion  11,038  calories,  or  31,650  B.  T.  U.  per  pound,  or  1,654  B.  T.  V. 
per  cubic  foot  at  o"  C.  and  760  mm.  pressure. 

One  kilo  of  C,H|  (propylene)  evolves  upon  complete  combustion  11,900 
calories,  or  31,430  B.  T.  U.  per  pound,  or  3,509  B.  T,  V.  per  cubic  foot  at 
o-  C.  Bud  760  mm.  pressure. 

One  kilo  of  C.Hn,  (quartane)  (butane)  evolves  upon  complete  combus. 
tioo  11.850  calories,  or  21.330  B.  T.  U.  per  pound,  or  j.447  B.  T.  U.  per 
cubic  foot  ato-  C.  and  760  mm.  pressure. 

One  kilo  of  C^H,,  (quintane)  (pentaue)  evolves  upon  complete  combus- 
tion 11,770  calories,  or  31,186  B,  T.  U.  per  pound,  or  4,350  B.  T.  U.  per 
cubic  fool  at  o'  C.  and  760  mm.  pressure. 

tiesoICH CaMi  arc  Imken  tram  ThomtOD'j  n/nm»- 

tktmu  VmlrrsiHki. 

^m      ud  Ihe  B.  T.  U.  prt  , 



One  kilo  of  C,H,,  (sextane)  evolves  npoo  complete  combustion  ti,6jO' 
calories,  or  20,916  B.  T.  U.  per  pound,  or  5,012  B.  T.  U.  per  cubic  foot  at 
o'   C.  and  760  mm.  pressure. 

One  kito  of  C,H,  (benzeue)  evolves  upon  complete  combustion  10.150 
calories,  or  1S.450  B.  T.  V.  per  pound,  or  4,010  B.  T.  U.  pet  cubic  foot  at 
o'^  C.  and  760  mm.  pressure. 

One  kilo  of  CuHj  (naphthalene)  evolves  upon  complete  combustion 
9,6ao  calories,  or  17,316  B.  T.  U.  per  pound,  or  6,176  B.  T.  U.  per  cubic 
foot  at  D-  C.  and  760  mm.  pressure. 

To  calculate  the  heat  units  of  a  gas  from  its  analysis,  multi- 
ply the  per  cent,  of  each  constituent  by  its  number  as  given  in 
the  above  table,  and  the  sum  of  the  products  will  represent  the 
British  thermal  units  evolved  by  the  combustion  of  one  cubic 
foot  of  the  gas.  Ordinary  gas  analysis  includes  as  combustibles 
only  hydrogen,  carbon  monoxide,  methane,  and  "illuminants, " 
the  latter  term  representing  the  hydrocarbons  that  are  deter- 
mined by  absorption  in  fuming  sulphuric  acid  or  bromine. 
This  has  proven  to  be  a  very  trustworthy  value  where  the  hy- 
drocarbons are  derived  chiefly  from  the  decomposition  of  mineral 
oil,  but  if  produced  by  the  distillation  of  coal,  this  value  is  too 
low.  owing  to  a  larger  percentage  of  benzene  vapors  contained. 
The  experinientai  conditions  necessary  to  give  these  theoreti- 
cal results  are  that  the  gas  be  measured  at  32°  F.  and  760  B.  and 
is  burned  with  exactly  the  proper  quantity  of  oxygen,  and  that 
the  products  of  combustion  are  reduced  to  the  initial  tempera- 
ture, the  water  being  all  in  the  liquid  state.  It  is  superfluous 
to  say  that  this  cannot  actually  be  done  ;  but  as  the  whole  mat- 
ter is  a  theoretical  discussion,  it  is  decided  to  adhere  to  the  sci- 
entific standard,  and  to  state  results  in  accordance  with  its  defi- 

But  in  order  to  obtain  figures  which  shall  more  nearly  agree 
with  practice,  many  persons  have  preferred  to  make  their  calcu- 
lations under  certain  assumed  conditions.  This  plan  is  not 
without  merit,  since  by  it  a  somewhat  better  idea  of  the  true  rel- 
ative values  of  fuel  constituents  is  obtained,  similar  conditions 
affecting  different  gases  unequally.  The  following  examples 
illustrate  this: 

The  temperature  for  standard  gas  measurement  in  this  coun- 
try is  60°  F..  and  this  point  is  usually  assumed   as  the  initial 



temperature.  As  a  final  temperature  in  this  case  let  the  tem- 
perature of  the  steam  be  100  pounds  absolute  pressure  per 
square  inch  (328®  F. ) ,  a  point  considerably  lower  than  the  average 
chimney  flue  heat.  Under  these  conditions  combustion  taking 
place  in  air,  not  in  oxygen,  we  must  add  the  heat  brought  in  by 
the  gas  and  air  at  60®  F.  and  subtract  the  heat  carried  away  by 
the  products  of  combustion  at  328*  F.,  and  since  the  volume  of 
gas  is  greater  at  60®  than  at  32°,  we  correct  for  this  by  multi- 

492  402 

plying  the  result  — ^ — -  or  -^^—  . 
'^  "^     **  492  +  28       520 

The  composition  of  air  is : 

By  volume.  By  weight. 

0 20.92  per  cent.  23.134  per  cent. 

N 79.08    ••      "  76.866 

1 1       It 

Hence  4.78  volumes  of  air  contain  one  volume  of  oxygen,  or 
one  volume  of  oxygen  is  accompanied  by  3.78  volumes  of  nitro- 

The  specific  heats  of  the  several  gases  are  as  follows  : 

Gaf.  Sp.  heat. 

H 3.4 

O 0.22 

N 0.24 

Air 0.24 

CO 0.25 

CO, 0.22 

CH4 U.60 

"Illuminants'' 0.41 

It  will  be  more  convenient  in  these  computations  to  make  use 
of  the  so-called  **  volumetric  "  specific  heats,  /.  ^.,  the  heat 
necessary  to  raise  the  temperature  of  one  cubic  foot  of  gas  from 
32^  F.  to  33'  F. 

Gaf.  Vol.  sp.  heat. 

H 0.019 

N 0.019 

0 0.019 

Air 0.019 

CO 0.019 

CO, .' 0.027 

CH4   0.027 

**  lUnminants  ** 0.040 


Then  : 

2H  +  O  +  3.78  N  =  H,0  +  3.78  N ;  on  cu.  ft.  H  -f  2.39  cu.  ft.  air 

I  cu.  ft.  steam  +  1.89  cu.  ft.  N. 

The  heat  gained  or  brought  in  is  as  follows  : 

H  =  I  X  0.019  X  [28=  (6o"-32'0]  =  0.53  B.  T.  U. 
Air  =  2.39  X  0.019  X  28  =  1.27  ** 

«4        (I 

«4        C« 

Total  gain     1.80" 

The  heat  lost  or  carried  away  by  the  products  of  combustion 
is  as  follows  :  Water  at  32*  F.  converted  into  steam  at  328®  F. 
absorbs  1,182  B.  T.  U.  per  pound,  and  one  cubic  foot  of  hydro- 
gen produces  when  burnt  0.0502  pounds  of  water. 

Steam  =  0.0502  X  1182  =  59.4  B.  T.  U. 
N  =  1.89  X  0.019  X  296  =s  10.6 

l(         Ct         4( 

Total  loss  =  70.0  '• 
Subtract  gain  =    1.8   ** 

Net  loss  =  68.2   **    *•    ♦• 

348  B.  T.  U.  less  68.2  B.  T.  U.  leaves  279.8  B.  T.  U.  and  cor- 
rected for  volume  gives  264  B.  T.  U.,  a  loss  of  twenty-four  per 

In  the  case  of  carbon  monoxide  no  water  is  produced  by 
combustion  and  the  former  value  is  consequently  much  less 

One  cubic  foot  of  carbon  monoxide  -(-  2.39'  cubic  foot  of  air= 
one  cubic  foot  carbon  dioxide  -|-  1.89  cubic  foot  nitrogen.  Heat 
gained  or  brought  in  by  carbon  monoxide  and  air,  the  same  as 
hydrogen  and  air  in  the  previous  case,  one  and  eight-tenths 
B.  T.  U. 

Heat  lost : 

CO.^  =  I  X  0.027  X  296  =  8.0  B.  T.  U. 
N  as  above  =  10.6  '*   '*    '* 

Total  loss    18.6  "   "    *' 
Subtract  gain       1.8  "   "    ** 

Net  loss     16.8  •*    •*    ** 

349.5  B.  T.  U.  less  16.8  B.  T.  U.  leaves  332.7  B.  T.  U. 

1  Refer  to  sample  in  hydrogen  combustion. 


This  corrected  for  volume  gives  315,  a  loss  of  only  ten  per 

For  marsh  gas,  i  cubic  foot  CH^  +4(2.39  cubic  foot  air)  = 
one  cubic  foot  CO,  -|-  2  cubic  feet  steam  +4(1.89  cubic  feet 

Heat  gained : 

CH4  =  I  X  0.027  X  28  (60^— 32"^)  =    o  8  B.  T.  U. 
Air  ^4Xi-27B.  T.  =5.1 

((    1 1    (< 

( <    ( (    i» 

Total  gain  =    5.9 
Heat  lost : 

CO,  =  I  X  0.027  X  296  (328^^—32^)  =        8.0  B.  T.  U. 
Steam  =  2  X  59.4  B.  T.  U.  =    118. 8 

N  =  4  X  10.6  B.  T.  U.  =      42.4 

Total  loss=     169.2  " 
Subtract  gain  =         5.9 

K         i< 
<(        < (         (( 

<(         l( 

Net  loss  =     163.3  ** 

1065  B.  T.  U.  less  163.3  B.  T.  U.  leaves  901.7  B.  T.  U. 

This  corrected  for  volume  gives  853  B.  T.  U.f  901.7  X  ^^  ) 

\  520   / 

a  loss  of  twenty  per  cent. 

For  '*  illuminants**  fifteen  per  cent,  is  taken  as  a  fair  loss,  and 
the  values  are : 

32°  p.  initial.  60*  initial.  Loss  in 

Gas.  32*  P.  final,  328"  final.  percent. 

H 348.0  B.T.U.              264  B.  T.  U.  24 

CO 349-5 315  "    "    "  10 

CH4 1065.0  ••    "    **  853  "    "    '*  20 

Illuminants 2000.0  *'    "    "  1700  "   "    "  15 

Natural  gas  has  been  taken  as  a  standard  for  heating  gases 
with  a  valuation  of  1000  B.  T.  U.  per  cubic  foot. 

At  ninety-four  per  cent.  CH„  which  is  not  far  from  the  aver- 
age, it  will  show  by  calculation  1000  B.  T.  U.  per  cubic  foot  and 
hence  the  numerical  result  obtained  by  this  method  for  any  fuel 
gas  will  indicate  also  its  standing  in  that  scale. 

Referring  to  the  analyses  of  gas  given  on  page  256,  the  *'  B. 
T.  U."  per  cubic  foot  are  as  follows  : 


Products  of  Combustion  Condensed. 

CO 28.0  per  cent.  X    349-5=   97.86  B.  T.  U. 

CO, 3.8 

(Illuminants)  CjH^,  &c 14.6    *'      *'      X  2000.0  s  292.00 '*    *'    ** 

H 35.6    "       *'      X    348.0^123.88*'    **    •• 

CH4 16.7    *•      *'      X  1065.0  =  177.85  ••   ••    •* 

ToUl  =  691.59  **    •*    ** 

The  **  B.  T.  U."  per  cubic  foot  of  gas  will  be  as  follows: 

Products  of  Combustion  in  Vapor  at  328^'  F. 

CO 28.0  per  cent.  X    315=    6820  B.  T.  U. 


(Illuminants)  CaH^ 14.6    "       '*      X  1700  =  248.20  *'    '*    ** 

H 35.6    •'       '*      X    264=   63.98*'    **    " 

CH, 16.7    *'       **      X    853  =  142.45**    **    •' 


•  •  • 

Total  =  552.83  *♦    **    ** 

In  determining  the  B.  T.  U.  per  pound  or  of  calories  per  kilo, 
the  analysis  of  the  gas  by  weight  is  taken  and  not  by  volume, 
as  just  instanced. 

The  B.  T.  U.  per  pound  of  the  gas  would  be  : 

Products  of  Combustion -Condensed. 

CO 45.1  per  cent.  X    4476  =  2018.6  B.  T.  U. 

(Illuminants)  CjH^,  &c  23.7    •'       "      X  21440  =  5081.2  **   **    ** 


H 4.1    •'       "      X  62100  =  2546.1  *'    **    ** 

CH,  ....   15.4    "       *'      X  23851  =  3673.0  **    •«    •* 

Total  =  13318.9  *•  *'    ** 

and  where  the  products  of  combustion  are  in  vapor  at  328''  F., 
as  follows : 

CO     45.1  percent.  X    3402.0  =  1534.3  B.  T.  U.  per  pound. 


(Illuniinants^  C,H^  23.7     ••      **      X  18224.0  =  4319.0  **    *'    **     " 


H         4.1     ••       ••      X  47196.0=  1935.0  *•    **    *•     ** 
CH4  15.4     "       •*      X  19081.0=2938.4**    '*    ••     **         ** 

Total  =  10726.7 



Nearly  all  of  the  carburetted  water  gas  in  the  United  States  is 
made  either  by  the  Lowe  process  or  the  Wilkinson  process, 
probably  four-fifths  by  the  former. 

Briefly  slated,  the  Lowe  water  gas  system  consists  in  the  de- 
composition of  steam  at  a  high  temperature  by  in  candescent  car- 
bon, thereby  producing  hydrogen  and  carbon  dioxide  :  2H,0  + 
C-2H,  +  C0,. 

In  an  excess  of  carbon,  the  carbon  dioxide  saturates  itself 
with  another  carbon  atom,  forming  carbon  monoxide  (CO,  -j-  C 
^  2CO.  making  the  finished  product  2H,  -|-  2CO). 

In  practical  working  the  reduction  of  carbon  dioxide  to  mon- 
oxide is  never  quite  perfect,  the  unpuritied  gas  usually  contain- 
ing about  three  per  cent  of  carbon  dioxide,  to  be  extracted  (as 
in  coal  gas)  by  lime  purification. 

As  the  gas,  in  the  process  of  manufacture,  passes  from  the 
generator  to  the  carburetters,  it  is  enriched  by  means  of  crude 
oil  or  cheaper  distillates ;  hence  the  name  carburelted  water 

The  generator,  carburetter,  and  super-heater  are  cylindrical 
steel  shells,  thickly  lined  with  special  fire  blocks,  between 
which  and  the  metal  are  annular  spaces  packed  with  non-con- 
ducting material.  The  generator  is  usually  supported  on  short 
columns,  as  illustrated,  leaving  cartage  room  under  the  hopper- 
shaped  ash-pit.  The  grate,  controlled  by  the  several  cleaning 
doors,  is  located  slightly  above  the  ash-pit.  and  the  fire  is 
charged  with  coke  through  the  door  in  the  extreme  top. 

The  generator  is  connected,  both  above  and  below  the  fuel- 
bed,  with  the  top  of  the  carburetter,  the  bottom  of  which  leads 
laterally  into  the  adjoining  super-heater.  The  carburetter  and 
super-heater,  often  referred  to  as  the  fixing-chambers,"  are  filled 
with  checker  work,  and  affording  such  an  enormous  heating 
surface  that  even  the  heaviest  distillates  can  be  permanently 
gasified  at  the  low  temperatures  necessary  to  the  highest  illumi- 
nating effect.  The  enriching  oil  is  introduced  at  the  top  of  the 

The  oil  heater  is  a  simple  and  practical  arrangement  for  pre- 

I  Humphreys  &  CIssgow  :  "  Carbuntud  Waltr  Gas."  189s. 



ieating  the  oil  on  its  way  to  the  carburetter  by  means  of  the  hot 
gas  escaping  from  the  super-heater. 

Operation. — A  fire  is  started  in  the  generator,  which  is  then 
deeply  charged  with  coke  and  opened  to  the  blast.  The  air 
enters  in  large  volume  below  the  grate  and  quickly  kindles  the 
fuel,  while  the  hot  products  resulting  from  the  partial  combus- 
tion pass  forward  through  the  carburetter  and  super- heater  and, 
after  parting  with  their  sensible  heat,  escape  into  the  stack.  As 
sewn  as  these  generator  gases  have  sufBciently  wamied  the 
checker-work,  supplies  of  secondary  air  are  admitted  to  the  topof 
the  carburetter  and  the  bottom  of  the  super-heater  respectively, 
and  the  combustion  regulated  to  give  the  requisite  temperatures 
in  the  two  vessels  simultaneously.  The  generator  fire  being  in 
proper  condition,  and  the  carburetter  and  super-heater  at  the 
desired  temperatures,  the  apparatus  is  ready  for  gas  making. 
The  blasts  are  shut  off  one  by  one,  beginning  with  that  of  the 
super-heater ;  the  stack  valve  is  closed ;  steam  is  admitted 
under  the  fuel  bed,  and  having  traversed  it,  passes  as  water  gas 
into  the  top  of  the  carburetter.  At  this  point  the  oil  is  intro- 
duced, and  encountering  the  heated  checker-work  is  vaporized 
and  ultimately  gasified  in  presence  of  the  hot  water  gas.  This 
process  continues  until  the  temperatures  of  the  fire  and  the 
checker-work  are  sufficiently  reduced.  The  oil  is  then  shut  off; 
uext  the  steam  ;  aud  the  stack  valve  being  opened  the  blasts  are 
again  admitted  and  the  energy  of  the  fire  and  the  checker-work 
recuperated  as  first  described.  The  generator  is  supplied  with 
fuel  at  inter\-als  of  from  fort>'-five  to  sixty  minutes,  and  cleaned 
usually  once  during  each  shift.  The  gas  passes  from  the  seal 
through  the  scrubbers  and  condensers  and  is  subsequently  de- 
prived of  its  carbon  dioxide  and  treated  for  its  slight  sulphur 
impurities  in  the  manner  common  to  coal  gas. 

tJncarbureUed  water  gas  has  the  following  composition  :' 

H 49-3*  percent. 

CH, 7-63     ■'       '■ 

■» 37-97    ■•      " 

:0, 0.14    ■•      '■ 

» 4.79    ■'      " 

> -     °-l3    "      " 

TotBl 100.00    '■      ■' 



and  after  carburetting 



CO 29.40 

O o.  10 


N 3.71 

Illuminants .^ 16.79 

38.05  per  cent. 
11.85    *• 






Total 100.00    '*      ** 

The  heating  power  of  the  uncarburetted  gas  per  cubic  foot 
would  be  : 

Products  condensed. 

H 0.4932  X     348.0    =  171.63  B.  T.  U. 

CH4 0.0765X1065.0    =    81.47" 

CO 0.3797  X  349-56  =  132.72  " 

«i    it 

it    i( 

Total 385.82*'    *•    " 

and  the  heating  power  of  the  carburetted  water  gas  per  cubic 
foot  would  be  : 

Products  condensed. 

H 0.3805  X     348.0    =  131.31  B.  T.  U. 

CH^ 0.1185X1065.0    =126.20*'    *'  *' 

CO 0.2940  X    349-56  =  102.77*'    "  ** 

^Ij^  ••••••••••  ••••  ••••  •••• 

^^  ••oa«*a*****  ••••  ••••  •••• 

^i  •.••••■•••••       ••••  ••••  •••• 

Illuminants..  0.1679  X  2000.0    =335.80'*   **   ** 


<(   li 

696.08  ** 

An  analysis  of  a  sample  of  London  (Eng.)  coal  gas  gives  the 
following  : 








Aqueous  vapor  ■ 

27.70  per  cent. 












Total 100.00 




The  heating  power  will  be,  per  cubic  foot, 

Products  condensed. 

H 0.2770  X     348.0    =    96.39  B.  T.  U. 

CH4 0.5000X1065.0    ^532.50" 

CO 0.0680  X    349-5^=    23.77  " 

C,H4 0.1300  X  1673.0    =217.49  **    '•    •* 

ToUl , 870.15'*    •*   *• 

and  when  the  products  of  combustion  are  in  a  state  of  vapor  (for 
instance  328*  F. )  the  heating  power  per  cubic  foot  will  be  : 

H 0.2770  X    264=   73.12  B.  T.  U. 

CH4 0.5000  X    853  =  426.50"  "  " 

CO 0.0680  X    315=   21.42"  "  " 

C,H4 0.1300x1400=182.00"  "  " 

Toul 703.04**  *•  " 

There  are  few  complete  analyses  of  purified  coal  gas  known,* 
f .  e.,  Heidelberg  gas  by  R.  Bunsen,  Konigsberg  gas  by  Bloch- 
mann,  and  Hannover  gas  by  Dr.  Fischer. 

Heidelberg^     Ronigsbergr    Hannover     Hannover 
gas.  gas.  gas.  gas. 

I.  II. 

C,He 1.33  0.66               0.69  0.59 

C,H^ 1. 21  0.72               0.37  0.64 

C,H4 2.55  2.01               2. II  2.48 

CH4 34.02  35.28  37.55  38.75 

H 46.20  52.75  46.27  47.60 

CO 8.88  4.00  II. 19  7.42 

CO2 3.01  1.40  0.81  0.48 

O 0.65  ....              trace  0.02 

N 2.15  3.18              i.oi  2.02 

Total 100.00  100.00  100.00         100.00 

In  the  Wilkinson  process  the  water  gas  is  made  by  the  com- 
bined generator  and  retort  process.  (A  full  description  of  a  recent 
plant  will  be  found  in  the  The  American  Gas  Light  Journal,  67, 

399,  401. 

An  analysis  of  a  sample  of  Wilkinson  water  gas,  made  by  the 

writer,*  gave  as  follows : 

1  'WmgneT'9  Manual  0/  Chemical  Technology,  (13th  edition,  1892)  p.  39. 
*  Wood's  '*  Thermodynamics,  Heat  Motors,  and  Refrigerating  Machines,  3rd  edition, 
pp.  a6o-a6i. 


H 39-50  per  cent. 

Heavy  hydrocarbons,   ^  „                  )  ^  ^ 

lUuminants,  C,H,  average  |   6.60 

CH, 37.30 

CO 4.30 

N 8.20 

0 1.40 

Impurities  (HjO,  CO,  HjS) 2.70 

ti  (( 

(I  <( 

tt  it 

n  tt 

tt  tt 

(t  tt 


One  cubic  foot  containing  755.31  B.  T.  U.  products  condensed. 
G.  Lunge'  gives  an  analysis  of  **  Tessie  du  Motay*'  gas,  as 
follows  : 


lUuminants 14.3  per  cent. 

0 0.6 

CO 27.7 

H 28.8 

CH, 25.5 

N 3.1 

<t  *( 

((  (( 

tt  i* 

tt  tt 

tt  tt 

Total loo.o    **      '* 

containing  827.62  B.T.  U.  per  cubic  foot,  products  condensed. 

For  complete  details  regarding  the  manufacture  of  coal  gas 
consult  King's  Treatise  on  Coal  Gas,'^  edited  by  Thomas  New- 
bigging,  London. 

Produckr  Gas. 

Sieman's  Anthracite         Soft  coal  pro- 

Constituents,  gas.  producer  gas.         ducer  gas. 

CO ...;...  23.7  27.0  27.0 

H 8.0  12.0  12.0 

CH4 2.2  1.2  2.5 

CO, 4.1  2.5  2.0 

N 62.0  57.3  56.5 

100.00       100.00       100.00 

The  heating  power  of  the  Sieman's  producer  gas  will  be 
134. 1  B.  T.  U.  per  cubic  foot:  of  the  anthracite  producer  gas 
153.7  S-  T.  U.  per  cubic  foot,  and  of  the  soft  coal  producer  gas 
168. 1  B.  T.  U.  per  cubic  foot  (products  of  combustion  con- 

1  "  Wassergasfabnkation  in  New  York.,"  Zetischri/'i  fitr angewandU  Ckemie^  1894,  pp. 




Oil  gas  is  usually  formed  by  vaporizatiou  of  mineral  oil  at 
high  temperatures.  Two  processes  are  in  use  :  the  "  Pintsch" 
and  the  "  Keith,"  the  former  probably  representing  ninety  per 
cent,  of  the  production  of  oil  gas. 

For  a  description  of  the  "  Pintsch"  oil  gas  apparatus  consult 
H'agner's  Chemical  Technology  (edition  of  1892),  p.  80,  also 
/.  Soe.  Chem.  Inditstry.  6,  March,  1887.  In  the  manufacture  of 
Pintsch  oil  gas,  in  the  United  States,  "mineral  seal"  oil  is  often 
used.  This  oil  is  a  petroleum  product  having  a  specific  gravity 
of  about  0.840,  flashing  point  266°  F.,  and  fire  test  311°  F. 

Several  analyses,  by  the  author,'  of  this  oil  give  carbon  83.30 
percent.,  hydrogen  13.20  per  cent.,  the  remainder  being  oxy- 
gen, nitrogen,  etc.,  and  the  analysis  of  the  gas  therefrom  gave  : 

CO .' 0.5  percent. 

CH 57.7    •■      - 

H 3-4    ■■      ■■ 

{Beniene  vapor.  CjH,  1 
Propylene  C,H,  \  .   38.1     "       ■■ 

Ethylene  C,hJ 

The  heating  power  would  indicate  1582  B.  T.  U.  per  cubic 
foot,  products  condensed. 

W.  Ivison  Macadam,  F.C.S.,_/.  Soc.  Chem.  Industry,  March, 
1887,  tabulates  the  results  of  a  series  of  his  tests  upon  the 
Pintsch  and  Keith  oil  gas,  as  follows  : 

Paraffin  Oil.  isto  Gas. 

Specific  gravity  of  the  oil  . . . 
Weight  of  one  gallon  of  the  o 
Number  of  gallons  per  ton  c 


Flashing  point 

Burning  point 

Gas  from  one  gallon  of  oil  ■  • '  • 

Candle  power  of  gas I 

Illuminating  value  of  I  cubic 
fool  in  grains  of  sperm 

lUumitiatiDg  value  of  i  ton  in 
lbs.  of  sperm 

.J  Am,<.  S>KUly  Mechanic. 


8,758  )bs. 


8.779  lbs. 


289^  F. 

347-  F. 
84.93  c-  ft- 
1,720  c.  ft. 
1.38  candles. 


=95^  F. 

354°  F. 

90.03  c.  ft. 

34,757  c.  ft. 

60.83  candles. 

1473  grs. 

I4S9  grs. 

4570  lbs. 

5160  lbs. 

lE-tgtnrfi,  [4, 

'893)  JSJ. 


Average  Average 

*                                                             of  trials  with  of  trials  with 

Keith's  Pintsch's 

apparatus.  appai^tus. 
Illuminating  value  of  i  gallon 

in  lbs.  of  sperm 17.876  lbs.  20.198  lbs. 

Heavy  hydrocarbons  absorbed 

by  bromine 39«o5  38.20 

Carbon  dioxide 0.27  0.08 

Dihydric  sulphide Decided.  None. 

Oil  gas,  compressed  to  six  or  eight  atmospheres,  in  iron  cylin- 
ders, is  extensively  used  for  the  lighting  of  railway  carriages. 
When  more  pressure,  say  ten  atmospheres,  is  used  the  gas  loses 
hydrocarbons  which  settle  out,  and  this  loss  in  illuminants  may 
cause  twenty  per  cent,  loss  in  the  illuminating  power  of  the  gas. 

References.— Qb.s  Manufacture  and  Analysis.  By  W.J.  Atkinson  Butter- 
field,  F.C.S.,  London,  1896. 

Methods  of  Gas  Analysis.  By  Dr.  Walther  Hempel,  translated  by  Prof. 
L.  M.  Dennis,  N.  Y.,  1892. 

Oil  Gas.  By  W.  A.  Noyes,  W.  M.  Blink,  and  A.  V.  H.  Mory,  /.  Am. 
Chem.  Soc,  x6,  688.   (A  report  of  a  very  complete  test  of  an  oil  gas  plant ). 

Technische  Gasanalyse.   By  C.  Winkler. 

Chemisch-Technische  Analyse.    By  Dr.  Julian  Post.  Braunsweig,  1890. 


The  increased  use  of  natural  gas*  in  the  metallurgical  indus- 
tries in  the  states  of  Pennsylvania,  Ohio,  and  Indiana,  has  given 
this  subject  an  enhanced  value. 

The  composition  of  the  gas  is  not  uniform  and  consequently 
the  heating  power  varies.  Chemists  are  not  in  agreement  with 
the  statements  of  the  results  of  analyses,  as  the  following  com- 
parisons show : 

Anai,ysis  of  Naturai,  Gas,  by  Dr.   G.  Hay,  for  the  Naturai,  Gas 


CO2 0.00  per  cent. 

CO i.oo  '* 

Heavy  hydrocarbons 0.50  *  *  •  * 

CH4 95.20  •'  •* 

H 2.00  "  " 

0 1.30  **  *• 

N 0.00  ** 

100.00    '*       ** 
B.  T.  U.  per  cubic  foot  =  1036.87. 

1  For  the  histor>'  of  the  development  of  natural  gas  in  Pennsylvania,  consult  Trans- 
actions  A  mer.  Institute  Mining  Engineers,  14,  423-439. 

2  Engineering  and  Mining  Journal,  39,  247. 


S.  A.  Ford  (chemist  to  the  Edgar  Thomson  Steel  Works) 
reports  analyses  of  natural  gas,  as  follows  : 

No.  I. 

No.  2. 

No.  3. 

No.  4. 

No.  5. 

No.  6. 







0.30  percent. 

CO  .... 





1. 00 



1. 10 




















CH4  . . . 







H  •    . .  • 



















i<       << 

Nos.  1-4  are  analyses  of  gas  from  the  same  well,  made  at  in- 
tervals of  two  months.  Nos.  5  and  6  are  from  two  different 
wells  in  the  East  Liberty  District,  Pa. 

W.  A.  Noyes*  gives  the  analysis  of  a  sample  of  natural  gas, 
Irom  New  Lisbon,  Ohio,  as  follows  : 

CH4 67.00  per  cent. 

C^He II. 10 

COj 1.20 

0 0.90 

N 19.80 

<  <       (< 

Total 100.00 

t(       <( 

The  number  of  B.  T.  U.  per  cubic  foot  amounting  to  917. 

The  percentage  of  nitrogen  in  the  gas  being  exceptionally 
high,  the  heating  power  is  correspondingly  reduced. 

Another  analysis  (Taylor,  Trans.  Amcr.  Inst.  Mining  Eng., 
18,  881)  is  reported  as  follows  : 

CO 0.50  per  cent. 

H... 2.18 

CH4*. 92.60 

CjH^ 0.31 

CO, 0.26 

N 3.61 

0 0.34 

n  t  t 

<(  <  ( 

<<  (I 

i  (  « I 

•  i  (< 

t  <  t  < 

Total 99.80 

£ach  cubic  foot  containing  1000.52  B.  T.  U. 

1  Pf^oceedings  Amer.  Asm.  Advancemtnt  of  Scttncf,  1&93,  p.  106. 

H  (( 



The  most  complete  investigation  regarding  the  chemical  com- 
position of  natural  gas,  has  been  made  by  Prof.  F.  C.  Phillips 
for  the  Geological  Survey  of  Pennsylvania.' 

Analysis  of  Fredonia  Natural  Gas.  (Phillips.) 

9.54  per  cent. 
0.41    *• 



C,H4,  etc> 

CO 0.00    ** 

Free  hydrogen 0.00 

NH, 0.00 

Hydrocarbons  of  paraffin  series 












Total 100.00 

•  < 


Sheffield  g&s.      Wilcox  g&».         Kane  f^aa. 

N 9.06 

CO, 0.30 

O trace 

H 0.00 

CjHi,  etc 0.00 

CO 0.00 

Paraffins 90.64 





9.79  per  cent. 
0.20    ** 











A  practical  test  of  the  fuel  value*  of  natural  gas  has  been  car- 
ried out  by  the  Westinghouse  Air-brake  Co.,  of  Pittsburgh,  Pa. 
Taking  the  usual  **  best**  quality  of  Pittsburg  coal,  it  was  found 
that  its  evaporating  duty  in  a  particular  boiler  was  10.38  pounds 
of  water  per  pound  of  the  solid  fuel.  With  the  same  boiler  1.18 
cubic  feet  of  natural  gas  evaporated  one  pound  of  water;  whence 
it  follows  that  one  pound  of  coal  is  equivalent  to  12.25  cubic  feet 
of  gas,  or  that  1000  cubic  feet  of  the  gas  were  as  good  as  8 iff 
pounds  of  coal.  According  to  calorimetric  tests,  55.4  pounds  of 
coal  contain  the  same  number  of  heat  units  as  1000  cubic  feet  of 
the  natural  gas. 

I  Report  on  the  Chemical  Composition  of  Natural  Gas,"  P.  C.  Phillips,  /.  Franklin 
Institute,  134.  24^-256,  358-375- 

^Journal  Iron  and  Steel  Inst.,  1887,  366-418.    "Fuels."  Mills  and  Rowan,  p.  39a. 




Practical  Photometry. 

The  illuminating  value  o£  any  source  of  light  is  determined  by 
comparing  it  with  some  source  of  light  of  known  value.  The 
illuminating  value  of  gas  is  measured  by  comparing  a  flame  that  is 
burning  at  the  rale  of  five  cubic  feet  an  hour  with  a  standard 
sperm  candle  that  is  burning  at  the  rate  of  120  grains  an  hour. 

The  amount  of  light  received  by  any  object  will  vary  inversely 
as  the  square  of  the  distance  of  that  object  from  the  source  of 
illumination,  hence,  if  the  light  wliose  power  is  to  be  determined 
illuminates  body  at  -V  inches  to  the  same  degree  that  a  standard 
candle  would  illuminate  that  same  body  at  y  inches,  the  illumi- 
nating power  of  that  light  will  be^^  candles. 

In  constructing  a  photometer  this  single  principle  is  kept  in 
view,  and  all  the  refinements  are  to  eliminate  errors  in  judgment 
and  to  allow  for  the  variations  in  the  rate  of  combustion  of  gas 
and  sperm. 

The  accompanying  illustration  shows  the  form  of  photometer 
known  as  the  Bunsen,  which  is  the  one  most  commonly  used  in 
Germany.  England,  and  America. 

It  consists  first  of  a  table  which  carries  the  apparatus  and 
on  which  the  distance  between  the  Hghts  is  accurately  laid  off  and 
marked  by  two  lines.  This  distance  is  generally  60  inches,  but 
2  meters  and  100  inches  are  also  used.  In  case  either  light  is 
changed  or  moved  for  any  reason,  it  may  easily  be  put  back  in 
place  by  placing  it  centrally  over  the  line  indicated  on  the  table. 
To  facilitate  the  adjustment  two  plumb-bobs  are  hung  over  each  of 
the  lines  at  the  ends  of  the  table,  so  it  is  easy  to  see  whether  the 
flames  are  properly  centered  in  one  direction.  In  the  other  direc- 
tion they  are  centered  by  sighting  along  the  bar.  The  bar  is 
placed  at  right  angles  to  the  two  lines  laid  out  on  the  table  and 
centrally  between  them.  It  is  laid  out  in  inverse  squares  50  that 
"  i"  is  in  the  center.  If  the  length  of  the  bar  is  I'and  the  dis- 
tance from  the  candle  is  A',  the  candle-power  is ^^ ■.    The 

mark  that  indicates  four  candle-power  is  twice  as  far  from  the 



light  to  be  measured  as  it  is  from  the  candle,  9  is  three  times  as 
far,  etc. 

The  bar  should  be  made  so  that  it  may  be  raised  or  lowered  at 
pleasure,  and  be  planed  to  a  thin  edge  011  top  so  that  up  light  will 
be  reflected  from  it  on  the  disk.  On  the  bar  is  a  sight-box  in 
which  a  paper  disk  is  placed  at  right  angles  to  and  centrally  over 
the  bar.  There  are  several  kinds  of  disks  used,  but  the  one  most 
commonly  preferred  in  this  country  is  made  by  taking  a  piece  of 
■white  sized  paper  of  medium  thickness,  and  cutting  out  of  the 
center  a  many-pointed  star  about  an  inch  and  a  half  in  diameter 
outside  the  points.  This  paper  with  the  star  cut  from  the  center 
is  then  placed  between  two  pieces  of  tissue  paper  and  the  three 
held  together  either  by  placing  between  pieces  of  glass  or  else  by 
being  fastened  with  thin  starch  water.  At  the  back  of  the  sight- 
box  are  two  mirrors,  so  placed  that  the  obser\-er  may  stand  in 
frout  of  the  bar  and  see  both  sides  of  the  disk.  On  the  front  of 
the  sight- box  a  hood  is  so  placed  as  to  partially  screen  the  eyes  of 
the  observer  from  the  lights. 

At  one  end  of  the  bar  is  the  light  to  be  tested.  This  is  con- 
nected to  a  pipe  sealed  in  mercury,  so  that  it  may  be  moved  back 
and  forth  or  raised  and  lowered  at  pleasure.  It  is  usually  arranged 
with  a  micrometer  cock  so  that  the  rate  of  flow  maybe  regulated 
as  closely  as  may  be  necessary. 

At  the  other  end  oi  the  bar  is  a  candle  balance.  The  balance 
is  usually  arranged  fortwo  candles  and  all  readings  are  multiplied 

I  by  2.  This  balance  is  so  constructed  that  the  position  of  the 
candles  may  be  adjusted  vertically  or  horizontally. 
This  end  of  the  bar  is  so  arranged  that  the  candle  balance  may 
be  removed  and  a  standard  burner  put  in  its  place.  The  standard 
burner  commonly  used  is  a  Sugg  Argand  burner,  size  D.  This  is 
covered  with  a  thin  sheet  metal  chimney  one  and  seven  eights 
inches  diameter.  This  chimney  has  an  opening  on  one  side  Jf 
inch  high  and  one  and  one  half  inches  wide.  On  the  opposite 
side  the  chimney  is  cut  away  to  prevent  light  being  reflected 
through  the  slot  in  front.  The  standard  burner,  like  the  one 
through  which  the  gas  is  tested,  is  so  arranged  that  it  may  be 
adjusted  in  all  directions. 
A  meter  to  measure  the  gas  is  necessary.     As  gas  is  burned 


at  the  rate  of  five  feet  an  hour  when  being  tested,  the  meter  is  s 
geared  that  one  of  the  hands  makes  a  complete  revolution  each 
time  a  twelfth  of  a  foot  of  gas  passes.  A  clock  is  attached  to 
the  meter  .with  a  large  second  hand,  so  when  the  meter  band 
mentioned  and  the  second  hand  move  together,  gas  is  passing 
at  the  rate  of  5  feet  an  hour.  In  addition  to  these  hands  are 
one  indicating  feel  and  one  minutes.  Some  meters  are  furnished 
with  a  third  set  of  hands  reading  feet  and  hundreds. 

The  meter  has  a  thermometer  to  show  the  temperature  of  the  1 
gas  and  a  universal  level  so  that  it  may  be  properly  leveled.  On  f 
the  side  is  a  glass  gauge  and  a  mark  indicating  the  height  of  the  J 
water,  which  should  always  be  constant. 

The  pipe  connections  to  the  meter  are  so  arranged  that  open-  , 
ing  a  cock  will  allow  the  gas  to  pass  around  instead  of  througfa,'| 
it.     This  permits  the  operator  to  start  or  stop  the  meter  at  pleai 
ure  without  interfering  with  the  light. 

A  pressure  gauge  connected  with  the  various  parts  of  the  ap-J 
paratus  enables  the  operator  to  ascertain  the  pressure  of  the  gai 
at  different  points.     One  of  these  connections  is  to  the  pipe  al 
short  distance  below  the  test  burner.     This  gives  the  pressun 
near  the  point  of  ignition.     The  pressure  is  read  in  inches  atu 
fractions  of  an  inch  of  water. 

A  gas  govenior  is  connected  before  the  inlet  to  the  meter»  1 
which  reduces  the  pressure  to  about  an  inch  and  a  half  of  water. 
Beyond  the  meter  is  a  smaller  governor  which  reduces  the  pres-  ' 
sure  to  about  nine-tenths  of  an  inch  and  prevents  alteration  of 
the  flow  of  gas  due  to  the  irregularities  in  the  meter.  , 

Black  screens  are  arranged  to  screen  the  eye  of  the  observer 
from  the  light.     These  are  sometimes  fixed  and  at  others  set  on  J 
the  bar.     The  latter  arrangement  is  preferable,  as  they  may  be.l 
moved  to  suit  different  positions  of  the  sight  box.  ^ 

For  testing  gas  of  not  over  eighteen  candle  power  the  Standard 
London  Argand  burner  is  used.  For  higher  candle  power  gas 
the  ordinary  sawed  lava  tip  is  best.  The  latter  is  commonly- 
known  as  the  batwing  burner. 

The  photometer  should  be  set  up  in  a  small,  light-proof  rooni>l 
with  dead  black  walls.  The  latter  can  be  hung  with  black  vel- J 
vet  or  painted  with  glue  and  lampblack.     Great  care  should  1 




taken  to  insure  proper  ventilation  without  draft.  The  tempera- 
ture of  the  room  should  be  kept  as  near  60°  F.  as  possible,  and 
the  air  should  not  be  allowed  to  become  vitiated  by  the  products 
of  combustion.  The  table  should  be  set  so  that  readings  may- 
be taken  from  both  sides  of  the  bar. 


When  one  starts  to  use  a  new  photometer,  or  one  with  which 
the  experimenter  has  not  previously  worked,  the  instrument 
should  be  carefully  verified. 

First,  make  sure  that  the  lines  defining  the  distance  between 
the  lights  are  the  proper  distance  apart  and  parallel,  and  that  the 
bar  is  perpendicular  to  andmidway  between  them.  Next  see  that 
the  bar  is  level.  The  disk  must  be  at  right  angles  to  the  bar, 
and  the  small  pointer  under  the  sight  box  in  line  with  the  disk. 
The  two  mirrors  should  be  made  of  the  best  plate  glass  and  well 
silvered.  They  should  be  kept  clean.  The  disk  should  exactly 
bisect  the  angle  made  by  the  mirrors.  The  bar  should  be  veri- 
fied so  that  the  operator  may  be  sure  that  it  is  properly  divided, 
and  the  meter  should  be  tested  with  a  meter  prover.  Iii  testing 
the  meter  be  sure  that  the  temperature  of  the  room,  of  the  water 
in  the  meter,  and  of  the  water  in  the  prover  are  the  same.  The 
pressure  gauge  should  be  verified  by  a  U-shaped  water  gauge.  The 
knife  edges  of  the  candle  balance  should  be  clean  and  sharp,  and 
the  lever  should  be  free  to  move  without  rubbing.  The  weight 
for  the  candle  balance  should  be  weighed  on  an  analytical  bal- 
ance to  be  sure  that  it  is  correct. 

For  testing  coal  gas  no  choice  is  allowed  in  the  burner,  but 
when  water  gas  or  any  high  grade  gas  is  to  be  tested  it  is  neces- 
sary to  get  a  burner  suited  to  the  gas.  The  most  suitable  burner 
can  be  quickly  determined  by  experiment,  and  the  greatest  effi- 
ciency is  usually  obtained  with  a  burner  of  such  size  that  the  gas 
is  almost  on  the  point  of  smoking.  When  the  photometer  light 
is  burned  continually,  as  is  usually  the  case  in  gas  works,  the 
tip  on  the  flat-flame  burner  should  be  changed  at  inter^-als  of  two 
or  three  weeks.  Care  should  be  taken  that  the  tip  is  smooth. 
Any  tips  that  are  chipped  on  top  or  rough  in  the  slot  should  not 
be  used. 



In  preparing  for  a  test,  the  burner  and  candles  should  be 
placed  in  their  proper  positions  and  at  such  a  height  that  the 
center  of  the  flames  will  be  on  a  level  with  the  center  of  the 
disk.  The  height  of  the  candle  flame  is  taken  when  the  candle 
end  of  the  balance  is  down.  The  gas  should  be  burned  long 
enough  to  be  sure  that  the  apparatus  is  cleaned  out  and  that 
fresh  gas  is  being  burned.  Before  starting  it  isnecessar>' to  con- 
trol the  pressure  under  the  burner  so  that  it  will  not  vary  during 
the  test.  The  governor  on  the  outlet  of  the  meter  will  do  this  il 
it  is  in  order.  Ifthepressurevaries,  the  governor  must  be  cleaned 
before  starting  the  test.  During  th'e  test  the  pressure  gauge 
must  be  shut  off,  as  in  case  there  is  change  of  pressure  it  will 
store  or  give  out  enough  gas  to  vitiate  the  result.  The  meter 
should  be  level  and  the  water  at  the  proper  height. 

The  wicks  of  the  candles  should  never  be  touched.  Thi 
candies  are  lighted  and  allowed  to  burn  until  the  wick  curls 
to  theedgeoi  the  flame  and  burns  away  as  the  candle  is  consumedt' 
The  end  of  the  wick  should  glow.  No  test  should  be  started^ 
until  the  wicks  are  bent  over  and  the  ends  are  glowing.  Th«] 
candles  should  always  be  burned  eight  or  ten  minutes  beforei 
starting  a  test.  A  common  practice  which  gives  good  results  is) 
to  allow  the  candles  to  burn  eight  or  ten  minutes  and  then  ex- 
tinguish them  for  two  or  three  minutes.  The  candles  are  then 
relighted  and  allowed  to  burn  about  two  minutes  before  starting 
the  test.  They  are  commonly  placed  in  the  holders  in  such  a 
way  that  the  ends  of  the  wicks  are  as  far  away  from  each  otheras 

When  the  apparatus  has  been  brought  to  the  proper  condition' 
for  testing,  the  flow  of  gas  is  adjusted  to  as  near  five  feet  an 
hour  as  possible,  and  the  meter  is  allowed  to  run  until  the  twelfth 
of  a  foot  hand  points  Co  o,  when  it  is  by-passed.  The  clock  is 
stopped  at  o.  The  candles  are  counterbalanced  by  the  sliding 
weight  ou  the  balance  lever  until  the  weight  almost  carries  the 
lever  down.  In  a  few  seconds  the  candles  burn  sufficiently  to 
allow  the  balance  to  fall,  and  at  that  instant  the  meter  and  clock 
should  be  started.  As  soou  after  as  possible  the  40-grain  w 
should  be  dropped  into  the  scale  pan,  which  brings  the  candli 
down  again.     The  operator  should  always  move  about  the  rooi 




■deliberately  so  as  to  avoid  as  far  as  possible  creating  currents  of 
air.  The  candle  flames  roust  be  still  before  begiuniiig  to  talte 

A  reading  should  be  taken  every  minute  for  ten  niinutes. 
When  the  screen  is  apparently  illuminated  equally  on  both  sides 
it  should  be  moved  a  little  to  the  right  and  to  the  left,  and  in 
each  case  the  illumination  on  that  side  should  increase.  Five 
readings  should  be  taken  on  one  side  of  the  bar  and  the  sight- 
box  turned  around  and  five  taken  from  the  opposite  side.  In 
case  the  barisaccessible  from onlyone  side,  the  readings  should 
be  made  with  one  eye  and  the  screen  turned  in  the  sight-box 
after  half  have  been  completed.  This  wilt  eliminate  the  errors 
due  to  possible  difference  in  eyes  and  in  the  sides  of  the  screen. 

The  last  reading  shonld  be  taken  during  the  first  half  of  the 
tenth  minute  and  the  times  noted  when  the  candle  balance  falls, 
atid  when  the  gas  hand  completes  its  tenth  revolution.  The  tem- 
perature of  the  gas  and  the  reading  of  barometer  should  also 
be  noted.  After  this  the  candles  may  be  extinguished.  They 
should  be  blown  out  and  the  ends  of  the  wicks  extinguished  with 
a  piece  of  sperm.  The  wicks  should  never  be  touched  with  any- 
thing else. 

If  the  candle  balance  falls  in  less  than  gj-  or  more  than  10^ 
minutes,  or,  if  the  gas  hand  takes  less  than  9I  or  more  than  loj^ 
minutes  to  make  lo  revolutions,  the  test  should  be  discarded. 
Long  practice  has  shown  that  within  these  limits  the  light  given 
by  the  candles  varies  approximately  with  the  consumption  of 
sperm  and  that  given  by  the  burner  approximately  with  the  gas 

If  the  candles  take  X  seconds  to  burn  40  grains  and  the  gas 
hand   V  seconds    to   make   ro   revolutions,  the    average  read- 

ingmultiplied  by  2  should  be  multiplied  by  /_   X  — y-  or-y. 

This  will  give  the  candle-power  of  the  gas  uncorrected  for  tem- 
perature and  pressure. 

The  standard  of  pressure  is  30  inches  of  mercury  and  the 

standard  temperature  is  60°  Fahr.   To  correct  the  pressure  multiply 

I  by  30  and  divide  by  the  barometric  reading.     In  correcting  for 


temperature  the  gas  is  assumed  to  be  a  perfect  gas  saturated  with 
water- vapor.  The  following  is  the  formula  for  correction  for 
pressure  and  temperature  : 

_  17.64  {h  —  a) 
'""       460  +  / 
71  =.  the  number  by  which  the  observed  volume  is  to  be 

multiplied  to  reduce  it  to  30  inches  and  60°. 
h  =  the  height  of  the  barometer  in  inches. 
/=the  temperature  Fahrenheit. 
a  =  the  tension  of  aqueous  vapor  at  /**. 

The  table  on  the  opposite  page  will  facilitate  corrections  for 
various  pressures  and  temperatures. 

Inasmuch  as  a  flame  is  nof  perfectly  transparent,  a  test  made 
with  it  at  right  angles  to  the  bar  does  not  give  the  mean  of  the 
light  that  is  emitted  horizontally.  The  richer  the  gas  the  greater 
is  the  difference  between  the  candle-power  measured  on  the  flat 
and  on  the  edge  of  the  flame .  A  gas  that  gives  25  candles  measure- 
ment flat  will  not  give  over  19.5  candles  measured  on  the  edge. 
When  the  flame  is  at  an  angle  of  10  degrees  with  the  bar  it  gives 
almost  as  much  light  as  when  it  is  measured  at  90  degrees. 

The  best  photometers  are  made  so  that  the  burner  may  be 
turned  on  its  axis  and  the  light  measured  at  all  angles. 

When  it  is  desired  to  measure  the  light  emitted  by  a  burner  at 
various  altitudes,  mirrors  are  used  to  reflect  the  light  to  the  disk 
as  the  latter  is  kept  vertical  and  in  the  same  horizontal  plane  as 
the  standard  burner.  In  such  cases  it  is  necessary  to  test  very 
carefully  the  amount  of  light  absorbed  by  the  mirrors  at  all  angles. 

There  is  a  popular  impression  that  photometrical  work  is  not 
accurate  and  therefore  not  to  be  depended  upon,  but,  if  care  is 
taken  by  the  operator  in  his  work,  and  all  the  apparatus  is 
properly  adjusted,  the  error  will  be  less  than  i  per  cent.  By 
taking  the  average  of  a  series  of  measurements  the  error  can  be 
reduced  to  a  point  where  it  is  inappreciable. 



M  p  p  p  p  pppppp>p^^S^«c«^S««^  pooooo9opoaooopcoooo 
b  ^  *0D^  <>cn  <jk  u*  V»  M  b  4?  co^  b>6i  ^  dt  M  M  b  >b  bc*^  9>u*  i^i»»  It  ''*  o 


Vt   M  00&  M  ^  ^  O  ^  &  O  0^0*  «  0«<«>  s5  Ov  w  ^  (A  kS  ODCn  M  00^  M  vj  4L  S 


H5|«55<2  S  8  8  2  2  2  8  8  8«|:S^M^^^1j^^ 


fi'%^£gaaa  s  s  s  2  2  2 « 8«««M^^^^^^^««JS 

&•  »  OO^  M^4^  O^^  O^C*t  O  0^(*<  "OO^MvO^M^Cyi  m  5c(/i  m  ooOi  m 


^^^555  S  8  S  2  2  2  8  8  8«««^f  ^:SiS5^?«^«««:e 


ii§mi  ill  ill  §iJitfii§H§ig?ii« 

p  p  p  p  p  p  o  p  p  o  o  o  o  o  o  o  o  o  o 

ilU  l§ia  sll  §lllt§HS?^HI«SI? 


mi  §22  III  §s||fgHaf l^tSlill 



li  §  1 2  2  y  §^ii^|!|§si:^i^siiisgii 


'  is  S  2  2  2  8  8  8«««itM^i*^§^^^^i^^^^t'S 


Cv^nC  9>0*^  OvUt^  QSO*^  OkM'f  O^M'd  9vwO  9tW«0  9^M^O  9^*>>C  9>W 



'—  C  '* 








p  p  p  p  p  p  p  p  p  p  p  p  p  o  p  p  p  p  p  p  p  p  p  p  p  c  c  p  c  p  c 

9k(**  O  $Oi  00«C**  O  ^'^  O  "^  («»O^U«0^i*«C'^<^0-''''»C->-'<C£-^  •* 


'';  I 














t:5  o 




^  < 

^  o 





^  > 












Hartley's  Calorimeter  for  Combustible  Gases. 

The  conditions  of  use  are  as  follows  :  From  a  small  cistern  A  ' 

(Fig.  S6)  water  flows  over  a  sensitive  thermometer^  thence  into 

a  case  surrounding  the  stem  of  a  suitable  burner  C.  onwards  to 

a  metal  casing  or  jacket  enveloping  the  calorimeter  D,  then 

makes  it  way  to  the  upper  part  of  the  latter,  descends  aftei 
traversing  a  series  of  shelves  which  present  a  very  large  surface,  fl 
and  finally  passes  out  to  the  collecting  tank/,  at  the  base  of  thel 
instrument.     The  burner  is  passed  upwards  into  the  center  of  fl 
cylindrical  chamber  at  the  bottom  of  the  calorimeter ;  and,  ; 
already  stated,  the  burner  stem  is  surrounded  by  a  casing  tbrougli 


which  the  supply  water  flows.  Loss  by  radiation  from  the  bur- 
ner is  thus  prevented. 

The  gas  is  measured  by  a  special  meter  M:  and  its  rate  of 
consumption  and  the  rate  of  water  flow  are  regulated  until  the 
issuing  water  is  found  to  be  a  tew  degrees  higher  in  temperature 
than  that  at  which  it  enters  :  and  the  temperatures,  as  indicated 
by  the  four  thermometers  employed,  are  found  to  be  steady. 
During  these  adjustments,  water  runs  to  waste  through  a  by- 
way cock.  When  all  is  ready,  and  with  the  meter  index,  the 
bye-way  cock  is  instantly  turned,  and  the  passage  of  the  out- 
flowing water  diverted  to  a  collecting  tank.  The  quantity  of  gas 
usually  burned  per  experiment  is  one-fourth  cubic  foot ;  and  the 
time  occupied  with  ordinary  coal  gas  leu  to  twelve  minutes. 
During  the  experiment,  the  temperatures  of  the  inlet  and  outlet 
water  should  be  frequently  observed,  and  now  and  Iheii  also  the 
temperature  of  the  jacket.  When  the  desired  quantity  of  gas 
has  been  burned,  the  water  flow  is  promptly  turned  to  waste. 

The  collected  water  is  next  measured,  and  its  weight  calculated, 
or  better,  weighed  directly,  The  weight  in  poundsmiiltipliedby 

the  number  oi  degrees  the  water  has  been  raised  gives  the  heat- 
ing power  due  to  the  quantity  of  gas  burned. 

Thus,  if  twenty  pounds  of  water  have  been  raised  8°  F,  by  one- 
fourth  cubic  foot  of  gas,  we  have  20  X  8  —  160  pounds,  Fahr. 
units  for  one-fourth  cubic  foot,  or  640  for  one  cubic  foot. 

Kote. — The  products  of  combustion  are  so  completely  reduced 

I  to  the  temperature  of  the  inflowing  water  that,  without  aspira- 

I  tion.  they  would  not  rise  through  the  instrument.     The  aspirator 

I   is  simply  a  copper  chimney  F,  heated  at  its  upper  part  by  a  ring 

gas-burner  G. 

Prof.  E.  G.  Love,  School  of  Mines   Quarterly,   13,  97,  gives 
the  result  of  an  analysis,  with  this  instrument,  of  a  sample  of  the 
aicipal  Gas  Co.  gas.  of  New  York  City,  as  follows: 

narometer 19.886     in. 

Temptrature  of  gas  burned 66.00°   F. 

Temperature  of  the  air  (a) 65-98""    " 

Temperalureof  the  water,  iolet  (b) 61.605°  " 

Temperature  of  the  water,  outlet  (c) 69.275°  " 

Temperature  of  the  water,  raised 7.670^  " 



Temperature  of  the  **  body"  (d) 63.795^'  F. 

Temperature  of  the  escaping  gases 64 .43^    ' ' 

Duration  of  test t 12.78  minutes. 

Gas  burned 0.25  cubic  feet 

Gas  burned,  corrected  to  60*^  F.  and  30  in  Bar- . . .     0.2452  **       ** 

Pounds  of  water  heated 23.228    **       ** 


(a—d)  2.185  X  0.025  X  1278  =  0.698°  gain. 
{c — a)  3.295  X  o.oi  X  12.78  ^0.421*^  loss. 

0.2770°  gain. 

23.228  -f-  7.67 — 0.28  =  177.88  -*-  0.2452  =s  725.3  heat  units  at  60*-  F.  and 
thirty  inches  barometer. 

The  coal  gas  of  London,  Eng.,  with  an  illuminating  power  of 
sixteen  to  seventeen  candles,  has  a  colorific  power  of  about  668 
B.  T.  U.  per  cubic  foot,  and  costs  from  sixty  to  seventy  cents 
per  thousand  cubic  feet. 

The  average  of  numerous  tests,  made  with  the  Hartley  calori- 
meter, upon  the  New  York  City  water  gas,  gives  710.5  B.  T.  U. 
per  cubic  foot.  One  thousand  cubic  feet  of  this  gas,  costing 
$1.25,  would  therefore  yield  710,500  heat-units,  which  would  be 
equivalent  to  568,400  B.  T.  U.  for  $1.00. 

^F                                JUNEER'S  GAS 

CALORIMETER.                                  287             ^H 

W                                                  XXXIV.                                                               U 

H                               Junker's  Gas  Calorimeter.                                        ^M 

H    Another  form  of  gas  calorimeter  is  the  Junker,  (Fig.  87).'              ^| 

H     The  above  sectional  drawing 

shows  the  instrument  to  consist          ■ 

■  1.  Cold  water  inlet. 


^    2,  Strainer. 


ft      1                  m 

3.  Overflow  to  calorimeter. 


i     !                 m 

4.  Upper  container. 


f                ■ 

5-  Waste  overflow. 

1  ^ 

i                H 

■    6  and  7.  Fall  pipe  and  joint. 

-,      i                ■ 

■  8.  Drain  cock. 

n(TteJi=a              H 

H'9-  Adjustment  cock. 

5  ti 

B^ii      ■ 

1     W.  Cold  water  t be rnio meter. 

-  ffa'      ■ 

rj.  Air  jacket. 

14.  Perforated  spreading  ring. 
15  and  16.  Water  jacket. 
17.   Baffle  plates  with  cross  slots. 
iS-  Lower  overflow. 


^  ^^7% 




_ii  h 

15.  Lower  container. 

30,  Hot  water  overflow. 



,2.  Gas  nipple. 


13.  Air  supply  regulator. 




14.  Gas  nozzle. 


15.  Clamp  for  burner. 


3b.  Burner  holder. 

J  US 

\'     1 

17.  Burning  cap. 




l8.  Combustion  chamber. 

»9.  Roof  of  combustion  chamber. 



30.  Cooling  tubes. 



31.   Receiver  for  combustion  gases. 

f   K^ 



31.   Outlet  for  combustion  gases. 

rJ4  ^^ 



33.  Throttle  for 

34.  Brass  base  tiug. 



35.  Condensed  water  outlet. 

«^          ■ 



37.  Utr  jacket. 

/                       1 

39.  Test  hole  in  air  jacket. 


43.  Hot  water  thermometer. 


ofacombustionchambersurroundedby  a  water  jacket,  the  latter          | 

filled  with  a  great  many  tubes 

To  prevent  loss  by  radiation         ^1 

^k        ly.  Sof.  Cktm.  /-d  /kJv,  M9J,  6ji. 



the  water  jacket  is  surrounded  by  a  closed  air  space.  The  whole 
apparatus  is  constructed  of  copper  as  thin  as  is  compatable  with 
strength.  The  water  enters  the  water  jacket  at  the  bottom,  and 
leaves  it  at  the  top,  while  the  hot  combustion  gases  of  the  Hame 
of  the  gas  that  is  on  trial  enter  the  tubes  at  the  top  and  leave 
them  at  the  bottom.  There  is  therefore  not  only  a  very- 
large  surface  of  thiu  copper  between  the  gases  and  the  water, 
but  the  two  move  in  opposite  directions,  during  which  process 
all  the  heat  generated  by  the  flame  is  transferred  to  the  water, 
and  the  water  gases  leave  the  apparatus  approximately  at  atmos- 
pheric temperature.  The  gas  to  be  burned  is  first  passed  through 
a  meter,  and  then  to  insure  constant  pressure,  through  a  pressure 
regulator.  The  source  of  heal  in  relation  to  the  unit  of  time  is 
thus  rendered  stationary,  and,  in  order  to  make  the  absorbing 
quantity  of  heat  also  stationary,  two  overflows  are  provided  at 
the  calorimeter,  making  the  head  of  the  water  and  the  rate  of 
flow  ofthesame  constant.  The  temperaturesof  the  water  entering 
and  leaving  the  appratus  can  be  read  at  the  respective  thermome- 
ters ;  as  shown  before,  the  quantities  of  heat  and  water  passed' 
through  the  apparatus  are  constant.  As  soon  as  the  flame 
lighted  the  temperature  of  the  exit  thermometer  will  rise  to 
certain  point  and  will  nearly  remain  there.  All  data  for  ascertai! 
ing  the  heat  given  out  by  the  flame  are  therefore  available 

All  that  is  required  is  to  measure  simultaneously  the  quantil 
of  gas  burned  and  the  quantity  of  water  pressed,  and  the  diffei 
ence  in  temperature  between  the  entering  and  leaving  water,^ 
Centigrade  thermometers  and  two-liter  flasks  are  required. 

The  meter  shows  one-tenth  of  a  cubic  foot  per  revolutioi 
the  large  hand ;  the  circumference  being  divided  into  loo  parts, 
so  that  o.ooi  can  be  read  accurately.  The  water  supply 
regulated  that  the  overflow  is  working  freely,  and  the  water- 
admission  cock  is  set  to  allow  two  liters  of  water  to  pass  in  about 
a  minute  and  a  half.  The  colorimeter  is  now  ready  to  take  the 
reading.  The  cold  water  as  a  rule  has  a  sufficiently  constant 
temperature  that  we  note  it  only  once  ;  it  is  now  17.2°  C. 
soon  as  the  large  index  o(  the  meter  passes  zero,  note  the 
of  the  meter  and  at  the  same  time  transfer  the  hot-water  tul 
from  the  funnel  into  the  measure  glass,  and  while  that  is  beinj 



filled  noie  llie  temperature  of  the  hot  water  at  say  ten  inten-als. 
to  draw  the  average. 

The  temperatures  are  43.8,  43.5,,43,9,43.8, 
43'7-  43-8,  and  43.7,  making  the  average  43. S. 

The  measure  glass  is  now  filled ;  turn  the  gas  out.  Find  from 
the  readings  of  the  meter  at  the  beginning  and  the  end  of  the 
experiment  that  there  was  burned  0.35  cubic  foot,  by  means  of 
which  the  temperature  of  the  two  liters  of  water  was  raised 
26.6'  C-,  viz.,  43.8°— 17.2°  =  26.6''C.  The  calculation  is  as  fol- 
lows : 

where  H=-  the  calorific  value  of  one  cubic  foot  of  gas  in  calories. 

W^=  the  quantity  in  liters  of  the  water  heated. 

T^  the  difference  in  temperature  between  the  two  thermome- 
ters in  degrees  C,  and  C^^the  quantity  in  cubic  feet  of  gas 
used, then 

^_2X26.6_  ^^^  calories  or  604  (152  X3. 968)    "  B.  T.  IT." 

per  cubic  foot. 

It  is  mentioned  before  that  the  effect  of  the  cooling  water  is 
such  that  the  waste  gases  leave  the  calorimeter  at  about  atmos- 
pheric temperature.  All  hydrocarbons  when  burned  form  a 
considerable  quantity  of  water,  which  in  all  industrial  processes 
escapes  with  the  waste  gases  as  steam.  The  latent  beat  of  this 
steam  is  therefore  not  utilized  when  fireing  a  stove  or  driving  an 
engine  with  gas  ;  in  the  above  result,  however,  the  latent  heal 
is  included,  because  in  the  copper  tubes  the  steam  is  condensed, 
and  its  heat  is  transferred  to  the  circulating  water  and  measured 
with  the  rest.  The  condensed  water  runs  down  the  tubes  which 
are  cut  off  obliquely  to  allow  the  drops  to  fall  off  easily,  and  is 
collected  in  the  lower  part  of  the  apparatus  from  where  it  runs 
through  the  little  tube  into  a  measure  glass.  In  condensing 
steam  gives  oS  six-tenths  calorie  for  every  cubic  centimeter  of 
water  formed.  If  therefore  a  graduated  (cc. )  cylinder  be  placed 
under  the  little  tube  the  amount  of  water  generated  by  burning 
say  one  cubic, foot  of  gas.  can  be  directly  measured. 

From  burning  one  cubic  foot  of  gas,  we  have  collected  27.25  cc. 



of  condensed  water,  and  must  therefore  deduct  16.35  calories  ] 
from  the  gross  value  found  above,  which  gives  the  net  calorific  1 
value  of  the  gas  tested  as  135.65  calories  or  538  B.  T.  U.  per  I 
cubic  foot. 

The  calorimeter  is  placed  so  that  one  operator  can  simulta-  . 
neously  obsen-e  the  two  thermometers  of  the  entering  and  e 
pingwatcr.  the  index  of  the  gas-meter,  and  the  measuring  glasses.  J 

No  draught  of  air  must  be  permitted  to  strike  the  exhaust  of  J 

_  the  spent  gas. 

junker's  gas  calorimeter.  291 

The  water  supply  tube  is  connected  to  the  nipple  in  the 
center  o£  the  upper  container  ;  the  other  nipple  is  provided  with 
a  waste  tube  to  carry  away  the  overflow.  This  overflow  must 
be  kept  running  while  the  readings  are  being  taken. 

The  nipple,  through  which  the  heated  water  leaves  the  calor- 
imeter, is  connected  by  an  india-rubber  pipe  with  the  large 
measure  glass,  and  the  water  must  be  there  collected  without 

The  smaller  measure  glass  is  placed  under  the  tube  to  collect 
any  condensed  water. 

After  the  thermometers  have  been  placed  in  position  with  their 
india-rubber  plugs,  the  water  supply  is  turned  on  by  the  cock, 
and  the  calorimeter  filled  with  water  until  it  begins  to  discharge. 
No  water  must  at  this  period  exude  from  the  smaller  pipe 
or  from  the  test  hole  under  the  air  jacket,  otherwise  this  would 
prove  the  calorimeter  to  be  leaking. 

Tablb  of  Resume  of  Tests  upon  London  Coal  Gas. 


:3a  i 

"^■2    I 



First  day.,  al.o''  15-33^  »6.II3  10.790.0407  ...  35.7  165.3  '5-4  149-9 
Second"  --  a».s°  IJ.9  37.68  14.780.0584  ...  17.4  165.9  '^-4  148.5 
Third  "  ••  17.5^  13.71  a8.6  14.89  0.1103  '7-5  *S-43  'M-8  15.86  148.94 
Ponrtb  "  ..  iT-S"^  '3-75    J8.53    14.780,1103  17.4  26.43  ">S-6  15-86  149.74 

Experiments    made   with    this    calorimeter  at   the   Stevens 
Institute,  are  recorded  in  the  Sf evens  Indicator,  October,  i8g6. 

The  gas  used  was   carburetted  water  gas  "  Lowe  Process" 
composed  as  follows  : 

. . .     J. 20  per  cent,  (by  volume). 



CH. 17.83 

H -.--■   37-95 



The  theoretical  heating  value  of  this  gas,  is  662  B.  T.  U. 
per  cubic  foot. 

The  heating  value  as  determined  with  the  Junker  calorimeter 
is  668  B.  T.  U.  per  cubic  foot. 

Liquid  Fuel. 

Petroleum  containing  eighty-six  per  cent,  of  carbon  has  an  evap- 
orative power,  as  estimated  by  Storer,  of  eighteen  pounds  of  water 
per  pound  of  petroleum.  Deville  has  determined  the  heating 
power  of  various  petroleums,  by  calorimetric  tests,  with  the  fol- 
lowing results : 

Heavy  oil  from  West  Virginia 10180  calories  per  kilo. 

Light    •*      **        **  " 10223 

Heavy"      **      Ohio I0399 

Light   "      "      Penn 9963 

Petroleum  from  Java 10831 

Petroleum  from  Alsace 10458 

Petroleum  from  £.  Galacia 10005 

Petroleum  from  W.  Galacia 10235 

Crude  shale  oil  from  Autun  (Prance)  9950 

Dr.  Paul  estimates  the  evaporative  power  of  liquid  hydrocar- 
bons as  the  sum  of  the  carbon  and  hydrogen  present,  on  the 
basis  that  when  oxidized  with  the  theoretical  proportion  of  air, 
each  pound  of  carbon  evaporates  11.359  pounds  of  water  at 
15.5°  C,  and  each  pound  of  hydrogen  41.895  pounds  of  water  at 
15.5°  C,  into  steam  at  100°  C.  The  following.table  gives  the  re- 
sults obtained  : 


























Hydrogen.  Oxygen. 


tion  power 

in  pounds 

of  water 

at  100*  C. 

tion  dutv 
of  water 
at  15.5*  C. 

C«HeO  (Phenol)....  76.6 





CtH^O  (Cresol)....  77.77 




II. 16 

C,oH„  (Naphthalin).  93.75 


•  •  •  • 



C^Hio  (Anthracine)  94.38 


•  •  •  • 



C\Hio  (Xylol) 90.56 


•  •  •  • 



C»H,2  (Cumol) 90.00 


•  •  •  • 



C,oHu(Cymol) 89.55 


•  •  •  • 



The  effective  heat  he  calculates  as  follows,  using  twice  the 
amount  of  ^air  required  by  theory  for  the  combustion. 

Combustion  of  One  Pound  of  Carbon. 


B.  T.  U. 
Heat  units. 

Total  heat  of  combustion 

Available  heat 

Waste  of  furnace  gases  at  315^  C 



Equivalent  evaporation 

of  water. 
At  loo'  C.     At  15.5'  C. 

15.0  lbs.       . .  lbs. 

3-6  " 


Effective  heat 1 1020 


i  ( 

9.8  ** 

Combustion  of  One  Pound  of  Hydrogen. 

Equivalent  evaporation 
of  water. 
B.T.  U.         Atioo'C.      At  15.5*  C. 

Total  heat  of  combustion 62032  64.2  lbs.      . .  lbs. 

Latent  heat  of  water  vapor 8695 



Available  heat 

Waste  heat  of  furnace  gases 


II.9       M 

43.3   " 


38   " 

Effective  heat 41817 

The  effective  heat  of  two  hydrocarbons  (containing  respectively 
carbon  eightj'-six  per  cent.,  hydrogen  fourteen  per  cent.,  and 
carbon  seventy-five  per  cent.,  hydrogen  twenty-five  per  cent.) 
are  thus  tabulated  : 

Hydrogen  Containing  Carbon   Eighty-Six  Per  Cent.,  Hydrogen 

Fourteen  Per  Cent. 

Total  heat 
of  com- 

c  =  0.86  X  14500 

H=o.l4  X  62032 

Equivalent  evaporation 
of  water. 
At  loo'  C.    At  15.5*  C. 

12470  B.  T.  U. 


Furnace  gases. 

CO, 3.16  lbs. 

Water  vapor 1.26    " 

N 11.45    " 

Surplus  air 14.37    " 

2 1 154 

Heat  units  in  fur- 
nace gases. 

411   B.T.  U. 



21.9  lbs.     18.8  lbs. 

1 1 






Total  heat  of  combustion 2 11 54 

Latent  heat  of  water  vapor 1217 

Available  heat 

Waste  in  furnace  gases 

Effective  heat 

Theoretical  evaporating  power 



<  ( 

i  t 


i  t 

'  < 

« ( 

2.2  lbs. 

<  t 

1.3  lbs. 


( t 

t  ( 


13.6  lbs. 



Hydrocarbon  Containing  Carbon  Sevknty-Fivb  Per  Cent.,  Hydro- 


0.75  X  14500 
0.25  X  62032 

GEN  Twenty-Five  Per  Cent. 

Total  heat 
of  com- 

EqniTalent  evaporation 
of  water. 
At  loo*  C.    At  15-5*  C 

10775  B.  T.  U. 


Furnace  gases. 


Water  vapor 


Surplus  air  • 


Heat  units  in  fur- 

nace gases. 

2.75  lbs. 

358  B.  T.  U. 

2.25    ** 


13.39    ** 


17.39    *• 

2483        ** 

27.1  lbs.    23.1  lbs. 

2.6  lbs. 

Total  heat  of  combustion . .  •  • 
Latent  heat  of  water  vapor* . 

Available  heat • 

Waste  in  furnace  gases 

Effective  heat 








2.2  lbs. 

Theoretical  evaporating  power 







16.5  lbs. 

The  theoretical  evaporative  efficiency  of  different  combustibles 
is  estimated  by  Rankine  from  their  chemical  composition  as  fol- 

-£*=  15C  +  64H — 80,  and  to  calculate  the  quantity  of  air 
required  for  combustion,  ^=i2C  +  36H  —  4|0,  from  which 
the  following  table  is  derived. 

Chemical  composition. 

Description  of  fuel.       C. 
Charcoal 0.93 


Rock  Oils  I 









Wood,  dry 0.58 



Ethylene,  CjH^.- 
Acetylene,  C4H2 
Peat,  dry 
















14. 1 



Evaporation  due  to 











Rankine  adopts  as  his  unit,  the  weight  of  fuel  required  to 
evaporate  one  pound  of  water  at  ioo°  C,  under  a  pressure  of  14.7 
pounds  per  square  inch  this  being  equivalent  to  966  B.  T.  U. 
The  results  were  reduced  as  follows  : 

Let  E  be  the  corrected  and  reduced  evaporation. 
e  =  the  weight  of  water  evaporated. 
7",  —  the  standard  boiling  point  (212  F.). 
7",  ^  the  temperature  of  the  feed  water. 
7"^  =  the  actual  boiling  point  observed  :  then 




7^— r,-t-o.3  (7;-7-,) 

966  F. 

Loss  by  chimney  = 

-  T.  (F'.) 

This  represents  the  number  of  times  its  own  weight  of  water 

that  a  fuel  would  evaporate  if  there  were  no  waste  of  heat,  as 

however  there  is  always  a  loss  of  heat,  the  efiBciency  of  a  furnance 

, ,  .     ^(available) 
would  be  — f-  ,.  ,  ,, — 
E  (total) 

The  loss  of  units  of  evaporation  by  waste  gases  Rankine  gives ; 

+  -^' 


where  \'\-  A'  equals  the  weight  of  burnt  gas  per  unit  of  fuel  and 
7"c  (F)  the  temperature  of  the  chimney  gases  above  that  of  the 

For  ordinary  coal  i  +  A'  ranges  from  thirteen  to  twenty-five, 
and  hydrocarbon  oils  it  is  16.3  if  no  excess  of  air  is  necessary 
above  what  is  required  for  the  combustion  of  the  fuel. 

Rankine  gives  the  theoretical  evaporative  power  of  hydrogen 
and  carbon  as  follows  : 


of  wcieht. 

Carbon,  solid  (charcoal)-.-  if 
Carbon  gas  in  3j  parts  CO..  \\ 
Carbon,  gaseous 3j 

In  1892.  from  tests  made  for  the  Engineer's  Club  of  Philadel- 
phia, the  relative  heating  value  of  coal,  gas  and  petroleum  are 
thus  stated : 


Lba.  of  wat«T.  ffom      ' 

■  uditii^F.  J 

I  lb,  anthracite  coal  evaporated 9.7  I 

1  ■■    oil  36- B 16.48  I 

I  cubic  foot  gas,  ao  C.  P i.aS  I 

E.  C.  Potter,  (Trans.  Am.  Inst.  Mining  Engineers,  Vol-  xvii,'! 
p.  807),  states  results  of  tests,  at  South  Chicago  Steel  Works^l 
of  heating  value  of  petroleum  and  block  coal,  as  follows;  M 

With  coal,  fourteen  tubular  boilers,  sixteen  feet  by  five  feet,. I 
required  twenty-five  men  to  operate  them:  with  fuel  oil,  six 4 
men  were  required,  a  saving  of  nineteen  men  at  $2.00  per  day 
or  $38.00  per  day.  For  one  week's  work  2,731  barrels  of  oil 
were  used,  against  848  tons  of  coal  required  for  the  same  work. 
With  oil  at  sixty  cents  per  barrel  and  coal  at  $2.15  per  ton.  the 
relative  cost  of  oil  to  coal  is  as  $1.93  to  $2.15. 

Valuation  of  Coal  for  the  Production  of  Gas. 
Take  100  grams  of  the  coal  in  small  lumps,  so  that  they  may 
be  readily  introduced  into  a  rather  wide  combustion  tube.    This 
is  drawn  out  at  its  open  eud  (after  the  coal  has  been  put  in)  SO 
as  to  form  a  narrow  tube,  which  is  to  be  bent  at  right  angles; 
this  narrower  open  end  is  to  be  placed  in   a   wider  glass  tube, 
fitted  tight  into  a  cork   fastened  into  the  neck  of  a   somewhatj 
wide-mouthed  bottle  ser\'ing  as  tar  vessel.     The  cork  alluded  to| 
is  perforated  with  another  opening  wherein  is  fixed  a  glass  tub« 
bent  at  right  angles,  for  conveying  the  gas.  first  through  a  cal'j 
cium  chloride  tube,   next  through  Liebig's  potash  bulbs  cow 
taining  a  solution  of  caustic  potash,  having  lead  oxide  dissolve^ 
in  it.     Next  follows  another  tube  partially  filled  with  drj-  i 
tic  potash  and  partly  with  calcium  chloride ;  from  this  last  tube  " 
8  gas-deliverj-  tube  leads  to  a  graduated  glass  jar  standing  over 
a  pneumatic  trough,  and  acting  as  gas-holder.     Before  the  igni- 
tion of  the  tube  containing  the  coal  is  proceeded  with,  all  the  1 
portions  of  the  apparatus  are  carefully  weighed  and  next  joined! 
by  means  of  india-rubber  tubing.     After  the  combustion  is  fin- I 
ished,  which  should  be  carefully  conducted  so  as  to  prevent  the! 




bursting  or  blowing  out  of  the  tube,  the  difierent  pieces  of  the 
apparatus  are  disconnected  and  weighed  again.  The  combus- 
tion tube  has  to  be  weighed  with  the  coal  after  it  has  been 
drawn  out  at  its  open  end,  and  with  the  coke  after  the  end  of  the 
combustion  when  it  is  again  cold,  and  for  that  reason  care  is  re- 
quired iu  managing  it.  We  thus  get  the  quantity  of  coke,  tar, 
ammoniacat  water,  carbon  dioxide  and  hydrogen  sulphide  (as 
lead  sulphide),  and  the  gas  is  measured  by  immersing  the  jar  in 
water,  causing  it  to  be  at  the  same  level  inside  and  out. 

Empty  the  Liebig's  bulbs  into  a  beaker  and  separate  the  lead 
sulphide  by  filtration,  wash  well,  dry  and  weigh.  From  the  lead 
sulphide  the  hydrogen  sulphide  present  is  calculated.  This  pro- 
cess, devised  by  the  late  Dr.  T.  Richardson,  of  Newcastle-on- 
Tyne.  was  found  by  him  to  yield  very  reliable  results,  so  as  to 
be  suitable  for  stating  what  quantity  of  gas  a  ton  of  coal  thus 
analyzed  would  yield.' 
Newbigging's  Experimental  Plant  for  the  Determination 
of  the  Gas-Producing  Qualities  of  Coal. 



A  description  of  the  apparatus  and  method  of  use  are  thus 
given : 

Retort — Cast  iron  1  five  inches  wide,  four  and  one -half  inches  1 
high,  two  feet  three  inches  long  outside,  and  one-half  inch  thick.  ( 

Ascension  pipe — Two  inch  wrought  tube. 

Connections — One  and  one-half  inch  wrought  tube. 

Condenser — Twelve  vertical,  one  and  one-half  inch  wrought  J 
tubes,  each  three  feet  six  inches  long. 

Washer — One  foot  long,  six  inches  wide,  six  inches  deep. 

Purifier^ — One  foot  two  inches  square,   twelve  inches  deep,J 
with  two  trays  of  lime. 

Gas-holder — Capacity  twelve  cubic  feet,  with  graduated  scale  1 

Amount  of  coal  to  be  taken  for  each  test  is  jTiVTr  part  of  a  ton, 
or  2.24  pounds.     Care  should  be  taken  to  obtain  a  fair  average 
sample  of  the  coal  to  be  operated  upon.     For  that  purpose  at 
least  fifty  pouuds  of  coal  should  be  broken  up  into  small  pieces   1 
and  thoroughly  intermixed,  and  from  this  three  different  charges  I 
are  to  be  taken.     The  retort  should  be  at  a  bright  red  heat  m 
before  the  introduction  of  the  coal  and  maintained  at  that  tem- 
perature during  test.     If  from  any  cause  the  temperature  is  much 
reduced,  the  test  wil!  not  be  satisfactory.     The  time  required  to 
work  off  the  charge  of  2.24  pounds  will  range  from  forty  to  sixty 
minutes,  according  to  the  character  of  the  coal.     The  illumina- 
ting power  of  the  gas  given  out  from  each  charge  should  be  as- 
certained by  the  Bunsen  photometer,  no  other  being  sufBciently 
trustworthy  for  that  purpose.     The  average  of  the  three  is  then 
taken,  both  for  yield  of  gas  and  coke  and  for  the  illuminating 
power  of  the  gas,  and  this  fairly  represents  the  capabilities  of 
the  coal.     The  further  conditions  to  be  observed  are  that  the  , 
holder  be  emptied  of  air  or  of  the  previous  charge  of  gas,  and   1 
that  the  condenser  be  drained  of  its  contents.     The  test  charge 
may  be  continued  until  the  whole  of  the  gas   is  expelled,   or 
otherwise,   depending   ou   circumstances.      In  comparing  t\vo 
coals,  an  equal  production  from  both  may  be  obtained,  and  the  n 
comparative  illuminating  power  then  ascertained. 

The  coke  and  "breeze"  should  be  carefully  drawn  from  thej 
retort  into  a  water-tight  receptacle  made  of  sheet  iron  closed  1 


a  lid.  This  is  then  placed  in  a  bucket  or  other  vessel  of  cold 
water,  and  when  sufEciently  cooled,  the  coke  is  weighed. 

For  ascertaining  the  quantity  of  tar  and  aminoniacal  liquor 
produced,  drain  the  yield  of  three  charges  from  the  condenser 
and  washer  and  measure  this  in  a  graduated  liquid  measure. 
The  number  of  fluid  minims  in  a  gallon  (English)  is  76,800. 

<Ti.-     »-h,  „f.  fTheoumberof)        f  J[?^   '?'"'-":  1 

I  The  weight  or)  I  _;_!_„  „f  ,_,  [  oer   0'    tniDims  I 

S.TsJthrce   charges  f  \  1340  ::  H  ,_j  V   „       li.    r  ■   ■!   of  Ur  and  liquor  > 

»"]        „,  oo.l'"      (  (■"'as.  "'■J       i  ""=•.■"    ■") 

and  this  amount  divided  by  76,800  gives  the  gallons  of  tar  and 
liquor  produced  per  ton.'  A  good  variety  of  gas  coal  should 
produce  from  2,240  pounds  of  coal  12,000  cubic  feet  of  gas,  illu- 
minating power  twenty  sperm  candles. 

Newcastle  coal  ou  an  average  produces   12,700  cubic  feet  of 

igas  per  ton  of  coal,  illuminating  power  of  fifteen  sperm  candles. 



.nalysis  of  Clay,  Kaolin,  Fire  Sand.  Building  Stones,  Etc. 

To  6e  Determined. — Silica,    (total),    (combined),    (free),    (hy- 

ated),  alumina,  lime,  magnesia,  potash,  soda,  ferrous  or  ferric 

ixide,   manganous  oxide,   titanic  oxide,   sulphur  trioxide  and 

bined  water." 

The  total  silica  is  determined  by  fusing  cue  gram  of  the  clay 
(previously  dried  at  100°  C.)  with  ten  parts  of  an  equal  mixture 
of  sodium  and  potassium  carbonates,  in  a  large  platinum  cru- 
cible. Fusion  must  be  complete  and  maintained  al  a  red  heat 
thirty  minutes. 

Allow  to  cool,  treat  with  an  excess  of  boiling  water,  make  acid 
with  hydrochloric  acid,  transfer  solution  to  a  four-inch  porcelain 
capsule  and  evaporate  to  dryness.  Take  up  with  twenty-five  cc. 
hydrochloric  acid,  add  water,  boil  and  filler  upon  ashless  filter, 
.Wash  well  with  boiling  water,  dry,  ignite  and  weigh  as  silica 

'bliKlDg's  Hsiidbook  tor  Cat  Engineer!,  p.  5;. 


The  forms  of  combination  of  the  silica  in  the  clay  are  deter- 
mined as  follows :  ^ 

Let  A  represent  silica  in  combination  with  bases  of  the  clay. 
Let  B  represent  hydrated  silicic  acid. 
Let  C  represent  quartz  sand. 

Dry  two  grams  of  the  clay  at  a  temperature  of  lOO®  C,  heat 
with  sulphuric  acid,  to  which  a  little  water  has  been  added,  for 
eight  or  ten  hours,  evaporate  to  dryness,  cool,  add  water,  filter 
out  the  undissolved  residue,  wash,  dry  and  weigh  A  -{-  B'\-C. 
Now  transfer  it  in  small  portions  at  a  time  to  a  boiling  solution 
of  sodium  carbonate  ( i  :  lo)  contained  in  a  platinum  dish,  boil 
for  some  time,  filter  off  each  time,  still  very  hot.  When  all  is 
transferred  to  the  dish,  boil  repeatedly  with  strong  solution  of 
sodium  carbonate,  until  a  few  drops  of  the  fluid,  passing  through 
the  filter,  finally  remains  clear  on  warming  with  ammonium 
chloride.  Wash  the  residue,  first  with  hot  water,  then  (to  en- 
sure the  removal  of  everj'  trace  of  sodium  carbonate  which  may 
still  adhere  to  it)  with  water  slightly  acidified  with  hydrochloric 
acid,  and  finally  with  water.  This  will  dissolve  A  -{-  B,  and 
leave  a  residue  Cof  sand,  which  dr>',  ignite  and  weigh. 

To  determine  B  boil  four  or  five  grams  of  the  clay  (previ- 
ously dried  at  100**  C.)  directly  with  a  strong  solution  of  sodium 
carbonate,  in  a  platinum  dish  as  above,  filter  and  wash  thor- 
oughly with  hot  water.  Acidify  the  filtrate  with  hydrochloric 
acid,  evaporate  to  dr>'ness  and  determine  this  silica.  It  repre- 
sents B  or  the  hydrated  silicic  acid.  Add  together  the  weights 
of  B  and  C  thus  found  and  subtract  the  sum  from  the  weight 
of  the  first  residue  A'\-B'\-C,  The  difference  will  be  the 
weight  of  A  or  silica  in  combination  with  bases  in  the  clay. 

If  the  weight  of  A  '\-  B-\'  C  found  here  be  the  same  as  that 
of  the  silica  found  by  fusion,  in  another  sample  of  the  clay  of 
the  same  amount,  the  sand  is  quartz,  but  if  the  weight  oi  A'\-  B 
+  Cbe  greater,  then  the  sand  contains  silicates. 

The  weight  of  the  bases  combined  with  silica  to  silicate^  can 
be  found  by  subtracting  the  weight  of  total  silica  found  in  one 
gram,  by  fusion,  from  the  weight  oi  A  -^  B -{-  C in  one  gram. 

1  From  Fresenius.  Quant.  Anal..  Cairn's,  p.  68. 



•^     D'D' 




u     A* 






9  9 

3  B 
P  < 

ST.  *5 

ft  O) 

o  P 

§8-0  8 
?*    i'S 

°  ft  > 

^   B     O 
O        «2 

B  O 

P   P   ^ 


BO-™         *2 

5*2.SL      §S! 
ft  t-^  I         II 

P9  § 















oo  3 
^  I 











P     B     M 

B  2:8 
<^  «  o 


»<  £. 

O  3*7 

Ȥ^>  ;? 

ft  P        ^ 

•-K  ft    O 

B        C 

O  3*^ 

B        O 
'^  -  "^ 

(l»    CP 

-      f*  CD 

^T  ■*•  N** 

ft  B  B 


B  B  > 

o  ST 

M  ^  p 

P  {D 

P  |e*  ft 

CO  3  5* 

ft  £3      "^ 

2  2  g 

c  -  3 


o  o 









^g  P 
ft   O   O 

B   ■•• 
P    r^   O 

p  2. 

3   ^> 


ft  ^ 

P  o 

3   P 

p  ■- 
CO   '^^ 




















»  2 
o  P 

^   o 








P  ^ 





s-  t^ 

p    '^ 

§-  ^  ft- 

<t  p    . 

W  i— 

'*  g 

CL  tr. 

«<  3 


Potash  and  Soda.  ' 

Take  oue  gram  of  the  dried  clay,  transfer  to  a  three-inch  plati-, 
iiuni  capsule,  add  ten  cc.  sulphuric  acid  and  twenty  cc.  hydro- 
fluonc  acid  and  heat  gently  until  the  silica  is  completely  dissi- 
pated and  the  excess  of  acid  added  driven  off.  Allow  to  cool, 
add  twenty  cc.  warm  hydrochloric  acid,  then  twenty-five  cc. 
water,  transfer  contents  of  platinum  capsule  to  a  No.  j  beaker, 
add  two  cc.  nitric  acid  and  boil.  Add  ammonia  to  alkaline  re- 
action, boil,  filter  off  the  alumina  and  ferric  oxide,  and  to  the 
filtrate  add  animoniuni  oxalate  to  precipitate  the  lime  :  allow  to 
stand  four  hours,  then  filter  ;  the  magnesia  is  separated  in  the 
filtrate  by  ammonium  phosphate,  and  the  filtrate  from  the  mag- 
nesium phosphate  precipitate  is  evaporated  to  dryness  and 
ignited  to  expel  ammonium  salts.  The  residue  is  treated  with 
hydrochloric  acid  and  the  potash  precipitated  by  solution  of 
platinic  chloride  as  usual,  and  weighed  as  K,PtCl,  on  counter- 
poised filters.  The  alcoholic  washings  and  filtrate  is  evaporated 
to  dryness,  the  platinum  compound  decomposed  by  heating  to 
redness  with  oxalic  acid,  boiled  with  water,  filtered,  a  few  drops 
of  sulphuric  acid  added,  then  evaporated  to  dryness,  ignited  t6> 
constant  weight  as  sodium  sulphate,  and  then  calculated  to  Na,0.j 
Sulphur  Trioxide 
Is  determined  by  fusing  one  gram  of  the  clay  with  sodium 
and  potassium  carbonates,  separating  the  silica  as  usual,  and 
precipitating  the  sulphur  trioxide  by  solution  of  barium  chlo- 
ride in  the  acid  filtrate.  (Consult  Scheme  XIII). 
Titanic  Oxide. 
Fuse  five  grams  of  the  dried  clay  with  an  excess  of  a  mixture'^ 
of  sodium  fluoride  and  sodium  bisutphate,  in  a  platinum  cruci- 
ble for  thirty  minutes  at  a  red  heat.  Treat  the  cold  mass  with 
cold  water,  about  200  cc,  add  potassium  hj-droxide  in  excess, 
filter  off  the  titanic  oxide,  wash,  dry  and  ignite  and  fuse  this 
titanic  oxide  with  about  twelve  times  its  weight  of  acid  sodium 
sulphate  :  allow  to  cool,  and  treat  with  concentrated  sulphuric 
acid.  This  is  now  added  to  600  cc.  of  water,  boiled  for  one  hour, 
and  the  precipitated  titanic  oxide  filtered,  dried  and  weighed, 
(Consult  Scheme  XIII,  Determination  of  Titani 






Water  of  Hydration. 

Take  two  grams  of  the  clay,  dried  al  100°  C,   transfer  to  a 

covered  platinum  crucible  and  ignite  over  a  blast-lamp  at  a  red 

heat  to  constant  weight.     The  loss  represents   the    combined 


Clays  or  fire  sands  that  are  to  be  used  in  the  manufacture  of 

fire  bricks,  retorts,  etc.,  should  contain  only  small  amounts  of 

easily  fusible  materials,  such  as  potash,  soda  or  iron  ;  le^  than 

one  per  cent,  of  either  alkali,  or  two  per  cent,  of  iron  oxide  be- 

,  iDg  allowable  in  the  best  fire  clays. 

COMPosmoN  OP  SoMK  Representative  Firs  Clays. 

I.  I.  3.  4.  3.  fi.  7-  S.  V. 

lom'd)  50.46    50.15  56.4a    65.10    39.94      ....     40.33    29.67  44.10 

[  A1,0, 35.90    35,60  26.35    "-a*    36-30      0-7*    38-54    ao.87  39.14 

!  BfO 12.74    i3-6i  lo-M      7'o    145*      0-35    i3a>     8.61  14.05 

K,0 0.48      0.18      0.42      O.C4      0.66      1.5s  0.25 

Ma,0 ,    ....      0.07  

CaO 0.13      0.11       0.60      0.14      0.21      0.08      

MgO 0.02      0.16      o.  18      o.  19      ...  -       0.38      0.30  .... 

Fe,Oi 1 .50      0.83  1 .33      1 .92      0.46      0.18      0.90      1.4s  0.45 

'  SiO,  Ucee) 4-9°    9831      S'S    36-41 

\  Hoisture 2.80      2,iS      3.26      0.90 

iTiO, 1. 15     1-14  1.05 

ISO,  0.14  

|Org.  matter.  ^^  j^  ^^  _o^   ^^_    ^^    ^^.    _^^  ^^. 

Total 100.7s  100.67  100.63    99-60    99.'8    99-9»    99-^4  100,00  100.14 

No.  I. — Mt.  Savage  fireclay,  Md. 

2.— Fire  clay,  Clearfield  Co.,  Pa. 
No.  3. — Glenboig  clay,  England. 
No,  4. — Stourbridge  clay,  England. 
No.  5. — Saaran  day,  Germany. 
No.  6. — "Dinas,"'  England. 
No.  7. — Zettlitz  clay,  Bohemia. 
No,  8, — Stoneware  clay.  N.  J. 
No.  9. — Paper  clay,  N.  J. 

Building  stone,  such  as  granite,  limestone,  sandstone,  slate, 
'  brick ,  etc. ,  are  generally  subjected  to  certain  mechanical  or  phys- 
ical tests  in  addition  to  a  chemical  analysis  to  determine  their 
[  relative  value. 

■  UMd  in  DukiDg  the  celcbrBlcd-DlDis"  Fire  Bricks,  Duted  tor  Ihclr  CDdurancc  at 
E'Ugli  beau  BOd  for  swcltins  and  uniting  URhl  roofs  for  (unudia 


These  physical  tests  generally  comprise  : 

1.  Crushing  strength. 

2.  Absorptive  power. 

3.  Resistance  to  the  expansion  of  frost,  by  saturating  the  stone 
with  water  and  freezing  a  number  of  times  to  produce  an  effect 
similar  to  frost. 

4.  Microscopical  examination. 

7.  Crushing  Strength, 

The  crushing  strength  is  generally  determined  by  applying  a 
measured  force  to  one-inch  or  two-inch  cubes  of  the  material 
until  they  are  crushed. 

These  compression  tests  are  comparative  only  and  give  no 
idea  of  the  crushing  strength  of  the  material  in  large  masses.  A 
Riehl6  U.  S.  Standard  Automatic  and  Autographic  Testing 
machine  is  used  for  this  purpose.*     (Fig.  90). 

Crushing  Strength  op  Various  Building  Stonbs; 

Ultimate  crushing  strength. 
Pounds  per  Tons  per 

square  inch.  square  foot 

Kinds  of  stone.         Minimum       Maximum.     Minimum.    M'«w<wnmi 

Granite 12000  21000  dSo  1510 

Trap  rock  of  N.  J 20000  24000  1440  1730 

Marble 8000  20000  580  1440 

Limestone 7000  20000  500  1440 

Sandstone 5000  15000  360  joSo 

Common  redbrick--**     2000  3000  144  216 

2,  Absorptive  Power, 

This  is  determined  by  drying  the  sample  and  weighing  it, 
then  soaking  it  in  water  for  twenty-four  hours  and  weighing 
again.  The  increase  of  weight  represents  the  amount  of  water 
absorbed.  A  close  fine-grained  stone  absorbs  less  water  than  a 
coarse-grained  one,  and  generally  the  less  the  absorption  the 
better  the  stone. 

Absorptive  Power  op  Stone,  Brick  and  Mortar. 

Ratio  of  absorption.! 
Kind  of  material.  Maximum.  Minimum. 

Granite    i — 150  o 

Marble  i — 150  o 

Limestone i — 20  ■    i — 500 

Sandstone  i — 15  i — 240 

Brick I — 5  I — ^50 

Mortar l — 2  i — 10 

1  Thus,  if  150  units  of  dry  g^ranite  weigh  after  immersion  in  water  151  units,  the  ab- 
sorption is  one  in  150  and  stated  i — 150. 

'^  For  description  of  this  apparatus  consult  T%e  Digest  qf  Physical  Tests  and  Laboratory 
Practice,  Vol.  i,  p.  248  (July  1896). 


_j.  Freezing  Test. 
Samples  of  the  weighed  material,  preferably  cut  in  two-inch 
cubes,  are  saturated  with  water,  then  placed  in  a  Tagliabue  freez- 
ing apparatus  ( Fig.  91 )  and  maintained  at  a  temperature  of  io°F. 

for  four  hours.  They  are  then  removed,  allowed  to  thaw  gradually 
to  a  temperature  of  about  65° ;  then  moistened  with  water  and 
placed  again  in  the  freezing  apparatus  and  maintained  at  a  tem- 
perature of  fo"  F.  for  four  hours.  This  process  is  repeated  at 
least  ten  times,  when,  after  the  samples  have  acquired  the  tem- 
perature of  the  room,  the  moisture  is  wiped  from  them,  then 
dried,  and  their  weight  carefully  determined.  The  loss  of  weight 
represents  the  material  broken  off  by  the  expansive  action  of 
freezing  the  contained  water.  The  following  method  of  making  the 
frost  test  of  building  stones,  is  from  "Uniform  Methods  of 
Procedure  in  Testing  Building  and  Structural  Materials"  by  J. 



Bauschingei  (Mecfaaiiisch-technischen  Laboratoriuni ,  Miiti- 
chen ) . ' 

The  examination  of  resistance  to  frost  is  to  be  determiued  from 
samples  of  uniform  size,  inasmuch  as  the  absorption  of  water 
and  actionof  frost  are  directly  proportional  to  the  surface  exposed. 
The  test  sample  should  be  a  cube  of  seven  cm.  (2,76  inches) 
length  on  edges. 

The  frost  test  consists  of  : 

a.  The  determination  of  the  compressive  strength  of  saturated 
stones,  and  its  comparison  with  that  of  dried  pieces. 

b.  The  determination  of  compressive  strength  of  the  dried  stone 
after  having  been  frozen  and  thawed  out  twenty-five  times,  and 
its  comparison  with  that  of  dried  pieces  not  so  treated. 

c.  The  determination  of  the  loss  of  weight  of  the  stone  after  the 
twenty-fifth  frost  and  thaw  :  special  attention  must  be  had  to 
the  loss  of  those  particles  which  are  detached  by  the  median- 
teal  action,  and  also  those  lost  by  solution  in  a  definite  quantity  of 

d.  The  examination  of  the  frozen  stone  by  use  of  a  magnifying 
glass,  to  determine  particularly  whether  fissures  or  scaling  oc- 

For  the  frost  test  are  to  be  used  ; 

Six  pieces  for  compression  tests  in  dry  condition,  three  norma! 
and  parallel  to  the  bed  of  the  stone,  six  test  pieces  in  saturated 
condition,  not  frozen  however  ;  three  tested  normal  to,  and  three 
parallel  to,  bed  of  stone. 

Six  test  pieces  for  tests  when  frozen,  three  of  which  arc  to  be 
tested  normal  to,  and  three  parallel  to,  bed  of  stoue. 

When  making  the  freezing  test  the  following  details  are  to  be 
observed : 

a.  During  the  absorption  of  water,  the  cubes  are  at  first  to  be 
immersed  by  two  cm.  (0.77  inch)  deep,  and  are  to  be  lowered 
little  by  little  until  finally  submerged. 

b.  For  immersion  distilled  water  is  to  be  used  at  a  temperature 
of  from  I5°C,  to  20°  C. 

c.  The  saturated  blocks  are  to  be  subjected  to  temperatures 
of  from  — 10°  to  —15°  C. 

iSlandnnlTesUondniethoilaofrejlingMntertiili    T'a-s.  Amir.  Society  Rftck.  Engi- 






d.  The  blocks  are  to  be  subjected  to  the  influence  of  such 
cold  (or  four  hours,  and  they  are  to  be  thus  treated  when  com- 
pletely saturated. 

c.  The  blocks  are  to  be  thawed  out  in  a  ^iven  quantity  of  dis- 
tilled water  at  from  15°  C.  to  zo°  C. 

The  Testing  0/  Brick. — i.  When  testing  bricks  as  found  in 
delivery,  the  least  burnt  are  always  to  be  selected  for  investiga*. 

2.  Bricks  are  to  be  tested  for  resistance  to  compression  ittl 
the  shape  of  cubical  pieces,  formed  by  the  superposition  of  tw*' 
half  bricks,  which  are  to  be  united  by  a  thin  layer  of  mortar 
consisting  of  pure  Portland  cement,  and  the  pressure  surfaces 
are  also  to  be  made  smooth  by  covering  them  wnth  a  thin  coal- 
ing of  the  same  material.     At  least  six  pieces  are  to  be  tested. 

3.  The  specific  gravity  is  to  be  determined. 

4.  In  order  to  control  the  uniformity  of  the  material, 
porosity  of  the  bricks  is  to  be  determined  ;  for  this  purpose  they 
are  to  be  thoroughly  dried  and  then  submerged  in  water  until 
saturated.  Ten  pieces  are  to  be  thoroughly  dried  upon  an  iron 
plate  and  weighed  ;  then  these  bricks  are  to  be  immersed  in 
water  for  twenty-four  hours,  in  such  a  waj-  that  the  water-level 
stands  at  half  the  thickness  ;  after  this  they  are  to  besubmerged 
for  another  twenty-four  hours,  then  to  be  dried  superficially  and 
again  weighed  ;  thus  the  average  quantity  of  water  absorbed  is 
determined.  The  porosity  is  always  to  be  calculated  by  volume, 
though  the  per  cent,  of  water  absorbed  is  always  to  be  stated  in 

5.  Resistance  against  frost  is  to  be  determined  as  follows  : 

a.  Five  of  the  bricks,  previously  saturated  by  water,  are  to 
tested  by  compression. 

b.  The  other  five  are  put  into  a  refrigerator  which  can  prodm 
a  temperature  of — i5°C.  at  least,  and  kept  therein  for  four  hours ; 
then  they  are  removed  and  thawed  in  water  of  a  temperature  of 
20°  C.  Particles  which  might  possibly  become  detached  are  to 
remain  in  the  vessels  in  which  the  brick  is  thawed  until  the  end 
of  the  operation.  This  process  of  freezing  is  repeated  twenty- 
five  times,  and  the  detached  particles  are  dried  and  compared 
by  weight  with  the  original  dry  weight  of  brick.     Particular 





attention,  by  using  a  magnifying  glass,  is  to  be  given  to   the 
possible  formation  of  cracks  or  laminations. 

c.  After  freezing,  the  bricks  are  to  be  tested  by  compression. 
For  this  test  they  are  dried,  and  the  result  obtained  is  to  be  com- 
pared with  that  of  dry  brick  not  frozen. 

d.  Thus,  freezing  the  bricks  does  not  give  a  knowledge  of  the 
absolute  frost-resisting  capacity  :  the  value  of  the  inve.stigation 
is  only  relative,  because  by  it  can  only  be  determined  which 
brick  can  be  most  easily  destroyed  by  the  action  of  frost. 

6.  To  test  bricks  for  Wi^ presence  of  soluble  salts,  five  are  selected , 
and  againthose  which  are  least  burnt,  and  then  such  which  have 
not  yet  been  moistened.  Of  these,  again,  the  interior  parts  only 
are  used,  (or  which  reason  the  bricks  are  split  in  three  direc- 
tions, thus  producing  eight  pieces,  of  which  the  corners  lying 
innermost  in  the  brick  are  knocked  off.  These  are  then  pow- 
dered until  alt  passes  through  a  sieve  of  900  meshes  per  square 
centimeter  (about  5,840  per  square  inch),  from  which  the  dust 
is  again  separated  by  a  sieve  of  4,900  meshes  per  square  centi- 
meter (about  31, 3^  per  square  inch),  and  the  particles  remain- 
ing on  the  latter  are  examined.  Twenty-five  grams  are  lixivi- 
ated in  250  cubic  centimeters  of  distilled  water,  boiled  for 
About  an  hour,  however  replenishing  the  quantity  evaporated, 
then  filtered  and  washed. 

The  quantity  of  soluble  salts  present  is  then  determined  by 
boiling  down  the  solution  and  bringing  the  residue  to  a  red 
heat  for  a  few  minutes.  The  quantity  of  soluble  salts  present  is 
to  be  given  in  per  cent,  of  the  original  weight  of  brick. 

The  salts  obtained  are  to  be  submitted  to  a  chemical  analysis. 

7.  Determinations  of  the  presence  of  calcium  carbonate,  py- 
rites, raica  and  similar  substances  are  to  be  made  on  the  un- 
burned  clay,  for  which  purpose  unburned  bricks  are  to  be  fur- 
nished. These  are  soaked  in  water  and  the  coarse  particles  are 
separated  by  passing  the  whole  material  through  a  sieve  having 
400  meshes  per  square  centimeter.  The  sand  thus  obtained  is 
to  be  examined  by  the  magnifying  glass  and  with  hydrochloric 
acid  to  determine  its  mineralogical  composition.  When  im- 
purities, such  as  carbonate,  pyrites,  etc..  are  found,  then  pieces 
of  brick,  such,   for  instance,  as  remained  from  the  determina- 


tion  oi  soluble  salts,  are  to  be  examined  in  a  Papin's  digester 
for  their  deleterious  influence.  They  are  to  be  so  arranged  in  a 
Papin's  digester  that  they  are  not  touched  by  the  water  directly, 
but  are  subjected  to  the  action  of  the  generated  steam  alone. 
The  pressure  of  steam  shall  be  one-quarter  atmosphere,  and  the 
duration  of  test  three  hours.  Possibly  occurring  disintegration 
is  to  be  determined  by  means  of  the  magnifying  glass. 
4.  Microscopical  Examination. 

This  consists  in  examining  under  the  microscope  their  sec- 
tions of  the  building  stone.  Important  results  are  often  obtained^  1 
especially  so  if  the  substances  used  as  matrix  are  indicated — the 
presence  and  amount  of  injurious  substances,  such  as  iron  py- 
rites, mica,  etc. 

Nearly  all  reports  upon  samples  of  building  stones  now  in^ 
elude  the  microscopical  examination. 

The  first  and  most  essential  test  applied  to  building  stone  isi 
to  determine  the  structure  and  character  of  a  stone,  to  know 
whether  it  be  a  granite,  syenite,  sandstone,  quaitzite  or  some* 
thing  else.  Although  an  expert  can  usually  determine  at  a 
glance  to  which,  if  any,  of  these  groups  a  particular  stone  be- 
longs, it  is  frequently  possible  to  determine  the  precise  litholog- 
ical  character  only  by  a  microscopical  examination.  Thus,  for 
instance,  there  is  a  class  of  Cambrian  rocks  commonly  calh 
Potsdam  sandstone,  that  are  not  sandstones  at  all,  but  are  hard, 
compact  rocks  known  as  quartzites,  which  have  been  deri\'ed 
from  sandstones  by  metamorphic  action.  The  essenti!!!  differ- 
ence between  a  sandstone  and  a  quartzite  lies  in  the  presence  of 
secondary  silica  between  the  quartz  granules  comprising  the 
latter ;  the  presence  of  this  secondar>-  silica  or  quartz  can  be 
determined  for  a  certainty  only  by  microscopical  means.  The 
microscope  is  not  only  useful  in  determining  the  structure  of  a 
stone,  but  it  has  an  even  greater  practical  value  in  making  it 
pos.<iib1e  to  detect  the  presence  of  deleterious  substances,  such  as 
pyrite  and  marcasite,  or  other  minerals  whose  chemical  compo- 
sition is  affected  by  atmospbenc  agencies  and  thus  exert  a  dele- 
terious effect  upon  the  stone.' 




Consult  : 

Tenth  Census  U.  S.,  1880.     "  Building  Stones  and  Quarry  Industry." 

Stones  for  Bnilding  and  Decoration.      By  G.  P.  Merrill.  1891. 

Building  Stone  in  New  York.     By   Prof.  J.  C.  Smock,  in  BulUtinof 
the Seiv  York  State  Museum,  1890. 

The  Testing  of  Material  of  Construction.    By  W.   C.  Unwin,  pp.  410- 


alise  on  Masonry  Construction.    By  I.  O.  Baker,  C.E-,  1893. 
iplete  description  of  the  methods  of  determining  the  fusibility  of 
ys  will  be  found  in  Trans.  Amer.  Inst.  Min,  Eng.,  14,  pp.  43-67. 

This  subject  may  be  divided  into  three  classes : 

1 .  Alloys  composed  principally  of  copper  and  zinc,  or  of  cop- 
per, tin  and  zinc. 

2.  Alloys  or  compositions  in  which  copper,  tin,  lead  or  anti- 
mony are  constituents. 

3.  Alloys  not  included  in  the  first  two  divisions. 

Alloys  of  the  first  class  may  comprise  brass,  bronze,  bell 
metal,  gun  metal.  Muntz's  metal,  etc.  The  analysis  may  be 
performed  as  follows  (if  composed  of  copper  and  zinc  only): 
Transfer  one  gram  of  the  brass  to  a  No.  3  beaker  covered  with 
a  watch-glass,  and  add  gradually  twenty-five  cc.  nitric  acid  ; 
when  solution  is  complete,  remove  watch-glass,  after  washing, 
allow  solution  to  cool,  transfer  it  to  a  one-quarter  liter  Hask  and 
add  water  to  the  containing  mark.  Mix  thoroughly  (the  solu- 
tion being  at  15°  C.)  and  transfer  fifty  cc.  of  the  solution  to  a 
No.  3  beaker,  dilute  sufficiently  with  water  and  precipitate  the 
copper  electrolytically,  as  in  Scheme  VI,  pages-  Upon  complete 
precipitation  of  the  copper,  the  platinum  cone  and  spiral  are 
renio%'ed  from  the  solution,  washed  with  water,  the  washings 
added  to  the  solution  in  the  beaker.  Add  a  few  drops  of  nitric 
acid  to  the  solution,  boil  and  precipitate  the  zinc  with  a  slight 
excess  of  sodium  carbonate.  Boil,  filter,  wash  well  with  hot 
water,  dry,  ignite  and  weigh  as  ZnO. 

Example:  One  gram  brass  turnings  taken.     Solution  350  cc. 

Fifty  cc.  of  solution  taken  ; 



Platinum  cone  +  Cu 28. 175  grams. 

Platinum  cone   27.995        ** 

Cu  =B  0.160        ** 

0.160  X  5  X  100       -  ^    _ 
=  80  per  cent.  Cu. 

Porcelain  crucible  -|-  ZnO 17-655  grams. 

Porcelain  crucible 17-605 

t  i 

ZnO  =     0.050 


0.050  X  65  __                   0.040  X  5  X  100  _  _  ^^  ^^„, 
« =  0.040  Zn,  ^  20  per  cent. 

Cu 80  per  cent. 

Zn 20    " 

Totel 100    •' 

Where  tin  is  also  a  component,  the  above  method  is  varied  as 
follows : 

Take  one  gram  of  the  fine  turnings  and  digest  with  nitric  acid 
as  above.  Evaporate  nearly  to  drj-ness,  add  fifty  cc.  warm 
water,  filter  by  decantation  into  a  one-quarter  liter  flask,  wash- 
ing the  precipitate  thoroughly  with  hot  water,  dr>^  it,  ignite  and 
weigh  as  SnO,  and  calculate  to  Sn. 

The  filtrate  is  made  up  to  250  cc.  ( 15**  C),  thoroughly  mixed, 
and  fifty  cc.  taken  for  copper  and  zinc  as  before. 

Porcelain  crucible  -f-  SnO, 16.6743  grams. 

Porcelain  crucible 16.5220       '* 

SuOj  =     0.1523       •• 
Sn  =  12  per  cent. 

Platinum  cone  H-Cu 28.115  grams. 

Platinum  cone 27.995       ** 

Cu=    0.120       ** 
Cu  =  60  per  cent. 

Porcelain  crucible  -|-  ZnO 17.6750  grams. 

Porcelain  crucible 1 7.6052 


ZnO  s    0.0698 
Zn  :=  28  per  cent. 




ALLOTS.  arj; 

Total loo    '■       " 

Examples  of  Alloys  op  the  first  Class. 

Bell  metal 13  78  ..     parts. 

Brass   72  18 

Brass  (yellow) 60  40         " 

Bronze  for  bearings..--    16  83  2         " 

Speculum  metal 33,4  66.6  ..        '■  ' 

Delta  ineUl'or"Sterro"   ..  60  38.3  (i.SFe)" 

Munti  metal 60  40         " 

Alloys  of  the  second  class  may  comprise  Babbitt  metal,  Britan- 
nia metal,  type  metal,  solder,  white  metal,  camelia  metal,  Tobin 
bronze,  ajax  metal,  car-box  metal,  manganese  bronze,  magnolia 
metal,  etc. 

Atialysis  0/  Babbitt  Metal.' 
Five  grams  of  drillings  in  an  eight-ounce  beaker  are  treated 
with  thirty  cc.  nitric  acid  (1.20  sp.  gr.)  and  heated  till  decom- 
position is  complete  and  the  free  acid  nearly  all  evaporated. 
When  about  five  cc.  of  the  solution  remain,  add  fifteen  cc.  o( 
water,  and  then  add  concentrated  sodium  hydroxide  solution  till 
nearly  neutral ;  fifty  cc.  of  sodium  sulphide  solution  are  then 
added,  the  mixture  well  stirred,  then  boiled  gently  for  half  an 
hour.  The  solution  then  contains  the  tin  and  antimony.  The 
precipitate,  which  contains  the  sulphides  of  lead  and  copper,  is 
filtered  on  a  nine  cm.  Swedish  filter,  and  washed  thoroughly 
with  water  containing  one  per  cent,  of  the  above  sodium  sulphide 
solution.     The  filtrate  is  received  in  a  300  cc.  beaker. 

Tin  and  AnlUnony. — The  filtrate  is  diluted  to  200  cc.  and 
boiled.  Crystalsof  oxalic  acid,  C.  P.,  are  cautiously  added  till 
the  sodium  sulphide  is  all  decomposed  and  a  milky  separation 
appears,  mixed  with  a  precipitate  which  is  usually  at  first  black. 
Boil  for  twenty  minutes.  Pass  hydrogen  sulphide  for  tea  min- 
utes.    Filter  rapidly  on  a  Gooch  crucible  and  wash  with  hot 

iSomcvarieil«8o[  Delta  nicUlconUiuoDc  10  two  pet  ceut.  of  Ila. 
»  Method  o(  E.  M.  Btiice,  modified. 



water.  Dry,  and  heat  crucible  and  contents  in  a  stream  of  car- 
bon dioxide  to  a  temperature  above  300°  C.  for  one  hour.  Cool 
in  carbon  dioxide,  remove  crucible  and  weigh  as  Sb,S,.  The 
Gooch  crucible  containing  the  Sb,S,  +  S  may  be  treated  with 
alcohol,  then  carbon  disulphide,  then  alcohol  (in  order  to  re- 
move the  sulphur),  dried  and  weighed,  instead  of  igniting  in 
carbon  dioxide.     Sb,S,  X  0.71390  =^  Sb. 

The  filtrate  from  the  Sb,S,  is  treated  with  thirt}'  cc.  concen- 
trated sulphuric  acid  and  boiled  down  till  all  oxalic  acid  is  decom- 
posed and  strong  fumes  of  sulphuric  acid  come  off.  Cool.  Dilute 
t cautiously  to  200  cc,  mix  well  and  filter  quickly.  Dilute  fil- 
trate to  300  cc,  warm  slightly  and  pass  hydrogen  sulphide. 
Filter  stannous  sulphide  and  wash  with  hot  water.  Dry,  ignite 
and  weigh  as  stannic  oxide  in  porcelain  crucible.  SnO,  X 
0.788  =  Sn. 

The  copper  and  lead  sulphide  precipitate  is  washed  off  the 
filter,  treated  with  dilute  nitric  acid,  warmed  till  decomposed, 
and  the  sulphur  filtered  off.  The  lead  is  then  separated  as  sul- 
phate by  evaporation  with  sulphuric  acid.  The  lead  sulphate  is 
filtered  on  a  Gooch  crucible,  washed  with  water  containing  five 
per  cent,  sulphuric  acid,  dried  and  ignited  over  a  Bunsen  bur- 
ner.    PbSO,  X  0.68298=  Pb. 

The  copper  is  separated  from  the  filtrate  by  hydrogen  sulphide. 
The  sulphide  is  decomposed  by  nitric  acid,  and  the  resulting^ 
solution  titrated  or  electrolyzed. 

Sodium  sulphide  solution  for  Babbitt  analysis  is  made  up  as 
follows :  One  pound  sodium  sulphide  crystals  are  dissolved  in 
two  liters  of  water.  Portions  of  this  are  from  time  to  time  satu- 
rated with  hydrogen  oxide  gas  and  filtered  for  use. 

Separation  of  Tin  and  Antimony  in  Alloys, 

Mengin  treats  the  alloy  (for  instance,  anti-friction  metal) 
with  nitric  acid  (1.15),  collects  the  insoluble  oxides  of  tin  and 
antimony,  washes,  carefully  ignites  and  weighs  them.  Af. 
The  mixed  oxides  are  next  suspended  in  hydrochloric  acid  and 
water  and  a  ball  or  plate  of  pure  tin  added,  whereupon  the  anti- 
mony is  reduced  to  metal  and  the  tin  converted  into  chloride  ; 
the  reaction  is  best  accelerated  by  heat,  about  three  hours  being 

ALLOYS.                                                        3iy'^H 

fill  m 




IP  ?« 
















'If  fill  Ii 

^      i^-fiif 



U  Pill 


ft  iS 














1   Ii* 

•  03  Its 









necessary  for  two  grams  of  the  oxides.  The  precipitated  anti- 
mony is  washed  by  decantation  with  water,  then  with  alcohol, 
dried  and  weighed  A.  There  is  no  appreciable  [oxidation  of 
the  antimony  and  the  method  is  very  exact.  The  tin  is  esti- 
mated by  difference. 

M — A  X  1.262  =  weight  of  tin  oxide  ;  the  latter  multiplied 
by  0.7888  gives  the  weight  of  tin  in  the  alloy.  An  alternative 
method  for  the  estimation  of  the  tin  is  to  precipitate  the  latter  by 
zinc.  The  following  figures  (indicating  grams)  of  an  analysis, 
show  the  accuracy  of  the  method  : 

SatnpleH  taken.  Oxides  found.  Metals  found. 

Sn  1. 162  \  (  Sn  1. 154 

Sb  1.312  /  3-131  \  Sb  1.309 

Examples  of  Ali«oys  op  the  Second  Class. 

Iron.  Tin.  Antimony.  Lead.  Copper.     Zinc.  Bismuth.  Phos. 

Babbitt  metal 45.5  13.00       40        1.5        

Britannia  metal 90.0  10.00      

Pewter   ••••  89.3  7.1        ....       1.8        ...-       1.8 

Argentine 85.5  14.5        

Ashbury  metal 77.8  19.4        2.8 

Soft  solder 50.0  .  •  •  •       50.0      

Anti-friction  metal 40.0  5.0        55.0      

Tobin  bronze 0.2  0.9  ....         0.4    61.2      37.3 

Phosphor-bronze* lo.o  ....         9.5     79.70      0.8 

Deoxidized  bronze. .  0.20  12.40  ....         2.2782.67      2.45      ..*•     0.005 

Magnolia  metal 4.75  15.0        80.0    trace      ....       0.25      .... 

Rose  metal 22.90  ....  27.10      50.0        .... 

Camelia  metal 0.55  4.25  ....  14.75     70.20     10.20      

Ajax  metal 10.98  ....       7.37    81.28      0.37 

Car-box  metal 0.61  ....  14.3S  84.33      * ** *       0.68      

Parson  *s  white  metal  ....  86.00  i.oo      2.00      2.00    27.00      

••  B  "  alloy.  P.  R.  R.'  ....  8.00  ....  15.00    77.00      trace 

Third  Class  Comprises  : 

Aluminum  bronze Al  7.3,  Si  6.5,  Cu  86.2  or  Al  10,  Cu  90 

Ferro-aluminum Al  i  .23.  Fe,  etc.  99.73  or  Al  12.50,  Fe,  etc.,  87.50 

Ferro-tungsten Fe  43.4,  W  53.1,  Mn  3.5^ 

German  silver Cu  50,  Ni  14.8,  Sn  3.1,  Zn  31.9 

Rosine Ni  40,  Ag  10,  Al  30,  Sn  20 

1  Detuilcil  insirtiotions  for  the  determination  of  phosphorus  in  phosphor  bronse  will 
be  found  in  fht-  AmrnVan  Frtgtnrrt  and  KailrxHiti Journal^  6S.  fzS. 

4  This  rtlloy.  aocordiuK  loC.  B.  Dudley  ^y.  FraHk:im  Inst..  Afarck^  1892^  p.  168).  is  the 
best  bcaiiiiK  nielal  known. 

»  Consult  experiments  on  ferro-tung»ten  :  J.  S.  I>eBenneville.  J  Am.  Ckem,  5ac.,  16, 

ALLOYS.  317 

Metalline Co  35,  Al  15,  Cu  30,  Fe  10 

AIumiBum  "bourbouni  " Al  85.74,  Sn  ti.94.  Si  1.31 

Silicon  bronie Fe,  etc.  86.59,  Si  13,41' 

Guthrie's  ■■  Enteclic" Cd  14.03.  Sn,  Pb  10.55,  Bi  50 

.Arsenic  bronze Cu  79.70,  Sn  10,  Pb  9.50.  As  0.80 

Manganese  bronze '  ■  -  ■  -  Cu  88,  Sn  10,  Mo  i 

Aluminum  bronze  can  be  analyzed  as  follows :  Take  one 
gram  of  bronze  in  fine  turnings,  transfer  to  a  No.  3  beaker  and 
add  gradually  twenty-five  cc.  of  aqua  regis.  Evaporate  to  diy- 
ness,  to  render  the  silica  insoluble,  take  up  with  twenty-five  cc. 
hydrochloric  acid,  twenty-five  cc.  water,  warm  and  filter.  Wash 
well.  The  residue  is  dried,  ignited  and  weighed  as  SiO,,  and 
calculated  to  Si.  The  filtrate  from  the  silicaisdiluted  to  250  cc. 
thoroughly  mixed  and  100  cc.  transferred  to  a  No.  3  beaker  and 
the  copper  precipitated  with  hydrogen  sulphide,  filtered,  washed 
with  hydrogen  sulphide  water,  the  cupric  sulphide  dissolved  in 
nitric  acid,  and  the  copper  determined  by  electrolysis  (Scheme 
VI).  The  filtrate  from  the  cupric  sulphide  is  boiled  to  expel 
hydrogen  sulphide,  a  few  drops  of  nitric  acid  added,  the  solution 
made  alkaline  with  ammonia,  and  the  alumina  determined  as  in 
Scheme  III,  and  calculated  to  Al. 

Determination  0/  Manganese  in  Majiganese  Bronze.*  Dissolve 
ive  grams  of  drilling  in  nitric  acid  of  i.zosp.  gr.,  using  a  large 
beaker  to  avoid  frothing  over.  An  excess  of  acid  must  be 
avoided  as  it  interferes  with  the  precipitation  of  the  copper  by 
hydrogen  sulphide.  When  solution  is  complete,  transfer  to  a 
500  cc.  cylinder  without  filtering  out  the  precipitated  stannic 
oxide.  Make  up  to  300  cc.  and  pass  a  rapid  current  of  hydrogen 
sulphide  from  a  Kipp's  apparatus  until  the  supernatant  liquid  is 
colorless.  Decant  off  through  a  dry  filter,  180  cc.  corresponding 
!o  three  grams  of  sample,  and  boil  rapidly  dowu  to  about  ten  cc. 
Transfer  to  a  small  beaker  and  add  twenty-five  cc.  of  strong 
nitric  acid.  Boil  down  one-half,  make  up  with  strong  nitric 
acid,  boil,  and  add  one  spoonful  of  potassium  chlorate.  Boil 
ten  minutes  and  add  another  spoonful  of  potassium  chlorate. 
Boil  till  free  from  chlorine,  cool  in  water,  and  filler  on  asbestos, 

■OB,  by  H    J.  Williami!,  Trai...  Amir,  Inil.  Mm.  EHg..  17,  S4i. 


using  filter  pump.     Wash  with  strongnitric  acid  through  which  I 
a  stream  ot  air  has  been  passed.     When  free  from  iron,  washl 
with  cold  water  until  no  acid  remains.     Place  the  felt  and  pre- 
cipitate in  the  same  beaker  and  dissolve  in  ferrous  sulphate, 
using  five  cc.  at  a  time.     Titrate  back  with  permanganate  until 
a  pink  color  remains.     Deduct  the  number  of  cc.  used  in  titra- 
ting back,  from  the  number  of  equivalents  of  ferrous  sulphate 
used,  and  the  remainder  shows  the  manganese  in  the  amount  of   ' 
sample  taken.  I 

Permanganate  Solution. — Dissolve  1-149  grams  of  potassium^ 
permanganate  in  cc.  water:  one  cc.  equals  0.00  igram 
manganese;  check  by  dissolving  0.1425  gram  ammonio- ferrous 
sulphate  in  a  little  water  and  acidulating  with  hydrochloric  acid. 
This  should  precipitate  ten  milligrams  of  manganese.  If  not, 
apply  factor  of  correction. 

Ferrous  Sulphate  Solution. — A  solution  of  ferrous  sulphate  inl 
two  per  cent,  sulphuric  acid  so  dilute  that  five  cc.  correspond) 
to  ten  cc.  permanganate  solution.     This  is  best  made  by  trial'J 
aud  solution. 

Analysis  of  Ferro- Aluminum. 

Five  grams  of   the  ferro- aluminum  are  transferred  to  a  500  c 
beaker  and  dissolved  in  seventy-five  cc.   sulphuric  acid   [sp. 
gr.  I. jo),  then  evaporated  to  dryness.     The  residue  is  treated 
with  fifty  cc.  dilute  sulphuric  acid  diluted  to  300  cc.  and  mixed 
well.      TOO  cc.  of  the  solution   {=■  r.666  grams)  are  filtered  off  < 
into  a  graduated  100  cc.  measure  ;  this  is  then  poured  into  335000.  I 
beaker  ;  about  five  grams  of  pure  iron  wire  are  now  added  and  the  1 
solution    boiled,  so    as    to    reduce    any  ferric  salt  formed  ;    the 
excess   of    acid  is  carefully  neutralized  with  solution  of  sodium 
carbonate  and  the  mixture  gradually  poured  into   150  cc.  of  a 
boiling  solution  containing  thirty  grams  of  potassium  hydroxide 
and  sixty  grams  of  potassium  cyanide :  the  mixture  of  potas- 
sium hydroxide  and  potassium  cyanide   with  iron  precipitated 
as   hydroxide  is  diluted  up  to  50c   cc.  in  a  graduated  measure,' 
and  300  cc.  (—  I  gram  of  sample)  filteredoff  intoasix-inchevap- d 
orating  dish;  200  cc.  of  a  standard  solution  o(  ammonium  ni-." 
trate  are  now  added  and  the  mixture  heated  forty  minutes  ;  I 

ALLOYS.  319 

filler  and  wash  the  precipitated  alumina  with  hot  water,  re- 
dissolve  iu  twenty-five  cc.  of  dilute  hydrochloric  acid,  dilute  to 
200  cc.  neutralize  with  ammonium  hydroxide,  add  a  slight  ex- 
cess, boil,  filter  and  wash  with  hot  water,  dr>-.  ignite  and 
weigh  as  A!,0,.  The  weight  obtained  multiplied  by  0.534  X 
100=  percentages  of  Al.  This  amount  subtracted  from  too 
per  cent,  gives  the  percentage  of  iron.     (Phillips.) 

German  silver,  Rosiue,  Aluminum  "  bourbounz."  Guthrie's 
■  'entectic' '  and  arsenic  bronze  all  require  solution  innitric  acid  to 
render  the  tin  insoluble,  which  is  then  separated  by  filtration 
from  the  other  components. 

The  determination  of  phosphorus  in  phosphor-tin  presents 
some  difficulty  on  account  of  the  insoluble  compound  which 
phosphoric  acid  forms  with  stannic  oxide.  Hempel,'  states  as 
follows : 

The  ordinary  way  of  analyzing  phosphide  of  tin  by  dissolving 
it  in  aqua  regia  and  precipitating  with  hydrogen  sulphide  is  not 
satisfactory,  as  a  considerable  quantity  of  phosphorus  is  thrown 
down  with  the  precipitated  sulphide  as  a  basic  phosphate  of  tin. 
It  is  easily  analyzed  according  to  Wohler's  method,  by  treating 
with  chlorine,  the  chlorides  of  tin  and  phosphorus  formed  being 
collected  iu  about  ten  cc.  of  concentrated  nitric  acid.  If  the 
apparatus  be  rinsed  with  a  solution  of  one  part  concentrated 
nitric  acid  and  two  parts  of  water,  no  trace  of  stannic  oxide  is 
precipitated.  The  phosphoric  acid  is  now  easily  precipitated  in 
the  usual  way  by  molybdic  acid. 

If  dilute  nitric  acid  is  taken,  a  part  of  the  phosphorus  separates 
with  the  stannic  oxide  and  the  result  will  be  too  low.  This  also 
applies  to  the  determination  of  phosphorus  in  phosphor-bronze. 

Qualitative  Tests  0/  A  Hoy  so/ Lead,  Copper,  Tin  and  Antimony.'' 
— For  lead,  dissolve  in  aqua  regia.  If  much  lead  be  present,  it 
will  separate  on  cooling  as  chloride:  if  only  a  small  amount  is 
present  it  will  be  detected  by  the  addition  of  four  volumes  of 
ninety-five  per  cent,  alcohol. 

For  tin.  dissolve  in  hydrochloric  acid,  concentrated,  and  be- 
fore the  portion  of  alloy  taken  is  completely  dissolved,  pour  off 

1  Be.  d.  CHiti.  C«..  11.  M7S,  /  Anal.  Chrm..  4.  S3. 

iConimuiiicaled  10  Ihc  nulhor  by  C.  W.  Thomiuon.  Chrmist  Nation«l  Lcid  Co..  N.V. 


the  supernatant  solution,  cool  to  separate  lead  as  chloride,  add 
four  volumes  of  alcohol,  filter  and  to  filtrate  add  slight  excess  of 
bromine  to  convert  stannous  to  stannic  chloride  ;  heat  to  expel 
free  bromine,  dilute  and  pass  hydrogen  sulphide  gas,  when  if 
tin  is  present  it  will  be  obtained  as  yellow  stannic  sulphide. 

For  antimony  treat  alloy  with  concentrated  hydrochloric  acid. 
Almost  all  the  antimony  is  left  undissolved.  Decant,  wash  the 
residue  with  water,  after  which  dissolve  in  hydrochloric  acid 
with  potassium  chlorate,  boil  to  expel  free  chlorine,  pass  hydro- 
gen sulphide,  obtaining  a  precipitate  of  Sb,S,,  if  antimony  is 
present.  If  copper  is  also  present,  it  will  be  precipitated  as  cop- 
per sulphide  and  may  obscure  the  color  of  the  antimonic  sul- 
phide ;  if  so,  filter  and  treat  the  precipitate  with  potassium 
hydroxide  solution,  which  will  dissolve  the  antimonic  sulphide. 

Filter  and  acidify  filtrate,  when  the  pure  color  of  antimonic 
sulphide  will  be  observed  if  antimony  is  present. 

For  copper,  treat  the  alloy  with  dilute  nitric  acid  in  a  porce- 
lain dish  and  evaporate  to  dryness  ;  if  copper  is  present,  it  will 
show  as  a  green  ring  where  it  crystallizes  out  as  nitrate  on  edge 
of  the  residue. 

For  arsenic,  dissolve  alloy  in  hydrochloric  acid,  with  additiotlj 
of  potassium  chlorate,  in  an  Erlenmeyer  flask,  boil  to  expel' 
chlorine,  add  some  more  concentrated  hydrochloric  acid  and  two 
grams  of  sodium  thiosulphate,  connect  flask  with  a  condenser 
and  distil,  following  in  principle  the  method  first  proposed  by 
Fischer.  Arsenic,  if  present,  will  be  found  in  the  distillate  by, 
passing  through  it  hydrogen  sulphide  gas. 
Quantila live  Analysis  of  Alloys  Containing  Copper,  Lead,  Anti- 
mony and  Tin'. 

One  gram  of  the  finely  divided  alloy  is  dissolved  by  boiling  in' 
from  seventy  to  loo  cc.  of  the  following  solution,  in  a  covered 

The  solution  is  made  by  dissolving  twenty  grams  of  potassium 
chloride  in  50a  cc.  of  water,  adding  400  cc.  concentrated  hydro- 
chloric acid,  mixing,  and  then  adding  100  cc.  nitric  acid  of  1.40 
sp.  gr.  No  decomposition  between  hydrochloric  acid  and  nitric 
acid  takes  place  in  this  solution  in  the  cold.     If  complete  solu- 

1  Method  ol  C.  W.  Thompson. 





tion  of  the  alloy  is  difiScuIt  in  the  amount  of  solution  taken,  more 
is  added  as  required.  Continue  boiling  until  solution  is 
evaporated  to  about  fifty  cc.  Cool  by  placing  beaker  in  cold 
water  until  the  bulk  of  the  lead  has  crystallized  out  as  chloride 
and  then  add  slowly,  with  constant  stirring,  100  cc.  ninety-Gve 
per  cent-  alcohol.  Allow  to  stand  about  twenty  minutes,  filter 
through  a  nine  cm.  filter  paper  into  a  No.  4  beaker;  wash  by 
decantation  three  times  with  mixture  (4  to  i)  of  ninety-five  per 
cent,  alcohol  and  hydrochloric  acid,  concentrated,  and  wash 
filter  paper  twice  with  the  same  mixture. 

Wash  the  lead  chloride  on  the  paper  into  a  beaker,  and  wash 
filter  paper  several  times  with  hot  water,  allowing  washings  to 
flow  into  the  beaker  with  rest  of  the  chloride.  Finally  wash 
twice  with  solution  of  ammonium  acetate,  hot,  (the  ammonium 
acetate  solution  is  made  by  taking  one  volume  of  ammonia 
water,  0.900  sp.  gr.,  adding  to  it  one  volume  of  water  and  then 
eighty  per  cent,  acetic  acid  until  the  reaction  is  slightly  acid  to 
litmus) .  heat  until  the  iead  chloride  is  dissolved,  then  add  fifteen 
cc.  of  a  saturated  solution  of  potassium  bichromate,  and  warm 
until  precipitate  is  of  good  orange  color.  Filter  on  weighed  Gooch 
crucible,  wash  with  water,  alcohol  and  ether,  dr>'  at  1 10°  C.  and 

Evaporate  filtrate  from  lead  chloride  by  heating  on  hot  plate 
and  finally  to  dryness  on  water-bath  ;  add  ten  cc.  solution  potas- 
sium hydroxide  (one  gram  to  five  cc.)  and  after  a  few  minutes 
twenty  cc.  peroxide  of  hydrogen :  heat  on  water-hath  for  twenty 
minutes,  add  ten  grams  ammonium  oxalate,  ten  grams  oxalic 
acid  and  zoo  cc.  of  water.  Heat  to  boiling,  pass  hydrogen  sul- 
phide with  solution  near  boiling  for  forty-five  minutes ;  filter  at 
once  and  wash  precipitate  with  hot  water.  Boil  filtrate  to  expel 
hydrogen  sulphide,  concentrate  if  necessary  and  electrolyze  over 
night,  using  a  current  of  about  one-half  ampere.  Usually  by 
morning  the  solution  will  have  become  alkaline,  in  which  case 
it  may  be  taken  for  granted  that  the  tin  is  all  precipitated  on  the 
cylinder.  The  cylinder  is  removed,  washed  twice  with  water 
and  then  with  ninety-five  per  cent,  alcohol,  dried  and 
weighed.  The  precipitate  of  antimony  and  copper  sulphides  on 
paper  is  washed  back   into  beaker  with  the  least  amount  of 


water  possible,  and  treated  with  ten  cc.  potassium  hydroxide 
solution  (1-5),  heated  on  water-bath  until  undissolved  matter  is 
distinctl}'  black  ;  then  filtered  through  the  same  paper  it  was 
washed  from  into  a  twelve-ounce  Erlenmeyer  flask,  washed,  etc. 
On  the  filter  the  copper  is  obtained  as  sulphide  with  a  small 
amount  of  lead  which  failed  of  precipitation  as  chloride.  If  it 
is  desired  to  determine  this  lead,  it  can  be  done  by  separation 
from  the  copper  as  usual ;  if  not,  dry  and  ignite  precipitate  in 
a  small  casserole,  dissolve  in  nitric  acid,  boil  to  expel  nitrogen 
dioxide,  neutralize  with  sodium  carbonate,  add  a  few  drops  of 
ammonia,  and  determine  volumetrically  with  potassium  cyanide 
standardized  against  pure  copper.  The  solution  of  antimony 
sulphide  in  potassium  hydroxide  should  not  amount  to  over  forty 
cc.  Add  one  gram  potassium  chlorate,  fifty  cc.  concentrated 
hydrochloric  acid,  boil  until  solution  is  colorless  and  free  chlo- 
rine is  driven  off ;  filter  through  mineral  wool ;  if  sulphur  has 
separated  into  similar  flask,  wash  out  original  with  concentrated 
hydrochloric  acid,  cool,  add  one  gram  of  potassium  iodide,  one 
cc.  carbon  disulphide,  and  titrate  for  antimony  with  tenth-nor- 
mal sodium  thiosulphate,  one  cc.  of  which  equals  0.0060  gr^m 
antimony.  This  systematic  method  assumes  the  absence  of 
other  metals  than  lead,  tin,  antimony  and  copper.  For  the 
determination  of  other  metals  we  offer  the  following  suggestions  : 
If  arsenic  is  present  it  will  be  separated  with  the  antimony  and 
will  liberate  iodine,  as  does  antimony.  One  cc.  of  tenth-normal 
thiosulphate  equals  0.00375  gram  of  arsenic.  Arsenic  is  prefer- 
ably determined  on  a  separate  portion  by  dissolving  in  hydro- 
chloric acid  and  potassium  chlorate,  boiling  to  expel  free 
chlorine,  and  distilling  after  the  addition  of  sodium  thiosulphate 
as  a  reducing  agent,  passing  hydrogen  sulphide  through  the 
distillate,  and  weighing  as  As,S,,  or  dissolving  in  potassium 
hydroxide  and  determining  volumetrically  as  in  the  case  of  anti- 

Bismuth  and  cadmium  sulphides  would  remain  with  copper 
after  treatment  with  potassium  hydroxide — this  renders  this 
method  very  suitable  for  fusible  metals.  Zinc  would  interfere 
with  this  method,  but  as  zinc  does  not  alloy  with  lead,  we  will 
not  speak  of  it  further.     Nickel  and  cobalt  alloy  but  slightly 



with  tiu,  and  if  present,  should  be  sought  for  both  in  the  pre- 
cipitate left  by  potassium  hydroxide  and  in  the  tin  precipitated 
on  a  cylinder.  Iron  will  also  be  precipitated  with  tin  if  present 
in  an  oxalic  acid  solution.  Phosphorus  is  best  determined  by 
Dudley's  method.' 

In  alloys  containing  only  lead  and  tin,  with  the  tin  under 
twenty  per  cent. ,  the  two  constituents  can  best  be  determined  by 
treatment  with  dilute  nitric  acid  in  a  porcelain  dish,  evaporating 
to  dryness  on  a  water-bath,  etc.,  and  determining  lead  as  chro- 
mate  and  tin  as  stannic  oxide.  In  samples  free  from  iron  and 
copper,  antimony  may  be  determined  directly  by  solution  in 
hydrochloric  acid  and  potassium  chlorate,  boiling  to  expel 
chlorine,  and  titrating  as  with  pure  antimony.  Antimony  in 
solders  may  be  detennined  very  accurately  by  dissolving  in  hy- 
drochloric acid  without  access  of  air  and  filtering  out  the  un- 
dissolved antimony  011  a  weighed  Gooch  crucible.  I  have  not 
found  that  a  weighable  amount  of  antimony  was  lost  as  stibine  by 
this  treatment.  In  theanalysisof  alloys  of  lead  and  tin,  Richards" 
scales,'  which  are  accurate  within  one  per  cent.,  may  be  used. 
In  the  examination  of  the  various  classes  of  alloys  described  at 
the  beginning  of  this  paper,  various  steps  in  their  analysis  may 
be  left  out  with  the  absence  oi  the  respective  metals. 

References:  "  Phosphorus  in  Phosphor  Bronie."  Ry  F.  Julian.,  /.  Am. 
Chem.Soc.  15.  rij. 

"  Analysis  of  American  Heline(i  Copper  (lictcrminatiou  of  Cu.  Ag,  Sc, 
Te.  Bi.  Sb.  As,  Fe,  Ni,  Co,    I'b]."/-  ^"i-  Chem.  Soc,  16,  785. 

'■  The  Commercial  Valuation  of  Tin-leaJ  and  Lead -antimony  Alloys." 
By  J.  W.  Richards,/.  Am.  Chem.  Soc.  16,  541. 

■'  Materials  of  Engineering."     By  R.  II.  Thurston,  Part  III. 

"  The  Testing  of  Materials  of  Engineering."     By  W.  C.  Unwiu,  p.  341. 

"  Das  mikroskopische  Gefuge  der  Metalle  uml  Legirungen."  By  H, 
Bebrens,  Hamburg.  1S95. 


Analysis  of  Tin  Plate. 

From  two  to  three  grams  of  the  tin  plate,  cut  into  strips  two 

to  three  cm.  long  by  three  to  five  mm.  wide,  are  placed  in  a  dry 

1  Consoll  A 1.  Ene-nirr  aM  K.  K.  Journal.  >.  <%*  1)«. 


bulb  tube.  A  current  of  carefully  dried  chlorine  gas  is  then 
passed  through  the  tube,  at  first  in  the  cold  ;  it  is  then  warmed 
gently  by  a  Bunsen  flame  at  most  three  cm.  high  and  placed  at 
least  fifteen  cm.  beneath  the  bulb,  the  object  being  the  complete 
chloridization  of  the  tin,  without  any  attack  upon  the  iron.  If 
the  temperature  be  unduly  high,  the  iron  will  be  violently  acted 
upon  and  the  experiment  spoiled.  The  excess  of  chlorine, 
laden  with  stannic  chloride  is  passed  successively  through  tw'O 
Peligot  tubes  and  a  small  Erlenmeyer  flask  containing  water,  in 
which  the  tin  is  retained,  partly  as  the  tetrachloride,  partly  as 
metastannic  acid.  The  connections  of  these  tubes  should  be 
entirely  of  glass  and  cork,  un jointed  with  rubber  and  the 
delivery  tube  of  each  part  of  the  apparatus  should  reach  nearly 
to  the  bottom,  to  prevent  undue  crystallization  of  the  tin  salt 
upon  the  moist  upper  walls  of  the  condenser.  The  current  of 
chlorine  must  be  so  regulated  that,  on  the  one  hand,  no  stannous 
chloride  is  formed,  whilst  on  the  other  hand,  no  tin  is  lost  by 
the  chloride  being  swept  through  the  washing  tube ;  it  is  con- 
tinued until  the  surfaces  of  the  strips  are  uniformly  brown  with- 
out white  spots.  Stannic  chloride  condensing  in  the  narrow 
portion  of  the  bulb  tube  is  carried  forward  by  the  application  of 
gentle  heat.  The  essentials  for  success  are  dry  chlorine  and 
the  minimum  temperature  possible.— y.  Soc.  Chem,  Ind.,  i8g§, 
p.  822. 

The  Analysis  of  Tin  Plate  for  Tin^  Lead,  Iroii  and  Manganese, 

The  following  volumetric  method,  depending  on  well  known 
reactions,  has  given  very  satisfactory  results : 

Dissolve  five  grams  of  tin  or  terne  plate  in  100  cc.  hydrochloric 
acid,  I. ID  sp.  gr.,  in  a  500  cc.  graduated  flask,  with  exclusion 
of  air. 

When  dissolved,  cool  and  fill  up  to  the  mark.  Transfer  fifty 
cc.  to  a  beaker,  and  after  adding  starch  paste  titrate  the  tin  with 
standard  iodine  solution. 

A  convenient  strength  of  iodine  is  made  by  dissolving  5.38 
grams  of  pure  iodine  in  strong  potassium  iodide  solution  and 
diluting  to  one  liter. 

For  the  iron  determination  add  mercuric  chloride  in  excess  to 



fifty  cc,  of  tia  plate  solution,  and  titrate  the  iron  with  standard 

The  determination  of  manganese  is  quite  important,  since  it 
shows  whether  iron  or  iieel  has  been  tinned. 

Treat  four  grams  of  tin  plate,  cut  into  small  pieces,  with  hot 
dilute  sulphuric  acid  for  about  fifteen  minutes. 

When  the  iron  has  dissolved,  leaving  the  layers  of  tin  and 
lead,  add  a  little  zinc  and  let  stand  for  about  two  minutes.  Fil- 
ter and  dilute  to  twenty  cc. 

Take  one-half  of  this  filtrate,  add  five  cc.  nitric  acid  of  1.20 
sp.  gr.,  and  treat  in  the  ordinary  way  with  lead  peroxide. 

The  lead  in  tin  plate  is  best  determined  as  sulphate  after  first 
separating  the  tin  by  nitric  acid.  However,  for  ordinary  work, 
it  is  sufficiently  accurate  to  take  lead  by  diSerence,  allowing  0.25 
per  cent,  for  phosphorus,  carbon,  sulphur,  silicon,  etc.,  in  addi- 
tion to  the  tin,  iron  and  manganese  previously  determined. 

In  order  to  test  the  accuracy  of  the  iodine  method  for  tin,  a 
weighed  quantity  of  pure  tin,  together  with  about  forty  times 
as  much  iron,  was  dissolved  and  the  tin  titrated. 

The  result  was  as  follows  : 

.Amount  taken.  0.1255  gram  tin. 

Amount  found,  0.1266  gram  tin. 

The  following  are  a  few  analyses  that  were  made  of  British 
teme  plate  used  for  roofing  : 

Tin    1.58  ».o8  1.40  3.37  1 .60  J.54  i .97  i .96      2.56 

L«d 7-97'  7-13'  889  U.98  1.48'  7.48"  8.tsi  7.09  10.23' 

Iron 89.84  90.13  88.10  84,18  95.31  S9.35  89.19  90.55  86.64 

MaugBQCBe       0.36  0.31  0.31  0.35  0.36  0.38  0.37  0.3a      0.3a 

Carbon 1 


Salphur  ...  ["''S  "■'5  0.25  0.25  0.25  0.25  0.25  0.25     0.25 

Silicon,  etc-  j 

The  iodine  method 
alloys  which  contain  i 

99,9s  100.00  100.00  100.00  100.17  100 
may  be  used  for  determining  tin  in  1 
o  metals  that  affect  iodine. 


,  when  the  percentage  of  tin  exceeds  ten  per  cent., 


as  in  the  case  of  solder,  the  following  method,  although  not  quite 
so  simple  or  rapid,  is  somewhat  more  accurate. 

In  principle  the  scheme  is  simply  a  revision  of  the  well  known 
stannous  chloride  titration  method  for  iron.  Dissolve  five  grams 
of  the  tin  alloy  in  strong  hydrochloric  acid  in  a  500  cc.  graduated 
flask,  as  in  the  case  of  tin  plate.  After  diluting  to  the  mark, 
fill  a  fifty  cc.  burette  with  the  solution.  Transfer  ten  cc.  of  a 
standard  ferric  chloride  solution  (ten  grams  iron  in  one  liter)  to 
a  four-ounce  flask  and  heat  to  boiling.  While  boiling  run  the 
tin  alloy  solution  cautiously  into  the  ferric  chloride  until  the 
yellow  color  disappears.  Cool  and  determine  the  excess  of  stan- 
nous chloride  with  standard  iodine  solution  (Fe,Cl,  +  SnCl,  = 
2FeCl,  -f-  SnClJ. — Proceedings  Eng.  Society  of  Western  Pa.,  82, 

Chrome  Steel.' 

The  Chrome  Steel  Company  designate  its  products  as  fol- 
lows : 

No,  I. — For  turning,  planing,  and  other  tools  used  for  pur- 
poses requiring  a  steady  cut. 

No.  I  A. — Special  for  punches,  heaters,  etc. 

No,  J. — For  all  kinds  of  fine-edged  tools,  chipping  chisels  and 
machine  shop  tools  ;  a  grade  well  adapted  for  general  purposes. 

No,  2. — Milder  than  No.  3,  for  heavy  or  drop  dies  of  all 
descriptions,  and  best  quality  sledges,  etc. 

Mill  Picks, — Special  for  mill  picks,  points,  etc. 

Rock  Drill. — Special  for  mining,  quarry  and  stone  cutting,  etc. 

Tap  and  Die  Steel, — For  tap  and  dies  of  all  kinds. 

Hammer  Steel. — Cast  Spring  Steel. 

Machinery  Steel. — Of  extra  toughness  and  strength,  capable 
of  enduring  great  friction  and  resisting  heavy  strains  ;  especially 
adapted  to  mandrils,  shaftings  for  rotary  pumps,  and  other  pur- 
poses where  great  strength  is  required. 

Round  Bars  for  Prisons  or  Burglar- Proof  Gratings. — These 
bars  consist  of  alternate  layers  of  steel  and  iron  welded  together 

1  Abstract  of  Thesis,  D.  P.  Hart.  Jr.,  «nd  J.  Cmlisch :  Skmm  tmdktt!lm\  %  4HiS- 





and  designed  tor  prisons,  bank  buildings,  etc.  The  gratings  or 
bars  are  first  fitted  and  then  hardened,  the  steel  receiving  atem- 
per  that  will  resist  any  saw,  file,  or  drill ;  while  the  iron  remain- 
ing soft  and  ductile,  will  not  fracture  under  heavy  blows.  This 
combination  of  iron  and  steel  is  also  made  in  special  shapes,  and 
is  largely  used  in  safe  building.  Chrome  steel  is  also  exten- 
sively employed  in  the  construction  of  large  bridges.  Chrome 
steel  possesses  great  strength,  as  the  following  table  of  testa 
indicates  (page  328).  Tests  are  made  by  Capt.  Eads  upon  sam- 
ples of  chrome  steel  furnished  in  the  construction  of  the  Illinois 
and  St.  Louis  Bridge. 


Chromium  Delertnination . — Dissolve  two  grams  of  the  sample 
in  seventy-five  cc.  of  hydrochloric  acid  (sp.  gr.  1.12)  in  a  500 
cc.  flask  fitted  with  a  rubber  cork  containing  a  glass  tube  and  a 
Bunsen  valve  (see  page  29);  heat  gently.  During  the  solution 
carbon  dioxide  is  passed  into  the  flask  slowly  to  prevent  oxida- 
tion of  the  iron.  When  solution  is  complete,  nearly  neutralize 
excess  of  free  acid  with  sodium  carbonate  and  render  slightly 
alkaline  with  powdered  barium  carbonate.  Add  distilled  water 
nearly  to  the  containing  mark,  cork  the  flask  tightly  and  allow 
to  stand  lor  twenty-four  hours,  with  occasional  shaking. 

AH  the  chromic  oxide  and  a  small  amount  of  ferric  oxiile  are  pre- 
cipitated, whilst  all  the  ferrouschloride.  manganese  chloride,  etc., 
remain  in  solution.  Filter  off  the  precipitate  together  with  the 
excess  of  barium  carbonate,  wash  with  hot  water,  transfer  filter 
paper  containing  the  precipitate  to  a  flask  and  dissolve  in  hydro- 
chloric acid  with  heat. 

Filter,  wash  well,  and  to  the  clear  filtrate  add  ammonium 
hydroxide  in  slight  excess  and  boil. 

The  chromic  oxide  and  the  ferric  hydroxide  are  thereby  pre- 
cipitated while  all  the  barium  remains  in  solution. 

Filter,  wash  well  with  hot  water,  dry.  ignite,  and  fuse  in  a 
platinum  crucible  with  sodium  carbonate  and  sodium  nitrate. 
Extract  the  fused  mass  with  hot  water,  boil  and  filter  off  the 
residual  iron  oxide. 

The  filtrate  contains  all  the  chromium  as  the  yellow  sodium 











OB     O 
^      In 



















0\     >H 

OvOO    "-•    W    ^«    »O0    "^"^u^iOiO 

xo    DDm    9JBnb8  _    _  _  _     ^  _  ^ 











iiDm  aaBnbs  a»d       ^    _-  ^-  /s    *<-  ^k  «*s    «     -    ^  .^  _i.  ^  #s    ^-  .<    -^  -r 

«    ^     »^     « 

'qam  9JBnbs 
J9d  6pnnod  ni 
Xipp«Bi»  JO  lira'n 

*8dt(3ai  ni  8aBno9 



IN.    "O    f>« 

00   5J   fo 




lOvO     ro**)NOvO    ^^    ^^lO^'^M^rOfOfOr^^ 

S*H>>»}         lOQvO    Q     Q     i-i    «-•    "-"vOvOOO    u^^OvO^O^O^O^O 

ni  a9)9mBi<i    *90    »oO    q    o    O    Ou^ir^io»o«o»OM[>io»or* 


o  «  o 

•sdqdni  ni  qiJ8n»i 

vO   00    ON  o 
>H     O   00     ON 

so  V)   O    O 

•O  1^    lO 

On  ON  VO 

•  •      •    ^ 

M  »N      O      O 


»oo    o    r*»oiovo    Q   »oo 






O  . 

c^  : 


CO  ; 

"o  o 

CI  4> 

g  s 



chromate.  Evaporate  this  to  dryness  with  hydrochloric  acid  in 
slight  excess,  and  treat  residue  with  hot  water. 

Filter  off  any  insoluble  residue  (generally  silica),  add  five  cc. 
hydrochloric  acid  to  the  filtrate,  then  sodium  sulphite  until  the 
yellow  color  disappears,  and  heat  to  boiling.  The  chromium 
trioxide  is  reduced  to  chromic  oxide.  To  the  boiling  solution 
add  ammonium  hydroxide  in  slight  excess,  boil  five  minutes, 
filter,  wash  well,  dry,  ignite  and  weigh  as  chromic  oxide,  con- 
taining 68,62  per  cent,  chromium. 

Manganese  Defermmation. — Dissolve  five  grams  of  the  steel 
in  150  cc.  of  nitric  acid  (sp.  gr.  1.20).  Boil  until  the  bulk  is 
about  100  cc.  Add  a  few  crystals  of  potassium  chlorate  whereby 
the  manganese  separates  as  manganese  dioxide,  insoluble  iu 
strong  nitric  acid.  Boil  for  a  few  minutes,  add  fifty  cc.  cold  con- 
centrated nitric  acid,  filter  on  an  asbestos  filter,  wash  three 
times  with  concentrated  nitric  acid,  and  four  times  with  cold 
water.  Place  the  asbestos  filler  containing  the  precipitate  into 
a  beaker,  add  an  excess  of  hydrochloric  acid  and  boil  until  all 
the  chlorine  is  driven  off.  Dilute  with  water,  filter  off  the 
asbestos,  wash  well,  add  ten  cc.  acetic  acid  to  the  hot  filtrate 
and  neutralize  with  ammonium  hydroxide.  Boil,  allow  the  basic 
acetate  of  iron  to  settle.  Filter  into  a  flask  and  to  the  filtrate 
add  a  few  cubic  centimeters  of  ammonium  hydroxide  and  then, 
carefully,  sufficient  bromine  to  precipitate  the  oxide  of  manga- 
nese. Cork  and  allow  to  stand  twelve  hours,  filter,  dry,  ignite 
and  weigh  as  Mn,0,  containing  72.08  per  cent,  ot  manganese. 

Silicon  Determination. — Five  grams  of  the  steel  are  dissolved 
in  thirty  cc.  dilute  sulphuric  acid  (one  part  sulphuric  acid  to 
two  parts  water).  When  solution  is  complete,  add  strong  nitric 
acid  until  no  more  effervescence  occurs.  Evaporate  to  dryness, 
moisten  with  hydrochloric  acid  and  dissolve  in  excess  of  boiling 
water.  Filter  off  the  silica,  wash  with  dilute  hydrochloric  acid 
and  hot  water,  dry,  ignite  and  weigh  as  SiO,  containing  46.7  per 
■  cent,  of  silicon.     This  process  is  used  when  tungsten  is  absent. 

Determination  of  Ttings/^i  .—Dissolve  five  grams  of  the  steel 
in  a  three-inch  porcelain  evaporating  dish  with  twenty  cc.  hydro- 
chloric acid  (strong)  and  fifty  cc.  of  strong  nitric  acid,  and 
evaporate  to  dryness. 



The  presence  of  tungsten  is  at  once  indicated  by  the  yellow 
color  of  the  tungstic  acid  (WO,),  which  separates  with  the  silica. 
Add  fifty  cc.  water,  ten  cc.  hydrochloric  acid,  warm  and  filter ; 
wash  with  water  acidulated  with  hydrochloric  acid  to  prevent 
any  tungstic  acid  passing  through  the  filter. 

The  tungstic  acid  is  then  dissolved  on  the  filter  in  hot 
ammonium  hydroxide  and  is  thus  separated  from  the  silica. 
The  filtrate  is  concentrated  so  as  to  allow  of  its  being  transferred 
into  a  weighed  platinum  crucible,  in  which  it  is  evaporated  to 
dryness,  ignited  and  weighed  as  WO,. 

Carbon ^  phosphorus  2ind  sulphur  are  determined  by  the  usual 
methods  in  steel  analysis. 

The  following  analyses,  made  in  the  laboratory  of  the  Insti- 
tute, show  the  composition  of  some  of  the  varieties  of  chrome 

steel : 

No.  I  Steel. 


P i 





Fe  (difference) 

1. 1077  per  cent. 












t  ( 

No.  3  Steel. 

C 0.7253  per  cent. 

P ooiS' 






Fe  (difference) 










**Magnet  Steel.'* 



Cr 0.4940 

W 0.6186 

Si 0.0550 

Mu 0.0167 

S 0.0043 

Fe  (difference ) 97.8021 

0.9571  per  cent. 
0.0522    ** 






•  < 


chemical  and  physical  examination  of  t 
•■  Rock  Drili,"  Steel. 

Fe  (difierence) 98.4, 

The  Chemical  and  Physical  Examination  of  Paper. 

This  subjeci  may  be  conveniently  divided  into  eight  sections : 

First. — Determination  of  the  nature  of  the  fiber. 

Second. — Microscopical  examination. 

Third. — Determination  of  free  acids. 

Fourth. — Determination  of  the  nature  and  amount  of  the  sizing 

Fifth. — Determination  of  the  amount  of  ash  and  its  analysis. 

Sixth. — Determination  of  the  weight  per  cubic  decimeter. 

Seventh. — Determination  of  the  thickness  of  the  paper. 

Eighth. — Determination  of  the  absolute  breaking  strength. 
First. -^Determination  of  the  Nature  of  the  Fiber. 

The  introduction,  in  late  years,  of  the  various  kinds  of  wood 
fibers  in  the  manufacture  of  paper  has  rendered  the  chemical 
examination  of  the  same  exceedingly  difficuU. 

This  is  more  especially  so  where  the  wood  fiber  has  been  sub- 
jected to  chemical  treatment,  as  in  the  "  sulphite  process"  or 
the  "  soda  process,"  before  being  incorporated  in  the  paper. 

Nearly  all  oi  the  chemical  reactions  for  the  recognition  of  the 
wood  fibers  in  paper  produce  certain  colors  with  the  various 
resins  in  the  wdod  when  the  reagent  is  added.  While  the  fiber 
prepared  entirely  by  the  "  mechanical"  process  can  beindicated 
without  difficuhy,  even  when  mixed  with  cotton  and  linen  in 
various  amounts,  the  conditions  are  greatly  altered  when  the 
wood  fiber  has  been  subjected  to  bleaching  and  chemical  treat- 
ment, since  the  latter  removes  much  of  the  resinous  matters  of 
'  "the  wood  and  increases  the  difficulty  of  the  qualitative  examina- 


The  chemical  reactions  of  the  fibers  produced  from  the  various 
woods  used  in  paper-making,  pine,  poplar,  and  spruce,  are 
identical,  qualitatively,  with  the  following  reagents. 

1 .  Hydrochloric  acid  and  phloroglucin  produce  a  red  color 
with  *'  mechanical* '  wood  pulp. 

2.  Aniline  sulphate  produces  a  yellow  color. 

3.  Naphtylamin  and  hydrochloric  acid  produce  an  orange  yel- 
low color. 

4.  Anthracene  hydrochlorate  produces  a  red  color. 

5.  Phenol  Hydrochlorate  produces  a  bluish-green  color. 

6.  Concentrated  hydrochloric  acid  produces  a  violet  color. 

7.  Pyrrol  and  hydrochloric  acid  produce  a  purple  red  color. 

8.  Pyrogallic  acid  and  zinc  chloride  produce  a  dark  violet 

9.  Nitric  and  sulphuric  acids  produce  a  red  color. 

10.  Haeraatoxylin  solution  produces  a  red  color. 

11.  Alcoholic  solution  of  cochineal  produces  a  blue  violet. 
Where  the  wood  pulp  is  composed  entirely  of  **  mechanical* ' 

wood  fiber  the  above  reactions  are  very  marked,  and  by  the  aid 
of  the  microscope,  the  varieties  of  wood  can  be  determined. 

Wood  pulp  produced  by  the  *'  soda**  or  by  the  **  bi-sulphite'* 
process  gives  a  much  weaker  reaction  with  the  chemical  reagents 
used  for  identification,  and  in  many  instances  where  the  pulp 
has  been  used  many  times  in  paper-making  will  give  no  color 
reactions  sufficient  for  recognition.  The  amount  of  **  mechanical 
fiber"  in  a  mixture  of  **chemical  fiber,*'  linen  fiber,  cotton  fiber 
and  **  mechanical**  fiber  in  a  paper  can  be  determined  as  follows : 

The  sample  of  paper  is  first  boiled  in  water,  then  in  alcohol, 
and  afterwards  digested  with  ether.  After  drying,  a  solution  of 
chloride  of  gold  is  added. 

Linen,  cotton,  and  **  chemical**  wood  fiber  have  no  reducing 
action  upon  the  solution  of  gold  ;  but  the  piechanical  wood  fiber 
immediately  reduces  gold  from  the  solution,  this  action  being 
due  to  the  ligno-cellulose  remaining  in  the  mechanical  wood 

TOO  grams  of  mechanical  wood  pulp,  under  above  conditions, 
will  reduce  14,285  grams  of  gold.* 

1  Handbuch  der  Technisch-Chem.  Untersuchunyen  (BOUXV).    6ttl  Avf^  P^S*  <<>(>7- 




If  a  sample  of  paper  be  submitted  for  examination  as  to  the 
fibers  used  in  its  manufacture,  the  following  preliminary  work 
is  requisite  :  The  rosin,  sizing,  filling,  etc, .  in  the  manufactured 
paper  must  first  be  removed.  Cut  the  paper  into  small  pieces, 
place  them  in  a  beaker  and  digest  with  a  solution  of  caustic 
soda  (one  part  caustic  soda  to  thirty  of  water},  at  a  moderate 
heat  for  ten  minutes.  Pour  off  the  liquid,  replace  with  dou- 
ble theamount  of  distilled  wat^r.  and  warm  ten  minutes  ;  pour  off 
this  liquid,  and  repeat  once.  Now  place  the  paper  in  a  solution 
composed  of  one  part  of  hydrochloric  acid  and  fifteen  parts  of 
distilled  water  and  digest  ten  minutes.  Wash  a  number  of  times 
with  distilled  water,  until  washings  are  no  longer  acid  :  then 

Suppose  the  sample  of  paper  so  treated  to  be  composed  of  a 
mixture  of  ■■mechanical"  chemical  wood  fiber,  linen,  and  cotton 
— a  mixture  to  be  found  in  many  samples  of  good  quality  of 
TBfH  ting- paper. 

A  sample  of  the  dried  paper  is  tested  with  solution  of  gold 
chloride.  If  no  reduction  of  gold  takes  place,  the  indications 
point  to  the  absence  of  mechanical  wood  fiber.  This,  however, 
I -is  not  absolute,  since,  if  the  paper  has  been  made  from  ■'  cut- 
.tings,"  "  old  paper  stock,"  etc.,  etc..  the  mechanical  wood  pulp 
juighl  have  been  treated  quite  a  number  of  times  by  chemicals 
in  the  production  of  the  finer  quality  of  paper,  and  its  ligno- 
cellulose  destroyed  or  modified  in  such  a  way  as  to  nullify  the 
gold  test. 

Generally  speaking,  however,  the  reduction  of  the  gold 
chloride  is  indicative  of  the  presence  of  mechanical  wood 

R.  Benedikt'  gives  a  method  for  the  determination  of  mechaui- 
wood  fiber  in  paper,  dependent  upon  the  methyl  numbers  of 
contained  in  it.  This  process  has  been  tested  by  W. 
■'  with  the  result  that  preference  is  given  to  the  use  of 
chloride  solution. 

aus  dtm  R.  K.  TccDiialotclecheii 


If  the  amount  of  mechanical  wood  fiber  in  a  paper  amounts  to 
about  ten  per  cent.,  Gottstein'  states  that  the  fibers  may  be 
counted  under  the  microscope,  after  the  fibers  have  first  been 
made  visible  by  a  treatment  with  an  alcoholic  phloroglucinol 
solution  and  hydrochloric  acid.  Fifteen  per  cent,  or  more  of  the 
mechanical  wood  fiber  in  the  mixture  renders  the  test  valueless. 
If  chemical  wood  fiber  be  present  in  a  paper  with  mechanical 
wood  fiber,  no  color  tests  for  the  •former  are  positive  in  the  pre- 
sence of  the  latter,  since  the  mechanical  wood  pulps  possess  a 
greater  tinctorial  power. 

Should  mechanical  wood  fiber  be  absent,  however,  a  solution 
of  resorcin  can  be  applied  to  a  properly  prepared  sample  of  the 
paper.  Chemical  wood  fiber  produces  a  violet  color,  whereas 
cotton  and  linen  are  without  action. 

A  solution  of  phenol  also  produces  a  violet  color  under  similar 

Second. — Microscopical  Examiyiation . 

By  careful  manipulation  of  the  microscope,  the  fibers  of  linen, 
cotton,  and  the  various  woods  can  be  recognized. 

The  distinction  must  be  noted  here,  however,  that  the  fibers 
from  paper,  no  matter  what  the  source,  do  not  have  the  appear- 
ance under  the  microscope  that  they  possessed  before  the  me- 
chanical and  chemical  treatment  required  in  the  manufacture  of 

The  chemical  process  in  paper-making  is  very  severe  upon  the 
various  fibers,  since  they  are  subjected  to  beating  and  cutting 
in  the  "  beating-machine,"  to  protracted  maceration  in  strong 
alkali,  to  digestion  in  boiling  water,  to  bleaching  with  chloride 
of  lime,  are  loaded  with  various  clays,  and  finally  are  sized,  and 
often  burnished. 

This  difference  between  linen  fibers  before  and  after  treatment 
is  shown  in  Figs.  92  and  93. 

A  comparison  shows  not  only  a  radical  change  in  the  form  of 
the  fibers,  but  a  difference  in  the  transparency,  due  to  removal 
of  soluble  portions  of  the  fiber. 

Poplar  wood  fiber  (Fig.  94)  made  by  chemical  process,  under 

1  Papier-Zeitung.  1884,  432. 


the  microscope  resembles  the  fibers  of  linen  more  than  does  any 
of  the  wood  fibers.  It.  however,  has  one  distinguishing  charac- 
teristic, even  among  the  disintegrated  pulps,  that  is.  the  tangen- 

tial fragments  have  among   them  particles  hearing  a   grate,  or 
screen-like,  appearance,  as  shown  in  Fig.  95. 

The  coniferous  woods  used  in  paper-making  show  peculiarities' 
in  structure  entirely  different,  under  the  microscope,  from  linen 
and  cotton,  the  most  distinctive  one  being  the  small  circular 


"pits'"  or  spots  along  the  center  of  each  fiber.  A  section  oi 
spruce  wood,  composed  of  fifteen  or  more  fibers,  is  shown  in 
Fig.  96.  J 

After  pulpiug  and  making  into  paper,  spruce  fiber  has  the  ap-"* 


pearance,  under  the  microscope,  .■ihown  in  Fig.  97.      it  s 
tains  the  peculiar  circular  markings,  and  is  readily  distinguishdj 
from  the  linen  paper  fiber,  Fig.  93,  or  from  cotton  fiber,  Fig.  ■ 
In  Fig.  99  is  shown  the  peculiar  "  center-marking"  of  coniffl 



erous  6ber,  as  taken  from  a  sample  of  writing  paper  sold  as  linen 
paper,  but  shown  by  both  chemical  and  microscopical  examina- 
tion to  be  composed  largely  of  spruce  fiber  and  linen.' 

The  microscope  will  thus  determine  the  differences  between 
the  various  fibers  used  in  paper-making,  and,  by  properly  ar- 
j  ranged  apparatus  connected  therewith,  the  percentage  of  each 
f  variety  of  fiber. 

According  to   the  German  official  directions,  the  sample  of 

paper,  after  removal  of  sizing,  etc.,  is  to  be  steeped  in  a  solution 

of  one-fifth  gram  of  iodine  and  two  grams  of  potassium  iodide 

Uin  twenty  cc.  of  water  and  then  examined  under  the  microscope. 

K^be  fibers  may  be  conviently  divided  into  three  groups  : 

.  Linen,  hemp,  and  cotton. 

,  Wood -cellulose  ("chemical"  wood  fiber),  straw-cellulose, 
land  esparto. 

,1-  Ground  wood- cellulose  and  jute. 

After  treatment  with  the  above  solution,  the  fibers  of  group  i 
I  are  stained  brown,  those  of  group  2  are  not  colored,  whilst  the 
I  strongly  lignified  fibers  of  group  3  are  colored  yellow.  But  it 
L  lias  been  found  that  this  method  is  somewhat  defective ;  the 
I'Cellulose  of  group  2,  for  example,  being  invariably  to  some  ex- 
I  tent  stained,  whilst  the  members  of  group  r  are  so  deeply  colored 
^hat  it  is  almost  impossible  to  distinguish  their  structural  char- 
lacters.  After  many  experiments,  the  following  method  was 
Kfoand  more  satisfactory. 

The  paper  is  placed  on  the  object-glass  of  the  microscope  and 
f  treated  with  iodine  solution,  the  unabsorbed  iodine  removed  by 
■means  of  filter  paper,  and  the  paper  covered  with  sulphuric  acid 
■.dilute.     The  solution  of  iodine  in  potassium  iodide  should  be  of 
Jguch  a  strength  that  a  layer  of  three  cc.  thickness  should  be  of 
"ruby-red  color  and  quite  transparent.     The  paper  is  now  removed 
and  boiled  with  a  solution  or  dilute  potassium  hydroxide,  washed 
thoroughly,  and  replaced  on  the  object-glass.     The  color  reac- 
tions are  as  follows  : 

I.  Cotton,  linen,  and  hemp  take  a  violet-red  or  wine-red  color. 

1  The  micro- phatOBiB phi  uicd  la  thii  article  an  from  apecimcai  made  during  ■<■ 
lovMlisalion  upon  fihers  of  papers  by  Charles  S.  Shullx  and  the  wrilet  la  1893,  and  rep- 


2.  Well  bleached  wood-cellulose  and  ordinary  bleached  strai 
cellulose  are  colored  gray-blue  or  pure  blue,  without  any  1 
of  red. 

3.  Unbleached  or  imperfectly  bleached  wood  fiber  absorbs 
little  iodine  and  remains  colorless. 

4.  Strongly  lignified  fibers,    such  as  ground  wood-cellu]< 
and  raw  jute,  are  colored  yellow. 

The  numbers  of  each  variety  of  fiber  are  now  carefully  counted 
by  means  of  the  microscope  and  an  eye-piece  micrometer  ruled 
in  squares.  This  chemical  treatment  and  microscopical  ex- 
amination is  to  be  repeated  upon  at  least  fifty  diSereiit  pieces  of 
paper  from  difierent  parts  of  the  sample,  and  an  average  taken 
By  this  means  approximate  percentages  of  each  variety  of  fibei 
in  the  paper  can  be  stated." 

Third. — Determination  of  Free  Acids  in  the  Papet 

Free  acids  in  the  paper  may  be  : 

1.  Chlorides,  from  the  hypochlorites  used  in  the  bleach ii 
and  which  have  not  been  removed  by  the  "  anti-chlor." 

2.  Sulphuric  acid,  from  acid  alums  used  in  the  sizing. 
Free  acids  are  exceeciingly  injurious  to  the  paper,  producing 

gradual  deterioration  in  the  breaking  strength,  and  alsoproduc 
ing  brittleness. 

The  amount  of  chlorides  can  be  determined  as  follows 
Take  five-tenths  gram  of  the  paper,  cut  into  small  portioi 
and  digest    with  fifty  cc.   of   boiling   distilled    water   for 
minutes,  then  filter.     The  filtrate  is  acidified  with  a  few  di 
of  nitric  acid,  and  the  amount  of  chlorine  determined  by  a 
tenth  normal  silver  nitrate  solution. 

The  free  sulphuric  acid  determination  requires  the  determina- 
tion of  the  combined  sulphuric  acid  in  the  alum,  since  in  the 
titration  with  soda  solution  the  combined  acid,  as  well  as  the 
free,  is  indicated.  The  combined  acid  is  determined  indirectly 
and  then  substracted  from  the  total  acid,  the  difference  being 
the  free  acid,  thus  :  If  the  alum  used  is  potash  alum,  the  per- 
centage of  potash  should  be  determined,  and  then  the  amount 
of  sulphuric  acid  and  alumina  calculated  from  the  formula  of 
the  alum  (anhydrous)  K,,A1,(S0.),. 






If  soda  or  ammonia  alum  be  used,  the  determination  of  the 
soda,  or  ammonia,  will  be  required.  Wliere  no  clay  has  been 
used  in  the  paper,  the  alum  can  be  determined'  instead  of  the 
other  base,  and  the  sulphuric  acid  necessary  to  form  the  alum 
calculated:  this  latter  is  then  deducted  from  the  total  acid. 
Total  acid  is  thus  determined: 

Two  grams  of  the  paper  are  cut  into  small  pieces  and  digested 
with  200  cc.  of  boiling  distilled  water  for  three  miuutes,  then 
filtered  and  a  few  drops  of  solution  of  litmus  added.  A  solution 
of  tenth-normal  soda  is  gradually  added  from  a  burette,  until 
the  red  color  of  the  solution  turns  to  blue,  when  the  amount  of 
alkali  used  is  noted  and  calculated  to  sulphuric  acid. 

From  the  total  amount  of  sulphuric  acid  is  subtracted  the 
combined  sulphuric  acid  already  determined  in  two  grams  of 
paper.  This  latter  amount  is  found  by  determination  of  any 
of  the  bases,  alumina,  potash,  soda,  or  ammonia,  and  calculation 
of  the  required  acid  necessary  to  form  the  alum  used  in  the  paper, 

If  aluminum  sulphate,  AI,(SO,),,  be  used  instead  of  alum, 
then  the  free  acid  and  combined  acid  will  be  the  same  in  amount, 
since  aluminum  sulphate  is  an  acid  salt,  and  titration  with  the 
soda  solution  will  give  the  amount  directly. 

Fourth. — Detenniitation  of  the  Nature  and  Amount  of  Sizing 

A  paper  sized  with  rosin,  when  extracted  with  absolute  alcohol, 
gives  a  solution  a'liich,  poured  into  excess  of  water,  yields  a 
milky  turbidity  due  to  precipitated  rosin.'  Another  test  is  based 
on  the  Raspail  reaction,  rosin  giving,  with  sugar  solution  and 
sulphuric  acid,  a  violet-red  color.  The  sugar  may  be  omitted, 
as  enough  is  formed  for  the  reaction  by  the  action  of  the 
sulphuric  acid  on  the  cellulose  of  the  paper. 

The  presence  of  animal  size  is  detected  by  treating  the  aqueous 
extract  of  the  paper  with  tannin.  The  following  fundamental 
distinction  between  papers  sized  with  rosin  and  gelatin  is  found 
to  exist.  In  the  former  the  rosin  is  distributed  uniformly 
throughout  the  substance  of  the  paper,  while  in  the  latter. 
whether  the  sizing  has  been  performed  in  the  pulp  or  sheet,  it  is 

339    H 
the    I 

»W.  Herwbirg  :  % 

•  ulphBte  ot 



always  found  exclusively  oil  the  surface  of  the  finished  product. 
This  peculiar  property  of  gelatin  can  be  shown  by  saturating  a 
plaster-of- Paris  slab  with  gelatin  solution  colored  suitably,  and 
breaking  it  when  dry,  on  which  it  will  be  found  to  be  colored  to 
a  trilling  depth,  the  inner  part  being  white.     On  these  facts  the 
following  test  is   based  :    A  half-sheet  oi    paper  is   repeatedly 
crumpled  and  unfolded,  and  when  the  surface  has  been  thoroughly 
chafed,  is  smoothed  out   and  written   upon  :   if  it  is  sized  with 
rosin,  the  inscribed  characters  are  but  little  blurred  ;  while,  if 
animal  size  has  been  used,  they  run  freely,  and  are  visible  from 
the  opposite  side  of  the  sheet.      Leonhardi  has  modified  this  test, 
removing  the  doubtful  element  introduced  by  the  manual  use  of 
pen  and  ink.     A  pipette,  of  which  the  exit  is  ten  cm.  above  the 
paper,  and  which  delivers  drops  weighing  0.05  graui  each,  is 
filled  with  a  solution  of  ferric  chloride  containing  1.531  per  cent, 
of  iron.     A   single  drop  is  allowed  to  fall  and   to  remain  on  the   I 
paper  for  the  same  number  of  seconds  that  onesq.  ni.  of  the  papwl 
weighs  in  grams,  when  it  is  removed  by  blotting  paper,  and  the  t 
under  side  of  the  paper  brought  in  contact  with  a  plug  of  wad- 1 
ding  wet  with  a  weak  solution  of  tannin  :  the  production  of  aa 
black  color  proves  the  iron  solution  to  ha\'e  penetrated,  and,! 
therefore,  shows  the  sizing  to  be  of  animal  origin, 

Schuman's  method  for  the  determination  of  rosin  in  paper  isl 
as  follows  :  Two  grams  of  the  paper  are  cut  into  fine  piecofl 
and  digested  below  boiling  fifteen  minutes  with  a  five  per  cent.^ 
solution  of  sodium  hydroxide,  and  filtered. 

The  filtrate  is  made  acid  with  dilute  sulphuric  acid,  the  rosiaj 
separating  and  rising  to  the  surface  of  the  liquid.  This  latter  isJ 
filtered  upon  a  weighed  filter,  dried  at  100°  C.  to  constant  weight,  T 
and  its  weight  carefully  determined. 

Starch  was  used,  formerly,  as  a  sizing  for  paper,  but  iu  recent  j 
years  it  has  been  largely  replaced  by  rosin  size.  It  can  be  de>J 
tected  as  follows : 

The  paper  is  cut  into  small  portions  and  is  digested  with  boit-'l 
ing  water  for  fifteen   minutes,  then  filtered.     To  the  filtrate  i 
added  a  drop  of  a  dilute  solution  of  iodine.     A  blue  coloration  v 
indicative  of  the  presence  of  starch. 

The  quantitative  determination  is  dependent  upon  the  conver-J 





sion  o(  starch  into  glucose  by  means  of  dilute  sulphuric  acid, 
aud  estimation  by  means  of  Fehling's  solution. 

Ten  to  fifteen  grams  of  the  paper  are  digested  with  250  cc.  of 
distilled  water,  to  which  has  been  added  two  per  cent,  of 
sulphuric  acid.  Two  or  three  hours'  heating  at  100°  C.  is  suffi- 
cient to  convert  the  starch  into  glucose,  the  exact  point  being 
determined  by  taking  a  drop  of  the  solution  and  adding  thereto 
one  drop  of  the  dilute  iodine ;  if  no  blue  color  is  shown,  the 
conversion  is  complete. 

The  solution  is  now  made  alkaline  with  soda,  diluted  with 
water  to  500  cc,  and  two  samples  each  of  150CC.  taken,  filtered, 
washed  well  and  treated  with  Fehling's  solution,'  as  usual  in 
the  determination  of  sugars.  Sad  tier  states  as  follows  regarding 
this  test : 

"  In  carrying  out  the  gravimetric  method  the  Fehling's  solu- 
tion remains  in  excess  (indicated  by  the  blue  color  of  the  solu- 
tion after  boiling),  while  the  cuprous  oxide  is  carefully  filtered 
off  and  further  treated." 

The  procedure  is  as  follows  ;' 

■'  Sixty  cubic  centimeters  of  the  mixed  Fehling's  solution  and 
thirty  cubic  centimeters  of  water  are  boiled  in  a  beaker,  and  the 
solution  containing  the  maltose  added  thereto  and  the  mixture 
again  boiled.  It  is  then  filtered  with  the  aid  of  a  filter-pump, 
upon  a  Soxhiet  filter  (  asbestos  layer  in  a  tared  funnel  of  narrow 
cylinder  shape),  quickly  washed  with  hot  water,  and  then  with 
alcohol  and  ether,  and  dried.  The  asbestos  filter,  with  the 
cuprous  oxide,  are  now  heated  with  a  small  flame,  while  a  cur- 
rent of  hydrogen  is  passed  into  the  funnel,  so  that  the  precipi- 
tate is  reduced  to  metallic  copper  It  is  allowed  to  cool  in  the 
current  of  hydrogen,  placed  for  a  few  minutes  over  sulphuric 
acid,  and  then  weighed." 

Fifth. — Detfrmination  of  the  Ash. 

Three  grams  of  the  paper  are  transferred  to  a  weighed  plati- 

IToUen's  lornmU  tor  FehliuK's  solulion  iaae  follows;  34.639  grarai  cr>-»lalli«ed 
copper  >d Ip bale  are  disulved  in  jooce.  water.  173  arams  Rocbelle  mIW  and  aiily  giarot 
■odium  hydroxide  an  dissolved  logelher  in  Joo  cc.  ol  water.  Hqual  I'oluinea  oi  these 
•olutloni  are  loiied  when  requited  lor  use.  Ten  cc  of  lhi»  fehling's  lolulion  correspond 
too.oSoi  gram  maltoae— or  0.0765  gram  starch, 

iSadlltr.-  Industrial  Organic  Chemislry,  p.  iji. 


num  crucible  and  ignited  until  all  carbonaceous  matter  is  con- 
sumed. The  amount  of  ash  is  indicative  of  the  use,  or  not,  of 
mineral  filling,  such  as  Carolina  kaolin,  to  increase  the  weight 
of  the  paper.  After  the  correct  determination  of  the  amount  of 
the  ash,  it  should  be  transferred  to  a  3-inch  porcelain  capsule, 
and  the  scheme  on  the  opposite  page  used  for  its  analysis. 

It  is  always  advisable  to  test  some  of  the  ash,  before  its 
analysis,  by  fusing  a  portion  on  charcoal  with  sodium  carbonate. 
By  this  means,  lead  or  chromium  can  be  detected,  and  then 
properly  separated  in  the  analysis  of  another  portion  of  the  ash. 
If  clay,  in  appreciable  quantities,  is  found,  it  will  be  necessary 
to  add  ten  per  cent,  of  its  weight  as  water,  since  most  clays 
contain  from  eight  to  twelve  per  cent,  of  water,  which,  in  the 
above  instance,  would  have  been  driven  off  during  the  ignition 
of  the  paper  to  determine  the  per  cent,  of  ash.  If  much  iron  be 
found,  Prussian  blue,  Indian  red,  Venetian  red,  or  ochre  may 
have  been  used.  If  the  color  of  the  ash  is  blue,  ultramarine  is 
present ;  if  white,  silica,  or  a  fine  quality  of  clay,  or  calcium  sul- 
phate, or  agalite*  may  be  present ;  the  chemical  analysis  readily 
showing  the  one  used  as  a  filler. 

If  the  ash  found  is  very  small  in  amount,  it  will  be  necessary 
to  subtract  the  amount  of  ash  corresponding  to  the  variety  of 
fiber  or  pulp  with  which  the  paper  is  made,  to  exactly  determine 
the  amount  of  ash  belonging  to  the  added  materials. 

Ash  in  Commercial  Pulps. 

Sulphite 0.48  per  cent. 

Sulphite,  bleached 0.42  '*  ** 

Soda 1.34  ** 

Soda,  bleached 1.40  **  •* 

Straw 2.30  ** 

Straw,  bleached 1.34  '*  ** 

Ground  wood  (pine'i 0.43  **  ** 

Ground  wootW, fir) 0.70  **  *• 

Ground  wooil  (aspen) 0.44  **  ** 

Ground  wooil  (^ lime) 0.40  **  •* 

Linen 0.76  **  ** 

Linen,  bloachetl 0.94  **  ** 

Cotton 0^1  *«  " 

Collon.  bleached 0.76  "  ** 

I  A  vnricty  of  tnlc—Mlicntc  of  iiiiigne$ium— in  a  finely  powdered  condition ;  it  has  a 
vcr>'  cxieusivc  use  nii  i>ni»er  filler. 



K    »  =  c    » 





on  g  Be 








■          111 






.  i 
















"0„"^?T.  ".  OS. 

































■  1 


""""'" J 



Ash  in  Fibers. 

Cotton O.I2  per  cent 

Italian  hemp 0.82 

Rhea 5.63 

Best  Manilla  hemp   1.02 

Sulphite  fiber 0.46 

Fine  Flemish  flax 0.70 

China  grass 2.87 

Jute 1.32 

Esparto 3.50-5.04 

Soda  fiber 1.00-2.50 

Sixth, — Determination  of  the  Weight  per  Square  Meter. 

It  is  best  to  use,  when  possible,  five  different  pieces  of  the 
paper  (from  different  packages  or  rolls) ,  each  piece  about  one 
square  decimeter. 

These  are  placed  in  a  drying  oven  and  exposed  to  a  tempera- 
ture of  105®  C.  until  the  weight  becomes  constant.  The  weight 
of  the  five  pieces,  multiplied  by  20,  gives  the  weight  of  one 
square  meter  of  paper.' 

Seventh, — Determination  of  the  Thickness. 

The  thickness  of  paper  can  be  accurately  determined  by  the 
apparatus,  a  sketch  of  which  is  shown  in  Fig.  100. 

By  means  of  a  delicate  spring,  a  lever,  ^„  is  held  against  5,, 
touching  5,  only  at  one  point.  S^  carries  a  toothed  segment, 
which  moves  a  pointer,  2,  along  an  arc  divided  into  500  parts. 
One  division  represents  0.002  mm.  of  thickness  of  the  paper  tested. 

Eighth. — Detenninatian  of  Breaking  Strength. 

By  the  strength  of  a  paper  is  understood  the  measurement  of 
the  resistance  it  offers  to  breaking  or  tearing  strains.  This  re- 
sistance is  always  greater  in  the  direction  of  the  length  of  the 
web  of  paper,  as  it  is  made  on  the  paper-machine,  than  across 
the  web.  On  the  other  hand,  the  amount  of  elongation,  which 
is  measured  while  determining  the  breaking  strain  is  greater  in  the 
direction  across  the  web  than  parallel  to  it.*  The  tensile  strength 
of  the  sheet,  both  across  and  parallel  to  the  web.  is  determined 
separately,  and  the  average  values  recorded.     To  ascertain  the 

1 1.citfaden  f iir  Paptcr-priifunif ,  W.  Hcrzberg.  Berlin,  iSSS. 

a  Vcrhandlung  dc*  Vercins  «ur  Befordening  des  Gewerbefleiascs  in  Preussen.  1885. 


direction  cor- 
responding to 
the  motion  of 
the  paper  ma- 
chine, in  any 
pie  of  ma- 
paper,  a  cir- 
cular piece 
is  cut  and 
placed  on  the 
surface  of  wa- 
ter, when  it 
will  be  o b - 
ser\'ed  to  roll 
up.  The  di- 
ameter of  the 
disk  where  it  is  not 
curved  indicates  the  di- 
rection of  the  length  of 
the  web.  The  strips  of 
paper  used  for  ascertain- 
ing the  tensile  strength 
and  elongation  are  cut 
to  the  following  size : 
ig  by  fifteen 
mm.  broad.  Five  strips, 
at  least,  are  taken  from 
different  sheets  and  rep- 
resenting the  length  and 
the  web,  in  order 
to  obtain  good  average 
values.  These  strips 
must  be  carefully  cut  ; 
the  edges  should  be  smooth  and  run  parallel.  Cutting  tools  are 
provided  for  this  purpose,  consisting  of  an  iron  ruler  and  plates 
of  zinc  or  glass. 

H      346 

QUANTITATIVE   ANALYSIS.                                             ^M 

■                 1^1 

Before  determining  the  tensile  strength    f 

■        m 

;      p. 

and  elongation,  careful  attention  must  be 

■       Jm 


paid  to  the  anionnt  ol  moisture   in  the 

1  -^m 


atmosphere.     The  breaking  strain  of  pa- 


per  decreases  with  increase  of  moisture     M 
in  the  air,  while  under  the  same  inSu-     H 


ence  the  percentage  amount  of  elongation    ^| 


increases.    The   humidity  of  the  atmos-     H 


phere  is  very  important    when    testings    ^M 
animal-sized  paper  and  should  on  no  ac-     ^ 
count  be  overlooked.  Indeed,  the  break- 


■'■  ■ 


ing  strain  values  cau  only  be  compared 



when  they  are  obtained  in  atmospheres 
oi  equal   humidity.     The  percentage  of 
atmospheric  humidity  chosen  is  65,  he- 
cause  it  is  much  easier  to  add  moisture 
to  the  atmosphere  than  abstract  moisture 
S  from  it.     The  former  is  done  by  boiling 


^  water  in  the  room.    The  instrument  in 

use  for  measuring  the  humidity  of  the 


air  is  the  Koppe-Saussure's  air  hygrom- 


eter.     Before  testing,  the  strips  of  paper 


are  placed  in  the  room  for  at  least  two 



hours.     The  principal  machines  in  use 



for  determining  the  breaking  strength  of 

paper  are  : 



The  Hartig^Rensch,  the  WendUr  and 


the   Chopper    Apparatus,  a  description 
of  the  WendUr  being   given   herewith. 

1     [1 

This   machine   is  used  for  ascertaining 

1    j| 

the  strength  and  elasticity  of  paper.      It 



consists  in  the  main  of  fourparts.    (Fig. 


1.  The  driver. 


2.  Apparatus  for  mounting. 


3.  Apparatusfortransmissionofpower.  ^H 

4.  Appa 

atus  lor  measuring  force  and  stretch.                             ^^M 



The  driving  is  produced  by  a  haud-wheel ,  a.  The  hub  of 
this  wheel  turns  in  the  bearing  /,  which  is  cast  in  one  piece  with 
the  bed  d.  The  screw,  6.  is  led  through  this  hub,  which  is  hol- 
low, and  is  fastened  to  the  slide  c,  and  through  its  agency  the 
slide  is  moved.  The  hand-wheel  is  equipped  with  a  boh-nut, 
consisting  of  the  shell  fi,  and  two  split  nuts,  which  may  be  opened 
or  closed  by  means  of  a  worm,  according  as  the  motion  of  the 
slide  is  to  be  produced  by  the  hand  alone  or  through  the  agency 
of  the  wheel. 

The  mounting  apparatus  consists  of  two  clamps  ii,.  the  first 
fastened  to  the  carriage  w,  the  second  to  the  slide  e.  Between 
the  jaws  of  these  clamps  the  paper  to  be  tested  is  stretched. 
The  jaws  of  these  clamps  are  normal  to  the  axis  of  stress,  wave- 
shaped,  and  are  lined  with  leather,  in  order  to  prevent  the  slip- 
ping of  the  strip  in  the  clamps.  The  jaws  are  pressed  together 
by  means  of  the  screws  j,  s,. 

The  transmission  of  the  force  is  done  in  this,  as  in  most  of  this 
class  of  machines,  by  means  of  a  spiral  spring,  those  of  Weudler's 
apparatus  possessing  respectively  a  maximum  force  ol  nine  and 
twenty  kilos,  The  spring  is  held  at  one  end  by  means  of  the 
shell  /',  which  is  fastened  to  the  bed  rf,  at  the  other  by  the  car- 
riage w,  and  passes  through  the  shell  /.  Fastened  to  the  bed  by 
means  of  screws  are  ihe  catches^,  which  work  in  the  teeth  of 
the  rack,  and  which,  as  soon  as  the  paper  tears,  prevent  the 
spring  from  flying  back. 

The  measurement  of  the  force  is  performed  as  follows  : 

By  means  of  the  lever  A  the  carriage  pushes  the  pointer  d  be- 
fore it,  which  travels  on  the  graduated  bar,  r.  The  pointer  has 
a  zero  mark  from  which,  after  the  breaking  o(  the  paper,  the 
breaking  strength  is  read  in  terms  of  kilograms. 

The  measurement  of  the  elasticity  is  done  by  reading  the 
movement  of  the  pointer  in  the  opposite  direction  along  the 
measuring  rod  c,  graduated  according  to  the  percentages  on  a 
strip  180  mm.  in  length.  After  the  breaking  of  the  paper,  the 
stretch  can  be  read  directly  in  per  cent. 

In  order  to  test  paper  with  this  apparatus,  one  adjusts  the 
force  measuring  rod,  by  raising  the  catches,  setting  the  spring 
in  oscillation,  allowing  it  to  come  to  rest  and  then  carefully 

■  348 



sliding  the  pointer  down  until  it  touches  the  lever.     Observe" 
whether  the  zero  of  the  pointer  agrees  with  that  of  the  measur- 
ing rod.     If  this  is  not  the  case,  the  latter  is  moved  until  both 
coincide.     The  spring  is  now  fastened  by  means  of  a  screw  /  a 
the  sled  is  moved  until  the  zero  marks  of  both  sled  and  stretch- 
measuring  rod  coincide.     Take  a  piece  of  the  paper  to  be  tested, 
previously  cut  to  standard  size,  clamp  it  in,  loosen  the  screw  /, 
drop  the  catches  and  begin  the  experiment,  giving  the  wheel  a 
slow  and  uniform  motion.     After  breaking  the  paper,  read  ofi 
the  loading  as  well  as  the  stretch,  relieve  the  spring  by  holding 
the  carriage  still  with  one  hand,  loosening  the  catches  with  the  J 
other  and  allowing  the  spring  slowly  to  sUde  back  into  place. 

In  order  to  insert  a  new  spring,  take  the  carriage  and  by! 
means  of  it  push  the  spring  in  the  direction  of  the  screw  /,  turn 
the  spring  through  90°  and  take  out  the  carriage  and  the  rack. 

In  conducting  the  experiments,  strips  180  mm.  long  and  fifteen 
mm.  broad  should  be  used,  and  not  less  than  five  cut  from  each  J 

In  order  to  render  the  result  independent  of  the  cross  sectionJ 
use  is  made  of  the  example  ol  Profs.    Reuleaux  and  Harti^.l 
Using  for  the  measure  of  strengh  of  paper  the  "  tearing  length, 
which  is  the  length  of  a  strip  of  paper  of  any  breadth  and  thick-j 
ness,  which,  if  hung  up  by  one  end,  would  break  in  conseqiiei 
ot  its  own  weight. 

Let  X  —  unknown  tearing  length. 

G—  wt.  of  the  torn  strip  (in  o.iSmm.  length),  in  grams. 

A'—  no.  of  kilos  necessary  to  tear  strip. 

„,       0.18      X  o.r8„, 

Then  — ^-3^-jir      or     jr=  -     -A. 

For  testing  materials  which  require  more  power  to  break  thaijB 
paper,  as  for  instance  cardboard,  Schopper  has  constructed  1 
more  powerful  apparatus,  which  has  a  maximum  force  of  150! 
kilos.     As  the   apparatus  is   buJll   on   the   same  fundamental^ 
principles  as  the  "  Wendler,"  a  description  here  is  needless. 

Referencei  :  "Haudbucli   der   Papier fabrikatiou."      By  S.  Mieniniiki,  1 

"A  Text-Book  on  Paper-Making,"     Cross  &  Bevan,  1888. 

"Tlie  Art  of  Paper-Making."     Alex  Watt,  1890, 



349  ' 

■•The  Chemisiry  o(  Paper-M siting."      By  R.  B.  Griffin  &  A.  D.  Little, 

"  Mittbeilungen  BUB  den  KoQiglicheo  techoischen  VersuchsanataUen 
lu  Berlin."  189J.  189J,  1893, 

Soap  Analysis. 

Soaps  may  be  conveniently  classified  into 

Toilet  soaps,  the  finest  grades  of  which  contain  no  impurities 
or  free  alkali ; 

Laundry  soaps,  in  which  rosin  and  generally  an  excess  of 
alkali  is  present  either  as  sodium  silicate,  sodium  carbonate, 
sodium  borate  or  free  alkali  ; 

Commera'al  soaps,  which    may  be   subdivided  into    (a)    soft   , 
soaps,  potash  being  the  base,  and  (i)   "  hydrated"  soaps,  soda 
being  the  base,   ("  marine  soap"  being  an  example)  formed  by 
caustic  soda  and  palmnut  oil  or  cocoanut  oil ;  and 

Medicated  soaps,  containing  medicinal  agents  such  as  carbolic 
acid,  tar,  sulphur,  etc.,  etc. 

The  complete  analysis  of  a  soap  often  presents  considerable 
difficulty — since  many  adulterants  may  be  used  in  the  cheaper 
grades,  and  many  substances  not  adulterants,  the  use  of  which 
is  permitted  as  colorants  and  for  perfume.  Allen  states  that  be- 
sides the  alkali  and  fatty  acids  and  water  requisite  for  the  forma- 
tion of  a  soap,  the  following  substances  have  been  found  in  the 
difierent  varieties — ochre,  ultramarine,  sodium  aluminate,  borax, 
resin,  vermilion,  arsenite  of  copper,  alcohol,  sugar,  vaseline, 
camphor,  gelatin,  petroleum,  naphthalene  and  creosote  oils, 
carbolic  acid,  tar,  glycerine  in  excess,  oatmeal,  bran,  starch, 
barium  sulphate,  sulphur,  steatite,  clay,  Fuller's  earth,  pumice- 
stone,  Icieselguhr,  chalk,  whiting,  etc. 

The  common  "  yellow  soap  is  formed  by  the  saponification  of 
tallow  or  palm  tree  oil  with  soda,  "  recovered  grease"  is  also 
used  in  the  cheaper  grades;  cotton-seed  oil.  olive  oil,  hemp-seed 
oil,  palm  oil,  cocoanut  oil,  castor  oil.  lard,  and  lard  oil,  are  all 
used  in  the  manufacture  of  soap. 

The  following  scheme  for  soap  analysis  is  by  C.  R.  Alder 
Wright  and  C.  Thompson.' 
















4>  (i 


•-  €8 



-  > 

SI . 











I  o 

S^:^  0  «  a  « 

v.S«  .bfia 
I  «  QTS  s  s  > 

^  OB  S  09  M  H-S 




.  « 
•S  « 


S  «  08 !!  *<z:— 
•2—  o  o-  «  2  . 

I  aS  o  o     - 

n  08  M 




I'd  u 


*•  C  08 

9  £.  o 

I       I     La      I 

•  3  0- 

fc-  a 

a"     • 

rt  O  3^11 

'•'7  crc 

5     tt  2 

*^  *"  ml  ** 


O  • 

—  B 

o  bo 

O  08 


Ti  0-.S 


08  a 

«  C  08  OiU'O*'  >  oC2  (8  «  O'S  OiV 

-  a- 

«*3  o 
V  a  > 
•O  3 

M  fl8>«« 

*.s  e  08 


SpS  CI 
OVj  e8 


•B  08 

w  U  4| 

3  u<i: 




lO^  3 

^ja  o 

C8  .    ♦* 

•o     n 

U  O  4; 
(A  U  i. 

•o  *5 

c  As 

^     —  3 
O  5"  K*-. 

i-i  w       O 
I  —  03  X 


6  4; 
-  3 


P  ^ 

a>  3  4, 

rt   08  ♦- 

«     be 
.£•0  a 


^*  •*  JO 
<  08  w 

•^  •*  jc 
I*  •*  o 

•-  a  o  c  **  3 

V  (13  3  a  ts  es'r: 
S  C  -  08  K  "  2 

.•g-  .bc«;*Sfi; 

I?'S   *-   «   **  O   /s 


3  3 



cros  be^  (u  be 

••=^3.:;*  3 

03  (0 


p.  5. 

3  cfl  «  i*  S         . 

0^^-.V  t  V  i» 

•otJ-i-J.  a> 


55'd  «' 
08  **»^  a 

.  S-5  08  3 


*<  0  2 

JC  (i  3 


•C       08 
o  «' 

X  a  «; 

SS-Sj-d  c  2 

•O  08  ^-^  • 
•2*2^  •' 

1  >.^^*- 

^  o— .^  g^ 

c  >  V  est: «  • 

•^  awo  S  08  0> 




;  u  <a  3  E2.o,c  aS.o'^iis 

3  g  -•a 

U  08 




/•  -'IT 

»  8»  .. 

o  o  t% 
•1  -"- 

(S  o 

rt  ^  *! 
X  S  2  c 

s  &'<  » 
•^  ff"C  g 

2  2  c 

rt  rt  '» 

ST  2.3  s 

f  3-0.0 
•«  .'•01 

9)  s  .«  . 

X  c 

Sis.*   a 

s  5  ^  et*i 

2  a»  ®  .-  S 

/»  «  s  5 

s  On  o.  I 

3  a.3*r-*  i 

a?  as* 































:l  X, 






















For  the  determination  of  water,  the  method  of  Lowe  is 

From  eight  to  ten  gramsof  the  soap,  (which  has  been  reduced  to 
very  fine  shavings,  and  represents  an  average  sample) ,  is  weighed 
out  between  watch-glasses  and  heated  in  the  air-bath,  at  first 
from  6o°-7o''  C,  to  avoid  melting,  then  at  ioo*'-i05®  C,  to  con- 
stant weight.  In  selecting  the  sample  in  this,  as  well  as  in  all 
subsequent  determinations,  it  is  essential  that  an  average  speci- 
men be  obtained,  since  the  content  of  water  in  the  different  parts 
of  the  bar  varies  considerably. 

This  is  best  effected  by  cutting  away  about  one-third  from  the 
end  and  evenly  scraping  the  cut  surface  of  the  remainder  until 
a  sufficient  amount  is  obtained  for  analysis. 

If  the  determination  of  free  alkali  is  of  considerable  importance, 
the  soap  should  be  dried  in  an  atmosphere  free  from  carbon  di- 
oxide. The  loss  at  105**  C.  represents  the  water  together  with 
other  volatile  constituents,  such  as  alcohol  and  essential  oils, 
which  may  be  present. 

Unsaponified  Matter. 

For  the  determination  of  unsaponified  matter,*  the  soap,  which 
has  been  dried  in  the  manner  indicated,  is  extracted  in  a  Soxh- 
let  extraction  apparatus  with  petroleum  ether,  which,  for  this 
purpose,  should  boil  below  80°  C,  and  should  leave  no  residue 
upon  evaporation.  After  the  extraction  is  complete,  the  petro- 
leum ether  is  distilled  off .  the  residue  dried  at  100®  C  and  weighed. 

In  a  boiled,  well-made  laundry  soap,  there  should  be  no 
unsaponified  matter  unless  the  same  has  been  added  subsequently. 

In  addition  to  unsaponified  fats,  foreign  matters  are  sometimes 
found  in  the  petroleum  ether  extract,  such  as  a  soft  paraffin  (so- 
called  "  Mineral  Soap  Stock"),  waxes,  hydrocarbon  oils,  phenol, 
etc.  If  waxes  are  found  to  be  present,  the  dried  soap  should 
be  extracted  with  boiling  toluene,  which  dissolves  the  same 
better  than  petroleum  ether. 

>  J.  F.  Schnaihlc.y.  Anal,  chrm.,  4,  147-156. 

■'*  Allen  :  Commercial  OrKanic  Analysis,  Vol.  2. 



Total  Alkali.     Fatty  Adds. 

The  dried  soap  thus  freed  from  uiisaponified   matter  is  next 
dissolved  in  hot  water,  preparatory  to   determining  the   total 
alkali  and  fatty  acids.  -   A  pure  soap  dissolves  completely  in  hot 
water,  and  no  ordinary  product  should  leave  more  than  a  slight 
residue.     If  the  article  examined  is  a  "  scouring  soap",  the  i 
soluble  residue  will  be  found  to  contain  quantities  of  fine  sand 
and  sometimes  taic.     The  residue,   if  appreciable,    should  be-' 
washed  by  decantation,  and  eventually  brought  upon  a  filter  with    i 
hot  water,  dried  at  ioo°  C,  and  weighed,  after  which,  if  desired, 
it  can  be  subjected  to  further  examination. 

To  the  aqueous  solution  is  added  an  excess  of  half  normal  sul- 
phuric acid  setting  free  the  fatty  acids  which  rise  to  the  surface. 
The  beaker  or  vessel  in  which  the  precipitation  was  effected  is 
next  cooled  with  ice-water.  When  the  fatty  acids,'  have  solidi- 
fied, it  is  best  to  decant  the  liquid,  remelt  with  hot  water  two  or 
three  times  to  remove  any  enclosed  mineral  acid,  again  cool,  fil- 
ter, and  wash  with  cold  water  until  the  washings  are  no  longer  I 
acid,  as  shown  by  litmus. 

The  filtrate  from  the  insoluble  fatty  acids  contains  the  total 
alkali  now  present  as  sulphate,  the  excess  of  sulphuric  acid  and 
any  glycerol  which  may  have  been  present  in  the  soap,  if 
saponification  was  effected  in  the  cold.  The  acid  liquid  may 
further  contain  a  small  amount  of  soluble  fatty  acids.  It  is  first 
titrated  with  half  normal  potassium  hydroxide  using  methyl 
orange  as  indicator.'  From  the  original  amount  of  sulphuric 
acid  added  and  the  number  of  cc.  half  normal  potassium  hy- 
droxide required  to  neutralize  the  excess  of  the  same,  the  total 
alkali  of  the  sample  can  be  determined. 

It  is  calculated  to  Na,0.  After  the  liquid  has  been  rendered 
neutral  to  methyl  orange  (which  indicates  the  mineral  acid), 
phenolphthalein  is  added  and  more  potassium  hydroxide  is  run  in. 
The  numberofcc.  of  potassium  hydroxide  required  for  neutralizing 
to  phenolphthalein  corresponds  to  soluble  fatty  acids  and  is  cal- 

(O  H   CO\ 
p'„"pqJO,  in  the  absence  of 


more  definite  knowledge  as  to  their  nature.  The  solution  is  now 
concentrated  and  tested  for  glycerol,  which  may  be  determined 
by  evaporating  to  dryness  and  extracting  with  ether-alcohol 
mixture' ,  or  else  by  oxidizing  to  oxalic  acid  by  means  of  per- 
manganate* (not  always  applicable).' 

In  soaps  containing  silicates  of  the  alkalies  (a  not  unusual 
constituent),  the  gelatinous  silicic  acid  which  separates  on  the 
.  addition  of  sulphuric  acid  remains  with  the  fatty  acids  on  filtra- 
tion. To  separate  the  fatty  acids  from  this  as  well  as  other 
impurities,  proceed  as  follows  : 

The  funnel  containing  the  filter  with  the  fatty  acids  is  placed 
in  a  small  beaker  and  heated  in  an  air  bath  (Allen's  method). 
As  the  filter  dries,  the  fatty  acids  pass  through  it  and  collect  in 
the  beaker  below,  while  all  impurities  (silicic  acid,  talc,  etc.)  re- 
main on  the  filter.  Of  course  it  is  necessary  to  wash  the  filter, 
which  remains  saturated  with  the  fatty  acids,  with  hot  redistilled 
alcohol  or  petroleum  ether,  or  else  exhaust  in  an  extraction 
apparatus.  The  alcohol  or  petroleum  ether  is  distilled  off  and 
the  residue  treated  in  the  same  way  as  the  main  quantity  of  fatty 

In  determining  the  fatty  acids  in  a  soap,  it  is  frequently  con- 
venient to  extract  with  ether  in  a  separatory  funnel.*  To  do 
this  the  soap  solution  is  placed  in  the  funnel  and  shaken  with 
sulphuric  acid  and  ether.  The  separated  acids  are  at  once  dis- 
solved in  the  ether.  The  aqueous  solution  may  be  drawn  off 
below,  the  ethereal  solution  washed  with  water,  the  ether  evapo- 
rated, and  the  residue  dried  at  ioo°  C,  and  weighed. 

Since  the  fatty  acids  exist  in  the  soap  as  anhydrides  and  are 
weighed  as  hydrates,  it  is  necessary  to  multiply  the  weight  found 
by  the  factor  0.97,  which  gives  the  weight  of  fatty  anhydrides. 
The  fatty  acids,  after  having  been  weighed,  may  be  titrated  with 
half  normal  potassium  hydroxide,  and  from  these  data  may  be 
ascertained  what  portion  of  the  total  alkali  exists  in  combination 
with  the  acid  as  soap. 

1  Chrm.  Z/a'..  8.  1667. 

2  Chrm.  Ztfr,,  g,  975. 

«  Allen  :  Com.  Ortf.  Anal.,  a,  7qo. 
♦  Chem.  Xeu's,  43,  21S. 


Free  Alkali. 


To  determine  the  per  cent,  of  free  alkali'  in  soap,  a  st 
portion  is  weighed  out  and  extracted  with  neutral  alcohol  in  an 
extraction  apparatus.  The  caustic  alkali  is  determined  in  the 
alcoholic  solution  by  titrating  with  half  normal  hydrochloric 
acid,  using  phenolphthalein  as  indicator.  If,  however,  the  soap 
contains  unsaponified  fat,  as  is  frequently  the  case  if  made  by 
the  so-called  "  cold-process,"  this  method  cannot  be  used,  since 
in  alcoholic  solution  unsaponified  fat  would  be  readily  saponi- 
fied by  the  free  caustic  alkali  present.  In  such  a  case  the  soap 
must  first  be  dried  in  an  atmosphere  free  from  carbon  dioxide 
at  lOo"  C.  the  unsaponified  matter  extracted  with  petroleum 
ether,  and  finally  the  soap  dissolved  in  alcohol  and  the  free  alkali 
determined  in  the  alcoholic  solution  as  before.  The  sodium  car- 
bonate, sodium  silicate,  borax,  and  everj'thing  insoluble  in  alco* 
liol,  remains  behind  in  the  extraction  tube  and  may  be  dried  at 
loo'  C.  and  weighed.  If  considerable,  it  may  be  further  treated, 
as  follows : 

First,  it  should  be  exhausted  with  boiling  water  ;  one-half  of 
this  solution  is  then  titrated  with  half  normal  hydrochloric  acid 
using  methyl  orange  as  indicator.  The  amount  of  acid  required 
corresponds  to  carbonate,  silicate  and  borate.  In  this  solution 
sulphates  may  also  be  determined  and  starch  and  gelatine  tested 
for.  The  other  half  of  the  solution  is  examined  qualitatively 
for  carbonate,  silicate  and  borate.  If  there  remainsaconsiderable 
residue  insoluble  in  water,  it  may  be  dried  at  ioo°  C,  weighed 
and  further  examined. 

Reiiii . 

Resin  is  a  very  common  constituent  of  soaps,  the  resinates  of 
the  alkalies  having  a  similar  action  to  soaps,  and  the  cheapness 
of  the  material  often  suggesting  a  partial  substitution  of  it  for 
the  natural  fats  and  oils. 

As  a  qualitative  test  for  resin,  Gottlieb's' method  is  reliable 
and  easily  made. 

The  soap  is  dissolved  in  water  and  heated  to  boiling.  A  strong 
solution  of  magnesium  sulphate  is  added  until  the  fatty  acids  are 


completely  precipitated.     The  magnesium  resiiiates  remain 
solution.     After  boiling  two  or  three  minutes,  the  solution 
filtered  and  the  hot  filtrate  acidified  with  dilute  sulphuric  acid. 
In  the  presence  of  resin   the  liquid  becomes  turbid,  due  to  the 
separated  resin  acids.     The  boiling  should  be  continued  forone- 
half  hour,  to  make  sure  that  the  turbidity  is  due  to  resin  aei« 
and  not  to  volatile  fatty  acids.     One  method  for  the  quantitative^ 
determination  of  resin  in  soap  is  that  of  Hiibl,'  as  follows 

One-half  to  one  gram  of  the  mixture  of  fatty  and  resin  acids 
is  heated  in  a  closed  flask  on  the  water-bath  with  about  twenty 
cc.  of  alcohol  to  complete  solution.  The  acids  are  neutralized 
with  alkali,  using  phenolphthalein  as  indicator.  The  alcoholic 
soap  solution  is  then  poured  into  a  beaker,  the  flask  rinsed  with 
water,  the  solution  diluted  to  300  cc.  and  silver  nitrate  added 
to  complete  precipitation.  The  precipitate  (consisting  of  the 
silver  salts  of  the  resin  and  fatty  acids)  must  be  protected 
from  sunlight.  It  is  filtered,  washed  with  water,  dried  at  100° 
C,  and  extracted  in  a  Soxhlet  tube  with  ether.  The  silveri 
resinates  dissolve  in  the  ether,  while  the  silver  salts  of  the  (altyj 
acids  remain  behind.  The  ethereal  solution,  as  it  leaves  the 
extraction  tube,  should  be  yellow  or  light  brown  in  color,  but 
not  dark  brown.  It  ia  filtered,  if  necessary,  and  the  filtrate 
shaken  with  hydrochloric  acid  in  a  separatory  funnel.  The  re- 
sulting ethereal  solution  of  theresinacidsisfiltered  from  the  sib 
chloride,  washed  with  water,  and  the  filter  and  separator  rin; 
with  ether,  the  ether  distilled  off,  and  the  residue  dried  at  100° 
As  the  resin  is  weighed  in  the  hydrated  form,  its  weight  must 
be  multiplied  by  the  factor  0.9732  to  obtain  the  weight  of  the 

Twitchell's method  for  the  determination  of  resin  in  a  mixtw 
with  fatty  acids  depends  upon  the  formation  (in  alcohol  solu 
of  the  ethereal  salts  of  the  latter  when  treated  with  hydrochloi 
acid,  resin  being  unacted  upon.  The  gravimetric  method  is  as 
lows  :'  Two  or  three  grams  of  the  mixture  of  fatty  acids 
re.sin  are  dissolved  intentimestheirvolumeof  absolute  alcohol  ai 
dry  hydrogen  chloride  is  passed  through  in  a  moderate 



ity  ■ 
:ed    ■ 

>y.  Ana:.  Afpl.  Oum.,  5,  j' 

: :  School  nf  Mines  QuarUrty.  i; 



the  flask  being  placed  in  a  vessel  with  water  to  keep  it  cool. 
The  gas  is  rapidly  absorbed,  and  after  about  forty-five  minutes 
the  ethereal  salts  separate  and  float  on  the  solution.  After 
waiting  for  half  an  hour  longer,  the  liquid  is  diluted  with  five 
times  its  bulk  of  water  and  boiled  until  the  acid  solution  is  clear, 
the  ethereal  salts,  with  resin  in  solution,  floating  on  top.  To 
this  is  added  some  light  petroleum,  and  the  whole  transferred  to 
a  separatory  funnel,  the  flask  being  washed  out  with  light  petro- 
leum. The  acid  liquid  is  then  run  ofl,  and  the  petroleum  ether  so- 
lution washed  once  more  with  water  and  then  treated  in  the  funnel  i 
with  a  solution  of  a  half  gram  of  potassium  hydroxide  and  live  cc. 
of  alcohol  in  fifty  cc.  of  water.  The  resin  is  immediately  saponi- 
fied, and  the  two  layers  separate  completely.  The  resin  soap 
solution  can  then  be  run  off,  and  the  resin  recovered,  as  usual 
by  the  addition  of  an  acid.  The  first  stages  of  the  volumetric 
method  are  similar  to  those  of  the  gravimetric,  with  the  excep- 
tion that  the  contents  of  the  flask  are  washed  into  the  separating 
funnel  witli  ether  instead  of  light  petroleum,  and  the  ethereal 
solution  is  then  thoroughly  washed  with  water  until  all  soluble 
acidity  is  removed  ;  fifty  cc.  of  neutral  alcohol  is  then  added,  and 
the  solution  titrated  with  standard  solution  of  sodium  hydroxide. 

Itis  frequently  of  interest 
to  know  the  origin  of  the 
fatty  acids  of  a  soap  which 
is,  however,  in  many  cases. 
a  problem  not  easily  solved . 
The  only  clues  are  to  be 
ssought  in  the  specific  grav- 
ity, combining  weight, 
melting  and  solidifying 
points,  and  iodine  number 
of  the  fatty  acids. 

The  values  for  the  specific 
gravitiesincolumn  III  page 
35S,  were  obtained  with  a 
Westphal  and  a  Reimann'.s 
balance  plummet,  with  a 
thermometer  of  a  range  95' 
ing  figure. 

s  shown  in  accompany- 



Ijhn  ujntei?  un 


|fi2j:  3S?p"hs  s;;s 




ffiila  ssliHasi  Sil 


IMII  llflilKH  llff 




|i*»ii  lii^llllli  l|ii 


SS88H  «5»|»SI|P  SS-S! 




^lISjs  :3JMi|i5|  IJ3S 


^i\m  \\\m\\i 'AW 


>^lll?ll  PHS1IEJ3  SSI* 







......    ........V.    ..J. 





III!!  ffi|l!!l!l  lit! 


l!!!!!  fflllllBI  fii! 



Iflfff  Ittlfftfil  IIH 



Ifilll  flllHIIII  till 




Ilflll  liftlHt:!  flli 

-        H 

•5  S 

iSliUis  iilllKssI  4|il 




Occasionally,  fats,  before  being  used  in  soap-making,  are 
bleached  by  various  cbemical  agents,  the  most  common  of  which 
are,  perhaps,  potassium  dichromate  and  hydrochloric  acid,  or 
sulphuric  acid.  H  now  such  a  mixture  is  heated  in  bleaching, 
as  is  frequently  the  case,  the  potassium  dichromate  acting  on 
the  hydrochloric  acid  liberates  chlorine,  and  under  favorable 
conditions,  the  chlorine  combines  with  the  unsaturated  acids 
present  in  the  fats  as  glycerides,  thus  utterly  destroying  the 
value  of  the  iodine  number,  the  most  definite  index  as  to  the 
origin  of  the  fats.  Again,  it  frequently  occurs  that  a  mixture  of 
two  or  more  fats  may  be  used,  the  combining  weights,  iodine 
number,  and  other  properties  of  which  closely  approximate  those 
of  an  individual  fat,  and  so  an  erroneous  conclusion  may  be  drawn 
from  an  examination  of  such  mixed  fatty  acids.  If,  however,  a 
mixture  of  two  fats,  in  their  natural  state,  without  having  under- 
gone any  bleaching  or  refining  process,  is  used,  it  is  generally 
possible  to  ascertain,  with  considerable  accuracy,  the  nature  of 
the  fatty  acids  by  means  of  the  iodine  number,  it  having  been 
found  by  actual  experiment  that  the  iodine  number  o(  a  mixture 
of  two  fats  corresponds  within  limits  of  analytical  error  with  the 
theoretical  numbers  calculated  for  the  pure  fats. 

Glycerine  in  fats  and  soaps  can  be  determined  as  follows;' 
three  grams  are  saponified  with  an  alcoholic  potash  solution,  the 
soap  solution  diluted  to  200  cc,  decomposed  with  dilute  acid, 
filtered  from  insoluble  fatty  acids,  and  the  filtrate  and  washings, 
which  should  amount  to  above  500  cc,  evaporated  rapidly  down 
to  250  cc.  sulphuric  acid  added  and  titrated  with  standard 
potassium  bichromate. 

For  the  titration  by  bichromate  the  following  solutions  are  re- 
quired : 

1 .  Bichromate  solution  containing  about  74.86  grams  of  potas- 
sium bichromate  and  150  cc.  strong  sulphuric  acid  per  liter. 
The  oxidizing  value  of  the  solution  must  be  ascertained  by  titra- 
tion with  solutions  containing  known  amounts  of  iron  wire. 

2.  Ferrous  ammonium  sulphate  solution  containing  about  240 
grams  per  liter. 

3.  A  bichromate  solution  one-tenth   as  strong   as  the   first. 

I  O.  HebncT  :  J.  Soc  Chim   /nd.,S,i. 



The  ferrous  solution  is  standardized  upon  the  chromate  solution, 
and  the  glycerol  value  of  the  chromate  (contents  of  bichromate 
divided  by  7.486)  is  calculated.  One  and  five-tenths  of  the  glyc- 
erol or  soap  lye  is  weighed  into  a  100  cc.  flask,  and  a  little  silver 
oxide  added  to  remove  any  chlorine  or  aldehydic  compounds. 
After  slight  dilution,  the  sample  is  allowed  to  stand  with  the  sil- 
ver oxide  for  about  ten  minutes.  Basic  lead  acetate  is  then  added 
in  slight  excess,  the  bulk  of  the  fluid  made  up  to  100  cc.  and  a 
portion  is  filtered  through  a  dry  filter. 

Twenty-five  cc.  of  the  filtrate  are  placed  in  a  clean  beaker, 
then  forty  to  fifty  cc.  of  the  standard  bichromate  solution  ac- 
curately measured,  are  added,  and  fifteen  cc.  strong  sulphuric 
acid.  The  beaker  is  covered  with  a  watch  glass  and  heated  for 
two  hours  in  boiling  water.  The  excess  of  bichromate  solution 
is  then  titrated  back  with  the  ferrous  ammonium  sulphate  solu- 

The  table  of  analyses  of  soaps  on  the  following  page  comprises 
in  each  instance  a  complete  analysis. 

In  most  analyses  of  soaps  the  following  determinations  only 
are  made  :  Water,  alkali  combined  as  soap  (Na,0),  alkali  free 
as  sodium  hydroxide,  sodium  carbonate,  and  total  fatty  acids  as 
anhydrides.  Thus,  an  ordinary  yellow  laundr>'  soap,  analyzed 
by  Schnaible,  gave  : 

Water 19.12  per  cent. 

(  Alkali,  combined  \  r  c-y     *«       •• 

\      as  soap,  Na,0      >    ^^' 

Alkali  free,  as  NaOH 0.20  •• 

"     NXCOs 0.20  *•       *• 

Insoluble  in  H,0    0.20  " 

Fatty  anhydrides 52.32  **       ** 

Resin 19-45  *''      ** 

Total  100.00     *•       ** 

IVas/tinj^  Powders, 

The  washing  or  soap  powders  contain  besides  powdered  soap, 
a  large  percentage  of  sodium  carbonate,  usually  in  the  form  of 
dried  soda  crystals.  These  powders  are  generally  prepared  as 
follows  :  Anhydrous  sodium  carl>onate  or  anhydrous  soda  ash  is 
added  to  a  "clear  boiled"   soap  paste,    and  after  thoroughly 


li      i 
■'I      I 

!  ■- 













j|  1  1 II 1  If 

anhydride""  " 

i?  5    i  sj  4    ?  1 

■%i   iU  i     il 


.VP         gPPP           gg 

:*  «     last,     Is 

Soda  Bi.  sillcxe. 

?<;■      P         P      P  P       ?           P      P 

38   s     5   sa    a      8.  5 

anil  hydrate. 

Sodium  chloride. 

pu.-             poop                pp 

.pp     pppp      pp 

Sodium  sulphate. 

pp    p      g    °°     ° 


3    £ 

oilde.°°    '™ 


8S    S       8    «S     «        8    8 


1     r        ?    ' 
i    S       S, 


L  %     ^  s 

Fatty  and  resin     ; 



mixing,  the  somewhat  stiff  material  is  drawn  o£f  into  cooling^ 
frames. '  The  cold  and  hard  soap  thus  formed  is  then  finely 

The  composition  varies  greatly.  Only  a  small  proportion  of 
resin  soap  can  be  used,  as  such  a  soap  is  sticky  and  cannot  be 

Olein  soap  is  generally  used  and  is  saponified  with  sodium 

References. — "Die  Darstellungder  Seifen,  ParfumerienundCosmetica.** 
By  C.  Deite,  1867. 

"  The  Art  of  Soap-Making."     By  A.  Watt,  1887. 

'•  The  Manufacture  of  Soap  and  Candles."     By  W.  T.  Brandt,  1888. 

•'  Lard  and  Lard  Adulteration."     By  H.  W.  Wiley,  1889. 

"  Die  Untersuchungen  der  Fette,  Oele  and  Wachsarten."  By  C.  Schaed- 
ler,  1890. 

*•  Analysis  of  Washing  Powders."     Am*  Chem.J.,  i4»  623. 

"  Soap  Powders."    Seifen^  Oel  und  Fett  Industrie y  3»  973. 

"  Oils,  Fats,  Waxes  and  their  Manufactured  Products."  By  Alder 
WriKht,  F.R.S.,  1894. 

'•  Oils,  Fats,  and  Waxes."  By  Dr.  R.  Benedikt  and  Dr.  J.  Lewkowitsch, 
F.C.S.,  1895. 


Technical  Examination  of  Petroleum. 

Tnis  is  usually  performed  by  fractional  distillation  of  the  petro- 
leum into  three  classes  of  distillates. 

1.  Light  oils,  distilling  over  up  to  150°  C. 

2.  Illuminating  oils  distilling  over  from  150**  C.  to  300^  C. 

3.  Residuum. 

The  methoil  of  Kngler,  which  is  largely  used  for  this  purpose, 
requires  a  ghiss  flask  of  the  form  shown  in  Fig.  103.  The 
measurements  j;iven  in  the  figure  are  stated  in  centimeters.  The 
flask  is  ciMuiecteti  with  a  condenser  in  the  usual  manner. 

uv^  00.  ot  the  oil  are  taken  and  the  temperature  in  the  flask 
si^  regulatevi  that  two  and  one-half  cc.  of  the  distillate  pass  over 
every  minute.  Chemists  vary  the  method  of  distillation,  some 
usinj;    ux^  ov\    v^t   the   oil   and  a   larger   flask   of  same    form. 




though  without  standard  rules 
respecting  the  number  of  dis- 
tillates to  be  obtained  :  thus 
A.  Bourgougnon  and  J.  Mon-  ' 
del'  report  the  aualysis  of  a 
sample  of  Ohio  petroleum  in 
which  the  distillation  was  in 
filty  parts,  each  part  repre- 
senting two  per  cent,  by  vol-_^ 
ume,  the  distillation  commen-' 
cingat  23°  C.  The  composi- 
tion of  the  oil  being  given  as 
sixteen  per  cent,  of  naphtha, 
70'  B.,  sixty-eight  per  cent, 
of  kerosene,  six  per  cent.  of.t~ 
paraffin  oil  and  ten  per  cent, 
of  residuum.  Durand  Wood- 
man' gives  an  analysis  of  a 
crude  petroleum  from  Ohio. 
300  cc.  of  the  oiTwere  taken 
andeighteendistillateseachof  "*        "'^   Pig.  1^, 

fifteen  cc.  (five  per  cent,  of  total)  were  obtained.     The  results  in 
detail  were  as  follows  : 



The  result  being 

Naphtha  10  per  cent. 

Illuminating  oil 50    "       ** 

Lubricating  oil 3a  *'       ** 

Residuum 10    **       •* 

Total 100   *•       ** 

A  distillation  of  a  Mexican  petroleum,  by  the  writer,  made  by 

the  Engler  method,  gave 

Naphtha lo.o  per  cent. 

Illuminating  oil 60.0 

Lubricating  oil 15.5 

Tar  and  Residuum 14.5  ** 

((        ft 

<(  4( 


Total 100.0  *•       *• 

Another  sample  of  the  same  oil,  submitted  to  a  somewhat 

higher  temperature  during  the  distillation,  using  a  similar  flask 

excepting  that  the  deliver>'  tube  was  one  and  one-half  inches 

higher  in  the  neck  of  the  flask  (requiring  higher  heat  upon  the 

petroleum  for  the  same  distillates  as  in  the  former  case),  gave  a 

lower  percentage  of  heavy  oils,  and   a  higher  percentage    in 

illuminating  oils,  the  result  being 

Naphtha  i  i.o  per  cent. 

Illuminating  oil 64.0   *•      '* 

Lubricating  oil 10.3 

Residuum 14.7 

Total    100.00  "      ** 

By  a  careful  regulation  of  the  heat,  the  amount  of  illuminating 
oil  can  be  increased  or  decreased  to  a  certain  percentage  as 

The  three  general  divisions  of  the  distillation  of  petroleum  are 
still  further  technically  divided  as  follows  : 

I .  Naphtha  group,  comprises  : 

Cymogene,  a  gas,  boiling  point  o^  C,  specific  gravity  116°  B. 
Rhigolene,  liquid,  boiling  point  18.3°  C,  specific  gravity  loo*^  B. 
Petroleum  ether.boiling  point  40"^  to  70^  C,  specific  gravity  85°  to  80^  B. 
Gasolene,  boiling  point  70"  to  90^  C,  specific  gravity,  80*^  to  75^  B. 
Naphtha  (Danforth  oil)  boiling  point  80^  to  110^  C,  specific  gravity 

76^   to  70^  B. 
Ligroine,  boiling  point  80*^  to  120''  C,  specific  gravity  67°  to  62*^  B. 
Benzene,  boiling  point  120'^  to  150-  C,  specific  gravity  62°  to  57°  B. 



2.  lUummating  oils.  The  various  varieties  of  kerosene,  boil- 
ing points  150°  to  300°  C. 

3.  Residuum,  (tar,  etc.)  boiling  point  300°  C,  and  above, 
from  which  is  obtained  :  Lubricating  oils,  paraffin  oils,  and  coke 
remaining  as  a  solid  body  in  the  retort. 

The  average  percentage  of  the  products  obtained  from 
Pennsylvania  petroleum  can  be  stated  as  : 

First  group  :  Naphthas.  16.5  per  cent. 

Second  group  :  Illuminating  oils,  fifty-four  per  cent. 

Third  group:  Lubricating  oils,  seventeen  per  cent.,  paraffin, 
two  per  cent.,  coke,  ten  per  cent." 

The  manufacture  of  vaseline,  petrolatum,  cosraoline, etc.,  from 
the  tarry  residuum  {vacuum  process,)  has  increased  largely  in 
the  last  few  years.' 

In  the  oil  trade  the  principal  mineral  oils  obtained  from  petro- 
leum are  as  follows : 

Benzenes  and  naphthas,  62",  65°,  75°,  88°,  90'  Baum&. 

Paraffin  gas  oil.      Paraffin  oils,  32°,   34°,  25°.  28°,  30°,  32°  B. 

Red  oil,  23°  and  24°  B.     Neutral  filtered,  32°,  34",  37°  B. 

"Extra  cold  test"  32°  B.     '"Wool  stock"  32°.    Blackreduced 
(25°  to  30°  F.  cold  test)    ( 15°  F.  cold  test),  28"  B.  zero   test. 
Black  reduced,  "Summer."     Cylinder,  light  filtered,  600°  F.  fire 
test.     Smith's  ferry  32°  to  34°  B. 

Dark  steam  refined.  West  Virginia,  natural  29°  ;  Franklin 
natural  29"  B. 

Kerosaie,  the  different  grades  and  colors. 

The  various  valve  oils,  car  oils,  engine  oils,  spindle  oils,  loom 
oils,  dynamo  oils,  etc.,  etc.,  are  usually  compounded  oils,  min- 
eral oil  of  some  variety  being  the  principal  constituent  v  1  ch 
varying  amounts  of  lard  oil.  tallow  oil,  lallow,  rape  oil,  etc., 
have  been  added. 

The  best  engine  oil  is  a  mixture  of  lard  oil  and  paraffin  oil  in 
equal  parts.  This  compound  has  been  in  use  by  the  Pennsyl- 
vania Railroad  for  the  past  ten  years,  and  after  many  experi- 
ments and  trials  of  different  substitutes,  still  remains  the  stand- 
ard.    Passenger  car  oil  is  usually  a  mixture  of  well  oil  and  lard 


oil  in  the  proportion  of  two-thirds  well  oil  and  one-third  lard 
oil.  Lard  oil  in  the  proportion  of  one  part  to  three  of  5cx>**  well 
oil  has  been  found  to  give  the  best  results  as  a  cylinder  lubri- 



The  Examination  of  Lubricating  Oils. 

The  generally  accepted  conditions  of  a  good  lubricant  are  as 
follows : 

I  St.  Body  enough  to  prevent  the  surfaces,  to  which  it  is  ap- 
plied, from  coming  in  contact  with  each  other. 

2d.  Freedom  from  corrosive  acids,  either  of  mineral,  animal  or 
vegetable  origin. 

3d.  As  fluid  as  possible,  consistent  with  **body.*' 

4th.  A  minimum  coefficient  of  friction. 

5th.  High  '* flash*'  and  **  burning  *' points. 

6th.  Freedom  from  all  materials  liable  to  produce  oxidation 
or  "gumming.** 

The  examinations  to  be  made  to  verify  the  above  are  both 
chemical  and  mechanical,  and  are  usually  arranged  in  the  fol- 
lowing order  : 

I  St.  Identification  of  the  oil,  whether  a  simple  mineral  oil, 
animal  oil,  vegetable  oil,  or  a  mixture. 

2d.  Specific  gravity. 

3d.  Cold  test. 

4th.  Viscosity. 

5th.  Iodine  absorption. 

6th.  Flash  and  fire  tests. 

7th.   Acidity. 

8th.   Maumene's  test. 

oth.  Coefficient  of  friction. 

It  the  oil  is  a  pure  mineral  oil,  the  tests  numbered  i,  5  and  8 
arc  omitted. 

The  first  tost,  the  nature  of  the  oil,  etc.,  is  performed  as  fol- 
lows : 

^  .'i.-  .V.;;>a:.:'.:-.;  Fnx'i^'f'Tfj:  .*.■•■' ■•■J.'.  64.  -5-: 26.    For  formulas  of  locomotive  and 
car  lulMioum>  .»^  u>c*!  on  ihr  Tdili\\ui>iii  Cterinany  consult:  "Die  Sctamiermittel." — Von 

lose!  lirOS>UI;lllll,  •.Si4. 



Ten  grams  of  the  oil  are  weighed  out  in  a  dry  tared  beaker 
(250  cc),  and  to  it  is  added  seventy-five  cc.  of  an  alcoholic  solu- 
tion of  potash  (sixty  grams  of  potassium  hydroxide  to  1,000  cc. 
of  ninety-five  per  cent,  alcohol),  and  the  contents  evaporated 
until  all  the  alcohol  is  driven  off.  In  this  process,  if  any  animal 
or  vegetable  oil  is  present,  it  is  formed  into  a  soap  by  the  potash, 
while  the  mineral  oil  is  unacted  upon.  Water  (seventy-five  cc. ) 
is  now  added  and  the  material  well  stirred  to  insure  complete 
solution  of  the  soap,  and  then  it  is  transferred  to  a  separatory 
funnel  (Fig.  104),  seventy-five  cc.  of 
sulphuric  ether  added,  corked,  the 
liquid  violently  agitated  and  allowed 
to  stand  for  twelve  hours.  Two  dis- 
tinct liquids  are  now  seen,  the  lower, 

e  solution  of  the   soap,   the  upper, 
the  ether  solution  (colored,  if  mineral 
oil  is  present,  colorless,  if  not).     The  | 
aqueous  solution  is  drawn  off   ir 

,  3  beaker,   the  ethereal  solution  ^ 
remaining    in    the    separatory    fun 

The    former    is    placed  on   a  \ 
water-bath,  heated  for  lial!  an  hour 
and  until  all  traces  of  ether  (which  is 
absorbed  by  the  water  in  a  very  small 
amount)  is  gone. 

The  solution  is  allowed  to  cool, 
diluted  somewhat  with  water,  and 
made  acid  with  dilute  sulphuric  acid. 
Any  animal  or  vegetable  oil  present 
will  be  indicated  by  a  rise  to  the 
;urface  of  the  liquid  of  the  fatty  acids. 
{In  this  reaction  the  sulphuric  acid 
decomposes  the  soap,  uniting  with 
the  potash  to  form  sulphate  of  potash 
and  liberating  the  fatty  acids  of  the  oil. ) 

If  it  be  desired  to  weigh  the  fatty  acids,  proceed  as  follows  : 

Weigh    carefully  about   five   grams  of  pure  white  beeswax, 
place  it  in  the  beaker  upon  the  surface  of  the  oil  and  water, 


and  bring  the  contents  nearly  to  boiling  ;  the  melted  wax  and  fatty- 
acids  unite ;  allow  to  cool,  remove  the  wax,  wash  with  water, 
dry  between  folds  of  filter  paper  and  weigh.  The  increase  in 
weight  of  the  wax  over  its  original  weight  gives  the  weight  of 
the  fatty  acids  of  the  animal  or  vegetable  oil  in  the  lubricating 

Another  method  of  determining  the  weight  of  the  fatty  acids 
after  saponification  is  given  on  page  354. 

The  weight  obtained  must  be  multiplied  by  the  factor  0.97, 
since  the  fatty  acids  exist  in  the  oil  as  anhydrides  and  not  as 
hydrates,  the  latter  being  the  form  in  which  they  are  weighed, 

Instead  of  weighing  the  animal  or  vegetable  oil,  some  chemists 
prefer  to  make  use  of  the  ether  solution,  determining  the  hydro- 
carbon oil  directly.     In  which  case  proceed  as  follows  : 

After  drawing  off  the  soap  solution  from  the  separatory  funnel 
the  ether  solution  is  run  into  a  weighed  flask  (about  250  cc), 
and  the  ether  distilled  ofl.  The  residue  in  the  flask  now  consists 
of  the  mineral  oil  and  some  water. 

It  is  quite  diflBcult  to  get  rid  of  all  this  water.  Direct  heat- 
ing is  inadmissible,  since  the  water  spurts  up  through  the  oil 
out  of  the  flask  and  is  lost.  This  can  be  overcome  by  placing  a 
glass  tube  through  the  stopper,  in  shape  of  the  letter  S.  Any 
oil  ejected  against  the  tube  or  cork  cannot  escape,  but  returns  to 
base  of  flask,  while  the  heat  is  gradually  increased  in  the  flask 
and  the  water  vaporized  and  passed  out  through  the  tube  ;  three 
or  four  weighings  are  generally  required  before  a  constant 
weight  is  obtained.  The  former  process  is  preferable,  since  it  is 
performed  much  more  rapidly  than  the  latter,  and  also  the  ani- 
mal or  vegetable  oil  is  positively  shown,  and  generally  can  be 
identified  ;  also  many  lubricating  oils  contain  as  high  as  twenty 
per  cent,  of  hydrocarbon  oil,  volatile  at  or  below  212**  Fahren- 
heit. It  is,  of  course,  in  the  ether  solution,  and  when  the 
water  is  expelled  from  the  oil,  after  the  ether  has  been  driven 
off,  a  large  proportion  of  the  volatile  hydrocarbon  is  also  vapor- 
ized. If  now  the  animal  or  vegetable  oil  is  not  also  determined, 
a  serious  mistake  would  be  made;  viz,,  reporting  twenty  per 
cent,  of  animal  oil  when  it  was  volatile  mineral  oil. 

The  fatty  acids  in  another  sample  of  the  oil  are  separated  and 



subjected  to  qualitative  tests  for  identification  of  the  oil  from 
which  they  are  derived.  These  tests  comprise  determination  of 
melting  point,  and  congealing  point,  page  337.  color 
reaction  with  nitric  and  sulphuric  acid,  iodine  ab- 
sorption, and  Maumenfe's  test,  rise  of  temperature 
upon  addition  of  sulphuric  acid. 

There  are  several  methods  of  determining  the  melt- 
ing point  ol  the  fatty  acids.  Where  a  considerable 
amount  of  the  fatty  acids  is  available  for  experiment, 
the  apparatus  in  Fig.  105  can  be  used.  The  glass 
cylinder  is  filled  one-half  with  fatty  acids,  the  cylinder 
closed  with  a  rubber  stopper,  through  which  a  ther- 
mometer is  inserted,  the 
bulb  of  which  is  covered 
by  the  fatty  acids. 

The  apparatus  is  sup- 
ported in  a  beaker  con- 
taining water.  (Fig.  106). 
If  the  fatty  acids  are 
liquid  at  ordinary  temper- 
atures, the  water  in  the 
beaker  must  be  cooled 
with  ice  until  the  fatty  acids 
arecongealed.  Theiceisre- 
moved,  and  the  water  grad- 
ually warmed  until  the  f^itt',- 
acids  become  melted  Al 
this  point  the  temptr.H'iu- 
is  taken  and  reim. 
Greater  delicacy  in  i'n 
termination  of  the  iiil!  .  ,^ 
Fig.  .OS,      point  is  obtained  by  vising  '■«    •■■'■ 

a  small  glass  tube,  sealed  at  one  end.  The  liquid  fatty  acids  a 
placed  in  this  tube,  then  congealed,  the  tube  then  tied  to  t 
thermometer(  Fig.  107)  and  both  inserted  in  a  beaker  of  water, 
as  shown  in  Fig.  108.  Another  method  is  to  cover  the  ther- 
mometer bulb  with  a  layer  of  the  solid  fatty  acids,  about  three 
mm.  thick  and  inunersing  it  in  water;  gradually  heat  the  water 


and  notice  the  temperature  at  which  the  fatty  acids  leave  tha 
thermometer  bulb  and  ascend  through  the  water. 

Fatt\-  Acids.  , 

Cotton-seed  oil 33. 

Olive  ■'    a6 

Rape-seed    "     10 

SeHmJ  "     16 

Cocoannt     "    >4 

L»rd 44 

Tallow  45 

Wool-fftt 4i.o 

Palm  oil , 48.0 

Spedfic  Gravity. 

Ill  the  chemical  laboratorj-  the  hydrometers  used  are  generallyl 
.  mailed  with  the  specific  gravity  direct.     In  the  oil  trade,  how«J 



ever,  and  in  general  commercial  practice  the  Baum& 
hydrometer  is  used,  and  the  following  precaution  is 

If  the  oil  is  not  liquid  enough  to  flow  easily,  it 
must  be  warmed  until  so.  and  then  tested  with  the 
hydrometer.  The  latter  should  move  easily  and  freely 
in  the  liquid .  As  all  specific  gravities  are  comparable 
at  60°  F.,  it  will  he  necessary  to  make  correction  for 
temperature :  if  the  temperature  of  the  oil  is  above  60' 
F..the  reading  of  the  hydrometer  is  too  large  ;  if  below 
60'  F.,the  readings  are  too  small.  Suppose  an  oil  reg- 
isters 28°  Baume  at  72"  F.,  we  make  use  of  the  table. 
on  p.  372,  and  find  the  corrected  reading  to  be  27.  j° 

To  convert  this  into  specific  gravity  the  following 
table  is  used : 

^a  ^^* 








"^^  10 

1. 0000 







^;?^      ji 










m        " 










I           '3 










1          '4 










1          '5 










>          16 










J          17 










1     \          iS 










1     1          19 










11   >          'O 










H  -    '' 










||            ia 










^^^y     and 

we  find  thai 

27,2°  Baumfe 

are  equal   to 


Wj      spec 

fie  gravity. 

fly            Figure  109 

represents  a 

Tagliabue  hydrometer  for 

JH        oils 

it  contains  a 


also  a 

scale  to  make 

^•I         the  readings 

at  60" 

F,     Subtract 


me  for  every 

■S'         '0°  F. 


60°  K.,  and  add  1° 


iumh  for  ev 

ery  10' 

F.  below  60°  F. 


V^^^^OcOfOfOfOfltf     W     W    M    MM     •-«    O    O     9^Ol0000t^t^^O     ir> 

•     ^M   m  \o  t>>  o^  ^    t<i  ifi  t^  Q\  !•<    rolor>lO^^   t'^voeo    «   '^oo    O    ^oo 

••••••••••»•••••••••,        ••■         •,     ^^ 

A      ••••«     •••■■     •••••     •■••«     •••     •«     « 




»o  VO 


































































































•            • 

^  so 





















•  «H 




•        • 













oooooooowoooooooooooo   r^i^t^t>»c^ 

•  •••••••••  •••  s  •  a 

qoqooqqoqqoqqoo    o 
dci^^oood  risdoo  o  'j-vo  p   M\d    o 


2.sii'fi.*S.o  cJnooo  d  ^vo  d  «\d   o 

t;     ^Ontj-vOoo     ON'«tvO00     OctTfvOOO     0C4nO000^nOOm\O     6 
O     ^«    «    «    «    M    »0»o»OfOcnTj-'^4'*j-^iO»nitnuotOsOvO    l^t^t^» 

'A*OtO^^^-^^ir>iO»0»OiOiOvOlONO>0    i>ii^r>i«>»oooOoo    o\  a^ 
t2    S^sooo     On    ^OOO     QW    ^nOoO     O    C4SO00O     -^NO    O     MVO     6 

^x^^io»oio»ovqvqvo«q    r>.i>it>>t>>oooooo    a\a\OkO    o   «    m   k^    « 
<4-i    v«o   «   Tt-vdoo   d   «'   "^sdoo    dc«   •xfvdoo    d   nvoooo   lArCi-I   rot^M 


u    xtoiONONO<o    t^r^oooooooo    onon0^o»o    o    «-•    ^    «    «   w   m  t<i  ■^-  in. 


.^      'A  NO     l^l>.l>.000O     0^0^0^0     O     O     •-'     m     «     p«     C»     tO«0*iOiONONO     e>»0O 

q>     TfC*    «    «    M   CI    fO«Ofo«ofo^^^^^ioio»oio«ovovo   r*r*«>»oo 


6t,t>-oooo   ONoto   o   •-«   M   «    «   «   «   toeo^^ioNO'O    1^00   o^a^o   ►* 

1.  ,      ^•••»    ••••••••••    ..,,•    .  ^   .   ,   ^   _ 

>J      ^2     f     "^^OOO     ►-     fOlOl^O»»-i     rOlor^OkM     rOl^O^"^     lOt^M     t*50O     « 

O     i*«    p*    N    «    ei    fn«orofOcOTtTj-Tj-^^ioir>io»o«OvOvo    i^r^rCoo 

'Z^     tO^ONCO»-i«-ii-iwc«e0rOTj-^^msOvO0000O»00««fO^^ 
J^       •••••••••••••••    •••••• ....^ 

nt      VO     '**     »0t>.0»'-'     tO»ot«*0»«-     tri    iTi    t^    a%    —     t*^    t^    O*    p4    sO    QO     fi     ^00     « 

C«      5n     m     m     m     CI     rOrOrorOC»)<^'^'^<«t^iOiOiOt^«Ov5>2    t^^rCS 

xO    o    •-«    N   n    «    fO^tTj-^ioiovovo   r>iOOoo    O    o   "^    n    «oioiovo    r^ 
"f.  '::'   »o»oi^o»«-    fotoi>.o»i-i   fo»oi^o»N-   fooo   o<*nooo   «   ^oo   « 

C5.N     «     M     M     CI     »OrOf0roC0rtTj-Tf^^iOlO»OvO»O^s3     r*t>.Ts»5 
r/M     ct     ct     rOr0^iOlONO<ONO     l>»t^O0     O»0     O     «     fOeO«ONO     1^00     OnO 

""•••••  ....••  ....  ....,  ..*  a  •        '? 

o!!!     *^»Or^O»M     rOlOr^2Mi     rOiOl^O»Ct     Tj-00     O     Mvpoo    «     "^00     tf>, 

:2o  .:::::::::::::::::::::::::    • 

•D-c  w  ->     V     o     o     -,     '.     O     o     O    '-     c>     o     o    o     o     -^     o     o     i.  o    o    o    o    o    o     o 
CC        O«r!-vO00OM'^NO00CN'««fvO00ONvO00Q^v©O«>OO 

i-io      «    <^    w   «   «    fOfoco«oco^"^^^^»i^io»oioOvo^   t>»r^T>ka> 


s   . 

•^    ti*MMO\9^oor^voiO'^coc«cti->Oo    o^oovo»o*«Moot^^« 

9tl<«    W    «     O    O»00    l^vO    iOio^«*5«*><^    «     M    000     «>.lO     ^CO"^    o    t^m 

•♦i    Vioo    O   w   '^ioc^a\'-»   tnirttN.a\^    fOiotN.a\c««   ^10    0   c>l^OQ0   ^   m 

6  •••••••••«••••,»•••«•••••• 

fr,  QO    Q    r«^«ooo    ON    fOmtN.a\'-'    »Oiftt^o»fOtor>iO    wv^oo    n' 

C      . 

O    Su^o   *9^"^«9**!   ***   *^,   9   ®.   <^oooq   t>.r>iVO   "p^toei   q   cr\i^t>.»o<o 

U      tU*^^     lOlO^'^'^W     C»     M     i-i     O     O     OnoOOO     t^tOIO^M     «     O     ONt^iO 

O    voo    O    «   ^loQO   O   w   '^looo    o    «    »oiot^ovto»or>i*-    tor^ioo   «io 


>    (JoM   «   w    «   M    «    «otOfnfOfo^Tt^7r^^»o»oiovovOvovO    «^r>. 

'C  VOO  d  «  'ii-Ood  d  N  ^10  00  d  W  'TIOOO  d  co»or>i«-I  rOt^Ovfrr^i 
2  dB"^  ^  *^  <*•  "  *^  »OfOcOeOfO^^'^^^iOiO»OiONOvOvOv5  i^r>i 


O     alO     O^O^000000     t^l^sOlO     lOtO*/".   ^^rOtO»-i     »^    O     OnOO    r^vO     »OeO 

Y%      tt"^     O     O     O     Oia\ON000O00     t>.t^vONO    iO»0^eO«*>CI     M     M     O     ON00«O 
Q.?«0^•-•    «o»i^O00    C    p*    tiooo    O    w    1*1000    O    ^vOoo    «    "^00    Ovfor^ 

Oj/Mi-ii-iMM    000ONOtON00Q0Q0r>ii>iNOiotom^<^c>icicok 

-»      •—^2     fOiO^ON'-'O     rtlOqOO     W     'tOOO     O     -tNOOO     P*     "^-00     O     'Tt^ 
V       t^   '^     W     W     W     «     «     fO»0»OfOfO^^'1'^^iOtO»OlONOvO^     t>.l>.C^ 

H^  i»o«owc«c«iMN^^MOoooNO»aNOOoot^r^vcw.  ^"^foei 
•  0\^  t*5iOr>iON^  •^iOt>.ON—  'O'^IOOO  O  't'OoO  ''•  toe  C  "TOO 
f^NN     M     «     M     M     W     fOtOfO«0»0^-1-^^-1-IOu^iOiOvOvCNC     l^t^r>i 

^e     :::::::::::::::::::::::::: 
r£w  :::::::::::::::::::::::::: 

13*  **  o    o    u    o_  o    0^  D    o    o    ">    :;     J    o    c-  ,     ■     j  j 

ma         OM^OOQ     O«"!fvr00     OCt-1-vOOO      C     «NvO00O     "«*vO     0«NvOQ 
mO         W     W     M     W     «     fOrOfOtOcO^^^TT^iO^OiOiOlOvOvO     t>.l>.t>.o5 


Thus,  if  the  hydrometer,  when  placed  in  the  oil,  reads  26* 
Bauni6  and  the  temperature  of  the  oil  80**  F.,  the  correct  reading 
will  be  24°  Baum6  at  60**  F.  The  specific  gravity  test  is  an  im- 
portant one  ;  by  it  an  admixture  of  certain  oils  with  mineral  oil 
is  indicated.  For  instance,  a  lubricating  oil  of  specific  grav- 
ity 0.915  was  found  by  qualitative  analysis  to  be  composed  of 
mineral  oil  and  menhaden  oil.  Knowing  the  kinds  of  oil  com- 
posing the  mixture,  an  approximation  of  the  per  cents,  would 
be  obtained  as  follows  : 

Mineral  oil Specific  gravity  =  0.890  (B) 

Menhaden  oil "  **        =  0.927  (A) 

Specific  gravity  of  mixture  =  0.915  (M) 

Let  A  — M=i  C     (0.927 — 0.915  =  0.012) 
M — B  =  D.     (0.915 — 0.890  =  0.025) 

Then  ^  ,    ^  =  per  cent,  oi  A\-^  ~\^ 

C^D      ^ V0.037- 


—  z=L  per  cent.  o\  B\- 


The  result  being 

Menhaden  oil    67.5  per  cent. 

Mineral        "     32.5     " 

^-j — ^  =  per  cent,  of  B\ I  • 

C-\-D       ^  \o.o37/ 


A  more  rapid  method  is  graphically  thus :  in  Fig.  no  let  the 
abscissas  represent  per  cents,  and  the  ordinates  the  specific 
gravities.  From  the  point  indicated  (on  the  line  A — B)  0.915 
the  specific  gravity  of  the  mixture  the  per  cents,  are  read  on 
abscissa  line  67.5  for  A  and  32.5  per  cent  for  B, 

Another  instrument  used  for  the  determination  of  the  specific 
gravity  of  oils  is  the  Westphal  balance. 

This  apparatus  (Fig.  iii)  is  ver>'  accurate  and  should  be 
used  as  a  check  determination  of  the  gravity  made  by  the 

If  the  oil  is  too  thick,  at  ordinar>'  temperatures,  for  the  deter- 
mination of  the  gravity,  it  should  be  heated  sufficiently  and  the 
modified  Westphal  balance  (Fig.  112)  used. 




-] — p- 

"^  i  1 

1  '  1   ■ 

1 1 1 1 

1  1  U 

1  1  1  1 

'  IJ 1 











^,  ■  ■ 

■  ■  '■ 




— — 

t    ■    ■    ■    ■ 






— " 











: , 







"■  - 






,  ■  1  , 






)     2 

0  3 


0           i 

0        6 



0         8 



0       ISO 

If  only  small  amounts  of  the  oil  are  obtainable  a  small  pic- 
nonieter.  or  an  Araeo-picnometer  of  Eiclihorn  can  be  used.  This- 
invention  isciescnbed  by  Dr.  H.  Hensoldt,  of  the  Petrographical 
Laboratory  of  Columbia  College.  New  York,  in  the  "Scientific 
American  SupplemeMl"  of  March  21,  1S91.  with  a  drawing.  The 
important  feature  of  this  instrument  consists  in  a  small  glass 
bulb  (attached  to  the  spindle),  which  is  filled  with  the  liquid 
whose  gravity  is  to  be  taken.  Thus  instead  of  floating  the  en- 
tire apparatus  in  the  test  fluid,  only  a  very  small  quantity  o£ 
the  latter  is  required.      (Fig.  1 13. ) 

The  glass  bulb,  when  filled  with  the  test  fluid,  is  closed  by 
mesns  of  an  accurately  filling  glass  stopper,  and  the  instrument' 
is  then  placed  in  a  glass  cylinder  filled  with  distilled  water  at 
■7Y  c. 

The  gravity  is  then  at  once  shown  ou  the  divided  scale  in 
upper  portions  of  the  spindle. 

The  following  tabic  converts  degrees  of  the  \-arious  bydrome-f 
ters  into  specific  gravity.     (Liquids  lighter  than  ^t'ater.) 







Gay-Lussac,  4   C  = : —  =  specific  gravity. 

ioo-\-n  r-  ^         J 

Beck,  12.5''  C.  = q- —  =  specific  gravity. 

T  ifi    R 

Cartier,  12.5"*  C.=  .^^  .  |-z  =  specific  gravity. 

Baum6  hydrometer,  at  15''  C. 

Brix  hydrometer,  Fischer 
hydrometer  at  15.6"*  C. 



= i —  =sp.  gr. 

400 +  «        ^   ^ 

n  =.  degrees  indicated  upon  the  spindle. 

Fig.  113. 

59 ;  fi»  65-73. 

Table  of  Specific  Gravity  of  Oils  Used  with  Mineral 
Oils  for  Lubricating  Purposes. 

Sperm  oil 0.883  specific  gravity. 

Olive  oil 0.916 

Cotton-seed  oil  (white) 0.925 

Cotton-seed  oil  (brown ) 0.930 

Castor  oil 0.960 

Dolphin  oil 0.922 

Neat's  foot  oil 0.915 

Lard  oil 0.915 

Tallow  oil 0.903 

Menhaden  oil 0.928 

Rape-seed  oil 0.916 

Rosin   oil 0.980  to  i  .05 

Blown  oils,  made  by  oxidation  of  rape- 
seed  oil,  cotton-seed  oil,  etc., 
(consult   Chapter  46) . .  .0.930  to  0.970 

References  on  the  specific  gravity  of  oils  : 

"On  Fluid  Specific  Gravity  Determinations  for  Practical 
Purposes."  By  C.  R.  Alder  Wright,  F.R.S.,  /.  Soc,  Chem. 
Ind.t  XI,  297. 

"On  the  Chemistry  and  .Analytical  Examination  of  Fixed 
Oils.'*     By  Alfred  H.  Allen,  F.C.S.,  /.  Soc.  Chem,  Ind.,  2,  49- 

<  < 

<  t 

<  t 



1 1 




•  < 



I  ( 

The  Cold  Test. 

The  degree  at  which  an  oil  becomes  semi-solid  and  refuses  to 
flow  freely  is  considered  the  cold  test,  and  is  performed  as  fol- 
lows : 

Fifty  cc.  of  the  oil  are  transferred  to  a  narrow  bottle  (capacity 



loo  cc),  Stoppered  with  a  rubber  Stopper,  through  which  is  in- 
serted a  thermometer,  the  bulb  of  which  reaches  an  inch  or  more 
into  the  oil. 

The  bottle  is  placed  in  a  mixture  of  ice  and  salt,  or  other 
freezing  compound,  and  retained  there  until  the  oil  becomes 
solid.  It  is  then  removed  and 
allowed  to  warm  until  the 
contents  become  somewhat 
thinner  in  consistence.  The 
bottle  is  inclined  from  side  to 
side  until  the  oil  begins  to 
flow,  when  the  temperature  is 

At  this  particular  tempera- 
ture the  oil  is  neither  at  its 
normal  fluidity,  nor  is  it  solid, 
and  while  this  method  does 
not  correctly  indicate  the  ex- 
act temperature  of  the  solidi- 
fying point,  it  does  show  the 
point  at  which  the  oil  ceases 
to  flow  readily,  the  import- 
ant one  to  the  oil  inspector. 

In  lubricating  oils,  to  be 
used  in  railroad  practice,  this 
cold  test  is  a  vital  one,  and 
receives  in  the  laboratories  of 
the  different  railroads  of  the 
United  States  considerable  at- 

A  mineral  lubricating  oil, 

non-paraffin,  of  good  quality, 

does  not  show  any  material 

difference   in  its  consistency 

at  25°  C.  or  10°  C,  but  a  radical  change  would  be  indicated  at 

10°  C.  i(  some  of  the  animal  or  vegetable  oils  were  a  component. 

While  it  is  true  that  no  proportion  of  one  or  the  other  can  be 

indicated  by  the  cold  lest,  and  that  this  test  might  not  properly 

■II     |l[[[[![fl^ 



f  be  classed  as  a  chemical,  but  rather  as  a  physical  one,  yet  so 
I  important  is  this  property  of  congealing  in   lubrication,  and  as 

.  all  laboratories  connected  with  railroad  work  rely  strongly  upon 
t,  it  is  included  as  one  of  the  principal  ones. 
In  connection  therewith  is  here  included  the  drawings  of  the 
I  apparatus  used  for  this  purpose  in  the  cheniicallaboratory  of  the 


Chicago,   Burlington  and    Quincy  Railroad  Co.,   Aurora,   111. 

Fig.  114  represents  the  glass  apparatus  with  the  thermometer 
arranged  for  the  cold  test. 

Fig.  115  represents  the  cold  box  to  contain  the  freezing  mix- 
ture and  in  which  the  oil  is  tested. 

The  following  determinations  of  the  cold  test,  made  in  my 
laborator>%  will  show  the  wide  range  in  this  regard  between  many 
of  the  oils,  used  in  lubrication  : 

Elain  oil 6^  C. 

Saponified  red  oil 5 

Prime  neat's  foot  oil 4 

White  neat's  foot  oil 4 

Pure  hoof  oil 6 

Prime  lard  oil 7 

No.  I  lard  oil 7 

XXX  lard  oil 3 

American  sod  oil i 

English  sod  oil 24 

Tallow  oil 26 

Dog  fish  oil 7 

Right  whale  oil  (Pacific) o 

Unbleached  bowhead  whale  oil  (Pacific) 7 

Bleached  whale  oil  (Pacific) 13 

Natural  sperm  oil  (Pacific) o 

Bleached  sperm  oil      "         4 

Herring  oil  "         o 

Natural  winter  sperm  oil  (Atlantic) i 

Bleached  winter  sperm  oil        "         4 

Natural  spring  sperm  oil  "         10 

Bleached  spring  sperm  oil        *'         8 

Natural  winter  whale  oil  "         2 

Bleached  winter  whale  oil        "         5 

Natural  spring  w^hale  oil  "         5 

Bleached  spring  whale  oil        **         2 

Prime  crude  menhaden  oil 4 

Brown  strained  menhaden  oil 7 

Light  strained  menhaden  oil 7 

Natural  winter  menhaden  oil 9 

Bleached  winter  menhaden  oil 12 

Extra  bleached  winter  white  menhaden  oil 11 

Bank  oil 4 

Straits  oil 7 

Sea  elephant  oil 5 


Black  fish  oil 8°  C. 

Rosin  oil,  ist  run 3 

'*        '*    2d  run 19 

•'        •*    3d  run 20 

Castor  oil 18 

Crude  cotton-seed  oil 7 

Prime  summer  yellow  cotton-seed  oil 5 

Off  quality  summer  yellow  cotton-seed  oil 6 

Prime  quality  winter  cotton-seed  oil 10 

Off  quality  winter  cotton-seed  oil 8 

Prime  quality  summer  white  cotton-seed  oil 3 

Off  quality  summer  white  cotton-seed  oil 8 

Prime  quality  winter  white  cotton-seed  oil 9 

Off  quality  winter  white  cotton-seed  oil 5 

No.  I  French  Degras  oil 25 

No.  2       •*  "        *•  25 

English  Degras  oil 18 

Olive  oil 3 

Oleo  oil 24 

In  the  specifications,  for  the  supply  of  oils  to  the  various  rail- 
roads, it  is  generally  stated  what  degree  is  required  for  the  cold 
test.     Thus  the  Pennsylvania  Railroad  Co.  requires  as  follows  : 

Lard  oil  8^  C.  November  i  to  April  i. 
Tallow  oil,  8^  C. 
Neat's  foot  oil. 8°  C." 

Baltimore  &  Ohio  Railroad  Co  ; 

Engine  oil  from  October  i  to  May  i,  below  9°  C.  . 

Passenger  car  oil  *'  **      **        "      "        "      **  •* 

Freight  car  oil       '*  "       "        "      "        '*      "  " 

Chicago,  Burlington  &  Quincy  Railroad  Co.   Black  Engine 

oils  : 

Summer  oil  must  flow  at  15*^  C.  and  above. 
25°  oil  ••        **     "     i^C.     ** 

15^  oil  *'        **     •*    9-C.     " 

Zero  oil  '*        "     ♦*  15^  C.     " 

Tagliabue's  standard  lubricating  oil  freezer  is  also  largely  used 
in  this  connection,  and  is  thus  described.  It  consists  of  a  semi- 
cylindric  metallic  stand,  neatly  japanned,  divided  into  three  com- 
partments.    (Apparatus  is  shown  in  Figs.  91  and  116). 

The  first,  y,  is  the  oil  cooling  chamber,  in  which  is  the  glass 
receiver,  adjusted  to  a  rocking  shaft,  ^,  to  facilitate  the  introduc- 

il  14  <t  <( 


lion  of  the  regulation  oil  cup  therein,  and  to  show  by  ils  motion 
whether  the  oil  is  congealing  or  not. 

The  second,  c,  is  the  ice  chamber  which  is  filled  with  ice  and 
rock  salt  for  the  cooling  process  ;  a  faucet,  k,  is  connected  with 
it,  to  allow  the  melted  ice  to  flow  out.  The  third,  a,  is  a  non- 
conductor jacket,  lined  with  mineral  woo!  filling,  to  maintain 
an  even  temperature  in  the  cooling  chamber,  and  to  prevent  a 
too  rapid  melting  of  the  ice. 

Three  thermometers,  d.  are  inserted    in    the    freezer,  one  ( 

each  side  of  the  cooHng  chamber,  to  denote  its  temperature  and 
a  third  one  in  the  center  so  adjusted  that  its  bulb,  penetrating 
the  middle  of  the  oil,  enables  one  to  .see  through  the  glass  door„ 
k,   (without  opening  the  same,)  at  what  temperature  the  C 

Two  stop-cocks, _/.  are  attached  to  the  bottom,  with  the  c 
ing  chamber,  to  force  therein  (by  either  opening  or  blovr 
through  them  with  a  rubber  tube)  atmospheric  or  warm  i 
whenever  it  is  desired  to  raise  its  temperature. 



The  first  instrument  tor  the  deter- 
mination of  the  viscosity  of  oils  was 
probably  Schubler's.  (Fig.  117).  It 
consisted  of  a  glass  cylinder,  open  at 
the  top  and  drawn  to  a  one  thirty- 
second  inch  tube  at  the  bottom.  Hav- 
ing filled  the  cylinder  with  the  oil  to 
be  tested,  the  time  required  fpr  100 
cc.  of  the  oil  to  flow  out  through  the 
aperture  was  noted,  and  this  figure 
compared  with  thatobtained  from  wa- 
ter under  similar  conditions. 

Thus,  Schubler  records,  among 
many  determinations,  the  following: 


■      A 

Colza  oil i6z.o 

Olive  oil 195.0 

Hcmp-seeil  oil 87.0 

Castor  oil   1830.0 

Distilled  water 9.0 

The  Pennsylvania  Railroad  Co.  viscosity  tests  are  made 

A  100  cc.  pipette  of  the  long  bulb  form  is  regraduated  to  hold 
just  100  cc.  to  the  bottom  of  the  bulb.  The  size  of  the  aperture 
at  the  bottom  is  then  made  such  that  100  cc.  of  water  at  100° 
F.  will  run  out  of  the  pipette  down  to  the  bottom  of  the  bulb 
m  thirty-four  seconds. 

Pipettes  with  bulbs  varying  from  one  and  three-fourths  inches 
to  one  and  one-half  inches  in  diameter  outside,  and  about  four 
and  oue-half  inches  long,  give  al  most  exactly  the  same  results,  pro- 
vided the  aperture  at  the  bottom  is  the  proper  size.  The  pipette 
being  obtained,  the  oil  sample  is  healed  to  the  required  tempera- 
ture, care  being  taken  to  have  it  uniformly  heated,  and  then  is 
drawn  up  into  the  pipette  to  the  proper  mark.  The  time  occu- 
ijcd  by  the  oil  in  running  out,  down  to  the  bottom  of  the  bulb 


gives  tile  test  figures.  A  stop  watch  is  convenient,  but  not 
essential,  in  making  the  test.  The  temperature  of  the  room 
affects  the  test  a  little.  The  limiting  figures  were  obtained  in  a 
room  at  from  70°  to  80°  F.  It  will  not  usually  be  possible  to 
make  duplicate  tests  without  readjustment  of  the  temperature 
of  the  oil. 

These  pipettes  are  in  use  in  many  railroad  laboratories  in  the 
United  States,  but  are  difficulttoclean.and  are  not  as  convenient 
as  the  Eugler  or  Redwood  viscosi meters. 

Engler's  viscosimeter  (original  form,  Fig.  118)  is  con- 
structed of  copper.and 
consists  of  .-?.  a 
chamber  holding  the 
oil  to  be  tested  ;  B, 
the  water  bath.  C.  a 
flask  graduated  so  as 
to  receive  200  cc.  of 
the  oil  a,  b,  ther- 
mometers ;  e the  open- 
ing through  which 
the  heated  oil  flows 
out  upon  the  with- 
drawal of  the  plug  d. 
I  n  using  this  instru- 
ment the  viscosity  of 
an  oil  is  stated  in 
seconds  required  for 
20OCC.  oftheoil  to  run 
into  the  flask  C  Heat 
can  be  applied  to  the 
water-bath,  the  \\s- 
cositY  being  deter- 
mined at  any  tempera- 
ture required  up  to  100°  C.  Any  temperature  up  to  360°  C.  can 
be  secured  by  filling  B  with  paraffin  instead  of  water. 

Engter  recommends  that  all  viscosity  be  compared  with  water 
thus : 


If  water  requires  52  seconds  for  delivery  of  300  cc.  into  the  re- 
ceiving flask,  and  an  oil  under  examination  requites  i30seconds, 

viscosity  of  2.5  times  that  of  water. 

This  instrument  has  been  for  many  years  the  standard  i 

Boverton  Redwood'  describes  a  -v-iscosimeter  (Fig.  119),  the 
general  principle  of  which  is  the  same  as  Engler's.  This  is  the 
standard  viscosimeter  for  the  English  oil  trade. 

The  septometer{  Figs.  120,  121 ),  originated  with  Dr.  Lepenau, 
is  used  for  the  direct  comparison  of  the  T.-iscosity  of  two  oils  j 
under  similar  conditions  at  the  same  moment-     It  consists  of  | 
two  cylindrical  vessels,  B.  B,  which  hold  the  oil.s  to  be  compared.  ' 

and  which  stand  in  the  same  water  bath.  A,  and  have  the  same 
temperature.  To  the  apparatus  the  holder,  A,  is  filled  with 
water,  which  can  be  heated  at  any  temperature  desired  below 
100°  C  ;  if  higher  temperatures  are  desired,  A  must  be  filled  with 
oil.  The  vessel.  B.  is  filled  with  the  oil  which  is  taken  for  the 
standard,  such  as  rape  oil  or  lard  oil,  and  the  second  one  is  filled 
with  the  oil  to  be  tested.  Since  the  heated  or  cooled  water  is 
stirred  regularly  the  oils  have  the  same  temperatures  which  are 
read  from  the  thenuomelers.  /,  t.  For  comparison  the  oils  are 
allowed  to  flow  out,  at  the  same  time,  for  the  same  length  of 
time.  The  relative  value  sought  is  found  then  by  measuring  or  g 
weighing  the  amounts  which  have  flowed  out. 

ly.  Sat.  C*"t.  fnd..  i,  IJS. 


Davidson's  viscosimeter  (Fig,  122)  is  designed  especially  for 
determining  the  relative  \'iscosity  of  oils  and  greases  when  heated 
to  the  temperature  of  locomotive  cylinders  (250°  to  375°  F.)- 

The  entire  apparatus,  except  the  glass  portion,  is  made  of 
copper  and  the  joints  brazed. 
I  The  oil  to  be  tested  isput  into  the  cylinder,  A,  and  the  cup.  R. 
'  which  are  connected  through  the  stop-cock  C.  The  cylinder,  A, 
is  also  connected  with  the  glass  gauge  through  the  tubes,  H,  and 
//,  so  that  the  height  of  the  oil  in  the  cylinder  can  be  seen. 
The  bottom  of  cylinder  A  is  covered  by  a  brass  plate,  through 
which  is  bored  a  hole  three  and  one-half  inches  in  diameter,  which 
can  be  closed  by  the  slide  valve.  £.  against  the  plate  by  a  spring. 
The  outside  of  the  plate  is  beveled  from  the  hole,  so  that  the 
hole  is  in  a  very  thin  plate,  and  thus  lateral  friction  is  reduced 
to  a  minimum.  A  long  thermometer  is  used,  so  that  the  bulb 
will  be  near  the  bottom  of  cylinder  A. 

The  cylinders,  B  and  B\  contain  the  lard  oil  bath  that  is  used 
for  conveying  heat  to  the  oil  in  cylinder  A.  Heat  is  applied  by 
lamp  or  gas  burner  at  the  base  of  cylinder  B',  and  the  hot  prod- 
ucts of  combustion  allowed  to  pass  through  tlje  cylinder  G.  As 
the  lard  oil  in  B'  becomes  heated,  it  rises  to  the  top  of  this 
cylinder,  and  passes  over  to  cylinder  B,  down  B,  passing  around 
the  cylinder  A,  and  back  to  B',  where  it  is  reheated  and  recircu- 
lated, as  shown  by  the  arrows.  The  oil  in  cup  /f  is  heated  by 
the  products  of  combustion  escaping  from  the  top  of  cylinder  C. 
and  in  case  of  a  high  temperature  bj-  an  additional  lamp  placed 
under  the  cnp  /i. 

When  the  oil  under  lest  in  A  and  ^  has  reached  the  desired 
temperature,  the  valve,  £.  is  opened  and  the  stop-cock  C  is 
adjusted  to  keep  the  height  of  oil  in  A  the  height  desired,  as 
shown  by  the  glass  gauge.  A  lOO  cc.  flask,  which  is  immersed 
in  hot  oil,  is  theu  placed  under  the  stream  of  oi!  flowing  from 
the  hole,  and  a  stop-watch  is  started  the  instant  the  oil  com- 
mences to  run  into  the  flask.  When  100  cc.  have  been  delivered 
into  the  flask,  the  watch  is  stopped.  The  number  of  seconds 
ftiequired  for  this  is  the  viscosity  of  the  oil  under  examination. 



Tagliabue's  vicosimeter  (Fig.  123),  consists  of  a  copper  basin, 
C  extending  bj'  means  of  the  coiled  tube  to  the  outlet  at  the  stop- 
cock on  the  outside  of  the  vessel. 

This  is  surrounded  by  the  water  bath,  B,  which  has  an  outer 
chamber  a  connected  bj-  two  tubes,  and  in  which  the  water  is 
poured  into  the  bath.  Z>  is  a  thennometer,  and  records  the 
temperature  of  the  water-bath. 

To  test  an  oil,  the  water-bath  is  filled  two-thirds  full  and 
heated  by  means  of  a  small  Bunsen  burner  or  alcohol  lamp. 
The  top  basin,  C,  lined  with  wire  gauze  is  filled  with  the  oil  to 
be  tested,  and  when  the  thermometer,  D,  indicates  100"  C,  the 
glass  measuring  flask,  E.  is  placed  under  the  faucet,  which  is 
opened  with  the  starting  of  the  watch. 

When  fifty  cc.  of  the  oil  Have  run  out  and  reached  the  mark 



upon  the  ueck  of  the  receiving  flask,  E,  the  watch  is  stopped, 

and  the  number  of  seconds  required  noted. 

The  ^-iscosity  of  the  oil  is  stated  in  seconds. 

This  viscosimeter  has  a  very  extended  use  in  the  oil  trade  but 
it  is  a  difficult  piece  of  apparatus  to  clean  when  any  particles  ol 


dirt  have  become  lodged  in  the  coil.  This  materially  interferes 
with  the  flow  of  oil  through  the  tube  and  gives  false  results. 
The  basin.  C.  as  well  as  the  coil,  cannot  be  removed,  as  they  are 
brazed  to  the  water-balh. 

For  this  reason,  and  also  when  used  at  higher  temperatures, 
the  faucet  being  metallic  and  not  heated  to  the  temperature  of 


the  oil,  the  oil  leaves  the  apparatus  much  cooler  than  the  tem- 
perature recorded  by  the  thermometer  of  the  water-bath. 

Gibb's  \'iscosiineter.  Fig.  124  (George  Gibbs,  M.  E.,  Chicago, 
Milwaukee  and  St.  Paul  Railroad),  was  designed  to  overcome 
some  objectionable  points  in  existing  forms  of  visco  si  meters. 

The  idea  being:  First. — To  have  a  large  body  of  hot  oil  as  a 
balh  surrounding  the  oil  to  be  tested  in  order  to  keep  the  latter 
at  a  perfectly  uniform  temperature. 

Second. — To  apply  a  forced  circulation  to  the  bath  by  means 
of  a  double  action  pump,  to  insure  equality  of  heat  in  all  parts. 

Third. — To  deliver  the  oil  to  be  tested  at  the  orifice  under  a 
constant  head,  which  is  accomplished  by  means  of  a  pneumatic 

Fourlk. — To  supply  convenient  means  for  accurately  measur- 
ing the  temperature  of  the  oil  near  its  delivery  point. 

The  large  reservoir  a  is  of  copper,  with  heavy  brazed  bottom. 
This  contains  the  cylindrical  inside  chamber  with  conical  bottom, 
B.  At  the  lower  end  of  this  is  the  gauged  aperture,  T.  Inside 
of  this  chamber  fits  the  inverted  reservoir,  C,  holding  the  oil  to 
be  tested,  lathe  interior  of  this  chamber  is  a  tube,/),  extendiug 
nearly  to  the  bottom  of  the  same.  This  tube  admits  air  to  deter- 
mine the  head  of  the  oil,  aud  also  to  admit  the  thermometer,  F. 
The  outside  bath,  a.  contains  the  deflector  plates,  O,  P  and  R 
to  obtain  proper  mixing  of  the  bath.  The  heating  of  the  balh 
is  done  by  a  lamp,  W.  set  underneath  the  separate  heating 
chamber,  G.     The  size  of  the  orifice  at  7"  is  one-sixteenth  inch. 

The  following  table  shows  the  result  of  viscosity  tests  upon 
various  oils  made  with  this  instrument. 






Per  cent 





300*  p. 

350*  p. 


Nat.  Ref*g  Co.,  Loco. 


Nat.  Ref  g  Co., German 
Perfection  valve  oil  •  •  • 
**  (another) 





Vacuum  valve  oil 

C.,M.  &St.  P.valveoil 
Extra  lard  oil  (average 

of  3  samples) 

St  d  Oil  Co., No.  I  stock 

















38  sec. 















Viscosities  expressed  in  seconds  for  50  cc. 


National  Ref'g  Co.,  car  oil 
Relief  Oil  Works, 
Galena  car  oil 



(  c 

Relief  Oil  Works,  engine  oil 
National  Ref'g  Co.,    " 






Per  cent. 























75"  P. 






no*  F. 




150'  F. 






Viscosities  expressed  in  seconds  for  50  cc. 


The  viscosities  of  a  oamber  of  other  oils,  at  the  temperature 
of  locomotive  cylinders,  as  made  by  this  instmment,  are  shown 
in  the  chart  of  curves.     (Fig.  125.) 




»   1 






\  \ 

C.M^.i^jST.pW   Ry.tjo. 


























\  \ 





,  \ 
















































A  viscosimeter  on  an  entirely  difierent  principle  than  the 
others  already  described  is  the  Perkins  instrument  (G.  H. 
Perkins,  Supt.  Atlantic  Oil  Refinery,  Phila.,  Pa.)     It  consists 


of  a  cylindrical  vessel  of  glass,  surrounded  by  a  proper  heating 
vessel,  and  fitted  with  a  piston.  This  piston  fits  into  the  cylin- 
der to  within  -j^^  of  an  inch. 

In  practice,  the  cylinder  is  filled  nearly  full  with  the  oil  to  be 
tested  and  the  piston  inserted.  The  time  required  for  the  piston 
to  sink  a  certain  distance  into  the  oil  is  taken  as  the  measure  of 
viscosity.  A  full  description  of  the  apparatus  will  be  found  in 
Transactions  of  the  American  Society  of  Mechanical  Engineers^  9, 

p.  375. 
J.  Lew',  introduces  an  instrument  not  only  for  the  viscosity 

but  also  to  include  the  internal  friction  of  an  oil.     By  these 

means  it  is  claimed  the  lubricating  value  of  the  oil  is  absolutely 


The  author  states  that  the  internal  frictional  resistances  are 
different,  and  vary  in  the  different  oils  at  various  temperatures. 
Formulas  and  methods  are  given  by  which  coefficients  are 
determined  and  used  in  the  examination  of  the  lubricating  value 
of  oils. 

Figure  126  represents  the  viscosimeter  designed  and  used  in 
the  chemical  laboratory  at  the  Stevens  Institute  of  Technology. 

It  consists  of  a  copper  bath,  B,  surrounding  the  vessel,  A^  also 
of  copper,  and  which  holds  the  oil  whose  viscosity  is  to  be  deter- 
mined. The  tube/ is  of  copper,  but  at  e  it  is  joined  to  a  glass 
tube,  which  is  extended  to  d — this  latter  is  used  for  measuring 
the  oil,  and  is  carefully  graduated.  Sizes  and  dimensions  of  the 
apparatus  are  given  in  the  figure. 

This  apparatus  was  designed  to  overcome  two  difficulties 
usually  occurring  in  the  use  of  other  viscosimeters ;  viz.  :  Firsts 
loss  of  heat  in  the  oil  during  its  passage  from  the  containing 
vessel  to  the  receiving  flask ;  and  second^  to  have  the  chamber,  A, 
of  size  to  work  small  quantities  of  oil.  First. — When  the  vis- 
cosity of  an  oil  is  taken  at  the  ordinary  temperature  the  measure- 
ment of  the  oil  in  the  receiving  flask  will  correctly  indicate  the 
amount  of  oil  delivered  through  the  aperture.  The  conditions 
are  altered,  however,  when  high  temperatures  are  required, 
since  the  oil  in  running  in  a  fine  stream  through  the  orifice  is 
chilled  in  contact  with  the  air,  and  if  its  temperature  be  taken 

1  Ding. poly.  J.y  1891,  aSo. 


Ml.»l  N0T|eO.«I.T. 

Fig.   136. 


at  the  moment  its  volume  is  read  in  the  receiving  flask,  a  notable 
difference  is  indicated,  depending  upon  the  temperature  of  the 
room  and  of  the  oil  before  delivery. 

In  this  instrument  provision  is  made  for  reading  the  volume 
of  the  oil  directly  in  the  chamber  A  without  any  graduated  re- 
ceiving flask,  as  follows : 

The  tube  fedis  graduated  so  that  when  the  oil  in  the  vessel 
A  is  at  the  proper  level,  the  oil  also  reaches  the  upper  graduated 
mark  upon  the  tube  d  e.  The  lower  graduated  mark  upon  the 
tube  indicates  when  twenty-five  cc.  of  the  oil  have  been  delivered 
from  A  through  the  orifice^. 

This  graduation  is  absolutely  correct  for  the  purpose,  and 
shows  accurately  the  viscosity  of  the  oil  at  an}'  temperature,  as 
indicated  by  the  thermometer  in  -^. 

None  of  the  oil  in  tube  from  ^  to  ^  passes  into  A  during  the 
delivery  of  the  twenty-five  cc.  through^,  since  the  tube/^  d  is 
only  partially  emptied  of  its  oil,  the  level  of  the  oil  in  A  after 
the  deliveryof  the  twenty-five  cc.  still  remaining  above  the  point 
where  the  tube /enters  A. 

Second, — Oftentimes  the  samples  of  oil  sent  for  examination 
do  not  exceed  loo  cc.  in  bulk,  an  amount  entirely  too  small  if 
other  tests  are  to  be  included. 

Many  forms  of  viscosimeters  require  loo  cc.  of  oil  for  the  vis- 
cosity test,  and  not  a  few  fifty  cc. 

I  have  found  twenty-five  cc.  to  be  ample,  provided  the  aperture 
at  g  is  small  enough  to  prevent  a  too  rapid  delivery  of  the  oil 
and  consequently  render  close  readings  and  comparisons  difficult. 
By  making  this  orifice  three  sixty-fourths  inch,  sufficient  time  is 
secured  to  obtain  accurate  results. 

If  the  operator  prefers  not  to  use  the  graduated  txxh^fed  to 
measure  the  oil,  a  receiving  flask,  properly  marked,  can  be 
placed  under  g^  as  in  other  forms  of  viscosimeters. 

The  plan  suggested  by  Schublerthat  viscosities  should  be  com- 
parable with  water  is  the  only  proper  one  and  in  the  following 
determinations  of  viscosity  the  comparison  is  included  : 


No.  I      ■'       " 

XXX     '■       " 

Prime  neat'*  foot  oil 

White       '    

Pure  hoof  oil 

Oleo  oil 

Horse  oil 

Gelatine  oil 

Roein  oil,  ist  rnu 

"   Id    '■   

'■        '■   3d    '■   

Dog-fish  oil  (Pacific) 

Right  whale  oil  (Pacific) 

Natural  bow  head  oil  ( Pacific)  -  - 

Bleached  whale  oil 

Natural  winter         ) 

Sperm  oil  (PacificH   

Bleached  sperm  oil  (Pacific).  ■■ 

Natural  spring  sperm  oil 

Bleached  spring  sperm  oil 

Natural  winter  whale  oil  (Atlan- 

Bleached  winter  whale  oil  (At- 

Extra  bleached  winter  whate  oil 

Natural  spring  whale  oiHAtli 

Bleached  spring  whale  oil  ■ 

Porpoise  head  oil 

Sea  elephant  oil 

Bank  oil 

Priniecrude  menhaden  oil- - 

Brown  strained     "  "  .. 

Light        '■  "  "  ... 

Natural  winter  menhaden  oil-. 

Bleached    "  " 

Extra   bleached    winter   white 
menhaden  oil 

Castor  oil 

"White  seal"  castor  blown  oil 

Prime    quality   summer    white 

Prime    quality     winter     white 

Herring  oil  ( Pacific) 


A  chart  of  a  few  of  the  above  oils  is  shown  on  the  followiag  page. 

quantitativb  analysis. 
Viscosity  Tbsts. 


■f— KmId  oil  (Kcood  run).  h— prime  qiul.  ■unincr  white  cottoc-ieed  oil 

(—Bleacbedvluleiril  (Pacific).  ■— Kape  oil. 

/—Sperm  oil  (Pacific).  o— OUtc  oil. 

(—Natural  winter  wtulc  oil  (Atlantic),  ^— "  Dcsru"  d1. 

*— Porpol»e  head  oil.  r— Roilnirfl  (flrtl  mn). 





















































'•■punn*'  *a*il>C 



An  examination  of  these  tables  and  cur\'es  brings  prominentlj' 
f  forward  the  following  facts  : 

That  at  high  temperatures  the  variation  in  the  viscosity  of  sim- 
[  pie  oils  is  ver\'  slight. 

That  "  blown"  oils,  and  "gelatine"  oils,  which  aremanufac- 
I  tured  especially  to  give  "  body"  to  compounded  oils  fail  in  their 
■  purpose  at  high  temperatures. 

This  is  shown  especially  in  Fig.  127,  by  thecur\'es  of  thecom- 
pounded  oil,  for  instance,  which  at  20°  C.  remains  solid,  like- 
wise at  50"  C.  and  100'  C,  but  at  [50°  C.  (302°  F.}it  indicates  a 
viscosity  of  360  seconds,  and  at  200  C . ,  a  viscosity  of  35  seconds. 

This  ■' gelatine"  oil  is  generally  a  compound  of  aluminum 
oleate,  lard  and  petroleum. 

Castor  oil  shows  the  highest  variation  of  any  of  the  simple 
oils,  while  sperm  oil  shows  the  least,  and  it  is  probably  this 
property  of  the  latter  that  has  given  it  the  reputation  as  the 
standard  oil  in  lubrication. 

Of  the  animal  oils.  lard  oil  ranks  first  in  lubrication,  followed 
in  order  by  neat's  foot,  horse  oil  and  tallow  oil. 

Generally  speaking,  the  marine  oils  are  the  better  lubricants, 
with  the  exception  that  acidity  often  rapidly  forms  in  them,  and 
so  renders  them  valueless  for  the  lubrication  of  many  forms  of 
machinery.  The  order  of  their  value  would  be  sperm  oil,  por- 
poise head  oil,  bleached  menhaden  oil,  whale  oil,  dog  fish  oil, 
sea  elephant  oil  and  herring  oil.  Of  the  vegetable  oils,  rape  oil 
is  the  recognized  standard  in  lubrication.  Its  use  for  this  pur- 
pose is  very  limited  in  this  country,  though  in  Germany  and 
Russia  large  amounts  are  annually  consumed. 

Olive  oil,  while  a  good  lubricant,  is  too  high  in  price  and  its 
place  has  been  taken  in  later  years  by  re6ned  cotton-seed  oil. 
This  latter  oil,  while  seldom  used  alone  in  lubrication,  is  added 
to  lard  oil  in  proportions  varying  from  twenty  to  fifty  per  cent., 
producing  a  mixture  that  lubricates  nearly  as  well  as  pure  lard 
oil.  though  acidity  more  rapidly  develops  than  in  lard  oil  alone. 
Castor  oil  is  largely  added  to  other  oils  to  give  high  viscosity  at 
ordinary  temperatures,  and  to  produce  "  body."  which  tt  loses  at 
high  temperatures.  Its  use  for  this  purpose  still  continues  in  Eng- 
land, while  in  this  countr}'  its  application  is  limited. 




The  so-called  "  seal  castors"  and  "  blown  oils"  are  made  from 
cotton-seed  oil,  and  are  used  in  place  of  "  gelatine"  oil  to  pro- 
duce high  viscosity,  at  a  much  lower  cost  than  "  gelatine"  oiL 
"The  Doolittle  torsion  viscosi meter"'  recently  introduced, 
(1893)  is  nsed  in  the  railroad  laboratories  of  the  Philadelphia 
and  Reading  Railroad  Co.     It  is  briefly  described  as  follows  : 

A  steel  wire  is  suspended  from  a  firm  sup- 
port and  fastened  to  a  stem  which  passes 
through  a  graduated  horizontal  disk,  thus 
measuring  accurately  the  torsion  of  the 
wire.  The  disk  is  adjusted  so  that  the 
index  point  reads  exactly  zero,  thus 
showing  that    there  is  no  torsion    in  the 


A  cylinder  two  inches  long  by  one  and 
half  inches  in  diameter,  having  a  slem 
stem  by  which  to  suspend  it,  is  then 
mersed  in  the  oil  and  fastened  by  a  thumb- 
screw on  the  lower  part  of  the  stem  to  the 
disk.  The  oil  is  surrounded  by  a  bath 
I  of  water  or  paraffin  wax  according  to  the 
temperature  at  which  it  is  desired  to  take 
the  viscosity.  This  temperature  being 
obtained  while  the  disk  is  resting  on  its 
supports,  the  wire  is  twisted  360°  by  means 
of  the  knob  at  the  top.  The  disk  being 
released,  the  cylinder  rotates  in  the  oil  by 
virtue  of  the  torsion  of  the  wire. 

The  action  now  observed  is  identical  with 
that  of  the  pendulum. 
If  there  was  no  resistance  to  be  overcome,  the  disk  would  re-' 
volve  back  to  zero,  and  the  momentum  thus  acquired  would  carry 
it  to  360°  in  the  opposite  direction.  What  we  find  is  that  the 
resistance  of  the  oil  to  the  rotation  of  the  cylinder  causes  the 
revolution  to  fall  short  oi  360",  and  that  the  greaterthe  viscosity 
of  the  oil  the  greater  will  be  the  resistance  and  hence  the  retarda- 

t/.^m.Ckim.  Joe.,  JJ,  ITJ, 






tion.     We  find  this  retardation  to  be  a  very  delicate  measure  of 
the  viscosity  of  an  oil. 

There  are  c  number  of  ways  in  which  this  viscosity  may  be 
expressed,  but  the  simplest  is  found  to  be  directly  in  the 
number  of  degrees  of  retardation  between  the  first  and  second 
complete  arcs  covered  by  the  pendulum.  For  example,  suppose 
we  twist  the  wire  360°  and  release  the  disk  so  that  rotation 
begins.  In  order  to  obtain  an  absolute  reading  to  start  from, 
which  shall  be  independent  of  any  slight  error  in  adjustment, 
we  ignore  the  fact  that  we  have  started  from  360',  and  take  as 
our  first  reading  the  end  of  the  first  swing.  Suppose  our  read- 
ings are  as  follows : 

Right,  350 ;  left,  336  ;  right,  328,  and  keeping  in  mind  the 
vibrations  ol  the  simple  pendulum  we  will  see  at  once  that  we 
have  read  two  complete  arcs  whose  difference  is  22°  computed  as 
follows  : 

1st  arc.  Right  350^  +  Uft  338'-"'  =  688^ 
rf  arc,  Left  338=  +  Right  328=  =  666° 

31°  retardatiou 
In  order  to  secure  freedom  from  error  we  take  two  tests — one 
by  rotating  the  wire  to  the  right,  and  the  second  to  the  left.  If 
the  instrument  is  in  exact  adjustment  these  two  results  will  be 
the  same,  but  if  it  is  slightly  out,  the  mean  of  the  two  readings 
will  be  the  correct  reading. 

It  will  also  be  noticed  that  if  the  exact  retardation  due  to  the 
oil  alone  is  to  be  obtained  we  must  subtract  the  factor  for  the 
resistance  due  to  the  air  and  the  wire  itself.  These  are  readily 
obtained  by  allowing  the  cylinder  to  rotate  in  the  air  and  deter- 
mining the  retardation  exactly  as  we  have  done  above.  This 
factor  remains  constant  for  each  instrument  and  is  simply  de- 
ducted from  all  results  obtained. 
I  Iodine  Absorption, 

The  determination  of  the  iodine  absorption  of  an  oil  is  prob- 

!#b]y  the  most  important  chemical  test  for  recognition  quanttta- 

'ely  in  a  mixture   of  animal  or  vegetable  with   mineral  oiJs. 

introduced   by    Hubl'  it   has    since    maintained    this    position, 

iziinr.^j'.y.,»ss,  jSi. 


though  Other  chemists  have  introduced  the  bromine  absorption 
and  others  of  similar  character.  They  have  not  been  adopted 
with  the  confidence  of  the  iodine  process. 

Warren*  draws  attention  to  the  fact  that  Chateau  in  his 
Essais  PersonnelUs^  p.  70,  used  the  iodine  absorption  in  a  manner 
similar  to  Hubl  many  years  previously. 

In  a  mixture  of  two  fatty  oils  with  a  mineral  oil,  the  best  re- 
sults are  obtained  by  saponifying  and  separating  the  fatty  acids 
from  the  mineral  oil.  The  iodine  absorption  of  the  mixed  fatty 
acids  is  then  taken,  and  where  the  nature  of  them  has  already 
been  shown  by  color  tests,  etc.,  their  proportion  can  be  indicated 
by  the  following  formula  : 

100  (/ — n) 

m — n 

Where  x  =  the  percentage  of  one  fat, 
y=    ''  **         **  the  other, 

/=  iodine  degree  of  mixture, 
w=      **  **      **fat;i:, 

n=      **  **      **  **  y. 

The  method  is  as  follows  :* 

Twenty-five  grams  of  iodine  and  thirty  grams  of  mercuric 
chloride  are  each  dissolved  in  500  cc.  of  ninety-five  per  cent, 
alcohol,  uniting  the  two  solutions,  and  allowing  to  stand  several 
hours  before  use. 

It  is  then  standardized  by  tenth  normal  thiosulphate  solution. 
The  process  of  the  determination  of  the  iodine  absorption  of  an 
oil  is  as  follows  :  One-tenth  to  five-tenths  gram  of  the  fat  or  oil 
is  dissolved  in  ten  cc.  of  purest  chloroform  in  a  well  stoppered 
flask,  and  twenty  cc.  of  the  iodine  solution  added.  The  amount 
must  be  finally  regulated  so  that  after  not  less  than  two  hours 
digestion  the  mixture  possesses  a  dark  brown  tint ;  under  any 
circumstances  it  is  necessary  to  have  a  considerable  excess  of 
iodine  (at  least  double  the  amount  absorbed  ought  to  be  present) , 
and  the  digestion  should  be  from  six  to  eight  hours.  Some 
potassium  iodide  solution  is  then  added,  and  the  whole  diluted 
with  150  cc.  of  water,  and  tenth  normal  thiosulphate  delivered 

1  Chrm,  A'ru's,  a6.  iSS. 

2  Sutton  :  Volumetric  Analysis.  343. 


in  till  the  color  is  nearly  discharged.  Starch  is  then  added,  and 
the  titration  finished  in  the  usual  way. 

If  more  than  two  fatty  oils  are  present  in  a  mixture  with 
mineral  oil,  the  method  of  Warren'  can  be  used. 

The  following  determinations  of  the  iodine  absorption  made  in 

my  laboratory  are  indicative  of  the  variations  of  the  absorption 

by  the  different  oils  : 

Prime  lard  oil 76.4  77.2 

No.  I      "     **    69.8  69.9 

XXX     *•     •*    65.1  65.6 

Oleo  oil 51.6  51.6 

Prime  neat's  foot  oil 80.1  82.0 

Horse  oil 82.3  82.5 

Natural  bow-head  whale  oil i30-5  131.  i 

**      winter            *'      "  121. i  126.0 

Extra  bleached  winter  white  oil  124.9  126.  i 

Bleached  spring      **            '*       126.1  126.2 

Crude  sperm  oil 82.3  82.3 

Prime  quality  winter  white  cotton-seed  oil  . .  114.2  ii4-9 

**       summer"         **          "      "  ..  110.2  110.6 

•*            **       winter  yellow  •'          "      ••  ..  115.9  118.6 

**      ,     **       summer    **       *'          "      "  ..  104.0  104.4 

Olive  oil 81.0  83.0 

Herring  oil 122. i  123.8 

Dog-fish  oil 102.7  104.7 

Porpoise  head  oil 28.9  29.1 

Rosin  oil,  second  run 92.1  93.4 

*•    third      "    90.4  92.2 

Flash  and  Fire  Test, 

The  flash  point  is  the  degree  oidemperature  at  which  ignitable 
volatile  vapors  are  given  off  by  the  oil,  producing  a  flash  when 
brought  in  contact  with  a  small  flame.  The  fire  test  is  a  contin- 
uation of  the  flash  test  until  the  oil  permanently  ignites.  A  sim- 
ple apparatus  that  gives  approximate  results  is  shown  in  Fig. 
129.  It  consists  of  a  porcelain  crucible  two  and  one-eighth  inches 
wide  at  the  top,  five-eighths  inch  wide  at  the  bottom  and  one  and 
one-half  inches  deep.  This  is  surrounded  by  an  asbestos  pad  three 
and  one-half  by  three  and  one-half  inches  and  one-eighth  inch 
thick.  This  prevents  the  direct  contact  of  the  gas  flame  upon 
any  portion  of  the  crucible   except  the  base.     The  oil  to  be 

I  Chem.  News,  6a,  2x5 :  /.  AnaL  Appl.  Chem.,  5.  215. 



tested  is  pUced  in  the  crucible,  a  thermometer  inserted  at  such 
a  depth  that  the  bulb  is  just  covered  by  the  oil,  and  the  heat 
applied.  The  rise  oi  temperature  in  the  oil  should  not  exceed 
2°  F.  per  minute. 

The  "  teit-flame"  (the  smallest  possible)  is  passed  over  and 
ncro9»thesurlaceof  the  oil  once  everj'  minute  beginning  at  ioo°P, 

Oils  that  fla&h  below   no"  F.  are  considered  unsafe  for  light- 

ing purposes,  and  for  lubricating  purposes;  oils  should  not  flash 
under  350*  F. 

The  Cleveland  cup  oil  tester  is  i-erj-  similar  to  this  instru- 
ment in  design  and  operation,  with  the  exception  that  the  porce- 
lain crucible  is  replaced  by  a  copper  one  of  (he  same  size  and 
heated  in  a  sand-bath  instead  of  being  surrounded  by  an 
asbestos  pad. 

Tagliahue's  open  tester,  has  a  ver>-  extensive  use  in  the  oil 
trade.      It  consists   (Fig.    130)   of  a  copper  cylinder,  B,  into 


which  fits  the  copper  water-bath,  .-/,  and  a  glass  cup,  D,  which 
contains  the  oil  to  be  tested.  This  apparatus  has  been  super- 
seded somewhat  by  another  form  of  open  tester.  The  "  Say- 
bolt"  which  is  used  by  the  chemists  of  the  Standard  Oil  Co., 
and  officially  adopted  by  the  N'ew  York  Produce  Exchange.  It 
consists  of  a  water-balh, 
Fy  (Fig.  131)  surround- 
ing an  inner  cup  con- 
taining the  oil.  An  in- 
duction coil,  E,  furnishes 
an  induction  spark  that 
passes  over  the  oil.  Bat- 
teries for  generating  the 
current  are  situated  un- 
der the  frame,  C. 

All  open  cup  testers 
give  higher  readings  for 
the  flash  test  than  closed 
testers  and  it  is  generally 
conceded  that  the  closed 
testers  admit  of  more  ac- 

,  curate  determinations. 
The  Abel  closed  tester 
Figs.  132,  133,  has  been 
adopted  by  the  English 
government,  and  in  a 
modified  form  (Pensky- 
Martens)  by  the  German 
government  as  the  official  instrument  for  this  purpose. 

The  specifications  for  this  instrument  require  that  the  oil  cup 
be  a  cylindrical  vessel,  two  inches  in  diameter,  two  and  two- 
tenths  high  (internal),  with  outward  projecting  rim  five-tenths 
Inch  wide,  three-eighths  inch  from  the  top.  and  one  and  seven- 
tieths inches  from  the  bottom  of  the  cup.     It  is  made  of  gun- 

r  I0U1  or  brass  (17  B,  W.  G.)  tinned  inside.      A  bracket,  con- 
il^  of  a  short  stout  piece  of  wire,  bent  upward,  and   termi- 

I  utti^tti  a  point,  is  fixed  to  the  inside  of  the  cup  to  serve  as  a 

|<£&tig*.     The  distance  of  the  point  from  the  bottom  of  the  cup  is 



one  and  a  half  inches.  The  cup  is  provided  with  a  close-fitting, 
overlapping  cover,  made  of  brass  (22  B.  W.  G.)  which  carries 
the  thermometer  and  test-lamp.  The  latter  is  suspended  from 
two  supports  from  the  side  by  means  of  trunnions,  upon  which 
it  may  be  made  to  oscillate  ;  it  is  provided  with  a  spout,  the 
mouth  of  which  is  one-sixteenth  of  an  inch  in  diameter.  The 
socket  which  is  to  hold  the  ther- 
mometer is  fixed  at  such  an 
angle,  and  its  length  is  so  ad- 
justed, that  the  bulb  of  the  ther- 


Fig.  131.  Flf.-i3J- ' 

mometer,  when  inserted  to  full  depth,  shall  be  one  and  a  half 
inches  below  the  center  of  the  lid.  The  coveris  provided  with 
three  square  holes,  one  in  the  center,  five-tenths  iocfa  by  four- 
tenths  inch,  and  two  smaller  ones,  three-tenths  inch  by  two- tenths 
inch,  close  to  the  sides  and  opposite  to  each  other.  These  three 
holes  may  be  closed  and  uncovered  by  means  of  a  slide  moving 
in  groves  and  bav-ing  perforations  corresponding  to  those  on  the 
lid.  In  moving  the  slide  so  as  to  uncover  the  holes,  the  oscilla- 
ting lamp  is  caught  by  a  pin  fixed  in  the  slide  and  tilted  in  such 
a  way  as  to  bring  the  end  of  the  spont  just  below  the  surface  of 


the  lid.     Upon  the  slide  being  pushed  back  so  as  to  cove: 
holes,  the  lamp  returns  to  its  original  position. 

The  flash  test  of  this  apparatus  is  about  27°  F.  lower  than  the 
open  cup  apparatus,  so  that  73°  F.  Abel  test  is  equivalent  to 
100°  F.  test,  open-cup  test. 

The  Pensky-Martens  closed  tester,  Figs.  134,  135,  in  action 

is  very  similar  to  the  Abel  closed  tester.  The  apparatus  of 
Treumann,  Figs.  136,  137  is  used  by  the  chemists  of  the  Prus- 
sian railways  for  the  determination  of  the  flash  and  fire  test  of 
both  illuminating  and  lubricating  oils. 

It  is  very  similar  in  construction  and  operation  to  the  Cleve- 
land cup,  in  use  in  this  counlrj'  for  the  same  purpose,  with  the  ex- 
ception that  the  oil  is  placed  in  a  porcelain  crucible,  a.  Fig.  137, 
instead  of  a  copper  one  as  in  the  Cleveland  cup. 


The  larger  containing  vessel  is  of  iron  and  contains  sufficient 
sand  to  raise  the  bottom  of  the  crucible  containing  the  oil,  one- 
half  inch  from  the  point  of  contact  of  the  flame. 

The  flash  and  fire  tests  are  required  of  all  lubricating  oils  as 
a  test  of  their  power  to  resist  combustion  by  overheating  in 
work.  Valve  oils  with  mineral  stock  are  especially  liable  to 
have  low  flash  points  caused  by 
imperfect  distillation  in  their 
manufacture.     They  should  be 

free  from  any  ol  the  lighter  oils  (naphtha,  kerosene,  etc..)  and 
should  not  flash  under  300°  F.  For  cylinder  oils  the  require- 
ment is  much  higher.  Animal  and  vegetable  oils  used  in  lubri- 
cation rarely  flash  under  400°  F. 


Acidity  in  oils  is  generally  due  to  a  partial  decomposition  of 
the  oil  with  liberation  of  fatty  acids.  These  latter  act  as  cor- 
rosive agents,  attacking  the  metal  bearings  of  machinerj',  form- 
ing "metallic  soaps"  and  producing  gumming  and  thickening 
of  the  lubricant. 

Properly  refined  mineral  oils  are  free  from  acidity, 
all  animal  and  vegetables  oils  possess  it  more  or  less. 

ty,  but  nearlir  ^| 



In  palm  oil,  lor  instance,  the  free  fatty  acids  vary  from 
twelve  to  eighty  per  cent.  In  eighty-nine  samples  of  olive  oil 
intended  for  lubricating  purposes,  D.  Archbutt'  found  from  2.2  to 
25.1  percent,  of  free  acid  (oleic)  the  mean  being  S.05  percent. 

Oleic  acid  cannot  be  present  as  a  constituent  of  a  pure  mineral 
oil :  still  the  acid  test  should  be  made,  since  poorly  refined 
mineral  oils  are  liable  to  contain  small  amounts  of  sulphuric 
acid  left  in  the  process  of  refining.  The  sulphuric  acid  is  easily 
indicated  by  warming  some  of  the  oil  with  distilled  water,  adding 
a  few  drops  of  hydrochloric  acid  (dilute)  and  solution  of  barium 
chloride.  A  white  cloud  or  precipitate  shows  the  presence  of 
sulphuric  acid. 

The  action  of  free  acid  on  journals,  bearing,  etc.,  as  a  corro- 
sive element,  has  led  many  engineers  to  include  a  test  of  free  acid 
direct  upon  copper  and  iron. 

This  is  done  by  suspending  weighed  pieces  of  sheet  copper 
and  iron  iu  the  different  oils,  for  a  number  of  days,  heating  if 
necessary,  and  determining  the  amount  of  metal  dissolved  by  the 

While  this  test  may  be  indicative  of  the  acidity  of  oils,  no  ratio 
exists  between  the  action  upon  copper  and  iron  or  even  between 
the  oils  themselves  in  this  respect,  owing  to  the  varying  quantity 
of  acid  in  the  same  oils. 

The  results  of  a  few  tests  are  shown  in  the  following  table: 

Kline  of  <^1.  icdny*.  Mdiy*. 

Linseed  oil  ■ .  -' 0.3000  grain.  o.oojogram. 

Oliveoil  o.ajoo     '■  0.0063  " 

Neat'*  fool  oil 0.1100     "  0.0875  " 

Sperm  oil 0.0030     "  0.0460  ■' 

ParaSa  oil 0.0015     "  0,0043  " 

Lard  oil 0.0J50  " 

The  following  is  the  method  for  determining  the  acidity  of  oils, 
18  used  in  many  of  the  railroad  laboratories  : 

Afaieriah  Required. 
One  fifty  cc.  burette,  graduated  to  tenths. 
Two  ounces  alcoholic  solution  phenolphthalein. 


Three  ten  cc.  pipettes. 

One  druggist's  graduate,  four  ounces. 

One  gallon  ninety-five  per  cent,  alcohol. 

One  dozen  four  ounce  sample  bottles. 

One  thermometer  graduated  from  15*  to  ^15**  F.,  and  bearing 
the  certificate  of  the  Yale  Thermometer  Bureau. 

Two  hydrometers  15"*  to  25**  and  25**  to  35"*  B.,  each  degree 
graduated  to  tenths  (Tagliabue*s.) 

One  hydrometer  jar. 

One  quart  caustic  potash  solution  of  such  strength  that  31.5 
cc.  exactly  neutralize  five  cc.  of  a  normal  solution  of  sulphuric 
acid  (contains  forty-nine  mgms.  per  cubic  centimeter  of  sulphuric 

Take  two  ounces  of  alcohol  and  warm  tq  about  100**  F.;  add 
ten  drops  of  alcoholic  solution  of  phenolphthalein.  Fill  the 
burette  to  the  top  of  the  graduation  with  the  caustic  potash  solu- 
tion ;  then  add  solution  drop  by  drop  to  the  alcohol  until  it  as- 
sumes a  pink  tint.  Add  ten  cc.  of  the  oil  to  the  alcohol,  refill 
the  burette  with  the  potash  solution  and  add  the  latter  until  the 
mixture  of  oil  and  alcohol  maintains  a  pink  color  after  thorough 
shaking.  Read  off  the  number  of  cc.  of  potash  solution  used, 
and  this  amount  divided  by  two,  gives  the  per  cent,  of  free  acid. 
For  example,  if  10.6  cc.  caustic  potash  solution  have  been  used, 
the  oil  contains  five  and  three-tenths  per  cent,  of  free  fatty  acid. 

Lard  and  tallow  are  very  liable  to  have  considerable  amounts 
of  free  acid.  The  specification  of  purchase  therefore  generally 
states  the  limits  of  free  acid  permitted. 

Maumem's  Test, 

The  rise  of  temperature  produced  when  sulphuric  acid  is 
brought  in  contact  with  certain  oils  was  first  investigated  by 
Maumene,  and  the  results  of  his  experiments  published  in 
Compfes  Rendiis,  35,  572. 

The  subject  has  been  investigated  by  Fehling,  Faist,  L.  Arch- 
butt,  C.  J.  Ellis,  A.  H.  Allen  and  others,  with  the  result  that 

I  Hydrometers  and  thermometers  should  be  procured  through  Chas.  A.  Tagliabue, 
New  York. 


this  test  has  been  generally  accepted  as  of  importance  in  the 
distinction  of  oils  in  mixtures. 

When  a  mixture  of  oils  has  been  analyzed  and  the  components 
recognized  the  proportions  oftentimes  can  be  determined  by 
this  reaction  ;  that  is  to  say,  suppose  the  oil  under  examination 
to  show  a  rise  of  temperature  of  80''  C,  and  the  oils  found  by 
analysis  to  be  lard  oil  and  menhaden  oil ;  their  relative  propor- 
tions can  be  determined  by  the  following  formula  : 

W^  =  proportion  by  weight  of  menhaden  oil. 

W^=  **  **         **      *»  lard 

W^  =  weight  of  mixture. 

/,  =  temperature  of  menhaden  oil. 

/.  =  *•  **  lard 

/,  =  **  **  mixture. 

The  method  is  as  follows : 

Fifty  grams  of  the  oil  are  placed  in  a  narrow  tall  beaker  and 
ten  cc.  of  C.  P.  sulphuric  acid  added  drop  by  drop  with  stirring 
and  the  rise  of  temperature  during  the  operation  noted. 

Lard  oil  alone  when  treated  with  sulphuric  acid  g^ves  a  rise 
of  temperature  of  40°  C;  menhaden  oil,  under  similar  conditions, 
a  rise  of  128®  C.  Using  these  values  in  the  above  formula  we 
obtain  54.6  per  cent,  lard  oil  and  45.4  per" cent,  menhaden  oil. 

In  the  mixture  containing  a  mineral  oil  mixed  with  animal, 
marine  or  vegetable  oil  the  distinction  would  be  even  more  pro- 
nounced, since  the  mineral  oil  shows  but  a  very  slight  increase 
of  temperature  (generally  from  2**C.  to  5"*  C).  The  increment  of 
temperature  would  be  dependent  upon  the  Other  oil  added  to  the 
mineral  oil. 

Briefly  stated,  the  rise  of  temperature  of  the  following  oils 
would  be : 



Lard  oil 

Tallow  oil 

Neat's  foot  oil 

Oleo  oil 

Elain  oil 

Sperm  oil 

Whale  oil 

Menhaden  oil 

Dog-fish  oil 

Cod  liver  oil 

Crude  cotton-seed  oil . . 

Rape  oil 

Castor  oil 

Olive  oil 

Rosin  oil 

Mineral  lubricating  oil 

Earth  nut 

Rosin  oil 

Sea  elephant 

Name  of  Observer. 


















9^   0 






























Attention  is  drawn  to  the  differences  in  the  determinations  in 
rosin  oil. 

Rosin  oil  of  the  first  run  is  a  white,  opaque,  thick  liquid  con- 
taining all  of  the  water  of  the  rosin  from  which  it  is  distilled, 
and  it  is  this  water  that  causes  the  rise  of  temperature  above 
10°  when  the  oil  is  mixed  with  the  sulphuric  acid. 

Rosin  oils  of  the  second  and  third  runs  are  clear,  limpid,  dark 
red  colored  fluids,  practically  free  from  water,  and  when  treated 
with  sulphuric  acid  do  not  indicate  more  than  10^  rise  of  tempera- 

From  these  tests  it  is  concluded  that  both  Schaedler  and  Allen 
tested  rosin  oil  that  was  a  mixture  of  the  first  and  second  runs, 
or  of  an  oil  not  properly  separated  into  the  different  distillates. 

Color  /Reactions  of  Oils  ivith  Nitric  and  Sulphuric  Acid. 

Of  the  many  color  tests  introduced  for  the  identification  of 
simple  oils,  preference  is  given  to  Heidenreich's  sulphuric  acid 
test  and  Massie's  nitric  acid  test. 

The  color  reactions  of  Chateau*  in  which  barium  poly-sulphide 

1  Spon's  BncycIoi>e<lia,  4,  1473-1475. 


zinc  chloride,  stannic  chloride,  phosphoric  acid  and  mercuric 
nitrate,  in  solutions,  are  used,  while  very  interesting,  seldom  are 
of  any  advantage  over  the  two  tests  noted  above.  Glassner's' 
nitric  acid  reactions  are  practically  the  same  in  results  as 
Masste's  so  that  no  advantage  would  be  obtained  in  including 
the  former. 

Heidenreich's  test  is  as  follows  : 

A  clear  glass  plate  is  placed  over  a  piece  ol  white  paper,  ten 
drops  of  the  oil  under  examination  are  placed  thereon,  and  one 
drop  of  concentrated  sulphuric  acid  is  added. 

The  color  produced  when  the  acid  comes  in  contact  with  the 
oil  is  noticed  as  well  as  the  color  produced  when  the  two  are 
stirred  with  a  glass  rod.  Many  oils  give  off  characteristic  odors 
during  the  reaction,  especially  neat's  foot  oil,  whale  oil  and 
menhaden  oil. 

Massie's  test  is  thus  performed  : 

Nitric  acid  of  specific  gravity  1.40,  free  from  nitrous  acid,  is 
mixed  in  a  test  tube  with  one-third  its  volume  of  the  oil,  and  the 
whole  agitated  for  two  minutes. 

The  color  of  the  oil  after  separation  from  the  acid  is  the  indica- 

In  mixtures  of  oils,  the  characteristic  colors  produced,  by 
either  Heidenreich's  or  Massie's  test,  are  often  clouded,  and  in 
many  instances  no  inferences  can  be  drawn,  yet  with  single  oils 
the  reactions  are  often  distinctive  and  sufficiently  strong  to  give 
confirmatory  results. 

In  cod  liver  oil,  or  whale  oil,  when  mixed  with  a  mineral  or 
even  vegetable  oil,  the  characteristic  brilliant  violent  color  pro- 
duced with  sulphuric  acid  cannot  be  mistaken.  This  color,  due 
to  the  presence  of  cholic  acid,  is  found  in  most  of  the  fish  oils, 
but  is  much  more  pronounced  in  cod  liver  oil. 

The  following  table  will  indicate  the  colors  produced  by  Hei- 

sreich's  and  Massie's  test. 

p.  Crmtrtl..  /S73.  $j. 



Heidenreich'i  test. 

Musie'i  test 

Before  stirring.             After  atirTiiig. 

T^&rrl  oil  ••••  •••••• 




Tallow  oil 




Neat's  foot  oil.  ••• 


Red  brown. 


Oleo  oil 




Elftttl     Oil  ••••••  ••• 

Light  g^een  (turn- 
ing to  brown). 

Orange  red. 


Sperm  oil 

Brown  with    pur- 
ple streaks. 
Rea  violet. 

Reddish  brown. 


Whale  oil 

Violet  brown. 


Menhaden  oil>«-« 



Dark  red. 

Doe-fish  oil 

Cod  liver  oil 


Dark  brown. 


Red  violet. 

Dark  brown. 

Orange  red. 

Crude  cotton-seed 

Brilliant  red. 



Ref 'd  cotton-seed . 

Reddish  brown. 


Orange  red. 

1ifl.n^  oil  ••>■••  .••• 

Yellow  brown  - 


Yellow  to 

OAfttor  oil ......  ... 

Lgt.  yel.  to  brown. 
Light  green. 

"Pale  Hromm 

Oliv^  oil....  ...... 

■Greenish  to  light 


RoRin   oil 




Earth  nut  oil 

Yellow  to  orange. 


The  oils  made  use  of  in  lubrication  can  be  separated  into  two 
groups:  saponifiable  and  unsaponifiable.  To  the  former  belong 
all  the  fatty  oils  ;  to  the  latter  the  mineral  and  rosin  oils. 

The  method  of  Lux*  is  made  use  of  to  determine  if  any  fatty 
oils  are  present  in  a  mineral  oil. 

If  rosin  otl  is  suspected  to  have  been  added  to  the  mineral,  it 
can  be  identified  by  the  method  of  Holde*  or  the  process  of  E. 
Valenta*  can  be  used. 

These  three  tests  will  indicate,  qualitatively,  the  presence  of 
any  fatty  or  rosin  oil  in  a  mineral  oil.  It  is  rarely,  in  the  bet- 
ter class  of  lubricating  oils,  that  more  than  one  oil  is  added  to  a 
mineral  oil,  such,  for  instance,  as  lard  oil,  or  tallow,  in  which 
case  saponification  easily  separates  the  two  oils,  and  identifica- 
tion of  each  by  special  tests  can  then  be  made. 

When,  however,  the  oil  added  to  the  mineral  oil  itself  contains 
an  adulterant,  such  as  lard  oil  to  which  cotton-seed  oil  has  been 
added,  then  the  fatty  acids  separated  by  saponification  will  re- 
quire a  more  extended  examination  to  prove  the  presence  of  both 
lard  oil  and  cotton-seed  oil. 

I  Zischr.  Anal.  Chem.,  34,  347. 

«  Mittheil  der  Konig.  tech.  Versuchsanstaltcn,  /890,  19. 

*  ZUehr.  anal.  Chtm.,  35,  441. 



The  following  skeleton  scheme  is  given  to  show  the  applica- 
tion of  the  above  upon  a  lubricating  oil  that  qualitative  analysis 
has  shown  to  contain  mineral  oil,  lard  oil,  and  cotton-seed  oil. 

Twenty  grams  of  the  oil  are  weighed  out  in  a  No.  3  beaker,  100  cc.  of  an 
alcoholic  solution  of  potash  (eighty  grams  potassium  hydroxide  to  one 
liter  alcohol  of  ninety-eight  per  cent. )  are  added,  and  heat  applied  with 
stirring  until  the  alcohol  is  all  driven  off ;  add  100  cc.  water,  heat  with 
agitation,  cool,  add  fifty  cc.  etheri  transfer  to  separatory  funnel,  stopper, 
shake  well  and  allow  to  stand  two  hours.     Draw  off  the  soap  solution. 

X.  Soap  solution  (containing  the 
fatty  acids  of  the  lard  and  cotton- 
seed oils).  Heat  ten  minutes 
nearly  to  boiling,  cool,  acidify 
with  dilute  sulphuric  acid,  allow 
to  stand  a  few  hours ;  collect  the 
separated  fatty  acids;  deter- 
mine their  weight,  then  test  as 
follows : 

First  portion :  Determine  the  "melt- 

Second  portion :  Determine  the 
'* iodine  absorption*'  and  their 
rates  by  formula : 

;r=ioo  (/— «) 
x=itn — «. 

a.  Ether  solution  remaining  in  the 
separatory  funnel  is  transferred 
to  a  flask,  the  ether  distilled  and 
the  mineral  oils  weighed. 

There  are  several  methods  for  the  quantitative  determination 
of  the  amounts  of  vegetable  and  animal  oils  when  mixed  with 
each  other,  or  when  the  mixture  is  incorporated  with  a  mineral 
oil.  The  determination  of  the  iodine  absorption  is  the  most 
delicate  and  correct  provided  no  fish  blubber  or  olive  oils  are 

If  the  fatty  acids  have  been  separated,  by  saponification,  from 
a  mineral  oil,  this  iodine  value  can  also  be  determined.  Consult 
soap  analysis,  for  table  of  constants. 

The  method  of  Salkowski^  depends  upon  the  fact  that  vegetable 
oils  (except  olive)  contain  phytosterol  and  that  animal  fats 
(butter  excepted)  are  free  from  it,  containing  cholesterol,  the 
latter  not  being  present  in  vegetable  oils. 

1  Benedikt :  Oils,  Pats  and  Waxes,  355. 


Fifty  grams  of  the  sample  free  from  mineral  oil  are  saponified 
with  alcoholic  potash  ;  the  soap  solution  is  diluted  with  a  liter 
of  water  and  exhausted  with  ether.  When  the  two  layers  have 
separated,  the  aqueous  layer  is  run  off  and  the  ethereal  liquid 
filtered  and  evaporated  to  a  small  bulk.  To  insure  complete 
absence  of  unsaponified  fat,  it  is  best  to  saponify  again  with 
alcoholic  potash  and  to  repeat  the  exhaustion  with  ether.  The 
ethereal  layer  is  then  washed  with  water  and  the  ether  evapo- 
rated in  a  deep  basin.  The  residue  is  next  dissolved  in  hot 
alcohol,  the  solution  boiled  down  to  one  or  two  cc.  and  the 
residue  allowed  to  cool.  If  phytosterol  or  cholesterol  be  present, 
crystals  will  separate  out.  They  are  dried  on  unglazed  porce- 
lain and  their  melting  points  determined. 

The  saponification  value  of  oils  is  often  made  use  of  for 
identification  :  but  as  this  value  varies  with  the  age  of  the  oil, 
it  is  extremely  difficult  to  obtain  concordant  results,  and  as  the 
majority  of  oils  have  a  saponification  value  of  193,  excepting 
rape-seed  oil  and  castor  oil  which  are  lower,  it  can  not  be 
relied  upon.  It  however  is  of  value  in  determining  the  amount 
of  liquid  waxes  in  the  presence  of  oils. 

Wool-grease  is  used  to  some  extent  in  the  cheaper  grades  of 
lubricants,  the  consumption  for  this  purpose  increasing  yearly. 
It  is  unsaponifiable  and,  if  present,  will  be  found  in  the  ether  ex- 
tract with  the  mineral  oil,  in  the  analysis  as  usually  conducted 
of  a  mixed  lubricating  oil. 

Degras  or  sod  oil  is  a  waste  product  obtained  in  the  chamois- 
ing process.  It  is  largely  derived  from  whale  oil  or  poor  quality 
of  cod  liver  oil  used  in  chamoising. 

The  English-German  method  of  treating  skins  produces  sod 
oil  as  a  waste  product.  The  French  method  produces  De- 
gras. These  fats  are  largely  used  in  the  production  of  cheaper 

Consult  Benedikt:  Oils,  Fats  and  Waxes,  589;  J,  Am,  Chetn,  Soc,, 
(Bush),  x6,  535. 

Bone  Fat  is  made  use  of  in  lubrication  mixed  with  mineral 
oils.  It  is  recovered  from  bones,  either  by  boiling  with  water 
or  extracting  with  solvents.  It  does  not  readily  become  nmcid. 
Its  examination  is  made  similarly  to  that  of  tallow. 


Coejffiden  I  of  Friction . 

The  ratio  of  the  force  required  to  shde  a  body  along  a  hori- 
zontal plane  surface  to  the  weight  of  the  body  is  called  the  coeffi- 
cient of  friction.  It  is  equivalent  to  the  tangent  of  the  angle  of 
repose,  which  is  the  angle  of  inclination  to  the  horizontal  of  an 
inclined  plane  on  which  the  body  will  just  overcome  its  tendency 
to  slide.  The  angle  is  usually  denoted  by  qt,  and  the  coefficient 

/=tan  <p.     (Kent.) 

Of  the  various  machines  used  for  this  purpose  nearly  all  are 
deficient  in  conducting  tests  under  extreme  pressure.  However 
as  all  the  tests  are  relative,  an  idea  of  the  value  of  a  lubricant 
can  be  formed  by  a  series  of  comparative  tests  upon  the  same 

G.  B.  Heckel  thus  describes  the  Thurston  and  Henderson- 
Westhoven  machines:  The  primary  idea  of  determining  dura- 

bility is  to  determine  how  much  rubbing  a  lubricant  will  with- 
stand before  exhaustion  of  its  power  to  maintain  the  friction  at 
some  agreed  minimum.  For  this  there  is  no  device  superior  to  the 
Thurston  oil-tester,  in  which  a  pair  of  brasses  are  forced  against 
a  journal  in  opposite  directions  by  a  spring  being  lodged  in  a 
lendulum  which  is  free  to  swing  about  the  journal,  the  friction 
eing  measured  by  the  inclination  to  the  vertical  of  a  line  join- 


ing  the  center  of  the  journal  and  the  center  of  gravity  of  the 
pendulum.  The  defects  of  this  machine  lie  in  the  infinitely 
variable  rate  of  metallic  wear  between  rubbing  surfaces,  which 
contaminates  the  oil  before  it  has  been  exhausted,  as  well  as  in 
the  escape  of  the  lubricant  between  the  surfaces. 

These  imperfections  have  been  overcome  in  the  Henderson 
machine  or  the  so-called  Henderson- Westhoven  machine,  a 
modified  Thurston  tester.     (Figs.  138,  139.) 

With  this  machine  lubricants  can  be  tested  at  the  same 
moment  for  the  degree  of  heat  developed  in  the  bearing  surfaces 
as  well  as  their  friction  reducing  qualities. 

The  journal,  A,  rests  upon  the  supporting  beds,  BB^  and  is 
revolved  by  the  puUy,  C,  This  journal,  A,  extends  on  both 
sides  beyond  the  supports,  BB,  and  the  projecting  ends  are  em- 
braced by  brass  boxes  DD,  to  which  are  fastened  the  pendulum 
parts  EE.  Strong  spiral  springs  mm^  in  the  interior  of  the 
pendulum  arms,  force  the  lower  pair  of  brasses,  DD,  against  the 
journal,  A,  and  the  pressure  of  these  springs  may  be  regulated 
by  means  of  the  screw,  N,  A  pointer  attached  to  the  movable 
block,  0,  indicates  on  the  scale,  Z',  as  in  a  spring  balance,  the 
thrust  of  the  spring  against  its  bed,  in  kilograms  per  cubic  centi- 
meter. By  the  revolution  of  the  journal.  A,  the  swinging  arms, 
EE,  are  actuated  by  friction  in  the  direction  of  the  motion,  and 
the  degree  of  their  deviation  from  the  vertical  is  read  by  means 
of  the  pointers,  FF,  on  the  quadrants  GG,  On  many  machines 
the  scales  give,  besides  the  deviation,  also  the  coefi&cient  of  fric- 
tion which  has  been  calculated  from  the  former. 

In  the  upper  brasses,  DD,  a.  thermometer,  //,  is  fixed  to  show 
the  degree  of  heat  developed  by  the  friction,  and  the  revolution 
counter,  /,  actuated  through  the  endless  screw,  g,  records  the 
revolutions  of  the  journal,  A.  The  column.  A',  through  its  two 
arms,  L,  carrying  the  boxes,  BB,  serves  to  support  the  entire 

In  operation  the  oil  to  be  tested  is  introduced  by  means  of  a 
small  glass  tube  or  pipette,  through  an  orifice  in  the  upper 
brasses,  DD,  the  journal  having  been  thoroughly  cleaned.  The 
position  of  the  thermometer  and  of  the  revolution  counter  are 
noted,  and  the  journal  is  then  put  into  motion  with  200  or  300 


'i  revolutions  per  minute.  At  each  succeeding  five-hundredth  or 
thousandth  revolution  the  temperature  and  the  degree  of  devia- 
tion of  the  pendulum  arms,  as  shown  by  the  quadrant,  are  noted, 
and  when  the  friction  has  raised  the  temperature  in  the  boxes 
about  30"  (usually  in  about  half  an  hour)  the  machine  is  stopped. 
In  figuring  up  results,  the  sample  of  oil  which,  with  an  equal 
rise  in  temperature  at  the  point  of  friction,  gives  the  slightest 
deviation  of  the  swinging  pendulum  arm,  and  the  greatest  num- 

ber of  revolutions,  is  regarded  as  the  best  The  advantages 
noted  in  this  device  are  its  facilities  for  testing  materials  under 
any  pressure,  even  up  to  the  load  limit  on  a  freight  car  axle  ;  the 
number  of  data  obtainable  at  one  time  ;  and  the  ease  with  which 
two  simultaneous  tests  of  competing  oils  can  be  made  on  the 
one  machine. 

The  apparatus  used  for  testing  lubricants  by  the  officials  of 
the  Paris-Lyon  Railway  is  shown  in  Figs.  140,  141.  Here  the 
conditions  are  maintained  as  nearly  as  possible  as  would  occur 

cur  ■ 


in  railroad  practice,  the  friction  being  determined  by  means  of 
two  (reight-car  wheels, 

The  heavy  cast-iron  frame,  .-J,  stands  upon  a  firm  stoire  founda- 
tion and  carries  the  shaft,  B,  on  which  are  fastened  the  two  fric- 
^tion  wheels,  CC.  These  are  placed  at  the  same  gauge  as  the 
railroad  track.  Two  ordinary  car  wheels,  DD,  with  axle, 
are  placed  above  and  in  contact  as  shown  in  the  figure.  The 
car  axle,  E,  is  fitted  at  each  end  into  the  axle  boxes,  mm.  The 
boxes  have  the  same  arrangement  as  those  in  the  railroad  cai» 

and  serve  for  the  reception  of  the  lubricant.  Resting  on  eadl 
side  of  the  axle  boxes  are  the  strong  springs,  ««,  Fig.  140,  oA 
the  end  of  which  the  weights,  FF.  work  by  means  of  the  levers^ 
OQ.  By  taking  off  or  putting  on  of  weights,  FF,  E  can  cany 
any  load  desirable. 

On  the  lower  shaft  is  the  driving  wheel,  G,  also  a  screw  by 
which  the  movement  of  the  shaft  is  earned  to  a  figured  dial. 
This  dial  sets  not  only  the  index  showing  the  number  of  revoln- 


tions  but  also  the  index  needle,  /,  in  motion  which  indicates  on 
the  scale,  a,  the  approximate  rapidly  of  the  wheel-rims  in  kilo- 

meters per  hour.  The  two  friction  wheels,  cc,  are  turned  eccen- 
trically about  two  and  five-tenths  mm,  that  by  the  motion  a  weak 
■vertical  oscillation  arises  which  is  communicated  to  the  upper 


wheels  whereby  the  rattling  of  the  wheels  upon  the  car  track  is 

In  making  a  trial,  the  lubricant  to  be  tested  is  placed  in  the 
thoroughly  cleaned  axle  boxes,  mtn^  the  springs  are  lifted  to  the 
utmost  release  of  the  upper  shaft  and  the  lower  shaft  is  placed 
in  rotation.  Hot  until  the  whole  is  in  motion  are  the  springs 
brought  down,  and  later  loaded  with  the  intended  weight.  The 
oil  which  by  this  test  carries  the  burden  with  the  greatest 
rapidity  without  heating  of  the  axle-boxes  is  to  be  considered 
the  best.  By  this  apparatus  it  is  possible  to  judge  of  the 
practical  working  of  an  oil  or  compounded  oil,  and  especially  if 
the  car  axles  would  become  heated,  a  point  of  vital  importance 
as  regards  the  use  of  the  lubricant. 

Another  instrument  of  a  similar  design  is  the  Riehl6,  (Fig. 
142,)  in  use  in  many  railroad  laboratories  in  the  United  States, 
for  testing  lubricants.  The  capacity  is  20,000  pounds  ;  it  deter- 
mines the  coefficient  of  friction,  the  pressure  per  square  inch  of 
journal  and  records  the  temperature. 

It  consists  of  a  Master  Car  Builder's  Axle  journal,  which  is  re- 
movable from  the  main  spindle.  This  journal  is  made  to  revolve  by 
cone  pully  at  different  speeds,  and  in  either  direction,  and  can 
be  loaded  to  different  pressures  per  inch  by  means  of  the  lever 
system.  The  oil  can  be  supplied  through  a  hole  in  the  top, 
which  is  tapped  to  receive  a  sight-feed  oiler,  or  funnel,  or  other 

The  friction  is  weighed  on  the  beams,  which  are  arranged  in 
double  system  to  balance  each  other,  allowing  the  machine  to 
be  run  in  either  direction.  The  opening  in  the  frame  over  the 
journal  is  made  large  enough  to  take  a  regular  car  box  if  desired. 

The  frame  and  beams  can  be  raised  bj'  rope  sling  and  hoist 
for  change  of  journal,  cleaning  up,  etc. 

There  is  an  end  motion  of  about  one-fourth  to  three-eighths 
inch  given  to  the  axle  by  the  gearing  shown  at  the  end,  giving 
a  natural  movement  like  cars.  The  weighed  end  of  spindle  runs 
loose  on  large  rollers,  to  avoid  friction  and  heating. 

An  oil  tested  upon  the  tester  may  show  a  fine  lubricant,  while 
put  under  practical  working  upon  a  freight  car  (for  instance) 
would  prove  vastly  inferior.     This  ver>-  often  happens,  and  it 



has  led  many  engineers  to  test  each  oil  by  a  long  run,  with  the 
particular  kind  of  machinery  upon  which  it  is  to  be  used. 

A  record-blank  used  by  the  engineers  of  the  Baltimore  and 
Ohio  Railroad,  for  testing  oils  upon  their  locomotives  is  given 
herewith.  It  is  a  point  in  instance.  After  experimenting  months 
upon  an  oU  its  work  is  established  so  that  a  practical  compari- 
son can  be  made  with  other  brands  of  similar  composition  for 
the  same  purpose. 

Baltimore  and  Ohio  Railroad. 

Engineer  of  Testt, 
DSAK  Sir  :  Below  pleaae  find  report  from  locomotives  inspected  this  day. 













•s  . 
















(Specifications  for  Compound  Oils.) 

DBTAIL  specifications. 

Engine  and  Passenger  Car  Oil. 
This  oil  mnat  conform  to  the  following  requirements  . 
I.  It  mnst  have  a  Sashing  point  from  October  i  to  May  I,  above  200*^  F,  ; 
from  May  i  to  October  i  the  flashing  point  must  be  above  150°  F. 
3.  From  October  i  to  May  i  it  must  have  a  cold  test  below  15°  F. 

3.  It  must  show  no  sediment  in  fifteen  minutes  when  five  cc.  are  mixed 
with  100  cc.  of  gasoline  of  %^'  B. 

4.  It  mast  contain  not  less  than  thirty  per  cent,  saponifiable  animal  oil. 

5.  Its  gravity  must  be  between  ifP  aud  30*^  B. 



Cylinder  OiL 

This  oil  must  conform  to  the  following  requirements : 

1.  It  must  have  a  flashing  point  above  440°  F. 

2.  It  must  contain  not  less  than  thirty-five  per  cent,  of  saponifiable  ani- 
mal oil. 

3.  It  must  show  not  more  than  six  per  cent,  of  fat  acid  or  its  equivalent. 

4.  It  must  not  show  any  precipitation  when  five  cc.  are  mixed  with  100 
cc.  of  gasoline  of  85^  B. 

Freight  Car  Oil. 
This  oil  must  conform  to  the  following  requirements : 

1.  It  must  have  a  flashing  point  from  October  i  to  May  i  above  200^  F. ; 
from  May  i  to  October  i  the  flashing  point  must  be  above  250^  F. 

2.  From  October  i  to  May  i  it  must  have  a  cold  test  below  15^  F. 

3.  It  must  show  no  sediment  in  fifteen  minutes  when  five  cc.  are  mixed 
with  100  cc.  of  gasoline  of  85°  B. 

4.  It  must  not  contain  less  than  ten  per  cent,  of  saponifiable  animal  oil. 

Special  Mixture. 
•  All  special  mixtures  of  oil  not  coming  under  the  above  specifications 
will  be  purchased  on  sample,  which  must  be  of  one  gallon.     Shipments 
will  be  required  to  conform  to  sample  in  every  particular.    Samples  must 
be  sent  as  the  purchasing  agent  may  direct. 

Chicago,  Borungton  and  Quincy  Railroad  Co.    Chem- 
ical Laboratory. 

Aurora,  III., 18.  • 

To Su  pt .  M .  P. : 

Dear  Sir  :  I  have  made  an  examination  of  sample  of  above  oil,  and 
have  obtained  the  following  results  : 

Flashing  point "^F.      Ash %.      Tar %. 

Burning      "       °F.      Cold  test at °F. 

Specific  gravity °B.      Viscosity  at °F.  100  cc.  oil 

Loss  at ^F.  for  3  hours %.  flows  from  instrument  in .  -seconds. 

Friction  Test  on  the  Thurston  Oil  Tester. 


Amount  used oz. 

Highest  reading. 
Range  of      " 

Time  run  in  minutes 

Total  revolutions 

Revolutions  per  minute 

Speed,  miles  per  hour 

Pressure,  total  lbs 

"  lbs.  per  sq.  inch 

Coefficient  of  friction 

Lubricating  value,  with  Extra  Lard 

Oil  as  too 

ist  trial. 



3d  trial. 








Received 18. .     Car  No.  and  Initials 

Tested 18. .     Tank  or  No.  Bbls 

Sample  No.  or  Letter Name  of  firm  supplying 

Blank  No Price ....  cents  per  gallon. 

Letter  Book  No Page 

Yours  truly, 


For  the R.  R.  Co. 

Chicago,  Burlington  and  Quincy  Railroad  Company. 

Specifications  for  Black  Engine  Oils, 

('*  Petroleum  lubricating  oils;"    **well  oils;"  *' petroleum  stock  oils;" 
or  **  passenger  and  freight  car  lubricating  oils.") 

Uses. — For  lubricating  the  journals  of  passenger  and  freight  cars  and 
locomotives,  and  for  miscellaneous  lubrication. 

Grades. — :  **  Summer,"  *'  25  degree,"  "  15  degree"  and  *'  zero." 

Requirements. — For  all  grades  : 

Specific  gravity,  between  26*^  and  29°  B. 

Loss  at  100^  F.  for  three  hours,  not  over  one-fourth  per  cent. 

Flashing  point,  for  all  but  *'zero"  oil,  not  under  300^  F. 

Flashing  point,  for  "zero"  oil,  not  under  250^  F. 

Burning  point,  for  all  but  *'zero"  oil,  not  under  375*^  F. 

Burning  point,  for  *'zero"  oil,  not  under  300^  F. 

Cold  Test — Summer  oil  must  flow  at  60°  F.  or  above. 
««         __      2^P         "  *'  "  xcP  '*  '* 

I «         ^_       I  c^         '  *  "  '  *  20*^  *  *  '  * 

All  these  oils  must  be  pure  petroleum  oils,  free  from  other  compounds, 
and  from  dirt,  grit,  lumps  and  specks  ;  transparent  and  greenish  or  red- 
dish (not  black)  in  tint^  when  spread  as  a  thin  film  on  glass  and  looked 
through  toward  the  light  ;  translucent  and  greenish  when  held  in  a  hori- 
zontal position.  Preference  will  be  given  to  those  oils  which  are  low  in 
tarry  matters  and  in  ash,  and  which  do  not  '*  froth"  when  tested  for  flash 
and  fire. 

Oils  differing  notably  from  above  requirements  will  be  rejected. 

Chicago,  Burlington  and  Quincy  Railroad  Company. 
Specifications  for  Cylinder  Stock. 

Use. — For  making  cylinder  lubricant.    One  grade. 

Require fnents. — Must  have  a  flashing  point  not  lower  than475^F.,  a 
burning  point  not  lower  than  575^  F.,  and  a  specific  gravity  between  25"^ 
and  28^  B.  Must  not  undergo  a  loss  greater  than  one-half  {\)  per  cent., 
when  exposed  for  three  (3)  hours  to  a  temperature  of  350^  F.     Must  be 


free  from  dirt,  griti  lumps  and  specks  ;  transparent  and  greenish  or  red- 
dish (not  black)  in  tint,  when  spread  as  a  thin  film  on  glass  and  looked 
toward  the  light. 

References  :  **  Measurements  of  Friction  of  Lubricating  Oils.*'  By 
C.  J.  H.  Woodbury,  Trans.  Am.  Soc.  Mech.  Eng,^  6, 136. 

"  On  the  Theory  of  the  Finance  of  Lubrication  and  on  the  Valuation  of 
Lubricants  by  Consumers."  By  R.  H.  Thurston,  Trans,  Am.  Soc. 
Meek.  Eng.y  7,  437. 

'*  Cost  of  Lubricating  Car  Journals.*'  By  L.  A.  Randolph,  Trans.  Am. 
Soc.  Mech.  Eng.t  10,  126-35. 

"Special  Experiments  with  Lubricants."  By  J.  E.  Denton,  Trans. 
Am.  Soc.  Mech.  Eng.,  la,  403-50. 

*'  Report  of  Committee  on  Lubrication  of  Cars  to  the  Master  Car  Build- 
er's Association  of  the  United  States  for  1893. ' '  The  /Railway  Car  Journal, 
4f  156.     (July,  1894.) 

"  History  of  Attempts  to  Determine  the  Relative  Value  of  Lubricants  by 
Mechanical  Tests."  Proceedings  of  the  American  Association  for  the 
Advancement  of  Science,  34*' 

"  Car  Lubrication."    By  W.  E.  Hall. 


Oils  Used  for  Illumination. 

Oils  used  for  illumination  may  be  classified  into  two  groups  : 

1.  Refined  products  from  petroleum,  such  as  naphtha,  gaso- 
line, kerosene,  signal  oil,  etc. 

2.  Certain  refined  oils  of  vegetable  and  animal  origin,  as  colza 
oil,  rape  oil,  lard  oil,  sperm  oil,  etc. 

/.  Rejined  Products  from  Petroleum. 

Kerosene  is  the  refined  product  from  petroleum  that  distills 
over  (in  the  refining  process)  after  the  lighter  oils,  naphthas,  etc., 
have  been  separated,  and  is  the  principle  oil  in  use  for  illumina- 
tion. In  color  it  varies  from  standard  white  to  water  white 
(colorless) ,  and  its  commercial  value  is  dependent  upon  its  flash 
and  burning  point.  In  the  oil  trade,  the  burning  or  fire  tests  are 
classified  as  110°  F.,  120°  F.  and  150"*  F.,  and  300''  F. 

The  150°  F.  is  known  as  headlight  oil  and  the  300°  F.  as  min- 
eral sperm  and  mineral  colza. 

The  requirements  for  mineral  oils  to  be  used  in  railroad 
illumination  are  as  follows  : 

OILS  nSED   FOR   Ii:.I.imiNATION . 

specifications  for  Petroleum  Burning  Oils. 

(CouditiaiiB  dI  shjpmeut  and  GEOFrnl  SpTdlEcatloas.} 

This  material  will  be  purcliased  by  weight.  Barrels  must  be  in  a  good 
condition  and  must  have  the  name  of  tile  contents  and  the  consignee's 
name  and  address  on  each  barrel,  and  plainly  marke<l  with  the  gross  and 
net  weight  which  will  be  subject  to  the  Company's  weight. 

When  received  all  shipments  will  be  promptly  weighed.  If  not  practi- 
cable to  empty  all  barrels,  tea  per  cent.  (io%)  will  be  emptied,  and  the 
losses  of  the  whole  shipment  will  be  adjusted  in  accordance  with  the  ten 
per  cent,  taken.  Should  the  net  weight  thus  obtained  be  less  by  one  per 
cent,  (t  %)  than  the  amount  charged  in  the  bill,  a  reduction  will  be  made 
for  all  over  one  per  cent. 

Prices  should  be  given  in  cents  or  hundredths  of  a  cent  per  pound. 

Shipments,  one  or  more  barrels  of  which  are  filled  with  oil  cloudy  from 
the  presence  of  glue,  or  which  contain  dirt,  water  or  other  impurities, 
will  he  rejected. 

Two  kinds  of  petroleum  burning  oils  will  be  used,  known  as  150^^  fire 
teat  for  general  use,  and  joa"  fire  test  for  use  in  passenger  cars. 
Detail  Spetifications. 

ISO-  Fire  Test  Oil. 
This  oil  must  conform  to  the  following  requirements  : 
1 — It  must  have  a  flash  test  above  lis''  P* 
5— It  must  have  a  fire  lest  not  below  ijo'-'  F. 
3 — It  must  have  a  cloud  test  not  above  o"  F. 
4 — It  must  be  a  "  water  white"  in  color. 
5 — Its  gravity  must  be  between  44-  and  48"  B.  at  60'-'  F. 
joo"  Pi  re  Teat  Oil, 
i  oil  must  conform  to  the  following  requirements  : 
I — It  must  have  a  Bash  test  above  130°  F. 
I— It  must  have  a  fire  test  not  below  300''  F. 
3 — It  must  have  a  cloud  test  not  above  31^  F. 
4 — It  must  be  a  "  standard  white"  in  color. 
S — Its  gravity  must  be  between  38°  and  41^  B.  at  60°  F. 

Method  of  Making  Tests. 

The  "Open  Tagliabue"  cup  is  used  for  determining  the  flash- 
ing and  burning  points  of  this  oil,  heating  the  oil  at  the  rate  of 
2°  F.  per  minute  and  applying  the  test  flame  every  degree  from 
120°  for  flash  atid  ever>'  4°  after  flash  for  the  burniug  point, 

The  ' '  Cleveland' '  cup  is  used  for  determining  the  flashing  and 
fttiuming  points  of  this  oil,  heating  at  the  rate  of  5°  per  minute 
rand  applying  the  test  flarae  every  5°  from  230°  F. 


Cloud  Test, 

The  cloud  test  is  made  as  follows :  Two  ounces  of  the  oil  are 
placed  in  a  four  ounce  sample  bottle,  with  a  thermometer  sus- 
pended in  the  oil.  The  bottle  is  exposed  to  a  freezing  mixture 
of  ice  and  salt  and  the  oil  stirred  with  the  thermometer  while 
cooling.  The  temperature  at  which  the  cloud  forms  is  taken  as 
the  cloud  test. 

The  requirements  for  the  flash  and  fire  test  for  illuminating 
oils  used  for  domestic  purposes  are  not  so  rigid  as  for  railroad 
practice.  In  fact  large  quantities  of  oil,  flashing  below  no**  F. 
are  used,  the  cheaper  price  being  the  incentive.  So  dangerous 
are  these  oils  with  low  flash  points,  that  many  states  have  passed 
stringent  laws  against  their  use.  An  oil  with  a  fire  test  of  no' 
F.  very  often  has  a  flash  test  of  90*  F.  and  many  oils  with  a  fire 
test  of  120°  F.,  flash  at  or  below  100°  F.  It  is  the  flash  point  of 
an  oil  that  makes  it  dangerous  and  while  the  refiners  of  oils 
mark  their  products  by  the  fire  test,  the  laws  as  passed  by  many 
of  the  states,  specify  the  flash  test  as  the  requisite. 

There  is  no  absolute  ratio  between  the  flash  and  fire  test  of 
an  oil,  since  while  many  illuminating  oils  have  a  high  fire  and 
flash  test,  others  may  have  a  high  fire  and  a  low  flash  test. 

The  instrument  that  gives  the  best  satisfaction  in  testing 
illuminating  oils,  not  lubricating  oils  (see  page 403),  for  the  flash 
and  fire  test  is  called  the  Wisconsin  Tester.     (Fig.  143.) 

It  is  thus  described  : 

On  the  left  side  of  the  figure  is  shown  the  instrument  entire.  It  con- 
sists of  a  sheet-copper  stand  eight  and  one-half  inches  high,  exclusive  of 
the  base,  and  four  and  one-half  inches  in  diameter.  On  one  side  is  an 
aperture  three  and  one-half  inches  high,  for  introducing  a  small  spirit- 
lamp,  A,  about  three  inches  in  height,  or  better,  a  small  gas  burner  in 
place  of  the  lamp  when  a  supply  of  gas  is  at  hand.  The  water-bath,  Z>, 
is  also  of  copper,  and  is  four  and  one-eighth  inches  in  height  and  four 
inches  inside  diameter.  The  opening  in  the  top  is  two  and  seven-eighths 
inches  in  diameter.  It  is  also  provided  with  a  one-fourth  inch  flange 
which  supports  the  bath  in  the  cylindrical  stand.  The  capacity  of  the 
bath  is  about  twenty  fluid  ounces,  this  quantity  being  indicated  by  a  mark 
on  the  inside.  C  represents  the  copper  oil-holder.  The  lower  section  is 
three  and  three-eighths  inches  high, and  two  and  three-fourths  inches  inside 
diameter.  The  upper  part  is  one  inch  high  and  three  and  three-eighths 
inches  in  diameter,  and  serves  as  a  vapor-chamber.  The  upper  rim  is  pro- 

Ided  with  a  small  flange  which 
The  oil  holder  contains  about  ten  fluid 
eighth  of  an  inch  of  the  Sange  which 
chambers.    In  order  to  prevent  reflectii 

bold  the  gtas 
ices,  when  filled  to  within  one- 
ns  the  oil  cup  and  the  vapor- 
tbe  otherwise  bright  sur- 
n  the  iaside  by  formini^a  sul- 


iS  blackened  o 

is  three  and  five-eighths  inches  in  diameter ; 
r  opening, 


face  of  the  metal,  the 

phide  of  copper,  by  mean 
Thecover,  C.isofglass. 

on   one   side  is   a  circuli 

closed  by  a  cork  through  which  the 

thermometer.  B.  passes.     In  front  of 

this  isa  second  opening  three-fourths 

of   an    inch  deep   and   the   same   in 

width  on  the  rim.  through  which  the 

flashing  jet  is  passed  in  testing.     The 

substitution  of   a   glass  for  a   metal 

cover  more  readily  enables  the  oper- 
ator to  note  the  exact  point  at  which 

the  flash  occurs.  A  small  gas  jet,  one- 
fourth  inch  in  length,  furnishes  the 

best  means  for  igniting  the  vapor. 
.Where  gas  cannot  be  had  the  flame 
^ftom  a  small  waxed  twine  answers 
■♦•ty  well, 

(l).  The  test  shall  be  applied  ac- 

.COrding  to  the  following  directions  : 

Remove   the   oil   cup  and    fill   the 

water-bath  with  cold  water  up  to  the 

iiiBrlc  on  the  iilside.     Replace  the  oil 

cup  and  pour  in  enough  oil  to  fill  it 

to  within  one-eighth  of  an  inch  of  the 

Range  joining  the  cup  and  the  vapor- 

chamher  above.     Care  must  be  taken 
oil   does  not  flow  over  the 

flange.     Remove  all  air  bubbles  with  Fig.  u; 

of  dry  paper.     Place  the  glass  cover  on  the  oil  cup, 

tbc  thermometer  that  its  bulb  shall  be  just  covered  by  the 

alcohol  lamp  is  employed  for  heating  the  water-bath,  the  wick 

Siiould  be  carefully  trimmed  and  adjusted  to  a  small  flame.      A  small 

Bunsen  burner  may  be  used  in  place  of  the  lamp.     The  rate  of  heating 

■hould  be  about  i-  per  minute,  and  in  no  case  exceed  3°. 

flash  torch,  a  small  gas  jet,  one-fourth  inch  in  length,  should  be 

employed.    When  gas  is  not  at  hand,  employ  a  piece  of  waxed  linen  twine. 

The  flame  in  this  case,  however,  should  be  small. 

When  the  temperature  of  the  oil  has  reached  85^  F.,  the  testings  should 

QOmmence.     To  this  end  insert  the  torch  into  the  opening 


passing  it  in  at  such  an  angle  as  to  well  clear  the  cover,  and  to  a  distance 
about  half  way  between  the  oil  and  the  cover.  The  motion  should  be 
steady  and  uniform,  rapid  and  without  any  pause.  This  should  be  re- 
peated at  every  2^  rise  of  the  thermometer  until  the  temperature  has 
reached  95^,  when  the  lamp  should  be  removed  and  the  testings  should 
be  made  for  each  degree  of  temperature  until  100^  is  reached.  After  this 
the  lamp  may  be  replaced,  if  necessary,  and  the  testings  continued  for 
each  2^. 

The  appearance  of  a  slight  bluish  flame  shows  that  the  flashing  point 
has  been  reached. 

In  every  case  note  the  temperature  of  the  oil  before  introducing  the 
torch.    The  flame  of  the  torch  must  not  come  in  contact  with  the  oil. 

The  water-bath  should  be  filled  with  cold  water  for  each  separate  test, 
and  the  oil  from  a  previous  test  carefully  wiped  from  the  oil  cup. 

(3).  The  instrument  to  be  used  in  testing  oils  which  come  under  the 
provisions  of  section  2  of  the  law  shall  consist  of  the  cylinder  D,  and  the 
copper  oil  cup  C,  together  with  a  copper  collar  for  suspending  the  cup 
in  the  cylinder,  and  an  adjustable  support  for  holding  the  thermometer. 

(4).  The  test  for  ascertaining  the  igniting  point  shall  be  conducted  as 
follows  :  Fill  the  cup  with  the  oil  to  be  tested  to  within  three-eighths  of  an 
inch  of  the  flange  joining  the  cup  and  the  vapor-chamber  above.  Care 
must  be  taken  that  the  oil  does  not  flow  over  the  flange.  Place  the  cup 
in  the  cylinder  and  adjust  the  thermometer  so  that  its  bulb  shall  be  just 
covered  by  the  oil.  Place  the  lamp  or  gas  burner  under  the  oil  cup. 
The  rate  of  heating  should  not  exceed  10°  a  minute  below  250°  F.,  nor  ex- 
ceed 5"^  a  minute  above  this  point.  The  testing  flame  described  in  the 
directions  for  ascertaining  the  flashing  point  should  be  used.  It  should 
be  applied  to  the  surface  of  the  oil  at  every  5°  rise  in  the  thermometer, 
till  the  oil  ignites. 

The  following  is  a  copy  of  the  law  of  the  state  of  New  York 
regulating  the  standard  of  illuminating  oils,  etc.: 

AN  ACT  to  regulate  the  standard  of  illuminating  oils  and  fluids  for  the 
better  protection  of  life  ^  health  and  property. 

Passed  June  6.  1882.  three-fifths  being  present. 

Section  i.  No  person,  company  or  corporation  shall  manufacture  or 
have  in  this  State,  or  deal  in,  keep,  sell  or  give  away,  for  illuminating  or 
heating  purposes  in  lamps  or  stoves  within  this  state,  oil  or  burning  fluid, 
whether  the  same  be  composed  wholly  or  in  part  of  naphtha,  coal  oil, 
petroleum  or  products  manufactured  therefrom,  or  of  other  substances 
or  materials,  which  shall  emit  an  inflammable  vapor  which  will  flash  at 
a  temperature  below  one  hundred  degrees,  by  the  Fahrenheit  thermome- 
ter, according  to  the  instrument  and  methods  approved  by  the  State 
Board  of  Health  of  New  York. 

§  2.  No  oil  or  burning  fluid,  whether  composed  wholly  or  in  part  of 

H.1  oil  and  petroleum  or  their  products,  or  other  substance  or  material, 
■which  will  ignite  at  a  temperature  below  three  hundred  degrees  by  the 
Fahrenheit  thermometer,  shall  be  burned  in  any  lamp,  vessel  or  other  sta- 
tionary fixture  of  any  kind,  or  carried  as  freight,  in  any  passenger  car, 
or  passenger  boat  moved  by  steam  power  in  this  State,  or  in  any  stage  or 
street  car  drawn  by  horses.  Exceptions  as  regards  the  transportation  of 
coal  oil,  petroleum  and  its  products  are  hereby  made  when  the  same  is 
securel}-  packed  in  barrels  or  metallic  packages,  and  permission  is  here- 
by granted  for  its  carriage  in  passenger  boats  moved  by  steam  power 
when  there  are  no  other  public  means  of  transportation.  Any  violation 
of  this  act  shall  be  deemed  a  misdemeanor  and  subject  the  offending 
party  or  parties  to  a  penalty  not  exceeding  three  hundred  dollars,  or  im- 
prisonment uot  exceeding  six  months,  at  the  direction  of  the  court. 

1 3.  It  shall  be  the  duly  of  the  State  Board  of  Health  of  New  York  to 
recommend  and  direct  the  nature  of  the  test  and  instruments  by  which 
the  illuminating  oils,  as  hereinbefore  described,  shall  be  tested  in  accord- 
ance with  this  act.  It  shall  be  the  duty  of  the  public  analysts,  who  may 
now  be  employed  by  the  State  Board  of  Health,  or  who  may  be  hereafter 
appointed,  to  test  such  samples  of  suspected  illuminating  oils  or  fluids  as 
may  be  submitted  to  them  under  the  rules  to  be  adopted  by  the  said 
board,  (or  which  service  the  said  board  shall  provide  reasonable  compen- 
sation at  the  first  quarterly  meeting  of  the  State  Board  of  Health  after  the 
passage  of  this  act  1  it  shall  adopt  such  measures  as  may  seem  necessary 
to  facilitate  the  enforcement  of  this  act,  and  prepare  rules  and  regulations 
with  regard  to  the  proper  methods  of  collecting  and  examining  suspected 
samplesof  illuminating  oils,  and  the  State  Board  of  Health  shall  be  author- 
ized to  expend,  in  addition  to  all  sum-!  already  appropriated  for  said 
board,  an  amount  not  exceeding  five  thousand  dollars  for  the  purpose 
of  carrying  out  the  provisions  of  this  act.  And  the  sum  of  five  thousand 
dollars  is  hereby  appropriated,  out  of  any  moneys  in  the  treasury  not 
otherwise  appropriated,  (or  the  purposes  of  this  section  as  provided. 

f4.  Naphtha  and  other  light  products  of  petroleum  which  will  not 
stand  the  flash  test  required  by  this  act  may  be  used  for  illuminating  or 
heating  purposes  only. 

In  street  lamps  and  open  air  receptacles  apart  from  any  building,  fac- 
tory or  inhabitated  bouse  in  which  the  vapor  is  burned. 

In  dwellings,  factories  or  other  places  of  business  when  vaporized  in 
secure  tanks  or  metallic  generators  made  for  that  purpose  in  which  the 
vapor  so  generated  is  used  for  light  or  heating. 

For  use  in  the  manufacture  of  illuminating  gas  in  gas  manufactories, 
situated  apart  from  dwellings  and  other  buildings. 

3  5.  It  shall  be  the  duly  of  all  district-attorneys  of  the  counties  in  this 
State  to  represent  and  prosecute  in  behalf  of  the  people,  withitj  their  re- 
spective counties,  all  casesof  offenses  arising  under  the  provisions  of  this 



gS.  Nothing  in  this  act  shall  be  so  construed  aa  to  interfere  with  the  pro- 
visions of  chapter  seven  hundred  and  forty-twoof  the  laws  of  eighteen  hun- 
dred and  seventy-one.  as  regards  the  duties  of  the  Bureau  of  Combustiblei 
of  the  city  of  New  York,  or  any  other  statutes  not  conflicting  with  this 
act.  provided  that  nothing  in  this  act  shall  be  deemed  to  interfere  with  or 
supersede  any  regulation  for  the  inspection  and  coutrol  of  combustible  ma- 
terials in  any  city  of  this  State  made  anil  established  in  pursuance  of 
special  or  local  laws  or  the  charter  of  said  city, 

S  7.  All  acts  or  parts  of  acts  inconsistent  with  this  act  are  hereby  repealed. 
§  8.  This  act  shall  lake  effect  sixty  days  after  its  passage. 
A  very  complete  report  upon  the  methods  and  apparatus  for 
testing  inflammable  oils  by  A. 
H.  Elliott,  Ph.D..  was  ren- 
dered to  the  New  York  State 
Board  of  Health  and  incorpor- 
ated in  their  annual  repwrt  for 

The  grades  of  color  of  an 
oil  are   noted   as   standard 
white,  prime  white,  superfine 
white  atid  water  white,'  and 
the  instrument  generally  used 
for  determination  of  the  color 
in  oils,  is  the  Stammer  Color- 
imeter (Fig,  144).    Tube /is 
closed  at  the  bottom    by  a 
t  transparent  glass  plate,  isopeii 
I  at  the  top,    and  a  projecting 
"  lip  on  the  side  whereby  the  oil 
F'K.  144-  to  be  tested  can  be  poured  in 

or  out.     The  tube  is  fastened  to  the  stand  by  two  screws.     The 
measuring  tube  ///is  closed  at  the  bottom  by  a  colorless  glass  I 
plate  and  is  movable  inside  of  tube  /. 

The  color-glass  cube  //  which  is  joined  firmly  to  the  meastir-  j 
ing  tube  ///,  is  open  at  the  bottom  and  at  the  top  contains  a  i 
colored  glass  plate,  which  plate  can  be  substituted  with  other 
tinted  glass  plates.     The  movement  of  the  joined  tubes // and 
///is  produced  by  inclosed  racket  work,  the  movement  of  the 


OILS  USHD  Foa  ili,uminat:on.  433 

tubes  being  read  on  a  scale  on  the  back  of  the  stand,  and  stated 
in  millimeters.  Since  the  color  of  a  liquid  is  inversely  propor- 
tioDal  to  the  height  of  the  column,  which  is  necessary  to  give 
the  standard  color,  and  since  this  color  is  here  expressed  by 
too,  the  absolute  number  for  expressing  the  tone  of  color  of  any 
oil  is  obtained  by  dividing  this  loo  by  the  number  of  millimeters 
read  oS  from  the  scale.     For  example  : 

Millimeter  scale  =  i.     Color  =  loo.oo 

I::    ;;  -=■  ;:  ='-" 
"    =7.      ■'    =   14.29 

■'    =  19.     ■■    =    5.26 

The  color,  tone  and  thickness  of  the  standard  glass  is  so  chosen 
that  the  scale  shows  the  following  values  for  the  ordinary  brands 
of  illuminating  oils. 

Standard  white  =   so.oinm. 
Prime  "      =   86.5     " 

Superfine      "     =  199.5    " 
Water  "     =300.00" 

Wilson's  calorimeter,  largely  used  in  England,  is  very  similar 
■in  construction  to  the  Sterner. 

2.    Vegetable  and  Anirnal  Oils, 

The  two  principal  oils  of  this  class  in  use  for  illumination  are 
colza  and  lard  oil. 

In  this  country  the  former  has  never  been  used  to  any  great 
extent,  its  use  being  confined  principally  to  Europe,  but  lard  oil 
and  sperm  oil,  in  former  years,  before  the  introduction  ol  the 
petroleum  products  for  this  purpose,  were  largely  used  as 
illuminants.  Except  in  railroad  practice  and  then  in  yearly  de- 
creasing amounts  their  use  now  is  very  limited  in  this  direction. 
In  the  matter  of  illumination,  the  methods  made  use  of  by  the 
railroads  are  worthy  of  study  and  comparison,  and  it  is  in  a  great 
measure  due  to  the  investigations  carried  out  in  their  interests 
that  the  great  advances  in  this  direction  are  due. 

George  Gibbs.  mechanical  engineer  of  the  Chicago  Milwau- 
kee and  St.  Paul  Railroad,  in  a  paper  recently  read  before  the 
Western  Railroad  Club  of  Chicago,  states  1 

There  are  about  eight  different  means  of  car  illumination ; 
•fiV.,  the  use  of 



I.  Vegetable  oils.  2.  Candles.  3.  Mineral  or  petroleiita, 
oils.  4.  Ordinary  coal  gas.  5,  Carburetted  water  gas.  6.  Ridk' 
or  oil  gas,  S.  Electric  light.  There  are  but  four  worthy  o(, 
consideration.     These  are : 

First.  Heavy  mineral  oils  in  lamps,  such  as  mineral  seal  whicb 
ranges  from  35°  B.  to  40°  B.  in  gravity,  has  a  fire  test  of  300°  P, 
and  gives  off  no  inflammable  vapor  below  230"  F. 

Second.  The  Pintsck  oil  gas.  This  is  by  far  the  most  promi- 
nent attempt  to  devise  any  economical  and  practical  gas-lighting 
system  for  railroad  service.  Its  primary  object  is  to  reduce  the 
bulk  of  stored  gas  necessary'  to  produce  an  adequate  illumina- 
tion for  a  considerable  length  of  time. 

The  Pintsch  system  has  largely  confined  its  attention  to  more 
efficient  gas,  which,  it  is  claimed,  is  supplied  by  the  use  of  a 
rich  permanent  oil  gas. 

Ordinary  city  or  coal  gas  when  burned  at  pressure  of  the  street 
mains,  one  to  one  and  one-half  ounces  may  be  taken  to  give  an 
illumination  of,  at  most,  four  candles  per  cubic  foot.  Oil  gas  at 
the  same  pressure  will  give  from  four  to  six  times  as  much,  say 
sixteen  candles  per  cubic  foot.  But  one  property  of  gas,  which 
vitally  affects  the  problem,  is  the  loss  of  light-giving  power  upon 
compression  and  storage.  This  is  true  of  all  illuminating  gas, 
and  is  due  to  the  deposition  of  the  rich  oily  hydrocarbons,  but 
is  not  true  to  the  same  extent  for  oil  and  coal  gas,  the  difference 
being  materially  in  favor  of  oil  gas.  Reliable  tests  for  this  loss 
of  light  by  compression  have  given  the  result  that  coal  gas  loses 
fifty  per  cent,  and  oil  gas  twenty-one  per  cent,  of  light-giving 
power  upon  compression  of  300  pounds  per  square  inch,  and  at 
225  pounds  per  square  inch  pressure,  the  quantities  required  for 
equal  illumination  would  be  about  as  five  of  coal  to  one  of  oil 

The  material  used  for  the  manufacture  of  Pintsch  gas  is  crude 
petroleum.  The  generation  of  gas  is  affected  by  vaporizing  the 
oil  at  a  high  heat  in  suitably  arranged  cast  iron  retorts,  the  pro- 
cess of  manufacturing  being,  on  a  small  scale,  essentially  that 
followed  for  city  gas.  From  the  storage  tank  pipe  connections 
lead  to  convenient  places  for  filling  the  car  tanks.  A  plant 
capable  of  making  sufficent  gas  for  500  cars  is  contained  in  a 


ea  m  a  ^m 

OIt,S   USED    FOR    ILLUMINATtON.  435 

one  story  building  26  ft.  X  36  ft.  The  outfit  on  the  cars  consists 
of  oue  or  two  cylinders  for  holding  the  compressed  gas,  a  pres- 
sure regulator  and  a  system  of  piping  to  the  lamps.  These  are 
of  special  design,  each  having  from  four  to  six  flames  ar- 
ranged beneath  a  procelain  reflector,  the  whole  encased  in  a 
glass  bell-jar ;  ventilation  is  suitably  provided  for  and  a  very 
steady  light  is  obtained. 

Mention  might  be  made  here  that  an  American  system,  the 
Foster,  appeared  a  few  years  ago  embodying  the  same  principles 
and  general  features  as  the  Pintsch. 

Third.  The  Frost  Systems.  In  the  Frost  and  all  other  simi- 
lar systems  the  principle  is  the  same,  being  the  property  possessed 
by  air  ol  holding  a  vapor  in  intimate  mixture  and  suspension, 
usually  the  vapor  of  gasoline.  The  amount  of  vapor  absorbed 
depends  upon  its  temperature;  thus,  at  14°  above  zero  (F.), 
about  six  per  cent,  and  at  68"  F..  twenty-seven  per  cent,  will  be 
taken  up.  This  is,  however,  a  mechanical  mixture  only  and  not 
a  permanent  gas.  The  vapor  thus  formed  is  capable  of  being 
burned  similarly  with  gas,  when  mixed  with  air  in  the  proper 
proportions,  giving  a  highly  luminous  flame.  This  principle 
has  been  utilized  for  many  years  for  making  gas  for  household 
purposes  in  places  where  city  gas  is  inaccessible,  a  simple  form 
of  air  pump  run  by  a  falling  weight  forcing  air  under  a  few  ounces 
pressure  through  a  tank  (generally  under  ground)  which  con- 
tains a  barrel  of  liquid  gasoline.  This  tank  is  divided  into  many 
compartments  in  which  absorbent  wicking  is  suspended,  dipping 
into  the  liquid  and  drawing  up  the  same  by  capillary  attraction. 
The  enriched  air  produced  in  this  carburetter  forms  the  gas  for 

The  difficulties  to  be  overcome  in  using  this  agent  for  safe  car 
lighting  are  as  follows :  First,  the  presence  of  liquid  gasoline. 
The  Frost  system  overcomes  this  objection  by  filling  the  carburet- 
ting  vessel  almost  completely  with  wicking  and  by  merely  satura- 
ting this  with  gasoline  and  drawing  off  the  superfluous  liquid. 
Second,  the  effect  of  variation  of  temperature  in  the  amount  of 
vapor  absorbed  by  the  air  current.  As  above  stated,  in  cold 
weather  only  a  small  percentage  is  absorbed,  too  little  to  produce 
a  good  light  and  in  warm  weather   too  much,  producing  a  rich 


but  smoky  light.  This  is  really  the  serious  stumbling  block  to 
this  system.  The  Frost  system  claims  to  overcome  it  by  placing 
a  small  generator  or  carburetter  above  the  light  on  the  roof  of 
the  car,  in  such  a  manner  that  a  portion  of  the  heat  generated 
by  the  burner  is  transmitted  to  the  carburetter,  insuring  a  uni- 
form temperature  at  all  times. 

The  system  in  detail  consists  of  an  air  storage  tank  underneath 
the  car,  containing  sufficient  compressed  air  to  supply  light  for 
six  hours.  This  compressed  air  is  obtained  directly  from  the 
train  pipe  of  the  air  brake  and  is  led  through  a  suitable  pressure 
reducer  and  a  regulator  to  the  carburetters  in  the  roof,  one  of 
these  being  placed  over  each  lamp,  and  thence,  after  passing 
through  them,  to  the  lamps  underneath.  These  are  now  con- 
structed on  the  Siemans  or  regenerative  principle  and  give  a 
brilliant  white  light  without  shadow.  The  supply  of  gasoline 
in  the  carburetters  is  sufficient  for  forty-three  hours  burning,  and 
then  can  be  recharged  by  filling  from  the  roof. 

Fourth.  The  Electric  System.  The  latest  phase  of  train  light- 
ing may  be  said  to  be  the  electric.  In  this  direction  numerous 
isolated  experiments  have  been  made  in  this  country  during  the 
past  five  years.  The  different  plans  suggested  for  obtaining 
electric  lighting  are  divided  as  follows : 

1 .  Primary  batteries ; 

2.  Secondary  batteries  or  accumulators  ; 

3.  Dynamo  machine  connected  to  car  axle,  with  or  without 
accumulators  as  auxiliaries ; 

4.  Dynamo  operated  by  special  steam  engine,  either  in  a  car 
or  on  the  locomotive,  and  supplied  with  steam  from  locomotive 
or  special  boiler  on  a  car  :  accumulators  used  or  not,  as  desired, 
as  equalizers. 

I .  The  first  method  has  been  tried  in  England  on  several  rail- 
ways, and  in  France  between  Paris  and  Brussels.  In  all,  a 
special  form  of  primary  battery  having  very  low  resistance,  great 
surface,  and  furnishing  a  constant  current  at  high  pressure,  was 
employed.  The  result  was  flat  failures,  on  account  of  the  enor- 
mous expense  of  the  electrical  energy  furnished  by  chemical 
means.  It  can  be  said  that  in  primary  batteries  chemicals  are 
expended  and  zinc  consumed,  instead  of  coal  under  a  boiler  to 


"produce  energy  ;  at  the  lowest  estimate,  the  former  is  forty  times 
as  expensive  as  the  latter. 

2.  111  England  the  London  and  Brighton  railway  made  an 
extensive  trial  on  a  Pullman  train  of  lighting  by  accumulators 
alone,  placing  batteries  under  each  car,  and  having  a  sufficient 
number  of  charging  stations,  with  hoilers.  engines  and  dynamos, 
to  charge  duplicate  sets  of  batteries  for  immediate  replacement. 

This  system,  alter  five  years  trial,  was  abandoned,  In  this 
couutry  the  Pullman  Company  gave  the  method  a  thorough  trial 
on  the  Pennsylvania  Railroad ' '  Limited ' '  between  New  York  and 
Chicago,  finallyabandoning  it.  It  was  also  tried  and  abandoned 
on  the  Baltimore  and  Ohio  and  Chicago,  Burlington  and  Quincy. 
Description  of  this  system  may  be  dismissed  by  briefly  stating 
that  each  car  carries  its  own  store  of  batteries  in  boxes  hung 
underneath,  arranged  so  that  they  can  be  readily  removed  at 
terminals  for  recharging  by  dynamo,  or  for  substitution  of  fresh 
cells,  The  weight  of  batteries  required  for  a  standard  coach  is, 
approximately,  one  ton. 

3.  Third  method.  A  favorite  method  for  obtaining  electricity 
at  a  low  cost  seems  to  have  been  to  connect  the  dynamo  to  a  car 
axle  ;  but  the  difficulties  of  obtaining  regular  motion  and  current 
and  providing  light  when  the  train  stops,  have  necessitated  the 
employment  of  accumulators  as  regulators  and  auxiliaries.  lu 
these,  automatic  appliances  are  provided  to  cut  off  the  current 
from  the  dynamo  when  the  speed  of  the  train  falls  below  a  cer- 
tain rate,  and  to  deliver  the  current  to  the  batteries  in  the  same 
direction.  The  main  difficulty,  with  Ibis  method,  and  one  which 
the  International  Railway  Congress  states  has  not  been  solved 
satisfactorily,  is  the  transmission  of  power  from  the  axle  to  the 

4.  The  fourth  method  is  the  only  prominent  electrical  one  in 
this  country  for  car  lighting.  It  consists  essentially  in  the  use 
of  a  dynamo  driven  by  special  steam  engine,  with  secondary 
batteries  for  reser\'e.  The  use  of  the  method  without  the  bat- 
teries as  auxilaries  has  been  often  attempted  without  success, 
but  recently  by  improvements  made  by  Mr.  Gibbs,  the  batteries 
are  dispensed  with  and  a  system  perfected  that  gives  general 

latisfaction  for  the  purpose.     The  plant  in  fact  is  made  an  exact 


duplicate  of  stationary  electric-lighting  plants.  The  engine  is  a 
15  horse-power  Westinghouse  automatic,  the  dynamo  a  150  light 
Edison  compound- wound,  connection  from  one  to  the  other  being 
made  by  belting.  In  the  summer  season,  when  steam  heat  is 
not  required  for  the  train,  this  outfit  is  placed  in  the  forward 
end  of  the  baggage  car,  occupying  twelve  feet  in  the  length  of 
the  car,  but  not  obstructing  passageway  through  it.  Steam  is 
taken,  at  sixty  pounds  pressure,  from  the  locomotive  boiler.  In 
winter  the  drain  upon  the  locomotive  for  steam  heat  is  often  ex- 
cessive ;  to  overcome  this  a  special  car  for  heating  and  lighting 
is  used.  Consult  Engineering  (London),  January  5,  1894,  for 
a  complete  description  of  this  system  as  now  used  successfully 
on  the  Chicago,  Milwaukee  and  St.   Paul  Railroad. 

Relative  Advantages  and  Disadvantages  of  the  Various  Systems, 

The  Electric  may  be  considered  adapted,  in  the  present  state 
of  the  art,  to  special  service  only.  It  fills  a  number  of  the  re- 
quirements for  a  perfect  light  in  a  manner  that  no  other  light 
approaches ;  it  is  cleanly,  cool,  safe,  allows  excellent  distribution 
and  is,  in  fact,  a  luxury  which  is  duly  appreciated  by  the  public. 
It,  however,  is  costly,  and  requires  great  attention  to  details; 
still,  in  many  instances  it  will  pay,  and  each  manager  must 
consider  whether   under  his  conditions  its  use  is   warranted. 

The  Frost  System  is  still  in  the  process  of  development.  It 
has  many  advantages  from  an  outside  point  of  view  ;  it  is  cleanly, 
the  light  is  good,  each  car  is  perfectly  independent  of  others  for 
its  supply  of  light,  and  it  requires  no  external  gas  works.  On 
the  other  hand  the  first  cost  is  excessive  ;  the  light  is  not  cheap 
for  running,  its  quality  is  not  uniform — due  to  the  effect  of  vary- 
ing temperature  and  quality  of  gasoline — the  apparatus  is  com- 
plicated, and  while  the  system  may  be  considered  safe  to  the  car 
itself,  the  use  of  gasoline  at  various  points  on  a  large  system  is 

The  Pintsch  System.  This,  in  spite  of  some  defects,  is  probably 
the  most  feasible  and  promising  method  in  the  direction  of  safety 
car  lighting.  It  is  safe  as  any  flame  method  of  lighting  can  be, 
is  cleanly  and  simple,  and  is  cheap  in  maintenance  and  running. 


It  is,  however,  very  high  in  first  cost,  and  is  not  universally 
applicable  on  account  of  dependence  upon  gas  works.  But  all 
main  line  tra£Bc  and  many  important  branch  lines  can  generally 
be  provided  for  by  this  system  at  a  moderate  cost  and  under  its 
rapid  extension  now  taking  place,  it  seems  likely  that  gas  works 
can  be  maintained  by  different  roads  at  many  points,  to  still 
further  reduce  the  individual  outlay. 

Oil  Lighting  by  Lamps.  Many  of  the  requirementsof  a  satisfac- 
tory car-lighting  system  appear  to  be  embodied  in  the  present  oil 
system,  or  might  be  with  some  improvements  which  are  readily 
attainable.  In  no  system,  with  the  exception  of  the  electric,  is 
it  possible  to  obtain  a  better  or  more  satisfactory  distribution  of 
light,  the  centers  being  of  moderate  intensity  ;  the  fuel  is  safe 
to  handle  and  may  be  obtained  without  delay  ;  each  car  is  inde- 
pendent of  the  others  :  it  is  cheapest  iu  first  cost  and  mainten- 
ance for  a  given  amount  of  light ;  it  is  simple  and  requires  bnt 
little  attention.  On  the  other  hand,  it  shareswith  flame  systems 
the  objections  of  giving  out  much  heat ;  the  quantity  of  light  is 
often  irregular  and  the  smell  objectionable  when  proper  care  is 
^not  exercised. 

The  possible  improvements  in  this  system  should  have  more 
iitteution  from  railway  officials. 

For  instance,  the  button  form  of  burners,  of  which  the  "Acme" 
Lis  a  good  example,  appears  to  solve  the  problem  of  sufficient  light 
Vss  has  been  done  in  the  other  flame  systems,  and  these  burners 
I  should  be  substituted  for  the  old  uneconomical  forms. 



-mi*  jlllSin^'^ioH ! '"  "^*r 



.,„„    8  88  88  88* 


■■""■B"  StMKI-! 





r.w»i4"i  !=  =  =-«-'• 

1 :5?SE 



The  Analysis  of  Lubricating  Oils  Containing  Blown  Rape- 
Seed  and  Blown  Cotton-Seed  Oils. 

Rape-seed  oil  has  long  been  the  standard  oi!  in  Europe  for 
lubrication.  Its  constancy  of  viscosity  at  varying  temperatures, 
its  non-liability  to  acidity  as  compared  with  other  seed  oils,  and 
its  low  cold  test,  unite  in  producing  the  results  required  of  a 
good  lubricant.  It,  however,  is  no  exception  to  the  rule  that 
vegetable  and  animal  oils  suffer  partial  decomposition  when  sub- 
jected to  high  temperature  produced  by  friction,  with  the  result 
that  fatty  acids  are  liberated  and  corrosion  of  bearings  produced. 

The  substitution  of  mineral  oils  in  varying  proportions  with 
rape-seed  oil  has  reduced  this  tendency,  this  reduction  being 
determined  by  the  percentages  of  mineral  oil  present,  as  the 
latter  liberates  no  free  acids. 

It  is  a  peculiar  fact,  however,  that  a  mineral  oil  alone  does  not 
give  as  satisfactory  results  in  lubrication  (especially  cylinder 
lubrication')  as  does  a  mixture  of  mineral  and  vegetable  or 
mineral  and  animal  oils,  one  of  the  primary  causes  being  that 
the  viscosity  o(  mineral  oils  rapidly  diminishes  at  high  tempera- 
tures, whereas  the  reduction  of  viscosity  of  vegetable  and  animal 
oils  is  very  much  less. 

If  it  were  not  for  this  peculiarity  between  these  two  classes  of 
oils,  mineral  lubricating  oils  could  easily  supplant  (on  the  score 
of  cheapness)  all  other  oils  used  in  lubrication. 

The  admixture  of  oils  then  being  required  for  the  better  class 
of  lubricants,  it  follows  that  in  England  where  rape-seed  oil  has 
been  the  standard,  its  use  should  be  continued  in  compounded 

The  proportion  of  rape-seed  oil  added  to  mineral  oil  varies 
from  five  to  twenty  per  cent.  Where  the  mineral  oil  is  a  clear 
paraffin  oil  twenty  per  cent,  of  the  seed  oil  is  used  ;  where  the 
mineral  oil  is  a  dark,  heavy  oil,  five  per  cent,  is  generally  added. 

The  separation  and  estimation  of  the  rape-seed  oil  in  these 
L  mixtures  presents  no  difficulty  to  the  analytical  chemist  when  no 
I  Other  seed  oil  is  present,  since  the  saponification  of  the  seed  oil, 

I  Tht  Failroad  and  EntinteringJuHinal.  it,  73-"* 


the  separation  of  the  fatty  acids  and  recognition  of  the  same  are 
a  part  of  the  usual  chemical  work  of  this  character.  The  recogni- 
tion of  the  constituents  of  a  mixed  lubricating  oil  by  analysis  is 
a  very  different  problem  from  giving  a  formula  by  which  the 
mixture  can  be  made.  This  is  evidenced  as  follows  : 
Suppose  the  analysis  shows 

Rape-seed  oil,  20  per  cent. 
Paraffin  oil,      80  per  cent. 

Paraffin  oil  varies  in  specific  gravity  from  0.875  to  0.921,  and  it 
is  essential  to  include  in  the  report  of  the  analysis  not  only  the 
amount  of  the  paraffin  oil  but  also  the  gravity,  since  paraffin  oil 
of  gravity  0.875  is  a  very  different  product  from  that  of  0.921 
gravity,  the  former  selling  at  seven  and  one-half  cents  and  the 
latter  at  twenty-three  cents  per  gallon.  This  determination  can 
be  made  by  taking  the  gravity  of  the  original  mixed  oil  (0.912), 
then  knowing  by  the  analysis  that  twenty  per  cent,  is  rape-seed 
oil  (gravity  0.918),  the  gravity  of  the  eighty  per  cent,  of  paraf- 
fin oil  is  easily  calculated.     Thus  : 

X  =  specific  gravity  of  rape-seed  oil  (0.918) 
y  =  specific  gravity  of  paraffin  oil 
;r  =  20  per  per  cent,  or  \ 
J/  ^  80  per  cent,  or  \ 
Then  \x  -\-  \y  =0.912 
0.183 +  f  J/ =0.912 
4^  =  0.729 
jr  =  0.910 
The  mixture  being  composed,  therefore,  of 
Paraffin  oil  (sp.  gr.  0.910),  80  per  cent. 
Rape-seed  oil  (sp.  gr.  0.918)  20  per  cent. 

The  direct  determination  by  analysis  from  the  ether  solution 
of  the  mineral  oil  in  the  mixture  does  not  give  an  oil  of  the  same 
specific  gravity  as  the  mineral  had  before  it  was  mixed  with  the 
seed  oil.  This  can  be  accounted  for  by  the  volatilization  of  a 
portion  of  the  lighter  hydrocarbons  of  the  mineral  oil  when  the 
ether  is  expelled  during  the  analysis.  For  this  reason  the 
determination  of  the  percentage  of  seed  oil  and  the  calculation 
of  the  mineral  oil  offers  less  liability  to  failure  than  finding  the 
mineral  oil  directly. 

The  introduction  of  blown  rape-seed  oil  instead  of  the  normal 


rape-seed  oi!  complicates  the  investigation  and  renders  the  use 
of  the  formula  above  given,  valueless.  Rape-seed  oil  has  a  grav- 
ity of  0.915  to  0.920.  Rape-seed  oil  blown  has  agravity  of  from 
0.930  to  0.960. 

Two  difficulties  are  immediately  presented  :  ( i )  The  chemical 
analysis  does  not  indicate  whether  the  rape-seed  oil  is  blown  or 
not ;  (2)  The  use  of  the  formula  given  without  the  correct  grav- 
ity of  the  blown  oil  would  give  false  results  regarding  the  par- 
affin oil.  To  overcome  this  difficulty  some  synthetical  work  is 
required . 

Suppose  the  specific  gravity  of  the  mixed  oil  is  0.922  and  the 
analysis  shows  twenty  per  cent,  of  rape-seed  oil.  It  will  be  nec- 
essary then  to  produce  a  mixture  in  these  proportions  that  will 
duplicate  the  original  sample.  A  check  upon  this  will  be  the 
viscosity  of  the  original  sample  as  compared  with  the  one  to  be 
made  by  formula.     Thus : 

The  original  oil  has  a  gravity  of  0.922,  contains  (by  analysis) 
twenty  per  cent,  of  rape-seed  oil,  and  has  a  viscosity  at  100°  F. 
of  335  seconds  (Pennsylvania  Railroad  pipette). 

First. — Make  a  mixture  of  paraffin  oil  (sp.  gr.  o.gro)  gen- 
erally used  in  this  character  of  lubricant,  eighty  per  cent,,  and 
rape-seed  oil  (unblown)  twenty  per  cent.  The  viscosity  is  165 
seconds,  showing  that  this  mixture  cannot  be  used  ill  place  of 
the  original  oil. 

Second. — Make  a  mixture  of  paraffin  oil  (sp.  gr.  0.910)  and 
rape-seed  oil  partially  blown,  (sp.  gr.  0.930)  in  the  same  propor- 
tions as  above.  The  resulting  viscosity  is  267  seconds,  showing 
that  the  compound  is  still  lacking  in  viscosity. 

Third. — Make  a  mixture  of  paraffin  oil  (sp.  gr.  O.910)  eighty 

parts,  and  rape-seed  oil,  blown  (sp.  gr.  0.960),  twenty  parts. 

The  viscosity  is  332  seconds. 

I      This  now  fulfills  the  conditions  required  and  the  synthetical 

I  sample  agrees  with  the  original  in  gravity,   composition  and 


The  use  of  blown  rape-seed  oil  -is  being  gradually  re- 
placed by  blown  cotton-seed  oil.  The  latter,  which  has  had 
but  a  limited  use  in  lubrication,  owing  to  its  liability  to  acidity, 


^  lias   been   greatly   improved    by   this   process  of   "  blowing,"  ^| 


which  is  nearly  complete  oxidation  of  the  oil  under  comparatively 
high  temperature. 

This  largely  prevents  the  occurrence  of  the  acidity  in  the  oil, 
and  thus  the  main  objection  to  its  use  in  lubrication  disappears. 
It  is  much  cheaper  than  rape-seed  oil,  since  it  costs  thirty  cents 
per  gallon,  to  sixty  cents  per  gallon  for  the  latter.  The 
chemical  reactions  of  the  two  oils  are  very  similar,  and  careful 
analytical  work  is  required  that  the  chemist  be  not  misled. 

The  following  table  of  comparisons  will  indicate  this  : 

Spbcific  GRAvrrv, 

Cotton-seed  oil 0.920  to  0.925 

Rape-seed  oil  *  0.915  to  0.920 

Blown  cotton-seed  oil 0.930  to  0.960 

Blown  rape-seed  oil 0.930  to  0.960 

Viscosity  (Pennsyi^vania  Raii,road  Pipette)  at  ioo°  F. 


Cotton-seed  oil  (sp.  gr.  0.925) 162 

Rape-seed  oil  (sp.  gr.  0.918) 210 

Blown  cotton-seed  oil  (sp.  gr.  0.960) 2143 

Blown  rape-seed  oil  (sp.  gr.  0.960)  2160 

Heidenrbich's  Test. 

Before  stirring:.  After  stirring. 

Cotton-seed  oil Paint  reddish  brown        Brown. 

Rape-seed  oil Yellow  brown  Brown. 

Massib's  Test. 

Cotton-seed  oil Orange  red. 

Rape-seed  oil Orange. 

Iodine  Absorption. 

Cotton-seed  oil 104  to  114 

Blown  cotton-seed  oil 93  to  103 

Rape-seed  oil 102  to  108 

Blown  rape-seed  oil 94  to  100 

In  the  comparison  of  the  two  oils,  when  not  mixed  with  a 
mineral  oil,  the  above  tests  can  be  used.  The  conditions  are 
altered,  however,  when  either  one  or  both  are  so  mixed,  since 
these  tests  apply  only  to  the  pure  oils  and  not  to  those  reduced 
with  large  percentages  of  mineral  oil.  After  the  separation  of 
the  seed  oil  from  the  mineral  oil  by  saponification  the  identifica- 
tion of  the  seed  oil  depends  upon  the  reactions  of  the  fatty  acids 
obtained,  and  a  careful  examination  and  comparison  of  these 


reactions  shows  that  the  melting  points  have  the  greatest  differ- 
ence and  thus  become  a  means  of  recognition. 

Thus,  the  fatty  acids  from  rape-seed  oil  melt  at  20°  C,  and 
from  cotton-seed  oil  at  30^  C.  Hence,  if  upon  analysis  of  a 
lubricating  oil  under  above  conditions,  the  fatty  acids  obtained 
show  a  melting  point  of  20^  C.  the  seed  oil  can  be  pronounced 
rape-seed  oil. 

If  the  melting  point  is  between  these  limits,,  say  23^  C,  the 

seed  oils  are  present  in  a  mixture,  the  proportions  of  which  can 

be  determined  by  the  following  formula : 

Wi  =  proportion  of  rape-seed  oil. 
Wf  ^  proportion  of  cotton-seed  oil. 
a;,  SB  weight  of  mixture  (20  per  cent.) 

ti  :=  temperature  of  melting  point  of  fatty  acids  of  rape-seed  oil. 
/,  =  temperature  of  melting  point  of  fatty  acids  of  cotton-seed  oil. 
/,  =  temperature  of  melting  point  of  mixed  fatty  acids. 

Then  Wj  =  ^tJ    J 

H — n 
Inserting  the  value  : 

2Z/i  as  20  ^3    3Q  -s  j^  per  cent. 
20 — 30 

ze/,  =  20  ?^IZ?2=K  6  per  cent. 
30 — 20 


Paraffin  oil 80  per  cent. 

Rape-seed  oil 14  per  cent. 

Cotton-seed  oil 6  per  cent. 

Total 100  per  cent. 

By  synthetical  work  upon  these  proportions,  with  comparison 
of  viscosities  of  the  sample  submitted  with  the  product,  the  re- 
sult will  be  not  only  a  correct  analysis,  but  a  working  formula 
can  be  given  by  which  a  manufacturer  can  duplicate  the  origi- 
nal oil. 

The  Analysis  of  Cylinder  Deposits. 

The  deposits  in  steam  cylinders,  formed  by  the  decomposition 
of  lubricating  oils,  may  be  classed  as  simple  or  compound,  de- 


pending  upon  whether  the  deposit  is  due  to  the  decomposition  of 
the  oil  alone  or  if  foreign  matters,  carried  over  in  the  steam  from 
the  boilers,  are  also  present. 

In  the  former  case,  carbon,  hydrocarbons,  oils  and  iron  oxide 
are  the  principal  constituents,  whereas,  in  the  latter,  oleate  of  lime, 
carbonate  of  lime,  and  silica  are  often  present  in  addition  to  the 

The  following  anal^'sis  of  a  sample  from  a  locomotive  cylinder 
would  indicate  a  simple  deposit. 

Moisture 2.28  per  cent. 

^.,       ,  , ,    .      ^.       r  Animal  10.54 

Oils  soluble  in  ether  {  ^.^^^^ „  ,^        .. 

Hydrocarbons  insoluble  in  ether 47'97        ** 

Fixed  carbon 23.73        ** 

FeO 2.83 

Undetermined 1.42 


Total 100.00 


And  the  one  given  below,  of  a  deposit  from  the  steam  cylinders 
of  a  large  stationary  engine,  would  show  that  scale- forming  mate- 
rial from  the  boilers  had  become  a  component. 

Moisture 13.12  per  cent. 

^..        ,   Ki    •      ^u     /Animal 8.15 

Oils  soluble  in  ether  \  Mi^^ral 7.86 

Soap 2.10 

Hydrocarbons  insoluble  in  ether 1.67 

Fixed  carbon 2.71 

Oxides  of  Iron  and  aluminum  6.81        " 

SiO, 3.65 

CaCOs  43.22 

MgCOs 10.17 

Undetermined  0.44 




Total 100.00 


In  many  samples  I  have  found  copper  and  zinc  in  the  de- 
posits, formed  by  the  corrosive  action  of  the  liberated  oleic  acid 
from  the  animal  oil  upon  the  brass  or  composition  bearings. 

This  corrosive  action  is  very  marked  where  a  poor  quality  of 
lubricating  oil,  composed  of  animal  or  vegetable  oil,  is  used, 
whereas,  a  pure  neutral  mineral  oil  has  no  acid  action  at  steam 


i  temperature.  Oftentimes  the  statemeiit  has  been  made  to  me, 
when  the  deposit  was  given  for  analysis,  "All  of  our  lubricating 
oil  is  pure  mineral  oil :  we  use  no  other."  And  yel.  upou 
analysis,  lard  oil  would  be  showu  in  comparatively  large 

This  is  accounted  for  from  the  fact  that  while  the  consumer 
believes  he  is  using  pure  mineral  oil — which  was  sold  to  him  as 
such — the  manufacturer  has  introduced  from  three  to  thirty  per 
cent,  of  lard  or  cotton-seed  oil. 

A  large  majority  of  the  so-called  "pure  mineral"  lubricating 
oils  for  cylinder  use  contain  at  least  five  per  cent,  of  animal 
oil :  and  it  is  the  exception  and  not  the  rule  to  find  a  "  pure 
mineral"  oil  for  cylinder  lubricating  purposes. 

An  analysis  of  a  deposit  from  the  steam  cylinder  of  a  large 
freight  steamer  gave  as  a  result : 

Moisture 16.16  per  cent. 

f  Castor  oil 26.19         " 

Oil.  K.lubl=,=  «h.,,  „.„„,, j,j„  ,. 

Fixed  t^arbon 7.91         ' "        • 

CuO 0.50 

FeO »5io 

Undetermined 1.63         " 

Total 100.00 

Pure  mineral  lubricating  oil  was  supposed  by  the  officers  of 
the  vessel  to  be  the  only  lubricant  used,  and  special  care  had 
been  taken  to  secure  it,  but  it  appears  that  the  engineer  added  a 
small  amount  of  castor  oil  to  the  mineral  oil,  as,  in  his  opinion, 
it  made  a  better  lubricant. 

The  decomposition  of  the  castor  oil  and  liberation  of  the  fatty 
acids  was  the  primary  cause  of  the  deposit. 

The  action  of  the  fatty  acids  upon  the  iron  and  metal  bearings 
results  in  different  products.  That  is  to  say.  while  the  copper 
when  present  has  generally  been  estimated  as  copper  oxide  the 
iron  may  exist  only  as  oxide  or  as  metallic  iron,  or  both. 

No  doubt  the  oleic  acid  acts  to  form  salts  of  these  metals,  but 
it  is  certain,  in  many  instances,  that  when  formed,  they  are 
immediately  decomposed  or  partially  so,  and  a  resulting  mixture 
formed  that  is  somewhat  difficult  of  analysis. 




In  the  analysis  here  given,  it  will  be  noticed  that  the  iron  was 
found  both  as  metal  and  as  oxide. 

Moisture 3.77  per  cent. 

^.,       ,  ^,    .       .      /Animal 21.27        ** 

Oils  soluble  m  ether  {  ^.^^^^^ ^^^        ,. 

Soap traces 

Fixed  carbon  10.90 

FeO  14.01 

Fe  27.85 

PbO 0.82 

CuO 1.07 

Undetermined  0.71        '* 

Total 100.00 


The  evolution  of  hydrogen  by  hydrochloric  acid,  from  the  de- 
posit, after  all  the  oils  and  fatty  substances  had  been  removed,  indi- 
cated the  presence  of  metallic  iron,  and  the  analysis  jf  the  resi- 
due, after  the  combustion  of  the  fixed  carbon,  gave  figures  by 
which  the  ratio  of  iron  and  iron  oxide  could  be  determined.  A 
portion  of  the  deposit,  after  extraction  of  oils  by  ether,*  is  dried, 
then  weighed,  the  hydrocarbons  driven  off  by  heat,  and  the 
amount  of  fixed  carbon  present  converted  by  combustion  with 
sulphuric  acid  and  chromium  trioxide  into  carbon  dioxide  and 
weighed,  this  weight  being  calculated  back  to  carbon. 

Another  portion  of  the  same  residue  is  ignited  in  a  platinum 
crucible  until  the  carbon  is  all  consumed,  then  weighed.  If  the 
amount  of  carbon  found  is  small  and  iron  large,  this  weight  may 
be  larger  than  the  original  weight  of  the  residue  taken,  owing 
to  oxidation  of  metallic  iron  to  ferric  oxide. 

Knowing  the  weight  of  carbon,  and  by  making  a  determina- 
tion of  iron  in  another  sample  before  ignition,  the  amount  of 
iron  oxide  is  easily  found. 

^The  Soxhlet  apparatus  as  shown  in  Fig.  145 is  well  adapted  for  this  purpose. 





















1  i=«r^ip 





















1  ai!i 











I  1- 

1  „  = 










s  i 

3    S 








.       J 



Where  qualitative  analysis  has  shown  the  deposit  to  be  a  sim- 
ple one,  the  analysis  can  be  stated  as  follows  : 

Moisture per  cent. 


Oils  soluble  in  ether  ( ^^{  i ' ; ;  ] ; ;  ]  [ ; ; ; ;  i^     .. 

Hydrocarbons  insoluble  in  ether ** 

Fixed  carbon ** 


Fe *• 


For  a  complex  deposit,  the  following  form  can  be  used  : 
Moisture per  cent. 

Oils  soluble  in  ether{^^^^^^\- ;;;;;;;;'  ;••  '/. 


Hydrocarbons  insoluble  in  ether ** 

Fixed  carbon •* 







MgO *• 



SiO,.  etc 


Where  the  lime  and  magnesia  exist  in  amounts  more  than 
necessary  to  combine  with  the  carbon  dioxide  and  sulphur  tri- 
oxide  present,  the  excess  may  have  united  with  oleic  acid  to  form 
soaps  insoluble  in  water,  but  soluble  in  ether. 

In  some  instances  the  lead  oxide  and  zinc  oxide  will  be  found 
only  in  the  ether  soap  solution  (3),  as  lead  and  zinc  oleates,  but 
in  others,  while  they  undoubtedly  first  existed  as  oleates,  they 
had  become  decomposed,  and  the  lead  and  zinc  oxides  would  be 
found  in  section  (8)  of  the  above  s(!heme. 

The  following  is  an  analysis  of  a  cylinder  deposit,  '*  Mica 
Grease''  having  been  used  as  a  lubricant : 


Moisture 11.23  percent. 

/Animal 9.05        ** 

Oils  soluble  in  ether  {  Mineral 6.28        - 

Soap  (CaO  +  MgO  united  with  fatty  acids)..  43.90 

Fixed  carbon 6.33 

Oxides  of  iron  and  aluminum 6.59 

CaO 3.15 

MgO 2.19 

CO, 6.27 

Silica  and  mica 5.01        ** 


Total 100.00 


References  :—'*T\i^  Production  of  Paraffin  and  Paraffin  Oils."    By  R.H. 

Brunton,  C.  E.,    Proc.  Inst,  Civ,  Eng,^  66,  180-237. 
*'  The  Russian  Petroleum  Industry."    By  Boverton  Redwood,  F.C.S.^ 

y.  Soc,  Chem,  Ind,^  4*  70-82. 
'*  On  the  Testing  of  Lard  for  Cotton-seed  Oil  and  Beef  Stearin."     By 

John  Pattison,  F.I.C.,/-  Soc.  Chem.  Ind,^  8,  30-31. 
"The   Manufacture  of  Paraffin   Oil."      By   D.   R.  Stewart,   F.C.S., 

Ibidt  8,  100- no. 
"  Wool-Fat,  and  the  Processes  of  Obtaining  It."  By  H.  W.  Langbeck, 

If>id,  9>  356-359- 
"Some  Experiments    on    Petroleum    Solidification."      By   Samuel 

Rideal,  F.I.C,     Ibid,  10,  889. 
"The  Flashing  Test  for  Petroleum."     By  F.  A.  Abel,  F.R.S.,   Ibid, 

I,  471-478. 
"  Report*  on  Lighting  Passenger  Equipment."     Master  Carbuilders* 

Association  for  1893.     The  Railway  Car  Journal,  July,  1894. 

Paint  Analysis. 

Paint  is  a  liquid  preparation  having  a  two-fold  use.  Pri- 
marily it  acts  as  a  protecting  coating  against  the  action  of  the 
weather,  and  simultaneously  as  a  decorative  agent. 

The  liquid  is  usually  linseed  oil  and  turpentine  and  the  color* 
ing  matter  or  body  some  solid  pigment,  such  as  finely  ground 
red  oxide  of  iron. 

1  This  report  includes  a  list  of  the  principal  railroads  in  the  United  States,  and  the 
methods  used  by  each  for  passen^^er  car  illumination  ;  vf>.,  oil  lamps,  oil  gas  under 
pressure  (**  Piutsch  gas"),  or  electric  light  (incandescent),  with  the  conclusion  that  the 
•'  Piutsch  gas  "  is  rapidly  being  adopted  in  preference  to  oil  illumination. 


It  is  essential  in  the  production  of  a  good  paint  that  the  oil 
used  should  be  one  that,  upon  drying  on  the  surface  applied, 
should  become  hard,  lustrous,  and  somewhat  elastic. 

Linseed  oil  excells  all  others  in  use  for  this  purpose,  and  any 
sophistication  thereto  only  deteriorates  the  quality. 

Four  qualities  are  essential  in  a  paint:  i.  Durability;  2. 
Working  Qualities ;  3.  Drying  Properties ;   4.  Covering  Power. 

The  following  list  of  pigments,  with  their  chemical  composi- 
tion stated,  will  give  an  idea  of  the  great  variety  that  can  be 
used  in  paints  for  outside  work.  The  list  would  be  largely  in- 
creased were  other  pigments  included  that  are  used  for  interior 
decorative  work  only. 

Red  Pigments, — Indian  red,  Tuscan  red  (Fe,0,),  vermilion 
(HgS),  red  lead  (Pb.OJ.  antimony  vermilion  (Sb,S,).  Iron 
oxide,  Indian  red,  and  Tuscan  red  can  be  analyzed  by  Scheme 
XIII,  p.  29. 

Brown  Pigments, — Umbers  (Fe,0,,  MnO,,  etc.),  Van  Dyke 
brown,  (Fe,0„  carbon),  manganese  brown  (Mn.OJ  and  sepia. 
The  composition  of  sepia  is  as  follows  : 

Melanin 78.00  per  cent. 

CaCO, 10.40 

MgCO, 7.00 

Alkaline  sulphates 2.16 

Organic  mucus i  .84 




White  />/:^»i^/j.— White  lead  (2PbC0..PbH,0,),  lead  sul- 
phate (PbSO^),  zinc  white  (ZnO),  sulphide  of  zinc,  white  (ZnS), 
**  lithophone."  Also  the  following,  added  oftentimes  as  fillers  : 
barytes  (BaSOJ,  **blanc  Fixe**  (artificial  barytes),  gypsum 
(CaSO,),  strontium  white  (SrSOJ,  whiting  (CaCO,),  China 
clay  (kaolin),  and  magnesite  (MgCO,). 

Yellow  and  Orange  Pigments, — Chrome  yellow  (2PbCrOJ, 
Chinese  yellow  (PbO.PbCrOJ,  zinc  chrome  (ZnCrOJ,  realgar 
(As,S,),  ''cadmium  yellow*'  (CdS).  **  King's  yellow**  (As,S,), 
yellow  ochre  (Fe,0„  A1,0„  SiO„  etc.),  and  Siennas  (Fe,0„ 
H,0,  Mn,0,)- 



Green  Pigments, — Chrome  green  (Cr,0,),  copper  green  (CuA), 
mineral  green  (malachite),  cobalt  green  (ZnO,  CoO,  P.O.,  etc.), 
manganese  green  (BaO,  MnO,,  etc.),  emerald  green  (**  Paris 
green, '*  7Cu2C,H,0,,  3CuAs,0J  and  Brunswick  green  (com- 
pounded of  barytes,  chrome  yellow,  Prussian  blue,  etc.). 

Black  Pigments, — Lampblack  (carbon),  bone-black  (carbon 
and  Ca.HPOJ,  vegetable  black,  Frankfort  black,  coal-tar  black, 
asphaltum  black,  and  graphite  black  (C).* 

Blue  Pigments, — Ultramarine  (SiO„  A1,0,,  Na,0,  S),  Prus- 
sian blue,  Chinese  blue,  or  Brunsw^ick  blue,  (Fe^Cj^N,,),  cobalt 
blue  or  smalts  (A1,0„  CoO),  Bremen  blue  (CuH,0,),  and  cop- 
per blue  (CuO,  CO,,  H,0). 

The  various  colored  lakes,  carmines,  aniline  lakes,  etc.,  have 
but  a  limited  application  in  Engineering  Chemistry.  Their 
methods  of  manufacture  and  assay  can  be  advantageously 
studied  by  reference  to  ''Painters*  Colors,  Oils,  and  Varnishes," 
by  George  H.  Hurst,  F.C.S.,  London,  1892,  pp.  249-282. 

The  analysis  of  a  white  paint,  ground  in  oil,  as  shown  in  the 
Scheme  on  page  455,  will  indicate  the  method  of  procedure  in 
analyses  of  this  character.  Where  qualitative  analysis  has 
shown  the  presence  of  a  few  constituents  onlj',  the  Scheme  can  be 
correspondingly  modified. 

"^  The  American  Engineer  and  Railroad  Journal ,  Nov.  1896.  p.  315,  states:  "Graphite 
mixed  with  an  oil  is  chemically  inert  and  in  drying.forms  a  coat  that  adheres  firmly  to 
the  metal  surface.  Its  resistance  to  the  action  of  acids  and  alkalies  has  been  proven  by 
numerous  tests  much  more  severe  than  the  conditions  of  service,  and  its  resistance  to 
the  penetrations  of  moisture  have  been  equally  satisfactory.  Heat  does  not  cause  it  to 
blister,  and  we  are  informed  that  steel  chimneys  painted  with  it  have  been  heated  to 
redness  without  decolorizintj:  the  paint.  The  paint  has  been  used  with  success  upon  the 
hulls  and  decks  of  steel  steamers,  and  there  seem  to  be  no  conditions  of  service  which 
it  does  not  successfully  meet." 

Ill  Hi 


si  =3 

s  I 


j5|l°  S 


^Spg'?  CO 

s  =  a  ■"  s. 


Analysis  of  White  Lead  Paints.^ 
(Dry,  not  ground  in  oil.) 

1.                  2.                  3'                  4-  5-                 6. 

PbO 86.35        85.93        83.77        84-42  86.5        86.24 

CO,.- 10.44        11.89        15.06        14.45  i>'3        11'^ 

H,0 2.95          2.01          i.oi          1.36  2.2          1.61 

Total ....  99.74        99.83        99.84      100.23  loo.o       99-53 

from  which  the  composition  of  the  white  leads  can  be  calculated 
to  be  : 

I.                  3.                  3.                 4-  S«                   ^ 

PbCO, 63.35        72.15        91.21        87.42  68.36        70.87 

PbHjO, 36.14        27.68          8.21        12.33  31.64        28.66 

Moisture  >.    0.25         ....  0.42         0.48 

• .  • . 

Total....  99.74       99*83       99.84      100.23    100.00       99.53 
No.  I.  English  make.     Made  by  Dutch  process ;  of  very  good  quality. 

No.  3.  Krems  white.  Made  by  precipitation  with  carbon  dioxide.  It  is 
deficient  in  body,  although  of  good  color. 

No.  4.  German  make.  Precipitated  by  carbon  dioxide ;  of  good  color, 
but  deficient  in  body. 

No.  5.  German  make.     Made  by  Dutch  process ;  a  good  white. 

No.  6.  German  make.  Made  by  precipitation  with  carbon  dioxide ; 
quality  fair. 

Lead  white,  ground  in  oil,  is  a  common  form  in  the  market. 
It  usually  contains  about  eight  per  cent,  of  raw  linseed  oil,  and 
has  an  extended  use  among  painters,  as  it  readily  mixes  with 
additional  oil  and  turpentine  to  form  liquid  paint. 

The  brown  pigments,  composed  principally  of  oxides  of  iron 
and  manganese,  can  be  analyzed  by  Scheme  XIII,  p.  29  ;  the 
yellows  and  greens  containing  chromium  require  a  special  pro- 
cess, as  follows  : 

1  Painters'  Colors,  Oils  and  Varnishes.    By  G.  H.  Hurst,  F.C.S.,  p„-3a 












S^    CM 

rt       W 

a  5 

S  ' 



^  *p 


0)  & 

s  i 

A    ts 


4.  Fl 


t»  1 







»  — 

^  •  r* 

^    & 

•n    »♦  O 

li'.    B) 
n     P 

r^    P« 

•      B 








*g  Q 


1?  " 


►n  ^ 


5-  ^* 



I  ^ 




2.  S 


2!  -. 

r»    3 


2    »* 


P«  r» 


<=^*  S 


B    w 


N'.      S 



ce  n 
of  Ra 


2p  G(S  ?  , 


8  '^o-S  ;ri 

jB  2  o*  o  o- 
r-  ?  V,!  B  p.  ' 


s  S 

Oq  CL'  S. 

S  "  I 

•n  B  g 

8  S  S. 

S  2  o 

«  o  5 

8  S  « 

O  »  'S 

r  o  s 

B  pa 

5  5^ 

2.  o  t^ 

>»  o  *» 

2  B*  '^ 

2  ^  ^ 

B  ►i     - 

p^  O 










3  o 

•O    B 


^    &  B* 
pa    A  v^ 

•    3  I 
B  S 

5   •^ 
B  ?5- 

0«Q  v< 

B     • 
*^»  B^ 

P*  • 

^     n 
»«•  D 

Oq    S* 

B*   P 
S   ^ 

CO  ^ 

-^  IS 




B      (g 

^^   •.     B 

C  O 

g   O  ^ 

rt      to* 

(0    p« 





o  o 

«.  C0 

'^  2 

8  * 


P.   ^ 































B     *^ 

(g      P. 
2.    CO 

2.  c 

1p  ^ 

r^    P<  » 

B*    P*  « 

B    B>  P* 

O   ^  B 

B*   rj-  P* 

2.  *^  B- 

P.  <  R 

2.  5 

pa    •< 

!«*•      M 

s : 


p.  rt 


r  B 

B    - 

o  i 

to    *" 

S*  B* 
o*  o 

B    ^ 
»    P 

r»    O 

r«  p 


2    to 

P     B* 


B     C  r^ 

S  2. 

p  - 

^  2 

r*  B 

p  3 

(fi  — 

p  ?^ 

3  H 

(0  C 

<»  B 

C  « 

'^  2 

»i  P 

R  B 

o  -*• 

a  e 

^  p 

«  «> 

o  B 

n»  » 

2  ^ 

P«  •I 

p  B^ 

p<  o 











§  i 

O  o. 












Occasionally  the  following  determinations  are  made  : 

Water. — Hygroscopic.  Heat  one-half  gram  at  105**  C.  in  an 
air-bath  to  constant  weight. 

Volatile  Matter. — Heat  one  gram  in  a  porcelain  crucible  to 
low  redness ;  loss,  less  water,  is  volatile  matter. 

Water  Extract. — (Acetates,  sulphates,  bichromates,  or  nitrates, 
indicating  imperfect  washing  in  manufacture.)  Treat  three 
grams  with  six  successive  portions  of  twenty-five  cc.  each,  of  cold 
distilled  water,  decanting  and  filtering  each  time,  and  evaporate 
the  filtrate  in  a  platinum  dish  to  dryness  on  a  water-bath. 

Analysis  of  Mixed  Chromate^  Sulphate  and  Carbonate  of  Lead. 
(Lemon,  Chrome  and  White  Lead.) 

Analysis  made  same  as  in  scheme  for  Lemon  Chrome ;  excess 
of  lead  is  to  be  calculated  to  white  lead,  2PbC0,+PbH,0,. 

Analysis  of  Red  Chromate  of  Lead. ^ 

For  the  lead  determination  take  one  gram  in  a  covered  casser- 
ole, add  twenty-five  cc.  concentrated  nitric  acid,  heat  to  boil- 
ing, and  while  boiliyig  add  half  a  dozen  drops,  one  at  a  time,  of 
alcohol,  by  means  of  a  pipette  ;  boil  a  while  longer,  add  water, 
and  all  of  the  chromate,  if  it  is  pure,  will  be  found  in  solution. 

Without  this  alcohol  treatment  great  difficulty  will  be  experi- 
enced in  getting  the  chromate  into  solution ;  with  it,  it  becomes 
very  easy.  Add  twenty-five  cc.  concentrated  sulphuric  acid, 
evaporate  to  white  fumes  and  complete  the  analysis  as  described. 
For  chromium  and  sulphur  trioxide  determinations,  boil  off  alco- 
hol and  proceed  as  previously  directed. 

1  Known  by  various  names,  as  scarlet,  dark  or  basic  chromate  of  lead,  chrome  red, 
Chinese  red,  American  vermilion  and  vermilion  substitute.  Formula :  aPbO.CrOt  or 
PbCr04  -H  PbO. 

■     Bg*B 

off-     «^2!fl 

g.  St  Q!Q   ^.  O    ►id 

*^  CThrt   O   CO 

-1  '•^   a* 

CO*  -.  S  ?  5* 

w  CJ  ST  w  '^  2 


P    g 

itf  o 

•      rt 





^  c  c 


-  < 

o  o 

-  » 


►t    C0 


00         O* 
a,  CO  ►i 

o  o*?r 



•-  **  O  e^ 

r*  f*  ►— Wi 

0  ^^? 

R  **  ET  ^ 
^^  O  »«•  CO 



O  ®  - 

CO  S 
O  o>  ® 

SB  O 

r«  crof 


SO  ^  o* 

BlJ.  rt  ^ 

P    »♦  CO 


f*    ^!f    CO 


r*         Ji  r*  «   »i 

*-•  —  ?' 

B»   r*'  P 

Pg  crcr 
-    □  r»  (V 





2  '^ 

2  <^ 

(0    ►- 


P   P 

2.  rt 

« a 

s-^  • 

^  p  ^ 

p  p  8 

S  ^  B 

So^  I 

CO    jj.    I 

o  S  :i- 

w     p 
"      CO 

3  B'co  O  d 

B  P  P  •^SLq* 
•B  »»  B  ^p^*^^ 

g  B  »  •  B*iS  I 

S?3  3::lB? 


•  -I « I 

p  Oo 
2  *  ffi^ 

^   B 

£  S  O  *  M'^Tl 
5  a  B  n^'^'-S 
CUB  «  L/i^-    ^ 

^q^g  095*3 

Sg^l?  to  ff 

<>  5p  -^  B-  ►u 
o  o>^  o  rt  "^ 

-•  •^^y-     CO  O 
•-K^  r»  p  •    P 
B  ?^  CO  _B 

O  «  B*  C  g. 

p  rt  rt  CO  —  ►U 

O  P  '^  3  B^" 

2.JioQ  2  w  o- 

-    *  g;^  o  o 

g-jz:---   B5==-* 
2.P  »  <- 

^  QB-rt  B  B5 

»^I*B  P%> 
p  »  "    o  B* 

•^  B  M  rS  p. 




P  s 


p   I 

(0  >:y 
CO   -^ 

ffi  p 
•*    3- 

P  3 
f^  B 

o  SL 

B  ?• 


•--  r*- 

CO  33 




r»  2 
B  ^ 






(0  o 


B  "< 

^  P 

r»  O 




B  g 




IT    _^ 

o  P 














B*    D 

o   SL 
B  v; 

P     CO 

r*  "-I 

p'  O 
5  « 

rt    • 






ft  A 

1  * 









Chrome  green,  in  which  the  coloring  matter  is  Cr,0„  is  sel- 
dom found  in  the  market  pure.  Usually  it  contains  from  twenty 
per  cent,  to  seventy-five  per  cent,  of  barium  sulphate. 

As  an  example  of  specifications  for  a  compound  chrome  paint, 
the  following  is  given  : 

Pennsylvania  Railroad  Company.    Motive  Power  Department. 

Specifications  for  Cabin  Car  Color, 

The  standard  cabin  car  color  is  the  pigment  known  as  scarlet  lead 
chromate.  It  is  always  purchased  dry.  The  material  desired  under  this 
specification  is  the  basic  chromate  of  lead  (PbCr04PbO),  rendered  brilliant 
by  treatment  with  sulphuric  acid  and  as  free  as  possible  from  all  other 

The  theoretical  composition  of  basic  lead  chromate  is  nearly  59.2  per  cent, 
of  the  normal  lead  chromate,  and  40.8  per  cent,  of  lead  oxide,  but  in  the 
commercial  article  it  is  found  that  a  portion  of  the  sulphuric  acid  added 
to  brighten  the  color  remains  in  combination  apparently  with  the  normal 
lea'd  chromate,  slightly  increasing  the  percentage  of  this  constituent. 

Samples  showing  standard  shade  will  be  furnished  on  application,  and 
shipments  must  not  be  less  brilliant  than  sample.  The  comparison  of 
sample  from  shipment  with  the  standard  shade,  may  be  made  either  dry 
or  by  mixing  both  samples  with  oil. 

Shipments  of  cabin  car  color  will  not  be  accepted  which 

1.  Contain  barytes  or  any  other  adulterant. 

2.  Show  on  analysis  less  than  fifty-seven  per  cent,  or  more  than  sixty 
per  cent,  of  normal  lead  chromate,  including  the  sulphuric  acid  combined 
as  above  stated. 

3.  Show  on  analysis  less  than  thirty-eight  per  cent,  or  more  than  forty- 
two  percent,  lead  oxide,  in  addition  to  the  lead  oxide  in  the  normal  lead 

4.  Vary  from  standard  shade. 

Office  oj  Gen,  Supi .  0/  Motive  Power ^  Attoona,  Pa,,  Feb,  18,  i8gi. 

The  various  red  paints,  Indian  red,  Tuscan  red,  and  other 
iron  oxides,  etc.,  used  in  general  practice  are  rarely  pure,  but 
contain  added  amounts  of  finely  pulverized  gypsum  and  calcium 
carbonate.  These  oxides,  when  properly  ground  and  mixed 
with  linseed  oil,  form  paints  that  cannot  be  excelled  for  dur- 
ability, permanence  of  color  and  cheapness.  Many  of  the  mix- 
tures contain  varying  amounts  of  japans,  but  as  the  japans  have 
been  subject  to  great  sophistications  of  late  years,  specifications 
now  generally  call  for  linseed  oil,  turpentine  and  pigment  only. 




Thus,  two  varieties  of  paints  might  be  roughly  classified  as  ; 
(1)   Paints  for  wood  work,  and  (2)  paints  for  iron  work. 

The  following  specifications  refer  to  class  1  ; 

Pennsylvania  Railroad  Company.    Motive  Power  Department. 
Specifications  for  Freight  Car  Color. 

Freight  car  color  will  be  bought  in  the  paste  form,  ami  the  paste  must 
contain  nothing  but  oil,  pigment  and  moisture. 

The  proportions  of  oil  and  pigment  must  be  so  nearly  as  possible  as  fol- 

Pigment  seventy-live  per  cent,  by  weight. 

Oil  twenty-five  percent,  by  weight. 

The  oil  must  be  pure  raw  linseed  oil,  well  clarified  by  settling  and  age. 
New  process  oil  is  preferred.  The  pigment  desired  contains  not  over 
one-half  per  cent,  of  hygroscopic  moisture,  and  has  the  following  compo- 

SesquioKide  of  iron,  fifty  per  cent,  by  weight. 

Fully  hydrated  calcium  sulphate  or  gypsum,   forty-five  per  cent,  by 

Calcium  carbonate,  five  percent,  by  weight. 

Samples  of  standard  pigment  showing  shade  will  be  furnished,  and 
Bhipments  will  be  required  to  conform  with  the  standard. 

The  shade  of  paint  being  afiected  by  the  grinding,  the  Pennsylvania 
Railroad  standard  shade  is  that  given  by  the  dry  sample  sent,  mixed  with 
the  proper  amount  of  oil  and  ground,  or  better  rubbed  up  in  a  small 
raortar  with  pestle  until  the  paste  will  pass  Pennsylvania  Railroad  test  for 
fine  grinding.  It  is  best  to  use  fresh  samples  of  the  dry  pigment  for  each 
day's  testing.  The  comparison  should  always  be  made  with  the  fresh 
material,  and  never  with  the  paint  after  it  haa  become  dry.  The  com- 
parison is  easiest  made  by  putting  a  small  hillock  of  the  standard  paste 
and  of  that  to  be  compared  near  each  other  on  glass,  and  then  laying  an- 
other piece  of  glass  on  the  two  hillocks,  and  pressing  them  together  until 
the  two  samples  unite.  The  line  where  the  two  samples  unite  is  clearly 
marked  if  they  are  not  the  same  shade.  Tbe  paste  must  be  so  finely 
ground  that  when  a  sample  of  it  is  miied  with  half  its  weight  of  pure  raw 
linseed  oil,  and  a  small  amount  of  the  mixture  placed  on  a  piece  of  dry  glass, 
there  will  be  no  separation  of  the  oil  (rom  the  pigment  for  at  least  half 
a.n  hour.  The  temperature  afiects  this  test,  and  it  should  always  be  made 
at  70"  P.  The  sample  under  test  runs  down  the  glass  in  a  narrow  stream 
when  it  is  placed  vertical,  and  it  is  sufficient  if  the  oil  and  pigment  do  not 
■eparate  for  an  inch  down  from  the  top  of  the  test. 

Shipments  will  not  be  accepted  which 

1.  Contain  less  than  twenty-three  per  cent,  or  more  than  twenty-seven 
per  cent,  of  oil. 

3.  Contain  moTC  than  two  per  cent,  of   volatile   matter,   the  oil  being 



dried  at  250^  P.,  and  the  pigment  dried  in  air  not  saturated  with  moisture 
at  from  60°  to  90°  F. 

3.  Contain  impure  or  boiled  linseed  oil. 

4.  Contain  in  the  pigment  calcium  sulphate  not  fully  hydrated,  less 
than  forty  per  cent,  of  sesquioxide  of  iron,  less  than  two  percent,  or  more 
than  five  per  cent,  calcium  carbonate,  or  have  present  any  barytes,  ani- 
line colors,  lakes,  or  any  other  organic  coloring  matter,  or  any  caustic 
substances,  or  any  makeweight  or  inert  material  which  is  less  opaque 
than  calcium  sulphate. 

5.  Varying  from  shade. 

6.  Are  not  ground  finely  enough. 

7.  Are  a  **  liver  **  or  so  stiff  when  received  that  they  will  not  readily 
mix  for  spreading. 

Altoona,  Pa,^  Office  Supt.  Motive  Power. 

As  an  example  of  the  composition  of  a  paint  for  iron  surfaces 
(Class  2)  the  following  mixture  as  used  for  painting  the  structure 
of  the  elevated  railroads  in  New  York  City  is  given  :' 

Boiled  linseed  oil 9    parts. 

Red  oxide  of  iron  finely  ground ^\    ** 

Turpentine 1    part. 

In  mixing  paints  for  iron  stuiaces,  it  is  of  the  first  importance 
that  only  the  best  materials  be  used.  Linseed  oil  is  the  best 
medium,  when  free  from  admixture  with  much  turpentine. 

The  large  percentage  of  linolein  formed  in  drying,  makes  the 
surface  of  the  paint  solid  and  of  a  resinous  appearance,  possess- 
ing toughness  and  elasticity.  Linseed  oil  does  not  crack  or 
blister,  by  reason  of  the  expansion  and  contraction  of  the  iron 
with  variation  of  temperature.  Another  important  characterise 
tic  is  its  expansion  while  dr>  ing,  which  adapts  it  to  iron  sur- 
faces. The  Metropolitan  Elevated  Railway  Company  experi- 
mented ver>*  thoroughly  with  the  various  kinds  and  colors  of 
paints :  their  labors  at  last  culminated  in  the  selection  of  a 
metallic  paint  for  the  first  coat  (formula  given  above)  and 
a  white  lead  paint  for  the  second  and  last  coat,  both  paints  to 
contain  the  best  linseed  oil  and  enough  turpentine  to  make  the 
paints  cover  well  and  facilitate  their  dr>-ing. 

The  formula  of  the  white  lead  paint  as  used  is  here  given  : 

1  On  the  Ccnstmctioci  of  the  Second  Arrnne  Line  ol  the  Mciropolitnn  Blernted  Rail- 
war  of  New  York.    Br  G.  Thomas  Hall.  CX..  TVvaj.  Am.  Sk.  CimiEa^.,  m^  xjol 

paint  analysis. 

White  Lead  Paint.    Olive  Color. 

147.43  Icilogratns  white  lead. 


n  blue. 

0.45  '■  burnt  umher. 

79.50  liters  boiled  Unseed  oil. 
S.67     "     turpentine. 
3.79      "      liquid  drier  (boiled  linseed  oil  and  leadoride). 

Some  engineers  prefer  red  lead  instead  of  iron  oxide  as  the 
pigment  for  paints  to  be  used  for  iron  structures. 

G.  Bouscaren,  C.E.,  states  with  regard  to  the  painting  of 
bridges,  that  having  used  both  varieties  of  paint,  be  gives  prefer- 
ence to  the  red  lead.' 

The  red  lead  paint  adheres  better  to  the  iron  and  fails  princi- 
pally by  wear  and  a  gradual  transformation  of  the  red  lead  into 
carbonate,  whilst  the  iron  paint  fails  by  scaling. 
Asphalt  Paint. 

Until  within  quite  recent  years  little  has  been  known  in  this 
country  of  the  valuable  properties  of  the  asphalt.  In  the  popu- 
lar mind  it  is  often  confused  with  certain  coal-tar  products, 
which,  though  similar  in  appearance,  differ  essentially  from 
asphalt  in  character.  Asphalt  oils  are  of  a  nearly  non-volatile 
nature,  and  are  therefore  permanent,  while  on  the  other  hand, 
coal-tar  is  volatile. 

The  so-called  asphalt  paints  which  have  been  used  in  the 
past  are  such  only  in  name.  They  contain,  at  best,  but  a  very 
small  per  cent,  of  asphalt,  which  is  incorporated  in  the  form  of 
a  pigment  and  which  serves  no  valuable  purpose.  Asphalt,  on 
the  contrary,  should  be  the  main  constituent,  since  the  value  of 
such  a  paint  depends  upon  the  presence  of  the  permanent 
asphalt  oils.' 

Fire  Proof  Paints.  Silicate  Paints,  Asbestos  Paints,  etc. 

The  principle  of  action  of  these  paints  is  not  to  render  wood 
{  work  or  similar  material  fire  proof,  but  to  retard  combustion. 

Wood  treated  with  a  solution  of  zinc  chloride,  or  with  a  solu- 


tion  of  sodium  silicate,  can  be  rendered  nearly  non-inflammable, 
and  after  such  treatment  and  drying,  paint  can  be  applied. 

Instead  of  using  the  ordinary  paints  for  this  purpose,  various 
compounds  are  incorporated  in  the  paint  itself  to  render  the  lat- 
ter non-inflammable.  Thus  the  preparation  of  Prof.  Abel  J. 
Martin,  of  Paris,  is  as  follows : 

Boracic  acid,  borax,  soluble  cream  of  tartar,  ammonium  sul- 
phate, potassium  oxalate,  and  glycerine  mixed  with  glue  and 
incorporated  with  a  paint.  It  is  the  result  obtained  after  many 
experiments  in  response  to  a  prize  of  looo  francs,  offered  by  the 
Society  for  the  Advancement  of  National  Industry  of  France. 
A  committee  consisting  of  Professors  Dumas,  Palaird  and  Troost, 
after  testing  the  materials,  consisting  of  painted  woods  and 
various  fabrics,  for  seven  months,  reported  in  favor  of  this  prepa- 
ration. The  municipalit>'  of  Paris  made  its  use  obligatory  in 
all  of  the  theatres  there  and  it  has  stood  the  test  of  the  last  six  years. 

Blue  Pigments.  Ultramarine  beinga  silicate,  can  be  analyzed 
by  Scheme  XIV,  page  37. 


SiO, 49-68  per  cent. 

A1,0, 23.00 

S 923 

SOj 2.46 

Na-0 12.50 

H-6 3.13 



Total loooo        " 

Composition  of  Commercial  Prussian  Blub. 

A1,0, 2.45  percent. 

Fe*0, 3.  ^i 

CaSO,-hSiO 89.86 

CN4-S 4- II 

H5O 0:27 


Total 100.00        *' 

CoMPv^siTiox  OF  Smalts. 

SiO 56.4  percent. 

AUO 3.5 

Fe.O 4.1 

Cob 16.0 

CaO 1-6 

K,0 13.2 

PbO 4-7 

TotAl 99-5 



Examination  of  Ike  Oil  after  Extraction  from  the  Paint. 

The  only  adulterants  used  in  linseed  oil,  in  this  connection, 
are  mineral  oil  and  rosin  oil.  Their  method  of  detection  and 
estimation  is  given  on  page  414. 

Turpentine,  when  present,  is  not  an  adulterant,  and  a  mixture, 
extracted  from  a  paint,  may  contain  linseed  oil,  mineral  oil, 
rosin  oil,  turpentine,  and  rosin  spirit.'  The  latter  is  quite  dis- 
tinct from  rosin  oil  and  when  properly  prepared  is  a  perfect  sub- 
stitute for  turpentine.  If  the  liquid  extracted  from  the  paint  is 
a  mixture  of  linseed  oil,  turpentine,  and  rosin  spirit  the  deter- 
mination of  the  amounts  of  each  is  somewhat  difficult. 

Turpentine  can  be  distinguished  and  determined  in  the  pres- 
ence of  rosin  spirit  by  the  action  of  the  former  on  polarized 
light,  rosin  spirit  being  inert.  Thus:  the  specific  rotation  of 
American  turpentine  varies  between  +8.8  to  +21,5.  The  bro- 
mine absorption  is  also  an  indication  : 

The  bromine  absorption  of  turpentine  varies  between  203  and  2  j6. 

The  bromine  absorption  of  rosin  spirit  varies  between  1 84  and  200. 

The  determination  of  the  amounts  of  petroleum  naphtha  and 
turpentine  in  a  mixture  can  be  made  by  the  method  of  H.  E. 
Armstrong,/.  Soc.  Client.  Ind.,  i,  480;  consult  also  Allen,  Com. 
Org.  Anal.,  2,  48-50. 

References .-  "  How  to  Design  a.  Paint."  By  C-  B.  Dudley,  Railway  and 
Eng.J.,f>S.  174,  318. 

On  the  Analysis  of  White  Paint."   By  G.  W.  Thompson./.  Soc.  Chem. 
'.,  rs.43J. 

"  Detection  of  Rosin  and  Roain  Oil  in  Oils  and  Varnishes."  By  F. 
Ulier,  Ibid.  15.  381. 

'■Technical  Analysis  of  Asphaltum."  By  L.  A.  Linton,  /  Am.  Chem. 
Soc..  16,  809. 

"  Rustless  Coatings  for  Iron  and  Steel.  Galvanizing,  Electro- Chemical 
Trealment.  Painting  and  Other  Preservative  Methods."  By  M.  P.  Wood, 
TraHi.  Am.  Soc.  Mec.  Eng.,  16,  /ipj,  350-450. 

'■  Preservative  Coatings  for  Iron  Work."  By  A.  H,  Sabin  and  A.  O. 
Powell,  Engineering  News,  Feb.  5.  1895,  p.  86. 

"  ChemischeOperattoneuder  Analyse  von  FarbstoSen."By  P.  Schmidt, 
Mitth.  Ualerei,  g,  iii. 

"  Chemistry  of  Paints  and  Fainting."   By  A.  H.  Church.  London.  1391. 

"  Figments,  Paints  and  Painting."    By  G.  Terry.     London.  1893. 

IConlUr  naphtha  ] 




Pyrometry,  or  the  art  of  measuring  high  temperatures,  has 
received,  in  the  past  few  years,  considerable  attention  from  en- 
gineers and  metallurgists. 

This  is  especially  so  in  the  direction  of  metallurgical  engi- 
neering, where  more  uniform  methods  of  heating  and  controlling 
heat  have  developed.  In  many  processes  of  melting,  refining, 
tempering,  etc.,  certain  temperatures  are  required,  from  which, 
should  much  variation  occur,  the  products  would  be  ruined. 

Many  forms  of  pyrometers  have  been  invented,  but  only  a  very 
few  have  accomplished  their  purpose.  Many  are  admirable  in 
design  and  construction,  and  prove  accurate  and  trustworthy  in 
the  laboratory,  but  fail  utterly  when  applied  in  practice  at  high 

Pyrometers  may  be  classified  according  to  the  principles  upon 
which  they  operate  :* 

1 .  Expansion  of  mercury  in  a  glass  tube.  When  the  space 
above  the  mercury  is  filled  with  compressed  nitrogen  and  a 
specially  hard  glass  is  used  for  the  tube,  mercury  thermometers 
can  indicate  correct  temperatures  to  1,000**  F. 

2.  Contraction  of  clay,  as  the  Wedgewood  pyrometers;  very 
inaccurate  as  the  contraction  of  the  clay  varies  with  the  compo- 
sition of  the  clay. 

3.  Expansion  of  air,  as  in  the  air  thermometer,  Siegert's 
pyrometer,  \Viborgh*s  pyrometer,  Uehling  &  Steinbart's  pyrom- 
eter, etc. 

4.  Pressure  of  vapors,  as  the  Spannung*s  pyrometer,  or  the 
Bristol  recording  thermometer. 

5.  Relative  expansion  of  two  metals,  as  Brown's  or  Buckley's 

6.  Specific  heat  of  solids,  as  iron-ball,  copper-ball,  or  plati- 
num-ball pyrometer. 

7.  Melting-point  of  metals,  alloys,  etc. 

8.  Time  required  to  heat  a  weighed  quantity  of  water — a 
water  pyrometer. 

1  Eni^imtfring  News.  34.  322  (Nov.  14,  1895). 


5.  Increase  in  temperature  of  a  stream  ot  water  or  other 
liquid  flowing  at  a  given  rate  through  a  tube  inserted  into  the 
heated  chamber,  as  the  Saintignon  pyrometer. 

10.  Changes  in  the  electric  resistance  of  platinum  or  other 
metal,  as  in  the  Siemen's  pyrometer. 

11.  Measurement  of  an  electric  current  produced  by  heating 
the  junction  of  two  metals,  as  in  the  Le  Chatelier  pyrometer. 

12.  Dilution  of  a  stream  of  hot  air  or  gas  flowing  from  a 
heated  chamber  by  cold 
air,  and  determination 
of  the  temperature  of  the 
mixture  hy  a  mercury 
thermometer,  as  in  Hob- 
son's  hot-blast  pyrome- 

13.  Polarization  and 
refraction,  by  prisms  and 
plates,  of  light  radiated 
from  heated  surfaces,  as 
in  Mesure  and  Noel's  op- 
tical pyrometer. 

The  standard  of  refer- 
ence for  all  temperatures 
above  ai3°  F,  is  the  air 
thermometer  and  all  py- 
rometers are  usually 
standardized  by  compar- 
ison with  it. 

The  various  forms  of 
air  theniiometers  are  de- 
scribed by  Prof.  R.  C. 
Carpenter  in  Engineer- 
ing Nfws,  Jan.  5,  1H93. 

The  air  pyrometer  of  ""«  '*"■ 

Messrs.  A.  Siegert  and  W,  Duerr,  Fig.  146,  consists  of  a  porce- 
lain cylinder  connected  by  a  thin  copper  tube  with  the  meas- 
uring portion  of  the  apparatus.  This  consists  of  a  bell  of  sheet 
I  brass,  the  lower  edge  of  which  dips  into   a   bath  of  petroleum. 



The  bell  is  attached  to  one  arm  of  a  balance  beam,  a  counter- 
poise being  carried  by  the  other  arm. 

The  porcelain  bulb  being  heated  to  the  temperature  to  be 
measured,  the  air  it  contains  expanding  enters  the  brass  bell, 
lifting  this  and  moving  the  beam.  The  movement  is  shown  on 
a  scale  and  the  temperature  read  direct  from  the  divisions,  into 
which  the  scale  is  divided.  (Consult /iwr.  Iron  and  Steel  Insl., 
1893,  p.  340- 

Wiborgh's  air  pyrometer  is  fully  described  in  Tram.  Am.  Inst. 
Mining  Eng.,  ai,  p.  59a. 

Hobson's  hot-blast  pyrometer  is  largely  used  for  measuring 

the  temperature  of  the  blast  in  hot-blast  iron  furnaces.  It  con- 
sists of  a  brass  chamberhavingthreearmsanda  handle,  Fig.  147. 
An  opening  through  a  jet  in  one  of  the  arras  admits  the  hot 
blast,  another  arm  admits  atmospheric  air,  while  the  third  arm 
is  for  the  discharge  of  the  mixture.  To  this  third  arm  is  at- 
tached a  thermometer  which  indicates  the  temperature  of  the 
mixed  current.  A  thermometer  is  also  attached  to  the  arm  ad- 
mitting the  atmospheric  air. 


This  pyrometer  can  be  used  constantly  to  2000°  P.,  without 
danger  of  injury, 

Bristol's  recording  thermometer  gives  a  continuous  graphical 
record  of  temperatures  up  to  600°  F.  It  consists  of  a  copper 
coil  which  takes  the  place  of  a  bulb  and  is  inserted  in  the 
heated  space.  The  bulb  is  partly  filled  with  alcohol,  which 
is  partly  vaporized  by  the  heat.  The  pressure  of  the 
vapor  is  transmitted  through  a  fine  flexible  copper  tube,  filled 
with  alcohol,  to  any  convenient  distance  not  exceeding  twenty- 
five  feet,  where  it  is  measured  by  a  recording  pressure  gauge. 
The  interior  of  the  gauge  contains  a  flat  Bourdon  spring  coiled 
into  three  complete  coils.  The  movable  end  of  the  spring  car- 
ries a  pointer,  which  contains  an  inking  pencil  at  its  outer  end. 
The  clock-work  revolves  a  paper  chart  once  in  twenty-four 
hours,  and  the  marker  thus  makes  a  continuous  record. 

The  metallic  pyrometer  of  Brown  or  Bulkley's  form  consists 
o(  the  well-known  copper  and  iron  tube,  and  is  based  on  the 
principle  of  the  difference  of  expansion  between  copper  and  iron. 
An  iron  tube  is