Field manual for engineers

Survival, Water, Medical Field Manuals

Military Manuals

Philbrick, P. H. (Philetus Harvey)

Document text

REESE>LIBRARY 
/ 

t 

UNIVERSITYTDF  CALIFORNIA. 

* 

ci(ecevved 


,  Occasion  No. 

Sj  —  imi  — 


.92418   •   C/jss 


r-u-u-ll1 


FIELD    MANUAL 


FOR 


ENGINEERS. 


BY 

PHILETUS    H.    PHILBRICK,   C.E.,   M.S. 

M.   AM.  MATH.   Soc., 
Chief  Engineer,  Kansas  City,   Wat\ins  and  Gulf  Railway, 

.A'.  Am.  Land  and  Timber  Co.,  etc.,  etc.; 

Sometime  Professor  of  Civil  Engineering  at  the 

State  University  of  Iowa. 


FIRST   EDITION. 
FIRST   THOUSAND. 

: 


extended,    and    extra    topi*, 
,10r  clearness,  considered  in  connection  with  the 
icrai  matter  of  the  text.  ..xpi<. 

Table  II  dispenses  with  all  calculation  Dy  numerous  readings 

the  last  figure. 
\j  &  4-  .L>aces,  has  bden  in  prepara- 


IV  PREFACE. 

as  shown  in  Chapter  IV;  Table  VII  greatly  simplifies  the 
finding  of  the  tangent  and  the  external  of  any  curve;  and 
Tables  XVIII  and  XIX,  and  some  others  to  which  the  remark 
can  apply,  are,  it  is  believed,  in  terms  of  the  proper  arguments 
and  in  the  best  form. 

In  Chapter  IV  the  laws  of  errors  in  field-work  are  demon- 
strated and  illustrated;  and  the  best  method  of  conducting  a 
preliminary  survey,  introduced  by  the  author  a  generation  ago, 
is  explained. 

In  Chapter  V  simple  and  exact  formulas  for  determining  the 
height  of  a  mountain  or  other  object  by  the  dip  of  the  horizon 
are  substituted  in  place  of  the  approximate  formulas  in  use. 

For  the  stadia,  as  well  as  for  the  telemeter,  new  formulas  are 
found  which  require  no  general  computation;  and  the  formula 
for  finding  the  proper  elevation  on  curves,  unlike  other  formulas, 
involves  no  large  factors. 

In  addition  to  a  very  general  treatment  of  Compound  Curves, 
Chapter  VI  includes  the  location  of  such  curves  of  any  number 
of  branches  (pp.  143-5),  as  well  as  easy  and  symmetrical 
formulas  for  finding  their  tangents  (p.  176). 

The  reader  interested  in  the  philosophy  of  mathematics  will 
find,  it  is  hoped,  an  elegant  and  fruitful  illustration  of  the  prin- 
ciples of  substitutions  in  determining  general  curves  to  fulfill 
required  conditions  on  pages  145,  146.  This  is  susceptible  of 
general  application.  There  has  also  been  added  a  general  treat- 
ment of  the  subject  of  Curves  tangent  to  Curves,  including  the 
"Wye  Problems";  and  also  that  of  Concentric  Curves  applic- 
able to  Parallel  Turnouts. 

The  finding  of  the  angles  between  the  rails  at  the  crossings 
of  curved  tracks  is  also  thought  to  be  a  valuable  addition. 

Chapter  VII  includes  a  variety  of  problems  in  Reversed  Curves. 
The  solution  of  Problems  IV,  V,  and  VI  were  first  given  to  the 
Senior  Class  (1869)  of  the  University  of  Michigan,  while  the 
author  was  in  temporary  charge. 

Chapter  VIII  treats  extensively  of  Turnouts.    The  distinction 

between  connecting  a  straight  line  and  a  track,  and  two  tracks, 

and  what  \^-^  iuired  in  each  case,  is  shown  on  pages  199,  200. 

of    certain    proposed    turnout    curves    is 

and  a  variety  of  methods  of  laying  out 


PREFACE.  V 

Chapter  IX  applies  entirely  to  the  author's  True  Transition 
Curve. 

Chapter  X  shows  that  the  formulas  for  the  computation  of 
earthwork  may  be  abridged  one  half  by  supposing  that  the 
side  slopes  are  produced  in  this  intersection.  A  general  relation 
between  the  "  end-area  "  volume  and  the  "  middle-area  "  volume 
is  shown  by  means  of  symbols,  and  new  formulas  are  given  for 
the  volume  of  a  frustrum  of  a  pyramid  and  the  frustrum  of  a 
cone.  A  formula  showing  the  true  correction  of  earthwork  for 
curvature  is  also  deduced,  and  the  best  method  of  computing 
earthwork  tables  explained  and  illustrated. 

Chapter  XI  contains  only  a  brief  exposition  of  the  subject 
of  approximate  and  abridged  computations,  which  it  is  thought 
may  serve  to  encourage  the  shortening  of  computations. 

Chapter  XII  describes  the  processes  incident  to  construction 
and  calls  attention  to  two  principles  that  aid  very  much  in 
"  staking  out  "  earthwork. 

The  logarithmic  tables  are  not  reproduced,  for  the  reason 
that  they  are  but  little  used  and  should  not  be  used  at  all — 
and  most  emphatically  so  in  this  line  of  work.  It  is  fair  to 
observe  that  the  space  required  for  such  tables  is  replaced  by 
numerous  useful  tables,  applying  directly  to  the  matter  in  hand, 
and  also  to  enlarging  the  subject-matter  of  the  book,  thus  sav- 
ing greatly  in  time  and  labor.  Furthermore,  it  should  be  stated 
that  there  is  no  problem  in  the  book  requiring  a  computation 
more  complex  than  to  find  the  cost  of  29  oranges  (say),  sup- 
posing that  17  oranges  cost  43  cents.  The  author  must  believe 
that  no  person — much  less  an  engineer — would  think  of  apply- 
ing logarithms  to  the  above  example;  and  if  so,  he  could  not 
with  any  propriety  apply  them  to  any  problem  in  the  Manual, 
since  the  nature  of  the  numerical  computation  to  be  made,  and 
not  the  subject-matter  of  the  problem,  furnishes  the  test  of 
methods. 

The  author  invites  criticism,  and,  should  another  edition  be 
called  for,  will  make  the  best  use  possible  of  any  suggestion 
that  may  in  good  faith  be  made  to  him. 

The  Tables,  all  but  three,  were  computed  expressly  for  this 
book,  and  scrupulous  care  has  been  taken,  by  numerous  readings 
and  checks,  to  make  all  tables  exact  to  the  last  figure. 

The  book,  as  indicated  in  a  few  places,  has  be"en  in  prepara- 


VI  PREFACE. 

tion  several  years  and  contains  matter  gathered  all  along  the 
paths  of  the  author's  experience;  and  is  the  result  of  a  belief, 
on  his  part,  of  his  ability  to  aid  his  professional  brethren  in  this 
direction. 

While  this  delay  has  not  been  to  the  advantage  of  the  author, 
it  lias  nevertheless  given  opportunity  for  due  reflection  and  re- 
consideration; and  therefore,  as  a  work  of  judgment  based  on 
experience,  the  volume  is  offered  to  his  brother  enquirers.  If 
the  book  even  partially  accomplishes  the  object  of  the  author's 
aims,  lie  will  feel  that  the  days  he  has  devoted  to  it,  though 
many,  have  not  been  spent  in  vain. 

P.  H.  PHILBKICK. 


CONTENTS. 


CHAPTER    I. 

PRELIMINARY  OPERATIONS. 

PAGE 

The    Reconnoissance 5 

The    Preliminary   Survey 5 

The  Location    5 

The   Organization   of  the  Transit    Party 8 

The  Compass:    What  kind  to  use  and  when  to  use  it 8 

Requirements  for  a  Successful  Reconnoissance 9 

Train    Resistances 10 

Total  Ascent  the  Main  Test  of  Gravity  Resistance n 

CHAPTER    II. 
ADJUSTMENTS,  USE,  AND  CARE  OF  INSTRUMENTS. 

The  Transit. 

Adjustments    12 

Use  and  Care  of  the  Transit 15 

Best    Way  to   Set  up   the   Instrument 15 

To  Measure  the  Angle  between  Two  Lines  or  Objects 16 

Hints  on  the  Care  of  Transits  and  Other  Instruments...                   .  16 


The  Level. 


Adjustments    

Use  of  the  Level. 


The  Compass. 

Adjustments    22 

Use   of   the    Compass 23 

vit 


Viii  CONTENTS. 

CHAPTER   III. 

PLANE  TRIGONOMETRY. 

PAGE 

Definitions   and   Explanations 25 

Fundamental    Relations 26 

Solution  of  Plane  Right  Triangles 31 

Table   for    Plane    Right   Triangles 32 

Fundamental    Relations   for    Oblique    Triangles 32 

Short  Solution  of  "  Tangent  Problem  " 34 

Solution  of  Plane  Oblique  Triangles 35 

Table  for  Solution  of  Oblique  Triangles 36 

List   of   Fundamental    Formulas 37 

CHAPTER    IV. 
SIMPLE  CURVES  CONNECTING  RIGHT  LINES. 

Properties  Relating  to  the  Circle 39 

Some  Elementary  Relations 41 

Notation 41 

Degree  of  Curve  Defined  and  Explained 41 

Rational  Treatment  of  Curves,  Example  Illustrating 42 

Difference  in  Lengths  of  Arcs  and  Subtended  Chords 43 

Huygens'  Formula  for  the  Length  of  an  Arc 44 

Table  Showing  Excess  of  Arcs  over  Subtended  Chords  when  the 

Arcs  are  Aliquot  Parts  of  100 44 

Table  Showing  Excess  of  Sub-chords  over  Aliquot  Parts  of  100, 

when  the  Chord  =  100 45 

Reason  for  these  Large  Excesses  Pointed  Out 46 

Formulas  for  Radius,  Tangent,  External  Secants,  Offsets,  etc 47 

Proper  Course  to  Pursue  in  Locating  a  Curve 49 

Long  Chords  and  Ordinates  to  Long  Chords 50 

Approximate  Value  of  Ordinates  to  Short  Chords 51 

Offsets  in  Terms  of  the  Degree  of  a  Curve 52 

Applications  of  Formula 53 

Laying  Out  Curves. 

A.  By  Deflection    Angles 53 

B.  By  Tangent   Offsets.     New   Method.     Without   Calculation 55 

C.  By  Ordinates   from   a   Long   Chord.     Without    Calculation 57 

D.  By  Chord  Offsets.     Without   Calculation 58 

E.  By  Middle    Ordinates.      Without    Calculation 59 

F.  By  Radial   Lines.     Without   Calculation 59 

Errors  in  Field-work — The  Nature  of. 

"A"   Method  of  Laying  Out 60 

"B"   Method   of  Laying  Out. 64 

"C"    Method   of   Laying   Out 64 

"  D  "   Method  of  Laying  Out 64 


CONTENTS.  IX 

PAGE 

"E"   Method  of  Laying  Out 65 

Fourteen    Problems    in    Simple    Curves - 65 

Obstacles  in  Surveying. 

To  Erect  a  Perpendicular  at  Any   Point  of  a   Line 76 

Table  and   Formulas  giving  Sides  of   Right   Triangles 77 

To  Drop  a  Perpendicular  from  a  Given  Point  to  a  Given  Line 77 

To  Draw  a  Perpendicular  to  a  Line  from  an  Inaccessible  Point....  77 

To  Prolong  a  Line  past  an  Obstacle  and  to  Measure  its  Length...  78 

Obstacles  to  Measuring  a  Line. 

When    One   End    is    Inaccessible 79 

When    Both    Ends   are    Inaccessible 80 

When   an   Inaccessible   Space   Intervenes 81 

Rest  Method  of  Making  a  Preliminary  Survey KJ 

Table    Illustrating    the    Same 83 

To    Replace    a    Broken    Line    between    Two    Points    by    a    Straight 

Line 85 

To  Find  the  Angle  between  Two  Straight  Lines   when  the   Point  of 

Intersection    is    Inaccessible 86 

To    Connect   Two   Tangents   by   a    Curve   when    the   Vertex   is    Inac- 
cessible   87 

To  Locate  a  Curve  when  the  Vertex  and  Both   Ends  of  the  Curve 

are    Inaccessible 87 

To    Pass    from    Any    Point    on    the    Curve    to    Any    Point    on    the 

Tangent    87 

To  Find  any  Desired  Point  on  the   Curve  when   Obstacles  preclude 

ti.e    Use    of    Ordinary    Methods 88 

CHAPTER    V. 
LEVELING,  STADIA  MEASUREMENTS,  ETC. 

Bench,    or    Bench-mark 90 

Form   of  Field-book  for  Level   Notes 92 

Proof  for   Level   Notes 93 

Benches:     Where    to    Establish    them 93 

The  Location  of  a  Level   Line 94 

The  Location  of  a  Grade  Line 94 

Correction  for   Curvature  and   Refraction 95 

Trigonometric    Leveling 97 

True  Simplified  Formulas  for  Heights  by  the  Dip  of  the  Horizon..  98 

The  Stadia. 

Formulas  for  the   Stadia — Simplified 102 

The  Gradient er. 

Formulas   for   the   Gradienter — Simplified 107 


Xll  CONTENTS. 


PAGE 

To  Connect  a  Straight  Track  and  a  Straight  Line  ...................   199 

Two  C'onnect  Two  Straight  Tracks  by  a  Curve  of  Three  I  tranches, 

the  Turnout  Curve  having  Radii  to  Suit  Given  Frogs  ............   199 

Formulas  Supposing  the  Switch-rail  to  be  a  Part  of  the  Turnout 

Curve    ................................................................  204 

A  Simple  Curve  Cannot  Meet  the  Conditions  Required  Above  ......  205 

Several   Methods  of  Laying  Out  Turnout  Explained  .................  207 

Double   Turnouts   from    a    Straight   Track  .............................  208 

To  Fit  a  Curve  to  a  Given  Middle  Frog  .............................  210 

Turnouts  from  Curves. 
Turnout  from   the   Inside  of  a   Curve  ..................................  212 

Turnout  from   the   Outside   of  a   Curve  ................................  213 

Double  Turnout  on   Opposite  Sides  of  a  Curve  .......................  214 

To   Find  Degree  of  a  Turnout  from  a  Curve  .........................  215 

Other   Turnouts   from   a   Straight   Track  ...............................  217 

To   Fit  the  Turnout  to  a  Given   Middle  Frog  ........................  219 

CHAPTER    IX. 
THE  TRUE  TRANSITION  CURVE. 


Reason  for  the  Need  of  Such   Curves 
Definition   and    Properties   of   Such    Curves 


Elementary  Relations. 
To    Find    the    Relative    Length    of    the    Offset    and    the    Transition 

Curve    ................................................................  225 

To    Find   Any   Tangent    Distance  ......................................  2_>6 

To    Find   Any   Offset  ....................................................  228 

Having  the  Offset,  t,  Supposing  the  Offset  Curve  of  the  Same  De- 

gree as  the  Original   Curve,   to   Find  the  True  Value,   /',   of  the 

Offset    ................................................................  230 

To  Find  the  Offset  iri  Terms  of  the  Cenrtal  Angle  and  the  Radius..  230 
To    Find   the   Angle   between    the   Tangent   and   Any    Chord    Drawn 

from    A  ...............................................................  23  1 

Cubic   Parabola  Not  Suitable  for  a  Transition   Curve  ................  232 

To    Find    Point    on    Curve    where    the    Tangent    is    Parallel    to    the 

Chord   of  the   Curve  .................................................  232 

To   Find  Tangents  at  the  Extremities  of  the  Curve  ..................  233 

To    Find    the    Length    of   Any    Radius    Vector,    or    Chord,    and    the 

Angles  between   these  Chords  ......................................  233 

To  Find  Deflection  Angle  at  Any  Point,  also  Any  Chord  ............  235 

To   Find  the   Exsec  dV  ',  also   TV,   etc  .................................  236 

To   Find  the   Radius  of  Curvature   at   Any   Point  .....................  236 

To  Lay  Out  the  Curve  by  Offsets  from  the  Tangent  AQ  ............  236 

Special  Problems  and  Examples. 

Given  Length  and  Degree  of  Kd,  to  Find  the  Offset  AK,  Tangent 
AO,  and  to   Lay  Out  the   Curve  ...................................  237 


CONTENTS.  xiii 


PAGE 

Given  the  Degree  of  the  Offset  Curve  and  Offset  AK,  to  Find  the 
Length  of  the  Transition  Curve,  etc 239 

Given  the  Degree  (Z)1)  of  Kd,  and  the  Tangent  AO,  to  Find  the 
Length  of  the  Transition  Curve  and  the  Offset  AK 239 

Given  the  Degree  of  the  Main  Curve,  and  the  Length  of  Kd,  to 
Find  the  Offset,  Tangent,  etc 241 

Given  the  Degree  of  the  Main  Curve  and  the  Length  of  Af  Re- 
placed, to  Find  the  Offset,  Tangent,  etc 242 

To  Replace  Each  Half  of  a  Simple  Curve  AfA1  by  a  Transition 
Curve  243 

To  Connect  Two  Tangents  by  Two  Transition  Curves,  Each  of  a 
Given  Length  5 244 

To  Connect  Two  Tangents  by  Two  Equal  Transition  Curves 
having  a  Common  Vertex  Distance  E 245 

The  Transition  Curve  Very  General  in   Use 245 

To  Lay  Out  the  Curve  from  Any  Point  on  it 245 

To  Substitute  a  Transition  Curve  for  Each  End  Portion  of  the 
Main  Curve  without  Changing  the  Rest  of  the  Curve 246 

The  True  Transition  Curve  Compared  with  Some  Others  and  His- 
torical Note 251 


CHAPTER    X. 

CALCULATION  OF  EARTHWORK. 

Prismoid   Defined 253 

Area  of  Level   Sections 254 

Area  of  Sections  not  Level 255 

Area  of  Irregular   Sections 256 

Formulas   for   Regular   Excavations   and    Embankments 256 

Error  of  the  "  End  Area  Volume  "  Always  Twice  the  Error  of  the 

"Middle  Area  Volume."     Demonstrated  by  Symbols 257 

The   End-area   Method    Simplified 258 

The    End-area  Volume    Generalized 259 

Special    Formulas   and    Cases 259' 

Formula    for    the    Volume    of    the    Frustum    of    a    Pyramid — Sim- 
plified      261 

Loaded  Flat  Cars,  Piles  of  Stone,  etc 263 

Ends  of  Embankments   or   "Dumps" 264 

Ground  Irregular  Laterally 264 

Mixed   Work,   Excavation,   and    Embankment 265 

Correction  of  Earthwork  for  Curvature 267 

Overhaul    268 

Monthly    Estimates 269 

Final    Estimates : 270 

Computation   of   Prismoids   Level   Laterally 270 


XIV  CONTENTS. 

CHAPTER    XT. 

APPROXIMATE  AND  ABRIDGED  COMPUTATIONS. 

PAGE 

Definitions    and    Notation 273 

The  Relative  Error 273 

Addition     274 

Subtraction    276 

Multiplication    and    Division 277 

Abridged    Multiplication 279 

Abridged    Division 281 

CHAPTER    XII. 

CONSTRUCTION. 

Clearing  and  Grubbing,   How  to   do  it 284 

Grade-line    284 

Surface   Ditches — Importance   of 285 

Cross-sections — Proper    Places   for 285 

Staking   Out   Earthwork   when   Ground   is   Level 286 

Staking  Out  Earthwork  when  Ground  is  Not   Level 287 

Two   Principles  to  Aid  in  Laying  Out   Earthwork 291 

Borrow-pits    - 291 

Shrinkage   of    Earthwork 294 

Retracing  the  Line 295 

Track-laying    296 

Culverts     297 

Location   of   Bridge   Piers 298 

Tunnels 298 

CHAFTEB    'Mil. 
EXPLANATION   OK  TARI.KS  AM.   MISCEI.LAKKOUS  TOPICS. 

To   Gauge  a   Stream  Approximately jo  i 

Transverse   Strength   of   Beams 302 

Safe    Bearing    Power   of    Piles 302 

TABLES. 

Table   for    Right    Triangles 32 

Table    for   Oblique    Triangles 36 

Table   Showing    Excess    of    Arcs   over    Subtended    Chords    when    the 

Arcs  are  Aliquot   Parts  of   100 44 

Table    Showing    Excess    of    Sub-cords    over    Aliquot    Parts    of    100 

when   the    Chord  =  100 45 

Table    Giving   the   Sides   of    Right    Triangles 77 

Table    for    Traverse    Survey 83 

Table  for  Level   Notes 92 

R.,  —  R, 
Table   Showing  the   Least  Value   of     T~-^—^-- 120 


CONTENTS. 


PAGE 

Table  Showing  Computation  of  Prismoids 271 

1.  Degrees,  Radii,  etc 304 

II.  Tangent   Offsets,   i   to   100   Feet 310 

III.  Offsets  for  Arcs  of  100  Feet 312 

Ilia.  Middle   Ordinate  Arcs  of   100  Feet 313 

Illb.  Chords  of  Arcs  of  100  Feet 313 

IV.  Long    Chords 314 

V.  Middle    Ordinates 316 

VI.  Turnouts  from  a   Straight  Track 318 

VII.  Tangents  and   Externals  of  a   i1   Curve 319 

VIII.  Arcs  of  Degrees,   Minutes,  and  Seconds  for  Radius  =  i 3.3 

IX.  Acres  for  Various   Lengths  and   Widths 323 

X.  Total    Grades 324 

XL  Correction   for    Curvature   and    Refraction 325 

XII.  Elevation   of  Outer  Rail 325 

XIII.  Coefficients  for  Stadia 3^6 

XIV.  Coefficients   for   Gradienter 327 

XV.  Offsets   for  Transition   Curves 328 

XVI.  Tangent    Distances    for    Transition    Curves 330 

XVII.  Deflection   Angles   for   Transition   Curves 332 

XVIII.  Earthwork  Tables,   Different   Slopes   and   Bases 333 

XIX.  Earthwork  Tables,  Two   Slopes  and  All   Bases 337 

XX.  Sines  and   Cosines 33'-$ 

XXL  Tangents  and   Cotangents 352 

XXII.  Versines  and   Exsecants 35<; 

XXIII.  Useful   Numbers  and  Formulas 382 

XXIV.  Conversion    of    Feet    into    Meters    and    Meters    into    Feet; 
also  Miles  into   Kilometers  and   Kilometers   into  Miles 383 


FIELD-MANUAL  FOE  ENGINEERS. 


CHAPTER  I. 

PRELIMINARY  OPERATIONS, 

1.  THE  engineering  operations  preparatory  to  the  construction 
of  a  railroad  are  : 

The  Reconnoissance  ; 

The  Preliminary  Survey  or  Surveys  ;  and 

The  Location. 

2.  The  Reconnoissance  is  a  general  but  incomplete  examination 
of  the  country  through  \vhich  the  proposed  road  is  to  pass,  made 
for  the  purpose  of  acquiring  data  upon  which  surveys  may  be 
made  and  compared,    and  the  best  possible  route  for  the  road 
selected. 

3.  A  Preliminary  Survey  consists  of  the  measurement  of  a  line, 
including  its  angular  deflections  ;  the  elevations  of  various  points 
upon  it,  the  determinations  of  the  topography  along  it  and  near 
it,  for  the  purpose  of  furnishing  the  data  from  which  the  line 
may  be  definitely  located;  or  the  survey  compared  with  other  sur- 
veys, for  the  purpose  of  selecting  one  from  which  the  location 
may  be  made. 

4.  The  Location  consists  in  placing  the  line  in  the  exact  posi- 
tion in  which  it  is  intended  to  be.     This  position  is  called  The 
Location. 

5.  It  is  convenient  to  carry  on  these  operations  concurrently. 
The  main  points  to  consider   in   the  location  are  the  relative 

cost  of  grading  and  bridging,  and  the  relative  grades  and  curva- 

K 


6  ITELD-MAKFAL  FOR  ENGINEERS, 

ture  of  the  lines.     In  grading,  the  character  of  the  soil  for  stabil- 
ity, in  both  "  cut "  and  "  fill,"  should  be  considered. 

It  is  sometimes  important  to  know  the  relative  value  of  property- 
traversed  by  different  lines  ;  and  if  the  lines  are  far  apart,  the 
probable  amount  of  traffic  that  the  respective  lines  can  command 
must  also  be  taken  into  account. 

6.  It  is  evident  that  the  best  possible  location  requires  the  least 
possible  grading,  bridging,  curvature,  etc.,  taken  together,  re- 
garding the   cost  and  the  expense  of  operating  the  road.     We 
can  afford,  therefore,  to  increase  the  curvature,  for  example,  if 
by  so  doing  we  can  at  the  same  time  decrease  the  earthwork,  and 
the  line  is  bettered  more  by  the  latter  than  it  is  damaged  by  the 
former. 

The  field-work  of  location  has  for  its  object  to  determine  the 
exact  position  of  the  selected  route  on  the  ground,  to  establish  the 
grade,  to  compute  the  amount  of  earthwork,  decide  upon  the 
amount  of  bridging,  etc. 

7.  A  railroad  line  usually  follows  the  valleys  of  watercourses 
or  the  dividing  ridges  between  watercourses,  or  crosses  valleys 
and  ridges  more  or  less  obliquely. 

8.  The  location  on  dividing  ridges  is  perhaps  the  simplest  of 
all.     In  this  no  bridges  and  few  culverts  are  required  ;  and  the 
elements  governing  the  location  are  mainly  the  amount  of  earth- 
work and  the  curvature  of  the  line.     In  this  case  a  sketch  of  the 
ridge,  especially  of  its  prominent  features  and  governing  points,  is 
made  while  walking  over  it;  and  the  preliminary  line  is  accurately 
run,  and  made  into  a  location,  if  the  route  is  adopted. 

9.  The  location  along  the  valley  of  a  stream  is  usually  more 
complex  than  the  former.     If  the  stream  is  so  small  that  the  cost 
of  bridging  it  would  be  plainly  less  than  the  advantage  to  the 
alignment  by  crossing  ;  or  if,  on  the  contrary,  the  river  is  so  large 
that  the  crossing  of  it  is  out  of  the  question,  the  -cost  of  bridging 
is  not  considered  and  the  problem  of  location  is  reduced,  in  the 
main,  to  that  of  making  the  best  alignment  within  the  limits  of 
the  valley  in  the  one  case,  or  upon  one  side  of  the  river  in  the  other. 
The  reconnoissance  and  surveys  would  be  made  as  already  de- 
scribed. 

Usually  the  most  favorable  ground  both  for  alignment  and  con- 
struction alternates  from  one  side  to  the  other  of  the  stream,  and 


PRELIMINARY  OPERATIONS.  7 

only  the  results  of  careful  and  scientific  surveys  can  tell  how 
many  and  where  the  crossings  rnust  be  in  order  to  make  the  align- 
ment and  grades  the  best  possible,  and  to  secure  the  most  favora- 
ble ground.  In  this  case  the  ground  on  both  sides  of  the  stream 
must  be  carefully  examined,  and  if  the  stream,  in  consequence  of 
banks  or  bottom,  is  not  easily  crossed  at  most  points,  the  most  suit- 
able crossings  must  be  found.  When  this  is  done  the  engineer 
will  mark  out  and  survey  one  or  more  lines  so  as  to  fit  the  chosen 
crossings  and  other  governing  points.  A  comparison  between 
different  lines  will  point  out  the  best,  and  generally  the  one  from 
which  the  best  location  can  be  made. 

10.  In  locating  a  line  across  valleys  and  ridges,  the  engineer 
must  find  the  best  crossings  of  the  streams,  and  the  gaps  or  notches 
in  the  ridges,  and  must  connect  such  of  the  former  with  such  of 
the  latter  as  will  furnish  the  best  line. 

11.  Sometimes  a  railroad  may  occupy  either  valleys  or  dividing 
ridges  for  the  greater  part  of  its  length  ;  in  which  case  a  choice 
must  be  made  between  the  higher  and  the  lower  line. 

The  higher  line  will  require  very  much  less  drainage  than  the 
lower  line,  which  is  an  important  advantage;  but,  on  the  contrary, 
probably  the  curvature  and  the  length  of  the  higher  line  will  ex- 
ceed that  of  the  lower. 

12.  It  is  not  to  be  supposed  that,  in  general,  a  railroad  will 
follow  either  a  valley  or  a  ridge;  or  that  it  will  cross  valleys  and 
ridges  obliquely  throughout  its   entire   length  ;   but  parts  of  the 
line  will  generally  do  so,  and  to  these,  and  therefore  to  the  whole 
line,  the  preceding  principles  will  apply. 

13.  The  regular  "  reaches  "  in  a  stream  furnish  the  best  cross- 
ings.    Throughout  these,  compared  with  other  points,  the  flow  of 
the  current  is  most  uniform;  the  wash  of  the  bottom,  and  the 
caving  and  the  shifting  of  the  banks,  are  least;  while  the  security 
of  the  foundations  of  bridges  and  of  approaches  is  greatest.     If 
ihe  bottom  of  the  stream  is  variable,  the  best  site  on  some  given 
"  reach  "  must  be  selected  with  a  view  to  the  kind  of  foundations 
f/.ui table  to  the  place.     Sharp  bends  in  streams  should  be  studi- 
ously avoided. 

.14.  The  engineer  should  freely  consult  the  best  maps  of  the 
(  (Kuitry  that  he  can  command,  and  he  should  prepare  a  map  on  a 


8  FIELD-MANUAL   FOR   ENGINEERS. 

convenient  scale,  upon  which  he  should  copy  the  principal  features 
of  the  country,  such  as  streams  and  lakes,  roads  and  towns,  and 
fill  in  the  details  as  he  progresses.  He  should  also  locate  on  the 
map  the  governing  points  of  the  route,  such  as  the  best  crossings 
of  streams,  the  "gaps"  in  the  ridges,  mountain  passes,  etc.  lie 
may  then  sketch  the  line  on  the  map. 

In  a  densely  wooded  country,  the  making  of  a  thorough  recon- 
noissance  is  comparatively  difficult.  In  this  case  it  will  often  be 
necessary  to  cross  and  recross  the  country  many  times  before  a 
comprehensive  knowledge  of  it  can  be  gained. 

15.  The  one  almost  indispensable  instrument  in  making  a  re- 
connoissance   is   a   pocket-compass.      Field-glasses,   hand-levels, 
telemeters,   and  other  instruments  are  sometimes  used,  but  are 
rarely  needed,   and   cannot  be  used  to  advantage.     Those  who 
cannot  make  a  proper  reconnoissance  without  them  would  better 
employ  some  one  who  can,  and  turn  their  attention  to  other  parts 
of  the  work. 

16.  For   preliminary  surveys   the  corps  of  engineers  may  be 
constituted  as  follows:  A  chief  engineer  or  engineer  in  charge,  an 
assistant  engineer  or  transitinan,  a  levelman,  a  rodman,  a  stake- 
man,  a  rear  flagman,  two  chainrnen,  and  one  or  more  axmen  accord- 
ing to  needs.     The  head  flag  should  be  carried  by  the  head  chain- 
man. 

17.  Since  a  survey  can  be  made  more  rapidly  with  the  compass 
than  with  the  transit,  the  compass  may  be  used  in  preliminary 
work  to  a  limited  extent.     Owing,  however,  to  the  inherent  in- 
accuracies of  the  compass,  a  line  run  with  it  is  generally  worthless, 
except  as  a  guide  to  a  transit  preliminary,  from  which  the  location 
may  be  made.     The  compass  is  therefore  of  very  little,  if  any,  use 
in  a  comparatively  level  country,  but  is  useful,  if  at  all,  for  run- 
ning the  first  preliminary  line  in  a  billy  region,  where  several  lines 
must  be  run.     The  compass  should  be  light,  should  be  mounted 
on  a  Jacob 's-staff,  and  should  have  a  narrow  slit  in  one  sight,  and 
a  fine  platinum  wire,  or  its  equivalent,  stretched  along  the  center 
of  a  wide  slit  in  the  other  sight.     A  self-reading  rod  is  best  for 
this  work  because  it  saves  much  time  over  the  sliding  rod;  and 
because,  too,  it  enables  the  levelman  to  do  his  own  reading.     The 
ax  for  driving  stakes  should  have  a  broad  head.    .  Stakes  shotild 
be  of  a  generous  length,  say  from  30  to  36  inches,  well  driven  into 


PRELIMINARY  OPERATIONS.  9 

the  ground,  and  projecting  above  the  grass  and  other  vegetation, 
so  that  the  line  may  be  easily  followed  or  recovered,  in  fields  or 
woods,  by  the  engineer  and  others.  Short  stakes  occasion  the  loss 
of  much  time,  and  those  shorter  than  about  30  inches  are  generally 
a  nuisance,  except  on  streets,  well-traveled  roads,  etc. 

18.  To  make  a  successful  reconnoissance  requires  a  good  eye 
for  distances,  elevations,  etc. ,  and  a  quick  and  clear  perception  of 
the  salient  features  of  the  country.     One  must  be  able  to  form  a 
mental  picture  or  image  of  the  country  along  the  proposed  line ; 
and  of  a  number  of  lines  crossing  and  recrossing  each  other  within 
the  limits  of  his  mental  map.     Thus  he  may  be  able  to  sketch  a 
proposed  line,  or  several  proposed  lines,  which  at  any  time  may  be 
tested.     The  comparison  of  lines,  to  test  their  relative  economy, 
is  a  question  of  science,  aided  by  mathematics. 

19.  It  is  plain  that  without  skill  in  making  the  reconnoissance, 
the  surveys  would  be  at  first  more  or  less  at  random,  and  the 
reaching  of  a  location  roundabout  and  expensive.     On  the  other 
hand,  the  greatest  skill  without  a  sufficient  knowledge  of  mathe- 
matics and  without  a  knowledge  of.  the  principles  of  engineering 
bearing  upon  the  question,  cannot  produce  the  best  location. 

A  happy  combination  of  the  qualities  necessary  to  the  successful 
locating  engineer  is  most  assuredly  found  in  comparatively  few 
individuals. 

The  two  main  obstacles  to  contend  with  in  building  a  railway 
are  grades  and  curves.  Since  both  affect  the  cost  of  construction 
and  the  expense  of  operating  the  road,  such  grades  and  curves  as 
will  render  the  total  cost  of  construction  and  operation  of  road  a 
minimum  are,  as  already  suggested,  the  best.  Since  the  resistances 
due  to  grades  as  well  as  to  curves  add  to  other  resistances  to  the 
movement  of  trains,  and  since  it  is  often  possible  to  lessen  grades 
at  the  expense  of  curvature,  and  vice  versa,  it  becomes  necessary 
to  briefly  consider  the  nature  of  the  resistances  that  a  moving  train 
encounters,  with  a  view  to  compensation. 

The  resistances  to  overcome  are,  first,  the  friction  of  the  moving 
parts  of  the  engine  and  train ;  the  friction  of  the  wheels  on  the 
rails  ;  impacts  and  oscillations,  and  the  resistance  of  the  air. 
These  resistances  vary  with  the  condition  of  the  rolling  machinery, 
the  road,  and  the  weather,  and  are  not  accurately  known.  Fric- 
tion is  nearly  independent  of  the  speed  of  the  train;  but  the  resist- 


10  I'lKI.iJ-MANUAL   FOR   ENGINEERS. 

ances  due  to  impact  increase  with  the  speed,  and  those  due  to  the 
atmosphere  increase  in  a  still  greater  ratio. 

The  sum  of  these  resistances  on  a  level  track  in  fairly  good  con- 
dition and  average  fair  weather  is,  according  to  Vose, 


in  which  r  is  the  resistance  in  pounds  per  ton,  and  v  is  the  velocity 
in.  miles  per  hour.  It  is  not  to  be  supposed  that  this  formula  is 
accurate,  though  perhaps  it  is  as  nearly  so  as  any;  and  probably 
the  resistances  are  no  greater  than  the  formula  indicates.  "With 
a  velocity  of  20  miles  per  hour  the  formula  gives  r  =  10.34  pounds 
per  ton  of  the  entire  weight.  Under  unfavorable  conditions,  as  a 
wet,  soft,  and  rough  track  and  high  wind,  the  resistance  might 
be  20  to  40  per  cent  more,  or  even  greater  still. 

The  second  resistance  is  due  to  grades  of  the  track.  The  force 
necessary  to  overcome  this  is  such  a  part  of  the  total  weight  of 
the  train  as  the  vertical  rise  of  the  grade  is  to  its  length.  Let 
r'  —  resistance  per  ton,  and  U  =  the  ascent  per  100  feet.  Then 

?•'  =  —  ;  •  X  2000  pounds  per  ton.     For  a  \%  grade  (52.8  feet  per 
1  UU 

mile)  h  =  1  and  r'  —  20  pounds  per  ton,  the  force  necessary  to 
overcome  the  load.  A  \%  grade  would  of  course  create  a  resist- 
ance of  10  pounds  per  ton,  or  approximately  the  same  as  the 
resistance  caused  by  friction,  impact,  resistance  of  the  atmos- 
phere, etc.,  at  a  speed  of  20  miles  per  hour,  as  stated  above. 

The  third  resistance  is  due  to  the  curves  of  the  track.  This  re- 
sistance is  not  accurately  known.  On  American  roads  with  American 
rolling  stock  it  is  probably  about  one  half  a  pound  per  ton  for  each 
degree  of  curvature.  Letting  ?•''  represent  the  resistance  per  ton 
per  station,  and  D  the  degree  of  curvature,  the  resistance  per  sta- 
tion is  r"  =  .5D.  We  observe  that  the  resistance  of  a  1°  curve 
is  but  the  ^  part  of  that  due  to  a  \%  grade,  or  equivalent  to 

K<)     Q 

-^-  =  1  .32  feet  per  mile.     The  resistance  of  a  5°  curve  is  equiva- 

lent to  6.6  feet  per  mile,  and  that  of  a  10°  curve  to  13.2  feet  per 
mile,  etc.  The  resistance  offered  to  a  train  in  moving  through  100 
feet  of  a  1°  curve  is  one-half  pound  per  ton  moved  100  feet,  which 

is  equivalent  to  the  total  load  lifted  -^  X  100  =  ^  =  .025  ft.; 
and  since  the  resistance  varies  as  the  product  of  the  degree  of 


PRELIMINARY  OPERATIONS.  11 

curvature  and  length  of  the  curve,  or  as  the  total  curvature, 
the  resistance  on  any  curve  for  each  degree  of  change  of  direc- 
tion is  equal  to  the  lifting  of  the  train  ^  of  a  foot.  Since  an  en- 
gine can  haul  over  a  road  only  what  it  can  haul  over  the  most 
unfavorable  part,  easy  curves  are  necessary  where  heavy  grades 
must  occur,  and  vice  versa  ;  the  object  being,  of  course,  to  keep  at 
all  points  the  sum  of  the  resistances  due  to  grades  and  curves  be- 
low the  allowable  maximum.  In  equating  for  grades  and  curves, 
however,  the  ruling  element  is  largely  the  total  ascent  and  not 
the  slope  merely,  as  is  too  often  considered  to  be  the  criterion. 
No  more  power  is  required  to  make  an  ascent  of  20  feet,  for  ex- 
ample, on  a  2/c  grade  1000  feet  long  than  is  required  on  a  \%  grade 
4000  feet  long. 

To  compare  grades  simply,  that  is,  the  rale  of  ascent  or  descent 
especially  on  short  grades,  as  is  usually  done,  is  meaningless  at 
least — it  is  absurd. 

On  light  grades  no  allowance  usually  need  be  made  for  curva- 
ture, since  the  momentum  of  the  train  would  carry  it  over  the 
ascent  even  without  aid  from  the  engine.  Grades  situated  near 
stations,  however,  or  other  points  where  a  train  must  stop  or 
"  slow  up  "  are  greater  obstacles  than  in  other  situations  to  the 
movement  of  trains,  and  upon  such  grades  an  allowance  for  cur- 
vature, as  above  shown,  should  be  made. 

Every  locating  engineer  is  presumed  to  make  himself  familiar 
with  the  principles  involved  in  locations  before  making  them; 
and  to  aid  in  this  the  reader  is  referred  to  Vose  and  others,  and 
especially  to  the  elaborate  treatise  of  Wellington  on  "Railway 
Location."  Wellington  gives  in  Table  118  the  "  Total  Energy, 
or  Potential  Lift,  in  Vertical  Feet  in  Trains  moving  at  Various 
Velocities,"  and  discusses  at  length  all  subjects  connected  with 
location. 


CHAPTER   II. 

ADJUSTMENTS,  USES,  AND  CARE  OF  INSTRUMENTS. 

THE  TRANSIT. 

The  following  are  the  usual  adjustments  of  the  transit : 

A.  To  Adjust  the  Levels  (that  is,  to  place  the  levels  in  a  plane 
perpendicular  to  the  upright  axis). — Set  up  the  instrument  upon 
its  tripod  as  nearly  level  as  may  be;  unclamp  the  plates  and  hring 
the  two  levels  above  and  on  a  line  with  the  two  pairs  of  leveling- 
screws.     Then  by  means  of  one  pair  of  screws  bring  the  bubble 
of  the  level  above  them  to  the  middle  of  the  opening.     Without 
moving  the  instrument  bring  the  other  bubble  to  the  middle  in 
the  same  way.     Since  in  moving  one  pair  of  screws  very  far  the 
other  pair  is  liable  to  become  cramped  and   the   corresponding 
bubble  somewhat  disturbed,  it  is  advisable  to  bring  the  bubbles 
in  succession  near  the  middle,  repeating  if  necessary,  and  ending 
by  bringing  them  exactly  to  the  middle. 

When  both  bubbles  are  in  place  turn  the  instrument  through 
about  180°;  if  the  bubbles  are  now  in  position  they  need  no  cor- 
rection ;  but  if  not,  turn  the  small  nuts  at  the  end  of  the  levels 
until  the  bubbles  are  moved  over  one  half  of  the  error.  Then 
bring  the  bubbles  again  to  the  middle  by  the  leveling-screws  and 
repeat  the  operation  just  described  until  the  bubbles  will  remain 
in  the  middle  during  a  complete  revolution  of  the  telescope.  This 
shows  the  adjustment  to  be  completed. 

B.  To  Set  the  Cross-wires  Vertical  and  Horizontal.—  Level  the 
instrument.     Move  the  telescope  upward  or  downward  and  note 
whether  the  vertical  wire  traverses  some  fixed  point  or  not.     If 
not,  loosen  the  four  cross- wire  screws  and,  by  the  pressure  of  the 
hands  on  their  head  outside  the  tube,  move  the  cross- wire  ring 
around,  what  seems  to  be  sufficient,  and  repeat  the  operation,  if 
necessary,  until  the  vertical  wire  will  traverse  some  fixed  point. 

The  cross-wire  screws  are  near  the  screws  of  the  centering-ring 

12 


ADJUSTMENTS,  USES,  AND    CARE   OF   INSTRUMENTS.     13 

of  the  eyepiece  (the  beads  of  which  are  usually  covered  by  an 
outside  ring),  but  between  them  and  the  axis  of  the  telescope. 

C.  To  Make  the  Line  of  Collimation  Perpendicular  to  tlie  Axis  of 
tlie  Wyes,  so  tliat  it  will  revolve  in  a  plane. — Set  up  the  transit 
at  some  point  0  near  the  middle  of  a  level  piece  of  ground 
and  level  it  carefully.  Let  AOB  be  a  straight  line,  AO  and 
BO  being  nearly  equal.  Direct  the  line  of  sight  to  A,  clamp 
the  instrument  and  revolve  the  telescope,  and  if  the  line  of 
collimation  is  not  perpendicular  to  the  axis,  the  line  of  sight 
will  determine  a  point  0,  say,  on  one  side  of  J3.  To  test  the  mat- 
ter, loosen  the  upper  clamp  and  turn  the  vernier  plate  almost  or 
quite  half-way  around,  so  that  the  line  of  sight  will  again  be  on 


FTG.  1. 

A,  and  clamp  the  plates.    Again  revolve  the  telescope  and  note  tbe 
point  D,  suppose,  thus  determined.     If  the  point  C  coincides  with 

B,  D  will  also  coincide  with  B  and  the  line  of  collimation  is  in 
adjustment.     If,  however,  C  is  on  one  side  of  B,  as  shown,  D  will 
be  equally  far  on  the  other  side,  showing  the  line  of  collimation 
out  of  adjustment. 

To  correct  the  error,  use  the  two  capstan-head  screws  on  the  side 
of  the  telescope,  to  move  the  ring  to  which  the  wires  are  fastened 
laterally,  and  with  it,  of  course,  the  intersection  of  the  wires. 

Having  moved  the  vertical  wire  until  by  estimation  one  fourth 
of  the  space  DC  is  passed  over,  return  to  A,  clamp  and  revolve 
the  telescope,  and  if  the  correction  has  been  carefully  made  the 
line  of  sight  will  now  cut  the  point  B.  It  will,  however,  generally 
be  necessary  to  repeat  the  operation,  the  adjustment  not  being 
perfected  at  first. 

Remember  that  the  eyepiece  inverts  the  position  of  the  wires, 
and  therefore  in  moving  the  ring  the  operator  must  proceed  so  as 
seemingly  to  increase  the  error. 

The  wires  may  now  be  brought  into  the  center  of  the  field  of 
view  by  moving  the  screws  of  the  centering-ring  of  the  eyepiece, 
which  are  slackened  and  tightened  in  pairs,  the  movement  being 
now  direct  until  the  wires  are  seen  in  their  proper  position. 


4  tftELft-MANUAL   FOR   ENGINEERS. 

It  is  proper  to  observe  that  the  position  of  the  line  of  collirna- 
tion  depends  solely  upon  that  of  the  objective,  so  that  the  eye- 
piece may  be  moved  in  any  direction,  or  replaced  by  another 
without  at  all  deranging  or  in  any  way  affecting  the  adjustment 
of  the  wires. 

I).  To  Adjust  the  Standards  to  the  Same  Height  so  that  the  Line, 
of  Collima'.ion  icill  Revrtve  in  a  Vertical  Plane. — Set  the  transit 
as  close  as  convenient  to  the  base  of  a  lofty  spire,  or  other  high 
object;  level  it  carefully  and  clamp  it,  and 
direct  the  telescope  to  the  top  of  the  spire  or 
other  elevated  and  well-defined  point,  as  A 
in  the  figure.  B  is  in  the  same  vertical  plane 
with  A  and  the  instrument,  that  is,  with  the 
line  of  sight. 

Turn  down  the  telescope  to  some  good  point 
on  the  ground,  either  found  or  marked,  as  C. 
Unclainp  the  plates  or  spindle,  revolve  the 
telescope  and  turn  it  half-way  around,  or  far 
enough  to  again  sight  to  the  high  point. 
Again  clamp  and  turn  down  the  telescope  to 
~Q  some  point  D  opposite  C. 

In  setting  C  the  left  standard  must  have 
been  too  high,  and  in  setting  D  the  same 
standard  (now  the  right  by  revolving  the  telescope)  shows  to  be 
equally  high,  the  errors  EG  and  BD  being  equal.  Correct  the 
error  by  raising  or  lowering  the  sliding-piece  at  one  end  of  the 
axis  by  means  of  the  screw  beneath  and  those  above,  so  that  when 
the  telescope  is  directed  to  A  and  lowered  it  will  cut  B  half-way 
between  C  and  D.  If  the  instrument  is  in  adjustment  it  will,  of 
course,  cut  B  instead  of  C  and  D  in  the  first  two  trials. 

E.  To  Adjust  the  Vertical  Circle. — Set  up  the  instrument  firmly, 
level  it  carefully,  bring  into  line  the  zeros  of  the  circle  and 
vernier,  and  with  the  telescope  find  some  well-defined  point,  from 
two  to  five  hundred  feet  distant,  which  is  cut  by  the  horizontal 
wire.  Turn  the  instrument  half-way  around,  revolve  the  tele- 
scope, fix  the  wire  upon  the  same  point  as  before,  and  observe  if 
the  zeros  are  again  in  line.  If  not,  loosen  the  capstan-head 
screws  which  fasten  the  vernier,  and  move  the  zero  of  the  vernier 
over  half  of  the  error.  Bring  the  zeros  again  into  coincidence 
and  proceed  as  before,  as  many  times  as  necessary,  until  the 
error  is  entirely  corrected. 


ADJUSTMENTS,  USES,  AND    CARE   OF   INSTRUMENTS.     15 

F.  To  Adjust  the  Level  on  Telescope.— First  level  carefully  and 
clamp  the  telescope  approximately  horizontal  by  the  eye. 

Then,  Laving  the  line  of  collhnation  previously  adjusted,  drive 
a  stake,  say  two  or  three  hundred  feet  away,  and  note  the  height 
cut  by  the  horizontal  wire  upon  a  staff  set  on  top  of  the  stake. 

Fix  another  stake  in  the  opposite  direction  and  at  the  same  dis- 
tance from  the  instrument,  and  without  disturbing  the  telescope 
turn  the  instrument  upon  its  spindle,  set  the  staff  upon  the  stake, 
and  drive  the  stake  into  the  ground  until  the  reading  is  the  same 
as  on  the  first  stake.  The  tops  of  the  two  stakes  are  equally  high 
however  much  the  telescope  may  be  out  of  level. 

Now  set  the  instrument  some  twenty  or  thirty  feet  from  one  of 
the  stakes  and  on  the  prolongation  of  the  line  joining  the  two. 
Level  the  instrument,  clamp  the  telescope  as  nearly  horizontal 
as  may  be,  and  note  the  readings  of  the  staff  on  the  two  stakes. 
If  they  agree,  the  telescope  is  level.  If  they  do  not  agree,  then 
with  the  tangent-screw  move  the  wire  over  fully  the  whole  error, 
as  shown  at  the  distant  stake;  repeat  the  operation  just  described 
a  <  many  times  as  may  be  necessary,  so  that  the  wire  will  give  the 
same  reading  at  both  stakes,  showing  that  the  telescope  is  truly 
horizontal.  Taking  care  not  to  disturb  the  position  of  the  tele- 
scope, bring  the  bubble  into  the  middle  by  the  little  leveling-nuts 
at  the  ends  of  the  tube,  which  will  render  the  adjustment  com- 
plete. 

The  above  are  all  the  adjustments  usually  required  at  the  hands 
ofthe  engineer  ;  for  any  others  see  the  Manual  of  W.  &  L.  E. 
Gurley. 

USE  AND  CAKE  OP  TRANSIT. 

The  instrument  should  be  set  up  firmJy  with  the  plates  nearly 
level,  and  thus  save  much  time  in  turning  the  leveling-screws, 
besides  the  worse  than  useless  wear  upon  them. 

To  do  this  :  Set  up  the  transit  approximately  over  the  desired 
point.  Hold  firmly  the  leg  of  the  tripod  on  the  left  with  the  left 
hand,  and  move  the  foot  of  the  leg  on  the  right  with  the  right 
hand  in  any  direction,  so  as  to  bring  the  lower  plate  of  the  tripod 
head  approximately  horizontal,  determined  by  placing  the  eyes  in 
the  plane  of  its  upper  surface  and  sighting  as  nearly  as  may  be 
to  the  distant  horizon.  It  will  do  no  harm  to  bend  the  body  the 
trifle  necessary  to  do  this.  This  sight  is  so  long  that,  according  to 


16  FIELD-MANUAL   FOR   ENGINEERS. 

the  principle  of  the  "  division  of  errors,"  the  plate  can  be  readily 
set  very  nearly  horizontal  in  almost  any  position  of  the  instru- 
ment. 

Now  place  the  left  hip  against  one  leg;  grasp  the  opposite  leg 
with  the  left  hand,  and  the  leg  on  the  right  with  the  right  hand; 
and  keeping  the  relative  position  of  the  legs  unaltered,  move  the 
instrument  laterally  over  the  desired  point  and  lower  it  to  the 
ground. 

Of  course  the  two  movements  of  tilting  and  sliding  may  be  done 
together  and  in  much  less  time  than  required  to  describe  the 
movement. 

To  MEASURE  THE  ANGLE  BETWEEN  Two  LINES  OR  OBJECTS. 

Level  the  instrument  carefully;  bring  the  zeros  of  the  verniers 
and  limb  together  by  means  of  the  upper  clamp  and  tangent-screw, 
and  direct  the  telescope  toward  one  of  the  objects  by  means  of 
the  lower  clamp  and  tangent-screw.  Upon  loosening  the  upper 
clamp  and  directing  the  telescope  toward  the  second  object,  the 
angle  desired  is  then  shown  upon  the  limb. 

Before  making  an  observation  with  the  telescope,  the  eyepiece 
should  be  moved  in  or  out  until  the  cross- wires  may  be  distinctly 
seen.  This  may  be  accomplished  with  the  greatest  precision  by 
directing  the  telescope  toward  some  white  object.  The  sky 
will  serve  very  well  for  this.  The  objective  is  then  adjusted  by 
moving  it  in  or  out  until  the  object  is  seen  clear  and  well  defined, 
and  the  cross- wires  appear  as  if  fastened  to  its  surface. 

Water  and  dust  are  very  destructive  to  instruments  ;  indeed, 
the  most  destructive  of  all  agents  is  dust.  An  instrument  may  be 
badly  bruised,  bent,  or  broken,  by  a  fall  or  otherwise,  and  yet  it 
may  be  repaired.  If,  however,  it  is  allowed  to  stand  on  its  tripod 
in  a  dusty  office — and  all  offices  are  dusty — its  life  will  be  short. 
The  author  has  known  of  instruments  so  choice  that  their  owners 
would  allow  no  one  to  handle  them  lest  they  might  become 
slightly  soiled,  or  some  other  mishap  befall  them;  and  yet  in  a 
few  months  they  were  ruined,  in  the  way  pointed  out.  A  minute 
quantity  of  dust  on  the  sockets  will  probably  cause  them  to  grind; 
and  more  of  it  will  increase  the  grinding  so  that  the  instrument 
will  soon  become  useless. 

Whenever  not  in  actual  use  the  eyepiece  should  be  covered  by 
the  lid,  and  the  object-glass  by  the  cap,  to  protect  them  from  dust, 
moisture,  rain,  etc, 


ADJUSTMENTS,  USES,  AND   CAKE   OF   INSTRUMENTS.    17 

"When  an  instrument  is  exposed  to  the  hot  rays  of  the  sun  for 
some  time  its  parts  are  subjected  to  very  unequal  expansion,  which 
throws  the  instrument  more  or  less  out  of  adjustment,  thus  to  some 
extent  vitiating  the  work  and  damaging  the  finer  parts  of  the  in- 
strument. To  prevent  this  it  should  be  shielded  by  an  umbrella  or 
screen  of  some  kind  supported  on  top  of  a  high  stake;  or,  if  nothing 
better  is  at  hand,  a  cloth  should  be  placed  over  the  telescope. 

When  an  instrument  is  in  the  office  or  in  transport  it  should 
be  clamped  and  placed  in  a  stable  position,  upon  proper  supports, 
in  a  tight  box,  well  cushioned  if  possible. 

If  the  box  cover  does  not  fit  upon  the  box  closely  it  can  be 
remedied  by  sticking  suitable  cloth  on  the  top  edge  of  the  box  or 
on  the  under  side  of  the  lid. 

When  handling  an  instrument  it  should  be  supported  by  plac- 
ing the  hand  under  the  lower  plate. 

Tangent  and  micrometer  screws  should  be  used  equally  on  all 
portions  of  their  length. 

Keep  the  tripod  legs  tight  enough  on  the  tripod  head  to 
secure  stability  by  tightening  the  nuts  on  the  bolts  when 
necessary.  Also  tighten  the  shoes  of  the  tripod  if  they  become 
loose.  Neglect  of  these  things  sometimes  prevents  an  engineer 
from  keeping  his  telescope  on  a  point,  and  causes  him  to  run 
a  zigzag  line  without  knowing  the  cause  of  it.  Secure  the 
instrument  well  to  the  tripod  head  before  using  it;  and  bring 
all  four  leveling-screws  to  a  bearing  and  cover  the  instrument 
with  an  oilcloth  hood  before  carrying  it.  Use  a  fine  camel- 
hair  brush  or  a  piece  of  old  and  soft  linen  to  clean  the  glasses 
of  the  telescope. 

Dust,  moisture,  perspiration,  etc.,  will  sometimes  cause  a 
film  to  form  on  the  lenses  of  a  telescope  which  may  greatly  im- 
pair the  sight  through  it. 

To  remove  the  film,  the  lenses,  after  being  carefully  brushed, 
should  be  gently  wiped  with  a  piece  of  chamois-skin  moistened 
with  alcohol,  and  the  lenses  must  be  wiped  dry  by  using  fresh 
portions  of  the  skin  on  separate  parts  of  the  lens. 

To  remove  dampness  in  the  main  tube  of  the  telescope,  take 
out  the  eyepiece,  cover  the  open  end  with  cloth,  and  leave  the 
instrument  in  a  dry  room  for  some  time. 

The  centers  of  an  instrument  should  aluavs  be  lubricated 
with  fine  watch-oil  only,  and  after  a  careful  cleaning.  First 
wipe  off  all  old  grit  and  oil  before  applying  fresh  oil, 


18  FIELD-MANUAL   FOR   ENGINEERS. 

If  dust  settles  on  the  cross-wires,  unscrew  the  eyepiece  and 
the  object-glass,  and  gently  blow  through  the  telescope  tube; 
cover  up  both  ends  ahd  wait  a  few  minutes  before  replacing 
the  eyepiece  and  object-glass. 

Be  sure  to  bring  the  object-glass  cell  to  a  firm  bearing  against 
its  shoulder,  and  then  examine  the  adjustment  of  the  lines  of 
collimation. 

To  clean  the  threads  of  a  leveling  or  tangent  screw  use  a 
stiff  tooth-brush  to  remove  the  dust,  then  apply  a  little  oil 
and  turn  the  screws  in  and  out  with  alternate  brushing  to 
remove  dust  and  oil,  until  it  moves  freely  and  smoothly.  Use 
such  a  brush  to  clean  the  object-slide.  Xo  screws  should  be 
strained  more  than  necessary  to  insure  a  firm  bearing,  and 
this  applies  with  special  force  to  the  cross-wire  screws. 

All  straining  of  such  screws  beyond  this  impairs  the  accuracy 
of  the  instrument  and  the  reliability  of  adjustment. 

These  remarks  in  reference  to  the  care  of  the  transit  apply 
substantially,  of  course,  to  all  instruments. 

If  obliged  to  work  with  an  instrument  of  faulty  graduation, 
it  is  best  to  read  each  angle  on  different  parts  of  the  circle, 
and  take  the  mean  value.  If  Are  take  the  mean  result  obtained 
by  both  verniers,  we  eliminate  the  errors  due  to  eccentricity 
of  the  vertical  axis,  and  also  reduce  the  errors  of  graduation. 

For  greater  accuracy  clamp  the  vernier-plate  to  zero  and  read 
the  angle  by  both  verniers.  Then  keep  the  vernier-plate 
clamped,  point  the  telescope  to  the  first  object  and  proceed  as 
before,  any  number  of  times.  Then  read  the  verniers,  adding 
360°  for  each  complete  revolution  which  has  been  made,  and 
divide  this  sum  by  the  number  of  times  the  angle  was  read. 
The  quotient  is  the  required  angle. 

The  cross-wires  can  be  illuminated  easily  by  placing  a  piece 
of  white  cardboard,  with  a  hole  through  it  for  the  line  of  sight, 
in  front  of  the  telescope,  and  in  an  oblique  positio'n,  so  as  to 
reflect  into  the  telescope  the  rays  of  a  lamp  or  of  a  lantern 
placed  back  of  the  object-glass  and  near  the  telescope. 

THE  LEVEL. 

The  principal  adjustments  of  the  level  consist  in  the  follow- 
ing: 

I,  Bringing  the  cross-wires  into  the  optical  axis  of  (he  lele- 


ADJUSTMENTS,  USES,  AND    CARE   OF   INSTRUMENTS.     ID 

scope  and  consequently  parallel  to  the  line  of  bearings  of  the 
wye  rings. 

%2.  Causing  the  line  of  bearings  to  be  parallel  to  the  plane 
of  the  level. 

3.  flaking  either  of  these  lines,  and  therefore  all  of  them, 
parallel  to  the  bar  and  consequently  perpendicular  to  the 
axis  of  the  instrument. 

1.  To  adjust   the   line   of  collimation,  that  is,   to  bring  the 
cross-wires  into  the  optical  axis,  so  that  their  point  of  inter- 
section will  remain  on  any  given  point  during  an  entire  revolu- 
tion of  the  telescope. 

Set  the  tripod  firmly,  remove  the  wye-pins  from  the  clips, 
so  as  to  allow  the  telescope  to  turn  freely;  clamp  the  instru- 
ment to  the  leveling-head,  and  by  the  leveling  and  tangent 
screws  bring  cither  of  the  wires  upon  the  clearly  defined  edge 
of  some  object.  Then  carefully  rotate  the  telescope  half-way 
around. 

!  f  the  wire  docs  not  coincide  with  the  line  observed,  bring 
it  half-way  back  by  means  of  the  capstan-head  screws  at  right 
angles  with  it.  always  remembering  the  inverting  property 
of  the  eyepiece.  Then  by  the  leveling  and  tangent  screws 
bring  it  again  upon  the  "edge,"  and  repeat  the  above  opera- 
tion if  necessary,  and  continue  to  do  so  until  the  telescope  may 
be  rotated  without  changing- the  position  of  the  wires. 

If  both  wires  are  much  out  of  position,  it  will  be  well  to 
approximately  adjust  the  wires  alternately  before  attempting 
to  make  the  adjustment  of  either  one  complete,  since  an  error 
in  one  somewhat  affects  the  other. 

It  may  be  advisable  to  center  the  eyepiece.  To  do  so  unscrew 
the  covering  of  the  eyepiece  centering-screws,  and  move  each 
pair  in  succession,  with  a  screw-driver,  until  the  wires  are 
brought  into  the  center  of  the  field  of  view. 

The  inverting  property  of  the  eyepiece  does  not  affect  this 
operation,  and  the  screws  are  moved  directly.  To  test  the  cen- 
tering, rotate  the  telescope,  and  if  an  object  observed  appears 
to  change  position  the  centering  is  not  perfect. 

In  all  telescopes  the  line  of  collimation  is  determined  by  the 
cross- wires  and  objective,  and  is  not  affected  in  any  way  by  the 
eyepiece. 

2.  To  make  the  line  of  bearings  parallel  to  the  bubble-tube, 


20  FIELD-MANUAL   FOR   ENGINEERS. 

so  as  to  insure  that  it  is  horizontal,  when  the  bubble  is  in  the 
center.    This  adjustment* embraces  two  parts: 

'  First,  to  bring  the  center-line  of  the  bubble  and  the  line  of 
bearings  in  the  same  plane. 

Second,  to  make  these  lines  parallel. 

To  effect  the  first:  Clamp  the  level  and  bring  the  bubble 
to  the  center  by  the  parallel-plate  screw*.  Now  rotate  the 
telescope  in  the  wyes  20°,  more  or  less.  If  the  bubble  runs 
toward  the  end,  it  shows  that  the  center-line  of  the  bubble  and 
the  axis  of  the  telescope  are  not  in  the  same  plane;  in  other 
words,  the  bubble-tube  lies  crosswise  of  the  telescope. 

To  correct  the  error,  bring  the  bubble  by  estimation  half- 
way back,  by  the  capstan-'head  screws,  which  are  set  in  either 
side  of  the  level-holder. 

Again  bring  the  level-tube  under  the  telescope,  the  bubble 
to  the  middle,  etc.,  repeating  the  operation  just  described  until 
the  bubble  will  keep  its  position,  when  the  telescope  is  re- 
volved. 

For  the  second  part:  Bring  the  bubble  to  the  middle  of  the 
tube  by  the  leveling-screws,  and  take  the  telescope  out  of  the 
wyes  carefully  and  turn  it  end  for  end. 

If  the  bubble  runs  toward  either  end,  lower  that  end,  or 
raise  the  other  by  turning  the  adjusting-nuts  on  one  end  of  the 
level  until  by  estimation  half  the  correction  is  made.  Again 
bring  the  bubble  to  the  middle  by  the  leveling-screws,  and 
repeat  the  whole  operation  just  described  until  the  reversion 
can  be  made  without  causing  tiny  change  in  the  bubble. 

3.  Having  made  the  previous  adjustments,  it  remains  to 
make  the  level-bubble  (and  therefore  the  line  of  bearings  and 
line  of  collimation)  parallel  to  the  bar,  and  therefore  perpen- 
dicular to  the  vertical  axis  of  the  level,  so  that  the  bubble  will 
remain  in  the  middle  during  an  entire  revolution  of  the  tele- 
scope. 

Place  the  level  over  a  pair  of  leveling-screws,  and  by  means 
of  them  bring  the  bubble  to  the  middle.  Turn  the  instrument 
half-way  around  horizontally.  If  the  bubble  runs  toward  either 
end,  bring  it  half-way  buck  by  either  pair  of  nuts,  at  the  ends 
of  the  bar.  Then  bring  the  bubble  to  the  middle  again,  by  the 
level  ing-screws,  etc..  repeating  the  operation  just,  described, 


ADJUSTMENTS,  USES,  AND   CARE   OF   INSTRUMENTS.     21 

until  the  bubble  will  remain  in  the  middle  of  the  tube  when 
the  instrument  is  revolved. 

In  making  this  adjustment  it  is  best  to  use  the  opposite 
pairs  of  leveling-screws  alternately,  thus  bringing  the  upper 
parallel  plate  of  the  tripod  head  into  a  position  as  nearly  hori- 
zontal as  possible,  so  that  the  error  caused  by  not  revolving  the 
instrument  precisely  180°  may  be  the  least  possible. 

This  adjustment  is  for  convenience  and  not  for  accuracy  in 
any  appreciable  degree. 

Xow  turn  the  telescope  in  the  wyes  until  the  pin  on  the 
clip  of  the  wye  will  rest  in  the  little  recess  in  the  ring  to  which 
it  is  fitted. 

Apply  the  horizontal  wire  to  any  level  line,  and  in  case  it 
does  not  coincide  with  it,  loosen  two  cross-wire  screws  at  right 
angles  to  each  other,  and  by  their  heads  outside  turn  the  cross- 
wire  ring  until  the  horizontal  wire  coincides  with  the  level 
line. 

The  line  of  collimation  must  then  be  adjusted  again.  In 
readjusting  the  line  of  collimation  none  of  the  lines  referred  to 
in  making  the  adjustment  is  disturbed,  and  the  adjustments 
are  complete. 

ADJUSTING  BY  THE  "  PEG  "  METHOD. 

By  this  method  the  main  adjustments  are  effected  at  once. 

Drive  two  pegs  several  hundred  feet  apart,  and  set  the  in- 
strument midway  between  them.  Read  the  rod  on  each,  keep- 
ing the  bubble  in  exactly  the  same  position,  preferably  at  the 
center.  The  difference  of  the  readings  is  the  difference  of  the 
heights  of  the  pegs,  no  matter  how  much  or  in  what  way  the 
level  may  be  out  of  adjustment.  Then  set  over  either  peg  and 
measure  the  height  of  cross-wires  above  top  of  peg. 

The  difference  of  heights  of  pegs,  added  to  this,  or  taken 
from  it,  according  as  the  instrument  is  over  the  higher  or  lower 
peg,  gives  the  height  of  the  cross-wires  above  the  other  peg. 
Set  the  rod  on  that  peg,  and  bring  the  horizontal  wire  to  that 
height  on  the  rod  by  the  leveling-screws,  keeping  them  at  a 
bearing.  Then  bring  the  bubble  to  the  center  by  raising  or 
lowering  one  end  of  the  level-tube. 

The  first  part  of  the  second  adjustment, — namely,  to  bring 


22  FIELD-MANUAL   FOR   ENGINEERS. 

the  level-bubble  and  line  of  collimation  in  the  same  plane, — 
also  the  third  adjustment,  should  be  made  as  heretofore. 

USE  OF  THE  LEVEL. 

The  instrument  should  be  set  up  firmly,  with  the  top  of  the 
tripod  as  nearly  level  as  may  be,  so  as  to  save  time  in  leveling 
and  the  wear  of  the  screws,  etc. 

The  setting  up  of  the  level  is,  of  course,  similar  to  that  of  the 
transit  already  described. 

The  bubble  should  then  be  brought  over  each  pair  of  level- 
ing-screws  successively,  and  leveled  in  each  position. 

Bring  the  wire  precisely  in  focus  by  the  eyepiece,  and  the 
object  distinctly  in  view  by  the  objective,  so  as  to  avoid  all 
"  traveling  of  the  wires  "  or  parallax. 

It  is  best,  where  practicable,  to  take  approximately  equal 
fore  and  back  sights,  so  as  to  eliminate  any  error  due  to  a  lack 
of  perfect  adjustment,  Athich  is  difficult  to  secure. 

For  precise  reading  the  rod  should  not  be  over  400  or  500 
feet  from  the  instrument. 

If  the  socket  of  the  instrument  sticks  in  the  leveling-head  so 
as  to  be  difficult  to  remove,  be  sure  that  the  instrument  is  un- 
clamped  and  the  leveling-screws  are  free.  Then  place  the  palms 
of  the  hands  under  the  wye-nuts  under  each  end  of  the  bar, 
and  give  a  sudden  upward  blow  to  the  bar,  and  take  care  also 
to  grasp  it  the  moment  it  is  free. 

To  ADJUST  THE  COMPASS. 

The  Levels. — First  bring  the  level-bubbles  into  the  middle 
by  the  pressure  of  the  hands  on  different  parts  of  the  plate: 
then  turn  the  compass  half-way  around.  If  either  bubble  runs 
toward  one  end  of  its  tube,  it  indicates  that  that  end  is  too 

high- 
Lower  it  by  loosening  the  screw  under  the  lower  end,  and 
tightening  the  one  under  the  higher  end,  until  the  error  is, 
by  estimation,  half  removed.  Level  the  plate  again,  and  repeat 
the  operation  until  the  bubbles  will  remain  in  the  middle  dur- 
ing an  entire  revolution  of  the  compass. 

The  Sight-ranes.— The  sights  may  next  be  tested  by  observ- 
ing through  the  slits  a  fine  hair  or  thread  made  exactly  verti- 
cal by  a  plummet. 


ADJUSTMENTS,  USES,  AND    CAKE    OF    INSTRUMENTS.     23 


If  either  slit  does  not  coincide  in  direction  with  the  hair  or 
thread,  it  must  be  made  to  do  so  by  filing  its  under  surface  on 
the  higher  side. 

The  Needle. — Having  the  eye  nearly  in  the  same  plane  with 
the  graduated  rim  of  the  compass  circle,  observe  whether  or  not 
the  ends  of  the  needle  cut  opposite  points  on  the  rim.  If  not, 
bend  the  center-pin  (by  means  of  a  small  wrench)  about  an 
eighth  of  an  inch  below  the  point  of  the  pin,  so  as  to  make  the 
ends  cut  opposite  points  on  the  rim. 

The  needle  now  may  be  supposed  to  occupy  the  position 
Npti,  the  pivot  p  not  being 
in  the  center  0.  Now 
keeping  the  needle  in  the 
same  position,  turn  the 
compass  half-way  around. 
The  needle  will  now  oc- 
cupy the  position  N'p'S'. 
Correct  half  the  error  by  El 
bending  (that  is,  straigJit- 
CH'UH/)  the  needle,  making 
it  cut  points  half-way  be- 
tween its  former  positions, 
so  that  it  occupies  the  posi- 
tion Ap'B  and  is  straight; 
and  correct  the  other  half 
by  bending  the  pin,  placing 

the  point  of  the  pin  at  0  and  giving  the  needle  the  position 
NOS. 

The  operation  should  be  repeated  until  perfect  reversion  is 
secured  in  the  first  position. 

Then  try  the  needle  on  another  quarter  of  the  circle,  and  if 
an  error  is  manifested,  correct  the  center-pin  only,  the  needle 
being  already  straightened  by  the  previous  operation. 

Do  the  same  on  other  quarters  of  the  circle  until  the  needle 
will  reverse  in  any  position. 

To  USE  THE  COMPASS. 

In  using  the  compass  keep  the  south  end  toward  the  person, 
and  read  the  bearings  from  the  north  end  of  the  needle. 
Mark  every  station  or  point  at  which  the  compass  is  set,  so 


24  FIELD-MANUAL   FOll   ENGINEERS. 

that  it  may  be  easily  found  for  verification  or  a  resurvey.  It  is 
much  more  important  to  have  the  compass  level  laterally,  or 
crosswise  of  the  sights,  than  in  their  direction;  since  if  it  is 
not  so,  on  looking  up  or  down  hill  through  the  lower  part  of 
one  sight  and  the  upper  part  of  the  other  the  line  of  sight  will 
not  be  parallel  to  the  N.  and  S.  or  zero  line  on  the  compass, 
and  an  incorrect  bearing  will  be  obtained. 

A  continuous  line  thus  run,  to  say  nothing  of  other  im- 
perfections of  the  compass,  would  be  a  zigzag  line  probably 
very  much  in  error. 

The  compass  cannot  be  leveled  by  the  needle,  for  the  dip  of 
the  needle  is  continually  varying.  If  the  needle  touches  the 
glass  when  the  compass  is  leveled,  balance  it  by  sliding  the  coil 
of  wire  along  it. 

The  vibrations  of  the  needle  may  be  checked  by  gently  rais- 
ing it  off  the  pivot,  so  as  to  touch  the  glass,  and  letting  it 
down  again,  by  the  screw  on  the  under  side  of  the  box. 

The  compass  should  be  smartly  tapped  after  the  needle  has 
settled,  to  destroy  the  effect  of  any  adhesion  to  the  pivot  or 
friction  of  dust  upon  it. 

The  glass  sometimes  becomes  charged  with  electricity  by 
carrying  it  against  clothing,  or  wiping  it,  etc.,  so  that  it  at- 
tracts the  needle  to  its  under  surface,  preventing  its  free  move- 
ment. 

The  difficulty  may  be  remedied  by  breathing  on  the  glass, 
or  touching  it  in  different  places  with  the  moistened  finger. 

Of  course  the  chain  and  all  other  metals  should  be  kept  away 
from  the  needle. 


CHAPTER  III. 


PLANE  TRIGONOMETRY. 


1.  Plane  Trigonometry  treats  of  the  relations  of  the  sides  and 
angles,  and  of  the  solution  of  plane  triangles. 

2.  Some  French  writers  divide  the  right  angle  into  100  degrees, 
the  degree  into  100  minutes,  the  minute  into  100  seconds,  etc. 
This  centesimal  system,  in  which  reductions  'are  made  by  simply 
moving  the  decimal  point,  is  altogether  preferable  to  our  sexagesi- 
mal system. 

3.  Two  angles  whose  sum  is  equal  to  90°  are  complementary. 
Two  angles  whose  sum  is  180°  are  supplementary. 

4.  Let  us  consider  a  series  of  right  triangles  ABC,  AB'G',  etc., 
having  the  common  angle  A.     The  triangles  are  equiangular  and 
therefore  similar,  and  we  have 

BC      BG'      B"G" 


AB  ~  AB' 


AB" 


BG 
AG 


B'C'      B"C' 


AG 


AB  _  AB' 

AC  ~  AC' 


AC' 


AB" 


FIG. 


Thus  it  appears  that  the  ratios  of  the  sides  are  the  same  in  all 
right  triangles  having  the  same  acute  angles;  and  therefore  if 
these  ratios  are  known  in  any  one  of  these  triangles,  they  will  be 
known  in  all  of  them. 

As  any  triangle  may  be  divided  into  two  right  triangles,  it  is 
evident  that  the  solution  of  oblique  triangles  may  be  made  to  de- 
pend upon  the  solution  of  right  triangles. 

The  above  ratios,  depending  upon  the  angle  alone  and  not  at  all 
upon  the  absolute  lengths  of  the  sides,  may  be  considered  as  indices 

25 


2G  FIELD-MANUAL   FOR  ENGINEERS. 

of  the  angle,  and  have  received  special  names,  which  we  will  ex- 
plain. 

5.  Let  us  represent  the  sides  and  angles  of  a  triangle  in  the 
usual  way,  shown  in  Fig.  5.  The 
side  opposite  an  angle,  divided  by 
the  hypothenuse,  is  called  the  sine  of 
that  angle.  Thus 


—  =  sin  A, 
c 


and     -  =  sin  B. 
c 


The  side  opposite  an  angle,  divided  by  the  adjacent  side,  is  called 
the  tangent  of  that  angle.     Thus 


The  hypothenuse,  divided  by  a  side  adjacent  to  an  angle,  is 
called  the  secant  of  that  angle.     Thus 


v  . 

—  =  sec  A; 
b 


-  =  sec  3. 
a 


6.  The  terms  cosine,  cotangent,  and  cosecant  are  convenient 
abbreviations  for  the  "sine  of  the  complement,"  "tangent  of  the 
complement,"  and  "  secant  of  the  complement, "  respectively.  The 
reader  must  not  suppose  that  there  is  such  a  thing  or  entity  as  co- 
sine, cotangent,  or  cosecant  of  an  angle. 

Since  the  acute  angles  of  a  right-angled  triangle  are  comple- 
mentary, that  is,  A  =  90°  —  B  and  B  =  90°  —  A,  it  follows  that 

cos  A  means  the  sine  of     (90°  —  A),  or  sin  B; 
cot  A  means  the  tangent  of  (90°  —  A),  or  tan  B; 
cosec  A  means  the  secant  of  (90°  —  A),  or  sec  B,  etc. 


Hence 

sin  J.  = 


cos  5=; 

c 


tan  A  —      cot  B  =•  —  > 


cos  A  =  sin  B  =  — ; 


cot  A  =  tan  B  =  — ; 
a 


(4) 


sec  A  —  cosec  B  =  r ;    cosec  A  =  sec  B  =  - . 
h  a 


PLANE   TRIGONOMETRY. 


Sin  ,4  =  cos  B  is  really  an  identical  equation,  since  cos  B  is  the 
sine  (90°—  B)  =  sin  A;  and  so  is  tan  A  =  cotB,  etc.  Furthermore, 

c  —  b 
vers  A  =  =  1  —  cos  ^L. 


exsec  A  =  - — ; —  =  sec  A  —  1  = 


cos  A 


1  —  cos  A      vers  A 


cos  J. 


cos  A' 


B 


7.  Plane  trigonometry,  applied  to  plane  triangles,  is  but  the  ap- 
plication of  the  one  well-known  proposition  in  geometry:  "  Equi- 
angular triangles  have  their  homologous  sides 

proportional  and  are  similar"." 

Comparing  Figs.   5  and  6,  we   observe  that 

the  above  ratios  -,  -,  etc.,  change  as  the  angles 

A  and  B  change.  These  ratios  have  been  com- 
puted, however,  for  all  values  of  the  angles, 
differing  by  single  minutes  or  less,  and  placed 
in  tables  under  the  corresponding  headings, 
sine,  tangent,  etc. ,  and  opposite  the  correspond- 
ing  angles. 

8.  From  the  above  equations  we  observe  that 


b 
FIG.  6. 


Hence 


(6) 
(7) 


sin  A  cosec  A  =  sec  A  cos  A  =  tan  A  cot  A  =  1.     ,    (8) 
We  also  have,  by  division, 

sin  A       a 


cos  A       b 


3  FIELD-MANUAL  FOR  ENGINEERS. 

Again, 

sin' 4  +  cos' ;t=^+i' =<;=!.       .     .     .     (11) 
Again, 


6»  62 

and 


sec  A  =  4/1  -f  tan2  A ; (12) 

cosec2  A  =  sec2  jB  =  —  = —  =  1  -f-  —  =  !-}-  cot2  ^4, 

and 

cosec  A  = 
From  (9), 

tan 
sin  A  =  cos  A  tan  J.  = 


From  (10), 

cos  J.  =  sin  J.  cot  J.  — 


cosec  . 


If  one  of  the  above  functions  of  the  angle  A  is  given,  all  the 
others  may  easily  be  found.  For  example,  if  sin  A  is  given,  we 
have,  from  (11), 


cos  A  =  |/1  —  sin5  A. 
Then 


sm  A  sin  A 

tan  J.  = T=  -  =  J  •     .     =     •     (16) 

cos  J.         /i  _  sins  j. 


sm  A  sin 


——  =  -—  :          —        ... 
cos  A       4/1  _  sin2  A 


(18) 


PLANE   TRIGONOMETRY. 


If  cos  A  is  given,  we  have,  from  (11), 


sin  A  —  4/1  —  cos*  A. 

Then  tan  A,  etc.,  as  above. 

Similarly  when  other  functions  of  the  angle  are  given. 

9.  The  sine  and  cosine  of  two  angles  being  given,  to  find  the 
sine  and  cosine   of  their  sum,    and  the 
sine  and  cosine  of  the  difference  of  their 
angles. 

In  Figs.  7  and  8  let  EOF  =  A,  and 
FOE=B;  then,  in  Fig.  7,  EOG=A+B, 
and,  in  Fig.  8,  EOO  =  A  -  B.  From 
any  point  F  in  OF  draw  FE  perpen- 
dicular to  OF;  also  EG  and  FH  per- 
pendicular to  OH,  and  FK  perpendicular 
to  EG.  Now  the  three  sides  of  the  tri- 
angle FEK  are  perpendicular  to  the  three  sides  of  the  triangle 
FOH,  and  therefore  FEK  =  FOH. 

No\v,  in  Fig.  7, 


FIG. 


!„  (A 


EK 


EO 


But 


and 


FH      FH   FO 


Ji/Ji 


=  cos  A  sin  B. 


.  '.     sin  (  A  +  B)  =  sin  A  cos  B  -f-  cos  A  sin  B,    .     . 
Again, 

.       OG      OH-KF     OH      KF 


.     (20) 


OR    OF       KF   EF 


=  C°S  A  Cos  3  ~  sin  A 


30  FIELD-MAKUAL   FOR  ENGINEERS. 

Then,  in  Fig.  8, 

_  EG  _  HF-KE_  HF     OF  _  KE    FE 
sm(A-B)=~gQ-        EQ       ~  OF' ^EQ~  ~YE'~EO 


=  sin  A  cos  B  —  cos  A  sin  B\ 


(22) 


OG 
cos  (A-B}  =         = 


FK_OH    OF^       FK^    FE 

1      ~   OF'  E0~^~  FE  '  OE 


—  cos  A  cos  B  +  sin  A  sin  B. 


(23) 


0  H          G 

FIG.  8. 

In  (20),  make  B  =  A  and  get 

sin  2A  =  2  sin  A  cos  A.  (24) 
In  (21),  make  B  =  A  and  get 
cos  2  A  =  cos2  A  —  sin2  A 

=  (1  —  sin2  A)  —  sin9  A  —  1  —  2  sin8  A. 
=  cos2  J.  -  (1  -  cos2  A)  =  2  cos2  ^.  -  1. 

(20)  and  (21)  give 

sin  A         sin  B 

sin  (J.  -j-  B)  _  sin  .4  cos  7? -f  cos  AsinB_      cos  A         cos  B 
cos  (A  +  B)  ~  cos  .4  cos  B  —  sin  .A  sin  B 


1  - 


cos 


X 


sin  B  ' 


or 


tan  ( A  -f  B)  = 


tnn 


tan  B 


I  —  tan  A  tan 


Similarly,  from  (22)  and  (23), 


tan  (A  ~  B)  =  ~ 


tan  A  —  tan  B 


1  -|-  tan  A  tan  B 


(27) 


(28) 


PLANE   TRIGONOMETRY.  31 

We  note  some  special  values  of  the  trigonometric  functions. 
See  Fig.  9. 

Let  COP  represent  any  triangle  ;  angle  COP  =  0.  Let  0  =  0. 
Then  CP  =  0,  PO  and  BO  coincide  and  are  equal. 

Hence 

SinO  =  ^  =  0,      tanO  =  JL  =  0,      «c  0  =  ™  =  I. 

Let  0  =  90°.    Then  PC  and  PO  coincide  with  B'O,  and  CO  =  0. 
Hence 

~£)f  f\  T>  ' /")  T>/  /^) 

sin  90°  =  —r  =1,     tan  90°  =  — —  —  oo ,     sec  90°  =  ~-  =  oo . 

_O    U  0 

Let  0  =  180°.    Then  PO  and  CO  coincide  with  B"0  and  PC  =  0. 
Then 

sin  180°  =^L  =  0,     tan  180°  =  ^  =  0,     sec  180°=  |^  =  1. 

Let  0  =  270°.  Then  PO  and  PC  coincide  with  B'"0  and  CO  =  0. 
Hence 

7?"'O  7?//7O  Ti"fO 

sin  270°  =  57^=1,     tan  270°=  —  -- =00  ,     sec  270°  =^r-^=  oo  . 
1>     U  0 

The  values  of  the  functions  of  360°  are  the  same  as  those  of  0°. 
We  need  not  consider  angles  greater  than  180°. 

SOLUTION  OF  PLANE  RIGHT  TRIANGLES. 

16.  In  order  to  solve  a  plane  right  triangle  it  is  only  necessary 
to  select  from  equations  (4)  an  equation  containing  the  two  given 
parts  aside  from  the  right  angle  and  the  part  sought.  By  trans- 
posing, if  necessary,  so  as  to  express  the  latter  in  terms  of  the 
former,  it  becomes  known.  There  are  two  cases  : 

CASE  I. — A  side  and  an  angle  given.  Given  A  and  c.  See 
Fig.  5. 

Example.— Let  A  =  35°  23',  and  c  =  874.8.     We  have 

B  =  54°  37'. 
Also  table  of  sines  and  cosines  gives 

sin  35°  23'  =  .57904,     and    cos  35°  23'  =  .81530, 

Hence  a  =  c  sin  A  =  874.8  x  .57904  =  506.54; 

b  =  c  cos  A  -  874.8  X  .81530  =  713.22. 


FIELD-MANUAL  FOR  ENGINEERS. 


CASE  II.— Given  two  sides. 

Example.— -Let  a  =  184.3,  and  c  =  246.     We  have 


sin  J.  =  _=  .74919. 
c 


We  find  in  the  table 

sin  48°  31'  =  .74915,      .•.  A  =  48°  31'  to  the  nearest  minute; 
B  =  90°  -  A  =  41°  29';    &  =  c  cos  A  =  246  X  .6624  =  162.95, 


b  =  Vc2  -  a8  =  162.95. 
TABLE  FOR  SOLUTION  OF  RIGHT  TRIANGLES. 


Give-u. 

Required. 

Formulas. 

1.  a,   b 

A,  B,  c 

tan  A  =  -, 
b 

B  = 

90°     -     Ay 

c  =  b  sec  ^1  =a  sec  .Z?. 

2.  a,    c 

A,  B,  b 

a 

sin  A  =  —  , 

B  = 

90°  -  A, 

5  =  c  cos  .4  =a  tan  B. 

3.  A,  a 

B,  b,  c 

B  =  90°  - 

Ay  b 

=  a  tan  B 

,  c—  <zsec  j5=&secu4. 

4.  A,  b 

B,  a,  c 

B  =  90°  — 

A,  a 

=  b  tan  A 

,  c=a  sec  _/?:=&  sec  A. 

5.        Ay      C 

B,  a,  b 

B  =  90°  - 

A,  a 

=  c  sin  A 

,  b=at&n.  B=ccosA. 

If,  in  the  second  case  above,  b  is  given  in  place  of  a,  then  a 
and  5  as  well  as  A  and  B  change  places  in  the  formulas.  However, 
A  =  90°  —  B  is  the  same  as  B  —  90°  —  A. 

17.  We  will  now  deduce  formulas  for  the  solution  of  oblique 
triangles. 

Draw  BD  in  Fig.  10  perpendicular  to  AC.  Then,  from  the  tri- 
angle ABD, 

BD  =  c  sin  A,     .     (29) 
and,  from  the  triangle  BCD, 
BD  =  a  sin  C.     .     (30) 

Also  CD  •=.  AC  —  AD  —  b  —  c  cos  A.     .     .    .     (31) 


PLAKE   TRIGONOMETRY. 

Equating  (29)  and  (30)  gives 

a      sin  A 

c  sin  A  =  a  sin  c,     or     —  =  -: — ~. 
c       sin  6 


Similarly,  or  by  analogy, 

a  _  sin  A 
b~  sin  B' 

From  the  figure, 


(32) 


b       sin  B 
and    c=^C'  J 


+  (CD)*. 
Substituting  for  BD  and  CD  from  (29)  and  (31),  we  have 

a'2  =  c1  sin2  A  +  62  —  2bc  cos  -4  -|-  c2  cos2  A\ 
or,  since  sin2  A  -(-  cos*  4  =  1, 

a».  —  £2  _|_  c->  _  2£c  cos  ^}     or    COB  A  ~ 
Similarly,  or  by  analogy, 

cos  B  =          ^C ,     and 


From  (29)  and  (31)  we  also  have* 

BD          c  sin  A 

tan  G  =  7=^  =  T- 


sin 


6  -  c  cos  A       b 

cos  A 


tan  5  = 


tan  J.  = 


sin  (7    . 

a 

- —  cos  U 

0 

sin  B 


(34) 


c 
a 

18.  Let  ABC,  Fig.  11,  represent  a  plane  triangle,  the  parts  be- 
ing represented  as  usual. 

Take  GE  =  CA,  and  draw  AD  and  EH  perpendicular  to  AE. 
We  have 

CAB  +  OEA  =  180°  -  (7  =  4  +  7?. 

and       £4#  =  (7^4  -  GBA  =  l(A  +B)  -  B  =  ±(A  -  B). 


FIELD-MANUAL   FOR   ENGINEERS. 


Also      GAD  =  90°  -  CAE,    and     CD  A  -  90°  -  (AEC '=  CAE). 
Hence  CD  =  AC  =  b. 

Now 

a  +  b      BD      AD      AEi&n  \(A  +B)      tan 


JEET 


.    .    (35) 


FIG  11. 


From  triangle  ABE, 


or 


BE  _  sin  BAE 
AB~  jsin  AEB' 

a  —  b  _  sin  \(A  —  B) 

c      ~  sin  \(A  +  B) 

In  triangle  ABD, 

BD  _  a  -f-  b       sin  BAD       cos  |(^4  —  B) 
AB  ~      c      ~  sin  ^IJW?  ~~  cos  ^(^4  +  B)' 


(36) 


•     (37) 


Eq.  (37)  divided  by  (36)  also  gives 


tan 


+  B) 


tan 


—  B)' 


(38) 


which  furnishes  another  demonstration  for  (35). 

A  slight  variation  of  the  above  solution  of  the  tangent  problem 
was  given  by  the  author  in  Vol.  I,  No.  1,  of  The  American  Math- 
ematical Monthly.  It  has  since  found  its  way  into  text-books  on 
trigonometry.  Still  another  solution  by  the  author  may  be  seen 
jn  Vol,  III,  No,  11,  of  the  same  journal, 


FLAKE   TRIGONOMETRY.  35 


SOLUTION  OF  PLANE  OBLIQUE  TRIANGLES. 

19.  There  are  three  cases. 

CASE  I.  —  In  this  case  two  of  the  given  parts  are  a  side,  and  the 
angle  opposite  ;  the  other  part  being  either  a  side  or  an  angle. 
Example  1.  —  Let  A,  a,  and  B,  Fig.  10,  be  given. 

C  =  180  -  (A  +  B). 
From  (32), 

sin  B  sin  C 

b  =  a  —.  —  —  ;    similarly    c  =  a  —  :  —  -T-. 
sm  A  sm  A 

Example  2.  —  A,  a,  and  b  given. 

(32)  gives       sin  B  =  sin  A  —  ;     then  the  table  gives  B. 
Now          C  =  180°  -  (A  +  B)  ;    and    c  =  a       --. 


CASE  II.  —  Given  two  sides  and  the  included  angle. 
Let  b,  c,  and  A  be  given.     We  have,  from  (34), 


sin  A 
tan  C  = 


Then 


cos  A 

c 


=  ISO8  —  (4  +  <7) ;     and    a  =  b  ~. 


Or,  from  (33), 

a  =  (&'  -f  e*  -  2bc  cos 


Then 


sin  B  =  sin  A  —  ,    and    C  —  180°  —  ( A  -f-  B). 


CASE  III.  —  Given  the  three  sides  a,  b,  and  c. 
Eq.  (33)  gives 


COS  J.  = 


36  FIELD-MANUAL  FOK   ENGINEERS. 

Then 

sin  B  =  sin  A-,     and     C  =  180°  -  (A  +  B). 


The  above  formulas  are  all-sufficient  for  all  practical  purposes. 
This  chapter  constitutes  a  complete  treatise  on  trigonometry, 

though  the  deductions  from  it  are  endless,  as  the  examples  in 

arithmetic  are  endless, 

TABLE  FOB  SOLUTION  OF  OBLIQUE  THIANGLES. 

(See  Fig.  10.) 


Given.         Required. 

Formulas. 

6.  A,B,a  C,  b,c 
7.  A,  «,  b  B,C,c 
8.  A,  b,  c  B,G,  a 

9.  a,  b,  c  A,B,C 

asinB 

a  sin  (A  -f  B) 

sin  A 
b  sin  A 

sin  A 
a  sin  (A  +  B) 

a 
sin  A 

sin  A 
b  sin  A 

tan  B   —  • 

-  —  cos  A 
b 

a*-(b-cy 

~    sin  B  ' 

(a+b-c}(a+c-b) 

2bc 
b  sin  A 

2bc 

a 

Use  the  first  form  for  vers  A  with  a  table  of  squares,  the  second 
without ;  they  are  the  best  formulas  known  for  this  case. 

The  following  are  the  best  formulas  known  for  the  area. 
10.  Area  lab  sin  C  —  %ac  sin  B  =  \bc  sin  A. 

It  is  never  necessary  to  compute  but  one  unknown  part,  and  in 
the  3d  case  none  at  all,  to  have  the  required  data  for  one  of  these 
equations  ;  and  the  computation  is  shorter  than  by  any  other 
formula. 

Observe  that  a  sin  C  is  equal  to  the  perpendicular  from  Bupon  b, 
b  sin  G  is  equal  to  the  perpendicular  from  ^1  upon  <(,  etc. 


PLANE    TLUGONOMKTRY.  37 


TRIGONOMETRIC  FORMULAS. 

tl.  sin  A       =  4/1  —  cos'J  A  — —  2  sin  ^A  cos  ±A. 

cosec  A 

12.  cos  A       —  \'\  —  sin*  A  =  —     -   =  cos2  IA  -  sin'2  \A 

sec  4. 

=  2  cos2  ^  -  1  =  1  -  2  sin2  \A. 

13.  tan  A      — ;  = r  =  cosec  2  A  —  cot  2 A 

cos  A        cot  .4 

/  1  -  cos  2.4  sin  2/1 


sin  2  A  1  +  cos  2^  ' 

14.  cot  ^4.       —  cosec  2  A  +  cot  24  —  the  reciprocal  of  any  expres- 

sion for  tan  A. 

15.  sec  A      —  --  —  the  reciprocal  of  any  expression  for  cos  A. 

cos  A 

16.  cosec  A   —  —  -  -  -  —  the  reciprocal  of  any  expression  for  sin  A. 

sin  A 

17.  vers  A    =  1  —  cos  A  —  2  sin2  \A. 

18.  exsec  A  —  sec  J.  —  1  =  —   -p. 

cos  4 

19.  sin  (.1  ±  B)  =  sin  .1  cos  5  ±  cos  A  sin  5. 

20.  cos  (A  ±  B)  =  cos  A  cos  .5  T  sin  A  sin  Z>. 

21.  sin  A.  +  sin  Z?  =  2  sin  |(4  +  5)  cos  \(A  —  B). 

22.  sin  A  —  sin  I?  —  2  cos  |(/1  -f  B)  sin  |(.l  —  7?). 

23.  cos  A  +  cos  B  =  2  cos  |(A  -f  .#)  cos  |(4  -  1?). 

24.  cos  B  —  cos  A  =  2  sin  \(A  +  5)  sin  £(J.  -  B). 

25.  sin2  4  -  sin2  B  =  cos2  B  -  cos2  /I  =  sin  (A  +  5)  sin  (.4  -  1?). 
20.  cos2  A  —  sin2  5=  cos  (4  -f-  B)  cos  (4  —  B). 

27.  tan  A  ±  tan  B=sin 

28,  cot  4  ±  cot  7?  = 


COS  4.  COS  5 

si"  <A  ±  B> 


38  FIELD-MANUAL    FOR   ENGINEERS. 

The  above  formulas  contain  the  practical  general  relations  exist- 
ing among  the  functions  of  an  angle.  By  writing  \A,  2A,  etc., 
in  place  of  A,  by  repetitious,  etc.,  the  formulas  may  be  greatly 
multiplied  without  producing  any  new  relations.  This  practice 
is  too  common  in  Trigonometries,  Field-books,  etc. 


CHAPTER   TV. 


SIMPLE  CURVES  CONNECTING  RIGHT  LINES. 

LET  ABODE  represent  a  circular  arc  joining  the  straigLt  lines 
A  V  and   E  V,   which  are 
tangent    to   the 


C' 


curve    at 
A  and  E. 

AV  and   JSV  are    tan-    V 
t^ents  to  the  curve,  A  and 
Z£are  tangent  points,  and 
the  angle  KVE  is  the  an- 
gle   of    intersection,    and 
shows  the  change  of  direc-    c 
tion  in   passing  from  one 
tangent  to  the  other. 

Vis  the  point  of  inter- 
section,  or  vertex. 

PROPERTIES  RELATING 
TO  THE  CIRCLE. 

The  following  proposi- 
tions rest  upon  elementary 
geometrical  principles, 
may  be  regarded,  for  the  most  part,  as  axiomatic. 

(a)  A  tangent  to  a  circle  is  perpendicular  to  the  radius,  at  the 
point  of  contact. 

(&)  Tangents  drawn  to  the  circle  from  the  same  point  are 
equal,  and  the  angle  between  these  tangents,  and  the  chord  join- 
ing the  tangent  points,  are  equal.  Thus, 

AV  -  EV,     and     VAE  -  VEA. 

(c)  The  central  angle  AOE,  subtended  by  a  chord,  is  equal  to 
the  exterior  angle  KVE  between  two  tangents  to  the  curve  at  the 
extremities  of  the  chord. 

(d)  The   angle   between   a  tangent  and  chord  is  equal  to  the 
angle  subtended  at  the  circumference  by  the  same  or  by  an  equal 
chord,     Thus,  VAB  -  BCA  =  BAG,  etc. 

39 


40 


FIELD-MANUAL   FOR    ENGINEERS. 


(e)  An  angle  between  a  tangent  and  chord,  or  an  angle  subtended 
at  the  circumference  by  that  chord,  is  equal  to  one  half  the  central 
angle  subtended  by  the  same  chord.  Thus, 


(/)  Equal  chords  subtend  equal  angles  at  the  center  of  a  circle, 
and  also  at  the  circumference.     Thus,  if  AB  --  BC,  etc., 

ACS  =  BAG,  etc.,     and    AOB  =  BOG,  etc. 


7 


FIG.  13. 


radius    perpendicular   to  a  chord  bisects  the  chord,  and 
also  the  angle  and  the  arc  subtended 
by  the  chord.     Thus,   if   OG  is  per- 
Px_^_ xN.T  pendicular  to'AE,  we  have 


AM  =  ME,     AOV  =  EO  V, 

and  AC=GE. 

(h}  Parallel  chords,  or  a  tangent 
and  parallel  chord,  intercept  equal 
arcs.  Thus,  in  Fig.  13,  if  PT,  P'T', 
and  P"I "  are  parallel,  then  PP" 
and  T'T"  are  equal,  also  PP'  and 
TT'  are  equal. 
(i)  The  exterior  or  deflection  angle  KBG,  Fig.  14,  between  any 

two  chords  AB  and  BG  is  equal  to  half  the  central  angle  AOG, 

subtended  by  the  chords.     This 

is  easily  shown  as  follows  : 

1.  Join   AC.      BAG  =  %BOC,  /     \i 
and   BGA   =  %AOB,    as    stated 

above.     But  /  / \C 

KBC=  BAG+BCA- 
.-.    KBC  =  ^AOB  +  iBOC 
=  \AOC. 

2.  Draw   the    tangent    EBL. 
Then 

KBC  =  KBL  +  LEG 
=  NBA  +  LBG 

Fio.  H. 


SIMPLE    CURVES    CONNECTING    RIGHT    LINES.          4l 

J3.  Since  the  chords  ^47?  and  BC  are  parallel  to  tangents  drawn 
at  the  'middle  of  the  arcs  AB  and  B C,  it  is  evident  that,  in  passing 
from  one  chord  to  the  other,  we  turn  through  an  angle  measured 
by  one  half  the  arc  AB  -\-  one  half  the  arc  BC,  that  is^  through  an 
angle  equal  to  one  half  AOC. 

SOME  ELEMENTARY  RELATIONS. 

In  Fig.  12  drop  the  perpendiculars  Bb,  Cc,  etc.,  upon  the  tangent 
AY. 

Ab,  be,  etc.,  are  called  tangent  distances,  and  Bb,  Cc,  etc.,  tan- 
gent offsets. 

Prolong  AB,  making  Bd  = '  AB.     dC  is  called  the  chord  offset. 

Then,  from  the  preceding  article,  we  have 

CBd  =  ^AOC  =  AOB. 

Drop  the  perpendicular  Bp  upon  Cd.  This  bisects  the  angle 
CBd  and  the  base  Cd.  Hence 

CBp  =  dBp  =  iCBd  =  \AOB  =  BAb. 

Since,  therefore,  the  triangles  CBp,  dBp,  and  BAb  have  an 
acute  angle  in  each  equal,  and  the  hypothenuses  also  equal,  they 
are  equal  in  all  respects  ;  and  therefore 

Cp  =  pd  =  Bb  ;     also     Cd  =  2Bb. 

Since  Ab  is  tangent  to  the  curve  at  A,  Bp  is  likewise  tangent  to 
the  curve  at  B. 

Represent  the  radius  AO  by  R,  the  tangent,  or  vertex  distance, 
A  Fby  T,  and  the  angle  of  intersection  K  VE,  as  well  as  the  central 
angle  AOE,  by  V.  Represent  the  chords  AB,  BC,  etc.,  by  c,  a 
long  chord,  as  AC,  by  C,  the  tangent  distances  Ab,  be,  etc.,  by 
d,  di ,  etc.,  and  the  tangent  offsets  bB,  cC,  etc.,  by  t,  ti ,  etc.  Then 
the  chord  offset  Cd  =  2t.  Represent  the  vertex  distance  GVby 
E,  and  the  middle  ordinate  of  a  long  chord  by  M.  Let  L  represent 
the  length  of  the  curve,  P. C.  (Point  of  Curve)  the  beginning  of 
the  curve,  and  P.T.  (Point  of  Tangent)  the  end  of  the  curve. 

The  degree  of  a  curve  has  been  defined  as  the  number  of  degrees 


FIELD-MANtJAL   FOR 


subtended  at  the  center  by  a  chord  100  feet  in  length.  This  defi- 
nition of  curvature  is,  however,  awkward,  arbitrary,  and  false. 
It  is  founded  on  error;  it  involves  unnecessary  labor  and  ends  in 
anomalous  and  erroneous  results. 

The  degree  of  a  curve  may  be  defined  as  the  change  in  its  di- 
rection between  one  point,  and  another  100  feet  from  the  first, 
measured  on  the  curve.  This  is  the  change  in  direction  which 
one  would  make  in  moving  100  feet  on  the  curve  from  one  point 
to  another.  Or,  as  the  angle  subtended  at  the  center  of  the  curve 
liy  an  arc  100  feet  in  length. 

Let  D  ==  the  degree  of  the  curve. 

The  circumference  of  a  1°  curve  is  therefore  360  X  100  =  36000 

OfjAAA 

feet,  and  its  radius  is  ~— —  —  5729.578  feet,  almost  exactly. 

Since  a  2°  curve  changes  its  direction  2°  in  100  feet,  its  circum- 
ference is  only  half  that  of  a  1°  curve ;  and  since  the  radius  varies 
directly  with  the  circumference,  the  radius,  too,  is  only  half  that 
of  a  1°  curve.  For  the  same  reason  the  radius  of  a  3°  curve  is 
precisely  £  the  radius  of  a  1°  curve;  and,  generally,  the  radius  of 
a  D°  curve  is  exactly  equal  to  the  radius  of  a  1°  curve  divided 

57*^9  57S 
by  1);  it  is  therefore  equal  to  1~ — . 

Referring  to  Fig.  15,  we  see  that  for  the  same  central  anglo 
AOE,  the  arc,  the  tangent,  the 
chord,  the  middle  ordinate,  the  er.- 
ternal  secant,  etc.,  vary  directly  with 
the  radius  or  inversely  with  the  de- 
gree of  curvature.  For  example  •. 
If  aO  =  %AO,  then  ape  -  \APE, 
ai  =  \AY,  etc.  If  APE  is  a  1°  or 
2°  or  3°  curve  of  length  L,  then  ape 
is  a  2°  or  4°  or  6°  curve  of  length 

Hence  to  compute    any  function 
of  any  curve,  the  tangent  or  exter- 
0  nal,   for  example,   from    the  corre- 

FIG.  15.  spending  function  of  another  curve, 

it  is  only  necessary  to  multiply  or  divide,  as  the  case  may  be,  by 
the  ratio  of  their  radii  or  degrees  of  curvature. 

Example. — Find  the  tangent  of  a  7°  13'  —  433'  curve,  the  cen- 
tral angle  being  37°  50', 


SIMPLE    CUHVES   CONNECTING  EIGHT   LINES.          43 

The  tangent  of  a  1'  curve,  by  Table  VII,  is  117812.     Then 

T=  117812  H-  433  =  272.1. 
Using  a  table  giving  functions  of  a  1°  curve,  we  have: 

Tangent  of  a  1°  curve  =  1963.6. 
Then        Tangent  of  a  1'  curve  =  1963.6  X  60  =  117816. 

Finally,  :i!7816  •+•  433  =  272.1. 

Or,  13'  T*|  60  =  0°.216, 

arid  therefore  7°  13'  =  7°. 216. 

Then  1963.6  -4-  7.216  =  272.1. 

With  a  table  giving  functions  of  a  1°  curve  there  is  no  escape 
from  dividing  by  60,  which  division  is  obviated  by  giving  the  func- 
tions of  a  1'  curve  as  in  Table  VII. 

It  is  often  desirable  to  know  the  difference  in  the  lengths  of  a 
chord  and  its  subtended  arc,  and  for  this  purpose  we  deduce  the 
following  formula  : 

d  =  .001269239249^*  -  .0000000048329—.  (8) 

nA  nb 

IP  D4 

=  .001269 .0000000048^,    very  nearly,  (9) 

ns  n5 

Z>2 
=  .001269—,    approximately (10) 

In  these  equations  D  =  the  degree  of  curvature,  d  =  the  differ- 
ence between  any  chord  and  the  subtended  arc,  and  n  =  arc  of  100 
feet  divided  by  this  arc. 

For  an  arc  of  100  feet  n  =  1  and  d  =  .001269Z*2,  nearly;  .  (11) 
For  an  arc  of  50  feet  n  =  2  and  d  =  .000158D2,  nearly;  .  (12) 
For  an  arc  of  25  feet  n  =  4  and  d  =  .00002Z)'2,  nearly.  .  (13) 

For  n  sub-chords  per  station  the  sum  of  the  difference  per 
station  is 

nd  -=  .001269  — (14) 


44 


FIELD-MANUAL   tOlt   ENGINEERS. 


Representing  the  central  angle  of  the  curve  l>y  V,  and  tlie  num- 
ber of  stations  in  the  curve  by  A",  we  have  N  ;  and   hence  tin- 

length  of  the  curve  exceeds  the  sum  of  the  lengths  of  the  sub- 
chords  by 

7)2        V  VD 

E=  Nnd=  .001269=^  X  -~  =  .001269—,..      .     (14') 


Hence,  in  laying  out  curves,  —  should  be  nearly  constant. 

Since  in  a  4°  curve  the  chord  of  an  arc  of  100  feet  is  99.98  feet, 
curves  from  0°  to  4°  can  be  properly  laid  out  with  chords  of  100 
feet,  and  with  the  same  degree  of  accuracy  we  may  lay  out  curves 
from  4°  to  16°  with  chords  of  50  feet,  and  those  from  16°  to  04° 
with  chords  of  25  feet. 

Very  sharp  curves  can  be  easily  laid  out  by  swinging  a  chain 
around,  while  one  end  is  held  at  the  center  of  the  curve. 

Let  s  represent  any  arc,  c  its  chord,  and  et  the  chord  of  one  half 
of  the  arc  s. 

Then,  from  above, 


This  is  said  to  be  Huygens'  approximation  to  the  length  of  an 
arc. 

The  following  table  shows  the  differences  between  arcs  of  25 
feet  and  of  50  feet,  and  the  chords  of  those  arcs.  (See  Fig.  16.) 


De* 

D 

Arc 

25 

Arc 

50 

Deg. 
D 

Arc 
s>5 

Arc 

50 

Deg. 
D 

Arc 

25 

Arc 

50 

1 

.000 

.000 

11 

.002 

.019 

21 

.000 

.070 

.000 

.001 

13 

.003 

.023 

22 

.010 

.077 

3 

.000 

.001 

13 

.003 

.027 

23 

.010 

.084 

4 

.000 

.003 

14 

.004 

.031 

24 

.011 

.091 

5 

.000 

.004 

15 

.004 

.036 

25 

.012 

.699 

6 

.001 

.000 

16 

.005 

.041  . 

26 

.013 

.107 

7 

.001 

.008 

17 

.006 

.046 

27 

.014 

.116 

8 

.001 

.010 

18 

.COS 

.052 

28 

.01C 

.124 

9 

.002 

.013 

19 

.007 

.058 

29 

.017 

.133 

10 

.002 

.010 

20 

.008 

\m 

30 

.018 

.143 

bIMI'LE    CURVES    CONNECTING    KKIHT    LINES. 


45 


II.   Suppose  the  chord  AB  =  100  feet,  and  AOB  =  D,  (Fig.  16). 
Then 

AOM=^,  MAtt  =  —  ,  etc. 

2  4 

Hence 
AM  =  AEsec  MAE 

=  50  sec  ~i. 

4 
Therefore 


AM  -  50  =  50 (sec  -£  -  1  j 


=  50  exsec  — . 

Similarly,     AK  =  AFsvc  —  =  25  sec  — -  sec  — ,  etc.,  etc. 
o  4  o 

The  following  table  gives  the  excess  of  AK  over  25  feet,  and  of 
AM  over  50  feet,  when  chord  AB  is  100  feet. 


Deg. 

AK 

AM 

Deg. 

AK 

AM 

Deg. 

AK 

AM 

D! 

-25 

-50 

A 

-25 

-50 

DI 

-25 

-50 

1 

.000 

.000 

11 

.036 

.058 

21 

.131 

.211 

2 

.001 

.00-2 

1-2 

.043 

.069 

22 

.144 

.231 

3 

.003 

.004 

13 

050 

.('81 

23 

.157 

258 

4 

.005 

.008 

14 

.058 

.093 

24 

.171 

.275 

5 

.007 

.012 

15 

.067 

.101 

25 

.186 

.29?. 

6 

.011 

.017 

16 

.076 

.122 

26 

.201 

.323 

7 

.015 

.0:23 

17 

.086 

.138 

27 

.217 

.349 

8 

.019 

.030 

18 

.096 

.155 

28 

.233 

.375 

9 

.024 

.039 

19 

.107 

.172 

29 

.250 

.408 

10 

.030 

.048 

20 

.119 

.181 

30 

.268 

.431 

Comparing  the  preceding  tables  we  learn  that,  when  the  chord 
AB  is  100  feet,  the  chord  AM  differs  nearly  four  times  as  much 
from  50  feet  as  the  chord  of  the  arc  of  50  feet  differs  from  50  feet. 
Furthermore,  that  the  chord  AK  differs  nearly  sixteen  times  as 
much  from  25  feet  as  the  chord  of  the  arc  of  25  feet  differs  from 
25  feet. 

Thus  let  us  first  suppose  A  MB  (Fig.  16)  to  be  a  10°  curve,  and 
the  chord  ^47?  =  100  feet.  Then  by  the  table  AK  =  25.030  feet, 
and  this  would  lead  to  an  error  of  .030  X  40  —  1.2  feet  in  laying 
out  a  curve  25  X  40  =  1000  feet  long,  taking  chord  AK  =  25  feet 
long. 


46  FIELD-MANUAL  FOR  ENGINEERS. 

Suppose,  secondly,  that  the  are  AMD  —  100  feet  and  therefore 
arc  A K  —  25  feet,  and  the  chord  AK  is  but  .002  of  a  foot  less  than 
25  feet. 

Hence  in  laying  out  the  above  curve,  taking  the  chord  AK '  —  25 
feet,  the  error  committed  would  be  only  .002  X  40  =  .08  of  a 
foot. 

Thus  we  see  that  when  the  chord  of  a  station  is  taken  =  100 
feet,  the  sub-chords  AM,  AK,  etc.,  differ  so  much  from  50  feet, 
25  feet,  etc.,  as  to  largely  vitiate  the  results,  whereas  such  is  by 
no  means  the  case  when  the  arc  AMB  is  taken  =  100  feet. 

Indeed,  when  the  chord  AB  is  100  feet,  the  shorter  the  sub- 
chords  used  in  laying  out  a  curve,  the  greater  the  discrepancy  in 
the  measurement,  on  this  basis.  Thus  for  a  10°  curve  the  difference 
between  four  equal  sub-chords  and  100  feet  is  .080  X  4  =  .120  of  a 
foot ;  whereas  the  difference  between  two  equal  sub-chords  and 
100  feet  is  .048  X  2  =  .096  of  a  foot.  The  reverse  is  of  course  the 
case  when  the  arc  AMB  is  made  the  standard  of  measurement. 

These  facts  are  evident ;  for  when  the  chord  is  made  the  stand- 
ard of  measurement,  the  sum  of  the  lengths  of  the  sub-chords  ex- 
ceeds more  and  more  the  length  of  the  chord,  the  shorter  they  are  ; 
whereas  when  the  arc  is  the  standard  of  measurement,  the  sum  of 
the  lengths  of  the  sub-chords  falls  short  of  the  length  of  the  arc 
less  and  less  the  shorter  they  are. 

Most  recent  writers  have  endeavored  to  obviate  the  inconven- 
iences and  inconsistencies  above  pointed  out  by  inconsistent  assump- 
tions, such  as  basing  the  curves  of  different  degrees  upon  chords 
of  different  lengths.  This  scheme  gives  the  values  of  some  of  the 
radii  quite  correct ;  but  it  causes  sudden  breaks  in  the  value? 
where  the  changes  are  made. 

For  example,  how  can  there  be  two  different  values  (819.02  an<* 
818.64)  of  the  radius  of -a  7°  curve? 

And  why  should  the  radius  of  a  7°  10'  curve  be  given  quite  cor- 
rect, while  that  of  a  6°  50'  curve  is  quite  incorrect? 

Again,  we  are  told  by  a  recent  author  that,  in  practice,  it  is  cus- 
tomary to  take  the  radius  of  a  1°  curve  as  5730  feet,  and  to  assume 
the  radius  to  vary  inversely  as  the  degree.  Thus  for  a  4°  curve 

the  radius  would  be  — j-  =  1432.5  feet.     This  is   rational   and, 
4 

moreover,  is  precisely  what  is  here  advocated,  except  that  the  true 
value  of  the  radius  of  a  1°  curve  is  used,  vi/.,  5729.58  feet. 


SIMPLE   CURVES   CONNECTING    RIGHT   LINES.         47 

FORMULAS. 
From  the  triangle  AOV,  Fig.  12,  we  have 

AV=AOt&nAOV,     or     T=Rtau^V.     .     .     (15) 
Transposing,  we  Lave 


From  the  triangle  AOF,  we  find 


A0  =     .Ani;,    or    R  =  -rn  =  &  cosec 
sin  ^107^  sin  \J) 

This  value  of  R  in  (15)  gives 


Measure  equal    distances  FIT"  and   VL  along  the   tangents  in 
Fig.  12  and  draw  HNL. 
Measure  #JVand  VN. 

A  V        VN  VN 


that  is,  T=R-.   ...  .....     (19) 

If  ^4J/  and  FJf  are  measured,  then 


Eqs.  (19)  and  (20)  serve  to  fix  geometrically,  without  measuring 
angles,  the  tangent  point  of  a  curve  of  given  radius  that  will  unite 
two  straight  lines  on  the  ground. 

In  Fig.  12  draw  VG'  perpendicular  to  A  V  to  meet  A  C  prolonged. 
Now  the  angle  AVG  —  \AVE  =  |(180  -  F)  =  90  -  |F.  Hence 

VCC'  =  AGO  =  90°  -  BOO  =  90°  -  4LF. 


48  FIELD-MANUAL    FOR    EXGJNKEkS. 

Also,   VG'C  =  90°  -  CA  V  =  90°  -  i  V; 

.-.     VC'C-  VCC',     and     VC'  =  VG  =  #. 
Hence 

7^7  /"fr   TT 

^-  =  -jy  =  laniF,     or     E  =  T  \&\\  ±V.    .     .     ("21) 

Substituting   T  —  ft  tan  IF  for  T,  from  (15),  gives 

^=  tftan^Ftan  £F.    ......    (21') 

Since  the  triangles  BCd  and   7?0(7are  similar,  \ve  have 
Gd        BC  BC* 


or 


and  Bb  =  t  =     ~  .........  (3:.) 

We  also  have 

56  =  AB  sin  7?^4&,     or     i  =  c  sin  |D,    .     .     .  (24) 

and  2t  =  2c  sin  \D  ........  (25) 

Example.  —  Given  R  =  1909.9  feet  and  c  —  100  feet,  to  find  the 
tangent  and  the  chord  offsets  for  100  feet. 
By  (23), 

1  0000 

=  2.618,    and    2t  =  5.236, 


or  Table  I  shows  that  1909.9  is  the  radius   of   a  3°  curve,   and 

\D  =  1°  30',     and     sin  \D  —  .02618. 
Then,  by  (24), 

t  =  .02618  X  100  =  2.618,     and    2t  =  5.236. 

The  tangent  offsets  are  given  in  Table  I,  and  the  chord  offsets 
are  twice  the  tangent  offsets. 


SIMPLE    CURVES    CONNECTING    RIGHT    LINES.          49 

In  laying  out  curves,  the  chain  is  stretched  from  point  to  point 
>n  the  curve,  and  coincides,  therefore,  with  chords  of  the  curve. 

Since  the  process  is  the  same  whatever  the  length  of  chain 
used,  we  will  assume  it  to  be  100  feet  long1. 


FIG.  17. 

The  length  of  the  curve  is  expressed  in  chains,  in  terms  of  the 
central  angle  AOF  =  V,  and  the  degree  of  the  curve  AOB  =  D. 

V 
The  number  of  chains  is  evidently  equal  to  —  . 

Thus,  in  the  figure,  if  V—  23°  and  D  =  5°,  the  curve  is  -2/  =  4f 
chains,  or  460  feet  long.  As  the  angle  AOE  =  5°  X  4  —  20°, 
EOF—  23°  —  20°  =  3°  ;  and  the  arc  EF  is  f  X  100  =  60  feet  long. 

It  is  usual,  in  laying  out  curves,  to  assume  the  radius  11,  and  to 
find  the  degree  of  the  curve  D  from  it  ;  or  to  assume  D  (usually 
in  degrees  and  minutes),  and  to  find  R  from  it.  Neither  way 
is  best. 

To  assume  a  value  of  R  or  of  D  does  not  aid  in  the  least  in 
properly  locating  the  curve.  Generally  the  surface  of  the  ground 
does  indicate  approximately  the  position  of  the  curve,  and  the 
proper  course  to  pursue,  therefore,  is  the  following:  Divide  the 
tangent  of  a  one-minute  curve  by  the  length  in  round  numbers  of 
the  desired  tangent,  and  neglect  the  decimal  in  the  quotient. 
This  give.s  the  degree  of  the  required  curve  in  minutes.  We  may 


..nzrociTV     11 


50 


FIELD-MANUAL    FOR    ENGINEERS. 


change  the  quotient  to  an  even  number  of  minutes,  or  to  some 
multiple  of  10  if  we  wish,  if  there  is  sufficient  latitude  to  be  taken 
in  the  position  of  the  P.C. 

Then  divide  the  tangent  of  a  one-minute  curve  by  the  corrected 
quotient  for  the  tangent  required. 

Example.  —  V—  46°  30',  and  the  tangent  should  be  1100  feet  or 
over.  Find  the  degree  of  the  curve,  and  the  length  of  the  tan- 
gent. 

Dividing  the  tangent  of  a  1   curve  by  1100  gives 

147697  +  1100  =  134'  =  2°  14'. 

Now  2°  10'  =  130',  and  147697-^-  130  =  1136.13,  the  tangent  re- 
quired. 

LONG  CHOKDS  AND  ORDINATES  TO  LONG  CHORDS, 

Let  A,  B,  C,  etc.,  represent  stations  upon  a  curve. 

Draw  the  lines  as  repre- 
sented, EX  being  a  perpendic- 
ular from  the  middle  of  the 
curve  E  upon  the  tangent 
AT.  Let  Jlf  =  EM,  M'  =  FN, 
etc. 

Draw  AE,  and  we  have  the 
angle  EA  M  equal  to  the  angle 
EAX,  and  therefore  the  tri- 
angles EAM  and  EAX  are 
equal  in  all  respects.  From 
this  we  learn  that  the  tangent 
offset  EX  for  any  arc  AE  h 
equal  to  the  versin  EM  of 
that  arc,  or  to  the  middle  or- 
dinate  EM  ot  the  chord  AK 
of  twice  the  arc. 

Or,  draw  the  tangent  Et  and  the  perpendicular  Kt  upon   it. 
Then  the  tangent  offset  Kt  of  the  arc  EK  =  the  versin  EM  of 
the  same  arc  =  the  middle  ordinal  e  EM  of  the  arc  AEK. 
To  find  the  middle  ordinate,  M.     (See  Fig.  18.) 


1°.  M  =  EM  =  EX  =  AM  tangent  EAM  =  \C  tan 
3°.  M  =  EM  =  EX-  ET  cos  TEX  -  #eos    V, 


.      (26) 
.     (26'  ) 


SIMPLE   CURVES   CONNECTING    RIGHT  LINES.          5l 

3°.  M  =  EO  -  MO  =  U  -  R  cos  |  F  =  #(1  -cos  £  F) 

=  R  vers  |  F.  .     .     .    (26") 

4°.  J/=  EO-MO=EO-  i/AO*-AM*=R-  \/ltr-iC-.   .  (26'") 

To  find  any  ordiuate  FN  distant  d  from  the  center  of  the  chord. 
Prolong  FN,  to  meet  OS,  drawn  parallel  to  AK,  and  join  FO. 


Now,     FS  = 


-  OS"  =  4/-R8  -  d*.  JVS  =  MO  =  R  -  M  . 


=  M+ 


-  -  d*  -  R. 


Other  methods    will   be   given    in  connection  with  laying  out 
curves  by  ordinates  from  a  long  chord. 


Approximate   Values  of  Ordinates  to  Short  Chords, 

^FG, 


Divide  the  chord  into  any  number 
of  equal  parts,  eight  for  example,  at 
z,  k,  ly  etc. ,  and  erect  the  ordinates 
Em,  Fn,  etc.,  and  prolong  them  to  A 
meet  the  curve  in  13',  F',  etc.  Let 
Em  =  m,  Fn  =  m',  etc.  We  have, 
from  geometry, 

_  Am  X  mB  __  \c  X  $c 
or,  approximately, 


»»=   2r-=55-  •  •  <27> 

6.K  oil 


E'F 
FIG.  19. 


^G 


JW 


Similarly, 


c  X  |c      15     c* 


^L 
WSR 


=  m    =-          =       .  ~  =      m  =  m  - 


Hq  =  m'"  = 


(28) 

(28') 
(28") 


For  any  other  equal  divisions  of  the  chord  we  have  similar  re-, 
suits. 


FIELD-MANUAL    FOR    ENGINEERS. 

tablisli  the  points  E,  F,  etc. 

Set  E  equally  distant  from  A  and  13,  and  at  the  distance  m  from 
the  point  m  ;  then  F,  equally  distant  from  k  and  13.  and  \im 
from  n  ;  then  G,  equally  distant  from  m  and  B,  and  £w*  from  p; 
lastly,  H,  equally  distant  from^)  and  B,  and  T7gw  from  q. 

This  method  involves  much  less  labor  than  that  of  drawing  sub- 
chords  to  find  the  points  F,  G,  etc.  If  we  draw  a  tangent  at  E, 
the  offsets  to  F,  G,  H,  etc.,  will  be,  according  to  the  preceding 
formulas,  fam,  yV^,  ~i%m>  e^c-  This  shows  a  convenient  way  of 
finding  the  points  on  the  curve. 

To  compute  tangent  offsets  and  middle  ordinates  by  means  of  a 
series. 

In  a  way  similar  to  that  pursued  in  finding  the  difference  be- 
tween an  arc  and  its  chord  we  find 

t  =  .872664625997Z>  -  .  000022 15240389.D3 

-f.00000000022493360386Z>5 -,  .     (29) 

or 

t  =  .872664626,0  -.000022152404£3-f-.000000000224933604Z)5.  (30) 

We  may  find  the  tangent  offset  for  m  stations  by  multiplying 
the  successive  terms  of  (30)  by  m1,  m4,  etc. 

We  thus  find  tang  offset  for  arc  of  50  feet,  or  —  stations, 

2 

t1  =  .2181G6156499Z)  -  . 000001 38452524D3 

+  .00000000000351458756Z)5-.     .     (31) 

For  1  foot  m*  =  (.Ol)'2  =  .0001,  m4  =  .00000001,  etc.     Then, 

The  tang  offset  for  1  ft. 

=  t,  =  .0000872664626Z)  -  .00000000000022152404D3  +  .  (32) 

From  (30),  we  have  t  =  .SID,  approximately. 

For  n  stations  tn  =  .87/iJX> (33) 

Wellington  in  Railway  Location,  ch.  xxx,  recommends  the 
equation  t  =  |»8A  ,  ,  , (33') 


SIMPLE   CURVES   CONNECTING   RIGHT  LINES.          53 

\vliicli  is  practically  the  same  as.  (33).  (33)  is  a  trifle  more  accurate 
than  (33'),  but  either  is  sufficiently  accurate  for  all  cases  in 
which  ii-D  does  no*  much  exceed  thirty. 

The  same  formuia  gives  the  middle  ordinate,  n  representing  the 
number  of  stations  on  either  side  of  the  center,  or  2n  the  number 
of  stations  in  the  arc.  Any  other  ordinates  desired  are  then  given 
by  eq.-?.  (27).  (28),  etc. 

Example.—  Find  six  offsets  of  a  3°  curve  at  points  50  feet  apart. 
(See  Fig.  18  ) 


For 

n 
n 
n 

n 
n 
n 

= 

1, 
1, 

t. 
2, 

2i, 
3, 

^    —   7 
t  = 

t   = 
t  — 

•  x 

X 
X 

\s 

x 

i 

1 

9 

•r 
4 

L'  5 
"T" 

6 

X 

x 

X 
X 

x 

X 

3 
:j 
3 
8 
3 
8 

1= 

9 

5 

10 

16 

2:' 

.66; 
.62; 
91; 
.50; 
.41; 
.62. 

These  results  are  of  course  equal  to  the  middle  ordinates  for 
1,  2,  3  ...  6  stations  of  the  same  curve. 

LAYING  OUT  CURVES. 

Since  in  laying  out  curves  the  operation  or  method  is  the  same 
whatever  the  length  of  the  chain  or  chord,  we  will  here  assume  it 
to  be  100  feet  long. 

A.     By  Deflection  Angles. 

Let  A  in  Fig.  17  be  the  P.C.  Set  the  instrument  at  A,  and 
turn  off  from  the  tangent  AVthe  given  deflection  angle  VAB  = 
\D,  D  being  the  degree  of  curve.  This  will  give  tlie  direction, 
AB,  and  measuring  100  feet  in  this  direction,  the  point  B  will  be 
determined.  Turn  off  the  additional  angle  B  AC  =  \T),  the  tele- 
scope being  now  directed  toward  C,  and  set  C  in  the  line  AC  &nd 
100  feet  from  B.  Turn  off  the  additional  angle  CAD  =  ^D,  and 
set  D  in  the  line  AD  and  100  feet  from  C.  Proceed  in  the  same 
way  for  other  stations  to  the  end  of  the  curve,  or  so  far  as  the 
stations  can  be  seen  from  A. 

It  is  usually  impossible,  on  account  of  obstructions  likely  to  be 
met  with,  to  lay  out  the  whole  of  a  curve  from  the  first  station. 
When  such  is  the  case,  we  determine  as  many  stations  as  conven- 
ient, remove  the  instrument  to  the  last  station  so  determined,  and 


54  FIELD-MANUAL    FOR   ENGINEERS. 

proceed  from  that  as  from  the  first  station.  For  example  :  Suppose 
B,  C,  and  D  to  be  found  with  the  instrument  at  A.  Remove  the 
instrument  to  Z),  sight  to  A,  turn  off  the  angle  ADV  '  =  DA  V  — 
\D,  and  the  line  of  sight  will  be  in  the  direction  of  the  new  tan- 
gent DV  Sit  D.  Reverse  the  telescope,  and  the  line  of  sight  will 
point  forward  along  the  same  tangent  VDG.  Now  set  E,  F,  etc., 
from  the  tangent  DG,  as  B,  C,  and  D  were  set  from  the  tangent 


In  setting  the  first  stake  from  the  new  position  of  the  instru- 
ment, as  E  from  D,  no  notice  need  be  taken  of  the  tangent  at  D, 
it  being  necessary  simply  to  turn  off  from  the  line  ADUihe  angle 
HDE  =  HDG  -f  GDE  =  AD  V+  GDE  -  4(iZ>). 

In  the  new  position  of  the  instrument  we  observe  that,  in  all 
cases,  the  deflection  from  the  line  pointing  to  the  back  station  to 
the  line  pointing  to  the  forward  station  is  as  many  times  the  de- 
flection angle  \D  as  there  are  chains  in  the  curve  between  the 
back  and  the  forward  station.  Of  course  this  applies  to  simple 
curves  only. 

The  beginning  of  a  curve,  as  well  as  the  end,  usually  falls 
between  regular  stations,  giving  short  chords  at  the  ends.  The 
deflection  angle  for  a  short  arc  is  such  a  part  of  the  full  deflection 
angle  as  the  short  arc  is  of  the  full  arc. 

Let  a  represent  the  arc  AB  or  BC,  etc.,  and  a,  represent  the 
arc  EF.  Also  EOF  =Di. 

Hence     DFE—^D,     and     EDF=$D1,     and  therefore 


—  =  —  ,     or 
\D        a 

In  beginning  a  curve,  the  instrument  being  at  the  first  station, 
it  is  convenient  to  place  the  zeros  of  the  instrument  plates  to- 
gether, and  direct  the  line  of  collimation  along  the  tangent  to  the 
curve.  The  reading  on  the  limb  for  any  station  will  then  be 
equal  to  the  total  deflection  angle  for  that  station. 

If  the  vernier  is  not  disturbed  while  laying  out  the  curve,  it  is 
plain  that  when  the  instrument  is  moved  to  its  second  position  D, 
Fig.  17,  and  the  line  of  sight  directed  to  A,  the  reading  will  be 
equal  to  the  total  deflection  from  A  to  D;  and  that,  an  additional 
angle  equal  to  this  deflection  being  turned  off,  the  reading  will  be 
equal  to  the  central  angle  AOD,  and  the  line  of  sight  will  be  in 
the  direction  of  the  tangent  at  D.  The  same  is  true  for  all  posi- 


SIMPLE    CURVES   CONNECTING   EIGHT   LINES.          ^ 

tions  of  the  instrument.  Since  the  deflections  for  the  last  section 
of  the  curve,  that  is,  from  the  last  position  of  the  instrument  to 
the  end  of  the  curve,  is  turned  off  but  once,  the  reading  on  the 
instrument,  when  the  curve  is  finished,  will  be  equal  to  the  total 
central  angle  V,  less  the  deflection  for  the  last  section.  This  fur- 
nishes a  convenient  check  for  the  work, 


B.  By  Tangent  Offsets.     New  Method. 

Let  ABCDEFGH  represent  a  curve  having  a  short  chord  AB 
--  c'  subtending   an  angle  AOB 
=  -  Di  at  one  end  of  the  curve. 

Define  the  tangent  A  V  by  set- 
ting stakes  upon  it.  Draw  Be, 
TX  etc.,  parallel  to  A  V,  and  BX, 
iJcX',  etc.,  perpendicular  to  AV, 
or  suppose  such  lines  to  be 
•drawn.  Now 

BX=ABsm£A  X=c'sm  \Di  =  t^ 

and  is  given  by  Table  II. 

Since  the  chords  BC,  CD,  etc., 
are  parallel  to  the  tangents  at  the 
middle  of  the  arcs  BC,  CD,  etc.,  FIG.  20, 

we  have 

CBc  =  D,  +  \D,     DCd  -  D,  +  \D,     EDe  =  A  -f-  ID,  etc. 
Hence 

Cc  =  c  sin  (D1  4-  |£),     Dd  =  c  sin  (D,  +  |Z>), 
Ee  =  c  sin  (D,  -f  |Z>),     etc. 

We  observe  that  Cc,  Dd,  Ee,  etc.,  are  respectively  the  tan- 
gent offsets  for  curves  of  the  degrees  (Di  -(-  ^D),  (Z>i  -f-  %D), 
(2>,  _|_  |/))>  etc.,  and  may  be  taken  from  Table  III.  Then 

CX'  =  BX  +  Cc,  DX"  =  CX'  +  Dd,  EX"'  =  DX'  +  Ee,  etc. 

Set  B  at  a  distance  c'  from  A,  and  ti  from  A  V;  then  C  a  dis- 
tance c  from  B,  and  CX'  from  AV;  D  a  distance  c  from  C,  and 
Dx"  from  A  V,  etc. 


FIELD-MANUAL    FOR   ENGINEERS. 


be  observed  that  in  this  method  we  avoid  constructing 
t  at  B,  in  consequence  of  the  short  chord  AE\  we  avoid, 
secondly,  the  finding  of  AX,  XX',  etc.;  thirdly,  the  errection  of 
perpendiculars,  at  X,  X',  etc.;  and,  lastly,  we  avoid  the  use  of 
the  radii  which  are  large  and  fractional.  Since  c  is  generally  100, 
though  sometimes  an  aliquot  part  of  100,  no  computation  is  gener- 
ally required,  and  none  to  mention  in  any  case.  Thus  we  see  how 
simple  and  short  this  method  is,  compared  with  the  method  of 
tangent  offsets  in  general  use. 

When  the  curve  begins  at  a  station  there  is  no  short  chord. 
Then 

A  B  =  c,  D!  —  D  ;  BX  =  c  sin  $D,  Cc  =  c  sin  |Z>,  Ed  —  c  sin  f  Z>, 

etc. 

It  is  best  to  lay  out  the  curve  from  each  end,  so  as  to  avoid  off- 
sets inconveniently  long.  Long  offsets  may  be  avoided  by  drawing 
a  tangent  at  any  station  and  continuing  the  curve  from  that  station 
precisely  as  from  A,  when  there  is  no  short  chord. 

To  draw  a  tangent  at  any  station,  draw  a  line  through  that  sta- 
tion and  at  a  perpendicular  distance  from  an  adjacent  station  equal 
to  the  tangent  offset  for  a  station  =  c  sin  \D.  Or,  draw  it  parallel 
to  the  chord  joining  adjacent  stations. 

It  will  be  noticed  that  stations  on  the  curve  are  not  opposite 
stations  originally  set  on  the  tangent. 

The  length  of  the  curve  gives  the  number  of  the  station  at  H. 

This  is  perhaps  the  best  of  all  methods  without  a  transit ;  but 
a  combination  of  methods  is  sometimes  advisable,  as  will  be  shown 
further  on. 

We  also  have 

AX—  CCOS^D!  ,  XX'  —  Be  =  c  cos  (Dl  -f  ^D),  etc., 

or  XX ',  X'X",  etc.,  are  respectively  equal  to  the  tangent  distances 
for  one  station  of  curves  of  the  degrees  D\  -f-  \D,  D\  -f-  f  Z),  etc. 

These  quantities  are  not  needed,  however,  and  it  is  to  be  ob- 
served that  the  points  X,  X',  etc.,  are  not  established  or  used. 


SIMPLE    CURVES    CONNECTING    EIGHT    LINES. 


C.   To  Locate  a  Curve  by  Ordinates  from  a  Long  Chord. 

Let  Fig-.  21  represent  a  curve  having  an  odd  number  of  full 
chords,  and  a  short  chord 
AB  —  c'  subtending  an 
angle  AOB  =  Dl}  and  a 
short  chord  KL  —  c"  sub- 
tending an  angle  KOL 
=  D<I.  Draw  the  tan- 
gent AT,  and  drop  the 
perpendiculars  BX  and 
CX'  upon  the  tangent. 
Draw  Be  parallel  to  AT. 

Since  Be  is  perpendic- 
ular to  AO,  and  BG  is 
perpendicular  to  a  radius 
bisecting  the  angle  BOG,  the  angle  GBc  =  J)l-\-  \D. 

Now  BX  is  found  in  Table  II  ;  and  Ce  may  be  found  from  the 
same  table  or  Table  III,  supposing  the  curve  to  be  of  the  degree 
A  +  \D.  Or, 

BX  —  c'  sin  IDj  ,     and     Cc  =  c  sin  (Dl  -f-  %D}. 
Suppose  c'  =  80,      D  =  4°. 


i=3°  12', 


and 

Table  II  gives 
and 


Cc  =  9.06. 
Y'  =  11.29. 


Place  B  80  feet  from  A  and  2.23  feet  from  the  tangent  A  T ;  then 
C  a  chain-length  from  B  and  11.29  feet  from  the  tangent  AT. 

Locate  K  from  the  tangent  at  L  in  the  same  manner  as  B  is 
located.  This  gives  the  line  CK. 

Suppose  the  stations  already  located  ;  the  long  chords  EG,  DH, 
etc.,  drawn,  and  also  the  perpendiculars  Dd,  Eel,  etc. 

Now  Cd=  Cm  -  Di  =  l(CK—  DH)  =  one  half  of  the  difference 
between  the  chords  of  six  and  of  four  stations.  See  Table  IV. 

dl  —  Di  —  Ef=  \(DH—  EG]  =  one  half  of  the  difference  be- 
tween the  chords  of  four  and  of  two  stations. 

lm  =•  one  half  the  chord  of  two  stations. 


58 


FIELD-MANUAL   FOR   ENGINEERS. 


Again,  Dd  —  Fm  —  Fi  =  the  difference  between  the  middle 
ordinates  of  chords  of  six  and  of  four  stations.  (See  Table  V.) 

El  —  Fm  —  Ff  '  =  the  difference  between  the  middle  ordinates  of 
chords  of  six  and  of  two  stations. 

Finally,  Fm  is  the  middle  ordinate  of  a  chord  of  six  stations. 

We  note  that  points  and  lines  on  the  right  of  Fm  are  of  course 
symmetrical  with,  corresponding  points  and  lines  on  the  left. 


Hence 


On  =  CK-  nK  =  CK  -  Cl 

Ch  =  CK-  hK  =  CK  -  Cd,  etc. 


Now  lay  off  Cd,  Cl,  Cm,  etc. 

Set  D  a  station  from  C  and  a  distance  Dd  from  d,  or  CK;  ther 
Ea.  station  from  D  and  a  distance  El  from  I,  or  CK;  then  F  i 
station  from  E  and  a  distance  Fm  from  m,  or  CK,  etc.,  etc. 

We  also  have 

Cd  =  c  cos  f  D,     dl  —  c  cos  f  D,     and     fan,  =  c  cos  \D  ; 
Dd  -  c  sin  %D,   Ee  =  c  sin  f  D,      and     .FJf  =  c  sin  £Z>. 

These  quantities  may  be  taken  directly  from  a  table  of  sines 
and  cosines,  as  already  pointed  out.  It  is  not  necessary,  as  shown 
under  the  preceding  method,  to  compute  and  lay  off  Cd,  Cl,  etc. 

D.     To  Locate  a  Curve  by  Chord  Offsets. 

Let  ABCDEF  be  the  curve,  having  short  chords  AB  =  c' ,  sub- 
tending an  angle  AOB  =  D\ 
and  EF  —  c",  subtending  an 
angle  EOF  =  D.,. 
Fa  Locate  B  as  in  the  las' 
T  method.  Prolong  AB,  mak 
ing  BCi  =  c.  The  angle  CBC, 
=  CAB  +  ACB  =  \D-\-\Di. 
Hence  CiCis  the  chord  offse 
2t'  for  a  curve  of  \D  -)-  \D 
degrees. 


=  2c  sin  i(Z>,  +  D}. 

Place  C,  therefore,  at  a  distance  c  from  B,  and  2t'  from  Ci,  2t'  beinjv 
twice  the  tangent  offset  which  is  given  by  Table  III  opposite  the. 


SIMPLE   CURVES   CONNECTING    RtGHf   LINES.          59 

degree  $(D  +  DI).  Prolong  BC  to  D,  ,  making  <7A  =  c.  Now 
D1D  —  2t  is  given  by  Table  III,  being  twice  the  tangent  offset. 
Place  D,  therefore,  at  a  distance  c,  from  (7,  and  2t  from  DI.  Place 
all  regular  stations  similarly.  C  may  be  placed,  also,  as  in  the 
preceding  method.  J^is  easily  placed  from  the  tangent  at  E,  as  B 
was  placed  from  the  tangent  at  A. 

It  may  be  observed  that  by  the  above  method  we  are  able,  by 
means  of  a  table  giving  offsets  for  full  stations  only,  to  locate  any 
station,  as  C,  by  chord  (or  tangent)  offsets,  though  the  chord,  as 
AB,  preceding  the  adjacent  chord  may  be  of  any  length. 

E.     To  Locate  a  Curve  by  Middle  Ordinates. 

Let  ABCDEFG  be  the  curve,  having  a  short  chord  AB  =  <•', 
subtending  an  angle  AOB  —  D\ , 
and  a  short  chord  FG  =  c",  sub- 
tending an  angle  FOG  =  D«. 

Locate  the  stations  B  and  C  from 
the  tangent  A x,  or  by  some  other 
method  as  already  explained. 
Then  set  off  on  CO  the  distance 
Cc  =  t  —  c  sin  \D  —  the  middle 
ordinate  of  a  chord  of  two  stations ; 
and  set  D  in  the  prolongation  of 
Be,  and  at  a  distance  c  from  C, 
Set  off  Dd  the  same  as  Cc,  and  set  E  in  the  prolongation  of  Cd, 
and  at  a  distance  c  from  D.  Locate  all  subsequent  stations  in  the 
same  way.  To  test  the  accuracy  of  the  work,  measure  the  per- 
pendicular Fy  from  the  last  regular  station  upon  the  tangent  at  G. 
This  distance  ought  to  be 


F.  To  Lay  out  a  Curve  by  Radial  Lines  from  the  Center. 

Consider  the  case  of  a  half-mile  race-track,  having  two  parallel 
sides,  each  600  feet  long,  connected  at  the  ends  by  semicircles,  as 

204Q 1900 

shown    in  Fig.  24.     Now    -    — — — '•  -  =  720,  the  length  of  each 

180° 
semicircle.     Hence  the  degree  of  each  curve  —  «-™  =  25°,  and 

the  radius  is  229.18  feet. 


66  FIELD -MAX  UAL    FOR   ENGINEERS. 

Set  the  instrument  at  0,  and  ran  radial  lines  01,  02,  etc.,  making 
6 


FIG.  24. 

angles  of  80°  with  each  other,  for  example,  and  set  stakes  at  1,  2, 
etc.,  229.18  feet  from  0.  These  stakes  are  100  f  <?  =  120  feet  apart, 
measured  on  the  arc. 

Determine  other  points  on  the  curve  by  middle  ordinates. 

The  inside  of  the  track  is  three  feet  inside  of  the  line  measured 
above. 

ERRORS  IN  FIELD-WORK. 

In  laying  out  curves,  as  well  as  in  all  field  operations,  errors 
will  sometimes  occur,  and  it  is  important  to  know  the  immediate 
and  the  ultimate  effect  of  such  errors;  to  know  when  such  errors 
are  increasing,  from  one  stage  of  the  work  to  another,  and  when 
they  are  decreasing;  when  they  are  too  small  to  be  of  importance, 
and  when  they  are  so  large  as  to  vitiate  the  result  if  not  cor- 
rected. 

It  is  important,  also,  to  know  the  law  of  increase  or  of  decrease 
of  such  errors;  for  in  that  case  the  error  at  any  point  may  be 
found  from  that  at  any  other  point,  the  end  of  the  curve,  for  ex- 
ample; thus  making  it  possible  to  properly,  that  is,  really,  correct 
the  curve  without  rerunning  it. 

I.   Curves  Laid  out  by  Deflection  Angles. 

A.  To  find  an  approximate  value  of  the  error  at  the  end  of  a 
curve  due  to  an  error  in  the  length  of  a  chord. — Suppose  A,  B,  C, 
etc.,  to  represent  stations  on  the  true  curve,  and  A',  B' ,  C",  etc., 
stations  on  the  false  curve. 

1°.  Suppose  the  first  stake  from  A  to  be  set  at  B' ,  a  distance 
BB'  =  e  beyond  B,  its  true  place. 


SIMPLE    CUliVK-i    COXXECTiXG    1UGHT   LltfES.          6'1 


Then  C'  will  be  pet  on  AC  prolonged  and  a  chain  from  B',  D' 
will  be  set  on  AD  prolonged  and  a  chain  from  C',  etc. 

To  prove  that  CC'  is  less  than  BB'.  Draw  GH  equal  and  paral- 
lel to  BE'.  Then  BB'CII  is  a  parallelogram,  and  B'H  —  BC  — 
B'C'  =  a  chain.  Hence  C'  is  on  the  arc  whose  center  is  B'  and 
radius  B'H.  With  C  as  center  and  radius  Ctl  =  e  describe  the 


FIG.  25. 

arc  HK  UK  lies  outside  of  the  arc  HG' ,  since  the  arcs  at  //  are 
perpendicular  to  CH  and  B'H  respectively,  and  are  limited  by  AC 
prolonged.  Hence 

CC'  <  CK  =  CII  =  BB'  =  e. 

For  the  same  reason  DD'  is  less  than  CC',  and  so  on  to  the  end 
of  the  i'U)i,  that  is,  to  the  last  stake  set  with  the  instrument  at  A. 

Suppose  the  instrument  moved  to  D'. 

Since,  in  setting  E'  from  D',  we  turn  off  fromD'A  the  same  angle 
that  we  would  turn  off  from  DA  in  order  to  set  E,  we  have  D  E' 
equal  and  parallel  to  DE,  and  therefore  EE'  is  equal  and  parallel 
to  DD' .  For  the  same  reason  FF'  is  equal  and  parallel  to  EE' 
or  DD' .  S  >  on  to  the  end  of  the  curve. 

2°.  Suppose  an  error  to  occur  in  some  other  chord  than  the  first — 
in  the  second  chord,  for  example.  B  is  supposed,  therefore,  to  be 
set  correctly.  Suppose  the  next  stake  set  at  C'  instead  of  at  C, 


FIELD-MANUAL   FOR   ENGINE 


Let  BCr  =  c  +  e.     Now  BC  -f-  CC'  >  BC',  or 

c  +  <7C"  >  c  +  e.        .-.   (7(7'  >  e. 

From  C"  the  errors  will  follow  the  same  law  as  in  the  former 
case. 

With  reference  to  these  two  cases  we  remark  that  if  the  error 
occurs  in  the  chord  adjacent  to  the  instrument,  the  errors  in  the 
stations  will  decrease  slightly  to  the  end  of  the  run,  and  from 
that  point  remain  constant  to  the  end  of  the  curve  ;  but  if  the 
error  occurs  in  a  chord  not  adjacent  to  the  instrument,  the  error 
in  the  station,  at  the  end  of  that  chord,  is  slightly  greater  than 
the  error  in  the  chord,  though  the  errors  slightly  decrease  from 
this  point  to  the  end  of  the  run,  and  remain  constant  from  the 
end  of  the  run  to  the  end  of  the  curve.  In  all  cases  the  error  at 
the  end  of  the  curve  may  be  regarded  as  equal  to  the  error  in  the 
chord,  whether  adjacent  to  the  instrument  or  not. 

3°.  Suppose  the  chain  is 
in  error,  in  which  case  all 
the  chords  will  be  in  error. 
Let  the  curve  AB  (R  = 
AO  =  radius,  and  D  —  de- 
gree) be  run  with  a  chain 
100  feet  long,  and  the  curve 
AB'(R'  =  AO'  =  radius) 
with  a  chain  100  -f-  e  feet 
long.  The  number  of 

V 

chords  is  equal  to  —  =   n. 

Then  for  the  difference  in 
length  of  the  curves  we 
have  AB'  —  (100  +  e)n. 
AB  =  ICOn,  and  therefore 


AB'  -  AB  =  en. 


(34) 


For  the  chord  we  find,  since    R'  —  lit— 


AB'  = 
and  therefore 


AB'  -  AB  =  BB'  = 


',     AB  =  21i 

ZRes'm  ^V 
~100        ' 


SIMPLE   CURVES   CONNECTING    RIGHT   LINES. 


But 


BB' 

360tfsin^y        114.60esin"-|-y 

(35) 

TtU                           D 

Also 

11'  =  (100  +  < 

<  180 

180.         57.3, 

(36) 

We  also  have 

TV  7»7. 

-BB'~'m  *  V       114-6^(sin*F)2 

(37) 

ul 

TV, 

57.3^  sin   y 
—   7?  7?  '  cos  i  y  —  •. 

(38) 

Example.— Let  e  =  .02,    T7  =  60°,  and  D  =  6°. 

The  error  of  the  curve  =  .02  X  — TT  =  -200. 

o 

114  6  X  1  X    02 
The  error  of  the  long  chord  =  —  —  =  .191. 

Also  11'  -  R  -  -   '—£—    ~  =  -19!  • 

114.6  X  .02  X 


([ 


=  .0955, 


B.  Suppose  that  the  first  stake  is  set  at  B',  Fig.  27,  instead  of 
at  B,  the  error  in  the  angle  being 
BAB'  =  A. 

Then  it  is  evident  that  Cf, 
D',  etc.,  will  be  set  at  the  same 
angular  distance  from  the  chords  p, 

AC,  AD,  etc.,  as  B'  is  set  from 
AB.  Bf,  C',  etc.,  are  on  the 
curve,  having  0'  as  center  and 
AO'  =  AO  as  radius.  OAO' 
=  BAB'. 

To  find  the  error  at  the  end  of 
the  curve. 

Suppose  .Z£the  end  of  the  true 
curve. 


C4  FIELD-MANUAL    FOll    ENGINEERS. 

Now  AE'  =  AE,  and  the  angle  EAE'  =  the  angle  BAB'. 
Hence        JE7i"  =  2^1#  sin  \EAE'  =  2AE  sin 


If  the  error  in  the  angle  is  corrected  at  the  end  of  a  run,  say  at 
IS',  then,  for  reasons  given  above,  the  error  in  the  position  of  the 
stations  is  constant  from  Er  onward  to  the  end  of  the  curve.  If 
the  error  is  corrected  during  a  run,  say  just  before  D'  is  set, 
then  that  station  will  be  set  at  Di  on  AD  and  100  feet  from  C'. 
Similarly  Ei  will  be  set  on  JJ^and  100  feet  from  2)lf 

It  may  be  shown  that  EEi  is  less  than  DDl  precisely  as  it  was 
shown  that  CO'  is  less  than  BB  '  in  Fig.  25.  Hence  in  this  case 
the  error  will  decrease  to  the  end  of  the  run,  and  remain  constant 
from  that  point  to  the  end  of  the  curve. 

II.     Curves  Laid  Out  by  Tangent  Offsets. 

If  in  Fig.  22  the  tangent  at  B  is  swung  through  an  angle  A,  say, 
then  all  stations  following  B  will  be  misplaced,  the  error  increas- 
ing to  the  end  of  the  "  run  "  in  the  manner  shown  in  Fig.  27.  If 
a  new  tangent  is  drawn  before  reaching  the  end  of  the  curve,  as 
at  E',  Fig.  27,  it  is  easy  to  see  that  such  tangent  would  make  an 
angle  A  with  the  tangent  to  the  true  curve  at  E,  and  that  the 
error,  therefore,  would  go  on  from  E'  forward  precisely  as  from 
B  to  E',  and  hence  the  error  would  increase  regularly  from  B  to 
the  end  of  the  curve. 

III.     Curves  Laid  Out  by  Ordmates  from  a  Long  Chord, 

In  this  case,  if  the  end  C  of  the  chord  CK,  Fig.  21,  is  misplaced, 
then  all  stations  from  D  to  H  inclusive  will  be  misplaced  in  pro- 
portion to  their  distances  from  K\  the  greatest  displacement  being 
less  than  that  of  G.  If  K  is  also  misplaced,  the  same  stations 
will  likewise  share  that  error,  in  proportion  to  their  distances 
from  C. 

IV.   Curves  Laid  Out  by  Chord  Offsets.     (Fig.  22.) 

If  in  this  method  any  station,  as  B,  is  set  at  one  side  of  its  true 
position,  then  all  subsequent  stations  will  be  in  error  in  the  same 
direction  ;  the  errors  increasing  regularly  to  the  end  of  the  curve, 
in  the  manner  shown  in  Fig.  27. 

If,  owing  to  an  error  in  some  chord,  some  station  is  set  forward 
or  backward  from  its  true  position,  then  all  subsequent  stations 
will  be  in  error  the  same  amount  and  in  the  same  direction. 


SIMPLE    CU-RVES    CONNECTING    RIGHT    LINES. 


65 


V.  Curves  Located  by  Middle  Ordinntes. 

In  Fig.  23  suppose  G'to  be  placed  a  distance  a  to  the  right,  say, 
(that  is,  along  CO,}  of  its  true  position.  Then  c  will  be  a  distance 
ft  to  the  right  ;  D  and  d  will  be  2a  to  the  right  ;  ^and  e  will  be 
'*i  to  the  right  of  their  true  positions,  etc. 

If  C  is  correct,  but  c  a  distance  a  too  far  to  the  right,  then  D 
•.nd  d  will  be  2a  to  the  right,  E  nud  e  will  be  4«  to  the  right,  F 
ind/  will  be  $a  to  the  right  of  their  true  positions,  etc. 

PROBLEMS  IN  SIMPLE  CURVES. 

I.  Given  the  tangent  distance  AB  =  d  and  the  tangent  offset 
BD  =  t,  to  find  the  radius  of  a  curve  that  will  pass  through  D 
ind  be  tangent  to  AB  at  A. 

Let  AD  =  c.     We  have 


From  this  we  have 


1  -  d\    .     .     (39) 


FIG.  28. 


and  d  =  4/t(2JZ  -  t).     .     . 

Second  Solution. — Let  AOD  =  A  ;  then 

BAD  =  ADC  =  %A. 
Now  the  triangle  ABD  gives 

d-  =  cot  $A.     .     . 


(40) 


Also 
and 


-  =  vers  ^4, 
—  =  cosec  A. 

a 


(41) 
(42) 
(43) 


66  FIELD-MANUAL    FOR    ENGINEERS. 

If  d  and  t  are  given,.  find  A  from  (41),  then  R  from  (42)  or  (43) 
If  t  and  .7?  are  given,  find  A  from  (42),  then  d  from  (41)  or  (43). 
Finally,  if  d  and  R  are  given,  find  A  from  (43),  then  t  from  (41) 
or  (42). 

Example  1.  —  Given  t  =  12  and  d  =  171.6,  to  find  It. 

tan  §A  =  p^-g  =  .06993. 
.;.     \A  =  4°,     and     A  =  8°. 

Then  R  =  _  -  _!_    -  --^_  =  1233.3. 

vers  .4         .00973 

Example  2.~  Given  72  =  1233.3  and  d  =  171.6,  to  find  t. 
cosec  A  =  -^8j|-  =  7.187.     .-.   4  =  8°. 

Then  t  =  1233.3  X  .00973  =  12. 

This  problem  is  useful  in  finding  points  on  a  curve  beyond  an 
obstacle. 

II.   To  find  the  distance  to  a  curve  in  a  given  direction  from  a 
given  point  on  a  tangent. 
We  have 

AO  =  11,     AS  =  d,     and     ABP  =  B. 

tan  ABO  =  ?  .  PBO  =  ABO  -  ABP; 
d 

sin  OPI  =  sin  OPB  =  ~  X  sin 
ji 

sin 


••  sin  ABO' 
Then 

1  sm(OP2-PBO) 

FIG.  29.  —faTPBO 

This  problem  furnishes  a  general  method  of  finding  any  desired 
point  on  a  curve  when  obstacles  preclude  the  usual  methods. 
(See  Problem  12.) 

III.  Having  run  the  curve  AD,  radius  AO  =  It,  Fig.  28,  to 
find  the  radus  R'  of  a  curve  that  will  puss  through  D',  given  by 
angle  BDD'  =  D,  and  DD'  =  E. 


SIMPLE    CURVES    CONNECTING    RIGHT   LINES.          67 

Let  AB  =  d,     and     BD  =  t. 
Draw  D'H  parallel  to  AB.     Now 

HD'  =  E  sin  D,     and    1W  =  E  cos  D. 
AB'  =  d+  #sin  D  =  dlt     and     B'D'  =  t  -  E  cos  D  =  ti. 


It  will  be  observed  that  when  AB'  <  AB,  E  sin  Z>  must  be 
subtracted  from  d  to  give  d* ;  and  that  when  B'D'  =  BH  >  BD, 
E  cos  D  must  be  added  to  t  to  give  £,. 

IV.  Having  run  a  curve  of  radius  R  and  tangent  T,  to  find  the 
new  tangent  T'  corresponding  to  a  new  radius  R' ',  or  to  find  a  new 
radius  h'  corresponding  to  a  new  tangent  T' ',  the  central  angle  re- 
maining constant. 

Eq.  (15)  gives 

T'  =  R'  tan  \V\     T  =  R  tan  % V\ 
.'.   T'  -  T—  (11'  -  R)  tan  *  F,     .     .     .     .     (45) 
or  72'  —  R  =  (T' —  T)  cot  \V.    '..'..     (46) 

Similarly,  from  eq.  (17),  we  get 

C'  —  C  =  2(72'  —  R)  sin  ^F.      ....     (47) 

These  equations  are  of  advantage  for  computing  the  change  in 
one  element,  T'  —  T  for  example,  from  the  change  in  another, 
R'  —  R  for  example,  when  the  given  change  is  small,  or  is  an 
aliquot  part  of  the  element  changed. 

Example  1. — Having  run  the  curve  of  radius  72  =  1910,  and  the 
central  angle  V=  46°  12',  and  tangent  distance  T  =  R  tan  \  V 
=  1910  X- 42654  =  814.7,  to  find  T'  when  72  is  made  equal  to  1900. 

Eq.  (45)  gives 

=  814.7  -  .42654  X  10  =  814.7  -  4.27  =  810.4. 
Example   2.—  Given  F  =  52°  04',    72  =  5730,   and    T  =  .48845 
X  .5730  =  2798.8,  to  find  72'  corresponding  to  T'  —  T  +  •-- . 
We  have 

R'  =  R  +  ^  =  2798.8  +  233.2  =  8032.0. 


68 


FIELD-MANUAL    FOR    ENGINEERS. 


V.  (riven  a  curve  joining'  two  tangents,  to  change  the  position 
of  the  P.  C.  so  that  with  the  same  radius  the  curve  may  end  in 

a  given  parallel  tangent. 

Let  ^47?  be  the  given  curve, 
and  IV  F'the  parallel  tangent. 
W' —a  shows  the  distance 
and  direction  that  all  points  of 
the  curve  are  moved.  The 
curve  will  therefore  begin  at 
A '  and  end  at  Bf ;  A  A  and  BB', 
as  well  as  00',  being'  equal 
and  parallel  to  YV.  It  is 
not  necessary  to  run  the  tan- 
gent B'  V  in  order  to  find  the 
distance  VV.  To  find  this 


FIG.  30. 


distance  run  a  line,  such  as  BB',  parallel  to  A  V  from  any  point 
on  BV  to  meet  B'V.     Then  make  A  A'  =  BB'. 

If  the  perpendicular  offset  Bh  —  7t  is  measured,  we  have 


AA  =  BB'  = 


V  being  the  vertex  angle. 

It  B'V  were  on  the  other  side  of  .BFfrom  that  shown,  the 
new  tangent  point  A'  would  fall  on  the  opposite  side  of  A  from 
that  shown  in  the  figure. 

If  the  new  curve  is  required  to  end  at  a  given  point  on  B'V, 


we  .have,   then,   the 
new  tangents, 


length   of   the 


A'V  =  B'V  =  T', 

which  gives  the  position  of  A  (and 
B'),  and  the  corresponding  radius, 

It'  =  7"  cot  IF, 
or,  by  (46), 

R'  =,  R  +  (T'  -  T)  cot  *F. 

FIG.  31. 

VI.  Given  a  curve  AB  joining  two  tangents,  to  find  the  radius 
of  a  curve  that  from  the  same  P,  C.  will  end  in  a  given  parallel 
tangent. 


SIMPLE    CURVES    CONNECTING    RIGHT    LINES.  09 

Let^F=£F  =  T,  AV'  =  B'V'  =  T't  AO  =  R,  A0f  =  R'. 

We  liave,  from  the  figure, 

R'       T'  7" 

R=-T>     °r     ll     =ET- 

Also,  from  eq.  (46), 

R'  =  R  +  (Tf  -  T)  cot£F. 

Or,  prolong  AB  to  B'  and    measure  BB'.       Let  ^45  =  c,  and 
A#'  =  c',  BB'  —  c'  -  c.      Then,  from  the  figure  or  from  (17), 


If  the  parallel  tangent  is  defined  by  a  perpendicular  offset,  as 
B'p  —  Ji,  draw  BG  parallel  to  AO.  Then 

Cp  =  BCcos  BCp  =  (R1  -  R)  cos  V. 

.-.  CB'  =  (R'-R)  =  (R'-K)coz  V+U,     or     (1?  -£)(!-  cos  V)=7i, 
or          (R'  -  R)  versin  V  =  h,     or     R'  =  R  +  -      A 

The  quantity  added  to  IMn  the  above  equations  must  be  sub- 
tracted from  it  to  find  R'  in  the  case  in  which  V  falls  between  A 
and  V,  that  is,  when  T'  is  less  than  T7.  • 

Example  1.  —  V  --  78°,  R  =  954.9.  T.7  may  be  computed  or  found 
by  Table  VII  to  be  773.3. 

Let  VV  =  20  feet.     Then 

R'  =R+  (T'  -  T)  cot  £F=  954.9  +  20  X  1.2349  =  979.6. 

Example  2.—R  —  1909.9,  F=  46°  38'.  T  may  be  computed  or 
found  by  Table  VII  to  be  823.2. 

It  is  desirable  to  move  the  vertex  from  Fto  V  about  100  feet. 
Find  the  new  radius  R'. 

823.2  -*  8  =  102.9; 
1909.9  H-  8  =  238.7, 

Hence  A  V  -    823.2  +  102.90  =  926.1, 

and  the  new  radius 

AO'  =  If  _  1909.9  +  238,7    =  2148.6. 


70 


FIELD-MANUAL   FOR   ENGINEERS. 


VII.  Given  a  curve  joining  two  tangents,  to  find  the  new  tan- 
gent points,  corresponding  to  the 
same  radius,  after  eacli  tangent  Las 
been  moved  any  distance  in  the  di- 
rection of  the  other. 

Let  A  Fand  .Z?Fbe  the  given  and 
A' V  and  B'V  the  required  tan- 
gents. Let  Hbe  at  the  intersection 
of  AV  &ul  B'V.  Let  VH  =  a, 
VII  =  b,  and  VV  =  c. 

We  observe  that 


B 


VHV  =  180°  -  F. 
.'.  sin  VII V  —  sin  V, 

FIG.  32.  and         cos  VHV  =  —  cos  F. 

Chap.  Ill,  formula  No.  (8),  gives 

sin  F  sin  F 


HV 


r-f  -j-  COS    V 


cos  V 


Then 


VV  = 


b  sin  V 


sinhVV 


We  observe  that  the  directions  of  F//and  HV  are  the  same 
as  that  in  which  the  tangents  B  V  and  .4  Fare  moved,  and  there- 
fore there  can  be  no  ambiguity  about  the  direction  of  these  lines 
or  of  FF',  which  is  the  line  joining  F  and  V. 

Since  R  and  F  are  not  changed,  it  is  evident  that  all  parts  of 
the  curve  are  moved  in  the  direction  VV  and  a  distance  equal  to 
FF'. 

Hence  make  the  angle  VAA'  =  HVV,  and  A  A'  =  VV. 

The  curve  will  begin  at  A'  and  end  at  B'  ,  BB'  as  well  as  00r 
being  equal  and  parallel  to  W  . 

If  the  distances  the  tangents  are  moved  are  given  by  perpendic- 
ular offsets  Vh'  =  hf,  and  Vh  =  h,  the  triangles  Vllh  and 
V  '  Hh'  are  similar  and  give 


HV 
HV 


sin  VHV 


HV 
HV' 


or 


SIMPLE   CURVES   CONNECTING    RIGHT   LINES. 

sin  V 
~  h 


71 


Now  the  triangle  hVV '  gives 

vv-    -7^- 

smHW 

VV  and  Vh  are  on  the  same  side  of  B  V;  also,   VV,  and  Vh> 
are  on  the  same  side  of  A  V. 

VIII.  Given  a  curve  AB  joining  two  tangents  AV  and  BV,  to 
change  the  curve  so  as  to  end  at  the 

same  point  as  before,  bat  in  a  tan-    A.  A'  V        V'     D 

gent    inclined  at  a  given   angle  A 
with  the  original  tangent. 

Let  A  V-  BV  =  T,  and  the  new 
tangents  A'V  =  BV  =  T'. 

Let  AO  =  /?,     and      A'O'  =  II'. 


Draw  BD  =  p  perpendicular,  and 
BMN  parallel  to  A  V. 
Let 

V'=BV'D=BVD+VBV'=V+A. 

Now     An  =  It  versin  V, 
and         A'm  —  11'  versin  V. 
But        A'm  =  An', 


B 


FIG.  33. 

. '.  R  versin  V  =  R'  versiu  V, 
R  versin  V 


R'  = 


versin  V 


With  this  value  of  R'  run  the  curve  back  from  B  through  the 
angle  V,  and  it  will  end  at  A',  tangent  to  A  V\  A'  V  being  equal 
to  BV. 

If  the-length  of  the  new  tangent  is  desired,  we  have,  from  the 
triangle  VBV, 


Then 


sin  V 
T' 


sin  V  ~  sin  V 


P 


tan 


sin  V  tan  \V  ~  vers  V 


72  FIELD-MANUAL    FOR    ENGINEERS. 

If  we  wish  to  run  the  curve  from  A,  we  have 

D  V  =  p  cot  V,     and     D  V  =  p  cot  V. 

.-.   VV  =p(cot  V-  cot  V), 
and         AA'  =  AV+  VV  -  A'V  =  VV  -\-T-T'. 

When  V  <  90°,  T  increases  as  V  decreases,  and  vice  versa; 
and  when  V>  90°,  I7  and  V  increase  and  decrease  together.  In 
all  cases  R  increases  as  V decreases,  and  vice  versa. 

IX.  Given    a    curve,    radius    AO  =  R,   joining  the   tangents 

A  V  and  BV,  to  find  the  radius 
AO'  =  R'  of  a  new  curve  start- 
ing from  A  when  the  forward 
tangent  VB'  takes  a  new  direc- 
tion from  the  vertex. 
"We  have 


and 


VA  =  R  tan  \Vt 
R'  =  VA  cot£F. 


tan 


FIG.  34. 


tan  |  V 


X.  Given  a  curve  AB,  radius  AO  —  R,  joining  the  tangents 
A  Fand  VB,  to  find  the  change 
in  the  P.C.,  the  radius  remain- 
ing the  same,  when  the  forward 
tangent  takes  a  new  direction 
from  the  vertex. 

We  have 

VA  =  5  tan  £F; 
VA'  =  Btan^F. 


FIG-  35. 


XL     Given  the  angle  of 
tersection    V  of  two  tangents  0 
4  Fand  BV>  tP  find  the  radius 


SIMPLE   CURVES   CONNECTING    RIGHT   LINES. 


R  and  tangent  distance    T  of  u   curve  joining  the  tangents  and 
passing  tb  rough  the  point  E. 

1°.  Let  E  be  given  by  VE=l, 
and  angle  EVO  =  A. 

Let    VEO  =  E,     VOE  =  0, 
and    AO  =  21,   AOV  =  $V. 

No\v 

i  v-  A0  -  E0  -  ^A. 

-Jd~'VO~ 


sin 


sin  E  — 


sin  A 


cos  \  V 
This  gives  E.     Then 

0  =  180°  -  (A  4  E). 
Moreover, 

EO     _  R  _   sin  A 

~EV  ~~  T  " ~    sin  0 


sin  A 

or      7?  —  I  -. — — 
sin  0 


Finally,  T= 

2°.   If  E  is  given  by   VZZ"  and  HE  perpendicular  to  each  other, 
EH 


then 


tan  E  VII  = 


VII' 


EVF  =  FVH  -  EVH  =  90°  -  £F -  EVH; 

VII 

~~  cos  EVH ' 

With  these  values  proceed  as  above. 

3°.  Let  Ebe  given  by  VD  —  a,  IXfiJ  =  ft,  the  angle  VDF  being 
rial  to  FO^l  =  4F.     Produce  i>^to  2^ and  (7. 


Let  DF  =  c.    Then  c  —  a  cos  ^F,  and  AD  =  \/DE  X  DO. 
But    DO  =  DF+FG  =  DF -\-  EF=2DF-  DE=2c-b. 
~1>),     and     VA=  VD  +  DA  =  T. 


.-,  AD= 
Now   It=  Tcot  \V, 
or         HE  =  DE  sin 


=  b  sin  i  F. 


74  FIELD-MANUAL   FOR  ENGINEERS. 

D1I  =  b  cos  £  V. 
Then  VH=  VD  —  DH  =  a  -  b  cos  \V. 

With  these  values  of  F/iTand  HE  proceed  as  above. 

XII.  Given  a  tangent  and  curve  (Fig.  36),  to  find  the  distance 
from  a  given  point  on  the  tangent  to  the  curve  in  a  given  direc- 
tion. 

Let  V  be  the  point,  and  suppose  the  direction  defined  by  the 
angle  EVA  =  B.  Let  AV  =  T.  We  have 

T 
taniF=-. 

Then        EVF  =  D  VF  -  EVA  =  (90°  -$V)-B  =  A,  say. 
Now  equation  under  Problem  XI  gives 

sin  A 

sin  E  =  --  T^pf. 
cos^F 

Then  0  =  180°  -(A  +  E), 


or, 


^,     and    EVF  =  FVD  -  EVD  =  FDV  -  B  =  A. 


Then  find  sin  J^,  then  0  and  7  as  before. 
Example.—  R  =  954.93,  T7  =  A  V  =  350,  ^L  VE=  40°.     We  have 


tan  AVO  =         -  2.72837;     .'.  J[FO  =  69°  52', 
and  AOV=±V=  20°  8',    and    #FF  =  29°  52'  =  F. 


0  -  180°  -  177°  50'  =  2°  10'; 

_  954.93  X  -03781  _  70  4 
.49798" 


SIMPLE    CURVES    CONNECTING    RIGHT    LIKES. 


75 


XIII.  To  locate  a  tangent  to  a  curve  of  given  radius  R  from  a 
given  point  V.     (Fig.  37.)   v  c 

1 .  If  the  curve  is  marked 
by  stakes  visible  from  the 
given  point,  a  tangent  can 
be  sighted  in  at  once. 

2.  If    the    curve    is   not 
visible  from  the  point,  run 

a   trial   tangent    VB,    and  FIG.  3T. 

measure  VB  —  A  arid  the  angle  VBO  =  B. 
Chapter  III,  formula  (8),  gives 

sin  B 


tan^FO  = 


Then          OV=OB. 
Lastly, 


sin  B 


—  —  cos  B 


and     sin  EVO  = 


sin  B  VO" 
BVE  =  EVO  -  BVO. 


EO 

or 


sin  A 


-  —  cos  A 


This  gives  the  angle  to  be  laid  off  from 
the  trial  tangent  to  give  the  true  tan- 
gent AE. 

XIV.  Given  two  curves  AB  and  AC, 
radii  AO  =  R,  AOi  =  R\,  subtending 
the  central  angles  AOB  =  0  and 
AO^C  =  0,,  to  find  the  length  of  the 
line  EC.  (Fig.  38.) 

Let  A,  B,  and  G  represent  the  angles 
of  the  triangle  ABC,  and  «,  b,  and  c  the 
sides  opposite.  We  have 

c  =  2R  sin  \0,     and    &  =  2R}  sin  |0,. 
Also, 

BAO  =  90°  -  \0,     CAOv  =  90°  -  £0, 
.-.  BAC=  A  =  1(0  -  0i). 

-,    (See  Chap.  Ill,  formula  (8).) 


and  then 


, 

a  =  b 


sin  A 
-  —  ~. 
sin  B 


FIELD-MANUAL   Foil    EXGIXKEKS. 


If  the  curves  are  run  through  an  integral  number  of  stations, 
Tables  IV  and  V  give  at  once  the  tangent  distances  AK  and  AH 
and  the  tangent  offsets  BK  and  CH.     Then,  drawing  CD  parallel 
to  A  //to  meet  BK'ni  D,  we  have 
CD  =  AH  -  AK  =  d,  say,     and     BD  =  BK  -  Ctf=  t,  say. 

Then  BC  =  f < 


and 


BC  = 


BD 


OBSTACLES  IN  SURVEYING. 

It  is  often  necessary  to  draw  lines  parallel  and  perpendicular  to 

other  lines.     Hence  the  follow- 
ing problems  : 

I.   To  erect  a  perpendicular  at 
any  point  cf  a  line.     (Fig.  39.) 

1.  Let  J.  be  the  point,  and  EC 
the  line.     Make  AE  =  AC,  and 
with  B  -and   C  as   centers   and 
any  radius  greater  than  AE  de- 
scribe arcs  intersecting  at  D  or 
at  E,  or  (with  a  different  radium) 
at  F.     Any  two  of  the  points  A, 
D,  E,  and  F  determine  the  per- 
pendicular required.  "! 

2.  Fix  any  two  points  of  the 
chain  at  B  and  at  C.     Take  hold 
of  the  point  midway  between  B 
and  G  and  stretch  the  chain,  the 

middle  point  being  at  D.  AD  is  the  perpendicular  required. 
We  may  find,  similarly,  other  points  E,  F,  etc.  Any  two  of  these 
points  A,  D,  E,  F,  etc.,  determine  the  perpendicular  required. 

3.  Let  C  be  the  point.     Take  any  point  D  as  a  center,  and  with 
a  radius  DC  describe  an  arc  BC.   Prolong  BD,  making  DH  —  BD. 
CIlis  the  perpendicular  required.     For  DA  (A  being  at  the  mid- 
dle of  BC)  is  perpendicular  to  BC,  and,  by  construction,  CH  is 
parallel  to  AD. 

4.  A  right  angle  may  be  obtained  by  laying  off  on  the  ground 
the  three  sides  of  any  of  the  triangles  represented  in  the  following 
table,  or  any  equimultiples  of  these   sides,   making  one  of   the 


FIG.  39. 


8IMPLK   CURVES   CONNECTING    RIGHT    LINES. 


77 


sides  adjacent  to  the  right  angle  (a  or  b)  coincide  with  the  line. 
Let  c  =  the  hypothenuse. 


No. 

u. 

b. 

c. 

No. 

a. 

b. 

C. 

1 

3 

4 

5 

6 

20 

SI 

29 

2 

5 

12 

13 

7 

12 

33 

37 

3 

8 

15 

17 

8 

9 

40 

41 

4 

24 

25 

9 

11 

60 

61 

5 

10 

24 

26 

10 

13 

84 

85 

B 


Thus,  in  Fig.  40,  using  70  links  of  the  chain,  hold  the  first 
end,  also  the  end  of  the  70th  link  of  the  chain,  at  A,  the  end  of 
the  21st  link  at  B,  and  the  end  of 
the  50th  link  at  C. 

If  in  the  three  expressions 
ra2  —  ir,  2mn,  and  m2  -j-  ifi  we 
assign  to  m  and  n  any  values  at 
pleasure,  in  being  greater  than  n, 
we  will  have  sets  of  numbers 
representing  the  sides  of  right- 
angled  triangles.  In  that  way 
the  above  numbers  were  found. 
Equimultiples  of  any  set  of  the 
above  numbers  will  represent  the 
sides  of  a  right-angled  triangle. 

II.  To  let  fall  a  perpendicular  from  a  given  point  to  a  given 
line. — Let //in  Fig.  39  be  the  point.  Measure  any  line  HB  to 
the  given  line.  At  the  middle  of  ///>  take  D  as  a  center,  and 
with  a  radius  DB  describe  an  arc  BC.  HO  is  the  required  per- 
pendicular. For  continuing  the  arc  to  II,  we  see  that  the  angle 
BCIIis  inscribed  in  a  semicircle. 

III.  To  let  fall  a  perpendicular  to  a 
line  from  an  inaccessible  point. — Let 
BC  (Fig.  41)  be  the  line,  and  P  the 
point.  Let  p  represent  the  perpen- 
dicular PK.  Then 

BK  —  p  cot  B,   and    CK  =  p  cot  C. 

BK  _  cot  B 
•''   CK  ~ 


21 
Fm.  40. 


FIG.  41. 


and 


BK 


cot  C' 

cot  B 
cot  B  -j-  cot  C ' 


78  FIKLD-MAXUAL    FOR    EXGTXEKRS. 

Since  BK  -\-  CK  =  BC,  we  have 
BK  =  BC 


cot  B  -f  cot  0  * 

Tills  gives  the  foot  of  the  perpendicular  K.  If  BC  is  taken  equal 
to  100  or  some  small  multiple  of  100,  BK  is  very  easily  found. 

This  problem  is  particularly  useful  in  locating  important  ob- 
jects, such  as  mills,  warehouses,  bridges,  etc.,  on  one  side  or  the 
other  of  a  railway  survey. 

In  this  case  B  and  C  represent  stations  or  points  on  the  survey, 
and  the  angles  at  B  and  G  can  be  measured  and  recorded  while 
the  instrument  is  set  at  B  and  at  C.  The  simple  divi-ion  required 
to  find  the  position  of  .STcan  be  made  at  any  time.  Of  course  the 
point  Pis  located  graphically  by  drawing  .Z?Pand  CP. 

IV.  To  prolong  a  line  AB  (Fig.  42)  past  an  obstacle  and  to 
measure  its  length. — This  is  easily  done  by  perpendicular  offsets, 
a  method  to  >  familiar  to  need  description,  but  not  the  best  way. 


B 


FIG.  42. 

Or,  measure  BC  in  any  convenient  direction,  and  at  C  deflect  any 
angle  FCD.  Draw  BD  and  the  perpendicular  CH.  The  angle 
CDS  =  FCD  -  CBD.  Hence 


sin  FCD 
also  <P-L>  =  ^^  sin  BBC' 

If  the  angle  BCD  is  made  equal  to  90°,  then 

CD  =  BC  tan  CBD, 

and  BD  =  BC  -f-  cos  CBD. 


SIMPLE   CURVES    CONNECTING    RIGHT    LINES. 


If  the  angle  FCT)  is  made  equal  to  2CBD,  tlien 
CDS  =  FCD  -  CBD  =  CBD. 
Hence         CD  -  BC,     and     BD  —  2BH  =  2BC  cos  CBH. 

OH  is  tlie  departure  of  the  line  BC,  or  of  DC,  from  the  line 
ABD.  It  is  also  the  approach  of  the  line  CB,  or  of  CD,  to  ABD. 

If  necessary  more  than  one  course  may  be  run  away  from  the 
main  line  ABD,  and  more  than  one  in  returning  to  it. 

To  recover  the  main  line  it  is  only  necessary  to  make  the  sum 
of  the  approaches  equal  to  the  sum  of  the  departures. 

The  distance  measured  on  the  main  line  is  obtained  as  above. 

V.  Obstacles  to  Measurement. — Methods  have  been  pointed  out 
in  connection  with  Fig.  42  for  rinding  the  length  of  obstructed  lines 
when  the  ends  are  accessible.  When  inaccessible  the  following- 
problems  apply. 

A.  When  one  end  of  the  line  is  inaccessible.     (Fig.  43.) 

1.  Let  AB  be  the  line  to  be  measured,  across  a  river  for  ex- 
ample. Measure  AC  in  any 


convenient  direction,  and  the 
angles  at  A  and  C.     Then 

AC  sin  C 

AB  =  — 


B 


sin  B 


=  AC-. 


sin  C 


sin  (A  +  C)' 

2.  If  the  angle  ACS  is 
made  equal  to  half  of  DAC, 
then 

OB  A  =  CAD  -  BCA 


.'.   AB  -  AC.  FIG.  43. 

3.  Or,  in  Fig.  43,  make  the  angle  BAG  =  90°.     Then 

AB  =  AC  tan  ACB. 

If,  in  this,  AC  =  100,  or  some  simple  multiple  of  100,  which  is 
usually  easy  to  effect,  the  formula  requires  no  computation 
whatever. 

4.  If  ACB  in  Fig.  44  is  made  equal  to  45°,  AB  —  AC. 


80 


FIELD-MAXUAL    FOR   EXGIXEERS. 


5.  If  at  G  we  make  the  angles  ACB  and  ACD  equal,  we  liavo 
AB  =  AD. 

When  the  river  or  other  obstruction  occurs  on  a  continuous 
survey,  as  a  railway  survey,  AD  is  a  measured  line,  and  this 
method  gives  AB  =  AD  without  any  computation  whatever. 


FIG.  45. 


6.  In  Fig.  45,  AB  being  the  distance  required,  run  and  measure 
any  line  AC  and  measure  the  angle  BAG  —  A.     Make 


Then 


ACB  =  90°  -  A  =  C. 
AB  =  AC  sin  C. 


B.  When  both  ends  of  the  line  are  inaccessible.     (Fig.  46.) 
Let  AB  be*  the  line  to  be  meas- 


ured. Find  the  distances  from  the    A- 
point  G  to  each  end  of  the  line  A 
and  B  by  preceding  methods,  and 
measure  the  angle  G.     Then 

sin  G 


tan  A  = 


AC 


(see  Chapter  III,  formula  (8).) 


C 
FIG.  46. 


. 

sin  A 


C.  To  erect,  at  a  given  point  A  (Fig.  47),  a  line  ylZf  perpendicu- 


SIMPLE    CURVES    CONNECTING    RIGHT    LINES. 


81 


\ 


lar  to  an  inaccessible  line 
BC,  and  to  draw  a  parallel 
AH  to  the  same  line. — Find 
.47?  =  c  and  AC  —  b  by  pre- 
ceding methods.  Then 

pin  A 
tan  B  =  -        . 

- —  cos  A 
I) 

Now  draw  A  K,  making 
BA  K  =  90°  -  B. 

A  K  will  be  the  required  per- 
pendicular, and  AH,  making  BAH  —  B,  will    be    the    required 
parallel. 

D.   To  find  the  length  and  relative  position  of  an  inaccessible  lino, 
AB,  from  an  accessible  line,  CD,  separated 
n  from  the  former  by  an  inaccessible  space.    - 
Measure  CD  and  the  angles  at  C  and  D. 
Example.— Let  CD  =  4000; 

BCD  =  126°  25V;    ADC  =  47°  53 V; 
ACB=      3°  10';      ADB=    3°  01'. 
.-.  ACD  =  129°  35V;    SJ)C  =  50°  54 1'. 
Also, 

CAD  =  180°  -  ACD  —  ADC  —  2°  31'; 
CBD  =  180°  -  BCD  -  BDC  =  2°  40'. 
Hence 

40=  400o5!2*™a.'  =  67«81.»; 


BC  =  4000 
tan  CAB  = 


sin  2°  40' 
sin  C 


=  66728.8; 


'AC 


FIG,  48, 


Finally, 


CAB  =  75°  28V. 
^sin    3°  10' 


-,  -  3807.8. 


82  FIELD-MANUAL    FOH   ENGINEERS. 

If  AD  is  accessible,  it  can  be  measured  as  a  check  on  the  com- 
putation. Such  is  the  case  when  it  is  a  tangent  of  a  railway  sur- 
vey adjacent  to  the  inaccessible  space. 

The  data  of  this  example  are  taken  from  an  actual  night  survey 
across  an  inaccessible  sea-marsh. 

Rockets  were  thrown  and  lights  then  exhibited  at  A.  and  B. 
which  were  observed  with  transits  from  the  tops  of  towers  at  C 
and  D. 

The  computed  and  the  measured  length  of  AB  agreed  within  a 
few  inches. 

The  best  method  of  making  a  preliminary  railway  survey 
through  a  wooded  region  is  by  a  suitable  adaptation  of  the  method 
of  traversing,  which  we  will  now  explain. 

So  far  as  known  to  the  author,  this  was  first  devised  by  him  in 
1869,  and  used  for  him  by  his  assistant,  Prof.  J.  B.  Davis  (now  of 
Michigan  University)  in  making  the  preliminary  surveys  of  the 
Owosso  and  Northwestern  Railway. 

Suppose  we  wish  to  run  from  A  in  the  direction  of  ABF,  which 
we  will  call  the  base  line;  and  upon  which  numerous  obstacles,  such 


FIG.  49. 

as  trees,  occur,  making  it  necessary  to  run  the  line  ABiCiDiE[F, 
called  a  traverse.  The  deflection  angles  at  Bi,  C\t  Diy  pud  E^  are 
supposed  to  be  small. 

From  Bi ,  d  ,  D\  ,  and  Ei  draw  perpendiculars  to  AB,  and  from 
Si,  C>,  and  Dl  draw  parallels  BiK,  dL,  and  D,P  to  AB  as 
shown.  Prolong  ABi  to  R,  and  B}  Ci  to  8. 

The  course  of  a  line  is  its  direction  with  reference  to  the 
base  line. 

The  departure  of  a  line  is  the  distance  that  a  point  recedes 
from  or  approaches  to  the  base  line  in  moving  from  one  end  of 
the  line  to  the  other. 

We  have  D1L  —  CiDi  sin  DiCiL.  Now  since  the  sines  of 
small  angles  vary  nearly  with  the  angles  or  with  the  number  of 
minutes  in  the  angles,  we  see  that  the  departure  of  a  line  varies 


SIMPLE   CUHVES   CONNECTING   1UGHT    LINES. 


83 


nearly  as  the  product  of  the  length  of  the  line  by  the  number  of 
minutes  in  the  course. 

The  departure  of  a  point  is  its  distance  from  the  base  line. 
Thus  the  departure  of  Z>,  =  DD,. 

The  departure  of  the  end  of  a  line,  as  B\  C\ ,  inclining  from  the 
base,  is  equal  to  the  departure  of  the  beginning  of  the  line  plus 
the  departure  of  the  line.  Thus  Cd  =  #77,  +  f,7i.  The 
departure  of  the  end  of  a  line,  as  C\D\ ,  inclining  toward  the  base, 
is  equal  to  the  departure  of  the  beginning  of  the  line  minus  the 
departure  of  the  line.  Thus  DDi  =  CC\  —  DiL.  The  departure 
of  the  end  of  the  line  D\E\  which  crosses  the  base  line  is  equal 
to  the  departure  of  the  line  minus  the  departure  of  the  beginning 
of  the  line.  Thus  EEl  =  E,P  -  DDt. 

The  record  of  the  survey  can  be  conveniently  kept,  as  shown 
in  the  following  table,  the  columns  of 'the  transit-book  serving 
the  purpose  perfectly. 


Angles  turned. 

Angles  with 
Main  Line. 

Departures. 

Stations. 

Each  Course. 

Total. 

Left. 

Right. 

Left. 

Right. 

Left. 

Right. 

Lcfr. 

Right. 

A  =      10 

10' 

10' 

11 

12 

£,=      13 

22' 

32' 

3000 

3000 

14 

15 

Ci  =  +  20 

4'2X 

10' 

7040 

10040 

16 

17 

18 

Z>!=      19 

3C' 

40' 

3800 

6240 

21 

E,=      24 

1°20' 

40 

20000 

137GO 

25 

26 

27 

F  =+44 

13760 

.00 

.00 

The  deflection  at  station  10  is  10'  R.>  and  at  station  13  it  is 
22'  R.,  etc.  The  course  from  10  to  13  is  evidently  10'  R. ;  from 
13  to  15  -f-  20  it  is  10  -f  22  =  32'  I*.;  from  15  +  20  to  19  it  ig 
42  -  32  =  10'  L.,  etc. 


84  FIELD-MANUAL    FOR   ENGINEERS. 

From  10  to  13  the  departure  is          300x10  =    3000  foot-minutes. 
"      13  "  15+20  the  departure  is  220x32  =    7040 
"      15+20  to  19    "  "          "380X10:=    3800 

'<      19  to  24  "          "500X40=20000       "         etc. 

The  aggregate  departures  are  : 

At  13  ................................  3000  R. 

At  15  +  20  ......  3000  +  7040  =  10040  R. 

At  19  ............  3000  +  7040  -  3800  =  6240  R.,  etc. 

The  distance  necessary  to  run  from  a  given  station  on  any  given 
course  to  reach  the  base  line  is  found  by  dividing  the  departure 
at  that  station  by  the  course.  Thus  from  station  24  forward  the 
course  is  40'  R.  Then  13760  -4-  40  —  344  feet,  showing  that  the 
auxiliary  line  (E^Fin  the  figure)  will  reach  the  base  line  344  feet 
beyond  station  24,  or  at  27  +  44. 

The  figure  represents  a  main  angle  at  F,  the  forward  tangent 
being  FIT,  and  which  may  be  followed  approximately  the  same  as 
AF  WAS  followed. 

Let  I  =  the  length  and  d  =  the  departure  of  any  line,  and  n  — 
the  number  of  minutes  in  the  course.  Then 

d  =  I  sin  n'  =  In  sin  1'  =  .0002909fo. 

Since  In  is  given  in  the  last  two  columns,  the  departures  in  feet 
are  found  by  multiplying  the  quantities  in  these  columns  by 
.0002909  or  .00029  nearly.  We  observe  that 

1  4-  .0002909  =  3438  nearly. 

Hence  the  departures  given  in  the  table  (in  foot-minutes) 
divided  by  3438  will  give  the  departures  in  feet. 

A  rough  approximation  for  the  purpose  of  keeping  sufficiently 
near  the  base  line  on  sideling  ground  is  generally  all  that  is 
needed.  This  being  the  case,  it  is  not  in  general  necessary  or 
advisable  to  find  the  total  departures,  except  when  it  is  desirable 
to  "  run  for  the  base  "  preparatory  to  turning  a  main  angle. 

Thus  to  find  the  departure  at   station   24.     The   sum   of  the 
product  to  the  left  is  ____  .............  3800  +  20000  =  23800 

and  the  same  to  the  right  is  ...........  3000  +    7040  =  10040 

The  difference  is  ..................................  13760  L. 

We  have  AR  =  AB,  cos 


~  4W  -  <-«s  #4#.)  =  AB  vers  BA 


SIMPLE    CURVES    CONNECTING    RIGHT   LINES.          85 

For  BABl  =  2°  34'  this  becomes  AB,  -  AB  =  .WlAB  nearly. 
This  shows  that  ABi  exceeds  the  true  distance  measured  along 
the  base  by  only  one  thousandth  part  of  its  length  for  an  angle 
of  2°  34'.  If  greater  accuracy  than  this  is  desired,  the  angles 
between  the  auxiliary  line  and  the  base  line,  or  the  "courses," 
may  usually  be 'made  smaller  than  2°  34'.  Since  the  error  is 
approximately  as  the  square  of  the  number  of  minutes  in  the 
angle, 

for  1°  1?'     it  is  nearly  .00025,  or  nearly  1  in    4000; 

and  for         0°  38£'  "  "      "       .00006,   "       "       1  "  16000,  etc. 

To  find  the  angle  in  minutes  between  the  base  line  and  a  line 
joining  any  two  stations. 

Divide  the  difference  or  the  sum  of  their  departures,  according 
as  they  are  on  the  same  or  on  opposite  sides  of  the  base  line,  by 
their  distance  apart. 

Thus  the  line  BiDi  makes  with  the  base  line  the  angle 
PA  -  BB,   _  6240-3000  _          _ 
BD  WO 

The  line  AE,  makes  with  the  base  an  angle 


In  platting,  the  auxiliary  lines  are  penciled  only,  so  as  to  plat 
observed  objects  in  proximity  to  the  line  necessarily  observed 
from  the  auxiliary  lines.  When  these  objects  and  the  base  lines 
are  mapped  the  auxiliary  lines  need  not  be  retained. 

This  method  yields  quite  accurate  results  when  the  angles  be- 
tween tbe  lines  of  the  survey  are  2°  or  3°,  as  we  have  seen. 

For  perfect  accuracy,  however,  use  the  following  method  : 

Problem. — Having  run  a  broken  line  ABCD,  to  find  the  angle 
between  the  first  course  and  the  direct  course  JIT).  Represent  the 
lines  run,  in  their  order,  by  a,  &,  and  c,  and  the  deflection  angles  at 


FIG.  50. 

B  and  at  C  by  B  and  C  respectively.     Draw  CR  parallel  to  AB, 
and  (7/7  and  KDK  perpendicular  to  AB. 

Angle  DCR  =  DCE  -  RCE  ~  C  -  B, 


86 


FIELD-MANUAL    FOR   ENGINEERS. 


Now 


Now 

BH=bcosB;     HK  =  c  cos  (C  -  B) ;     AK  -  AB  +  BH+  HK 
CH  =  b  sin  B;     DR  =  c  sin  (C  -  B);    DK  =  CH  -  DR. 

DK 

4jr 

Also  AD  =  AK  -*-  cos  DAK. 

Of  course  the  method  is  applicable  whatever  the  number  of  lines 
run.      All  but  the  last  line  could  usually  be  taken  equal  to  a  whole 

number  of  chains,  which 
would  reduce  the  required 
computation  to  a  simple  mul- 
tiplication. 

E.  To  find  the  angle  of 
deflection,  i",  between  two 
straight  lines  A  V  and  VM, 
when  the  point  of  intersection 
is  inaccessible;  and  the  dis- 


A  M  tances  of  the  intersection  from 

FlG-  51-  given  points  on  the  lines. 

1.  Run  and  measure  a  perpendicular  PK  to  one  of  the  lines. 
Measure  the  angle  VPK  =  P.     Then 

V=  90°  +  P; 
KV=  A'Ptan  F, 
and  PV  -  KP  •  t- co*  P. 

2.  Run  and  measure  any  line  PL  from  one  line  to  the  other. 
Measure  also  the  angles  VPL  =  P'  and  PL  V  =i  /,.     Then 

V=  P'  +  L. 

Hence,  in  the  triangle  PVL,  PL  and  the  angles  are  known,  to 
find  FPand  VL. 

3.  If  obstructions  prevent  the  use  of  the  former  methods,  run 
and   measure  any  broken  line  ABCD.     Prolong  AB  and    BC  to 
meet  VM  at  M  and  N. 

Measure  the  deflection  angles  CBM  =  B,  DCN  =  C,  and  CDN 
=  D. 
Let  AMV  =  M,  and  CNV  =  N.     Then 


sTiHv  '  tto.(0+J5' 

=  BC  -f  CN.     Also     M  =  W  -  B-  C  -f  D  -  /?; 


SIMPI.I-;  CURVES  CUNXKCTI.NI.T  RIGHT  LINES. 


BM  = 


sin  N 


AM  =  AB  +  BM. 


Now  we  have  ^43/arid  the  angles  at  A  and  M,  to  find  vl  F,  MV, 
and  the  angle  F  =  A  4-  Jf.  A  similar  explanation  will  apply  to 
any  case. 

Wlie.n  a  broken  line  must  he  used,  the  above  method  involves 
fewer  computations  than  any  other. 

F.  To  locate  a  curve  joining  two  tangents  when  the  vertex  is 
inaccessible. 

Find   by  the  last    problem   the 
distances  Va  and  Vb  to  convenient 

points   on    the   tangents,    and    the  /  \    \,  V 

angle  V.  Then  assigning  or  com- 
puting the  tangent  T  =  A  For  BV 
from,  the  radius,  we  have 


and 


aA  =  T  -  a  V, 
bB  =  T-  bV. 


We  now  have  the  tangent  points 
and  can  run  in  the  curve  as  usual. 

GK  To  locate  a  curve  of  radius  U  or  tangent  T  when  the  vertex, 

the  beginning,  and  the  end  of  the 
curve  are  inaccessible.  Find,  as 
shown  with  Fig.  52,  the  angle  F 
and  the  distance  Fa  to  any  point, 
«,  on  A  V.  Then 

a  A  =  T  -  a  F, 
and 

cd        a  A 


FIG.  53. 


Also 

ac  =  Ad  =  J?versin  AOc. 
Drawing  the  tangent  cb,  we  have 

abc  =  AOc,     or     acb  =  90°  -  AOc. 

This  gives  the  direction  of  the  curve  at  c,  and  it  may  be  run  in 
each  way  from  c. 

To  pass  from  any  point  c  on  the  curve  to  any  point  n  on  the 
tangent. 


We  have 


1an  (dcu  =  an<<)  = 


88  FIELD-MANUAL   FOR   ENGINEERS. 

Set  the  instrument  at  c  and  turn  off  from  the  tangent  cb  an  angle 

ben  =  bed  —  den 
=  dbc  —  anc, 


,  Us  IV 

and  measure  en  =  —      — . 

cos  anc 

H.  To  find  any  desired  point  on  a  curve  when  obstacles  precl  udo 
the  use  of  ordinary  methods. 

(1.)  In  Fig.  28  measure  any  convenient  tangent  distance  AB  =  d. 
Then,  as  shown  in  Problem  1,  eq.  (43), 

cosec  A  =  — ;     then     t  —  d  tan  ±A. 

d  and  t  give  the  point  D  on  the  curve. 

It  is  important  to  note  that  if  AB  is  made  equal  to  one  half  the 
long  chord  for  any  number  of  stations  given  by  Table  IV,  BD  = 
AC  is  the  corresponding  middle  ordinate  and  may  be  found  in 
Table  V. 

Example. — Let  AD  be  a  4°  curve,  and  AB  —  one  half  the  chord 

of  four  stations  =  — ^—  =  199.35.     Then  Table  V  gives 

BD  =  13.94. 

(2.)  Problems  2  and  12  of  this  chapter  furnish  general  methods 
of  overcoming  obstacles  on  curves. 

(3.)  We  can  find  points  on  the  curve  as  follows: 
Let  b  be  a  station  near  the  obstacle.  Deflect  from  the  tangent 
at  b  some  small  multiple  of  the  deflection 
angle  for  one  station  ^D,  giving  the  line 
bd.  The  length  of  bd  may  be  taken  at 
once  from  Table  IV  and  measured  off, 
giving  d,  a  station  on  the  curve  beyond 
the  obstacle.  Taking  bm  =  \bd,  and 
measuring  off  the  middle  ordinate  me 
taken  from  Table  V,  gives  also  a  point  c 
on  the  curve. 

If  more  convenient,  make  bx  =  cm,  and 
xc  =  bm,  which  also  gives  c. 

Again,   run  the  tangent  b  V  =  d'  any 
jrIG  54  convenient  distance.     Then 


SIMPLE   CURVES   CONNECTING    RIGHT   LINES.          89 

Deflect  at  V  an  angle  equal  to  2bOV,  and  make  V'd  =  bV. 
d  will  be  a  point  on  the  curve.  The  number  of  stations  from  b 
is  equal  to 

bOd       2bOV 


bV  sliould  usually  be  taken  equal  to  a  whole  number  of  chains, 

7) 

in  which  case  —^-f  is  very  readily  found. 

The  lines  bV"  and  V"d  lying  on  the  inside  of  bd  may  be  run 
instead  of  bV  and  V'd. 


CHAPTER   V. 

LEVELING,  STADIA  MEASUREMENTS,  ETC. 

THE  field  operations  in  connection  with  the  level  are  more 
simple  than  those  required  with  the  transit,  but  they  require 
greater  skill  and  facility  in  manipulation  in.  order  to  produce 
correct  results. 

It  is  to  be  observed  that  the  elevation  of  points  is  a  relative 
matter.  The  elevation  of  some  point,  from  which  all  others  are 
to  be  found,  is  arbitrarily  assumed  to  be  100,  or  some  other 
number  sufficiently  large,  so  that  the  elevation  of  all  points 
to  be  considered  will  be  greater  than  zero. 

Near  the  coast,  and  in  fact  wherever  practicable,  it  is  im- 
portant to  refer  the  levels  to  the  mean  level  of  the  sea,  calling 
this  zero,  or  100,  or  some  other  number,  taking  care  to  estab- 
lish from  it  some  convenient  and  permanent  reference-point 
called  a  bench-mark,  or  bench. 

All  points  having  the  same  height  as  this  bench  are  some- 
times said  to  be  on  a  level  surface  called  the  datum.  This, 
however,  makes  no  difference  with  the  work  and  serves  no  use- 
ful purpose,  and  need  not  be  considered. 

All  elevations  thus  found  become  of  much  importance  in 
determining  the  relative  elevations  of  the  country,  and  in  the 
construction  of  physiographical  maps,  etc. 

Having  established  the  first  bench,  and  recorded  its  elevation, 
the  rod  man  stands  squarely  on  both  feet  behind  the  rod,  and 
rests  it  on  the  bench  as  nearly  in  a  vertical  position  as  possi- 
ble, which  is  best  done  by  simply  steadying  it  with  the  thumbs 
and  fingers,  taking  care  not  to  grasp  it. 

The  levelman  sets  up  his  level,  preferably  in  the  direction 
that  the  line  extends,  in  any  position  from  which  he  can  well 
see  the  bench,  as  well  as  points  to  be  afterwards  observed. 

He  then  makes  sure  that  the  instrument  is  in  adjustment, 
and  is  focused;  levels  it  carefully  and  sights  to  the  rod.  He 

90 


LEVELING,   STADIA    MEASUREMENTS,  ETC.  91 

may  plumb  the  rod  laterally  by  means  of  the  vertical  cross- 
wire  of  the  level,  and  the  rod  may  be  waved  gently  on  each 
side  of  the  vertical  toward  and  from  the  instrument,  the  short- 
est reading  being  the  true  reading. 

The  line  of  sight  on  the  rod  covered  by  the  horizontal  cross- 
wire  is  then  on  a  level  with,  or  at  the  same  height  as,  the  wire 
itself,  and  the  latter  is  therefore  higher  than  the  bench  by  the 
distance  intercepted  on  the  rod  between  the  line  of  sight  and 
the  bottom  of  the  rod.  This  is  called  the  reading  of  the  rod, 
or  simply  the  reading.  Adding  this  reading  to  the  height  of 
tho  bench,  we  obtain  the  height  of  the  cross-wire,  technically 
called  the  height  of  instrument,  and  designated  by  the  initials 
H.  I. 

Having  obtained  the  height  of  instrument,  the  elevation  of 
any  other  point  upon  which  the  rod  can  be  read  can  be  found 
by  taking  a  reading  of  the  rod  upon  it.  Of  course  the  point  is 
below  the  instrument  an  amount  equal  to  the  reading,  which 
must  therefore  be  subtracted  from  the  height  of  instrument  to 
give  the  elevation  of  the  point.  The  elevations  of  any  number 
of  points  may  be  thus  obtained. 

In  order  to  obtain  the  elevation  of  points  above  the  instru- 
ment, or  below  it  more  than  the  length  of  the  rod,  the  instru- 
ment must  be  moved  from  its  present  position  to  one  higher 
or  lower  as  the  case  may  require. 

Before  the  instrument  is  moved  to  a  new  position  a  temporary 
henfJt,  tailed  a  turning-point  (and  designated  by  T.  P.  or 
"Peg"),  must  be  established  and  its  elevation  ascertained  with 
care,  since  any  error  in  its  elevation  is  carried  forward  through- 
out the  whole  line  of  levels.  A  turning-point  must  be  firm 
and  definite  and  not  easily  disturbed  or  lost.  A  small  stake 
or  "  peg "  driven  with  its  upper  surface  about  flush  with  the 
surface  of  the  ground  is  generally  used.  The  top  of  a  rock  may 
well  serve  the  purpose. 

Benches  and  turning-points  are  of  course  the  same  in  prin- 
ciple, but  the  more  or  less  permanent  point  taken  as  the  basis 
of  the  elevations  of  the  Survey,  and  also  those  made  usually 
along  and  near  the  line,  for  future  reference,  whether  used 
as  turning-points  or  not,  are  usually  called  benches. 

From  this  new  turning-point  we  proceed  precisely  as  before, 
by  getting  a  new  height  of  instrument,  etc.,  and  it  is  important 


FIELD-MANUAL    FOR    ENGINEERS. 


to  note  that  the  operation  just  described,  of  obtaining  a  height 
of  instrument  from  a  bench  or  turning-point,  and  then  obtain- 
ing the  heights  of  any  number  of  desired  points  within  range 
of  the  instrument,  including  a  new  bench  or  turning-point, 
includes  the  whole  subject  of  leveling. 

Since  the  cross-wires  must  be  higher  than  any  point  upon 
which  a  reading  is  taken  it  must  be  remembered  that: 

1.  The  reading  on  a  point,  added  to  its  elevation,  gives  the 
height  of  instrument. 

2.  The  reading  on  a  point  subtracted  from  the  height  of  in- 
strument gives  the  elevation  of  the  point. 

In  other  words:  We  must  add  a  reading  (to  the  height  of 
some  point)  to  get  a  height  of  instrument,  and  must  subtract 
a  reading  (from  a  height  of  instrument)  to  get  the  height  or 
elevation  of  some  point. 

The  theory  of  leveling  requires,  therefore,  only  a  simple 
application  of  addition  and  subtraction,  and  it  is  not  easy,  it 
would  seem,  to  go  wrong  in  it. 


Station. 

+  8 

H.  I. 

-  S 

Elevs. 

Remarks. 

EM 
0 
1 
Peg        2 

3.46 

203.46 

7.29 
5.34 

0.81 

200.00 
196.17 
198.12 
202.65 

W.  Oak  60  ft.  R.  of  Station  0 

+  40 
8 
4 
5 

Peg  +  60 
6 

4.17- 
6.18 

206.82 
211.93 

1.12 
3.16 
6.09 
4.14 
1.07 
3.13 

205.70 
203.66 
200.73 
202.68 
205.75 
208.80 

The  accompanying  table  shows  a  convenient  form  of  field- 
book  for  keeping  the  level  notes  of  a  railway  or  other  survey. 
The  first  column  contains  the  stations  and  benches.  The  second 
the  plus  readings  taken  on  points  whose  elevations  are  assumed 
or  already  determined.  The  third  column  contains  the  heights 
of  instrument  recorded  one  line  below  the  elevation  of  the  turn- 
ing-point (or  bench)  from  which  it  is  calculated.  The  fourth 
column  contains  the  minus  readings.  The  fifth  column  con- 
tains the  elevations  of  all  points  observed.  The  right-hand 
page  is  reserved  for  remarks  describing  the  benches  and  their 
location,  also  objects  crossed  by  (or  near)  the  line,  as  roads, 
streams,  ditches,  etc. 


LEVELING,  STADIA    MEASUREMENTS,  ETC.  93 

It  is  to  be  observed  that  for  any  series  of  levels  the  sum 
of  the  plus  sights  less  the  sum  of  the  minus  sights  (omitting 
those  for  determining  intermediate  points  on  the  ground)  is 
equal  to  the  difference  between  the  first  and  last  elevation. 

Thus  to  prove  station  3,  we  have 

3.46  -f  4.17  —  0.81  —  3.16  =  203.66  —  200  —  3.66. 
To  prove  the  H.  I.,  211.93,  we  find 
3.46  -f-  4.17  4-  6.18  —  0.81  —  1.07  =  211.93  _  200  —  11.93. 

In  practice  it  is  best  to  check  each  page  of  the  field-book  by 
comparing,  as  above,  the  first  turning-point  or  height  of  instru- 
ment (brought  over  from  the  preceding  page),  with  the  last 
turning-point  or  height  of  instrument  on  the  page. 

To  facilitate  this  work  some  engineers  use  two  columns  for 
the  minus  sights,  placing  those  which  determine  the  turning- 
points  in  a  column  by  themselves. 

This  practice  is  commendable. 

Benches  should  be  established  at  short  distances  apart  along 
the  line,  taking  care  to  locate  them,  so  far  as  possible,  near  the 
crossings  of  roads,  streams,  railways,  etc.,  and  at  all  points 
where  their  need  can  be  foreseen,  in  the  location  of  cattle- 
guards,  culverts,  bridges,  etc.  Of  course  an  extra-good  bench 
should  be  established  at  the  end  of  the  survey.  An  extra-good 
turning-point  or  bench  should  also  be  established  at  the  end 
of  each  day's  work. 

The  object  of  obtaining  a  line  of  levels  is-  to  furnish  a  profile 
of  the  line  surveyed,  showing  the  undulations  of  the  surface 
over  which  it  passes. 

The  elevations  are  platted  on  profile  paper,  the  horizontal 
scale  being  about  400  feet  to  an  inch,  and  the  vertical  scale 
about  25  feet  to  an  inch.  This  distortion  of  scale  magnifies  the 
vertical  measures  about  ****/.,?,  =  16  times,  so  that  the  slight 
changes  in  the  elevation  of  the  surface  may  be  distinctly  seen. 

In  running  a  line  of  "  flying  "  levels  no  readings  are  taken 
except  on  turning-points.  If  the  difference  of  levels  of  the  ex- 
treme points  only  is  desired,  it  is  necessary  to  find  the  differ- 
ence only  between  the  sum  of  the  plus  and  of  the  minus  read- 
ings, as  already  explained.  This  is  very  convenient  for  testing 


94  FIELD-MANUAL  FOR  ENGINEERS. 

a  line  of  levels  already  run;  in  which  ease  it  is  best  to  touch 
on  the  benches  only,  and  if  found  correct,  the  intermediate  ele- 
vations may  be  regarded  as  correct  also. 

No  line  of  levels  should  be  taken  as  correct,  and  so  used, 
without  first  being  carefully  checked. 

The  Philadelphia  rod  is  the  most  convenient  and  best  rod 
in  use.  It  is  plainly  lettered  and  easy  to  use,  and  may  be 
read  by  the  levelman  when  desirable,  and  at  a  distance  of  sev- 
eral hundred  feet. 

To  Locate  a  Level  Line.— Set  a  peg  at  the  desired  height,  as 
a  starting-point,  and  take  a  reading  of  the  rod  thereon.  Send 
the  rod  forward  in  the  desired  direction,  and  have  it  moved  up- 
ward or  downward  along  the  slope  of  the  ground  until  a  point 
is  found  which  gives  the  same  reading  as  before. 

Of  course  the  reading  is  taken  on  a  peg.  This  second  peg 
is  of  the  same  height  as  the  first.  Find  in  the  same  way  :i 
third  peg  from  the  second,  etc. 

In  this  way  stakes  may  be  set  at  points  on  the  ground  level 
with  the  top  of  a  proposed  dam,  or  with  the  supposed  top 
of  water  flowing  over  the  dam.  Then  joining  these  stakes  by 
lines,  the  area  thus  inclosed  may  be  measured. 

The  water  behind  a  dam  is  not  level,  but  is  curved  con- 
cavely  upward  and  so  increases  in  height  back  of  the  dam, 
and  sets  back  farther  than  if  level. 

For  the  subject  of  backwater,  Works  on  Hydraulics  must  be 
consulted. 

Other  applications  of  the  level  line  are  to  obtain  "  contour 
lines "  for  topographical  maps,  for  levees  in  irrigated  rice- 
fields,  etc. 

To  Run  a  Grade-line This  consists  in  setting  a  series  of 

pegs  so  that  their  tops  shall  be  points  in  a  line,  which  shall 
have  any  required  slope  ascending  or  descending. 

First  drive  pegs  at  each  end  of  a  line  to  the  heights  required. 
These  heights  may  differ  by  a  given  amount,  or  this  difference 
may  be  undetermined. 

Set  the  level  over  one  of  the  pegs  and  measure  the  height,  a.  of 
the  cross-wires  above  the  top  of  the  peg. 

Set  the  rod  on  the  other  peg,  and  make  the  reading  on  the  rod 
equal  to  the  height  a. 

Without  disturbing  the  level  drive  any  desired  number  of  pegs 


LEVELING,  STADIA   MEASUREMENTS,  KTC.  95 

along  the  line,  so  that  the  reading  on  each  will  also  be  equal 
to  a. 

A  line  of  uniform  grade  or  slope  is  not  a  straight  line. 

Calling  the  globe  spherical,  this  line  when  traced  in  the  plane 
of  a  great  circle  would  be  a  logarithmic  spiral.  On  a  length  of 
six  miles  the  distance  of  its  middle  point  from  the  middle  of  its 
straight  chord  would  be  six  feet  almost  exactly. 


CORRECTION  FOR  THE  EARTH'S  CURVATURE  AND  FOR 
REFRACTION. 

This  is  necessary  for  long  distances. 

Let  AB  (Fig.  55)  represent  a  portion  of  a  section  of  the  earth's 
surface.  Then  if  a  level  be  set  at  A,  the  line  of  sight  of  the  level 
will  be  the  tangent  AD,  while  the 
true  level  will  be  the  arc  AB.  The 
difference  BD  between  the  line  of 
sight  and  the  true  level  is  the  cor- 
rection for  the  earth's  curvature  for 
the  distance  AB.  This  must  be  sub- 
tracted from  the  reading  of  the  rod 
at  B,  or,  what  is  the  same  thing, 
added  to  the  height  of  B,  as  given 
by  the  reading  of  the  rod. 

Let  AE  =  R,  AB  =  1),  and  BD 
=  E.     By  geometry, 


AD*  =  BD(BD  4- 
A& 


BD  = 


FIG  55. 


BD  + 


Omitting  BD  in  the  right-hand  member,  since  it  is  small  com- 
pared with  2R,  and  supposing  AD  —  AB  =  J),  we  obtain 


22t       2  X  20913650 


rx  =  .0000000239087)5.    . 


(1) 


This  formula  gives  a  result  or  value  for  E  slightly  too  small; 
but  the  relative  error  is  only  about  one  in  24,000  for  a  distance  of 


^Q  =  39.6  miles,  or  arc  of5-    m 


=  34'  22".65. 


96  FIELD-MANUAL    FOR    ENGINEERS. 

In  observing  distant  objects,  a  ray  of  light  traversing  tlie  air 
from  an  object  to  the  eye  or  instrument  is  refracted,  and  takes  a 
curved  path  which,  for  points  near  the  surface  of  the  earth,  is 
practically  the  arc  of  a  circle,  concave  downward,  and  whose 
radius  is  7.K. 

Thus  a  point  at  C  (Fig.  55)  would  appear  at  D  higher  than  it 
really  is  by  an  amount  CD.  This  may  be  found  from  the  above 
formula  by  substituting  7R  for  2L 

Hence  the  correction  for  refraction  is 


E'  =  ~_  =  .  00000000341  5D2  .....     (2) 

The  correction  for  curvature  and  refraction  is 

IP        D2        3  /)2 
E"  =  £C=BD-CD=:  —  -  ~—rt  =  -  -    =  .000000020492D8.  (3) 


This  must  be  added  to  the  apparent  elevation  of  the  observed 
object  to  give  the  true  elevation. 

Table  XI  gives  the  value  of  the  correction  for  the  value  of 
R  =  20911790  feet. 

When  it  is  possible  to  set  the  level  midway  between  the  points 
whose  heights  are  required,  the  corrections  will  balance  each 
other  and  may  be  omitted. 

The  above  equations  may  be  put  into  a  form  sometimes  more 
convenient  as  follows  : 

The  length  of  arc  on  the  earth's  surface  subtending  angle  of 
one  minute  is 


Then  ~  =  .12638 (!') 

=  the  correction  for  refraction  for  distance  6083  ft.  or  arc  of  1'. 
Also  ~~    =  .8846 (2') 


=  the  correction  for  curvature  for  the  same  distance  or  arc  1', 


LEVELING,  STADIA   MEASUREMENTS,  ETC, 


97 


and 


i   1  .75828 (3') 

7/t 


=  the  correction  for  curvative  and  refraction. 


TRIGONOMETRIC  LEVELING. 

First  Method. — When  the  point  C  (Fig.  56)  can  be  seen  from 
two  points  A  and  B  on  the  same  level,  then 

AD  -  CD  cot  CAD,     and    BD  =  CD  cot  CBD, 

Subtracting  gives 

AB 

AB=OD(cot  CAD-  cot  CBD),  or  ^=calCAD_^cB1)-      W 

Second  Method. — Let  A  and  B  (Fig.  57)  occupy  any  positions 

C 


B 
FIG.  56. 


except  in  line  with  C.     Measure  AB  and  the  angles  at  A  and  B\ 
also  the  angle  of  elevation  GAD. 


C=1SQ°-A-B,  AC  = 


sin  C 


,  and  CD  =  AC  sin  CAD.  (5) 


If  A,  B,  and  C  are  in  the  same  vertical  plane,  the  solution  is  in 
no  wise  affected. 

Unless  the  distance  AD  =  D  is  short  it  is  necessary  to  add  to. 
the  correction  found  by  the  preceding  formulas  the  correction 

O   7^0 

for  curvature  and  refraction,  namely,  —  —  —  .0000000204927)2, 


FIELD-MANUAL    FOR    ENGINEERS. 

To  find  the  height  of  instrument  by  an  observation  of  the  horizon 

(Fig.  58). 

First  Method. — Let  C  be  the  place  of  the  transit,  and  BAA  a 

portion  of  a  section  of  the 
earth's  surface. 

Were  there  no  refraction 
the  line  of  sight  would  be 
the  tangent  CA\  ACE—  C 
would  be  the  angle  of  de- 
pression or  dip  ;  and  we 
would  have 

BC  =  BOX  exsec  CO  A 
=  It  exsec  C. 

Owing  to  refraction,  how- 
ever, the  line  of  sight  would 
be  a  curve  concave  down- 
ward whose  radius  =  7/£; 
and  it  would  therefore  ex- 
tend from  C  to  a  point  A', 
say  a  distance  A  A'  beyond 
A. 

Draw  CDF  tangent  to  this 
curve  at  C  to  meet  the  ra- 
dius AO  prolonged  in  F. 
Draw  also  the  tangent  A't. 

Let  COA=N,  COU=I1', 
and  CO  A'  =  0.  Then 


FIG.  58. 


CO  =  r  sec  //  = 


cos  H' 


Let  E,  on  A'O  prolonged   but  not  shown,  be  the  center  of  the 
arc  A'C.     Now 

#0V_  CO9 


cos  0  =  -  cos  COE  =  - 


2CO .  EO 


-  ___ 

TsT^eclT^  12 

Clearing,    substituting   1  —  vers  0  for   cos  0,    1  —  vers  H  for 
cos  II,  and  1  -f  exsec  7?"  for  sec  H,  we  find 


13  vers  0  —  13  vers  //  +  exsec  Jf. 


LEVELING,  STADIA    MEASUREMENTS,  ETC.  99 

Or,  writing  versines  for  exsecants  or  vice  versa,  we  have 

vers  0  =  £  vers  II, 
or,  approximately, 

exsec  0  =  £  exsec  H. 

In  the  triangle  CEO, 

49  r5  -f  36  ?•*  -  ~U6l  _  85r2  -  r2  sec9  H  _  85  -  sec2  H 
2~x~  6r"x~7V  *  84?"  ~~84~~ 

Also  sin  E  =  sin 


.-.  sin  #sin  0  =  sin2  0'  ----  =  (1  —  cos'2  Oj—  —  — 

_  1  70  -  169  cos2  H  -  sec2  H        sec  H 

~l44~  ~~T"~ 

Now  since  OA'  is  tlie  prolongation  of  EO,  and  JS'C'and  OZ)  are 
perpendicular  to  CD  and  therefore  parallel,  we  have 

CO  A    -  1)0  A'  =  COD,     or     0  -  E  =  II'; 
.-.  cos  H'  =  cos  (0  —  E)  =  cos  0  cos  E  -\-  sin  0  sin  E. 

Substituting  in  this  the  above  general  values  of  cos  0,  cos  E, 
and  sin  7?  sin  0,  and  expanding  and  reducing,  we  find 


13  cos  E  +  sec  H  ,    (sec  H  —  cos  //) 


Putting  cos  H=  1  —  vers  //,  sec  7/=  1  +  exsec  IT,  etc.,    we 
find 

(exsec  //  —  vers  H) 
vers  /f  =  f  vers  77  --         —  —  —        —  ,     .     .     (a) 

or  vers  77'  =  |  vers  77,  very  nearly,  ......     (6) 

or  exsec  //'  =  f  exsec  //,  very  nearly  ......     (7) 

exsec  H  '  —  vers  //' 

Hence       vers  //'  =  f  vers  H  --  —  . 

i& 

From  this  we  have 

,   £(exsec  H'  —  vers  H'} 
vers  H  —  \  vers  //'  -f  ^  —  ,     .      (ft) 


100  FIELD-MANUAL   FOR    ENGINEERS. 

or  vers  //=  |  vers  //',  very  nearly,      .....     (6') 

or  exsec  H=  |  exsec  II',  very  nearly  ......     (7') 

Since  the  versines  of  small  angles  are  very  nearly  in  the  ratio 
of  the  squares  of  the  magnitudes  of  the  angles,  we  have,  from  (6), 


H—  4/JZT  =  1.08//',     approximately.    .     .     .     (?') 

T 

Supposing  r  =  4000  miles  and  AB  =  —  —  50  miles,  then    it  is 

80 

easy  to  show  that  the  error  of  eq.  (7)  is  less  than  .00000002,  and 
the  error  of  (6)  is  about  .0000000004. 

Ki'<n/'iple  1.  —  The  observed  dip  of  the  sea  horizon  is  //'  .—  24'. 
What  is  the  height  of  the  instrument  above  the  sea? 

We  have 

BC  =  r  exsec  H-  r|  exsec  //'  =  20914000  X  .00028467 

'=  595.35  feet,  exactly. 

Example  2.—  Let  r  —  4000  miles,  and  AB  =  50  miles.  What  is 
the  observed  angle  of  depression  H',  and  what  is  the  height  of 
the  observer  above  the  sea  ? 

We  have 

Kf\ 

H=  j^  X  57°.29578  =  0°.7162  =  42'  58". 

.-.  H'  =  0.716  -4-  1.08  =  0°.663  =  39'  47". 

Also, 

7i  =  r  exsec  H=  4000  X  .0000781  =  .3124  miles  =  1649.47  feet. 

The  exact  relations  between  II  and  H',  shown  above,  would 
seem  to  be  more  satisfactory  than  the  approximate  equations  in 
general  use  even  if  these  were  regarded  as  sufficiently  accurate. 

THE  STADIA. 

The  stadia  is  a  compound  cross-wire  ring  or  diaphragm  having 
three  horizontal  wires. 

The  two  outer  ones  are  called  stadia  wires,  and  distances  deter- 
mined by  means  of  them  are  called  stadia  measurements. 


LEVELING,  STADIA    MEASUREMENTS,  ETC.  101 

The  stadia  wires  are  adjusted  so  as  to  intercept  a  certain  space 
on  a  rod  at  a  given  distance  from  the  transit  and  perpendicular  to 
the  line  of  sight. 

Let  C  (Fig.  59)  =  the  distance  of  the  object-glass  from  the  axis 
of  the  transit,  and/  =  the  focal  length  of  the  object-glass. 


FIG.  59. 

This  focal  length  is  equal  to  the  distance  of  the  cross-wires 
from  the  object-glass  when  this  is  focused  for  a  distant  object. 

This  focal  length  may  be  found  also  by  removing  the  ohject- 
glass,  exposing  it  to  the  rays  of  the  sun,  and  noting  at  what  dis- 
tance from  the  center  of  the  glass  the  rays  form  a  perfect  and 
minute  image  of  the  sun  on  a  smooth  surface. 

Let     Cm  =  I',     Cn  =  I,     DE  =  S',     and    HK  =  8. 

The  focal  distance  OF  is  constant,  but  Co  =  c  varies  with  the 
position  of  the  object-glass,  and  hence  CF  is  also  variable. 

Let  Co  =  c'  when  the  rod  is  at  DE,  and  Co  =  c  when  the  rod  is 
at  HK. 

Now,  from  the  figure, 

* 

' 


—  - 

Fm  ~  DE'  I'  -  (c'  +7) 


8'  is  usually  assumed  =  1  foot,  and  Fm  =  I'  —  (c'  -\-  f)  —  100 
feet;  and  the  stadia  wires  are  then  adjusted  accordingly. 

c'  is  measured  on  the  telescope  when  the  object-glass  is  focused 
on  the  rod  at  the  assumed  distance. 

To  measure  any  other  distance  the  rod  is  again  observed  at  the 
desired  point  and  the  space  8  noted,  which  placed  in  (8)  gives 
I  —  (f'  +/)  —  ^i»  sav-  We  may  then  measure  c  on  the  telescope. 
Then 


102 


FIELD-MANUAL   FOE   ENGINEERS. 


Since,  however,  c  has  but  a  small  range  of  values,  it  will  usually 
be  sufficient  to  assume  it  to  be  constant  and  equal  to  some  mean 
value. 

Suppose  that  in  (8)  c  =  c'  =  Ci,  and  solving  we  find 

SI'  -  S'l 


If  we  observe  S'  and  8  corresponding  to  any  two  distances  I' 
and  I  and  substitute  in  (9),  we  have  Ci  -(-/. 

Having  found  c,  +/,  lay  off  Cm  =  100  +  c,  -f  /,  or  Fm  =  100, 
and  adjust  the  stadia  wires  to  subtend  DE  =  just  one  foot  at  that 
distance. 

Then  from  (8),  writing  c,  for  c  and  c',  we  have 


or,  omitting  accent, 


(10) 


Example.  —  Suppose  at  £'  =  100  we  find  S'  =  1,  and  at  I  =  500 
we  find  S  =  5.0453. 
Then  eq.  (9)  gives 

504.53-500  _ 
~ 


Then,  from  eq.  (10),  I  =  100$  -f-  1.12,  provided  the  stadia  wires 
are  spaced  so  as  to  intercept  1  foot  at  101.12  feet  distance  from 
the  center  of  the  instrument. 

The  foregoing  formulas  must  be  modified  when   the  line   of 

collimation  is  oblique  to  the 
horizon,  which  is  usually  the 
case.  Thus,  in  Fig.  60,  let 
jyE'  =  the  space  intercepted 
on  the  rod  when  the  line  of 
collimation  FH  is  horizontal, 
and  let  DE  =  S  =  the  space 
intercepted  when  the  line  of 
collimation  Fn  makes  an  angle  nFH  =  a  with  the  horizon. 
Let  DFE  =  D'FE'  =  0. 


Ff0 


LEVELING,  STADIA   M-EASUREMENTS,  ETC.  103 

In  Fig.  60, 

8  =  DH  -  EH  =  HF  [tan  (a  +  |0)  -  tan  (a  -  £0)]. 
The  horizontal  reading  desired  is  D'E'  =  2JBFtan  £0. 

D'^'  2  tan  40 

Dividing  gives      — ^—  =  —  — r^. 

8          tan  (a  -f-  £0)  —  tan  (a  —  £0) 

2  sin  40 
But  2  tan  40  =  - 

COS  ^U 

Also,  by  Chapter  III,  formulas  (26),  (27), 

sinO 


tan  (a  -f  £0)  —  tan  (a  —  £6)  = 


cos  (a  -j-  -|0)  cos  (a  —  i0) 

2  sin  |0  cos  4.0 
cos8  a  —  sin2    0* 


Substituting  these  values,  we  obtain 

D'E'        cos-  a  —  sin2 


cos2 


(11) 


If  we  neglect  sin2  |0,  or,  what  is  equivalent,  add  sin8  ^0  to  the 
numerator,  we  introduce  a  relative  error  of 

sin2  40  sin2  4,0 

— —  —  —  =  sin*  |0  sec2  a,  very  nearly. 

cos8  a  —  sm-  ^^         cos^  a 

Again,  if  we  add  sin*  |0  to  the  denominator,  making  it  unity, 

we  introduce  a  relative  error  of  — — ^-r  =  sin2  40  sec2  40. 

cos2  |0 

The  first  of  these  errors  increases  the  fraction,  and  the  second 

decreases  it.     Moreover,  since  0  =  — —  —  —  34'   22". 65,    we 

J(JO 

have  in  all  practical  cases  a  >  |0,     or     sec2  a  >  sec2 10. 

Hence  the  fraction  is,  by  the  double  approximation,  increased 
sliglitly  more  than  it  is  decreased.     Hence 

TV  W  T)'  W 

— —  <  cos8  a  ;  but  —    —  —  cos*  a,  almost  exactly.   .     (12) 

o  o 

The  total  relative  error  is 

e  —  sin2  ^0  (sec2  a  —  sec2  4;0) 
=  sin3  £0  (tan2  a  —  tan2  £0). 


104  FIELD-MANUAL   FOR   ENGINEERS. 

Since  tan1  ^Q  is  very  small  compared  with  tan2  a,  we  have 
e  =  sin5  -|0  tan-  a,  very  nearly, 

=  .000012  tan2  a,  very  nearly. 
Hence 

D'E' 


=  cos5  a(l  -  .000012  tan2  a) 

o 

=:  cos2  a  —  .000012  sin8  a (13) 

Since  sin  £9  =  tan  ^0  =  — —  =  .005.  very  nearly,  the  quantity 

neglected  above  is  only  (.005)4  =  .00000000625  and  does  not  come 
within  the  range  of  the  table. 

The  above  is  the  coefficient  of  reduction  by  which  to  multiply 
the  observed  space  DE  —  S  in  order  to  produce  the  true  space 
D'E'  which  would  be  observed  at  the  same  distance  if  the  line  of 
colliination  were  horizontal. 

Hence  we  have,  from  (10), 

I  =  (1005  +  c  +/)(cos2  a  -  .000012  sin2  a). 

By  eqs.  (25)  and  (26),  Chap.  Ill, 

1  4-  cos  2a 
cos2  a  =  ~  —  SB  .  J  • 


and  Bin'  a  = 

Hence 
I  =  (100S+  «  +/)l  -  ~-  -  .000013  ,  .     (14) 


I  =  (100/8  +  c  +/)  (l  -   ver*2(l  \  very  nearly.     .     .     .    (14') 

\  / 

These  coefficients  may  be  read  from  a  table  of  versed  sines 
without  any  computation  whatever. 

The  last  equation  is  quite  accurate  enough,  but  the  coefficients 
of  Table  XIII  are  calculated  by  the  exact  formula. 

JSxample. — Find  the  coefficient  for  a  =  7°  20'. 

We  write  half  the  vers  14°  40' =  .01629 

and  subtract  from  unity  and  find  coefficient —  .98371 

Another  method  of  procedure  is  that  in  which  the  rod  is  held 
perpendicular  to  the  line  of  collimation. 


LEVELING,  STADIA   MEASUREMENTS,  ETC. 


105 


To  secure  this  position  of  the  rod  a  bar  is  attached  to  it  having 
sights  upon  it,  through  which  the  rodina
…[truncated]