CAMBRIDGE LIBRARY COLLECTION

Books of enduring scholarly value

Mathematical Sciences
From its pre-historic roots in simple counting to the algorithms powering
modern desktop computers, from the genius of Archimedes to the genius of
Einstein, advances in mathematical understanding and numerical techniques
have been directly responsible for creating the modern world as we know
it. This series will provide a library of the most influential publications and
writers on mathematics in its broadest sense. As such, it will show not only
the deep roots from which modern science and technology have grown, but
also the astonishing breadth of application of mathematical techniques in the
humanities and social sciences, and in everyday life.

Statics
A.S. Ramsey (1867-1954) was a distinguished Cambridge mathematician and
President of Magdalene College. He wrote several textbooks ‘for the use of
higher divisions in schools and for first-year students at university’. This book
on statics, published in 1934, was intended as a companion volume to his
Dynamics of 1929 and like the latter was based upon his lectures to students
of the mathematical tripos, but it assumes no prior knowledge of the subject,
provides an introduction and offers more that 100 example problems with
their solutions. Topics include vectors, forces acting at a point, moments,
friction, centres of gravity, work and energy, and elasticity.
Cambridge University Press has long been a pioneer in the reissuing of
out-of-print titles from its own backlist, producing digital reprints of
books that are still sought after by scholars and students but could not be
reprinted economically using traditional technology. The Cambridge Library
Collection extends this activity to a wider range of books which are still of
importance to researchers and professionals, either for the source material
they contain, or as landmarks in the history of their academic discipline.
Drawing from the world-renowned collections in the Cambridge
University Library, and guided by the advice of experts in each subject area,
Cambridge University Press is using state-of-the-art scanning machines
in its own Printing House to capture the content of each book selected for
inclusion. The files are processed to give a consistently clear, crisp image,
and the books finished to the high quality standard for which the Press
is recognised around the world. The latest print-on-demand technology
ensures that the books will remain available indefinitely, and that orders for
single or multiple copies can quickly be supplied.
The Cambridge Library Collection will bring back to life books of enduring
scholarly value across a wide range of disciplines in the humanities and social
sciences and in science and technology.
 Statics
A Text-Book for the use of the Higher Divi-
sions in Schools and for First Year Students at
the Universities

A rt hur Stan l ey R am sey

 C A M B R I D G E U N I V E R SI T Y P R E S S

Cambridge New York Melbourne Madrid Cape Town Singapore São Paolo Delhi

Published in the United States of America by Cambridge University Press, New York

www.cambridge.org
Information on this title: www.cambridge.org/9781108003155

© in this compilation Cambridge University Press 2009

This edition first published 1934

This digitally printed version 2009

ISBN 978-1-108-00315-5

This book reproduces the text of the original edition. The content and language reflect
the beliefs, practices and terminology of their time, and have not been updated.
STATICS
 LONDON
Cambridge University Press
FETTEE LANE

I « W YORK • TORONTO
BOMBAY • OALOTTTTA • MADBAS
Macmillan
TOKYO

Maruzen Company Ltd

All rights reserved

 STATICS
A Text-Book for the use of the
Higher Divisions in Schools
and for
First Year Students at the Universities

by
A. S. RAMSEY, M.A.
President of Magdalene College
Cambridge

CAMBRIDGE
AT THE UNIVERSITY PRESS
1934
PBINTED IN GBEAT BBITAIN
 PREFACE
THIS book has been written as a companion volume to my
book on Dynamics published a few years ago and is intended
mainly for the same class of students, namely, for mathe-
matical specialists in the higher divisions of schools and for
students preparing for a degree in mathematics in the Uni-
versities. It is based in part upon courses of lectures given
during many years to first-year students preparing for the
Mathematical Tripos; and though many readers will already
possess some knowledge of the subject, no such knowledge is
assumed and an attempt has been made in the early chapters
to present the subject in as simple a way as possible and with
very detailed explanations.
The book deals with all those parts of the subject which are
usually covered by the term Elementary Statics, with special
attention to Graphical Statics, Friction and Virtual Work.
For the use of more advanced students there are also chapters
on the statics of flexible strings and the bending of rods, and
the book concludes with a brief account of force systems in
thr6e dimensions.
There are nearly five hundred examples for solution taken
mainly from papers set in either Scholarship, College, Inter-
collegiate or Tripos Examinations, and their sources are
indicated by the letters S, C, I and T. More than a hundred
examples are solved in the text, sometimes by alternative
methods.
It has become the fashion of late to express mechanical
relations in the symbolism of vector algebra and to use the
methods of vector algebra in proving mechanical theorems.
The method has its advantages; but I have not adopted it in
this book, because to most readers it would represent a new
teohnique. The effort necessary to acquire this technique
would in many cases be a hindrance rather than a help to the
grasping of the mechanical principles which the book is
intended to teach, and should, I think, be made at a later
Vi PREFACE
stage. I have therefore made no use of vector analysis, but the
way in which forces and couples obey the vector law of addition
is fully explained, and the chapter on Vectors from the book
on Dynamics appears here again in amplified form.
Readers who are familiar with the books of the late Dr
Routh and of Sir Horace Lamb will realize something of
my indebtedness to both these authors, but I am conscious
of a greater debt than is apparent and I should like to take
this opportunity of.expressing my gratitude.
In conclusion I desire to thank the printers and readers of
the University Press for their excellent work in setting up
the book and eliminating mistakes, and in so far as the book
still contains errors I shall be grateful to anyone who will
point them out.
A. S. R.

November 1933
CAMBRIDGE
 CONTENTS
Chapter I: INTRODUCTION
ART. PAGE
1-1. Force 1
1-2. Rigid Bodies 2
1-3. Types of Forces 2
1-4. L a w of R e a c t i o n 2
1-5. Principle of Transmissibility of F o r c e . . . . 2

Chapter I I : V E C T O R S
2-1. Vectors a n d Scalars 4
2-2. Composition of Vectors 5
2-3. Orthogonal Projections 6
2-4. Analytical Method of Composition 6
2-5. Multiplication a n d Division b y Scalars . . . . 8
2-6. Centroids or Mean Centres 8
2-7. Centroid Method of Compounding Vectors . . . 9
Examples 10

Chapter I I I : F O R C E S A C T I N G A T A P O I N T
3-1. Parallelogram of Forces 12
3-2. Triangle of Forces 14
3-21. Lami's Theorem 15
3-22. Polygon of Forces 16
3-3. Composition of Forces 17
3-31. Expressions for t h e R e s u l t a n t 18
3-4. Conditions of E q u i l i b r i u m 19
3-42. Oblique Resolution 20
3-5. Worked Examples . 21
3-6. Systems of Particles. I n t e r n a l a n d E x t e r n a l Forces . 22
3-7. Constraints. S m o o t h a n d R o u g h Bodies . . . 23
Examples 26

Chapter I V : M O M E N T S . P A R A L L E L F O R C E S . C O U P L E S
4-1. Moment of a Force a b o u t a P o i n t 31
4-2. Theorem of Moments 31
4-21. Moment of a Force a b o u t a Line 32
4-3. Parallel Forces 33
4-31. Couples 35
4-4. Theorem of M o m e n t s for Parallel Forces . . . 35
V1U CONTENTS
ABT. PAGE
4-5. Centre of Parallel Forces . 36
4-51. Centre of Gravity 37
4-53. Analytical Formulae for Centre of Parallel Forces . 38
4-6. Couples 40
4-61-3. Equivalence of Couples 41
4-64. Specification of a Couple 44
4-65. Composition of Couples 45
Examples . 50

Chapter V: COPLANAR FORCES

5-1. Reduction to a Force at any Point and a Couple. . 56
5-2. Conditions of Equilibrium 57
5-3. Analytical Method 58
5-32. Worked Examples 60
Examples 63

Chapter V I : T H E SOLUTION OF PROBLEMS

6-1. Equations of Equilibrium 67
6-2. Constraints and Degrees of Freedom . . . . 67
6-21. Three Forces.. .Coplanar and Concurrent or Parallel . 68
6-3. Problems of two or more Bodies 72
6-4. Reactions at Joints 75
6-42. Working Rules 77
6-5. Chain .of Heavy Particles 81
6-52. Chain of Heavy Rods 83
Examples 85

Chapter V I I : BENDING MOMENTS

7-1,7-2. Stresses in a Beam 95
7-3. Relations between Bending Moment and Shearing
Force 98
7-4. Worked Examples 100
Examples 104

Chapter V I I I : GRAPHICAL STATICS

8-1. Graphical Determination of Resultant . . . . 107
8-2. Pole of Force Diagram 109
8-3. Parallel Forces Ill
8-4. Graphical Representation of Bending Moment . . 113
8-5. Reciprocal Figures 116
8-6. Frameworks 117
 CONTENTS ix
ABT. PAGE
8-63. Bow's Notation 119
8-7. Method of Sections 124
8-8. Distributed Loads 125
Examples 128

Chapter I X : FRICTION
9-1. Laws of Friction 134
9-15. Angle of Friction and Cone of Friction . . . . 136
9-17. Rolling Friction 137
9-2. Problems 138
9-22. Indeterminateness of Problems 139
9-3. Initial Motion 141
9-4. Friction in Unknown Directions 145
9-5. Miscellaneous Problems 149
Examples 152

Chapter X : CENTRES OF GRAVITY

10-1. Formulae 160
10-2. Rod. Parallelogram. Triangle. Quadrilateral. Tetra-
hedron. Pyramid. Cone 161
10-3. Centres of Gravity by Integration 165
10-31. Curves 165
10-32. Areas and Surface Distributions 167
10-341. Volumes of Revolution 171
10-35. Zone of the Surface, of a Sphere 172
10-4. Orthogonal Projection 174
10-41. Quadrant of an Ellipse 175
10-5. Theorems of Pappus 176
10-6. Lagrange's Formula 177
Examples 181

Chapter I X : WORK AND E N E R G Y

11-1. Work done by a Force 187
11-14. Conservative Field of Force 188
11-2. Virtual Work for a Single Particle 190
11-23. Reactions which do no Work 193
11-3. Virtual Work for a System of Bodies . . . . 195
11-31. The Converse Theorem 197
11-33. Applications 198
11-34-6. Worked Examples 200
11-4. Determination of Unknown Reactions . . . . 205
X CONTENTS
ABT. PAGE
11-5. D e d u c t i o n of Conditions of E q u i l i b r i u m . . . 209
11-6. Potential Energy 211
11-62. E n e r g y T e s t of Stability 213
11-7. R o o k i n g Cylinders a n d Spheres 219
11-8. Hooke'sLaw 220
Examples 221

Chapter X I I : F L E X I B L E C H A I N S A N D S T R I N G S
12-1. E q u a t i o n s of E q u i l i b r i u m 230
12-11-14. String on S m o o t h a n d R o u g h Curves . . . 231
12-2. T h e C o m m o n C a t e n a r y 236
12-3. T h e Parabolic Chain a n d Suspension Bridge . . . 240
12-4. Chain of Variable D e n s i t y or Thickness . . . . 242
12-5. C a t e n a r y of Uniform S t r e n g t h 243
12-6. Elastic Strings 244
12-62. T h e Elastic C a t e n a r y 245
12-7. Miscellaneous E x a m p l e s 246
Examples 249

Chapter X I I I : ELASTICITY
13-1. Isotropic Bodies. Uniform E x t e n s i o n . . . . 255
13-2. E x t e n s i o n of B a r s . Y o u n g ' s Modulus . . . . 255
13-21. Variable E x t e n s i o n 256
13-23. Elastic E n e r g y of L o n g i t u d i n a l Strain . . . . 257
13-3. Bending-of B a r s . - 258
13-4. Applications 260
13-45. T h e o r e m of T h r e e M o m e n t s 264
13-5. Combined E x t e n s i o n a n d B e n d i n g . . . . . 266
13-6. Elastic E n e r g y 267
13-7. General E q u a t i o n s for a T h i n R o d b e n t in One Plane . 268
13-8. Euler's Strut 269
13-9. L o a d e d Column 270
Examples 271

Chapter X I V : FORCES I N T H R E E DIMENSIONS

14-1. R e d u c t i o n t o a Force a n d a Couple . . . . 275
14-2. Conditions of E q u i l i b r i u m 277
14-3. P o i n s o t ' s Central Axis. T h e W r e n c h . . . . 278
14-33. Condition for a Single R e s u l t a n t F o r c e . . . . 279
14-4. The Invariants 281
 CONTENTS XI
ART. PAQH
14-41. Geometrical Representation of the Moment of a Force
about a Line 282
14-5. Wrench Equivalent to Two Forces 283
14-51. Resultant of Two Wrenches 284
14-6. Nul Points, Lines and Planes 285
14-7. Conjugate Forces 286
14-8. Equilibrium of Four Forces 289
Examples 291
 STATICS
Chapter I
INTRODUCTION
1*1. Statics is that branch of Mechanics which is concerned
with the conditions under which bodies remain at rest relative
to their surroundings. In such circumstances bodies are said
to be in a state of equilibrium. It is assumed that they are acted
upon by 'forces' which balance one another. The primary con-
ceptions of Statics are forces and the bodies upon which they
act. In Dynamics, force is denned as that which changes or
tends to change the state of motion of a body, but in Statics
we are not concerned with motion save in so far as a force
would cause motion if unbalanced by another force. We get
our ideas of force from the ability which we ourselves possess
to cause, or to resist, the motion of our own bodies or of other
bodies. We are conscious of measurable efforts required to lift
bodies or to support bodies which would otherwise fall to the
ground; we speak of these measurable efforts as forces which
we exert, and we compare them with the 'weights' that we
associate with bodies, by which term we mean 'the forces with
which the Earth attracts them'. Thus the phrase 'a force of
x pounds weight' means a force that would support a body
weighing x pounds.
When holding the string of a kite flying in a gusty wind we
are conscious of a pull or 'tension' on the string at the point
where it leaves the hand that holds it, and we realize that this
'force' is something measurable and of varying measure, and
that it acts now in this direction and now in that as the string
moves hither and thither with the kite. This serves to illustrate
the fact that a force possesses magnitude and acts in a definite
line (that of the string) and may be regarded as acting at a
definite point of that line (the point where the string leaves
the hand). Thus in so far as a force possesses magnitude and
2 INTRODUCTION [i

direction it is a vector as denned in the next chapter, but

whether forces obey the vector law of addition must be dis-
cussed later.
1*2. It is necessary to specify at the outset what it is upon
which forces act. They cannot act upon nothing. Forces act
upon material bodies, and unless the contrary is stated we shall
assume that in every case the bodies are rigid, i.e. that the dis-
tance between each pair of particles that compose a body
remains invariable. Actual bodies are all more or less elastic
and capable of compression, extension or distortion under the
action of forces, and the assumption of perfect rigidity is neces-
sary in order to simplify the building up of the elementary
theory of the subject.
We regard solid bodies as agglomerations of particles held
together by forces of cohesion. We do not put any limit to the
number, large or small, of particles that go to form a body, and
there is no difficulty in the conception of a force acting upon a
minute body or 'a single particle'.
1*3. The forces with which we shall be concerned are of
three types: (a) a push or thrust; (6) a pull or tension; (c) an
attraction such as the 'weight' of a body or 'the force with
which the Earth attracts it'.
1*4. We shall assume (i) that if two equal and opposite
forces act upon a body in the same straight line they have no
effect upon the body's state of rest or motion, i.e. they balance
one another. This is a statement the truth of which can easily
be tested by experiment.
We shall also assume (ii) that the forces mutually exerted
between two bodies are always equal and opposite. This is the
law of reaction enunciated by Newton in the words 'Action and
Reaction are equal and opposite'. It means that if a body A
exerts a force F upon a body B, then B exerts an equal force
F upon A but in the opposite direction.
1*5. A consequence of assumption (i) of 1*4 is what is some-
times called the Principle of Transmissibility of Force, viz.
that a force may be supposed to act at any point in its line of
1-1-1-6] INTEODUCTIOK 3
action provided that the point is rigidly connected with the
body on which the force acts.
For if P and Q are equal and opposite forces acting upon a
body at the points A and B,
then we may say that Q
balances P no matter at
what point B in the line AB
the force Q acts; i.e. a force
Q acting at A produces the
same effect as a force Q acting at B in the same straight line.
1*6. In the following chapter we shall discuss some of the
common properties of a class of physical magnitudes called
vectors and in Chapter i n we shall give further reasons for in-
cluding force in this class and then develop the consequences.
 Chapter II
VECTORS
2*1. The physical quantities or measurable objects of rea-
soning in Applied Mathematics are of two classes. The one
class, called Vectors, consists of all measurable objects of
reasoning which possess directional properties, such as dis-
placement, velocity, acceleration, momentum, force, etc. The other
class, called Scalars, comprises measurable objects of reasoning
which possess no directional properties, such as mass, work,
energy, temperature, etc.
The simplest conception of a vector is associated with the
displacement of a point. Thus the displacement of a point from
A to B may be represented by the line AB, where the length,
direction and sense {AB not BA) are all taken into account.
Such a displacement is called a vector (Latin veho, I carry). A
vector may be denoted by a single letter, e.g. as when we speak
of 'the force P', or 'the acceleration/', or by naming the line,
such as AB, which represents the vector. When it is desired to
indicate that symbols denote vectors it is usual to print them
in Clarendon type, e.g. P, and to write them with a bar above
the symbol, e.g. P, AB.
Since the displacement from B to A is the opposite of a dis-
placement from A to B, we write
BA = -AB,
and take vectors in opposite senses to have opposite signs.
Since two successive displacements of a point from A to B and
from B to 0 produce the same result
as a single displacement from A to
0, we say that the vector AC is
equal to the sum of the vectors AB,
BG and write
1 ~ (1),
and further, if A, B, G, ... K, L are any set of points,
=AB + B~C+... + ~KL (2).
2-1, 2-2] COMPOSITION OF VECTOES 5
Vectors in general are not localized; thus we may have a
displacement of an assigned length in an assigned direction and
sense but its locality not specified. In such a case all equal and
parallel lines in the same sense will represent the same vector.
On the other hand, vectors may be localized, either at a point,
e.g. the velocity of a particle; or in a line, as for example enforce
whose line of action (but not point of application) is specified.
2*2. Composition of Vectors. A single vector which is
equivalent to two or more vectors is called their resultant, and
they are called the components of the resultant. Vectors are
compounded by geometrical addition as indicated in (1) and
(2) of the preceding Article.
A vector can be resolved into two components in assigned
directions in any plane which contains the vector; for if AC
be the vector, and through A, C two lines are drawn in the
assigned directions meeting in B, then AB, BC are the com-
ponents required.
When a vector is resolved into two components in directions
at right angles to one another, each com-
ponent is called the resolved part of the
y
vector in the direction specified. Thus if a
vector P makes an angle a with a given
direction Ox, the resolved parts of P in the
direction Ox and in the perpendicular direc-
tionO*/are
Further, if Ox, Oy, Oz are three lines mutually at right angles,
and the line OF. represents a
vector P, if we construct, z
a rectangular parallelepiped c
with OF as diagonal and F/
edges OA, OB, 00 along Ox,
A
Oy, Oz, as in the figure, then 0 / X

OF=OA+AD+DF
or
And, if P makes angles a, /?, y with Ox, Oy, Oz, we have, since
OAF is a right angle, OA = P cos a, and similarly OB = P cos j8
6 VECTORS [n
f
and OC = Pcosy, so that in this three-dimensional resolution
of a vector its resolved parts in the three mutually perpen-
dicular directions are
Pcosoc, PCOSJS, Pcosy.
It is clear therefore that the resolved part of a vector along
a given line is the orthogonal projection of the vector upon that
line.
2-3. Let AB, BC, CD, ... KJJ be a set of vectors forming
sides of a polygon. Their resultant
is the vector AL which completes
the polygon. Let a, b,c,d, ...k,l
be the orthogonal projections of
the points A, B,G, D,... K, L on
any straight line Ox. Then, with
due regard to signs,
ab + be + cd +... + hi = al.
But these projections ab, be, ...
are the resolved parts of the
vectors AB, BG, ... in the direc-
tion Ox, therefore the algebraical sum of the resolved parts
of a set of vectors in an assigned direction is equal to the
resolved part of their resultant in the same direction.
2*4. Analytical Method. To compound n vectors
P , P , ... P
(i) When the vectors are in the same plane.
Let the vectors make angles <x1( <x2, ... <xm with an axis Ox.
Each vector may be resolved into two components, one in the
direction Ox and the other in the perpendicular direction Oy.
The components in direction Ox are equivalent to a single
vector
X = P t cos ax + P 2 cos a2 +... + PTC cos a.n = 2 (P cos a),
and the components in direction Oy are equivalent to a single
vector
Y = 1?1 sin <xx + P 2 sin a2 +... + P B sin art = 2 (P sin a).
2-2-2-41] COMPOSITION OP VECTORS 7
The two vectors X, Y can now be compounded into a single
vector R making an angle 6 with Ox, such that
Rcos0 = X and Rsin0=Y,
and therefore
i?2 = X 2 + 7 2 and t a n 0 = r / X (1).
(ii) When the vectors are not all in the same plane.
As in 2*2 take three axes Ox, Oy, Oz mutually at right angles
and let the vectors make angles cf.x, <x2,... <xn with Ox, /?x, )S2,... pn
with Oy and y1, y 2 , ... yn with Oz. Each vector may then be
resolved into components of the types
P cos «, P cos /?, P cos y
in the directions Ox, Oy, Oz. The components in direction Ox
are equivalent to a single vector
X = P x cos ax + P 2 cos a2 +... + P m cos <x.n = 2 (P cos a),
similarly the components in directions Oy and Oz are equiva-
lent to single vectors

and
Z = P 1 cos y1 + P 2 cosy 2 +...+P n cosy M = S ( P cosy).
The three vectors X, Y, Z can now be compounded into a single
vector R making angles 6, <j>, tjj with Ox, Oy, Oz, such that
Rcos0 = X, Rcos<£=Y andRcosi/r = Z ...(2),
and by squaring and adding
i?2 = Z 2 + Y2 + Z2 (3).
When the magnitude of B has been found from (3) its direc-
tion is determined by (2).
In obtaining (3) we have assumed that cos2 6 + cos 2 (j> + cos2 >fi = 1 ;
that this is true is seen from the figure of 2*2, where 6, <j>, <\> may denote
the inclinations of OF to the axes, then
OA2 OB2 OC2
2
+ COS ^ + + l
2*41. The method of obtaining the resultant in 2*4 is based on the
fact that if the vectors are all resolved in any assigned direction Ox, then
the resolved part of the resultant in that direction is equal to the alge-
braical sum of the resolved parts of the given vectors.
When the vectors are not all in the same plane each vector is resolved
into three components in the directions of three rectangular axes
8 VEOTOES [II
Ox, Oy, Oz chosen arbitrarily, so that for any direction Ox in space the
resolved part of the resultant is equal to the algebraical sum of the
resolved parts of the given vectors. Also, if each vector be resolved into
two components only, one along Ox and the other in the perpendicular
plane y Oz, the latter components taken together are equivalent to the
resolved part of the resultant in the plane y Oz.

2*5. Vectors may be multiplied and divided by scalar num-

bers. Thus, if we take n equal vectors AB and compound them
together, we get a vector AC such that AG = nAB; and con-
versely A B =—A G.
n
Note that relations of the form
, or
imply that the points A, B, G are in the same straight line.
2"6. Centroids or Mean Centres. If tn1, m 2 , m3, ... mn be
a set of scalar magnitudes associated with a set of points
Alt A2, A3, ...An> the centroid or
mean centre of the points for the
given magnitudes is the point ob-
tained by the following process:
Divide AXA2 at Bx so that
m1A1B1 = m2B1A2;
divide BrA3 at B2 so that

divide B2At at B3 so that (mt+m2 + m3) B2B3 — miB3Ai.

Proceed in this way until all the points have been connected,
then the last point of division Bn_1, usually denoted by the
letter G, is called the centroid or mean centre.
2*61. In order to shew that this process leads in general to
a unique point, i.e. that the point determined by the process is
independent of the order in which the points A X,A 2 ,... Anaxe
joined, we shall first prove that
...+mnAnG = 0 (1).
Assume that this formula is true for the first r points, i.e. that
mxA-L Br_x + m%A2Br_^ + ... + mtAtBT_x=0.
2-41-2-7] CENTROID METHOD 9
Now the next step in the process is to divide Br_1Ar+1 at Br
so that ( m i + m a + . . . + mr) B~Z[Br = mr+1 B^Ar+1,
therefore, by adding the last two lines,

It follows that if the formula (1) is true for the centroid of r

points it is also true for the centroid of r + 1 points; but it is
true for two points, since, by hypothesis,

Therefore the formula (1) is true for the centroid of any

number of points.
Now, if by taking the points in a different order we arrive at
a centroid G', we can shew similarly that
nA~J}'= 0 (2);
and by subtracting (1) from (2) we get
... +mn) GG1 = 0.
Hence G' must coincide with G unless m1 + m2 +... + mn = 0.
In the latter case there is no centroid at a finite distance,
because the last step in the process of finding the centroid
consists in dividing a line in the ratio mn :m1 + m2+ ... + imn_x,
i.e. in the ratio 1: — 1.
2*7. Centroid Method of Compounding Vectors. To
shew, with the notation of 2*6, that, if 0 be any other point,
the resultant of n vectors m1OA1, m2OA2, ... mn0An is
(m-L + m^ ... +mn) 00, where O is the centroid of the points
Ax, A2, ... An for the magnitudes m1,m2, ... mn.
This follows at once by substituting
0Ax=0G+GAx, OA2 = OG + GA2, etc.,
so that
...+mn0An
= (m1+m2+...+mn) OG + (m1GA1 + m2GA2+...+mnGAn),
and by 2*61 (1) the sum of the terms in the last bracket is zero,
therefore
10 VECTORS [II
2*71. When reference is made to the centroid of a set of
points without mention of any associated magnitudes it is
understood that the magnitudes are equal; thus the centroid
of a triangle A BG is a point 0 such that

2'7&. It may be noticed that if P, Q, R are vectors in the lines OA,

OB, 00, then the resultant vector is

\OA + OB+O
where O is the centroid of the points A, B, C £or the magnitudes Pj OA,
p
Q/OB, RjOC; for a vector P in the line OA is the same as -g-j OA.
2*8. We began this chapter with a statement that certain
physical quantities were to be classed together as vectors and
then proceeded to define the properties of vectors and shew
how they can be compounded. In order, therefore, to satisfy
ourselves that a physical quantity such as force or acceleration
is rightly described as a vector, we need adequate reasons for
stating that
(i) it possesses direction,
(ii) it conforms to the laws AB = — BA,
and ~AB + B~C=AC.

EXAMPLES
1. ABC is a triangle. Prove that the magnitude of the resultant of
vectors AB, 2BC and 3 OA is (62 + ca + 26c cos A )* and that its direction
is that of the diagonal through A of the parallelogram of which AB,
AC are adjacent sides.
2. ABCDEF is a regular hexagon. Prove that
A~B+AC + iD + AS+iF= 3AD.
3. AA', BB', CO', DD' are parallel edges of a parallelepiped, of
which A C is a diagonal. Prove that

4. Prove that, if Q is the middle point of AB and O' is the middle

point of A'B', then Z F + B~W= 2OW.
5. Prove that, if O is the centroid of n points Ax, A%, ... An, and
O' is the centroid of n points B1, B2, ... Bn, then
Wn = n GG'.
2-71-2-8] EXAMPLES 11
6. Prove that, if A, B, C are any three points and G is a point such
that AG + BG+CG=0, then G is the intersection of the medians of the
triangle ABC.
1. Prove that, if A, B, C, D are any four points and G is a point such
that AG + B~G+ CG+DG = 0, then G lies on the lines which join each
point to the centroid of the other three.
8. Prove that the lines which join the middle points of the opposite
edges of a tetrahedron meet in a point and bisect one another.
9. Prove that, if H is the orthocentre and 0 is the circumcentre of
a triangle ABC, then
(i) AH+BH+CH='.
(ii)
(iii) ZO sin 24 + BO sin 2B+CO sin 2(7 = 0.
10. Shew that, if m()A+nO~B+pOC = 0 and m + n+p = 0, then
the points A, B, C are collinear.
 Chapter III
F O R C E S A C T I N G AT A P O I N T
3*1. In Chapter i we claimed that the idea of force is a
primary conception attained by experience, and we inferred
that a force possesses magnitude and direction and a definite
line of action, and that its point of application may be assumed
to be any point in the line of action, provided that the point is
rigidly connected with the body upon which the force acts.
3*11. In order to proceed further in constructing a theory
of forces we need some further knowledge of the properties of
force, and it is open to us either to appeal to experiment and
base our theory on so-called 'experimental law', or frankly to
state that the system of Mechanics which we are about to
develop needs a certain hypothesis or postulate concerning the
nature of force, and that, taking for granted this hypothesis,
the system of Mechanics which we erect upon this basis leads to
results in conformity with everyday experience so far as it can be
tested, and thence infer that our basic hypothesis is a sound one.
3-12. If we were to adopt the first alternative we should
start from the 'law of moments' de-
rived from experiments with the lever:
i.e. the property of the lever known to |
Archimedes, that if AB is a straight T
lever which can turn about a fulcrum
G, and weights P and W suspended from A and B keep the
lever at rest in any position, then P.AC = W.BG.
It would be possible by a generalization of this result to
build up a complete system of Statics.
3*13. On the ground of simplicity in developing the theory
we prefer, however, to take the second alternative and adopt
as our further postulate about force the statement that a force
is a vector, so that the magnitude and direction of the resultant
of two forces is determined by the vector law of addition; but
3-1-3-15] PARALLELOGRAM OF FORCES 13
inasmuch as a force is a vector localized in a line due care must
be exercised in order to determine in what line the resultant
force is localized. This determination is embodied in the pro-
position known as the Parallelogram of Forces, which states
that: If two forces acting at a point are represented in magnitude
and direction by the sides of a paral-
lelogram drawn from that point, then
their resultant is represented by the
diagonal of the parallelogram drawn
from that point.
A
Thus if OACB is a parallelogram °
and OA, OB represent forces acting at 0, then OG represents
the resultant of the forces OA, OB.

3*14. The enunciation of the foregoing proposition refers

to forces 'acting at a point'. We have already stated that forces
can only act on a body, so that the 'point' must be a point of
a body, and there is no reason why the body should not be a
single particle.

3*15. Consider two forces acting upon a body and repre-

sented by the lines AB, BG.
In accordance with the vec-
tor law of addition, the re-
sultant force is represented
in magnitude and direction
by the vector AG, but this
is not the position of the re-
sultant. B is the point of
intersection of the forces and
the only point on their lines of action at which both forces can
be considered to act. Hence the resultant is represented by a
line through B equal and parallel to AG; and this accords with
the parallelogram law, for in order that the forces may be repre-
sented by lines drawn from a point we must produce ABtoE
so that BE = AB, then the forces AB, BG are equivalent to
BE, BG and have a resultant represented by the diagonal BF
of the parallelogram BEFO, which is easily seen to be equal
14 FORCES ACTING AT A POINT [ill

and parallel to A C. It is necessary to warn the student against

assuming that the vector law

gives anything more than the magnitude and direction of the

resultant.
3*16. We propose then to take the parallelogram offorces as
our fundamental hypothesis or postulate, and the evidence for
the truth of the hypothesis is not to be found in a formal proof
based upon some other assumption about force, but in the fact
that the theory of Statics built up from this hypothesis accords
with experience so far as it can be tested.
3*2. Triangle of Forces. When a body is in equilibrium
under the action of Ijhree forces, it follows that if two of the
forces intersect, the third force must pass through the point of
intersection, because it must be equal and opposite to the
resultant of the other two.
A simple deduction from the parallelogram of forces is the
following theorem, known as the Triangle of Forces: when three
forces acting at a point can be represented in magnitude and
direction by the sides of a triangle taken in order the forces are in
equilibrium.
Let the lines OA, OB, 00 represent the three forces P, Q, B
acting at 0. Complete the paral-
lelogram OADB. Then AD being
equal and parallel to OB represents
the force Qin magnitude and direc-
tion, and if the forces P, Q, B can
be represented by the sides of a
triangle taken in order, then OAD
is such a triangle because OA, AD represent P and Q. There-
fore DO represents E. But OD is the resultant of P and Q.
Therefore B is equal and opposite to the resultant of P and
Q and the three forces are in equilibrium.
Conversely, if three forces acting at a point are in equilibrium
and a triangle be drawn with its sides parallel to the directions of
the forces, then the lengths of the sides will be proportional to the
3-15-3-211] LAMIS THEOBEM 15
magnitudes of the corresponding forces. For with the same figure
and construction, since the forces are in equilibrium, therefore
R balances the resultant OD of P and Q; therefore DO repre-
sents R, while OA and AD represent P and Q. Hence OAD is
a triangle having its sides OA, AD, DO parallel to the forces
and also proportional to their magnitudes.
3*21. Lami's Theorem. If three forces acting at a point are
in equilibrium, then each force is proportional to the sine of the
angle between the other two.
For, making use of the result and figure of the preceding
theorem, if the forces P, Q, R acting at 0 are in equilibrium,
we have p:Q: R = OA:AD:DO
= sin ODA : sin DOA : sin OAD
= sin BOG: sin COA : sin A OB,
using supplementary angles.
3'211. Example. The ends of alight string ABCD arefixed.Weights
are fastened to the string at B
and O, and the parts AB, BO,
CD are inclined to the horizontal
at angles of 60°, 30° and 55°. The
weight at B is 22 lb., determine
that at C.
This problem can be solved
graphically by drawing tri-
angles of forces or directly by
calculation.
Thus, at B there are three
forces in equilibrium, namely
the weight of 22 lb. vertically
downwards, which we represent
on a convenient scale by a
length Be, and the tensions of
the strings BA, BC. From e we draw a parallel to BC to meet AB
produced i n / and thus get a triangle Bef with its sides parallel to the
forces. Then by the converse of the triangle of forces ef, fB represent
the tensions in BC and BA and by measurement they represent 22 lb.
and 38 lb.
Next, at the point C there are three forces in equilibrium, namely a
weight acting vertically downwards and the tensions of two strings
CB, CD whose directions are known, and we have found that the ten-
sion in CB is 22 lb. So we construct a triangle of forces at C, by marking
off on CB a length Cg=fe and drawing a vertical line gh to meet DC
16 FORCES ACTING AT A POINT [ni
produced in h. Then the sides of the triangle Cgh are in the directions
of the three forces at C and we find by measurement that gh = 38 and
hC =33. Hence, to this degree of accuracy the weight at C is 38 lb. and
the tension in CD is 33 lb.
Alternatively, we may use Lami's Theorem. Thus, if we denote the
tensions in AB, BC, CD by Tlt Tit Ta, we have at B
fji rjn 99
= =
sin 60° sin 150° sin 150°'
whence we get T2 = 22 lb. and Tt = 38-1 lb.
Again, if W denotes the weight at C,

sin95o = sinl20 o = sinl45°'

giving W= 38-2 lb. and T3 = 33-2 lb.
3*22. Polygon of Forces. If any number of forces acting at
a point can be represented in magnitude and direction by the sides
of a polygon taken in order, then the forces are in equilibrium.
D

Let the forces P, Q, R, S, T acting at 0 be represented in

magnitude and direction by the sides AB, BC, CD, DE, EA
of the polygon ABCDE, then the forces are in equilibrium.
For if we compound the forces by the vector law, step by
step, we have
P+Q=AB+BG=AC
P+Q+R=AC+CD=AD
P+ V

where in each equation the resultant, on the right, of the forces

named on the left acts at the point 0. It follows that the forces
are in equilibrium.
3-211-3-3] COMPOSITION OF FORCES 17
We have set out the composition of the forces step by step
but we might have said at once that vectors AB, BG, CD, DE,
EA have a zero resultant, and hence the given forces are in
equilibrium.
3"23. It should be noted that in 3*2, 3*22 it is essential that the forces
should act at a point. It will be seen later that the theorems 3*2, 3'22
are not true if the forces are represented in position as well as in magni-
tude and direction by the sides of the triangle or polygon.
There is no converse proposition to the polygon of forces—if there
were, then by analogy from the converse of the triangle of forces it
would be that if a number of forces acting at a point were in equilibrium
and a polygon were drawn with its sides parallel to the forces the
sides of the polygon would be proportional to the magnitudes of the
corresponding forces; but this is clearly not true because equiangular
polygons are not necessarily similar.
3*24. It should also be noted that in the polygon of forces the forces
need not necessarily all be in the same plane.
3'3. Composition of Forces acting at a point. There are
three methods of finding the resultant of a number of forces
acting at a point and they are the same as the methods given
in Chapter n for finding the resultant of a number of vectors.
(i) The Graphical Method.
Let P, Q, R, S be the forces acting at 0. Starting from any
point A, draw lines AB, BG, CD, DE to represent P, Q, R, S

in magnitude and direction. Then by vector addition AE re-

presents the resultant of P, Q, R, S, and a force equal and
parallel to AE acting at 0 is the required resultant.
(ii) The Analytical Method as in 2*4, resolving the forces
along rectangular axes, adding up the components along each
axis and then compounding the results.
18 FORCES ACTING AT A POINT [ill

(iii) The Gentroid Method, which may be enunciated as

follows: / / any number of forces acting at a point 0 are repre-
sented by the vectors m1OA1, m2OA2> m3OA3, ..., then their
resultant is represented by (m1 + m2 + m3 + ...)OG, where G is the
centroid of the points Alt A2, A3, ... for the multiples mx, m2,
m3,....
The proof of this theorem is contained in 2*6 and 2'7.
We observe that it is not necessary that the points
0, Ax, A2, A3, ... should all lie in the same plane.

3*31. Expressions for the Resultant. Let P1, P 2 , P3, ...

be any number of forces acting at a point and let drs denote the.
angle between the forces Pr and Ps, then the resultant force B is
given in magnitude by
B2 = 2 P r 2 + 22P r Ps cos drs,
where the latter sum includes the products of all the forces taken
in pairs.
(i) When the forces are coplanar: let a.1> oc2, oc3, ... be the
angles which the forces P1,P2,P3,... make with a fixed direc-
tion Ox. Then as in 2*4 (i)
B* = {S (Pr cos ar)}a + {S (Pr sin a,)}2
= S{P r 2 (cos 2 a r + sin2ar)}
+ 22 {P r P g (cos ar cos a3 + sin ar sin ocs)}
2
= SP r + 2SP r P s cos(oc,-a s )

(ii) When the forces are not all in the same plane, as in
2*4 (ii), let a typical force P r make angles <xr, /?r, yr with three
axes Ox, Oy, Oz mutually at right angles. Then
i?2 = {L(Pr cos «r)}2 + {S (P r cosft)}2+ {S (Pr cos yr)f
= S {Pr2 (cos2 <xr + cos2ft+ cos2 yr)}
+ 22 {P r P s (cos ar cos ag + cosftcosft+ cos yr cos ys)}.
Then as in 2-4 (ii)
cos2 ar + cos2ft+ cos2 yr = l,
3-3-3-4] CONDITIONS OF EQTTILIBBITJM 19
and it can easily be proved* that
cos xr cos a8 + cos /?, cos j8s 4- cos yr cos ys = cos 0rs.
Therefore i?2 = 2 P r 2 + 2SP r Pa cos 8n.

3*4. Conditions of Equilibrium of Forces acting at a

point. Forces acting at a point are in equilibrium when their
resultant is zero.
(i) Graphically—if the forces are represented in magnitude
and direction bylines forming the successive sides of a polygon,
the polygon must be closed.
(ii) Analytically—if the forces are resolved along rect-
angular axes, then for forces in one plane as in 2*4 (i)

and R can only be zero if X and Y are both zero; and for forces
not all in one plane as in 2*4 (ii)

and B can only be zero if X, Y and Z are all zero. Hence a

necessary condition of equilibrium is that the sum of the
resolved parts of the forces in any direction is zero—any direc-
tion because the choice of direction of the axes in relation to
the forces is arbitrary.
(iii) Centroid Method—forces acting at a point 0 represented
by vectors m1OA1, mz0A2, m3OA3, ... are clearly in equili-
brium if and onlyif Ois the centroid of the points AX,A%, A 3 ,...
for the multiples m1,m2,m3, ....
* Let OP, OQbe the directions of the forces Pr,Ps. Take OP of unit length and
draw PN perpendicular to OQ, PM per-
pendicular to the plane xOy and ML per-
pendicular to Ox. Then OL, LM, MP being
equal to the projections of OP on the axes
Ox, Oy, Oz are of lengths cos ocr, cos f!r,
cos yr; and 6rs is the angle POQ, therefore
cos 9rs = 0N = projection of OP on OQ
=sum of projections on OQ of
OL, LM, MP
= O L cos a3 + LM cos /3S + MP cos ya
=cos ar cos as + cos )3r cos /J4
+ cosy r cos ys.
20 FORCES ACTING AT A POINT [ill

3*41. In many examples we are only concerned with three

coplanar forces. If the forces are not all parallel, two of them
must meet in a point and for equilibrium it is necessary that
the third force shall balance the resultant of the other two, but
this resultant passes through the point of intersection of the
first two, therefore the third force must pass through the point
of intersection of the first two. Hence if three forces are not all
parallel they can only be in equilibrium when they meet in a
point. The further condition necessary for equilibrium being
that the magnitudes of the forces must be proportional to the
lengths of the sides of any triangle drawn parallel to them (3*2),
or each force proportional to the sine of the angle between the
other two (3*21). The case in which the forces are parallel will
be considered later.

3*42. Oblique Resolution. A force may be resolved into two

components in any assigned directions
which are coplanar with it. Thus if OG
represents a force M and Ox, Oy are any
lines making angles a, jSwith OG and in
the same plane, by drawing GA, GB
parallel to Oy, Ox we form a parallelo-
gram and the force B, or OG is equivalent
to components P, Q represented by OA, OB such that
P:Q:B = OA:AG:OC
= sin /J: sin a: sin (a + j8);
D sin/3 _ , _ sin a
or P= . . Q,-R a n d Q = -7— — B.
sin(a + j8) ( + 8)
3*43. The centroid method may also sometimes be used
conveniently for resolving a force into components along
assigned directions. Thus a force B acting along a line 00
may be resolved into three forces Px, P2, P3 along lines OAl,
OA2, OA3 if we can find points Alt A2, A3 on these lines and
multiples m1,m2, m3 such that G is the centroid of A1, A2, A3
for these multiples; for in such a case
P-y _ J*2 _ P3 -B
m3OA3
3-41-3-5] EXAMPLES 21
3'431. Example. It is easy to prove that the orthooentre H of a
triangle.4 .BC is the oentroid of the points
A, B, G for multiples tan A, tan B, A
tan G. Hence if 0 he any other point,
a force R along OH is equivalent to
forces Plt P 2 , Ps along OA, OB, 00
such that

OAtanA OBtaxiB OCt&nC

R R
:
(tan A + tan B + tan G) OH' OH tan A tan Stan G'
The point 0 need not be in the plane ABO.

3*5. Examples, (i) Two forces P,Q act at a point along two straight
lines making an angle a. with each other and R
is their resultant: two other forces P', Q' acting
along the same two lines have a resultant R'.
Prove that the angle between the lines of action
of the resultants is
cos-1 {PP' +QQ' +cos* (PQ' + P' Q)}/RR'.
[S.]
Let the resultants R, R' make angles 8, 8' with the line of action of
P and P'. By resolving along and perpendicular to this line, we get
i? cos 0 = P + Q cos a, .Rsinfl^Qsina,
and .R'cos0'=:P'+Q'cosa, R'sin8'= Q'sina.
Therefore
) ( )
= (P + Q cos a) (P' + Q' cos a) + QQ' sin2 a
= PP' +QQ' + cos a {PQ' + P'Q),
whence the result follows.
(ii) Two forces given in magnitude act each through afixedpoint, and
are inclined at a constant angle 8; shew that their resultant also passes
through afiscpd point A.
If 8 varies, shew that the locus of A is a O
circle. [S.]
Let B, C be the given fixed points, and P,
Q the magnitudes of the given forces. On BG
construct a segment of a circle containing an
angle BO G=8. Then forces P, Q acting along
OB, OC will have a resultant along a line
OA which cuts the circle again in a point A.
Then so long as the angle BOG between
the forces is constant, since the forces are of given magnitudes the
22 FORCES ACTING AT A POINT [ill
angle which their resultant makes with either force is also constant,
i.e. the angle BOA is constant and therefore A is a fixed point.
Again, when the angle 6 varies, we still have
BA_smBOA_Q
AG~BinAOG~P
but B and C arefixedpoints, therefore the locus of A is a circle. (Circle
of Apollonius.)
(iii) ABCD... is a polygon of n sides, and forces act at a point parallel
and proportional to AB, 2BC, 3CD, etc. Shew that their resultant is
parallel and proportional to
(n-l)OA,
where O is the centroid of all the points B,
C, D, ... excluding A. [S.]
Let LA be the nth side of the polygon;
then, as regards magnitude and direction,
since AB=AO~+ OB,
BC=BO + OO,
and so on, therefore we have

But since 0 is the centroid of the points B, C, D, ... L, therefore

Bd~+~Cd+Dd+...+TJO = 0 (3-4 (iii)),
and the required resultant is represented by (»— 1) OA.
3*6. Systems of Particles. Internal and External
Forces. So far we have only considered the case of forces
acting 'at a point' or on a single
particle. It is a simple step further
to the case of two or more particles
in equilibrium under the action of
given forces and the interactions
of the particles upon one another. Q
For example, let A, B, C be three
particles connected by light threads
which are kept taut by applying
forces P, Q, B to the particles A, B, C respectively. The inter-
action between each pair of particles is then a tension in the
string joining them, which acts in opposite directions on the
two particles. If we denote the tensions in BC, CA, AB by
3-5-3-7] CONSTBAINTS 23
Tlt T2, T3, then, considering each particle separately,
A is in equilibrium under the forces P, T2, T3,
B is in equilibrium under the forces Q, T3, Tx,
and C is in equilibrium under the forces R, Tx, Ta.
And the system of three particles A, B, C is in equilibrium
under the action of the three forces P, Q, R, because when we
add together the nine forces previously named they include
three pairs of equal and opposite forces which cancel out.
This example serves to illustrate the difference between
'internal' and 'external' forces. Considering the three par-
ticles as one system the forces P, Q, R are 'externally applied
forces', i.e. forces applied by some external agent, and the ten-
sions Tx, T 2 , T3 are 'internal forces' since they constitute the
mutual reactions of the particles upon one another. But if, on
the other hand, we consider a single particle, say A, as a self-
contained system, then we do not classify some of the forces
P, T2, T3 as internal, they are all regarded in the same way as
externally applied forces.
There is no hard and fast rule that such and such a force
should always be regarded as either internal or external. Thus
the weight of this book, or the force with which the Earth
attracts it, is an external force when we are considering the
equilibrium of the book as a separate entity; but when we
consider the Earth and all things on the Earth as an entity
attracted by the sun and moon, then the weight of the book and
the counterbalancing reaction of the body which supports it
are internal forces which cancel out.

3*7. Equilibrium under constraint. Smooth and Rough

Bodies. The freedom of a body is often restricted by passive
obstacles, e.g. a donkey tethered to a post, a bead threaded on
a wire; or by conditions imposed in mathematical language,
e.g. a particle shall remain on a given material curve or surface,
the ends of a rod shall lie upon a given curve and so forth. In
such cases the conditions imply the existence of a force or
forces of constraint, such as the tension of the tethering rope,
the pressure of the wire on the bead, of the curve or surface on
24 FORCES ACTING AT A POINT [in
the particle, and of the curve on the ends of the rod. Such forces
are passive in the sense that no more force will be called into
play than is necessary to maintain equilibrium.
When two bodies are in contact at a point, the mutual re-
action B between them is not in general at right angles to their
common tangent plane at the point of
contact but inclined to the normal at an
angle a say. This force B can then be
resolved into components B cos a along
the common normal and B sin a in the
tangent plane. The former component
is called the normal reaction and the
component that lies in the tangent plane is called Friction; the
latter is a measure of the resistance which the bodies them-
selvesofferto any attempt to make the one slide over theother.
Friction depends upon the nature of the substances in contact
and will be discussed fully in a later chapter. Bodies which
are incapable of exerting any frictional force or between which
the force of friction is so small as to be negligible in comparison
with the other forces in action are said to be smooth. In every
case of contact of smooth bodies the mutual reaction between
them must act along the common normal at the point of
contact.
When a rod passes through a smooth ring the mutual re-
action between them is at right angles to the rod, and the ring
can only be in equilibrium if the other forces acting upon it
when resolved along the rod balance one another, for other-
wise the ring would slide along the rod.
When a particle is to remain in contact with a smooth
material surface the reaction of the surface on the particle can
only be along the normal in one direction and it is necessary
that the external forces acting on the particle shall press it
against the surface and have a resultant in no other direction
than the normal, for otherwise the particle would slide on the
surface.
3'71. Examples, (i) A bead of weight Wean slide on a smooth circular
wire in a vertical plane. The bead is attached by a light thread to the highest
point of the wire, and in equilibrium the thread is taut and makes an angle
3-7, 3-71] EXAMPLES 25
6 with the vertical. Find the tension of the thread and the reaction of the
wire on the bead.
Let B be the bead, AB the thread, ^400 the vertical diameter of the
circle, and 0 the centre. Then the angle
0BA = 0AB = 6, and BOC=28.
Hence, if T denotes the tension and R the
reaction, by Lami's Theorem
T: R: W=sin (R, W): sin (W, T): sin (T, R)
= sin 26: sin 6: sin 8,
therefore T = 2Wcos0 and R=W.
Alternatively, BA 0 is a triangle whose sides
are parallel to the forces, so that
T:R:W=BA:0B:A0,
whence 'B/o = 2Wcos8, and R=W.

(ii) The ends of a light string are attached to two smooth rings of
weights w, w' and the string carries a third smooth ring of weight W which
can slide upon it; the rings w, w' are free to slide on two fixed rods inclined
at angles a. and fl to the vertical. Prove that, if <f> be the angle which either
part of the string makes with the vertical, then in equilibrium
cot <}>:tan0: t a n a = W: W + 2w': W + 2w.
Let P, Q, R be the rings of weights W, w, w' respectively. It will
appear in a later chapter that the tension of a string is not affected by

w w'

passing through a smooth ring, so that the tensions in the parts PQ, PR
of the string are of equal magnitude T say. Then since two of the forces
acting on the ring P are equal they must be equally inclined to the
third force, namely the weight, therefore PQ, PR are equally inclined
to the vertical at an angle <f> say.
By resolving vertically for the ring P, we get
2Tcos^=TF (1).
Consider next the ring Q. It is in equilibrium under the action of its
weight w, the tension T of the string QP and the reaction of the smooth
rod AB on which the ring can slide. This reaction is at right angles to
the rod and since it balances w and T its line of action (produced back-
wards) must fall between w and T. This fact tells us how to place the
26 FORCES ACTING AT A POINT [ill
rod AB in the diagram and by a like argument we find the position of
the rod CD.
Then, by resolving along BA for the ring Q, we have
wcosa+Tcos(a + <£) = 0 (2).
By eliminating T between (1) and (2), we get

or (2w+TF)cot^=PFtana.
Similarly (2wr + W) cot <j> = W tan p,
and these are the required results.
It will be observed that we might easily find the reactions of the rods
on the rings Q and R by using Lami's Theorem; but it is not neces-
sary to introduce them into our equations for the purpose of the
problem, and we avoid doing so by resolving along the rod. Of course
a partial application of Lami's Theorem to get the ratio of T to w gives
the same result.

EXAMPLES
1. Find graphically and by calculation the resultant of two forces
of 5 lb. and 3 lb. acting at an angle of 100°, and determine the angle
which the resultant makes with the larger force.
2. Forces of 7 lb., 10 lb. and 12 lb. acting at a point are in equili-
brium; determine graphically and by calculation the angles between
their lines of action.
3. AB is a light string 24 in. long; its upper end A is fastened to a
fixed point, and B is attached to a weight of 10 lb. Determine the force
required to hold B at a distance of 12 in. from the vertical through A,
when the force is applied (i) horizontally, (ii) at right angles to AB.
4. A square A B CD, of side 4 ft., is acted on by a force of 8 lb. along
the diagonal A C and a force of 4 lb. along the diagonal DB. These two
forces can be balanced by a third force acting through a point Pin BC.
Find graphically or otherwise the length of BP.
5. A force of 12 lb. is resolved into two components, one of which is
5 lb. in a direction making an angle of 30° with the former force. Find
the other component in magnitude and direction.
6. Three ropes, all in the same vertical plane, meet at a point, and
there support a block of stone. They are inclined to the horizontal at
angles 35°, 100° and 160°. The tensions in the first two ropes are 200 lb.
and 150 lb. Find graphically the tension in the third rope and the weight
of the block of stone, and verify by calculation.
7. A string ABCD hangs from fixed points A, D, carrying a weight
of 12 lb. at,B and a weight W at C. AB is inclined at 60° to the hori-
zontal, BC is horizontal and CD is inclined at 30° to the horizontal.
By drawing triangles of forces, or otherwise, find W.
 EXAMPLES 27
8. Two forces act at a point and are such that if the direction of one
is reversed the direction of the resultant is turned through a right angle.
Prove that the two forces must be equal in magnitude.
9. Forces of 2, V3, 5, V3, 2 lb. respectively, act at one of the angular
points of a regular hexagon towards the five others in order. Find the
magnitude and direction of the resultant. [S.]
10. Equal forces P act at a point parallel to the sides of a triangle
ABC taken in order the same way round. Prove that the resultant B
is given by iJ2 = P 2 (3-2cos .4 - 2 cos B - 2 cos 0).
11. A weight of 10 lb. is supported by two strings which make angles
of 30° and 60° with the vertical. Find the tensions in the strings.
12. A bead free to slide on a smooth circular wire in a vertical plane
is attached by a fine taut thread to a given point in the vertical line
through the centre of the circle. Shew that the pressure of the wire on
the bead is independent of the length of the string.
13. Two beads of weights w and w' can slide on a smooth circular
wire in a vertical plane. They are connected by a light string which
subtends an angle 2|3 at the centre of the circle when the beads are in
equilibrium on the upper half of the wire. Prove that the inclination
a of the string to the horizontal is given by
, w ~w' .
tan a = -. tanr 8.
w+w
14. Weights w and w' are fastened to the ends of a light rod AB
which is suspended from a point O by strings OA, OB and the vertical
through 0 cuts AB in O. By considering triangles of forces, prove that
w.A0=w'.BG,
and that the tensions in OA, OB are in the ratio
AO.BC-.BO.AC.
15. Three given weights P, Q, B, any two of which are together
greater than the third, are attached to the ends of three strings the
other ends of which are knotted together at a point 0. The strings that
carry the weights P, Q pass over two smooth pegs A and B and the
weight B hangs between the pegs. Give a geometrical construction
for finding the position of the point O.
16. A weight is supported by a light string passing over a smooth
pulley and gently lowered on to a smooth inclined plane. Shew that if
the string is slowly paid out so that the weight slides down the plane the
pressure on the plane increases and the tension of the string decreases.
17. A small ring is capable of motion along a wire of circular shape,
and is attracted by forces varying as the distance, and of equal absolute
intensities, to two given external points in the plane of the circle.
Shew that in any position of the ring the resultant attraction passes
through afixedpoint; and give a geometrical construction for a position
of equilibrium. [S.]
28 FORCES ACTING AT A POINT [HI
18. A weight of 10 lb. hangs by a string from afixedpoint. The string
is drawn out of the vertical by applying a force of 5 lb. to the weight.
In what direction must this force be applied in order that in equilibrium
the deflection of the string from the vertical may have its greatest
value? What is the amount of the greatest deflection?
19. 0 is any point in the plane of a triangle ABO; D, E, F are the
middle points of the sides. Prove that the resultant of forces repre-
sented in magnitude and direction by OE, OF, DO is represented by
OA. [S.]
20. P is any point in the plane of a triangle ABO, and D, E, F are
the middle points of its sides; prove that forces AP, BP, OP, PD, PE,
PF are in equilibrium. [S.]
21. A system of n forces acting at a point is represented in magnitude
a n d direction b y t h e lines A1 Bx, A2B2,... AnBn, w h e r e A1, A2, ••• An
and Blt JE?2,... Bn are the corners of two regular polygons. Find a line
representing the resultant force in magnitude and direction. [S.]
22. Any two points E, F are taken on the sides AB, CD of the
parallelogram A BCD; O, H are the middle points of AC, EF respec-
tively; prove that the resultant of forces acting at a point represented
by EC, ED, FA, FB is represented in magnitude and direction by
4HG. [C]
23. Three equal forces are represented in magnitude and direction
by OA, OB, OC, where O is the circumeentre of the triangle ABC.
Prove that the resultant is represented in magnitude and direction by
the line joining O to the orthocentre of the triangle.
24. ABO is a triangle; prove that six forces represented by AH,
BH, OH, AA', BB', CO' will be in equilibrium; where H is the ortho-
centre and AA', BB', OC are diameters of the circumcircle. [S.]
25. Forces P, Q, R acting at a point O are in equilibrium and a
straight line meets their lines of action in A, B, C respectively; shew
that, with certain conventions of sign,

26. Four straight lines in a plane intersect, two at a time, in the six
points A, B, C, D, E, F, and 0 is any other point. Find a geometrical
construction for the resultant of the forces represented by OA, OB,
OC, OD, EO, FO. [S.]
27. Three forces acting at a point are parallel to the sides of a
triangle ABO, taken in order, and proportional to the cosines of the
opposite angles; shew that their resultant is proportional to
(l-8cos.Acos.BcosC) i . [S.]
28. P is a point in the plane of a triangle ABO, forces act at P
towards the angular points represented by PA.sin A, PB.aiaB and
 EXAMPLES 29
PC. sin C respectively: shew that the resultant is
._. A B O
4PI. cos -_- cos g- cos -^,
where / is the incentre of the triangle. [S.]
29. A particle P is attracted towards each of four points A, B, O,D
by forces equal to fj,xPA, fi2PB, ix.3PC, ^PD. Shew that it will rest in
equilibrium only at the centroid of A, B, C, D, for multiples fix, fi2>

Shew also that if a force Q be applied to the particle its position of

equilibrium will be displaced from this position a distance
Ql(p\ + / 2 + H* + / « )
m the direction of Q. [S.j
30. If O, A, B, C, D, ... Zhe any fixed points, and if any points P,
Q, R, ... be taken in AB, BO, CD,.... YZ, ZA respectively, so that
AP _BQ_GR_
P~B~QG~RD '
shew that the resultant of the forces represented by OP, OQ, OR,... is
constant in magnitude and direction. [I.]
31. Forces act at appoint P, along the lines joining P to the vertices
of a triangle ABO, and the magnitudes of the forces are
(sin 25 + sin 20) AP, (sin 2C +sin 2.4) BP, (sin 2A + sin 2JB) OP
respectively; prove that their resultant passes through the centre of
the 'nine points' circle of the triangle ABO, and find its magnitude.
[S.]
32. ABGD is a quadrilateral inscribed in a circle whose centre is O.
Prove that forces along AO, OB, CO, OD proportional to the areas
BCD, CDA, DAB, ABO are in equilibrium.
33. I is the centre of the inscribed circle of a triangle ABO and D is
any point. Prove that a force ID is equivalent to three forces -JJ- AD,
-BD
34. Prove that, if E, F. are the feet of the perpendiculars from two
corners of a triangle ABO upon the opposite sides, a force P acting
along EF can be replaced by forces Pcos.4, PcosiS, PcosC acting
along the sides. [S.]
35. Forces are represented by the lines joining one corner of a
parallelepiped to the middle points of the edges which do not meet in it.
Prove that their resultant is represented by 5£ times the diagonal
through that corner. [S.]
36. Two forces are represented by radii vectores drawn from the
focus of an ellipse to the curve, the sum of the two being given. Shew
that the locus of the extremity of the resultant is a straight line parallel
to the minor axis. [S.]
30 tfOBCES ACTING AT A POINT [ill
37. JfABCD be a plane quadrilateral such that the triangles ABC,
ADC are equal in area, O any point in the plane, H, K the middle points
of A C, BD, and a, b, c, d the middle points of the sides, then forces
10a, iOb, \Oc, \Od and OH are equivalent to forces OA, OB, OC,
OD and KO. [S.]
38. The line of action of a force P cuts the sides of a triangle in given
points. Shew how to find three forces acting along the sides of the
triangle which shall have P for their resultant. [S.]
39. Forces P, Q, B, S, T acting at a point form a system in equili-
brium. Iftheanglesbetween(iJ, £),(&, T),(T, P) be a, /J,y respectively,
find, by the polygon of forces or otherwise, what two forces, in the
directions of P and B respectively, will balance Q. [S.]
40. O, O' are inverse points with respect to a circle centre C and P
is a particle anywhere on the circle acted upon by forces ^ p a along
PO' and Q ~ along OP. Shew that the resultant force on the particle
is normal to the .circle, provided

41. Three small particles are placed in a narrow, smooth, circular

tube and repel each other in such a way that the force between any two
of them is proportional to the product of their masses and their distance
apart. Shew that when they are in equilibrium their masses are pro-
portional to the sides of the triangle formed by drawing tangents to
the tube at the positions of the particles. [S.]

ANSWERS
1. 5-36 lb., 33° 24'. 2. 92° 5', 123° 37', 144° 19'.
3. (i)J3Qv'31b., (ii)51b. 4. 5 | ft.
5. 8-06 lb. inclined at 18° 4' to the resultant. 6. 146-6 lb., 312-6 lb.
7. 41b. 9. 10 lb. along the diagonal. 11. 5 -\/3 lb., 5 lb.
18. At right angles to the string; 30°.
21. nOO', where O, O' are the centres of the polygons.
31. 8PN sin A sin B sin C, where N is the 'nine points centre'.
39. {Psin(a + ^ + y) + Tsin(a + j8) + >S'sina}/sin(a + ^ + y)
and {.Rsini
 Chapter IV
MOMENTS. PARALLEL FORCES. COUPLES
f
4 i. The Moment of a Force about a Point is denned to
be the product of the force and the perpendicular distance of
its line of action from the point.
Thus if P be a force and p the distance of its line of action
from a point O, then the moment of
P about 0 is Pp.
Also if the force P is represented
by AB, then Pp = AB .p = twice the
area OAB.
This gives a geometrical representation of the moment of a
force by an area.
The moment of a force about a point may be regarded as a
measure of the tendency of the force to cause rotation about
the point.

Thus if P and Q are forces applied to a disc in its plane and

0 is a point about which the disc can turn, in fig. (i) P and Q
tend to cause rotation about O in the same (counter-clockwise)
sense, and we should regard their moments as of the same sign;
but infig.(ii) P and Q tend to cause rotation about 0 in oppo-
site senses and we should regard their moments as of opposite
signs. Either sense—clockwise or counter-clockwise—may be
chosen as the positive sense.
4*2. When any number of coplanar forces act at a point the
algebraical sum of their moments about any point in their plane is
equal to the moment of their resultant about the same point.
32 MOMENTS. PARALLEL FORCES. COUPLES [IV
Let any number of forces Plf P2, P3, ... act at a point A,
and let R be their resultant, and 0 the point about which
moments are to be taken. Let the
forces R, Plt P 2 , P3, ... be at the
distances r, plf p2, p3, ... from 0
and make angles 9, a ls a2, a3, ...
with AO.
For simplicity only one of the
forces P 1 ; P 2 , P3,... is shewn in
the figure.
The algebraical sum of the moments about 0 of Px ,P2,P3,...

= OA x alg. sum of resolved parts of the forces along a line Ax

at right angles to OA
= OA x resolved part of R along Ax
= OA x Raind = Rx OAsin0 = Rr
= moment of R about 0.
Cor. The algebraical sum of the moments of any number of
coplanar forces which meet at a point about a point on the line
of action of their resultant is zero.
4*21. The Moment of a Force about a Line is defined to be the
product of the resolved part of the force at
right angles to the line and the shortest
distance between the force and the hne.
Thus if P be a force and Oz be a line
which does not intersect P, MN=p the
shortest distance between P and Oz, and N
8 the angle between P and a hne through M Psinfl
parallel to Oz, then P sin 6 is the resolved
part of P at right angles to Oz and Pp sin 8
is the moment of P about Oz.
If P intersects the line Oz or is parallel "
to Oz, then the moment of P about Oz is
zero, because in the one case p = 0 and in /
the other sin 8 = 0.
4*22. If through any point A on the line of action of a force P a plane
xOy be drawn at right angles to any line Oz meeting it in 0, then the
moment of P about Oz is equal to the moment about 0 of the projection of
the force P on the plane xOy.
Let NM =p be the shortest distance between P and Oz, and let DM
4-2-4-3] PARALLEL FORCES 33
parallel to Oz make an angle 9 with P and meet the plane xOy in D.
Then the projection of the force P
on the plane xOy is a force P sin 8
acting in the line AD, and OD is
the perpendicular from 0 to AD,
so that the moment about 0 of Psinfl
the projection of P on the plane
xOy
= 0DxPsin8
=NM x Psin 8 = Ppsin 0
= moment of P about Oz.

4*23. When two forces act at a

point the algebraical sum of their
moments about any line is equal to
the moment of their resultant about
this line.
Let P, Q be the forces acting at A, R their resultant and Oz the given
line.
Through A take a plane xOy at right angles to Oz and let P', Q', R'
denote the projections of the forces on the plane xOy.
Then the algebraical sum of the moments of P and Q about Oz is
equal to the algebraical sum of the moments of P' and 0/ about 0 (4*22)
= moment of R' about 0 (2'41 and 4'2),
= moment of R about Oz.

4*3. Parallel Forces. Let two forces P, Q acting at a point

0 have a resultant R, and let any
straight line intersect the lines of action
of the forces in A, B and 0.
We may use the centroid method as
in 2*72 and 3*43 to express the relations
A
between P, Q, R and the lines in the ° / p
figure, or we may proceed as follows:
Take moments about A; then since P has no moment about
A, we have
mt. of Q about A = mt. of R about A
or Q.ABsinB = R.ACamC
or Q.AB.OG=R.AG.OB
Q _ R
or
OB.AC~OC.AB'
34 MOMENTS. PARALLEL FORCES. COUPLES [iV

Similarly by taking moments about B, we find that

P JR
OA.GB OG.AB'
., , P Q R ,,,
therefore _ _ _ _ _ « _ _ (i).
Now let the point 0 move to an infinite distance while A, B
remain fixed and the magnitude of the forces P, Q remain un-
altered. Then OA, OB, OG become parallel lines, so that the
resultant of two parallel forces is parallel to its components.
Also since the ratios OA : OB: OG tend to equality, therefore
(1) becomes T> n r>
GB AC AB
But AB = AC+CB,
therefore B =P +Q (3).
Figure (ii) shews what (i) becomes when 0
moves to an infinite distance. The direction in
which 0 moves becomes the direction of the
parallel forces but it does not affect the rela-
tions (2) and (3).
Next consider the case in which the transversal ABC cuts
P on the other side of 0 as infig.(iii). By the same argument
as before we obtain the relation
P Q * (4)
OA.BC~OB.AG~OG.AB
where BC, AC, AB are all positive lengths measured in the
same sense.

O
(iii)

Then keeping A, B and the magnitudes of P, Qfixedas

before, let 0 move to an infinite distance in any convenient
4-3—4-4] COUPLES 35
direction so that P, Q, R become parallel as in fig. (iv) (where
the direction in which O is moved in (iii) is along BO).
As before, since the ratios OA: OB: OG tend to equality,
(4) becomes P n p
K
BG AC AB • '•
But AB = AG-BG,
therefore R = Q-P (6).
Parallel forces are said to be like when they act in the same
direction as in fig. (ii), and unlike when, as in fig. (iv), they act
in opposite directions.
It follows from (3) and (6) that, if parallel forces are counted
as positive or negative according to the sense in which they act,
then the resultant of two parallel forces is equal to their
algebraical sum; and from (2) and (5) it follows that the re-
sultant divides AB in the ratio Q: P, internally or externally
according as P and Q are like or unlike. Kg. (iv) shews the case
Q > P; if however P> Q the point G will lie on BA produced
and R will act in the same sense as P.

4*31. Couples. The method of 4*3 fails to determine a

resultant for two equal unlike parallel forces, for when Q = P
and the forces are unlike, (6) gives R — 0 and (5)gives BG = AG,
implying that the point G is at an infinite distance. Such a pair
of equal parallel forces acting in opposite senses is called a
couple. A torque is another name for a couple. We shall discuss
the properties of couples later in this chapter.

4*4. The theorem of moments of 4*2 can now be extended

to include the case of parallel forces.
For this purpose we shall prove that o
the algebraical sum of the moments of
two parallel forces about any point in
their plane is equal to the moment of
their resultant about the same point.
Let 0 be any point in the plane of the parallel forces P, Q.
From 0 draw a line OA CB at right angles to P and Q, meeting
3-z
36 MOMEKTS. PARALLEL FORCES. COUPLES [iV
them in A, B and their resultant R in 0. Then the algebraical
sum of the moments of P and Q about 0
= P.OA + Q.OB

= (P + Q)OC, since P . A O = Q . C B ,

= moment of R about 0.
This proof can be adapted for the case of unlike parallel
forces, and for the case in which 0 lies between the forces.
It follows by the process of addition, step by step, that the
algebraical sum of the moments of any number of coplanar parallel
forces about any point in their plane is equal to the moment of
their resultant about the same point; so that the theorem of 4*2
now holds goqd for all coplanar forces that have a resultant.
4*41. Since parallel lines project orthogonally into parallel lines and
the ratio of lengths on a line is unaltered by orthogonal projection, it
follows that the resultant of two parallel forces projects orthogonally
into the resultant of their projections, and the same is true for the
resultant of any number of parallel forces. Hence by the aid of 4'22 we
deduce that the algebraical sum of the moments about a line of any number
of parallel forces is equal to the moment of their resultant about the same
line.
For, with the figure of 4*22, the algebraical sum of the moments
about Oz of parallel forces of which P is the type = algebraical sum
of moments about 0 of the projections of the forces on the plane
xOy = moment about O of the resultant of the projections of the
forces = moment about 0 of the projection of the resultant force
= moment about Oz of the resultant force.

4*5. Centre of Parallel Forces. Let n parallel forces

Ply P2, ... Pn act at points Ax, A2, ... An. Then, whether the
forces are all in the same sense or not, if we compound them to-
gether step by step, the resultant of P1 and P 2 is P1 + P2 acting
at a point Bx on AXA2 such that P x . A1B1 = P 2 . .5^42; next the
resultant of P x + P 2 at J5X and P 3 at A3 is Pt + P2 + P 3 acting
at a point B2 on B±A3 such that (P1 + P2)B1B2 = P3.B2AS
and so on. But this process is that defined in 2*6 as finding the
centroid of the points Ax, A2,... An for multiples P1,P2,... Pn,
and as proved in 2*61 it leads to a unique point 0 through
4-4-4-51] CENTRE OF PARALLEL FORCES 37
which the resultant of the parallel forces P 1 ; P 2 , ... Pn acts,
save in the exceptional case in which
P 1 + P 2 + . . . + P re = 0 (1).
The vanishing of the algebraical sum of the parallel forces
does not necessarily imply that the resultant is zero, because
the forces may be equivalent to a couple. In order to determine
whether this is so or not, we may divide the forces into two
groups according as they represent positive or negative terms
in (1). The resultants of the two groups are then equal in
magnitude and opposite in sign, and because neither of these
separate sums is zero we can use the centroid method above to
determine definite points on the lines of action of these re-
sultants of the two groups of forces. The system is thus either
in equilibrium or equivalent to a couple according as the two
resultants are found to act in the same straight line or not.
When the system is not equivalent to a couple, the unique
point 0 determined as above is called the centre of parallel
forces. Its position depends only on the positions of the chosen
points A1,A2,... An and the relative magnitudes of the forces.
It does not depend on the direction of the parallel forces, so
that the forces can be turned about their points of action
remaining parallel to one another, and so long as they remain
parallel the point 0 remains fixed.
4*51. Centre of Gravity. The weights of the particles which
compose a material body constitute a set of forces directed to-
wards the Earth's centre and therefore parallel with sufficient
accuracy for our purpose. Since they are like parallel forces they
have a resultant equal to their sum and this is called the weight
of the body and it acts through a definite point in the body, viz.
the centre of the parallel forces, which is called the centre of
gravity or centre of mass of the body, or its centroid.
As thus defined the weight of a body is a force acting through
a definite point of the body irrespective of the orientation of
the body in reference to surrounding objects; and in writing
down relations between the forces acting on a body the
weights of its elements may in general be replaced by a single
force acting at the centre of gravity
38 MOMENTS. PARALLEL FORCES. COUPLES [IV
4*52. Gentrobaric Bodies. When we speak of the weight of a
body we mean 'the force with which the Earth attracts it'. In accord-
ance with Newton's Law of Gravitation 'every particle in the universe
attracts every other particle with a force directly proportional to the product
of their masses and inversely as the square of the distance between them'.
Newton proved that the resultant force of attraction exerted upon an
external particle by the particles of a uniform sphere is the same as if
the whole mass of the sphere were collected into a particle at its centre,
and that the forces of attraction exerted upon the particles of a uniform
sphere by an external body have a resultant passing through the centre
of the sphere. But this does not hold good for bodies in general of
various shapes. It follows that there is nothing approximative in the
statement that the weight of a uniform sphere acts through its centre.
It is an exact statement, and such a body is called centrobaric in that it
has a real centre of gravity, but bodies in general only possess centres
of gravity subject to the approximation that the Earth's attractions
on the particles which compose the body may be regarded as parallel
forces.

4'53. Analytical Formulae for Centre of Parallel Forces.

(i) Coplanar forces. Take rect-
angular axes Ox, Oy in the plane y
and let (x1, t/1), (x2, y2), ... be the
co-ordinates of the points of ap-
plication Als A2, ... of the parallel
A, •A 3
forces P1, P2, ... and (x, y) the co-
ordinates of the centre G. Since the
position of G does not depend on the A2
direction of the parallel forces, we
may suppose them to act parallel to the axis Oy, then by taking
moments about 0, we get

or x
Similarly, by taking the forces parallel to Ox, we get

(ii) When the forces are not coplanar. Let (xt, yx, zx), (x2, y%, z a ),... be
the co-ordinates of the points of application A1,A2, ... and (x, y, z) the
co-ordinates of the centre O of parallel forces P1,Pi,P3,... referred to
rectangular axes Ox, Oy, Oz.
Since the position of O is independent of the direction of the parallel
4-52-4-541] CENTRE OP PARALLEL FORCES 39
forces, we may suppose them to be parallel to Oz, then by taking
moments about Oy and using
4-41 we get

or
Similarly

and 2=S(Ps)/SP.
The method only fails to give
a result when S P = 0, i.e. when
the forces are either in equili-
brium or equivalent to a couple.

4'54. Position of equilibrium of a heavy body suspended from

a given point without any other constraint.
The body is in equilibrium under the action of two forces
(i) its weight which acts •vertically through its centre of gravity
and (ii) the reaction at the point of support. Hence in the
position of equilibrium the line joining the centre of gravity to
the point of support must be vertical.

4*541. Examples, (i) A uniform triangular lamina ABC is sus-

pended from the corner A, and in equilibrium the side BO makes an angle a
with the horizontal. Prove that
= cotB~cotC. [T.]
It can be shewn that the centre of gravity
of a uniform triangular lamina is at the inter-
section of the medians. Hence the vertical
through A bisects BC in D.
Let AE be perpendicular to BO, then
EAD = a and

= (EB-CE)IEA
= cot B — cot 0.
(ii) A heavy uniform rod of length 2a turns freely about a pivot at a
point in it, and suspended by a string of length I fastened to the ends of the
rod hangs a smooth bead of equal weight which slides on the string. Prove
that the rod cannot rest in an inclined position unless the distance of the
pivot from the middle point of the rod is less than a2 /I. [T.]
Let AB be the rod and O the ring. Then since the bead is smooth the
tensions in both parts of the string are the same so that CA and OB
must be equally inclined to the vertical CD, or the vertical CD bisects
40 MOMENTS. PARALLEL FORCES. COUPLES [IV
the angle ACB. The weights of the rod and the ring are equal and
parallel forces and they are balanced by the
reaction of the pivot O, therefore we must
have GO = OD, where G is the middle point of
the rod.
Then because CD bisects the angle ACB
therefore
AD_IW a+GD a-GD
AC~CB' ° r A(
., , 2GD 2a
therefore AC-CB'AC+CB'
or GD^AC-CB).
But AC- CB must be less than AB or 2a, therefore GO or \GD
must be less than a2/l.
4*55. When a heavy body is placed upon a horizontal plane or on a
plane inclined to the horizontal and sufficiently rough to prevent
sliding, we may determine whether the body will stand or fall over as
follows:
If the part of the body in contact with the plane has a convex boun-
dary line—such as a triangle or a convex polygon—take this as 'the
boundary of the base'; and in other cases, e.g. if the body were a chair,
tie a string tightly round the parts of the body in contact with the plane,
e.g. the feet of the legs of the chair, and take the line of the string as
defining 'the boundary of the base'. The weight of the body acts
vertically downwards through its centre of gravity and in equilibrium
must be balanced by the pressures of the plane on the body, and how-
ever these pressures are distributed their resultant could not act-out-
side the boundary of the base. Hence it is necessary for equilibrium that
the vertical through the centre of gravity shall intersect the plane
within the boundary of the base.

4*6. Couples. Couples play an important part in the general

theory of systems of forces and we shall now establish some of
their principal properties.
Since a couple consists of two equal and opposite parallel
forces, the algebraical sum of the resolved parts of the forces
in every direction is zero, so that there is no tendency for the
couple to produce in any direction a displacement of trans-
lation of the body upon which it acts; and, as we saw in 4*31,
the couple cannot be replaced by a single force.
The effect of a couple must therefore be measured in some
other way, and, since it has no tendency to produce transla-
4-541-4-61] COUPLES 41
tion, we next consider what tendency it has to produce rota-
tion.
Let the couple consist of two forces of magnitude P. It is of
course assumed that they are both acting upon the same rigid
body. Let us take the algebraical sum of the moments of the
forces about any point 0 in their plane as the measure of their
tendency to turn the body upon which they act about the
point 0.
From 0 draw a line at right angles to the forces meeting
them in A and B.

(ii) (iii)

Then the algebraical sum of the moments about 0 (taken in

the counter-clockwise sense in the three figures, which illustrate
all cases) is
in fig. (i) P.OB-P.OA = P.AB,
infig. (ii) P.OB + P.AO = P.AB,
and in fig. (iii) P.AO-P.BO = P.AB.
Thus the measure of the tendency of the couple to cause
rotation about a point 0 is independent of the position of 0 in
the plane and is measured by the product of either force and
the distance between the forces; this distance is called the
arm of the couple and the product is called the moment of the
couple, and it maybe counted as positive or negative according
to the sense in which the couple tends to cause rotation.
4*61. The conclusion arrived at in the preceding article is of
great importance. We may summarize it thus: the only way of
measuring the effect of a couple is by considering its tendency
to cause the rotation of a body; this tendency can be expressed
by taking moments about a point, and the result obtained is
independent of the point chosen and is a definite constant for
each couple, namely the product of either force and the dis-
42 MOMENTS. PARALLEL FORCES. COUPLES [iV

tance between the forces, which product we call the moment

of the couple.
Further, the effect of a couple does not depend upon the
absolute measure of its forces, but only upon its moment, i.e.
the product force x distance; and if forces are measured in
pounds and distances in feet, a couple is measured in foot-
pounds.
When a body is acted upon by a couple and we consider its
tendency to turn about a point in the plane of the couple, we
have seen that the effect of the couple is independent of the
relative positions of the point and the couple and depends only
on the 'moment of the couple'; therefore couples of equal
moments in the same plane are equivalent, and couples of equal
and opposite moments in the same plane balance one another.
4*62. The equivalence of couples of equal moment is im-
portant because it means that we can represent a couple of
given moment in a given plane by forces of any magnitude we
like to choose and place one of them to act at any chosen point
of the plane in any assigned direction, provided that we place
the equal parallel force at the distance necessary to make a
couple of the prescribed moment.
Thus a force and a couple in the same plane can always be
reduced to a single force. For let P be a
force acting at 0 and 0 the moment of a
counter-clockwise couple, then we can re-
present 0 by two forces of magnitude P at
a distance p apart such that Pp = 0, and
place them so that one of them balances
the given force P acting at 0, and then the
resultant of the given force P and the
couple G is the single force P parallel to the given force and
at a distance G/P from it.
4*621. Example. Make use of 4?62 tofindthe resultant of two parallel
forces.
Let P, Q be the given parallel forces and let a line at right angles to
them cut them in A and B.
At A introduce two equal and opposite forces of magnitude Q in the
same straight line as P. This does not affect the resultant of the system,
4-61-4-63] COUPLES 43
but we may now consider the system to consist of a force P+Q acting
at A and a couple of moment Q.AB (clockwise infig.(i)). This couple
can then be represented by equal and opposite forces of magnitude
P + Q> one of which can be placed at A to balance the former force P + Q

P+Q/

P+QY VP+Q
(i) (ii)
acting at A (fig. (ii)), and the other will then cut AB at a point C such
that
(P+Q)AC=Q.AB,
or P.AG=Q.GB,
and the resultant of the original forces P at A and Q at B is this force
at C.

4*63. Couples of equal moment in parallel planes are equi-

valent.
Consider a couple of forces P at the ends of an arm AB. In
any parallel plane take a line A'B' equal and parallel to AB,

and let AB' and BA' intersect at 0. The effect of the given
couple is not altered by introducing at 0 two equal and
opposite forces of magnitude 2P parallel to the forces of the
44 MOMENTS. PARALLEL FORCES. COUPLES [iV

given couple; but we can now compound P at A and 2P

(upwards) at 0 into a single force P at B', since
AB':0B' = 2P:P.
Similarly P at B and 2P (downwards) at 0 have a resultant P
at A'. We are therefore left with a couple of forces P at the ends
of the arm A 'B' in a plane parallel to that of the original couple,
so that any couple is equivalent to a couple of equal moment
in any parallel plane.

4*64. Specification of a Couple. In order to specify a

couple it is therefore only necessary to know (i) the direction
of the set of parallel planes in any one of which the couple may
be supposed to act, (ii) the magnitude of its moment, (iii) the
sense in which it acts (clockwise or counter-clockwise).
These three properties of a couple can be represented com-
pletely by drawing a straight line
(i) at right angles to the set of parallel planes to indicate
their direction,
(ii) of a measured length to indicate the magnitude of the
moment,
(iii) in a definite sense to indicate the sense of the couple in
accordance with a convention to be stated.
For example, a measured length MN on a certain straight
line will represent a couple whose moment on a certain scale is
of magnitude MN acting in a plane at
right angles to MN in a certain sense,
it being understood that a measured
length NM on the same straight line
will represent a couple of equal numeri- M
cal moment but acting in the opposite ^— g
sense. It is open to us to adopt any ^>
convention of signs; e.g. a clockwise
couple in the plane of the paper may be represented by a line
at right angles to the paper on either side, i.e. towards or from
the eye of the observer provided that a counter-clockwise
couple is represented by a line drawn in the opposite sense.
4-63-4-65] COMPOSITION OF COUPLES 45
A line drawn to represent a couple is called the axis of the
couple.
4*641. A convenient method of specifying the sign of a couple is to
take a set of three mutually perpendicular axes Ox, Oy, Oz and agree
that a couple in the plane xOy which would cause rotation from Ox
towards Oy shall be represented by a positive length along Oz. Simi-
larly a couple in the plane yOz which would cause rotation from Oy
towards Oz will be represented by a positive length along Ox, and a
couple in the plane zOx which would cause rotation from Oz towards
Ox will be represented by a positive length along Oy.

(i)

Figures (i) and (ii) represent 'right-handed' and 'left-handed' sets of

axes respectively, so-called because if what is known as a 'right-handed
screw' were placed along Oz with its head at 0 in (i), it would have to be
twisted in the sense of rotation from Ox towards Oy to drive it along Oz.

4*65. Composition of Couples. A couple is a vector.

We have seen that a couple can be represented by its axis, i.e.
a straight line of definite length in a definite direction; in order
to shew that a couple is a vector it only remains to prove that
couples can be compounded by applying the vector law of
addition to their axes.
The composition of couples in the same or in parallel planes
is effected by the algebraical addition of their moments, be-
cause we can choose a common arm for all the couples and we
then have only to compound by algebraical addition the sets of
forces which act at the ends of this arm. In terms of the axes
of the couples the composition is effected by the addition of
lengths measured on the same straight line.
When the couples are not in the same plane nor in parallel
planes, consider two couples in planes which intersect. We
46 MOMBKTS. PARALLEL FORCES. COUPLES [IV

may place the forces anywhere we please in the given planes

and take them to be of any convenient magnitude provided
that we adjust the,arms of the couples suitably to make couples
of given moments. Take a length AB on the line of inter-
section of the planes together with a parallel length CD to
represent the forces of one couple, and let the other couple
be represented by forces BA and EF; and let AD, AE at right
angles to AB be the arms of the couples.

Then since the forces AB and BA balance one another, the

resultant couple consists of the forces CD and EF, with an arm
DE at right angles to both. Since the forces of all three couples
are equal, the lengths of their axes will be proportional to their
arms. Let XA equal to AD and at right angles to the plane
ABCD represent the first couple, and A Y equal to AE and at
right angles to the plane BAEF represent the second couple.
Then the triangle XAY is equal to the triangle DAE turned
through a right angle in its own plane, so that XY is equal to
DE and at right angles to the plane GDEF, and therefore XY
represents the resultant couple on the same scale as XA and
AY represent the component couples; and since the vector
XY is the sum of the vectors XA and A Y, therefore the axes
of couples obey the vector law of addition.
It follows that when couples are represented by their axes
they may be compounded and resolved in the same way as
4-65-4-68] COUPLES 47
forces, and that any formula that has been established for the
resultant of forces holds good for the resultant of couples.
There is, however, an important distinction, that a force is a
vector localized in a line, but a couple is a non-localized vector
and all equal and parallel straight lines can be used indiffer-
ently to represent the same couple.

4*66. Three forces represented in magnitude, direction and

position by the sides of a triangle taken the same way round are
equivalent to a couple.
Let the lines A B, BG, GA represent the forces. By the vector
law of addition the forces GA, AB
have a resultant GB acting at A.
Hence if AD be drawn equal and
parallel to GB, then the couple BG,
AD is equivalent to the three given
forces. Since the arm of the couple
is the perpendicular from A to BG, its moment is represented
by twice the area of the triangle.

4*67. Forces completely represented by the sides of a plane polygon

taken the same way round are equivalent to a couple whose moment is
represented by twice the area of the polygon.
This theorem follows from the last, if
the polygon is divided up into triangles by
joining one corner to each of the others,
and equal and opposite forces are intro-
duced represented by the lines so drawn.
By this means we obtain a number of sets
of three forces acting round triangles to
each of which the last theorem applies, and
by adding the equivalent couples we obtain
a single couple whose moment iss represented by twice the area of the
polygon.

4'68. Examples, (i) A stairway is made of n equal uniform rect-

angular blocks placed on top of each other and each projecting the same
distance at the back beyond the one below. The top block is supported from
below at its outermost point. Shew that the stairway can stand without
mortar if, and only if, 2l>(n—l)a, where 21 is the width of each block
and a is the width of the tread. [T.]
Let r blocks counting downwards from the top be supported by a
force P acting at A, the outermost point of the top block, and a force
48 MOMBKTS. PARALLEL FORCES. COUPLES [iV
Q between the rth block and the (r + l)th, which must act either at B
or to the left of B in the figure, where BO is the top of the (r+ l)th
block. If this condition holds whatever number of blocks we take, then
the staircase will stand and not otherwise.

B
r+1

If x is the distance of Q from A, we must therefore have

x^ra (1).
But by moments about A

(r-l)} (2).
But P + Q = rW, and, by considering the top block alone, P is not
greater than \W: for P is at a distance I from the middle point of the
block, and, if P were greater than \W, the other force supporting the
block would have to act at a distance greater than I from the middle
point of the block which is impossible.
Therefore P<W>
and Q>(r-i)W.
Hence, from (2),

or rl+$ar(r-l)^x(r-l) (3).
But from (1) the necessary and sufficient condition for equilibrium is

hence from (3) rl + \ar (r — 1) > ar (r — £),

or
or
This condition is to hold good for all values of r from 1 to n — 1, therefore
the staircase stands if, and only if,

(ii) Two spheres, whose radii are alt a 8 , rest inside a smooth hollow
vertical right cylinder, of which the external radius is c, and the internal
radius is b, where b<a1 + ai. Prove that, if the sphere % is the lower, the
4-68] EXAMPLES 49
cylinder will not overturn if its weight exceeds w[2b — a1 — ai)/c, where w
is the weight of the upper sphere. [T.]
Since the cylinder is smooth the reactions between it and the spheres
are horizontal forces, and the only vertical
forces acting on the spheres are their
weights and the reaction of the plane on
which the lower one stands. This reaction
passes through the point of contact and
though it is equal to the sum of the
weights w and w' of the spheres it cannot
balance their resultant, so the vertical
forces form a couple, viz. w downwards
through the centre of the upper sphere
and w upwards in the vertical through
the centre of the lower sphere. The
moment of this couple is w(2b — a1 — a2).
This is a couple acting on the spheres and if they remain in equilibrium
this couple must be balanced by another, therefore the mutual re-
actions between the spheres and the cylinder constitute a couple of
the same moment.
Considering next the equilibrium of the cylinder, when it is just
about to overturn the reaction of the horizontal plane upon it will pass
through a point C on its outer edge and form with the weight W of the
cylinder a couple of moment We tending to keep the cylinder upright,
and the cylinder will not overturn if the moment of this couple exceeds
that of the other couple which acts in the contrary sense, i.e. if
W > w (26 -a1-ai)/c.
(iii) ABCD is a skew quadrilateral and forces are completely repre-
sented by the lines AB, BC, CD, DA; prove that they are equivalent to
a couple of moment
AC x BD x sin (angle
between AC and BD).
We have A~B + B~C = AC acting at
B and CD + DA = CA acting at D.
These two forces AC at B and CA at
D form a couple whose arm is
x
BD x sin (angle -^
between A C and BD), ^_
s
and hence the required result.
(iv) A body of weight W is suspended by two equal threads AP, BQ;
the points of support A, B are on the same level at a distance a apart; and
the threads are fastened to two points P, Q on the body so that PQ is hori-
zontal and PQ — b. A couple O is applied about a vertical axis and the
body is deflected through an angle 6 with PQ at a depth h below AB. Shew
that
G = lW{abjh) sin 6. [T.]
50 MOMENTS. PARALLEL FORCES. COUPLES [IV
Let MN be the proj ection of A B on the horizontal plane through P Q,
and O the middle point of MN and PQ. Let T denote the tension in
either thread and <f> its inclination
to the vertical. The weight W is
supported by the vertical com-
ponents of the tensions, therefore

The couple 0 must balance the

horizontal components of the ten-
sions which act along PM and QN,
therefore
G=Tsin<j,xDE,
where DOE is perpendicular to PM and QN.
But OD.PM = 2 area POM = $a. J6 sine,
and PM = h t a n <j>,
therefore DE —i(abjh) sin Boot <f>,
and G = \T cos $ (abjh) sin 6 = \W (ab/h) sin 9.

EXAMPLES
1. A uniform beam 20 ft. long and weighing 100 lb. is supported at
its ends. If weights of 30 lb., 40 lb. and 50 lb. are placed on the beam at
distances of 6 ft., 9 ft. and 15 ft. from one end, find the pressure on each
support.
2. Two weights of 10 lb. and 2 lb. hang from the ends of a uniform
lever a yard long and weighing 4 lb. Find the point in the lever about
which it will balance.
3. ABOD is a straight line, AB = 2 ft., BO = 4 ft., CD = 3 ft. Forces
of 2 lb. and 6 lb. act vertically upwards at A and C, and forces of 7 lb.
and 4 lb. act vertically downwards at B and D. Find the magnitude
and line of action of the resultant.
4. A loaded horizontal plank 8 ft. long rests on supports at its ends.
A load of 20 lb. is removed from a point 1 ft. from one end of the plank
and placed at the middle. By how much is the pressure on each support
altered.
5. A uniform rod AB, of weight W, rests horizontally on props at
M, N, where AM = \AB, AN = \AB. The rod remains at rest when
weights P and Q hang from A and B. Prove that
6P+W>iQ>2P-W.
Prove that, if Q = P, an additional vertical force which, applied at
 EXAMPLES 51
B, will disturb equilibrium cannot be less than £ (2P + W), whether it
act upwards or downwards.
6. A uniform rod AB, 6 ft. long, weighing 10 lb., can turn freely
about a hinge at A. It carries a weight of 7 lb. at B, and is kept in
equilibrium by a horizontal string CD, which is attached to a point D
of the rod 5 ft. from A and to a point C 3 ft. vertically above A. Find
the tension of the string.
7. Prove that if four forces acting along the sides of a square are in
equilibrium they must be equal in magnitude.
8. A uniform ladder 13 ft. long, weighing 20 lb., rests against a
smooth vertical wall with its lower end on rough ground 5 ft. from the
bottom of the wall. Prove that the pressure of the ladder on the wall is
a force of 4$- lb.
9. Two horizontal wires inclined to one another at an angle 2a are
attached to the top of a vertical post movable about its lower end,
which is fixed. The post is supported by a stay inclined at an angle /?
to the vertical and fastened to a point two-thirds of the way up the
post. Find the tension in the stay when the tensions in the wires are
both T. Find also the horizontal and vertical components of the
reaction at the lower end of the post.
10. A gate is supported by two hinges in a vertical line at a distance
3 ft. apart. The breadth of the gate is 5 ft. and its weight is 100 lb. The
upper hinge exerts a horizontal force only and will yield when this
force exceeds 250 lb. If a boy weighing 140 lb. stands on the gate
without the hinge yielding, what is the greatest possible distance of
his centre of gravity from the line of the hinges ?
11. Three like parallel forces P,Q,B act at the corners of a triangle
ABC; prove that their resultant passes through the circumcentre of
the triangle for all directions of the forces if P, Q, R are in the ratios
acos^4. :6cos J3:ccosC [C]
12. A, B, C are fixed points and D is on a fixed straight line. Find
the locus of the centre of four equal parallel forces acting at A, B, C, D
respectively. [S.]
13. A table, whose surface is horizontal, is in the shape of an equi-
lateral triangle ABC, having a leg at each corner. If we neglect the
weight of the table itself, find where a weight must be placed in order
that the pressures on the legs may be as 1:2:3. [S.]
14. A uniform rectangular board A B CD of given weight is supported
in a horizontal position on three pegs placed at the corner B and the
middle points of the sides AD and CD. Find the pressures on the pegs,
and prove that they are independent of the form of the rectangle. [S.]
4-2
52 MOMENTS. PARALLEL FORCES. COUPLES [IV
15. A light horizontal beam, freely jointed at O, is supported and
loaded as shewn. Determine the reactions at the supports.

Determine the greatest weight which does not destroy equilibrium

no matter where it is placed on the beam. [S.]
16. Parallel forces of 1, 2, 3 lb. weight act at the corners A, C, E of
a regular hexagon ABCDEF, and forces 4, 5, 6 lb. weight act at the
corners B, D, F in a direction parallel to the first three but in the
opposite sense. Find the point of application of their resultant. [S.]
17. Forces of 1, 2, 3, 4, 5, 6 lb. weight respectively act at the corners
of a regular hexagon inscribed in a circle of radius R, in directions per-
pendicular to the plane of the circle and in the same sense. Shew that
2 7?
the resultant cuts the plane at a distance -=- from the centre. [S.]
18. A uniform rectangular brick 9 in. by 4£ in. by 3 in. is placed on
an inclined plane which is so rough that sliding does not take place, the
longest edges being horizontal and the shortest perpendicular to the
plane. Find the greatest angle to which the plane can be tilted without
the brick toppling.
19. A uniform solid right circular cone stands on an inclined plane
of angle 30°. Prove that the cone will upset if its height exceeds 4-y/3
times the radius of its base.
[The distance of the centre of gravity of such a cone from its base is
one-quarter of its height.]
20. Two uniform rods AB, BC, rigidly jointed at B so that ABC is
a right angle, hang freely in equilibrium from a fixed point at A. The
lengths of the rods are a and 6 and their weights are wa and wb. Prove
that, if AB makes an angle B with the vertical,
(,2
tan0=-
21. The ends A, B of a light rod A B are j oined by light inextensible
strings AO, BO to a fixed point 0, AO, BO being equal in length and
at right angles to one another. Weights Wt and W2 are suspended
from A and B. Shew that the rod will take up an inclination 8 to the
horizontal, where

22. A sphere of weight W and radius a is suspended by a string of

length I from a point P, and a weight w is also suspended from P by a
 EXAMPLES 53
string sufficiently long for the weight to hang below the sphere. Shew
that the inclination of the first string to the vertical is

23. A point P is taken inside a triangle ABC, AP cuts BC in D.

Equal parallel forces in the plane of the triangle act at A, B, C, P, and
an equal but opposite parallel force acts at D. Find the position of P
so that the resultant may always pass through P, independently of the
direction of the forces. [T.]
24. P, Q are two like parallel forces. If two equal and opposite
forces S in any two parallel lines at a distance 6 apart in the plane of
P. Q are combined with them, shew that the resultant is displaced a
distance bS/(P + Q). [T.]
25. Forces act along the sides BC, CA, AB of a triangle. Shew that
they reduce to a couple only if the forces are proportional to the sides.
26. If a couple of given moment and three forces in its plane are in
equilibrium, shew how the magnitudes of the forces can be obtained,
graphically or otherwise, when their lines of action are known. [I.]
27. If forces completely represented by the sides of a triangle taken
in order are in equilibrium with three equal forces acting at the corners
of the triangle along the tangents to the circumcircle the same way
round, prove that the triangle must be equilateral. [S.]
2 8. Three forces act along the sides of a triangle A B C taken in order
and are represented by these sides on a certain scale. They are in
equilibrium with three other forces acting along the sides of the pedal
triangle. Prove that these forces must be represented by the sides of
the pedal triangle on a scale which is to the first as
2 cos ^L cos B cos 0 : 1 . [T.]
29. Forces Xa, Xb, Xc parallel to the sides of a triangle ABC act at
A', B', C, the centres of the escribed circles: shew that they are
equivalent to a couple of moment 2A (a + 6 + c) -R, where B is the radius
of the circumscribed circle. [S.]
30. ABCD and A'B'C'D' are two coplanar parallelograms. If
forces act along AA', B'B, CC, D'D proportional to these respective
lengths, shew that they reduce to a couple, and find its moment. [S.]
31. Three forces Xa, Xb, Xc act along AO, BO, CO respectively, where
0 is the orthocentre of the triangle J 4 . S C If they are rotated through
the same angle 8 about A, B, C respectively, shew that they become
equivalent to a couple whose moment is 4AA sin 6, where A is the area of
the triangle ABC. [S.]
32. Three coplanar forces are represented by straight lines AA',
BB', CC. Shew that the forces will reduce to a couple if the c.o.'s of
the triangles ABC, A'B'C are the same, and if the straight lines are
not concurrent at a finite or infinite distance. [S.]
54 MOMENTS. PARALLEL FORCES. COTJPLES [ IV
33. AB, A'B' are two equal lines in the same plane, C and C their
middle points. Prove that forces represented by AA' and BB' are
equivalent to a force represented by 2CC and a couple whose moment
is 2AC2 multiplied by the sine of the angle between AB and A'B'.
[S.]
34. P, Q, R are taken on the sides BG, GA, AB of a triangle, dividing
each in the Same ratio 1 + A: 1 — A in the same sense round the triangle.
Prove that forces represented by AP, BQ, GR are equivalent to a
couple whose moment is 2AA, where A is the area of the triangle. [S.]
35. Three uniform heavy rods OA, OB, OC mutually at right angles
are rigidly connected at O. Their lengths are a, b, c respectively and
they are suspended freely from 0. Find the angle that OA makes with
the vertical. [S.]
36. A circular tray of radius a stands on a single circular foot of
radius b. If w is the whole weight of the tray and its support,findhow
far from the centre a weight W can be placed without the tray falling
over. [S.]
37. Two smooth spheres each of radius r and weight W are placed
inside a hollow cylinder of radius a, open at both ends, which rests on a
horizontal plane with its axis vertical: prove that, in order that the
cylinder may not upset, its weight must at least be
2w(l-rA, r>\a. [S.]
38. A uniform bar of weight W and length 2a is suspended, from two
points in a horizontal plane, by two equal strings of length I, which are
originally vertical: shew that the couple, which must be applied to the
bar in a horizontal plane, to keep it at rest at right angles to its former
Wa*
direction is , [S.]

39. A body of weight W is suspended by two equal threads of length

I, fastened to two fixed points, in the same horizontal line distant 2a
apart and also tied to two fixed points in the body, distant 26 apart.
Calculate the couple required to turn the body through an angle 8;
and prove that when I is large compared with a or b, the couple is
W(abjl) sin 8. [C]
40. Two straight uniform beams A CD, GDB of weights W^ and W2
are braced together at C and D and the compound beam thus formed is
supported at its ends A and B. Prove that the bracings have to sustain
tensions and thrusts equal respectively to
lADjBD.W^BG.Wj) , 1 BC(AD. W^ + AG. W2)
2 BC.AD-BD.AO 2 BG.AD-BD.AG '
Prove that, if the weights of the beams are inversely proportional to
their lengths, these forces constitute a torque acting on each beam. [T.]
 EXAMPLES 55
41. Equal lengths AA', BB' are marked off in the same direction
along a given straight line, and equal lengths CC, DD' along another
given line. Prove that forces completely represented by AC, C'A',
OB, B'C, BD, D'B', DA, A'D' are in equilibrium.
42. OA BC is a tetrahedron. Shew that couples represented by the
areas OBG, OCA, OAB can have a couple represented by the area
ABC as resultant.
43. ABCD is a skew quadrilateral and forces represented com-
pletely by AB, BC, CD, DA act in those lines; give a geometrical con-
struction to determine the moment of the system about any line and
prove that the moments about any two lines are as the cosines of the
inclinations of those lines to the common perpendicular to A 0 and BD.
[C]

ANSWERS
1. 114-5 lb., 1055 1b. 2. 9 in. from one end. 3. 3 1b.,
4 ft. 8 in. from A. 4. 1\ lb. 6. 19-2 lb. 9. STcosoccosec0;
Tcosa, 3Tcosacot£. 10. 3fft. 12. A parallel at three-fourths
of the distance from the centroid of A, B, G. 13. Take D on BO
sothat BD = \BC, and (? on AD so that AQ=%AD. 14. One-
third of the weight on each peg. 15. 6 1b. 16. IfM, N axe
the middle points of AF and CD, O divides MN in the ratio 7:11.
18. 56° 19'. 23. The middle point of the median through A.
30. The moment may be represented by the area
ABCD - A'B'C'D' - 2ABB' + 2DCD'.
35. cos- 1 {a a /V'( a4 + 6 4 + c*)}- 36. b{
39. (Wab sin B)ly/(ll-aa - 6 2 + 2ab cos 6).
 Chapter V
COPLANAR FORCES
5*1. A system of coplanar forces acting on a rigid body can be
reduced to a single force acting at an arbitrarily chosen point in
the plane of the forces together with a couple.
Let O be the chosen point and P any one of the forces at a
distance p from 0. At 0 introduce a pair of
equal and opposite forces equal and parallel
to P. This does not affect the body, but the
three forces P may now be regarded as a
force P at 0, parallel to the original force P
and in the same sense, together with a
couple of moment Pp. Hence any force P
can be transferred to act at 0 parallel to its
original direction by the introduction of a suitable couple
whose moment is the moment of P about 0.
Let all the forces of the given system be transferred to act
at 0 in this way by the introduction of suitable couples. Then
the forces acting at 0 can be compounded into a single force R
and the couples can be compounded into a single couple of
moment G by the algebraical addition of their moments. Also
R is the resultant of all the given forces moved parallel to them-
selves to act at 0, and G is the algebraical sum of their
moments about 0.
5*11. Since R is the resultant of all the forces moved
parallel to themselves to act at a point, it follows that M has
the same magnitude and direction no matter at what point 0
in the plane it acts. But G is the algebraical sum of the mo-
ments of the forces about 0 and therefore depends in general
upon the position of 0 in the plane.
5*12, When neither R nor G is zero (fig. (i)) the system can
be reduced to a single force R. For the couple can be repre-
sented by two forces R at a distance GjR apart and so placed
5-1-5-2] CONDITIONS Or EQUILIBRIUM 57
that one of them balances R at 0 (fig. (ii)) as in 4-62. The
system has then been reduced
to a single force R acting at a
point 0' at a distance G/R
from 0.
5-13. When G = 0 and R is
not zero. The system is then / ' ' 6'
equivalent to a single force R,
or the chosen point 0 is a point
about which the algebraical sum of the moments of the forces
is zero. It follows that when a system of coplanar forces has a
single resultant the algebraical sum of the moments of the
forces about any point on the line of action of the resultant is
zero.
5*14. When i? = 0 and G is not zero. The system is then
equivalent to a couple G. Also the vanishing of R does not
depend upon the choice of 0, and therefore in this case G is
independent of 0 or the algebraical sum of the moments of the
forces about every point in the plane is the same.
5-15. When R = 0 and G = 0 the system is in equilibrium.
5*2. Conditions of equilibrium of a system of coplanar forces.
When the system of forces is reduced to a single force R at
an arbitrarily chosen point 0 and a couple G, it is necessary
and sufficient for equilibrium that R=0 and G = 0.
These conditions can be expressed in other ways, thus:
(i) Since R represents the resultant of all the forces moved
parallel to themselves to act at a point, R will vanish if and
only if the algebraical sums of the resolved parts of the forces
in two perpendicular directions in the plane are separately
zero. Again G = 0 means that the algebraical sum of the
moments of the forces about the point 0 is zero. Hence it is
necessary and sufficient for equilibrium that the algebraical sums
of the resolved parts of the forces in two perpendicular directions
in the plane and of the moments of the forces about a point in the
plane should separately be zero.
(ii) From 5*13 it follows that when the algebraical sum of
58 COPLANAR FORCES [V

the moments of the forces about a point is zero if there is a

resultant force it must pass through this point. Hence another
way of stating the conditions of equilibrium is that the alge-
braical sums of the moments of the forces about three non-collinear
points 0, 0', 0" should separately be zero. For the vanishing of
the sum of the moments about 0 means that G = 0, and if there
is a resultant E it must pass through 0; but in like manner it
must pass through 0' and 0" and this is not possible because
0, 0' and 0" are not collinear, therefore R = 0.
(iii) If the algebraical sums of the moments of the forces
about two points 0 and 0' are zero, then G = 0, and if there is
a resultant R it must pass through 0 and 0'. This latter pos-
sibility will be excluded if we know that the algebraical sum of
the resolved parts of the forces in any direction not perpen-
dicular to 00' is zero, for a force along 00' would have a com-
ponent in every direction save perpendicular to 00'. It follows
that the system of forces is in equilibrium if the algebraical sums
of the moments of the forces about two points and of the resolved
parts of the forces in any direction not at right angles to the line
joining the points are separately zero.

5*3. Analytical Method. The conclusions that we have

established so far in this chapter can also be obtained analytic-
ally as follows:
Let P be any one of a system of
coplanar forces which acts at a
point A whose co-ordinates are x, y p(
referred to rectangular axes in the
plane. Let Px, Py be the com- ^ o
ponents of P parallel to the axes
V
Ox, Oy. At 0 introduce equal and
opposite forces Py along the axis of y. One of these forces
together with Py at A forms a couple of moment xPy in the
counter-clockwise sense in the figure. Similarly by introducing
at 0 equal and opposite forces Px along the axis of x and
combining one of these forces with Px at A we get a couple yPx
in the clockwise sense. The two couples are equivalent to a
single couple xPy — yPx in the sense of rotation from Ox
5-2-5-31] CONDITIONS OF EQUILIBRIUM 59
towards Oy, and we have in addition the forces Px, Py at 0
parallel to their original directions. If we proceed in the same
way with all the forces of the given system and compound
together by addition the forces along Ox and along Oy and
add the couples, we get a force at 0 with components
X = SP3 along Ox
and Y = T,Py along Oy,
and a couple G = 2 (xPy — yPx).
The forces X, Y can be compounded into a single force B
given by R 2 = X 2 + Y2
and this force is clearly the resultant of all the given forces
moved parallel to themselves to act at 0, while G is the alge-
braical sum of the moments about. 0 of the forces.
It is evident that G depends upon
the choice of origin 0 and that B y
does not.
If we reduced the system to a o'
force B at a point 0' of coordinates
(|, 17) instead of at 0, we can deduce
the value of the corresponding
couple G' from the formula obtained for G by subtracting $,
7) from x, y respectively, so that

= 2 (xPv - yPx) -
or G' = G
Again, since the algebraical sum of the moments about any
point on the line of action of the resultant is zero, 0' is on the
resultant if G-£Y+ r)X = 0,
and, if we regard £, rj as current co-ordinates, this equation is
the equation of the line of action of the resultant force.
5'31. Conditions of Equilibrium.
(i) It is clearly necessary and sufficient for equilibrium that
B = 0 and G = 0,
or that X = 0, 7 = 0 and G = 0.
60 COPLANAR FORCES [V
These conditions expressed in words are the same as those
of 5-2 (i).
(ii) An equivalent set of conditions is that the algebraical
sums of moments of the forces about each of three non-
collinear points should be zero. For if the sums of the moments
about 0, 0', 0" are zero, we may write
0 = 0,

and G" = O-
where g, ij are the co-ordinates of 0', and £', rj' those of 0".
Since 0, 0', 0" are not collinear, therefore rj/g =(= ij'/f' a n ( l * n e
equations can only be satisfied by X = 0 and Y = 0 so that
these conditions, i.e. (ii) of 5*2, are equivalent to (i).
(iii) Lastly, conditions (iii) of 5*2 are represented by equa-
tions 0 = 0,

and X = 0,
taking the direction of resolution to be the axis of x and the line
00' not perpendicular to Ox, i.e. £ is not zero. The equations
then require that Y = 0, and so the conditions are equivalent
to (i).
5*32. Examples, (i) ABCDEF is a regular hexagon and 0 is its
centre. Forces of magnitudes 1, 2, 3, 4, 5, 6 act in the lines AB, OB, CD,
ED, EF, AF in the senses indicated by the order of the letters. Reduce the
system to a force at O and a couple, and find the point in AB through
which the.single resultant passes.
Take the axis Ox along BE, and y
a perpendicular axis Oy as in the A
.6 F

A
figure, then using the notation of
5 "3 we have /H
X=-l.i-2.£ + 3-4.£
-5.J + 6
= 3, O / x
and
+
*•• 2 2 2

C ^3 D
5-31-5-32] EXAMPLES 61
Hence the single force at 0 is R, where i?2 = 9 + 3=12, so that
Y 1
R — 2-\/3, and it makes with Ox an angle tan" 1 = , i.e. tan" 1 —^ or 30°.
Again Gf=sum of moments about O of the forces

= -3a,
in the counter-clockwise sense, where a is the distance of 0 from a side
of the hexagon. The couple is therefore 3 in the clockwise sense and it
can be represented by two forces of magnitude R so placed that one of
them balances R at O and the other, the single resultant, is parallel to
R and at a distance O/R, i.e. 3a/2\/S or V3a/2, from 0. It is easily
seen that the line of action passes through the middle points of AB
and AF.
Here we should note that there is a much more direct method of
finding the point in AB through which the resultant passes, if that fact,
and nothing more about the resultant were required. Thus let H be the
required point, and let AH = z. Then, since H is a point through which
the resultant passes, the algebraical sum of the moments of the forces
about H is zero; therefore
6zsin60 o -5(a + zsin60°) + 4.2a-3(2a-zsin60°)
+ 2(a-zsin60°) = 0,
or V3z-a = 0.
Therefore
(ii) AB, CD are segments of twofixedcoplanar lines. IfABbe of fixed
length and likewise CD, shew that the resultant offorces represented by A O,
DB has a fixed direction and magnitude.
We have a force
AC=AB+ B~O acting at A,
and DB = I)C + C~$ acting at D.
Then the forces BO at A and CB at D
constitute a couple and no matter what
its moment may be a couple cannot affect the magnitude or direction
of the resultant. There remain the two forces AB and DC which are
of given magnitudes and act in given straight lines so that the resultant
must be fixed in direction and magnitude.
(iii) Forces P, Q, R act along the sides BC, CA, AB ofatriangle ABO,
and forces P', Q', R' act along 0A,0B,0C, where 0 is the centre of the
circumscribing circle. Prove that, if the six forces are in equilibrium,
P cos A + Q cos B + RCOBC=0,
, PP' , QQ' , RR' n rq ,
and f- —• -i = 0. [>-J
a b c
Since the angle BOO is 2A, therefore the distance of the force P from
62 COPLANAR FORCES [V
0 is OB cos A, with similar expressions for the distances of the forces
Q and R. Hence, by taking moments about
0 for the six forces, we get

Again the moment about A of a force

represented by BO is twice the area ABC,
therefore the moment about A of the force
P along BC is ^ x 2AABC. Hence, by
-DO
taking moments about A, we get
~ x 2AABC+ — x 2AA0C-^5 x 2AA0B = 0.
UL> \J\j UJtj
Similarly, by taking moments about B and G, we get

UJx OC
and ~^-rx2AC0A =(
Now multiply these equations respectively by P', Q', R' and add,
and we get the result
PP' QQ' RR'_

(iv) If a system offorces in one plane reduces to a couple whose moment

is 0 and when each force is turned round its point of application through a
right angle reduces to a couple H, prove that when each force is turned •
through an angle a the system is equivalent to a couple whose moment is
O cos a + H sin a. [S.]
Let any force P of the system act at the point (x, y) and make an
angle 8 with the axis Ox. The components of
P are P cos 9 and P sin 9 and the sum of theis 2/
moments about the origin O is
xP sin 9—yP cos 9.
Hence, since the system reduces to a couple
O, we must have
x
(? = S(a;Psin0-2/Pcos0). ^
In like manner, when each force is turned through a right angle, we
have rr
' = S {xP cos 9 + yP sin 9);
and when each force is turned through an angle a, the moment of the
couple is
S {x'P sin (0 + a) - yP cos {9 + a)}
= cos a 2 (xP sin 9 — yP cos 9) + sin a 2 {xP cos 9 + yP sin 9)
 EXAMPLES 63
(v) Prove that, if a variable system of forces in a plane have constant
moments about two fixed points in the plane, the resultant passes through
another fixed point and its minimum value is the difference of the moments
divided by the distance between the two points. [S.]
Let O, O' be the twofixedpoints about which the moments are given
constants G, G'. Take 00' as axis of x and a
perpendicular through O as the axis of y, and let
O' be the point (f, 0). Then as in 5*3, if X, Y are
components of the resultant force,
G'=G-£Y (1).
But G, G' and £ are all definite constants,
therefore Y is a constant. Hence X is the only
variable part of the system of forces. But the equation of the line
of action of the resultant is
G-xY + yX = 0, (5-3)
and this passes through the fixed point ( ^ , 0) whatever be the value
oiX.
The minimum value of the resultant is got by taking X = 0, the
resultant then being Y, and from (1) this is equal to (G — G')/£, or
(G-G')IOO', which proves the proposition.

EXAMPLES
1. A BCD is a square of side a whose diagonals intersect at O.
Forces 1, 2, 3, 4, Q^/2 act along AB, BC, CD, DA, AC. Find the
magnitude and direction of the single force at 0 and the magnitude of
the couple which together are equivalent to the given forces. Also find
at what distance from A the resultant of the five forces cuts the line
AB.
2. Forces 1, 3, 5, 7, 8-\/2 act along the sides AB, BC, CD, DA and
the diagonal BD of a square of side a. Taking AB and AD as axes of
x and y respectively, find the magnitudes of the resultant force and the
equation of its line of action.
3. ABCD is a square of side 8 in. Forces 3, 2, 5, 4 act along AB, BC,
DC, DA in the senses indicated by the order of the letters. Find the
resultant force and the point in which it cuts CD.
4. ABC is an equilateral triangle of side a and forces 4, 2, 2 act along
AB, AC, BC in the directions indicated by the letters. Find the re-
sultant force and the distances from A at which it cuts AB and AC.
5. A force lies in the plane of the triangle ABC and its moments
about A, B,C are L, M, N respectively. If the force is the resultant of
three forces acting along the sides of the triangle, determine the
magnitudes and senses of these forces.
64 COPLANAR FORCES [V
6. Forces P , 2P, 3P, 4P act along the sides AB, BC, CD, DA re-
spectively of the square ABCD. Find the magnitude of their resultant
and the points in which it cuts the sides AB and BC. [S.]
7. Three forces act along the sides of an equilateral triangle, and
vary so that their resultant is constant in magnitude and direction.
Shew that the difference of the magnitudes of any two of the three
forces is constant. [S.]
8. Forces P , 2P, 3P, 4P, 5P act respectively along the lines AB,
CA, FC, DF, ED, where ABCDEF is a regular hexagon. Find the
forces along the lines BC, FA, FE which will keep the system in
equilibrium. [S.]
9. Forces P 1 ( P 2 , P 3 , P 4 , P 5 , P e act along the sides of a regular
hexagon taken in order. Shew that they will be in equilibrium if
£P = 0 and P1-Pi = P!l-Ps = P5-P2. [S.]
10. A BC is an equilateral triangle and D, E, F are the middle points
of the sides BC, CA, AB. Forces P, 2P, 3P act along BC, CA, AB and
forces 4P, 5P, 6P along FE, ED, DF. Find the line of action of the
resultant. [I.]
11. Three forces P, Q, R act along the sides of the triangle formed
by the lines , , . o
•* x + y=l, y-x=l, y = 2.
Find the equation of the line of action of their resultant. [I.]
12. Forces ka, kb, kc act along the sides a, 6, c of a triangle, and a
fourth force P is given in magnitude. Prove that, if the line of action
of P touches a given circle of radius r, the line of action of the resultant
will touch a fixed circle; and find its radius. [S.]
13. Prove that a force acting in the plane of a triangle ABC can be
replaced uniquely by three forces along the sides of the triangle.
If each of a system of coplanar forces be replaced in this way by
forces of type p.BC, q.CA and'r.^4B, shew that the necessary and
sufficient conditions that the system reduces to a couple are that
Sp = Sg = Sr. [S.]
14. Forces P, Q, R act along the sides BC, CA, AB of a triangle.
Find the condition that their resultant should be parallel to BO and
determine its magnitude. [S.]
15. Three forces each equal to P act along the sides of a triangle
ABC in order; prove that the resultant is

P f l - 8 sin- sin- sin - X *

and find the distance of its line of action from one angular point, and
where it cuts one side of the triangle. [S.]
 EXAMPLES 65
16. Forces P , Q, R act along the lines x — 0, y — 0, and
xcoad+ysa\d=p.
Find the magnitude of the resultant and the equation of its line of
action. (The axes of co-ordinates are rectangular.) [S.]
17. Forces P , Q, R act respectively along the lines BC, CA, AB;
prove that, if 8 is the magnitude of their resultant,
82 = P 2 + <22 + iJ2 - 2QR cos A - 2RP cos B - 2PQ cos C,
and that the line of action of this resultant will pass through the ortho-
centre of the triangle ABC, if

18. Three forces P , Q, R act along the sides BC, CA, AB of a

triangle ABC. Their resultant lies in the line joining the centre of the
circle inscribed in A BC and the centre of gravity of the triangular
lamina ABC. Shew that
P:Q:R = {b-c)a:(c-a)b:{a-b)c. [S.]
19. Forces P, Q, R acting along the sides BC, CA, AB of a triangle
are in equilibrium with forces P ' , Q', R' acting along AO, BO, CO, the
lines joining the vertices to the centre of gravity of the triangle. Prove
that PP
' . QQ' i BB' -n rqi
AO.BC BO.CA + C0.AB~
+ L J

20. A system of forces X, Y, Z acting along the sides of the triangle

ABC is equivalent to a system P , Q, R along the sides of the pedal
triangle. Prove that 2P _ F / c o s B + Z/cos G. [S.]
21. Three forces act at the angular points of a triangle ABC, each
parallel to the opposite side. If they have a resultant acting along the
line joining the in-centre and- orthocentre of the triangle, prove that
they are proportional to
ab (a + b) cos B — ac (a + c) cos C
and the two similar expressions. [S.]
22. Forces P , Q, R, 8 acting in the sides AB, BC, CD, DA of a
quadrilateral A BCD are in equilibrium: shew that
PxR _ Qx8
AB.CD~BC.DA'
Will this condition alone ensure equilibrium? [S.]
23. Four forces in equilibrium act along the sides of a quadrilateral
inscribed in a circle. Prove that each force is proportional to the sine
of the angle in the segment of the circle cut off by the side opposite to
its line of action. [S.]
24. Four forces act along the sides of a quadrilateral A BCD and
are represented in direction and magnitude by BA, BC, AD, CD.
Prove that their resultant is parallel to one diagonal and bisects the
other. [I.]
RS 5
66 COPLANAR FORCES [V
25. ABCD is a quadrilateral in which AD, BC are parallel. Prove
that forces represented by AB, BC, CD, DA, AC, DB have a resultant
of magnitude 2EF, where E, F are the middle points of AD, BC, and
find its position. [T.]
26. A lamina is acted on by four forces represented in magnitude,
direction and position by the sides AB, BC, CD, DA of a crossed
quadrilateral, the lines AB, CD crossing one another. If the lamina is
pivoted at 0, the point of intersection of AC and DB, and can turn in
its plane, prove that it can be kept in equilibrium by a force in BC of
., , AC.BD „„ rT,
magnitude BQ g Q BC. [I.]
27. Four forces act in the sides AB, BC, CD, DA of a plane quadri-
lateral and are in equilibrium; shew that
XAB/AB ADAB AABC=XCBIBC AABC ABCD
= XCD/CD ABCD A CDA = XADIAD A CDA ADAB. [C]
28. Prove that in all cases a system of coplanar forces can be re-
placed by two forces, one of which acts through a given point and the
other along a given straight line. [S.]
29. A system of coplanar forces is such that the sum of their mo-
ments about any point in a given straight line is always the same. State
what inferences may be drawn about the resultant of the system. [S.]

ANSWERS
1. 4V2 along AC, 5a; 5a/4. 2. 4V10;
3. V68; 20 in. fromZ). 4. 2^/1, \a, a.
5. £a/2A, Mb/2A, JVc/2A. 6. 2i/2P; \AB from B.

10. Parallel to CB dividing DA in the ratio 1:5.

11. P(x + y-l) + Q(y-x-l)-BV2(y-2) = 0.
12. r + 2/fcA/P. 14. Qc = Rb; P-Qa/b.
15. Distance from A = 2PA/aR, where resultant R divides BC ex-
ternally in the ratio c: b.
16. ^/(P*+Q* + Rt-2QRsin0+2RPcosd);
Px— Qy + R (x cos 0 + y sin 9-p) = 0.
25. R cuts DA produced in L so that AL = FB and is parallel to EF.
29. R is either zero or parallel to the given line.
 Chapter VI
T H E S O L U T I O N OF P R O B L E M S
6*1. We have found that conditions sufficient to ensure the
equilibrium of a rigid body under the action of coplanar forces
are three in number, and may be summarized as—
two resolutions and one equation of moments,
or one resolution and two equations of moments,
or three equations of moments.
The student will do well to bear in mind that, in the solution
of a problem, no additional information can be obtained by
writing down more than three such equations of moments or
resolutions for one and the same body.
It follows that many statical problems are indeterminate or
insolvable without further hypotheses as to the elastic pro-
perties of the body acted upon. For example, if a beam or bar
under the action of given forces has its ends fixed, it is not
possible to determine the magnitude and direction of the
reactions at the fixed ends, because this means four unknown
quantities, two magnitudes and two directions, and the con-
ditions of equilibrium give only three equations between the
four unknown quantities.
6*2. Constraints and Degrees of Freedom. The position of a
plane body in its plane is determined by three quantities, such as the
rectangular co-ordinates of one specified point of the body and the
angle which a linefixedin the body makes with a linefixedin the plane.
Hence a body free to move in a plane has three degrees offreedom.We
may also express this fact by saying that it is free to have a motion of
translation with components in either or both of two perpendicular
directions and also free to rotate.
The number of degrees of freedom may be reduced by the imposition
of constraints: thus for a body with two points fixed no motion is
possible, i.e. it has no degrees of freedom; but with one point fixed,
rotation is possible, i.e. the body has one degree of freedom; and if a
point of the body is constrained to lie on a given line, then the point
may move along the line and the body turn about the point so that it
has two degrees of freedom.
5-2
68 THE SOLUTION OF PROBLEMS [VI
In general the loss of a degree of freedom means the imposition of a
constraining force. There are three equations of equilibrium and these
constitute relations which should enable us to determine the three co-
ordinates which specify the position of an unconstrained body in
equilibrium under the action of given forces. The three equations
should also enable us to determine the smaller number of co-ordinates
of a body deprived of one or two degrees of freedom together with
some of the elements of the constraint.
Thus the equations determine in the case of
complete freedom three co-ordinates of position,
two degrees of freedom two co-ordinates of position and one
element of constraint,
one degree of freedom one co-ordinate of position and two
elements of constraint,
no freedom three elements of constraint.
But as stated in 6*1 the fixing of two points in a body under the
action of given forces imply four unknown elements, viz. the magni-
tudes and directions of two unknown constraining forces, so such
problems are indeterminate unless some fact about one of the con-
straining forces is prescribed.
6#21. There is a large class of problems in which a body is
in equilibrium under the action of three forces. We can shew
that in such circumstances the forces must be coplanar and either
concurrent or parallel.
We assume that since the forces are in equilibrium they can
have no tendency to turn the body on which they act about
any straight line, or that the algebraical sum of their moments
about any straight line is zero. A formal proof of this statement
will be found in 14*2.
Let P, Q, R be the forces. Take any points A, Bon the lines
of action of P and Q. Then the three forces
being in equilibrium the sum of their mo-
ments about AB is zero; but P and Q inter-
sect AB and have no moment about it,
therefore the moment of R about AB is zero.
It follows that R is either parallel to AB or
intersects it in a point C. We may reject the
former alternative since the points A and B
are chosen at random on the lines of action
of P and Q. Hence AB intersects R in C. Similarly if D be
any other point on the line of action of Q, AD must intersect
6-2-6-23] EXAMPLES 69
R in a point E. And for a like reason CD must intersect the
force P in a point F. The three forces P, Q, R have therefore
been proved to lie in the plane of the triangle ADC.
Again if no two of the forces are parallel, two of them, say
P and Q, must meet in a point 0, and since R is equal and
opposite to the resultant of P and Q, therefore R also passes
through 0 and the forces are concurrent. But if two of the
forces are parallel, then the third force which balances the
resultant of the first two is also parallel to them.

6*22. When the number of forces acting on a body in equili-

brium is three, or can be reduced to three, the position of equi-
librium can be determined by making use of the fact that these
forces must be concurrent or parallel, the expression of this
fact geometrically involving relations between the parts of the
figure which serve to determine the position of equilibrium.
This is called the geometrical method of solution. In the
alternative analytical method we write down the three
equations between the forces obtained from the conditions of
equilibrium, i.e. by resolving and taking moments, and use
these equations to obtain the required solution.
We will exemplify these methods by the solution of some
problems.

6*23. Examples, "(i) A triangular lamina A BCis suspendedfrom a

point O by light strings fastened to the points A, B and hangs so that the
side BG is vertical. Prove that, if a, /? are the angles which the strings AO,
BO make with the vertical, then
2 cot a - cot j8 = 3 cot B.
The three forces acting on the lamina are the
tensions of the strings OA, OB arid the weight. The
weight must therefore act through O; but it is a
vertical force, therefore OO is vertical, where 0 is the
centre of gravity (intersection of medians) of the
triangle.
Now, if AD is the median through A,
A

and 00 and BC are parallel, so that by projecting

on the horizontal
AO since: BO Binp = AG:GD = 2:1 (1).
70 THE SOLUTION OF PROBLEMS [VI
Also by projecting on the horizontal and vertical

and ^Ocosa — BO cos £ = 4 B cos B (3).

Then, by substituting from (1) in (2) and (3), we get
3B0sinfi=4BsinB,
BO
and -j—(2cosasin 8 —sinacos 6)=-4BcosB;
therefore, by division 2 c o t a - c o t S = 3cotB ...(4).
Alternatively, if we approach the problem analytically, we call the
tensions T, T' and the weight W.
By resolving horizontally and vertically, we get
T sin a = 2 " sin 0 (5),
and 2 7 cosa+2 v cosfi=TT (6).
We can also obtain information by taking moments about a point
in the plane, and since T and T' pass through O we choose O for this
purpose as leading to the simplest result. The result is that the moment
of W about O is zero so that W must act through 0. The reader should
note that we cannot, by taking moments about any other point or
resolving in any other direction, obtain any additional information
than is contained in (5), (6) and the fact that W acts through O. Also
that (5) and (6) suffice only to express T and T' in terms of W, a. and /?,
and are equivalent to an application of Lami's Theorem

and that for the geometrical fact expressed by (4) we are compelled to
resort to the geometrical solution given above.
(ii) A uniform rod has a ring at one end which slides along a smooth
vertical wire; the rod rests touching a smooth cylinder of radius r whose
axis is horizontal and at a distance cfrom the wire. If 9 is the angle the
rod makes with the horizontal, find its length in terms of c, r, and 6. [S.]
Let AB be the rod and O its middle point,
XY the vertical wire and A the ring. It is
assumed that the vertical plane through the
rod cuts the cylinder in a circle of centre C
and radius r. The rod touches this circle at
H. The rod is in equilibrium under the action
of three forces, namely, its weight acting
vertically downwards through G, the re-
action of the smooth cylinder along the
radius through H, and the reaction of the
smooth wire at right angles to the wire, i.e.
horizontally through A, and these three
forces must meet in a point 0.
Then, if CM is perpendicular to XY and 21 denotes the length of the
rod, we have C=OM=AO + COsm6=lcos6+ COsinfl,
and Zsin0=O<?=O.ffsec0=(»--CO)sec0;
6-23] EXAMPLES 71
therefore, by eliminating GO
lsin6eos0 = r-CO
= r — (c—l cos 9) cosec 9,
or Icoa36=c — rsinfl.
The reactions at A and H can then be expressed in terms of the
weight by using Lami's Theorem.
We leave it to the reader to shew that there is another solution
Icos3 9 = c+rsin9,
in which the horizontal reaction of the vertical rod on the ring is
reversed in direction.
(iii) ABCD is a uniform lamina, in shape a rhombus with sides of
length a and the angle A = 2a.. P and Q are smooth pegs, PQ being of
length I and horizontal. Find the angle which AC makes with the vertical
if the lamina can rest with points on the sides AB, AD in contact with the
pegs and with AC not vertical. Shew that such a position of equilibrium
occurs only if aoo*adn.«<l<a<xxfia. [S.]
The reactions of the smooth pegs at P and Q are at right angles to
AP and BQ and meet in O, and
since the only other force acting
on the lamina is its weight, the
vertical OL through the centre Q
of the rhombus must pass through
O. Let AQ make an angle 9 with
the vertical in equilibrium; then
DB makes an angle 9 with the
horizontal, and the angle

and AQP=ABJD-6,
therefore
APQ-AQP=29.
Since APO and AQO are right
angles, therefore AO is the dia-
meter of a circle APOQ, and
6 = AOsinA0L=A0sin(A0Q-QOL)
= AOsin(APQ-AQP)
= A0 sin 29.
Hence, when 9 is not zero,
AG=2A0 cos 9.
But AO = a cos a, and, since AO is the diameter of a circle in which
a chord PQ (= 1) subtends an angle 2a at the circumference, therefore
AO = 1 cosec 2a.
Hence a cos a = 21 cosec 2a cos 9,
or Z cos 0 = a cos2 a sin a.
72 THE SOLUTION OS PROBLEMS [VI
It follows at once that there is such an angle 8, if and only if
a cos2 a sin a <l.
Again, if AL is a horizontal line through A, then GAL is the com-
plement of 9, and OAQ is a, and AQ must be above AL so that

or sin a < cos 0;

therefore 1 sin a < I cos 9,
i.e. I sin a < a cos2 a sin a,
so that £<acos 2 a,
i.e. another necessary condition.
6*3. When a problem relates to the equilibrium of two or
more bodies in contact under the action of given external
forces, we assume the existence of mutual actions and reac-
tions at the points of contact of the bodies, and we also assume
that when the bodies are smooth these forces act at right angles
to the common tangent planes at the points of contact. We
may then write down the equations of equilibrium for each
body separately, including in such equations the given external
forces which act.upon the body and the unknown reactions of
the other bodies upon it. By eliminating the unknown reac-
tions we then get a smaller number of equations connecting
the given external forces and the co-ordinates of position and
so arrive at a solution as in a problem of a single body. In some
cases it is useful to write down the equations of equilibrium
for two bodies regarded as though they were a single body and
thus avoid the introduction of the unknown reaction between
them. Sometimes both methods are combined in the solution
of the same problem. It is essential, however, to remember that
the force between two bodies is always a mutual action and
reaction, and to keep clearly in mind in writing down an
equation to what body or bodies it refers so that all the forces
which act on the body or bodies are included and none others.
The examples of the next article will serve to illustrate the
methods described above.
6*31. Examples, (i) Two smooth spheres of equal weight rest inside a
fixed smoothrightcircular cone which has its vertex downwards and its axis
inclined at an angle j8 to the vertical, each sphere touching the cone at one
6-23-6-31] EXAMPLES 73
point only. Prove that the common normal to the spheres makes an angle
_, f sin 2/?
cot f
(cos 2 a + 008 2,8/
with the vertical, where a is the semi-vertical angle of the cone. [S.]
A, B are the centres of the spheres and G, D their points of contact
with the sides of the cone. They need not have the same radius, but
they have the same weight, so that, if we regard the two as one body,
their weight acts in the vertical through O the middle point of AB.
There is a mutual action and reaction between the spheres along the
line of centres AB, but, if we regard the two as forming one body, the
only forces in addition to the weight are the pressures of the sides of the
cone along the radii of the spheres at O and D.

Since we are now considering the equilibrium of a body under the

action of three forces these forces must be coplanar and concurrent, so
that CA and DB meet in a point 0, and 0 must lie on the vertical
through G.
We are now in a position to obtain a geometrical solution as follows.
Let GO, CO and DO meet the axis of the cone in L, M and N. Then the
a^S16 OLM = fi and OMN=ONM = £77-a;
therefore AOG^LOM — ^TT — X — ^,
and BOG—OLN+ONL = ^ + \TT — CL.
Let OGB =6, then
sin A OG_AG_BG sinGOB
l
sinOAG~ ~0G~~ sinOBG'
cos (a + ft) cos (a— j8)
•*-a)'
- cos (a + j8+ 6) _ cos (a - ft - 6)
or cos (a + ft) ~ cos (a —ft) *
74 THE SOLUTION OP PROBLEMS [VI
Therefore
- cos 8 + sin 8 tan (a +ft)= cos 0 + sin 8 tan (a — ft),
2 cot 0=tan (a + ft)-tan (a-ft)

"cos(a + ft)cos(a-ft)'
whence fsin2ft_l
[cos 2a + cos 2ft/'
(ii) Inside a fixed hollow cylinder of radius R, whose generators are
horizontal, there are placed symmetrically two equal cylinders, each of
radius r: a third cylinder equal to each of the latter is placed symmetrically
on them; shew that equilibrium cannot exist unless R<r(l + 2Vi). [S.]
Taking a vertical section through the centres of the axes of the
cylinders, let O, A, B, C be the centres and let W denote the weight of
each of the smaller cylinders. The figure is symmetrical about the
vertical 00. Let X denote the mutual action and reaction between the
two lower cylinders, T the mutual reactions in the lines CA and OB,
and S the pressures of the fixed cylinder on each of the two lower ones
acting along AO and BO.
O

Let the angle AOB = 28. The angle AOB is 60°.

By considering the three smaller cylinders as one body and resolving
vertically, the weight 3W is balanced by the vertical components of
the forces S, therefore
2Soos8=BW (1).
Then, resolving vertically for the upper cylinder alone, we get
221cos30°=TF, or V3T=W (2);
and, resolving horizontally for the cylinder centre A, we get
X + T cos 60° = £81110 (3).
Now equilibrium could not exist if X were negative, because X re-
presents the pressure between the two lower cylinders; hence we must
have
T cos 60° =s:,S sin0.
6-31-6-41] REACTIONS AT JOINTS 75
On substituting for S and T in terms of W from (1) and (2), we get
W 3

But

so the required condition becomes

or
or
6*4. Reactions at Joints. There are a large number of
problems in which two bodies are described as 'smoothly
hinged' at a point. In such a case the hinge may be regarded as
a pin passing through cylindrical holes in the bodies, closely
fitting and so smooth that each body can turn about the pin
without friction. When the hinge or joint is smooth the reaction
of the pin on either body reduces to a single force, because, no
matter how many points of contact there may be between the
pin and the cylindrical hole in the body, the reaction at each
of these points acts along the common normal and therefore
passes through the centre of the pin (considering only forces in
one plane) and all such forces can be combined into a single
force through the centre of the pin. When the pin connects two
bodies A and B only, then the pin is subject to two forces only,
namely the reactions of A and B upon it, and in equilibrium
these must be equal and opposite. But the reactions of the
pin on the bodies are equal and opposite to the former forces,
so that the result of the smooth joint is to set up equal and
opposite forces on the bodies A and B and it is unnecessary to
consider the precise form of the joint, because it is sufficient to
know that, as the result of the smooth joint, there is a pair of
equal and opposite forces between the bodies at a certain
point and that the bodies are so constrained that the only
possible relative motion is one of turning about this point.

6*41. Example. Two uniform rods AB, BC of equal weight are

smoothly hinged at B. The end A can turn about a fixed point and BG rests
76 THE SOLUTION OF PROBLEMS [VI
across a smooth horizontal peg. If in equilibrium both rods make angles of
60° with the vertical, prove that the reaction at B divides the angle ABC
into angles whose tangents are as — 1:14. [S.]
The rod A B is acted upon by three forces, its weight W acting through
its middle point O, the mutual reaction R at the hinge B, and a con-
straining force at A; and these three forces being in equilibrium must
meet in a point D.
The rod BO is also acted upon by three forces, its weight W acting
through its middle point H, the mutual reaction R at the hinge B, and
the pressure of the smooth peg P acting at right angles to BG. These
three forces must also meet in a point E.
The reader is advised to draw figures with the rods or reactions in
other positions in order to convince himself that the position shewn in
the figure is the only one in which the directions of the two sets of
three forces acting at a point can possibly result in equilibrium, and in
particular that it is not possible for B to be lower-than A and C.

Let 0 denote the angle ABB.

Then at D we have ^ = s , by Land's Theorem, and at E we
, W
W sinBEP
have -^ = for the same reason; therefore
sin BEP _ sin ADB _ ADsvuADB
sin PEH ~ sin ADO ~ AD sin ADG
_ABainABD
~AGsinAGD'
cos(6O° + 0)_ 2sin0
or sin 30° ~ sin 60°'
Therefore V3 cos (60° + 8) = 2 sin 0,
so that tan0=V3/7.
taxi ABD _ tanfl _ tang(l--y/3tan0)
Then
tanDBC7~tan(12Oo-0)~ ^/3

=
~l4"
6-41-6-43] REACTIONS AT JOINTS 77
6*42. Working Rules. When more than two bodies are con-
nected at the same joint, then it is necessary to exercise care in
regard to the reactions because what is described as 'the
reaction at the joint' usually depends on the way in which the
attachments are made.
Before considering further problems we will state some
working rules which serve to simplify the process in many cases.
(i) When a framework and the forces acting upon it are both
symmetrical about a line passing through a joint, then the reaction
at that joint is perpendicular to the line of symmetry.
This follows because no reason could be assigned for sup-
posing the reaction to act on one side of the perpendicular
which would not by symmetry also be a reason for supposing
it to act on the other side of the perpendicular.
(ii) When a body is in equilibrium under the action of forces
acting upon it at two points A, B, only, then the forces at these
points must act in the line AB and be equal and opposite.
For otherwise the forces could not maintain equilibrium.
This applies to all cases of rods which form parts of jointed
frameworks and are acted upon by no forces save the reactions
at the joints; the reactions in any such case must be along the
rod.
In many problems a 'light' rod or a string is used to
connect two points in order to limit the freedom in a frame-
work. A 'light' rod means one whose weight is negligible.
Since the only possible effect of such a rod or string is to
produce forces along its length at the points which it connects,
we may frequently effect a simplification in the consideration
of the nature of a joint by substituting for the rod or string two
equal and opposite forces acting on the framework at the ends
of the rod or string in its direction. We may treat such forces
as external forces applied to the framework without affecting
the solution of the problem.
We shall make some use of this artifice in the problems that
follow.
6'43. Examples, (i) Four rods are smoothly jointed at their ex-
tremities to form a quadrilateral ABCD, and the opposite corners A, C
78 THE SOLUTION OF PROBLEMS LVI
and B, D are joined by tight strings. Prove that, if T, T' denote the
tensions in AC and BD, then

0 OCJ \B0 ODJ'

where 0 is the intersection of the diagonals.
In obtaining the required result we can avoid all difficulties about the
joints by regarding the strings as fastened to the rods AD and BC, and
regarding the rods AB and CD as replaced by the reactions Rlt R3
which they produce. The rod AD is then in equilibrium under the
action of four forces T and R1, at A, and T' and R3 at D.
By taking moments about D, we get
T. AD sin CAD = R1.ADsin BAD.
Similarly by taking moments about C for the rod BC, we get
T'. BCsin CBD = Rx. BCsiaABC.

A B A R , R,

Therefore T sin A BC sin CAD = T' sin BAD sin CBD.

Using AABC to denote the area ABC, this gives
T.AABC.ACAD_T'.ABAD.ACBD
AB.BC.AC.AD ~ AD.AB.BC.BD '

BO.OD'
1 ! 1 I
i.e.
V BO + OD
It will be noticed that in this solution we have avoided altogether
the question of the nature of the joints and the
reactions at the joints. A simple method of repre-
sentation would be to suppose the rods connected
by pin joints and the strings attached to the pins.
Each rod can then be supposed to end in a small
ring which passes over the pin, and the forces act
only along the rods and along the strings.
The figure shews these three forces at the corner
B, and their ratios are given by Land's Theorem.
6-43-6-45] REACTIONS AT JOINTS 79
This affords an alternative method of obtaining the relations between
T, T' and iJ t , used in the foregoing solution.
(ii) ABCD is a rhombus formed by four light rods smoothly jointed at
their ends and PQisa light rod smoothly jointed at one end to a point P in
BG and at the other end to a point. Q in AD. Two forces each equal to F are
applied at A and C in opposite directions along AC. Prove that the stress
inPQis F.AB.PQ/AC(AQ~BP). [S.]
If the j oints are regarded as pins fixed
to the rods AD and BG passing through
holes in the rods AB and CD, while the
forces F are applied to the pins all the
conditions are fulfilled and the part
played by the rods AB and CD is that
of exerting equal and opposite forces
at their ends in the direction of their
lengths.
Let T denote the stress in PQ. Con-
sider the equilibrium of the rod AD and
resolve at right angles to AB; then
T sin e-FsinCAB,
where 8 is the inclination of PQ to AB.
Then, if PM, QN are perpendicular to AB, we have
MP~NQ_BP~AQ
sin 6 — siaABC.
PQ ~ PQ
Therefore T= F.PQ sw.GAB_ F.PQ.AB
BP-AQ'sinABC AG(BP-AQ)'
6*44. In order to simplify the solutions of problems it is
important to make use of symmetry wherever it exists and to
introduce no more unknown reactions than is necessary.
The following problem could be solved by first inserting in
the figure unknown horizontal and vertical reactions at every
one of the six joints, but it will be noticed that in the solution
given no unknown quantities are introduced save the two
that are to be compared.
6*45. Examples, (i) A hexagon ABCDEF is formed of six equal
rods of the same weight W smoothly jointed at their extremities. It is sus-
pended from the point A and the regular form is maintained by light rods
BF and GE. Prove that the thrust in the former is five times that in the
latter.
Suppose that the rod BF is attached to the two upper rods and the
rod GE to the two lower rods. Let R and S denote the thrusts in BF and
CE. Then since the only effect of these rods is to produce thrusts at
 80 THE SOLUTION OF PROBLEMS [VI
their ends, we may ignore these rods if instead of them we suppose
horizontal forces R to act outwards on AB at B and on AF at F, and
horizontal forces 8 to act outwards on CD at C and on DE at E. Begin
by inserting these forces in the figure.
Then consider the equilibrium of the rod CD. The reaction at D is
horizontal because there is symmetry about the vertical through D.
But the only horizontal forces on CD are the force S at C and the
reaction at D, so that this reaction at D must be equal and opposite to S.
Then as regards vertical forces: the weight W acts vertically down-
wards through the middle point of CD and the only other vertical
force can be at C, therefore there is a reaction at C which acts vertically
upwards and is equal to W. Insert this in the figure; and, since it is
produced by the rod GB, also insert an equal and opposite force W
downwards acting at C on CB.

We can now express S in terms of W by taking moments for the rod

CD about C or about D, or, what is the same thing, equating the
moments of the two couples that act upon the rod. We find that
S. CD sin 30° = W. £ CD cos 30°,
or S=W-s/3/2.
Then, returning to the figure, consider the rod BC. It is in equili-
brium under the action of its weight W, a downward force W at C and
the reaction at B. This latter force must therefore act vertically up-
wards and be equal to 2W. Insert this force in the figure and also the
equal and opposite reaction 2W at B on AB.
Then consider the rod AB. It is in equilibrium under the action of its
weight W, the horizontal and vertical forces R and 2W at B and the
reaction at A. It is not necessary to specify the latter because we can
take moments about A; by so doing we find that

or It
If the reactions at A were required, their precise form would depend
upon the method of support, but so far as ourfiguretakes us, it is clear
6-45-6-5] CHAIN OF PARTICLES 81
that AB must be acted upon, at A by a horizontal force R towards AF,
and by a vertical force 3W; also that the total supporting force at A is
equal to the total weight 6 IF.
(ii) A regular pentagon ABODE formed of five uniform rods, each of
weight W, freely hinged to each other at their ends is placed in a vertical
plane with CD resting on a horizontal plane and the regular pentagonal
form is maintained by means of a string joining the middle points of the
rodsBC and DE. Prove that the tension in the string is

foot J+3 cot ^ W . [S.]

V 5 5/
It is only necessary to consider the reactions at the corners A and B.
By symmetry that at A is horizontal
and equal say to X. The rod AB is
also acted on by its weight W and
the reaction at B. The latter must
therefore have a horizontal com-
ponent X and a vertical component
W upwards. Insert in the diagram
forces at B acting upon BC in the
opposite senses. Then by taking
moments about B for the rod AB,
since the rod AB makes an angle \n
with the horizontal, we get
Xsin|=£IFcos| (1),

after dividing by the length of the rod.

Again if T denotes the tension in the string which joins the middle
points of BC and DE, by taking moments about C for the rod BC,
which makes an angle £n with the horizontal, we get

^ ^ ^ (2),
after dividing by the length of the rod.
On substituting for X in terms of W from (1), we find that

T= IF (cot |+ 3 cot ^

6*5. Chain of Heavy Particles. Let a number of heavy particles

of weights wltwitw3, ...he fastened to a light string the ends of which
are held fixed. Let the successive portions of the string be inclined to
the vertical at angles 6t, 82, 6a,... and let the tensions in these portions
be denoted by 2 \ , Ta, Ta, ... as indicated in fig. (i).
By considering the equilibrium of the successive particles and
resolving horizontally, we get
Tx sin 8X = T% sin 02 = Tt sin B3 =...,
82 THE SOLUTION OF PROBLEMS LVI
or the horizontal component of tension is constant and
= Tsay .(1).
Then by resolving vertically for each particle we get
T 2 c o s 02 ~ ^ 1 o o s $ i = wi >
T3 cos 03 - Tj cos 02 = w%,
Tt cos 04 - ^ cos 03 = w3,
etc.,

and on dividing these equations by the horizontal tension they become

cot 0a — cot 6X = -jk,

cot 0 S -cot 02 = -^,

etc.,
so that if the weights are equal the cotangents of the inclinations to the
vertical of the successive portions of string increase in arithmetical
progression.
6*51. We can also treat the problem graphically by constructing
triangles of forces for the three forces acting on each particle. Thus in
fig. (ii) OBA is a triangle of forces for the particle wx, in which BA
represents wltAO represents 2\ and OB represents Ta; then OCB is a
triangle of forces for the particle wa, in which CB represents w%, BO
represents T2 and OC represents Ta; similarly ODC is a triangle of
forces for the particle w3, in which DC represents ws, CO represents
T3 and OD represents T 4 . It is seen at once that the points A,B,C,D
lie in a vertical line and that the perpendicular ON on this line repre-
sents the horizontal component of the tension.
When the weights are equal we have AB — BC=GD=... and the
cotangent relation is obvious.
6-5-6-52] CHAIN OF RODS 83
6*52. Chain of Heavy Rods. Let a series of uniform rods be
smoothly jointed at their extremities so as to form a continuous chain
hanging in a vertical plane with its ends fixed. Let AB, BO, CD,... re-
present consecutive rods of weights Wt, Wt, W3, The form of the
chain would not be altered by supposing half of the weight of each rod to
act at each of its ends. The weights would then act at the points A, B,O,...
and the rods would play the part of the portions of string connecting
the heavy particles in 6*5 and the form would be determined by the
method of 6*5 or 6"51 with the difference that instead of the weight of a
particle we have the mean of the weights of the two rods that meet in
the corresponding point.

If, however, we want to determine the reactions at the j oints we must

replace the weights at the centres of the rods; then we may proceed
as follows.
Let X, Y be the horizontal and vertical components of the reaction
between the rods of weights Wx and W2, and let d1, 02 be the inclinations
of these rods to the vertical. By taking moments for the lower rod about
its lower end and for the upper rod about its upper end, we get
Y sin 6X - iWx sin Qx = X cos 91,
and Fsin0 2 + 4IF2sin02 = .X'eos02,
where the equations of moments have been divided by the lengths of
the rods. Solving for X and Y, we get
X(oot6i-cote1) = i(W1 + Wi) (1),
and r(tan0 1 -tan0 2 ) = j(l^ 1 tan0 1 + JF2tan02) (2).
It is evident that the horizontal component of reaction X is the same
at every joint, and the result (1) is precisely the cotangent equation of
6'5 if the weight at the j oint be assumed to be J (W± + W%) and X denotes
the horizontal tension.
Again the X and Y at each joint compound into a single force R,
and each rod is in equilibrium under the action of its weight and the
reactions at its ends. The reactions at the ends of the rods must there-
6-2
84 THE SOLUTION OF PROBLEMS [VI
fore intersect in pairs on the verticals through the middle points of the
rods. The lines of the reactions there-
fore form the sides of another polygon
whose vertices are on the lines of
action of the weights of the rods.
Thus if AB, BC, CD, ... are con-
secutive rods, abed... is the sub-
sidiary polygon along the sides of
which act the reactions at the joints
A, B, C, ...; and these forces along
the sides of the polygon abed..., if
regarded as acting at the corners
6, c, d,..., i.e. at points on the lines of
action of the weights Wt, W2, Ws,...,
play the same part as the tensions of
the strings in 6'5.

6'53. E x a m p l e . A chain of four equal heavy rods, each of weight w,

is held up at points A, B in the same horizontal line. If the points of
suspension are drawn apart until the horizontal component of the pulls at
A, B are each 2w>, shew that ^ . 5 = 3-54 times the length of a rod and
determine the slope of each rod. [T.]
Since we are not concerned with the reactions at the joints but only
with the form of the chain, we can suppose the weight of each rod to be
equally divided between its ends.

1w

w w

The vertical supporting forces at A and B must be equal by sym-

metry, so that each is half the total weight, i.e. 2w. The distribution of
the weights at the ends of the rods is as shewn in the figure. These
separate weights are now regarded as supported by tensions in the rods,
and the horizontal component of tension is the same throughout and
equal to 2w, as we find by resolving horizontally at A or B.
Now construct a triangle of forces for the forces at the lowest joint
D. Ocd is such a triangle in which cd represents w and Of the perpen-
dicular from O to cd represents the horizontal component of tension 2w.
By symmetry Oc and Od parallel to DE and CD respectively are
equally inclined to the horizontal so that cf=fd = \w. If we then
draw a vertical be equal to w, Obc will be a triangle of forces for the
forces acting at E. The inclinations oiDE and EB to the horizontal are
6-52, 6-53] EXAMPLES 85
1 1
the angles cOf and b Of or tan" £ and tan- f respectively, i.e. 14° 3'
and 36° 54'.
Then AB being the projection of the four
rods on the horizontal, we have
AB
length of a rod = 2 cos 14° 3' + 2 cos 36° 54'
= 3-64.
Alternatively we may solve the problem
directly. Let a, fl denote the inclinations of
the lower and upper pairs of rods to the
horizontal. By symmetry the reaction at D
is horizontal, and by resolving horizontally
for each rod we see that this reaction must
be equal to the horizontal reaction at A, i.e.
2w. Insert equal and opposite horizontal
forces 2w at the ends of each rod. Then by resolving vertically for the
rod CD, we see that the vertical reaction at C is w upwards on CD and
downwards on CA.
Now take moments about C for the rod CD and we get
2w.CDsa\ a, =
or tan a = £.
And, by taking moments
about A for the rod AC,
2w. AC sin p = w. AC cos p
+ w.$AC cos p,
or 1w
The remainder of the solu- V)
tion follows as before.
By comparing the two solutions it is evident that though the artifice
of replacing the weight of a rod by equal weights at its ends may seem
a simplification, it does not in point of fact reduce the amount of work
necessary to the solution of such a problem as this one.

EXAMPLES
1. A sphere, of weight W, is in equilibrium on a smooth plane of in-
clination a to the horizontal, being supported by a string, which is of
length equal to the radius and is fastened to two points, one on the
sphere and one on the plane. Prove that the tension of the string is
}V3Vsina. [I.]
2. A heavy uniform rod AB rests with one end A in contact with a
smooth vertical wall and with a string A CB attached to its ends and
passing through a ring G vertically above A. Prove that
[S.]
86 THE SOLUTION OF PROBLEMS [VI
3. A sphere, of weight W and radius r, is hung by a string of length
I from a point, and a uniform rod, of weight W and length 2a, has one
end attached to the same point and can freely turn about it. If the rod
rests touching the sphere, shew that the angle 8 made by the string with
the vertical is given by the equation
Wa cos2 a
W r + Wa sin a cos a
where cos a = r/( I + r).
Find the tension of the string in terms of W, W, r, a, and a. [S.]
4. A rod is in equilibrium resting over the rim of a smooth hemi-
spherical bowl,fixedwith rim horizontal, one end of the rod resting on
the curved surface. Shew that the inclination, 8, of the rod to the
horizontal is given by 4r cos 28 = 1 cos 8, where I is the length of the rod,
and r the radius of the bowl. Shew also that for the rod to rest in this
manner I must be greater than —K—T. [I.]
o
5. A triangular lamina ABC, right-angled at A, rests with its plane
vertical, and with the sides AB, AC supported by smooth fixed pegs
D, E in a horizontal line. Prove that the inclination 8 of AG to the
horizontal is given by
ACcos8-ABsin8 = 3DEcos28. [S.]
6. Two equal rods A B, A G hinged at A rest with their ends B, G on
a smooth table and their plane vertical. A string is attached at points
L, M of AB, AG respectively. Shew that the tension of the string is
W (BC)/4:p, wherp j> is the perpendicular from A on LM and W is the
weight of each rod. [S.]

7. Twoequalrods,^4.B,4(7freelyjointedat 4 , have their ends B, G

connected by a light string. B and G rest symmetrically on two smooth
planes inclined at angles a to the horizon, with their slopes facing one
another, the plane of the rods being vertical. Shew that the tension of
the string is \~W (tan 8 — 2 tan a), W being the weight of a rod and 28
the angle BAG. [S.]

8. Two uniform equal rods are hinged together and placed in a

vertical plane, with their extremities on a smooth horizontal plane.
They are kept from sliding by a horizontal rod the ends of which slide
without friction on the other rods. Find the position of equilibrium
and shew that the horizontal rod will rest close to the ground if its
weight = W tan2 8, where W is the weight of either of the other rods, and
8 their inclination to the vertical. [S.]

9. Two uniform rods, each of weight W and length a, are freely

jointed at A, and each passes over a smooth peg at the same level.
 EXAMPLES 87
From A a weight W is suspended. Shew that in the position of equili-
brium the inclination 6 of the rods to the horizon is given by
cos3 6 = c{2W+W')l2 Wa,
o being the distance between the pegs. [S.]
10. A circular disc of radius r and weight w is placed inside a smooth
sphere of radius R(>r), and at a distance d from the centre of the disc
is fixed a weight W. Shew that in the position of equilibrium the plane
of the disc makes with the horizon an angle whose tangent is
Wdj(W+w)VB2-r2. [S.]
11. Three equal smooth pencils of weight W are tied together by a
string and laid on a smooth table. Find the tension of the string when
there is a pressure P between the pencils in contact with the table.
[S.]
12. Two spheres each of radius a and weight W lie in contact in a
spherical bowl of radius na: shew that in the absence of friction the
pressure between them is IF/A/JI2 — 2n. [S.]
13. A square lamina rests in a vertical plane on two smooth pegs at
a distance c apart in the same horizontal plane. If 8 be the angle the
diagonal through the lowest corner makes with the vertical, x and y
the distances of the centre of gravity from the sides which touch the
pegs, prove that
(x + y)sin6=(x~y)cos6+V2csin28. [S.]
14. Three smooth heavy cylinders A, B, O lie on a table, with B
between A and C and touching each of them. A and C have equal radii
a and B has weight W and radius b (< a). The cylinders A, C are pressed
towards each other by equal forces P acting horizontally through their
axes. Find the pressure of B on the table, if 6>a/4 and the outer
cylinders do not lift. [S.]
15. Three cylinders rest in equilibrium each touching a table so that
two of the same radius a are in contact with the third of smaller radius,
the two outer cylinders being pressed towards each other by equal
forces P acting in horizontal lines at a distance a from the table. If
the radius of the smaller cylinder be such that its pressure on the table
is the greatest possible, then this pressure is §P + W, where W is the
weight of the smaller cylinder. [I.]
16. An equilateral triangle, of side a, is supported by a string which
has its extremities fastened to two angular points and passes through
two small smooth rings fixed at a distance c from each other in a hori-
zontal line. Prove that if the triangle hangs with one side vertical, the
length of the string is y/Za — c. Shew also that c cannot be greater than
[I.]
88 THE SOLUTION OF PROBLEMS [VI
17. A string of length 21 has one end attached to the extremity of a
smooth uniform heavy rod of length 2a, and the other carries a weight-
less ring which slides on the rod. The rod is suspended by means of the
string from a smooth peg: prove that if 0 be the angle which the rod
makes with the vertical, then Zcos0 = asin 3 9. [S.]
18. A uniform smooth plank of weight W and length 2a is hinged to
the bottom horizontal edge of a smooth plane, which is fixed at an
inclination a to the horizontal, and a sphere of radius 6 and weight W
is placed between the plank and the plane; prove that, in the position
of equilibrium the angle, 8, between the plank and the plane is given by
the equation
and the direction Of the action at the hinge is inclined to the plank at
an angle whose tangent is
/6. [S.]
19. A ladder of length 2a and weight W, with its centre of gravity
three-eighths of the way up it, stands on a smooth horizontal plane
resting against a smooth vertical wall, and the middle point is tied to a
point in the wall by a horizontal rope of length I; find the tension of
the rope. [S.]
20. A ladder A B rests against a smooth vertical wall at A and is kept
from slipping down by a cord PQ, fastened at Q to the ladder and at P
to the line of intersection of the ground and the wall. Prove that the
reactions at A, B and the tension in PQ are given by the following
construction: Let the reactions at A, B meet in I, and let the vertical
through the centre of gravity of the ladder meet AI in O and PQ in H:
also let PQ meet BI in K. Then if the weight is represented by IK, the
other sides of the quadrilateral OHIK represent the reactions and
tension. [S.]
21. A uniform triangular lamina ABC rests between two smooth
pegs D and E in the same horizontal, its vertex A being downwards.
Shew that, if <j> be the angle which the bisector of the angle at A makes
with the vertical in the position of equilibrium,
3DE_smj(B-C) A eosj(B-C) A
BG ~ sin^ s m 2 " + ^ cos
2* [
°-]
22. Two circular cylinders, of radii Sa and a, are glued together along
a generating line and the compound body thus formed rests with the
larger cylinder on a horizontal plane and both cylinders in contact with
a third, of radius a, which itself rests on the plane. If all the bodies are
smooth and the generating lines are all parallel and horizontal and the
weight of the upper small cylinder is W, prove that the total pressure
between the two lower cylinders will be —jir~ ^- [S-]
 EXAMPLES 89
23. A uniform rod, of length c, rests with one end on a smooth elliptic
arc whose major axis is horizontal and with the other on a smooth
vertical plane at a distance h from the centre of the ellipse; the ellipse
and the rod both being in a vertical plane. Prove that, if d is the angle
which the rod makes with the horizontal, and 2a, 26 are the axes of the
elli se
P -
where
and explain the result when
a = 26 = c, h = 0. [S.]
24. A uniform rod of length 2a and weight W hangs in an oblique
position, supported by a light inextensible string of length 2l(l>a)
whose ends are fastened to the ends of the rod, and which passes over
a smooth peg, and a weight w is attached to the rod at a distance d from
its middle point. Prove that the lengths of the string on the two sides
of the peg are // d\
d\ f d
1(1-—)
1\ ( 1an) and J ( l +
) ) and \ an)
where nw=W+w. [S.]
25. A B, BC are two uniform heavy rods of equal length and weight
W. The rod AB can move freely about A, there is a hinge at B, and at
C there is a ring which can move freely along a fixed rod through A
which is inclined downwards at an angle a to the horizontal. Shew that
in equilibrium tan BAG = % cot a, and that the horizontal component
of the action at B is f IF sin 2a. [S.]
26. Two uniform rods AB and CD each of weight W and length a
are smoothly jointed together at a point O, where OB and OD are each
of length 6. The rods rest in a vertical plane with the ends A and C on a
smooth table and the ends B and D connected by a light string. Prove
aW
that the reaction at the joint is -g, tan a, where a is the inclination of
either rod to the vertical. [S.]
27. A uniform rod of length 21 is attached by smooth hinges at its
ends to a vertical wall, and to a uniform smooth sphere of the same
weight as the rod, which rests against the wall. Prove that the co-
tangents of the angles which the rod, and the actions at the hinges,
make with the vertical are proportional to 3:2:4. Prove also that, if
the rod is inclined to the vertical at an angle cot" 1 2, the radius of the
sphere is Vlil. [S.]
28. Two light rods AO, OB are jointed at O. The rod OA is attached
by a hinge at A to afixedrod AD and B is attached to a ring which can
slide along AD. A force P acts at O towards AB at right angles to AB
and a force Q acts at B in the direction BA. The joint, hinge and ring
are frictionless and the angles OAB, OBA are acute; prove that in
equilibrium PoosOABoosOBA^QsinAOB. [S.]
 90 THE SOLUTION OF PROBLEMS [VI
29. Two equal beams AB, AC each of weight W connected by a
hinge at A are placed in a vertical plane with their extremities B, C
resting on a horizontal plane; they are kept from falling by strings con-
necting B and C with the middle points of the opposite sides; shew that
the tension of either string is
—
o
where 9 is the inclination of either beam to the horizon. Shew also that
the action of the hinge on either rod is fJPcot 0. [S.]
30. Two equal uniform rods A B. A C, each of weight W, are smoothly
jointed at A: D, E are points on AB, AC respectively, such that
AD = AE — X. AB and two light equal strings connect B to E and C
to D. The system stands in a vertical plane on a smooth horizontal
table. Shew that the tension in either string is represented by BE or
CD, provided that W is represented by 8A times the line joining A to
the mid-point of BC. [S.]
31. Two equal rods of weight w are freely j ointed and their free ends
are attached by strings to a fixed point. A circular disc of weight W and
radius r rests in the angle between the rods, and the whole hangs in a
vertical plane. If 2a be the length of each rod and each string, and 20
the angle between the rods, prove that
r = 2asin*0tan0-3^±^. [S.]

32. Three rods of the same material are freely jointed so as to form
a triangle ABC. The middle point of AB is supported; shew that, when
the triangle is in equilibrium, AB makes an angle with the horizon
\ (A — B) and that the action on either rod at the joint C makes the
same angle with the horizon and is of magnitude
i {(W - WY + ( W + W)* cot 8 * (A + B)}*;
W, W being the weights of the rods AC, BC. [S.]
33. An elliptic wire is fixed with major axis vertical. Two smooth
rings slide on it, being connected by a string passing through a smooth
ring at the upper focus. Shew that if the weights of the rings are equal
equilibrium is possible in all positions, but if they are unequal one ring
must be at the highest or the lowest point, according as the length of
the string is less or greater than the major axis. [S.]
34. A hollow right circular cylinder with smooth interior and open
at both ends is placed on a horizontal plane. A series of n uniform rods
of equal weight (w) and of equal length (greater than the diameter of
the cylinder) are placed one on the top of another and resting against
the inside of the cylinder. Shew that the cylinder will not fall over, if
its weight is greater than nw. [I.]
 EXAMPLES 91
35. A triangle ABO hung up by the corner A is formed of uniform
rods cut from the same uniform bar and smoothly jointed at A, B, G.
The rod AB is loaded so that BG is horizontal. Find the action on the
rod AB at B, and prove that its direction is inclined to the vertical at
an angle whose tangent is
ootofl + ^ f V [S.]
36. One end of a uniform rod is hinged to a fixed point, a cord
attached to the other end passing over a pulley sustains a weight, the
rod and pulley being in the same vertical plane. If the system be in
equilibrium with the rod inclined at a given angle i to the horizontal,
and the weight attached to the cord is the least possible, shew that the
pulley lies on a fixed straight line, and find the position of this line.
Also find the ratio of the attached weight to the weight of the rod. [S.]
37. A smooth lamina in the form of an equilateral triangle ABC is
placed on a horizontal table so that each of its sides touches one of
three smooth pegs P, Q, B, where
AB: RB= BP: PG= CQ: QA = n: 1.
A couple O acts on the lamina; shew that the pressure on each of the
P]

38. A triangle ABO is formed of three light rods smoothly jointed to

each other at their ends. D is the foot of the perpendicular from A on
BO. A string at tension T joins AD. Find the stresses in the rods AB
and AC. [S.]
39. ABOD is a square formed of four light rods jointed together,
the diagonal AC being a fifth light rod. Weights P and Q are attached
to the corners B, D respectively, and the system is hung up by the
comer A. Find the inclination to the vertical of the rod AC and also
the stress in it. [S.]
40. Four equal rods hinged together form a rhombus ABCD, which
is hung in a vertical plane from a fixed point E by two equal rods EB
and ED and a third EA, of such lengths that the angle at G is 60° and
each of the angles at E is 15°; the rods being without weight and a
weight W attached to C, prove that the force in the rod EA is

41. ABOD is a rhombus formed of freely jointed light rods. AC is

vertical, A being the higher end, and B, D are tied by strings of equal
length to a fixed point in the line AC. Weights W, W (W> W) are
suspended from A, C. If the rhombus is constrained to remain in a
vertical plane, prove that in the position of equilibrium the fixed point
divides AC in the ratio W: W. [S.]
92 THE SOLUTION OP PROBLEMS [VI
42. A rhombus ABCD of smoothly jointed rods rests on a smooth
table with the rod BC fixed in position. The middle points of AD, DC
are connected by a string which is kept taut by a couple L applied to
the rod AB. Prove that the tension of the string is
2LI(AB)cos(iABC). [S.]
43. Four light rods AB, BC, CD, DA are freely jointed together;
AB—BC and CD = DA. The rod AB is fixed horizontally and masses
P, Q are suspended from C, D respectively. Shew that in equilibrium
the angles DAB, ABC will be both acute or both obtuse, and that if a,
j8 are the angles which AD, CD make with the vertical

i^i +^ rsi
+ Lb>J
sin2a P-
44. Five weightless rods form a regular pentagon ABCDE, and the
framework is stiffened by other weightless rods BD, CE. The system is
placed in a vertical plane with CD on a horizontal table, and a weight
W is hung from A. Prove that the thrust in BD or CE is IF cot \n.
[S.]
45. Two corners of a regular pentagon of light freely jointed rods
are connected by a string in tension, and equilibrium is maintained by
another string also in tension, connecting one corner to the middle
point of the opposite side, the strings being perpendicular to each
other. Shew that their tensions are in the ratio 2 sin £ ir: 1. [S.]

46. Six equal weightless rods are smoothly jointed at their ends so
as to form a regular tetrahedron ABCD. Prove that if equal weights
W are attached to the joints C, D, while the rod AB is supported in a
horizontal position, the thrust along the rod CD will be W/V2. [S.]
47. Six equal rigid weightless bars, freely jointed at their ends, form
a regular tetrahedron A B CD. It is suspended from A, and three weights
each equal to W are hung from B, C and D respectively. Find the
stresses in all the bars. [S.]
48. A framework in the form of a tetrahedron is formed of six equal
light rods freely jointed to each other at their ends. A string at tension
T joins the middle points of two opposite rods. Prove that the stress
in each of the other rods is T/2-\/2. [S.]
49. A smoothly jointed quadrilateral of rods lies on a smooth hori-
zontal table and is enclosed in a smooth circular hoop which presses
tightly at each of the hinges. Prove that the pressures at the hinges are
proportional to the sides of the circumscribed quadrilateral which
touches the hoop at the hinges and that the stresses in the rods are
inversely proportional to their distances from the centre of the hoop.
[S.]
 EXAMPLES 93
50. AB represents the piston-rod of the fixed cylinder of a steam-
engine, and CD is a crank turning an axle D, BC being a connecting-
rod. DE is drawn perpendicular to AB meeting BC in E, and CF is
the perpendicular from C on AB; CO is perpendicular to BC.
Shew that if the thrust in AB is given,
(i) the couple exerted on the axle is proportional to DE;
(ii) the transverse pressure on the guides constraining the joint B
to move along AD is proportional to CO;
(iii) the thrust in BC is inversely proportional to BF.
[SJ
c
E

51. Four equal uniform rods are freely jointed together so as to

form a square ABCD, and the system hangs from the point A, the
square form being maintained by an inextensible string connecting the
middle points of AB and BC. If W is the weight of each rod, prove that
the tension of the string is 4TP, that the stress at-Bis£TF-\/17in the
direction inclined to the vertical at the angle tan- 1 £, that the stress at
C is JTF\/5 in the direction inclined to the vertical at the angle cot" 1 2,
and that the stress at D is $W in a horizontal direction. [S.]
52. A framework of four heavy rods, of length a, hinged together
to form a rhombus is supported by a smooth cylinder of radius c.
Shew that, if the rods are in equilibrium when each makes an angle 30°
with the vertical, then a = 4V3c. Find the ratio of the actions at the
top and bottom hinges. [S.]
53. Four equal uniform straight rods are freely jointed together to
form a rhombus ABCD, the rod AD being fixed vertically with A
uppermost. A string CE joins the corner C to a point E in AB; prove
that the tension of the string bears to twice the weight of a rod the
ratio of CE to BE. Find also the magnitude and direction of the stress
at the joint B. [S.]
54. Four heavy uniform rods A B, BC, CD, DA, of which AB = AD
and BC= CD, are jointed at their extremities, and the corners A and
C joined by a string of such a length that AB and BC are at right
angles. If the framework is suspended from A, prove that the tension
of the string is
where I is the length of the string, a the inclination of the upper rods to
the vertical, and w is the weight of the rods per unit length. [S.]
94 THE SOLUTION OP PROBLEMS [VI
55. A pentagon ABCDE is formed of rods whose weight is w per
unit length. The rods are freely jointed together and stand in a vertical
plane with the lowest rod AB fixed horizontally, while the joints C, E
are connected by a string. If BC and AE are of length a, GD and DE
of length b and the angles at A and B are each 120° and the angles at
C and E are each 90°, shew that the tension of the string is
a + Sb r<J i
W [ ]

56. Five equal uniform rods, each of weight W, are freely jointed at
their extremities, so as to form a pentagon. The pentagon is suspended
in a vertical plane by a string attached to one corner, and the two
adjacent corners are connected by a light rod of such length that the
pentagon is regular. Shew that the stress in this rod is
W (tan 18° + 2 tan 54°),
and find the reactions at the corners of the pentagon. [S.]
57. A regular hexagon ABODEF is made of six equal uniform rods
jointed freely. The hexagon rests in a vertical plane having AB in
contact with a given horizontal plane and C, F are connected by a
light inextensible string; shew that the tension of the string is W\/3,
where W is the weight of a rod, and shew that the action at the joint E
[ S ]
"

ANSWERS
3.
11. P+TF/2i/3. 14. W + P(a-b)W(ab). 19.
35. \V{w2 + \w + w'Ycot2 G), where w, w' are the weights of BG, CA.
36. Perpendicular to the rod through its free end. J c o s i .
38. TcosecAcosG, TcosecAcosB.
39. tan.-H(P-Q)/(P+Q)}, T/2PQ/V(P*+Q*).
47. TF/V6, WV(3/2). 52. 11:1.
vert ca
2 "*" BE ) ^ ^y» where w = wt. of a rod.
 Chapter VII
B E N D I N G MOMENTS
7*1. In the preceding chapter we considered a number of
problems about the equilibrium of rods or beams, but beyond
an occasional reference to the fact that a light rod may be re-
garded as a 'tie' or a 'strut' exerting equal and opposite forces
at its ends along its length, no reference has yet been made
to the sort of stresses that are set up in a rod or beam by
externally applied forces. We shall discuss this subject in the
present chapter.
7*2. Let AB be a beam in equilibrium under the action of
any given system of external forces. Imagine the beam to be
divided into two parts AC, GB by a cross section at any point
C; and consider the equilibrium of one part of the beam, say

J L

GB. We may assume that some, but not all, of the given ex-
ternal forces act on the part GB, and that these forces which
act upon CB would not in general be in equilibrium of them-
selves, because the other forces which act upon A G are needed
to balance them. But the part CB of the beam is in equili-
brium, therefore the external forces which act upon it must be
balanced by stresses at G exerted by the part AC upon the
part CB.
It follows that the stresses in the beam at any cross section C
are a system of forces that would balance the external force
system which acts upon either portion into which the beam is
divided at G.
Now confining ourselves to coplanar forces, we saw in 5*1
that any system of coplanar forces can be reduced to a single
force acting at an arbitrarily chosen point together with a
couple. Hence the external forces acting on the part GB are
equivalent to a single force acting at C together with a couple.
96 BENDING MOMENTS [VII
Let the force be resolved into components T along the beam
and S at right angles to it, and let M be moment of the couple.
Then T is called the tension or thrust according as it is a force
tending to lengthen or shorten the beam; 8 is called the
shearing force, and M is called the bending moment at C.

This diagram shows the external forces on CB and on A C

respectively each reduced to act at the point C and constituting
equal and opposite systems.
Now the forces exerted by AC on CB must balance the
external forces on CB, so that we may represent the forces
exerted by AC on CB thus:

and the equal and opposite forces exerted by CB on AC thus:

By reference to 5*1 we see that the shearing force and the

tension or thrust may be found by moving all the external
forces that act upon CB (or AC) parallel to themselves to act
at C, and that the bending moment at C is the algebraical
sum of the moments about C of all the external forces which
act upon CB (or AC). The sense of the shearing force and of
the bending moment are determined in this way.
7-21. Examples, (i) A light beam of length I is supported at its ends
and carries a load W concentrated at a point at a distance a from one end.
Let AB — I be the beam, and let AC = a.
By taking moments about B, we find that the supporting force at A
is W (1-T) , Biid therefore that at B is IF?.
•H)-
 7-2, 7-21] EXAMPLES 97
Let P be any point on the beam at a distance x from A, then when P
lies between A and C,
or for 0 < x < a, we have
S=w(l-j) upwards,
and M=W \\—-j)x clockwise;
\ IJ »vv
and when P lies between C and B,
or for a < x < I,
S is the resultant of W. (1 — j ) upwards and W downwards,

i.e. = TF y downwards,

and M is the sum of the moments of the same two forces,

i.e. M=w(l-tf)x-W(x-a)=Wa(l-j\ clockwise.

These values of 8 and M have been found from the forces on A P; if

we used instead the forces on PB we should clearly get the same
numerical values but acting in the opposite sense.
Graphical representations of the values of the shearing force S and
the bending moment M are
useful as shewing at a glance
the way in which they vary c B
along the beam. 'W
Ordinates are drawn at
points along the beam to.re- (
present S or M as the case
may be. In this example the Graph c
graph of 8 is two lines ac, c'b of S c
parallel to the beam with a
jump in the value in passing c b
the load W; and the graph
of M is two straight lines
c
which pass through A and B Graph
and meet on the vertical of M
B
through O, shewing no dis- c
continuity in M in passing
the load W, though there is a discontinuity in the gradient of M.

(ii) A heavy uniform beam of length I supported at one end and at a

distance \lfrom the other.
Let ABC be the beam supported at Af w I
B and C. Let w denote the weight of A p R
unit length, so that wl is the total '—I—'—A
weight. Then by taking moments *wx
98 BENDING MOMENTS [VII
about C we find that the supporting force at B is %wl, so that that at
C is ±wl.
Let x denote the distance from the end A of a variable point P. We
propose to find the values of the shearing force S and the bending
moment M at P as P moves along the beam from A to C.
So long as P lies between A and B, i.e/for §<x<\l, the only force
on AP is its weight wx acting at a distance $x from P, so that
S=wx downwards,
and M=£wx* counter-clockwise.
As P moves to the right of B we have to include in the forces acting
upon AP the upward force \wl
at B as well as the weight wx,
so that for \l<x<l B
8 = wx—\wl downwards,
and M = \wx*

counter-clockwise. G paph
If we draw the graphs of S o f s
and M, we notice that the graph
of S consists of parts of two
parallel straight lines, with a
discontinuity at B; that the
graph of M consists of arcs of
two parabolas, and that M
vanishes at the ends and the Gnaph
middle of the beam and has no of M
point of discontinuity in its
y \

values though there is a dis-

continuity in the gradient at the point B.
7-3. The examples just considered shew that the shearing
force and the gradient of the bending moment are liable to
discontinuities. We will now formulate some general relations
between shearing force and bending moment for a straight
beam and see why and when such discontinuities are to be
expected to arise.

MD
Sc

Let us disregard tension and thrust and suppose that any

7-21, 7-3] BENDING MOMENT AND SHEARING FORCE 99
length AC ( = x) of a beam is acted upon by external forces
which give rise to a shearing force 8C and a bending moment
Mc at C. As it is necessary to adopt some convention about
signs, we will suppose that Sc acts downwards and Mc in the
clockwise sense on the portion of the beam on the left of C, and
of course in the contrary senses on the portion of the beam on
the right of C, and that a like convention is adopted all along
the beam.
Consider a short length CD of the beam. The shearing force
and bending moment at C are denoted by Sc and MG, and in
like manner those at D are denoted by SD and M^ acting in
the senses indicated in the figure in accordance with our con-
vention. We may consider three cases:
(i) When CD is subject to no forces but shearing force and
bending moment: by resolving vertically,
8» = 80 (1),
and by taking moments about D,

which, as the length CD (or dx) tends to zero, gives a differentia

« -a (2,.

(1) shews that in this case the shearing force is constant,

and (2) shews that at all points of CD the gradient of the bend-
ing moment is equal to minus the shearing force.
(ii) When CD carries a uniformly distributed load, such as its
weight: by resolving and taking moments as above, we get

and MD-Mc o \
where w is the load per unit length.
These, as the length of CD diminishes, give differential
relations JO
(3)

T-s <*>•
7-2
100 BENDING MOMENTS
so that the gradient of the graph of the shearing force is minus
the load per unit length and is constant for a uniformly dis-
tributed load. We also see from (4) that the gradient of the
bending moment is equal to minus the shearing force whether
the beam is loaded in this manner or not, and that a discon-
tinuity in the shearing force implies a discontinuity in the
gradient of the bending moment.
These facts are exemplified in the examples of 7-21; as also
the fact that the vanishing of the shearing force is accompanied
by the vanishing of the gradient of the bending moment.
(iii) When CD carries a finite load W concentrated at a point
on CD: by resolving and taking moments as above, we get
8D-80=-W (5),
and MD-Mo + Sc.CD-W.eCD = 0 (6),
where 0< e< 1.
It appears from (5) that a concentrated load gives rise to a
discontinuity in the shearing force, as is otherwise obvious;
but if in (6) the points C and D move up to the point at
which the load is concentrated so that CD^-0, we get
MD = MC, so that no discontinuity arises in the bending
moment, but inasmuch as there is a discontinuity in the shear-
ing force there will also be a discontinuity in the gradient of the
bending moment.
7-4. Examples, (i) A light rodABCD,in whichAB = 5ft.,BC — 10ft.
and CD = 5 ft., is supported at B and C and carries uniformly distributed
loads of 1 Ib. per foot along AB and CD and a load of 20 Ib. concentrated
at the middle point E of BC. Draw diagrams to represent the shearing
force and bending moment at all points of the rod.
ii 5 .15

"20
Since there is symmetry about the centre and the total load is 30 Ib.,
therefore the supporting forces at B and C are each 15 Ib. If a; denotes
the distance from A of a variable point P which moves across the beam
from left to right, then the shearing force and bending moment at this
point are given by the following set of values:
whenO<a;<5, <S=a;lb.,
M=%x*ft. Ib.;
7-3, 7-4] EXAMPLES 101
5<x<10, £=5-15=-101b.,
M=5(x-%)-15(x-5)
= (_10a; + i|i)ft. lb.;
£=5-15 + 20=10 lb.,

t. lb.;
15<a;<20,

= |(20-a;) 2 ft. lb.

It will be noticed that in this procedure so long as P lies between A
and B, the effective load on AP is x lb., but when P is to the right of
AB the whole load on AB contributes to the shearing force and bending
moment; and that when P is to the right of C we have to take into
account the load on CP, viz. (re—15) lb.
Also, when P lies on CD, S and M could easily be written down by
considering the force on PD which is simply the load (20 —a;) lb.; we
may use this as a check on the results obtained above, noting the
difference of sign according as the force and couple are considered to be
acting upon AP ov upon PD.
>
A B E C D

Graph
of S

Graph
of M \ /

\ /

(ii) A uniform rod of length SI and weight w per unit length rests upon
two supports at its points of trisection. A peg of small diameter is fixed
horizontally through the centre of the rod and a couple of moment \wP is
applied to the axis of the peg. Sketch carefully diagrams of shearing force
and bending moment. [T.]
102 BENDING MOMENTS [VII
Let OB be the rod, A, B the points of trisection and C the middle
point. Then the total weight is Swl, and if B and Q denote the supporting
forces at A and B, we have JJ _|_ Q _ 2wl,
and by moments about B

whence l = wl and Q = 2wl.

H=wl C

Then for the shearing force and bending moment at a point P at a

distance x from 0 we have the following values:
whenO<a:<Z, S=wx,

S=w(x-l),

S-wx-R-Q= -
- 8 (x-

This is an example in which there is a discontinuity in the value of

the bending moment. It arises from the fact that there is an externally
applied couple at a particular point of the beam.
1

A C B

Graph
of S

-"
Graph
of M

(iii) A uniform beam of length 21 rests symmetrically on two supports

which are a distance 2a apart in a horizontal line; prove that the beam is
7-4] EXAMPLES 103
least liable to break if a = l(2— \ / 2 ) , it being assumed that the beam is
liable to break if a definite bending moment is exceeded at any point. [S.]
Let AB be the beam, G its middle point and D, E the points of sup-
port, so that AD — EB — I —a. If w is the weight of unit length, the
supporting forces are each wl.
wl
A
A D" C E B
(i)

If x denotes the distance of a point P from A, then for the bending

moment at P , we have that
when O^x^l — a,
M1 = $ wx2 with a greatest value \w (l — a)2&tD.
And when I — a^x^l + a,
iW2 — kwx^ — wl {x — I + CE)

Now if I > 2a, then Jkf2 cannot vanish and its values vary from a
greatest value $w (I — a)2 at D to a least value \w (I2 — 2al) at C.
The diagram of bending moments is as shewn infig.(ii).

(ii)
In this case the beam must be able to stand a bending moment
£w(l — a)2, the value at D and E; this decreases as a increases, but
since a< \l, it cannot be less than \wl2, which is its value when a = \l.
Again if I < 2a, then M2 is negative at C and the diagram of bending
moments is as shewn infig.(iii).

(iii)
The numerical values of -M2 at D and at C are then \w(l — a)2 and
\wl(2a — l). The former is decreased and the latter is increased by
increasing a. It follows that the value of a which will give the lowest
maximum to M2 is found by making the above two values equal, i.e.
when
(lay = 2all\
or when a = l(2~\/2),
104 BENDING MOMENTS [VII
2 2
and this makes the greatest bending moment %wl (\/2— I) which is
less than %wP, so that this value of a gives the lowest value to the
maximum bending moment along the beam.

EXAMPLES
1. A uniform heavy rod AB is supported at each end. If w be the
weight per unit length, prove that the bending moment at any point
PialwAP.PB.
2. A light horizontal rod 16 ft. long is supported at points distant
2 ft. from one end and 3 ft. from the other, and is loaded with a weight
of 6 lb. at its middle point and weights of 3 lb. at each end. Draw
diagrams shewing the shearing stress and bending moment at all points
of the rod. [T.]
3. A beam AD, 35 ft. long and supported at points B and O distant
10 and 5 ft. respectively from the ends A and D, carries a load of 5 tons
at D. If the weight of the beam is neglected, draw diagrams shewing
the shearing force and bending moment at all points of the beam. [T.]
4. A light rod is maintained in a horizontal position by supports at
its ends. Find expressions for the bending moment as a concentrated
load is moved slowly along the rod, and shew that the tendency to
break is greatest at the load and has its maximum value when the
load is in the middle of the rod.
5. A light rod is supported at A and B and carries a load at Q.
Prove that the bending moment at P is proportional to AP.QB or to
AQ. PB according as P lies on AQ or on QB.
6. A uniform rod ABC of length a is smoothly hinged at A and
supported at B at an inclination of 45° to the horizontal, where
AB = f a. Find expressions for the bending moment at all points of the
rod and shew that it vanishes at a distance f o from the end A.
7. A light rod passes through two smooth closely fitting fixed rings
at its ends and is acted upon by equal and opposite forces at its points
of trisection. Draw a diagram to shew the bending moment at every
point of the rod.
8. Two uniform heavy rods AB, BG of the same weight per unit
length are smoothly jointed at B and maintained in a horizontal
position by supports at A, D and C, where AD = 3DB = SBC. Find the
pressures on the supports and draw diagrams shewing the shearing
force and bending moment at all points of the rods.
9. Two equal uniform beams AB, BC of length a and of the same
weight per unit length w are smoothly hinged at B and supported in a
horizontal line by props at A and D, where BD = \DC. Find expres-
sions for the shearing force and bending moment at any point of each
beam and draw graphs to represent the variations in their values.
[S.]
 EXAMPLES 105
10. A uniform beam AB of length I and weight w per unit length is
smoothly hinged at A, and is kept at an inclination of 45° to the upward
vertical through A by a light horizontal rope which joins a fixed point
to the point of the beam distant 1/3 from A. Find expressions for the
thrust, shearing force and bending moment at any point of the beam.
[SJ
11. Two uniform rods AC, CB of the same weight per unit length
are smoothly jointed at C, and smoothly hinged to two points A, B in
the same vertical line. Prove that the reaction at G bisects the angle
ACB, and that the bending moments at two points P, Q on AC and
BC respectively are in the ratio
AP.PCsiaA to BQ.QCsinB.
12. A uniform horizontal beam, which is to carry a uniformly
distributed load, is supported at one end and at some other point; find
where the second support should be placed in order that the greatest
possible load may be carried by the beam, and shew that it will divide
the beam in the ratio 1 to -y/2— 1. [T.]
13. A uniform semicircular hoop hangs freely from one of its ends.
Prove that the bending moment is greatest at the point where the
tangent is vertical.
14. A beam A B of uniform material proj ects horizontally from a wall
at A. It is of length I and its cross section tapers uniformly from o2 at A
to 62 at B. Prove that the greatest bending moment is

where w is the density of the material.

15. Two uniform rods ACB and CD are smoothly jointed at O the
middle point of AB and the ends A, D are smoothly hinged to points
in the same vertical line with D below A and such that DA — %DC.
Determine the values of the tension or thrust, the shearing force and
the bending moment in AB and CD,
(i) when the rods are light and a load W is suspended at B;
(ii) when the rods are of weight w per unit length.

ANSWERS

6. On CB, . / 9 W ; on BA, -p—r* (3x — a) (x — a), x measured from C.

8. £Wa,tA, ifWsXD, \W at C, when W=wt. of BC.
9. 0<x<$a, S = %w(a — 2x), M=\wx(a — x);
%a<x<2a, S — w(2a — x),M= —\w (2a — x)2, x measured from A.
10. 0<x<%l, T = S = wxj^/2, M = wxil2^2;
%l<x<l, T = w{3l+2x)/2-v'2, S = -w(3l
M = w(2l — x)(l — x)j2il/2, x measured from B.
106 BENDING MOMENTS [VII
15. (i) On CB, thrust
T=~WcosDAB, S=Wsia.DAB,
MP=W .BP .ainDAB;
on AC, T=-WCOBDAB-4WOOBACD,
S=-WsinDAB, MQ = W (BQ-2CQ)sinDAB;
on CD, T = 4W, S = 0, M = Q.
(ii) x is measured from B on ACB and from C on CD.
On CB, Tx = - wx cos JOA B, SX = wx sin DA B,

on AC, Tx= -
Sx= -w{2a-x)sinDAB, Mx=iw(2a-x)2sinDAB;
on CD, Tx = wxcoaADC + wa(4: + coBACD),
Sx= ~w(b-x)sinADC, Mx- ~ \ wx (26 - * ) sin ADC,
where A C = a and AD = b.
 Chapter VIII
G R A P H I C A L STATICS
8*1. To determine graphically the magnitude, direction and
position of the resultant of any number of coplanar forces.
Let P 1 ; P 2 , P 3 , P 4 , P 5 be the given forces in fig. (i). The
method is a general one, but five forces are a sufficient number
for consideration.
From any point L in fig. (ii) draw LB12 to represent P l 5
B12B23 to represent P 2 and so on, Bi5M representing P 5 . Join
LM to complete the polygon. This diagram (fig. (ii)) is called
the Force Diagram and it will be shewn that LM represents
the resultant force in magnitude and direction.

B,2

Now take any point 0 in the plane and join it to the corners
of the polygon fig. (ii). 0 is called the pole of the force diagram.
Then return to fig. (i) and take any point Ax on the line of
action of the force P x and construct a polygon thus:
Draw
AXA% parallel to B12 0 to meet the line of action of P 2 in A2,
A2A3 parallel to B23 0 to meet the line of action of P 3 in As,
A3At parallel to BSi0 to meet the line of action of P 4 in Ait
AtA5 parallel to Bis 0 to meet the line of action of P 6 in A5;
108 GRAPHICAL STATICS [VIII

lastly, from A5 draw a parallel to MO, and from Ax a parallel

to OL to meet in A6. Then we shall prove that the resultant
force is equal and parallel to LM and acts through Ae.
The proof is as follows:
The force P1 acting at A1 = LB12
= LO along A6AX + OB12 along A2A1;
The force P 2 acting at A% = B12 B23
= B120 along AXA2 + OB23 along A3A2;
The force P 3 acting at ^43 = -B23-B34
= J? 23 0 along A2A3 + OB3i along
The force P 4 acting at A^ = B3iBiS
= BMO along A3At+ OBi5 along ^ 5 ^ 4 ;
The force P 5 acting at As — Bi5M
= 5 4 5 0 along ^4^5 + Oif along AeA5.
With the exception of the first and last, these forces are all
equal and opposite in pairs round the sides of the polygon
Ax, A2, A3, .... There remain only the forces LO along A6A1
and OM along AeA5, and these have a resultant E represented
in magnitude and direction by LM and acting through the
point A6.
The polygon A1} A2,... Aa is called a Funicular Polygon.
It is assumed that the given forces are acting upon the same
rigid body, but it is clear that if the points Alt A2, ... Ae were
joined by threads or by smoothly jointed light rods forming
the sides of a polygon the force R reversed in direction and
the given forces P 1 ; P 2 , ... would maintain the polygon in
equilibrium.
8-12. Special case of a system equivalent to a couple. When
the force polygon is closed by the coincidence of the points
L, M, the resultant force is zero, but we are not entitled to
assume that the given forces are in equilibrium. We must care-
fully construct a funicular polygon; then, since L and M are
coincident, the line through A5 parallel to MO and the line
through A-L parallel to OL are either coincident or parallel.
In the former case the force system reduces to two equal and
 8-1, 8-2] FUNICULAR POLYGON 109
opposite forces LO, OM in the same straight line and the
system is in equilibrium, and in the latter case the system
reduces to a couple of equal and opposite forces LO, OM in
parallel lines, and the moment of the couple is LO x distance
between these parallel lines.
8*2. It is evident that for a given system of forces by taking
different positions for the pole of the force diagram a great
variety of funicular polygons can be constructed and we have
the following theorem:
/ / the pole of a force diagram moves along a straight line the
sides of the funicular polygon turn aboutfixedpoints which all lie
on a parallel straight line.

Let LM be the side of the force diagram that represents the

force P i , and when 0 is the pole let A6AV AXAZ be sides of the
funicular polygon that meet on the force Px, parallel respec-
tively to LO, OM.
Let 0' be another position of the pole and through any point
X on AeAt draw X4 X ' parallel to LO' and through Ax' draw
A{Ai parallel to O'M meeting AtA2 in Y. Then the force P x
is equal either to
LO acting along XAX and OM acting along YA1}
or to
LO' acting along XAt' and O'M acting along
110 GRAPHICAL STATICS [VIII
Therefore these two pairs of forces are equivalent, from
which we deduce that
O'L + LO acting at X = O'M + MO acting at Y,
or O'O acting at X = O'O acting at Y,
and this can only be true if XY is parallel to 00'.
Similarly, if the sides A2Aa and A2'A3' of the two funicular
polygons which have 0 and 0' as poles intersect in Z, we can
prove that YZ is parallel to 00', and therefore X, Y, Z,... are
collinear points and the theorem follows. It is clear moreover
that the choice of the point X on the line A6 Ax is arbitrary, but
having chosen a point X then as 0 moves along a given line
00', the sides of the funicular polygon will all turn about their
intersections with a parallel to 00' drawn through X.
8*21. Reverting to the figures of 8*1, for different positions
of the pole O, the lines A±A6 and A5A9 always intersect in a
point on the line of action of the resultant force; so that for
different positions of the pole the locus of the intersection of a
pair of sides of the funicular polygon is a straight line. This is
true of any pair of sides, for example AXA2 and AtA5 intersect
on the line of action of the resultant of the forces P2, P3, P 4 .
8*22. For a given system offorces to construct a funicular polygon two
of whose sides shall pass through given points.
Suppose that the sides AXA2 and AtAb are to pass through given
points X, Y. We have seen (8"21) that these two sides intersect on a
fixed line, viz. the resultant of P 2 , P3, P 4 . Let SS' be this line. It is
parallel to 2J12 B45 in the force diagram. On it take any point K and
join KX, KY.
11B45

B12

Then, in the force diagram, draw parallels to KY and XK from

8-2-8-31] PARALLEL FORCES 111
and Bu intersecting in O, and O is the pole for a funicular polygon
satisfying the required conditions. For the resultant of the forces
P2, P3, Pt is B12 Bif acting along S'S and this is equivalent to forces
B12O acting along KY and OBi5 acting along KX.
For different points K on 88' the locus of 0 is a straight line 00'
parallel to YX (8-2).
8*3. Parallel Forces. Exactly the same method applies
when the given forces are parallel, with the simplification that
the sides of the force diagram all lie in the same straight line.

B46

B23

B34
B12

(i)

Thus if Plt P2, P3, P 4 , P 5 are five parallel forces acting in

given lines, we draw LB12, B12B23, B23B3i, B3iBi5, Bi5M to
represent them in order, take any point 0 for the pole of the
force diagram and join it to the points L, B12, B23, ... M.
Taking any point Ax on the line of action of P x we construct a
funicular polygon AtA2, ... Ae as in 8*1, and it follows by the
same proof as in 8*1 that the resultant of the five given forces
is a force B of magnitude and direction LM acting through A6.

8f31. Given a system of parallel forces to determine the magni-

tudes of two parallel forces which shall act in given straight lines
and be together equivalent to the given system.
Let Px, P2, P3 be the given forces and XX', YY' the given
straight lines. It is sufficient to consider a system of three
forces as the same method will apply for any number. Con-
struct a force diagram LB12, B12B23, B23M to represent the
112 GRAPHICAL STATICS [VIII
forces and take any point 0 as pole. From any point K on
XX' draw KA1 parallel to OL, then AXA2 parallel to 0B12,
A2A3 parallel to 0J323 and A3H parallel to OM to meet 7 7 ' in
H as infig.(i). Join KH and from 0 infig.(ii)draw ON parallel
to KH to meet LM in N. Then we have the
Force Px at Ax = LB12
= LO along AlK+OB12 along A1A2;
Force P 2 at A2 = B12B23
= B120 along A2A± + OB23 along ^42^43;
= ~B^M
Force P , at A»3 — -JS,,if
"23

= B23 0 along A3A2 + OM along A3H.

(i) Y' (ii) M

The forces in the lines AtA2 and A%AZ are equal and
opposite pairs, and there remain the forces
LO along AXK and OM along j4 3 i?.
We may suppose these to act at K and H respectively, and
write TO acting at K=ZF along XX' + NO along #iT
and OM acting &tH=ON along ILff + i O f along 7 7 ' .
The forces in the line KH are equal and opposite, so that the
system is equivalent to a force LN along XX' and a force NM
along 7 7 ' .
The method established above is important for it enables us
8-31-8-4] BENDING MOMENT 113
to find graphically the supporting forces when a beam is
supported at two points and carries given loads at other points.
8*32. Example. Given a system ofparallel forces P1,Pi,Pi,Pi,P6,
to construct a funicular polygon to pass through two given points, X on the
line of action of Px and Y on the line of action of P 4 .
Draw a force diagram LM for the five forces with any point 0 as pole,
and starting from X on P x draw consecutive sides of a funicular polygon
XAA^A^^. Then a line OK parallel to XAX meets LM in K and the

A4
A5

B34

B45

M
four forces Plf P2, P3, P 4 are equivalent to a force LK at X and KBa
at AA. Draw KO' parallel to YX, then if we construct a funicular
using any point 0' on KO' as pole'starting from X it will pass through
Y, for since O'K is parallel toXYwe shall have the four forces P1,P2,
P3, P 4 equivalent to a force LK at X and KBis at Y.
8*4. Graphical Representation of Bending Moment. A mea-
sure of the moment of a force or system of forces about a point
may be obtained from the funi- .
cular polygon.
Let P be a force represented in
a force diagram by the line ML.
Let 0 be the pole of the force
diagram and AB, AC the sides
of the funicular polygon parallel
to OL, OM. Thenifwe require the
moment of P about a point X,
we draw through X a parallel to LM cutting the sides of the
funicular polygon in B and C.
114 GRAPHICAL STATICS [VIII
Now the force P or ML is equivalent to MO along AC and
OL along BA. Therefore the moment of P about X = sum of
moments about X of MO along AC and OL along BA. Now
if OK be perpendicular to ML and we regard the last two
named forces as acting at G and B respectively, in calculating
their moments about X we need only consider their components
at right angles to BG which are KO and OK respectively.
Therefore the moment of P about X
= KO.CX+OK.BX
= OK.BG.
And OK is the same for all positions of X, so that the required
moment is proportional to the intercept BC.
In like manner the sum of the moments about X of any
system of forces may be measured. For the sum of their
moments = the moment about X of their resultant; and the
first and last sides of the funicular polygon intersect on the
resultant, so if we take P to represent the resultant and draw
a parallel to it through X cutting the first and last sides of the
funicular polygon in B and G the sum of the moments of the
forces about X will be proportional to the intercept BG.
8'41. E x a m p l e s , (i) As an application of the method of 8*4 consider
the case of a horizontal beam supported at its ends and carrying loads
Plt P2, P3 at given points.

A2
B
A, A3
\ J

dL •—-
P
3
8-4, 8*41] EXAMPLES 115
We construct a force diagram LBU B2SM with a pole 0 and a funi-
cular polygon YA1A2A3Z as in 8*31; the line ON parallel to YZ
determining the supporting forces NL and MN which act through Y
and Z. To find the bending moment at any point X on the beam, let
the vertical through X cut the sides of the funicular polygon in B, C;
then the bending moment at X is proportional to BC and is measured
by BC x OK, where OK is the perpendicular from the pole to the line
LM in the force diagram. For the bending moment about X is the
sum of the moments about X of the forces at the corners A3 and Z of
the funicular polygon, and, infindinggraphically the resultant of these
two forces, the first and last sides of the corresponding funicular
polygon would be A2A3 and ZY.

(ii) Two beams ABC, CD are smoothly jointed at C and supported

horizontally at A, B and D. Determine graphically the supporting forces
atA,B and Dfor given loads Pt and P 2 , and represent the bending moment
graphically.

Draw the force diagram LBnM so that LB12 and B12M represent
the forces P x and P 2 . Take the verticals through A, B, C,D and con-
struct a funicular polygon by the following steps: from any point H on
the vertical through A draw HA1 parallel to OL to meet P x in Ax, then
AiAa parallel to 0B12 to meet P 2 in A2, and A2K parallel to OM to
meet the vertical through D in K. We then make use of the fact that
there is a smooth joint and therefore no bending moment at C, so that
the funicular polygon must cross itself on the vertical through C. Let
AtA2 cut this vertical in Q. Join KO cutting the vertical through B in
N, and complete the funicular polygon by joining NH. In the force
diagram draw OB, OS parallel to HN, NK.
Then the reactions at A, B and D are equal to EL, SB and MS
respectively. This may be proved as follows:
8-2
116 GRAPHICAL STATICS [VIII

= LO along A1H+ OBla along AtA2

+ B120 along AsAt+OM along AtK
= L0 acting at H+ OM acting at K
= ZB along AH + SO along NH
+ OS along NK + SM along DK
= ~LR along AH + ItS along BN + Sid along DK,
and the reactions are equal and opposite to these three forces.
The bending moment at any point of the beam will be proportional
to the intercept made by the funicular polygon on the vertical line
through the point.
We notice that in addition to vanishing and changing sign as we pass
the joint C, the bending moment also vanishes and changes sign at a
point between B and the force Px.
8*5. Reciprocal Figures. A force diagram and the corresponding
figure containing the forces and funicular polygon are so related that
to every line in one figure there corresponds a parallel line in the other
figure. This is a reciprocal property, and is one of the properties of a
class of figures to which the name Reciprocal Figures was given by
Clerk Maxwell* who defined them thus:
"Two plane figures are reciprocal when they consist of an equal
number of lines, so that corresponding lines in the two figures are
parallel, and corresponding lines which converge to a point in one
figure form a closed polygon in the other.
"Since every polygon in one figure has three or more sides, every
point in the other figure must have three or more lines converging to it;
and since every line in the one figure has two and only two extremities
to which lines converge, every line in the other figure must belong to
two and only two closed polygons."

DA

The simplest plane figure fulfilling the conditions is that formed by

the six lines which join four points in pairs as shewn in the two figures
above. The reciprocal figure consists of six lines parallel to the original
six, points in one figure corresponding to triangles in the other.

* Scientific Papers, vol. i, p. 514.

8-41-8-61] RECIPROCAL FIGURES 117
8'51. Not every figure has a reciprocal. It is necessary that the
same relation shall hold between the numbers of points, polygons and
sides as between the numbers of corners, faces and edges of apolyhedron,
viz. G + F = E+2, where C, F and E denote the numbers of corners,
faces and edges: in fact that the figure may be regarded as the plane
projection of a polyhedron.
In the above figures there are four corners, four triangles and six
sides, so that G + F = E + 2, and this is the simplest pair of reciprocal
figures. If we derive another such figure by drawing additional lines,
the effect of drawing a line from a corner H to a point K unconnected
with the figure is to increase both G and E by unity; but if we draw a
line from one corner H to another corner K, while adding nothing to C,
we increase both F and E by unity. It follows that the relation
C + F = E + 2 continues to hold for all figures built up in this way.
A simple way of constructing a reciprocal to a figure composed of
triangles is to join the centres of the circumscribing circles of all the
triangles and then turn the figure through a right angle.
8*6. Frameworks. By a framework we mean a number of
bars or bodies of any shape jointed together. We shall assume
that the framework lies in one plane and that the joints are
formed by smooth pins passing through small holes in the
bodies, a 'single' joint being a point where two bodies only are
connected, at a 'double' joint three bodies are connected and
so on. Each bar or body is connected with the others at two
points only. The external forces are applied to the pins so that
the only forces acting on a body are the two forces at its joints
and these must be equal and opposite, so the body may be
adequately represented as a bar connecting the two joints.
8-61. A framework is said to be stiff or rigid if it cannot be
deformed. For example, a triangle of jointed rods is a stiff
frame; a quadrilateral of four rods AB, BO, CD, DA is not

stiff, but can be made so by the addition of a diagonal rod AC;

while two diagonals AC and BD would make it overstiff, for
if the length of BD, say, were such as to cause a tension or
118 GRAPHICAL STATICS [VIII

thrust between the points B and D there would be proportional

forces along all the other rods (6*43 (i)). A framework may be
described as just stiff when the removal of a single bar would
render it deformable. There is a connection between the num-
ber of joints of a framework and the number of bars necessary
to make the framework stiff. Let us begin with a single bar
connecting two points A and B; two more bars are required to
connect a third point rigidly with A and B, and so on for every
joint added to the framework two bars must be added. Hence
if the framework contains n joints in all, i.e. n — 2 in addition
to A and B, the number of bars necessary to make it stiff is
1 + 2 (n — 2) = 2n — 3. Otherwise, using plane co-ordinates, to
fix the position of n points in a plane, 2n co-ordinates must
be given; but this wouldfixnot only the shape of the frame-
work but also its position in the plane, and since only its
shape is to be fixed and it can move as a whole with three
degrees of freedom, therefore 2n — 3 data are required to
determine its shape, and these would be provided by the
lengths of 2n — 3 bars, so that 2n — 3 bars are sufficient to
make the framework stiff.

8-62. Determination of Stresses in a Framework. Con-

sider a framework of n joints and 2ra — 3 bars so that the frame-
work is just stiff. By resolving in two directions for the forces
at each joint we obtain 2n equations connecting the stresses in
the bars and the given external forces; but the external forces
alone have to satisfy three conditions of equilibrium, which
must be embodied in the 2n equations already obtained.
Therefore there are 2^ — 3 independent equations for the
determination of the stresses in the 2n — 3 bars, and the
problem of finding the stresses in a framework which is just
stiff is in general a determinate one. There may, however,
be exceptional cases in which the 2n — 3 equations are not
independent or are inconsistent and the problem is then
indeterminate.
For example in the case of three bars AB, BC, GA smoothly
jointed and such that
AB = AC+CB.
8-61-8-63] FRAMEWORKS 119
The jointed structure would be in equilibrium under the action
of equal and opposite forces P applied to the pins at A and B.
But if the stresses in AC, AB and BC are denoted by B, S and
T, we have R + s = P, B = T and S+T = P
and learn no more by resolving in any other direction, so that
the stresses are in this case indeterminate.

If a framework is over-rigid the stresses are indeterminate

because the number of unknown stresses exceeds the number
of independent equations, and, as in the example of the over-
stiff quadrilateral (6*43 (i)), the members may be in a state of
stress without the application of external forces. Such a frame-
work is said to be self-stressed. The three rods in the last figure
might constitute a self-stressed frame if, say, AB were in a
state of thrust 8, while the other two rods were in a state of
tension. The equations at the joints being S = R, R=T and
T = S, and the actual values indeterminate.

8*63. The Problem of a Framework. A framework is in

equilibrium under the action of given external forces and pos-
sibly also certain constraints; for example, it may carry certain
loads attached to some of its joints and it may be supported or
constrained at other joints; but, as explained in 6*2, the con-
straints must not involve more than three unknown elements
or the problem is indeterminate. It is required to determine
the stresses in the different members of the framework. We
shall assume in the first place that the problem is determinate,
in that it is possible to find by graphical or analytical methods
all the supporting forces or forces of constraint. We shall also
assume that the framework is not over-stiff and that at least
one of the joints at which an external force is applied is a single
joint. The stresses in the members meeting in this joint can be
determined by a triangle of forces; we then proceed to the
joints at the other ends of these members and can construct
120 GRAPHICAL STATICS [VIII
triangles or polygons of forces for these joints provided the
members that meet in a joint do not involve more than two
unknown stresses, and continue the process until all the stresses
are found. In practice it is not necessary to construct separate
triangles or polygons for the different joints after the first, but
merely to add, for each joint, additional lines to the figure
already obtained. The figure constructed will in general be
reciprocal to the figure consisting of the framework and the
external and constraining forces, in that it will contain the
same number of lines such that corresponding lines are parallel,
and corresponding to lines which meet in a point in the frame-
work there will be lines forming a closed polygon in the other
figure. There is not, however, complete reciprocity in the sense
defined by Maxwell because the lines that represent external
forces applied to the framework are not sides of closed polygons.
The method will be illustrated by a few examples. It will be
found that there is sufficient reciprocity between the two
figures for us to adopt a method of lettering introduced by
R. H. Bow in 1873. The lines that represent external forces or
supporting forces are placed outside the framework and, if we
imagine them to be produced outwards to the edges of the paper,
we see the latter divided up into triangles or polygons by the
members of the framework and the forces applied to it. We
then designate each polygonal area by a letter. If a line separ-
ates two areas A and B, then we place the letters A and B at
the ends of the corresponding line in the force diagram, and so
we obtain a complete correspondence between the lettering of
the diagrams, showing the way in which a polygonal area in
one figure corresponds to a point in the others.
8*64. Examples, (i) Fig. (i) represents a framework, hinged at
O, further supported by a horizontal force P of unknown magnitude at
the corner above O and carrying a single load W. To find P, the force
at the hinge O and the stresses in the members of the framework.
Firstly we letter the framework by the method described in 8'63, so
that each line infig.(i) has a lettered region on each side of it. The joint
to which W is attached is a single joint and we construct first a triangle
of forces FED for this joint taking any convenient length FE to
represent W. It is important to notice at this stage that this triangle of
forces is named FED (not FDE), because FE is the sense of the force
W, and that ED, DF give not only the magnitudes but also the senses
8-63, 8-64] FRAMEWORKS 121
in which the forces act along the rods at the j oint FED in the framework.
We mark these senses by inserting arrow heads in the framework, up-
wards in ED and right to left in DF. We then insert arrow heads to
indicate the equal and opposite forces which act at the other ends of
the same rods.
Then moving along the framework we notice that at the upper joint
FDGB there are four rods and only one known force, viz. that in FD,
so we are not yet in a position to deal with this joint. But at the joint
below, viz. DEC, there are only two unknown forces, viz. those in EC
and CD; so we construct a triangle of forces for this point, starting with
the known force DE, then drawing EC horizontal and CD vertical; and
then put arrow heads to indicate the senses in which the forces act at
this joint, viz. EC left to right and CD upwards, and in the contrary
directions at the opposite ends of the rods.

A B 0/
/ C /
O
1 w
CO
We may now proceed to the joint FDCB, where there are now only
two unknown forces CB and BF; we construct a polygon of forces for
this joint beginning with the known forces FD, DC, then drawing CB
and BF in the required directions and noting that two sides FD, BF
of the polygon overlap. We then put in arrow heads in the framework
to indicate the senses in which the newly determined forces act at this
joint, viz. CB upwards and BF right to left, and in the contrary
directions at the other ends of the rods.
Proceeding next to the joint above the hinge, we see that there are
two horizontal forces AF (or P) and FB and a vertical force BA, and,
since the latter would be unbalanced, it must be zero. So in the force
diagram the point A coincides with B and AF (or P) is equal and
opposite to FB.
There is now only one unknown force, the reaction at the hinge O.
Its line of action separates the regions E and A and the polygon of
forces for this joint is ABCEA, so that the reaction is EA in magnitude,
direction and sense.
The arrow heads now serve to shew that the rods BC, DE, EC are
in a state of thrust, while BF, FD and DC are in a state of tension.
It would have been difficult to make a blunder in this very easy
example, but in more complicated cases difficulties may arise in con-
structing the force diagram through failure to adopt a consistent con-
vention in the order in which the different members that meet in a
joint are taken. Thus in this example a clockwise convention was
adopted at every joint, viz. FED, DEC, FDCB and so on.
122 GRAPHICAL STATICS [vm
It should be noted that in this and the following examples the arrow
heads in the members of the framework are only inserted step by step
as their senses are determined by the construction of the force diagrams.
(ii) The rough sketch represents a jointed framework of light bars,
loaded as indicated; AC, CE are horizontal and the angles at A, G, E are
each 45°; the lengths AC, CE are
each 20 ft. and CF is 16 ft.
Calculate the reactions at A, E 200 600
assuming them to be vertical; and
draw a force diagram to give the
stresses in the bars, distinguishing
the ties (in tension) from the struts
(under compression). [C]
By taking moments about E
or A we find that the supporting forces are 500 and 700 at A and M
respectively.
We first re-letter the framework according to Bow's method. When
a figure like the given one has to be copied, the simplest method is
to prick through the corners with a pin on to the paper to be used.

400

200

500

The force diagram is then constructed, taking the joints in the

following order: MNG, GNOH, OPKH, PQLK and QML.
We find that the points G, H coincide, so that there is no stress in
the bar OH, and by measurement, or, more accurately, by using the
theorem of Pythagoras, that the stresses are as follows:
Ties Struts
GM=500 NG = nO1
LM = 100 OH =583
HK = 200 PK= 583
KL = 283
8-64] EXAMPLES 123
(iii) The figure represents a framework of smoothly jointed rods which
can turn about a pivot at G. It carries loads ofZlb.atA and N, and is kept
in equilibrium by a force P at B parallel
to AC The rods are all either vertical,
horizontal, or inclined at 30° to the vertical.
Determine the force P graphically or other-
wise. Construct a force diagram and find
the stresses in the sides ofthetriangleLMN, 3lb.
stating whether they are tensions or thrusts.
By taking moments about O we find
that P = 3\/3 lb.; or graphically, the
total load is 6 lb. bisecting NA, and the
resultant of this force and P passes
through C, from which it is easy to find
the same value for P by a triangle of
forces.
After re-lettering the framework we
now find that there are two single joints
at which we know the external forces, but that the neighbouring
joints are all triple joints, so that it is only by making use of all
our data that we are able to get a solution. Begin by constructing

,3

E / ' G
/
/F

H / C

the triangle BCG with the side BC—3 units. Passing next to the
centre top joint ABGFE, we have found the stress BO and we can
draw a line AB in the force diagram to represent the load of 3 lb., but
we axe unable to complete the pentagon because three sides are of
124 GRAPHICAL STATICS [vni
unknown lengths. But if we pass on to the left-hand top joint we know
that the line DA is to be 3-\/3 units of length and inclined at 30° to the
vertical. To construct this length make CBX a right angle and
BCX-6O0, then XB = .BCtan60° = 3 y 3 . Then in the force diagram
make the triangle of forces for the corner DAE, using the point A
already obtained and making DA equal to XB. This determines the
point E, and we can now return to the joint ABOFE and complete the
pentagon, of which we have already got the points O, B, A, E, by
drawing the sides OF, FE.
Take next the joint FOCH, two of the sides FO, 00 of the quadri-
lateral having already been found; then the joint DEFH, for which
DE, EF, FH have already been found. The force diagram is now
complete, and CD represents the force at the pivot.
The required stresses are found thus from the force diagram
, and 00=
therefore FE= 0H=^/3 and this is a thrust;
then FH= \FE=JV3 and this is a tension;
lastly = EH=^ FE= 1-5 also a thrust.

8-7. The Method of Sections. If a straight line section be

drawn across a framework intersecting a number of the bars,
then the stresses in these bars must together be in equiKbrium
with the external forces or constraints applied to either part
into which the framework is divided by this section; and if the
cutting line intersects not more than three bars, the stresses
in these bars can be found from the equations of equilibrium.
We will illustrate the method by a simple example.
The framework shewn in the figure, in which the angles are all either
30° or 120°, is hinged to a wall at A and F and carries a load of 50 Ib. at
D. To find the stresses in the bars using the method of sections.

iX

50 Ib.

The section XX' which cuts the bars CD, DE gives no more informa-
8-64-8-8] METHOD OF SECTIONS 125
tion than that the three forces at D are in equilibrium, and by Lami's
Theorem we find that
(ED) = 1001b. and (DC) = 50^/31b.,
where (ED) denotes the force along ED.
The section YY' gives that the forces in CD, CE and EF balance the
load of 50 lb. By resolving vertically we get
(.EC)sin30° = 50, therefore (EC)- 1001b.;
and by resolving horizontally, we get
(FE) = (EC) cos 30° + (DC) = 100 V3 lb.
The section ZZ' gives that the forces in BO, CF and FE balance the
load of 50 lb. By resolving vertically we get
(.FC)sin30° = 50, therefore (JO) = 1001b.;
and by resolving horizontally we get
(CB) = (FC) cos 30° + (FE) = 150-v/31b.
A section VU', which cuts AB alone, merely serves to point out that
the moment of (BA) about F is equal to the moment about F of the
load at D, from which we find that

Finally a section VV, which cuts AB, BF, CF and EF, gives that
the four forces in these bars balance the load of 50 lb. at D; so that by
resolving horizontally we get
(BF) cos 60° + (BA) cos 60° = (FC) cos 30° + (FE),
giving (BF) = 150-v/31b.
We have determined all the stresses by using sections, but the last
two forces might have been deduced at once because the angles at B
are all 120° so that
We have said nothing about the senses in which the stresses act.
These are easily determined from the fact that they have to be such as
to keep the joints in equilibrium; and it is well to bear in mind that in
every case we were considering the equilibrium of the portion of the
frame on the right of the section, so that the reactions were always
from or towards joints on the right of the section.
8*8. Distributed Loads. In the foregoing problems we
have only considered cases in which the loads are applied at the
joints. When it is required to take into account the weights of
the bars or of loads distributed along the bars we may proceed
as follows: resolve the total load on each bar into two forces
acting at the joints at the ends of the bar and deduce by the
foregoing methods a system of stresses in the bars which to-
gether with these concentrated loads would maintain all the
126 GRAPHICAL STATICS [VIII
joints in equilibrium. Then to find the tension or thrust, shear-
ing force and bending moment in any bar AB under the given
distributed loads we argue thus:
We want tofindthe reactions of the rest of the system on the
bar AB and these reactions are independent of whether the
load on AB is a distributed load or represented by its con-
centrated components, say P at A and Q at B. Suppose that
the stress in AB, in the problem of concentrated loads, is a
thrust T and that the resultant reaction of the rest of the
system on AB at A is a force R. Then since the joint A is in
equilibrium under the action of R, T and P, we have a vector
equation

or R=-P-T.
Therefore the resultant reaction at A of the rest of the system
upon the bar AB is simply the resultant of P and T reversed
in direction, similarly the resultant reaction at B is the
resultant of Q and T reversed in direction.
8*81. Example. The pin-jointed framework shewn in the figure is
supported by vertical reactions at D and E, and the loads consist of a
weight of 2 tons suspended at F and the weight of the curved member BC,

which may be taken as 2 tons, uniformly distributed along the horizontal

line BC. Prove that the bending moment at the middle of the member BC
ts
a(y/Z-\)tons-feet,
where a is the length, in feet, of the members AB, BD, AC, CE, and the
centre line of BO is a circular arc of radius a with A as centre. [T.]
By symmetry it follows that the vertical reactions at D and E are
8-8, 8-81] DISTEIBUTBD LOADS 127
each one-half of the total load, i.e. 2 tons. Also since FD, FE make
angles of 60° with the vertical, the stresses in FD and FE are tensions
of 2 tons. Then considering the joint at D, where stresses in DB and
DA are unknown, and either by constructing a polygon of forces or by
resolving, we find that there is a thrust of 2 tons in DB and no stress
in DA.
We then proceed to the joint at B, and suppose that half the weight
of the member BG acts at each point B and G, i.e. a force of 1 ton
vertically downwards. Having removed
the load, the only forces that now act upon
BG are forces at B and C and these must
be equal and opposite, so that in the pro-
blem in which loads are concentrated at
the joints the reaction T of BC on the joint
at B is a horizontal force.
Then, either by constructing a polygon
of forces or by resolving vertically for the
forces acting at the joint B, it is at once apparent that there is no stress
in AB.
If we followed the procedure described in 8*8 we should now reverse
the force T and compound with it a force of 1 ton vertically upwards at
B, in order to find the reaction at B of the rest of the framework on the
member BG when its load is re-distributed. This, however, is unneces-
sary because in this case the resultant of the two forces last named is
clearly the thrust of 2 tons in the
member DB.
The member BG is therefore in
equilibrium under the action of its
weight and the thrusts of 2 tons each
in the members DB and EC.
For the bending moment at M, the
middle point of the arc BG, we take
the moments about M of the forces
acting on the portion BM, viz.
2cos30°.JVilf

1V
= a (\/3 — |) tons-feet.
This example serves to indicate that, if a member is regarded as
perfectly rigid, its curvature makes no difference to the distribution of
stress in the other members of the framework but only in itself.
128 GRAPHICAL STATICS [VHI

EXAMPLES
1. Four forces acting in one plane are in equilibrium, their lines of
action are all given and also the magnitude of one of them; give a
geometrical construction to find the magnitudes of the others. [S.]
2. A light rod 4 B, 10 ft. long, is supported at A and at another point.
A load of 1 lb. is suspended from B, and loads of 5 lb. and 21b. at points 2 ft.
and 6 ft. from A. (i) If the second support is at the middle point of AB,
find graphically the pressures on the supports, (ii) If the pressure on
the support at A is required to be 4 lb., find graphically where the
second point of support must be. [S.]
3. A uniform plank of weight 40 lb. and length 8 ft. carries a weight
of 12 lb. at a point 3 ft. from one end. If the plank is supported at its
ends in a horizontal position, determine graphically the pressures on
the supports. [I.]
4. ABCDE is a light rod supported horizontally at B and D and
carrying loads of 10, lfiand 201b. &tA, CaxidE,wheieAB = DE = 2ft.,
BC = 3 ft. and CD = 4 ft. Determine graphically the pressures on the
supports and construct the diagram of bending moments.
5. AC and CF are light rods smoothly jointed at C, supported at
B, D and F and carrying loads of 8 and 30 lb. at A and E, where
AB=BC=CD=DE = EF.
Determine graphically the pressures on the supports and construct the
diagram of bending moments.
6. AB, BC, CD are equal light rods smoothly jointed at B and C
and supported horizontally at A, G, H, K, where O, H are points of
trisection of BC and K is the middle point of CD. They carry loads of
4, 20 and 2 lb. at D and the middle points of BC and AB. Determine
graphically the pressures on the supports and construct the diagram
of bending moments.
7. A portion of a Warren girder consists of three equilateral triangles
ABC, ADC, BCE, the lines AB, DCE being horizontal, and AB
uppermost. It rests on vertical supports at D and E, and carries 3 tons
at A and 1 ton at B. Find the reactions at the supports and the stresses
in the four inclined bars, stating whether they are in tension or com-
pression. [S.]
8. Five weightless rods AB, BC, CD, DA and BD, smoothly jointed
at their ends, form a framework: AB is vertical with B above A,
BAD = S0°, ABD = 60°, BC is horizontal and BD = DC. A weight W
 EXAMPLES 129
is suspended from C and the framework is supported by a horizontal
force at B and a force at A. Find the stress in each of the rods, dis-
tinguishing between thrust and tension. [S.]
9. A framework is made of light rigid rods smoothly jointed. Its
configuration may be obtained by describing squares ABDE, ACFO
on the sides AB, AC of an equilateral triangle ABC, the squares being
external to the triangle, and joining EO, so that EAO is an isosceles
triangle with an angle of 120° at A. The points FD are pressed together
by external agency with a force of magnitude P: determine, by
graphical methods or otherwise, the stress in the rod EO. [S.]
10. AB CD is a rhombus of freely j ointed rods in a vertical plane and
B, D are connected by a rod jointed to the rhombus. A and B are fixed
so that AB is horizontal and below the level of CD. The acute angle A
of the rhombus is a. If a weight W is hung from C, draw the force
diagram and find the stress in the rod BD in terms of W, a. [S.]
11. A regular pentagon ABCDE, formed of light rods, jointed at
the angles, is stiffened by two light jointed bars AC, AD. Two equal
and opposite forces, each equal to 3 lb. weight, are applied at B and E:
find graphically or otherwise the stress in each bar of the framework,
stating whether it is tensile or compressive. [S.]
12. A,frame of five light rods is formed by two congruent triangles
ABC, ABD on the same base AB and on opposite sides of it so that
angle ^lBC=:angle ABD = 4:5° and angle BAC= 120°. If this frame
be acted upon at C and D by equal and opposite forces, each of magni-
tude 56 lb. weight, determine by a force diagram the stresses in the
rods. [S.]
13. A regular hexagon ABCDEF formed of light rods, connected
by smooth joints, stiffened by light rods FB, FC, FD, and suspended
from the point A, has weights each equal to W attached to the joints
B, G, D, E, F. Find graphically the stress in each rod. [S.]
14. AB, BC, CD are three sides of a square, E is the point of inter-
section of AC and BD, and AEDF is a square; nine straight rods
without weight occupy the positions AB, BC, CD, DF, FA, AE, EC,
BE, ED, and are jointed together at their ends; the frame so formed is
in a vertical plane the point B being vertically above C, a given weight
is attached to F, C is hinged to a fixed point, and equilibrium is main-
tained by a horizontal force acting at B; find the tension or thrust in
each of the rods. [S.]
15. A framework, lying upon a smooth horizontal table, is composed
often rods. Of these AB, BC, CD, DA form a rhombus with angles of
60° at B, D, and AE, EC, CF, FA form a square of which E is the
corner nearest to B. The remaining two rods connect B, E and D, F.
RS 9
130 GRAPHICAL STATICS [VIII
Shew by graphical methods that, if all the joints be smooth and the
points A, Che pushed towards one another with force P, the stress in
BF and DF will be J P (\/3 + 3). Is this a tension or a thrust ? [S.]
16. A rigid plane framework of five jointed bars forming two equi-
lateral triangles BAC, CDA is in equilibrium under the action of
three forces on the joints at A, B, D. Prove that the directions of these
forces must be parallel or concurrent, lying in the plane of the frame-
work.
If the force at B be 4 cwt. acting perpendicularly to CD and the force
at A be in the direction BA, find the force at D and the stresses in the
bars by a graphical method. [S.]
17. ABC is a horizontal line such that AB = 5 ft. and BC — 15 ft.
D is a point vertically over B such that BD=. 10 ft. E bisects CD.
AB, BC, CE, DE, AD, BD, BE are seven freely jointed rods forming
a framework. Loads of 10 cwt. each are applied at D and E, and the
system is supported at A and C. Find graphically or otherwise the
stresses in the various rods, stating which are in tension and which in
compression. [S.]
18. A framework of seven freely jointed light rods is in the form of
a regular pentagon ABCDE and its diagonals A C, AD. The framework
is in a vertical plane with the lowest rod CD horizontal and is supported
at C and D and weights W, 2 W, W are suspended at B, A, E respectively.
Draw a force diagram to exhibit the stresses in the rods, state which of
them are in tension and which in thrust and shew that the tension in
2JFtanl8°(l-cos36°). [T.]
19. Thefigurerepresents a pin-jointed framework of light rods. The
rods BC, ED are horizontal and the length of each of the rods AB,
BC, ED is equal to AE. The whole is attached to a vertical wall at A
and E and the angle XAB is 60°. Determine by graphical methods the
stress in each rod. [S.]

1001b.

20. Draw a force diagram for the framework shewn in the diagram,
 EXAMPLES 131
in which all the rods except AB, AC are of the same length. Determine
the stresses in AB, AC. [S.]
I5

21. Thefigureshews a roof truss, which is loaded at B with a weight

W and is supported at A and C. The angle ABCiaa, right angle and is
trisected by BD and BE; the angles A and C are each 30° and BA — BC.
Draw a force diagram to shew the tension or compression in each
member. [T.]

22. The diagram represents a framework of smoothly jointed rods,

loaded at CDE, and supported at A and B. If a = 60°, draw a diagram
giving the stresses in the rods. Shew that if tan <x = 3 there is no stress
in the rod FD. [8.]

t W
l w
t
w framework of light rods freely
23. The figure represents a stiff
j ointed. The framework is supported at B, C and weights of 3 and 4 lb.
are suspended from A, D. Draw a stress diagram for the figure and
state whether the stress in each rod is a tension or a thrust. [S.]

9-2
132 GRAPHICAL STATICS [vni
24. The figure represents a Warren girder, which is hinged at A and
supported by a smooth horizontal plane at B: equal weights are
attached to the points a, b, c. Draw the force diagram, and determine
which members are struts and which are ties. [I.]

25. Draw a force diagram for the frame sketched, which consists of
light rods freely jointed, is supported at A and B, and supports weights
of 1, 2 and 5 tons at the corners O, D and E respectively. [T.]

26. In the jointed frame of light rods shewn below, equal and
opposite forces are applied at A and B in the line AB. Draw a force
diagram for the frame, and state which members are in compression
and which in tension. [8.]

27. A parallelogram ABOD formed of heavy uniform rods freely

jointed at their extremities is suspended by the joint A, and kept from
collapsing by a weightless rod forming the diagonal BD; prove that the
thrust in BD equals (W+W) BD/A C, where W and W are the weights
of AB and AD respectively. Find also the magnitude of the action
at O. [L]
 EXAMPLES 133
28. A regular hexagon formed of six uniform freely jointed rods is
hung up by one corner and held in shape by horizontal forces R applied
at the upper ends of the vertical sides and similar forces S applied at
the lower ends. Shew graphically that B is five times S and hence find
where to place a light bar horizontally which will alone preserve the
shape. [C]
29. A regular hexagon ABODEF is formed of heavy rods. A, B, C,
E, F are joined to the centre 0 by light rods and the hexagon is hung
from A. Draw the force polygon and determine the thrusts in the
light rods. [I.]
30. A symmetrical arch of span a and height h is to be constructed
of straight massless jointed rods, to carry seven equal weights w at
horizontal distances Ja apart, in such a way that there shall be no
bending moment at any point of any rod. Shew how the form of the
arch may be determined by graphical construction, and prove that the
horizontal forces necessary to keep the ends in position are aw/h. [C]

ANSWERS
2. (i) l-61b., 6-41b.; (ii) 8 ft. from 4 . 3. 27-5 lb., 24 •51b.
4. 15-71b., 29-31b. 6. 161b., 31b., 191b.
6. lib., 161b., lib., 81b.
7. 2-5 tons at D, 1-5 tons at E. Thrusts 4I> = 2-88 tons, AC =
•58 ton, BE= 1-73 tons; tension BC= -58 ton.
8. Tensions AB = $W, BC=\/3W; thrusts AD=^/3W, CD = 2W,
BD=W. 9. P. 10. IF cos a secret.
11. Tensions 4 0 = 4.0=1-85 lb.; thrusts AB = AE = S lb., CD =
M61b., 5 0 = 1-85 lb.
12. Tensions AB=AG=AD= 1531b.; thrusts BC=BD= 1081b.
13. Tensions AB = AF=5W, BC=1-5W, CD--5W, EF=W;
thrusts BF = ±-33W, CF = -5W, DF=-87W.
14. Tensions AF = DE=BE=WI^/2, AB=W; thrusts CD = W,
BG=\W, AE = EC = DF=W/-y/2. 15. Thrust.
16. 8cwt. Tensions AB = ^\/3vwt., AC = § \/3cwt.; thrusts BC =

17. Tensions 4 5 = 5-625, 50=13-125, BD = 5cwt.; thrusts AD =

12-5, i)JE7 = 6-75, EC-U-il, BE=9owt.
18. Tensions AB, CD, EA; thrusts BC, AC, AD, DE.
19. Tensions 4 5 = 200, 50=57-7, 5Z>=100; thrusts DE= 57-7,
CD- 115-4, 5.67 = 230-8lb. 20. Thrusts 4 5 = 6J, 4 O = 7 } .
23. Tensions AE, ED; thrusts 4 5 , BC, CD, BE, CE.
24. Tensions Aa, ab, be, cB, cD, bE, aG, bF; thrusts BD, DE, EF,
FG, GA, cE, aF.
26. Tensions 4 0 , CE, ED, DB; thrusts FA, FC, FE, FD, FB.
27. (W2.4X)2 + TF' i! .4S 2 -2TFTF'45.4 J Dcos54Z))*/24O.
29. Tension 0 4 = 2W; thrusts 0 5 = 0 . F = 3IF, OC= 0E= W, where
W = weight of a rod.
 Chapter IX
FRICTION
9*1. Friction between two bodies in contact at a point was
defined in 3*7 as that part of the mutual reaction between the
bodies which lies in the tangent plane at the point of contact.
When two bodies are in contact over a plane area on the sur-
face of each body, the mutual reaction between them is in
general a force acting at some point of this area, and its com-
ponent, if any, in the plane of contact is called Friction.
The mutual reactions between bodies are passive forces in
the sense that they only exist because of other forces applied
to the bodies. Thus a body may stand alongside a vertical wall
and in contact with it without any pressure between them
until an external force presses the body against the wall and
then there is a mutual reaction between the body and the wall
and its amount is no more than is necessary to balance the
externally applied force.
Friction, being a component of the mutual reaction, is a
passive force. It prevents or tends to prevent the motion of one
body across the surface of another, i.e. the relative motion of
the points of contact.
9*11. Laws of Friction:
(i) Friction acts in the direction opposite to what would be
the direction of relative motion if the friction did not exist; or,
in the case of relative motion, the friction opposes the relative
motion.
(ii) The magnitude of the friction is always just sufficient to
preserve equilibrium or to prevent relative motion of the
points of contact, provided that sufficient friction can be
obtained.
There is a limit, however, to the amount of friction that can
be called into play; when this limit is reached the friction is
called limiting friction, and limiting friction is controlled by
the following further laws:
9-1-9-14] LIMITING FRICTION 135
(iii) Limiting friction between two bodies bears a constant
ratio to the normal reaction between them; i.e. if R denotes
the normal reaction, then pR denotes the limiting friction,
where JU. is a constant depending on the nature of the materials
in contact and called the coefficient of friction.
(iv) Limiting friction is independent of the areas in
contact.
(v) When motion takes place the foregoing laws of limiting
friction are still true, but the coefficient of friction /x is slightly
less for bodies in motion than for the same bodies at rest; and
the friction is independent of the velocity.
9*12. The foregoing laws are based upon experiments. It
must not be supposed that they are rigorously true under all
conditions. If the pressure between two bodies were increased
to such an extent as to crush the parts of the bodies in contact,
the coefficient of friction would not remain invariable.
9*13. Experimental Verifications. A box resting on a horizontal
plane is fastened to a string which
passes, at first horizontally, over a
smooth pulley and carries a scale
pan.
Weights W are placed in the box / \
and a gradually increasing load P of
sand or other suitable material can
be placed in the scale pan and sub-
sequently weighed.
There is then a normal reaction
R equal to W, and a frictional force
F equal to P. The load P can be increased until the box just slides, and
a verification obtained that in limiting equilibrium F/R is independent
of R. By using boxes of different sizes it can be verified that /x is in-
dependent of the areas in contact. Results so obtained will not be very
precise because it is impossible to elimi-
nate friction entirely from the pulley.
9*14. A second method of experiment-
ing is to place a box containing weights
on a horizontal plane which can be tilted
about a horizontal line. It will be found
that no matter what may be the total
weight W the angle A through which the
plane can be tilted before sliding begins
is always the same for the same two vW
136 FRICTION [IX
substances in contact. Then if B be the normal reaction and F the
friction, by resolving along and perpendicular to the plane, we get
^=TfsinA and i?=TFcosA,
so that FjR = tan A.
It follows that the coefficient of friction is equal to the tangent of the
angle of inclination of the plane to the horizontal when the equilibrium
is just about to be broken by sliding.

9*15. Angle of Friction and Gone of Friction. When two

bodies in contact at a point are in limit-
ing equilibrium so that the greatest pos-
sible amount of friction is called into play,
the angle which the resultant reaction
makes with the common normal at the
point of contact is called the angle of
friction.
If R denotes the normal reaction, then pR denotes the
limiting friction and, if A is the angle of friction,
tan A=fiR/R=p;
so that the angle of friction is the angle whose tangent is the
coefficient of friction.
It follows that in 9*14 the greatest angle through which the
plane can be tilted before the body slides is the angle of
friction as denned above.
The resultant reaction between the bodies is, as explained
in 9*1, a passive force called into play to balance external forces
and possibly other constraints. The question arises whether
this passive force can adjust itself so as to balance any given
force system, and the answer imposes two important limita-
tions:
(i) the resultant of the other forces acting upon either
body must pass through the point of contact of the two bodies;
(ii) the same resultant must not make with the common
normal at the point of contact an angle greater than the angle
of friction, because such a state would imply a component
force in the tangent plane whose ratio to the normal pressure
exceeded the coefficient of friction, and this is not possible in
a state of equilibrium.
9-14^9-17] CONE OF FRICTION 137
It follows that, if, with the common normal NN' to the
surfaces as axis, we construct a cone of
semivertical angle A, the angle of friction,
then if the resultant of the other forces
acting on either body falls within or along
the surface of this cone it can be balanced
by the reaction of the other body; but, if
it falls without the cone, equilibrium is
not possible because there is not enough
friction available to balance the tangential
component of the other forces acting on
either body. This cone is called the cone
N
of friction. '
9*16. Numerical Values. The values of the coefficient of friction
for various substances have been found by experiment and tabulated:
About
Timber on stone 0-4
Timber on timber 0-5 to 0- 2
Timber on metals 0-6 to 0- 2
Metals on metals 0-25 to 0- 15
Masonry on dry clay 0-51
These figures are taken from Rankine's Applied Mechanics. It will
be noticed that all the above values of n are less than unity. The only
substances stated by Rankine to have a coefficient of friction as great
as unity are
earth on earth, damp clay 1-0
and earth on earth, shingle and gravel 0-81 to 1-11
9'17. Rolling Friction. We saw in 5-1 that a system of coplanar
forces acting on a rigid body can always be
reduced to a single force acting at a specified
point, together with a couple. So far in this
chapter we have assumed that, when two
bodies are in contact, the other forces acting
upon either body apart from their mutual
action and reaction are equivalent to a single
force acting through the point of contact, and
therefore such as could be balanced by the
mutual reaction when sufficient friction can
w 'W+W
be called into play. It is easily conceivable,
however, that the system of' other forces' acting upon either body might
not be equivalent to a single force at the point of contact, but to a force
and a couple, and in that case the mutual action and reaction between
138 FRICTION [IX
the bodies must consist oi a, friction couple as well as a force in order to
maintain equilibrium.
Consider for example a solid uniform cylinder of weight W resting on
a horizontal plane, with a cord passing over the cylinder and having
weights attached to its ends whifch may pass through a slit in the plane.
Suppose that the weights suspended from the cord are at first equal
to w and that one of them is gradually increased until, when it becomes
w + w', equilibrium is just about to be broken by the rolling of the
cylinder. The external forces are all vertical and may be represented
by a force W + 2w + w' downwards at the point of contact A, together
with a couple of moment w'r, where r is the radius of the cylinder. The
reaction of the plane at A therefore consists of an upward force

and a couple of moment w'r, and this couple is called the friction couple.
If the cylinder were also acted upon by a force with a horizontal com-
ponent there would also be a friction force along the plane at A.
It is found by experiment that, when equilibrium is about to be
broken by rolling, w' varies directly as R and inversely as r, or that the
friction couple w'r is proportional to the normal pressure R.
The explanation of the existence of the friction couple is that the
cylinder is not perfectly rigid and the contact is really over a small area;
that, when rolling is about to begin, at points where the surfaces tend
to separate they also tend to adhere, and the normal reactions on the
cylinder at these points and at points where the tendency is to produce
compression are in opposite senses and so give rise to a couple. This
kind of friction is sometimes described as the friction of cohesion.
9*2. Problems. Problems about friction are of very varied
kinds. The following example may be taken as typical of a
kind of problem in which the coefficient of friction and the
configuration are given and it is required to find a certain
external force.
A body of weight W is placed on an inclined plane whose inclination a
is greater than the angle offriction; to find
the least force which will prevent the body
from sliding down the plane.
Let R be the resultant reaction of
the plane on the body. Since the body
tends to slide down the plane the fric-
tion on the body acts up the plane, so
that in limiting equilibrium R makes
an angle A (the angle of friction) with
the normal to the plane on the upper
side of the normal. But the normal to
the plane makes an angle a with the
vertical and a > A, therefore R lies be- 'W
9-17-9-22] PROBLEMS 139
tween the normal and the vertical. Let P be a force which will balance
R and W, and let P make an angle 8 with the plane.
Then, by Lami's Theorem, we have
P R W
sin (a —A) cos (a— 8) cos (8 — A)*
It is clear that the values of P and R both depend upon 8, and that P is
least when cos (8 — A) is greatest, i.e. when 8 = A. So the least value of P
is IT sin (a — A), and it occurs when P is at right angles to R.

9*21. Another type of problem is one in which the data are

a given configuration in limiting equilibrium under given
external forces, to find the coefficient of friction.

Example. A uniform cylinder rests on a horizontal plane with its axis

horizontal. A plank lies across the cylinder with one end resting on the
plane. Shew that if the plank is just about to slide on the cylinder the angle
offriction between them is half the inclination of the plank to the horizontal.
Let the figure represent a vertical section through the central line
of the plank. AB denotes the plank
touching the cylinder at D which also
touches the horizontal plane at E.
There is assumed to be such symmetry
that all the forces act in the vertical
plane of the diagram.
Consider the equilibrium of the
cylinder. It is acted upon by three
forces: (i) its weight in the vertical line
CE, (ii) the reaction of the horizontal
plane through E, and (iii) the reaction of the plank at D. Since the
first two of these forces pass through E, the third force must also pass
through E. Therefore the resultant reaction between the plank and
the cylinder acts along DE, but in limiting equilibrium it is inclined
to the normal DO at the angle of friction. Therefore ODE is the angle
of friction and this is easily seen to be half the angle DAE.
We notice that this result is independent of the weight and length of
the plank, so that it would have been less simple to try to obtain a
solution by first considering the equilibrium of the plank as this would
inevitably have introduced several unknown quantities which would
have had to be eliminated.

9*22. Indeterminateness of Problems. Problems con-

cerning the reactions between rough surfaces are indetermi-
nate when more friction is available than is necessary to
maintain equilibrium.
140 FRICTION [IX
Consider for example the case of a uniform rod with its ends resting on
two equally rough inclined planes, in a vertical plane at right angles to the
line of intersection of the inclined
planes.
Let AB he the rod. Let the
planes make angles a, (S with the
horizontal and let the normals
to the planes at A and B in-
tersect at N. Draw lines at A
and B on both sides of the nor-
mals making with them angles
equal to the angle of friction A;
viz. the lines AML, AHK and
BHM, BKL. Then so far as
availability of friction goes, the
reaction at A may lie anywhere
within the angle KAL, and the
reaction at B anywhere within
the angle MBL: so that these
reactions could intersect at any
point within the quadrilateral
HKLM.
But since there is only one other force acting on the rod, namely
its weight acting vertically through its centre of gravity O, therefore
the reactions at A and B must intersect on the vertical through 0.
The only part of this vertical line which falls within the area HKLM is
the part PQ in the figure. It follows that for the given position of the rod
the reactions at A and B are indeterminate to the extent that they may
intersect at any point of the vertical segment PQ, and their ratios to the
weight of the rod are then given by Lami's Theorem. It is clear that
the position of the rod would not be a possible position of equilibrium
if the vertical through O did not
meet the area HKLM.
In a position of limiting equi-
librium the rod must be just
about to slide down one plane
and up the other, so that the
reaction on the rod must be
above the normal at one end and
below it at the other, i.e. the
figure must be such that the ver-
tical through ©passes through a
corner M or K of the quadri-
lateral.
Suppose that the rod is just
about to slide down the plane
AC and up the plane CB, then the reactions at A and B intersect at
M on the vertical through O, and if AN, BN are the normals at A
9-22-9-31] INITIAL MOTION 141
and B the angles MAN and MBN are both equal to the angle of
friction A.
Then since AN makes an angle a with the vertical, the angle AMO
is a — A, and for a similar reason BMO is /?+A; and the inclination 8 of
the rod AB to the vertical is given by
ainMAG _MG _MG _sinMBG
sin AMQ ~ A Q<~ BG ~ sin BMG'
s i n ( g - g + A)
01
sin(ot-A)
which reduces to
2cot 6—sin (2A + /5—a) cosec (a — A) cosec (j8 + A).

9*3. Initial Motion. Sometimes a system is in equilibrium

under the action of a gradually increasing force and it is re-
quired to determine how the system will begin to move when
the conditions of equilibrium are no longer satisfied. When the
circumstances are such that equilibrium is broken either by
the turning of one body about a point of contact with another
or by the sliding of the one over the other, then the former will
happen if the forces acting on the body reduce to a single force
at the point of contact falling within the cone of friction at that
point. That this is so follows from the fact that when the body
turns the friction required to prevent sliding is less than the
maximum friction available, whereas if sliding took place the
maximum friction would be called into play; and it is implied
in the definition of friction as a passive force that the amount
of friction called into play is never more than is necessary to
prevent sliding.
We may summarize this conclusion by saying that, when it
is a case of rolling or sliding, then rolling takes place, if there is
sufficient friction to prevent sliding.
The method of solving such problems is therefore to assume
that equilibrium will be broken by the turning of a body about
a point and examine whether the reaction at the point lies
within or on the cone of friction at the point. If it does so, the
assumption is justified. If not, then equilibrium will be broken
by sliding.
9*31. E x a m p l e s , (i) A heavy cubical block of edge 2a is placed on a
rough table with one face parallel to the edge of the table and at a distance
a cot a from it; to the centre of this face a light smooth rod of length I is
142 FRICTION [IX
freely jointed, it passes over the smooth edge of the table and carries a weight
W at its end. Shew that as W is increased the equilibrium of the block is
broken by its tilting about an edge if
I cos a sin2 a ,-„ -.
^ a + l sin a cos a (sin a — cos a)"
Let G be the edge of the table and DOB the rod jointed to the block
at B. The forces acting on the
rod are the weight W at D,
the reaction at G at right
angles to the rod since the
edge of the table is smooth,
and the reaction at B. The
two former meet in a point E,
therefore the reaction at B
on the rod is along EB. Let
EB cut the vertical through
the centre of gravity G of the
cube in F. Then since two of
the three forces acting on the
cube, viz. its weight and the reaction at B, pass through F, therefore
the third force, the reaction of the table, must also pass through F. And
from 9*3 equilibrium will be broken by turning about A if the force
along AF falls within the cone of friction at A, i.e. if

Now CA = a cot a, therefore BCA = a; and DB = I, therefore

DC = 1 — a cosec a;
and, if CBE=$,
, _ EC __(l — a cosec a) cot a
CB a cosec a
l
= - cos a — cot a.
a
Whence we find that
, .. I cos2 a sin a — a
t a n (vf6 —a)— , . „2 .
^ ' Zsin acosa
Now

l-tan(jS-a)
_ I sin2 a cos a
~ a + l sin a cos a (sin a — cos a)'
and, if /J. is greater than this expression, as W increases the block will
turn about the edge through A. If p has a smaller value than the fraction
stated, the block will eventually slide.
(ii) A right circular cylinder, radius a, whose centre of gravity is at a
distance cfrom the axis, is placed in the angle between a horizontal and a
vertical plane, so that its axis is horizontal. If the planes be of equal rough-
9-31] EXAMPLES 143
ness, find in what positions it will roll, and in what positions it will slip,
and shew that it will just not be able to slip at all if

where fi is the coefficient of friction. [S.]

Taking a vertical section through the centre C of the cylinder, let
A, B be the points of contact with the
horizontal and vertical planes. Draw
lines APQ, ASB, BBQ, BSP on both
sides of the normals AC, BC, making
with the normals angles equal to the
angle of friction A.
Then since
CAP=CBP = X, and CAB=CBB = X,
therefore the points P and B lie on the
circle ABC, i.e. the circle on AB as
diameter, so that
BPA - BBA = BCA = a right angle.
Now there are only three forces acting on the cylinder, its weight and
the reactions at A and B, and the latter cannot intersect outside the
quadrilateral PQBS if equilibrium is maintained. But this alone is not
a sufficient test, for it must also be possible for the reactions to be
directed so that they can have a vertical resultant.
Suppose that the line of action of the weight lies between B and B
in the figure. It has a counter-clockwise moment about C, but the
cylinder cannot roll in this sense, nor is there enough friction to main-
tain equilibrium, therefore the cylinder slides.
If the cylinder is placed so that the line of action of the weight passes
between B and C, it will be in equilibrium, because the reactions at A
and B can intersect on the line of action of the weight and balance it.
But if the line of action of the weight falls to the right of C, it is no
longer possible for the reactions to balance the weight although they
fall within the cones of friction at A and B. There will in fact then be
no reaction at B and equilibrium will be broken by the rolling of the
cylinder at A.
It follows that slipping will not take place unless the vertical
through the centre of gravity lies between B and B, and this will just
not be possible if
c = a — BB cos A
— a— BA sin ( ~ — A ) cos A
\4 /
= a — a (cos A — sin A) cos A
= a (sin2 A + sin A cos A)

where * = tan A.
144 FRICTION [IX
(iii) A uniform circular cylindrical log of radius a and weight W lies
with its axis horizontal between two rough parallel horizontal rails at the
same level and at a distance 2a sin a apart; shew that, if a gradually in-
creasing couple be applied to the log in a plane perpendicular to the rails
and axis, the log will turn over one of the rails when the couple is of
magnitude TFasina, provided the angle of friction A is greater than a.;
but otherwise the log will turn about the axis when the couple is
Wa sin A cos A sec a. [S.]
Taking a vertical section through the centre of the cylinder at right
angles to its axis, let C be the centre and A, B points of contact of the
cylinder with the rails. Then
A B = 2a sin a, so that the angle A GB = 2a.
Assuming that equilibrium is about to be
broken by turning about the rail at A, there
is then no reaction at B and the cylinder is
acted upon by the couple O, its weight W and
the reaction at A, and the latter must there-
fore be vertical and equal to W, forming with
the weight a couple Wa sin a, so that when O
exceeds Wa sin a the cylinder will turn over
the rail, but only provided that the reaction at
A falls within the cone of friction at A. The
axis of the cone is the normal AC and the re-
action at A is W making an angle a with AC, therefore the condition
is satisfied if A>a.
If A < a, the reaction at A, as determined above, falls outside the cone
of friction, so that the assumption of turning about the rail at A is not
justified, and we must assume that the cylinder remains in contact
with both rails.
There are now two methods of completing the solution. We may
draw the resultant reactions at A and B inclined to the normals at the
angle of friction A and meeting at O. The resultant of these reactions
together with the weight W must form a couple which, in limiting
equilibrium, balances the couple O.
Therefore, in fig. (ii),
G=W.MN = bW(BN-NA)
= \W {BO sin (a + A) - OA sin (a - A)}.
But O lies on the circle ABC, since the angle OBC = \-= OAC, and
the diameter of this circle is
AB/ainACB = a sec a;
therefore B0 = a sec a cos (a — A) and OA = a sec a cos (a + A),
and C = \Wa sec a {cos (a — A) sin (a + A) — cos (a + A) sin (a — A)}
= Wa sec a sin A cos A.
9-31, 9-4] FRICTION IK UNKNOWN DIRECTIONS 145
The alternative method of procedure is to put in the normal com-
ponents of reaction R and 8 at A and B and the frictions pR and pS,
as in fig. (iii).

Then the couple which, in limiting equilibrium, balances the couple

O is found by taking the sum of the moments of all the other forces
about any point in the plane. Taking moments about C we get

But by resolving horizontally and vertically, we get

fj, (R + S) cos a = (R — S) sin a,

Therefore (1 + JU.2) (R + S) cos a = W,

T ~, tJL(Z -nr
and (x = j-^1—2 " see a
»
or, since ^ = tan A, Q=Wa sec a cos A sin A.

9*4. Friction in Unknown Directions. A different kind of

problem from that discussed so far is one in which the direc-
tions in which the different points of a body begin to move
when equilibrium is broken are not known a priori but have to
be determined. The directions in which the frictions act are
therefore also unknown. The method of procedure is to assume
some position for the instantaneous centre of rotation* of the
body which is about to move. The direction of motion of each
point is then at right angles to the line joining it to the instan-
taneous centre and the frictions therefore also act at right
angles to the lines joining points to the instantaneous centre.
We then have the usual three equations of equilibrium and,
• v. Dynamics, 5-41.
146 FRICTION [IX
when the problem is a determinate one, these serve to deter-
mine the position of the instantaneous centre of rotation and
a relation between the forces.
9*41. Examples, (i) A uniform heavy rod A B lying on a rough table
has a force applied at the end A which is gradually increased until the rod
is just about to move in such a way that the initial motion will be rotation
about a point C in AH. Prove that the force at A must be applied at right
angles to AB and that the point C must divide AB so that
. [S.]

IB
x dx

Suppose that the rod is just about to rotate about C in the counter-
clockwise sense. Then the friction at every point must act so as to have
a clockwise moment about C. If dx denotes the length of an element
at a distance x from A, its weight may be denoted by wdx, so that the
vertical reaction of the table upon it is wdx and the friction acting upon
it is pwdx, where /x is the coefficient of friction. The only forces acting
on the rod in the horizontal plane are the frictions and the force applied
at A, and since the frictions are all at right angles to AB, therefore in
equilibrium the force at A must also be at right angles to AB.
The position of C may then be determined by taking moments about
A. Let AB — l, and AC = a, then noting t h a t the frictions o n i C and
CB are in opposite senses, we get
fa fl
I ixwxdx= I p.wxdx,
JO Ja
or i/j.wa2 = iix.w(P — a2),
so that V2a = l.
Therefore AC:CB = a:l-a

=V2+1:1.
We observe that the three equations of equilibrium are made use of
as follows:
(a) a resolution in the direction AB leads to the fact that the force
applied at A has no component along the rod;
(/?) the equation of moments about A determine the position of the
instantaneous centre O;
9-4, 9-41] EXAMPLES 147
(y) a resolution at right angles to the rod will now determine the
least force at A that will just move the rod, say
a (I

/
uwdx— I
0 Ja

(ii) A uniform rod AB is lying in a horizontal position upon two

parallel horizontal equally rough rods and is at right angles to them. The
points of contact are C and D, C being nearest to A. A gradually increasing
force is applied to the rod AB at A parallel to the other two rods. Shew
that if T g < - r = + - r f r equilibrium will ultimately be broken by the rod
jQJli J\\J jQlS
turning round D and slipping at O. [S.]

Let W be the weight of the rod AB acting at its middle point G.

Then by taking moments about C or D we find that the forces of sup-
port at C and D are TF 7=715 and W-^j respectively.
If the rod were to turn about any point but C or D the maximum
amount of friction would be called into play at both O and D, and if it
is possible for there to be a state of limiting equilibrium in which an
amount of friction smaller than the maximum is called into play at C
or at D this will be the actual state.
Suppose then that when the force at A is P, the friction at C is
k. W jrj=: and that at D is k'. W -^y > these forces being at right angles to
AB, the former opposed to P and the latter in the same sense as P in
the equilibrium state.
By moments about A we have

(1),

and, by resolving at right angles to AB, we have

lr w ®® h' W —P (2\
lc w K
- CD CD~ >'
It follows from (1) that k and k' axe always in a definite ratio and then
from (2) that k and k' increase with P. Hence, if fi be the coefficient of
friction, as P increases, equilibrium will be maintained until k or k'
reaches the value p, andif k is thefirstto become equal to JLC, equilibrium
148 FRICTION [IX
will be broken by sliding at C and turning at D. This will be so if h > k',
i.e. if CG.AD>QD.AC,
or if (IAB-AC)AD>(AD-IAB)AC,

OTif
^ ^
(iii) A uniform triangular lamina rests on a rough horizontal plane,
supported upon short pegs of equal length at its corners. To find the least
couple which will cause the lamina to turn in its plane.
It is easy to shew that three equal parallel forces acting at the
corners of a triangle have a resultant
passing through the centre of gravity,
so that if W be the weight of the tri-
angle the pressures on the pegs will
each be $W.
In limiting equilibrium the forces
due to friction just balance the re-
quired couple, and their directions
are at right angles to the lines joining
the corners A, B, C to the instan-
taneous centre of rotation.
Hence if there is an instantaneous centre I inside the triangle, so
that slipping takes place simultaneously at A, B and C, the frictions
are all limiting and equal to JftTF acting at right angles to IA, IB, IC.
Since these forces are equivalent to a couple, they would balance one
another when turned through a right angle to act along IA, IB, IC;
but they are equal forces, so by Lami's Theorem the angles AIB, BIG,
CIA are equal. Hence the position of I can be found by drawing on AB
and BC segments of circles containing angles of 120°. The couple
required is then the sum of the moments about I of the frictions, i.e.

It is easy to see that the frictions could not be equivalent to a couple

if the instantaneous centre were outside the triangle, so that the above
solution would not hold good if the triangle had an angle greater than
120°.
To consider the possibility of the triangle turning about one corner,
say the corner A. The frictions at B and C would then each be £ft W, and
that at A should be less than the limiting friction, i.e. less than \ft.W,
and the three should be equivalent to a couple; so that, as above, when
turned through a right angle they should be three forces acting at A in
equilibrium. But the resultant of ipW along AB and ^^W along AC is
cosjA and this can only be in equilibrium with a force less than
, if cos \A is less than J; i.e. if A > 120°.
It follows that the lamina will only turn about a corner if the angle
at that corner is greater than 120°, and if this be the angle at A, the
least couple required is
9-41, 9-5] MISCELLANEOUS PROBLEMS 149
9*5. Miscellaneous Problems, (i) A cart-wheel.
Consider a wheel which turns upon an axle whose circular section
is only slightly smaller than the aperture in the wheel, so that the
wheel and axle may be regarded as
having a common centre G.
Let a be the radius of the wheel
and b that of the axle.
Let A be the point of contact of
the wheel with the ground and sup-
pose the wheel about to move from
left to right in the figure. There are
only three forces acting upon the
wheel:
(a) its weight acting along GA;
(/?) the reaction of the ground
at A, acting on the left of the
normal AC, because the friction
opposes the tendency to slide at
A. (The friction at A is not limiting
friction.)
(y) The third force is the pressure of the axle on the wheel, and
since this force must balance the other two acting at A, therefore it
acts along BA, where B is the point of contact of the wheel and axle,
and B must be on the left of GA.
Also since the friction between the wheel and axle is limiting friction,
therefore BA makes an angle with the normal OB equal to the angle
of friction A. Then the angle 9 between BC and GA is given by
sin (A — 8) = - sin A.
Also, if P be the horizontal force applied to the axle in the direction
of motion, R the reaction between the wheel and axle and W the load,
then P = i?sin(A-e) and TF=.Rcos(A-0).
(ii) A window sash with a broken cord.
Let ABCD represent a loosely fitting
window sash of breadth a and height b.
The cord at B is broken and that at A
passes over a pulley and carries a counter-
poise of weight %W, where W is the
weight of the sash. It is required to find
the least coefficient of friction p that will
prevent the sash from sliding down.
Since the sash is loosely fitting there
will be contact with the frame at opposite
corners B and D only, and by resolving
horizontally we find that the reactions
there must be equal and opposite, of mag-
nitude R say.
150 FRICTION [IX
The friction forces fiR act vertically upwards at B and D. Then
neglecting the slight deviations of the sides of the sash from the vertical
and taking moments about B, we get
\Wa- W\a + fiRa- Rb = 0,
so that fj, = b/a.
(iii) Braking of a carriage on an inclined plane.
A carriage stands on an inclined plane with its axles horizontal. To
determine which pair of wheels should be locked in order to admit of the
greatest possible gradient for a given coefficient of friction.
Let O be the centre of gravity of the carriage. Assuming the carriage
to be symmetrical about the ver-
tical plane through O at right
angles to the inclined plane, we
can treat the problem as a plane
one. Let A, B be points of contact
of the wheels with the inclined
plane, let the vertical through O
cut the plane in C and let a be the
inclination of the plane to the
horizontal.
Let R and S be the normal reactions at A and B. In limiting equili-
brium the friction between the locked wheels and the plane will be ju,
times the normal reaction, and the friction between the wheels that
are free to rotate and the plane will be negligible in comparison. Hence
the friction is /xi? or pS according as the wheels at A or B are locked.
By taking moments about O for the equilibrium of the carriage, we
8et R.AC = S.CB.
Hence R is greater or less than S according as AC is less or greater
than OB.
But the friction pR or faS balances the resolved part of the weight
down the plane, viz. TFsina, where W is the weight of the carriage.
Hence the greatest possible gradient for a given fi corresponds to the
greater value of R or S. Therefore the wheels at A or at B should be
locked according as A O or CB is the smaller.
(iv) A pair of compasses.
A pair of compasses is used to describe circles on a horizontal piece of
paper, and has legs (of equal length I) always in a vertical plane. The
handle at the joint is always vertical. If$be the angle of friction between
the com/pass pencil and the paper, 2a the angle between the legs, and W the
vertical pressure on the handle, shew that a horizontal couple Wl sin a tan <j>
must be applied to the handle in drawing a circle, the joint being supposed
to be clamped.
When the joint is clamped, the angle 2a at the joint is capable of being
increased elastically by a small amount kB when a bending moment B is
applied to the joint in the plane of the legs. When a circle is about to be
9-5] PROBLEM OF A PAIR OF COMPASSES 151
drawn as above, it is found that this increase in 2a takes place as the
weight W is applied, provided a.>cj>. If a<<f>, the increase takes place
just after the start. Shew that, as a result, the radius of the circle drawn in
the latter case is increased by ^kWl* sin 2a. [T.]
Let A be the joint, O the centre of the circle and P the point of the
pencil. By taking moments about O in the vertical plane we find that
the vertical reaction at Pis ^W. It follows that when P is describing
the circle the friction is %Wt&n (j> along the tangent to the circle, and
since the radius of the circle is 21 sin a, therefore, by moments about 0,
the horizontal couple necessary to overcome the friction is Wl sin a tan <j>.
Next consider the case in which the joint can yield elastically to a
bending moment. Suppose that a vertical pressure W is applied to the
handle; then the vertical reaction at P is \W as above, and before any
horizontal couple is applied the joint tends to yield and the pencil
point P tends to slip outwards along OP.

t a n (f>

The friction F is therefore along PO and, by taking moments about

A for the leg AP, we get
Fl cos x = $Wl sin. a, or F=^W tan a.
This is the amount of friction necessary to maintain equilibrium; but
the maximum amount available is %W tan cj>, since <j> is the angle of
friction; therefore, if a > ^ , equilibrium will be broken by the pencil
point slipping outwards along OP while the pressure W is being
applied.
Next suppose that <f>>tx; then sufficient friction will be called into
play to prevent the pencil point from slipping along OP, and no motion
will take place until a horizontal couple is applied to make P move at
right angles to OP. As soon as this motion begins the friction ceases to
act along PO and acts at right angles to PO opposing the motion. The
only force at P in the plane OAP of the legs is now the vertical reaction
\W and this produces at A a bending moment

The joint then yields so that the angle OAP becomes 2a. + kB, and
the radius of the circle becomes
152 FRICTION [IX
Expanding this sine and neglecting higher powers of k than the first,
we get for the radius of the circle
21 sin a + 21. \kB cos a;
so that the increase in the radius is
ikBooaa. or ikWl2ain2x.

EXAMPLES
1. A particle rests on a rough horizontal plane under the action of a
horizontal force P. A gradually increasing horizontal force is applied
to it perpendicular to P. Determine the direction in which the particle
will begin to move. [S.]
2. Find the magnitude and direction of application of the least force
necessary to drag a particle up a rough inclined plane. [S.]
3. Prove that the least force, which applied to a uniform heavy sphere
of weight W will maintain it in equilibrium against a rough vertical
wall, is IF cos A provided A the angle of friction is less than

oo^pi. p.]
4. A circular hoop hangs over a horizontal peg and a weight 1-5
times that of the hoop hangs tangentially from it. Shew that if the
hoop is about to slip on the peg the coefficient of friction is -75. [I.]
5. A solid hemisphere of weight W rests in limiting equilibrium with
its curved surface on a rough inclined plane, and its plane face is kept
horizontal by a weight P attached to a point in its rim. Prove that the
coefficient of friction is p
. [S.]
VW(2P+W)
6. A heavy uniform rectangular block whose base is a square of side
2a and whose height is 26 stands on a rough horizontal plane of coeffi-
cient fj.. A horizontal force is applied to the middle point of the top of
one of the vertical faces and is increased until equilibrium gives way;
shew that this happens by the block tilting about an edge of the base
or by slipping according as p is > or < -^r • [S.J
7. A square is placed with its plane vertical and one side resting
along the line of greatest slope on a rough inclined plane, the coefficient
of friction being p, a string is attached to the upper corner and pulled
in a direction parallel to the line of greatest slope up the plane; shew
that if the tension be gradually increased the square will slide or tilt
according as the angle of inclination of the plane is less or greater than
tan- 1 (l-2 A i). [S.]
 EXAMPLES 153
8. A sphere rests on a rough inclined plane, the coefficient of friction
being p, and is supported by a horizontal string attached to the highest
point of the sphere and to the plane. Find the inclination of the plane
when the sphere is on the point of slipping. [S.]
9. Two weights, Wt and W2, rest on a rough plane inclined at an
angle a. to the horizon, being connected by a string which lies along a
line of greatest slope. If ^, fj,2 a r e their coefficients of friction with the
plane, and ^ > tan a > ft2, prove that, if they are both on the point of

10. Spheres whose weights are W, W rest on two inclined planes.

The highest points of the spheres are connected by a horizontal string
perpendicular to the common horizontal edge of the two planes and
above it. If the coefficients of friction JX, JX are such that each sphere is
on the point of slipping down, then
n'W'. [S.]
11. A uniform straight rod rests in a vertical plane with one end
resting against a rough vertical wall and the lower end on a rough
horizontal plane. If the friction is limiting at both ends when the in-
clination to the horizontal is a, and the coefficient of friction is the same
for both contacts, prove that the angle of friction is -g — -=. [S.]
12. A heavy circular hoop is hung over a rough peg. A weight equal
to that of the hoop is attached to it at a given point. Find the coefficient
of friction between the peg and the hoop so that the system may hang
in equilibrium whatever point of the hoop is placed in contact with the
peg. [S.]
13. A cylindrical hole of radius a is bored through a body and the
body is suspended from a rough horizontal peg passing through the
hole. Prove that in equilibrium the inclination to the vertical of the
plane through the centre of gravity of the body and the axis of the hole
is not greater than sin-1 —3—, where A is the angle of friction and d is
the distance between the centre of gravity and the axis of the cylinder.
14. A hoop of mass M hangs from a rough peg in a vertical plane.
An insect of mass m starts from the lowest point of the hoop and crawls
slowly upwards. Prove that the insect can reach the peg if sin A > -vy ;
but that the hoop will slip on the peg when the insect has traversed an
arcual distance a A+sin-1) sin A) if sinA<^ F -—. A is the
|_
angle of friction between the\ hoop /J peg, and aMis+the
m and the m radius of
the hoop. [S.]
154 FRICTION [IX
15. A semicircle, of radius a, is fixed with its diameter on the ground
and its plane vertical. In the same vertical plane a uniform heavy rod,
of length I, rests touching the semicircle with one end on the ground.
The coefficient of friction between the rod and both the ground and the
semicircle being tan" 1 A, shew that 6, the inclination of the rod to the
horizon in the limiting position of equilibrium, is given by
= osin2A. [S.]
16. A uniform rod of weight W and length 26 carries a weight P at
its upper end and passes over one peg and under another, the pegs
being at a distance a and the line joining them making an angle a. with
the horizontal. Prove that when sliding is just about to take place, the
distance of the upper end of the rod from the point midway between
the pegs is .„ „ + a ^ncc (^ being the coefficient of friction). [S.]

17. A string of length I has two light rings fastened at its ends, these
slide upon a rough straight wire fixed at an inclination /? to the horizon;
a heavy smooth ring has the string passing through it, and hangs in
equilibrium; shew that the greatest possible distance between the
rings on the wire is —^ r-^, /i being the coefficient of friction
between the rings and the wire, and being greater than tan j8. [S.]
18. A uniform rod AB, whose weight can be neglected, is placed
with its middle point G at the highest point of a fixed rough horizontal
circular cylinder, the rod being at right angles to the axis of the
cylinder. Two weights W, W are now suspended from the ends A, B,
and the rod is rotated on the cylinder, without sliding, until it is in a
new position of equilibrium. Prove that this is impossible unless the
I W ~W
angle of friction is greater than - .„ . „ , , where 21 is the length of the
ct w ~\~ w
rod and a the radius of the cylinder. [I.]
19. A uniform rod of length 21 rests within a hollow sphere of radius
a in a vertical plane through the centre of the sphere. The sphere is
rough, the angle of friction being A. Shew that, if I < a cos A, the greatest
inclination to the horizontal at which the rod can rest is given by the
equation ^p _ a^ s m 0 _ a2 s i n ( 0 _ 2A) = 0.
How can the rod rest if I > a cos A ? [S.]
20. A heavy beam inclined at an angle a to the horizontal rests with
one end against a vertical wall and the other on the ground, the coeffi-
cient of friction in each case being tan A. The beam is in a vertical plane
perpendicular to the plane of the wall. Shew that if the beam is kept
from slipping down by a horizontal string of tension T attached to the
lower end, or by a vertical string of tension T' attached to the upper
end, then T'= Ttan(a + A). [S.]
 EXAMPLES 155
21. A rough chain lies across the ridge of a double inclined plane,
stretching down the line of greatest slope on either side of the ridge.
Prove that it will be in limiting equilibrium when the line joining its
two ends is inclined to the horizon at the angle of friction. [C]

22. Two beads A, B, whose weights are w1,wi, are tied to the ends of
a string, on which is threaded a third bead G of weight W. The beads
A, B can slide on a rough horizontal rod, whose coefficient of friction
with each bead is ft. If, when A, B are as far apart as possible, the
strings AC, BC each make an angle 6 with the vertical, prove that, if
Wl>Wa>
taxi 6 ^IM {I + 2 (wJW)}. [S.]

23. Three rods AB, BC, CD, of equal length (I) and weight, are
loosely jointed together at B and C. Two light rings are fastened at A
and D and slide on a rough horizontal rod AD. If f* is the coefficient of
limiting friction between the rings and the rod, shew that the greatest
length of AD is
(*!LJ). [C]

24. A uniform rod rests within a fixed vertical circle, subtending an

angle 2a at the centre; its upper end is smooth, and its lower end rough
(coefficient of friction = tan A); shew that the angle which the rod makes
with the horizon cannot be greater than 6, where
, „ sin A ro ,
tan6 r- ,. , , . [S.I
L
cos A + cos (A +o 2a) '

25. A uniform solid rectangular block leans against a rough vertical

wall, with one edge resting on a rough horizontal plane. If 8 is the angle
which the plane containing the two supporting edges makes with the
vertical, and A is the angle of friction for both edges, prove that, in
order that the friction may be limiting at both edges, the thickness of
the block measured perpendicular to this plane must bear to the dis-
tance between the edges the ratio of
tan0-tan2A to sec2A. [S.]

26. A uniform heavy elliptic lamina rests with its minor axis (26)
vertical on a rough horizontal plane. A string is attached to the centre
and is pulled horizontally in the plane of the lamina until the major
axis (2a) of the lamina is vertical. Shew that if there is no slipping the
coefficient of friction between the lamina and the horizontal plane
cannot be less than (a2 - 62)/2a6. [S.]

27. A cube of weight W and edge 2a rests on a rough horizontal

plane. A ladder of weight W and length 21 rests against a smooth
vertical face of the cube, making an angle a with the plane. If the cube
were fixed, a man of weight w could ascend the ladder to a vertical
156 FRICTION [IX
height h before it slips. Prove that, if he attempts to ascend, the cube
will turn round its further edge before he reaches this height, if
h > (Wa - W'l cos a) tan tx/w;
the coefficient of friction between the cube and the ground being
> a/21 sin a, and 21 sin a > a. [S.]
28. A uniform rod AB 24 in. in length is lying in a horizontal posi-
tion upon two parallel horizontal equally rough rods and at right angles
to them. The points of contact are C, D. AC is 11 in., AD 18 in. A
gradually increasing force is applied at A to the rod AB parallel to the
other two rods. Shew that the equilibrium will ultimately be broken
by the rod turning about C and slipping at D. [C]
29. A uniform rod AB of weight W and length I lies on a horizontal
plane whose coefficient of friction is p. A string is attached to B and is
pulled in a horizontal direction perpendicular to the rod. As the
tension is gradually increased, shew that the rod begins to turn about a
SI
point in it whose distance from A is approximately j ^ , and the tension

of the string is then about *t • [S.]

30. A uniform rod AB of weight W rests horizontally on two equally
rough supports at A and C. Prove that the least horizontal force
applied at £ in a direction perpendicular to BA which is able to move
the rod is ^pW or yiW ~ r, according as 36 is greater or less than 4a,
where AB = 2a, AC = b, and /j, is the coefficient of friction. [S.]
31. A uniform rod of length 2a rests on a rough plane whose inclina-
tion to the horizon (a) is greater than the angle of friction (A) and can
turn about a point in it distant b from the centre of gravity of the rod.
Prove that for equilibrium the greatest possible inclination of the
rod to the line of greatest slope is
. tanA/6 + a \ ra
sin ' T ( , . [S.J
L J
t a n a \ o2ab J
32. A uniform square lamina ABCD is supported by three small
pegs rigidly attached to the lamina at the corners B, C and the middle
point of the side AD, the pegs resting on a rough horizontal plane.
Find the least force in the line AD that will move the lamina and find
the point about which it begins to turn. [S.]
33. A solid cylinder of weight w and of radius R rests with its axis
vertical on a rough horizontal plane. If the coefficient of friction be-
tween the surfaces in contact be /J., shew that the couple required to
rotate the cylinder about its axis is f pwR, assuming that the normal
pressure on the base of the cylinder is uniformly distributed over the
area of the base. Find also what couple will be required if it be assumed
that the normal pressure per unit area of the base varies inversely as
the distance from the axis of the cylinder. [S.]
 EXAMPLES 157
34. A uniform rod of mass M is placed horizontally on a rough
inclined plane of angle a ( < t a n - 1 / i ) . A string fastened to one end is
pulled downwards in the direction of the line of greatest slope, the pull
being slowly increased until the rod begins to move. Find the point
about which it will commence to turn, and shew that in that case the
tension of the string

where A = V l + tana/ft. [I.]

35. Two equal heavy uniform circular cylinders lie on a rough
horizontal plane and touch along a generator. A third equal rough
cylinder is placed on the top of the former and touching each along a
generator. Shew that in order that equilibrium may be possible in this
configuration the angle of friction at any point of contact cannot be less
than 15°. There is no action between the two lower cylinders. [I.]
36. Two rough planes are equally inclined at an angle a to the
horizontal. A cylinder of radius a, whose centre of mass is at distance c
from its axis, rests between them. If A be the angle of friction between
the cylinder and each plane (A < iv — a), shew that it is impossible for
the cylinder to be placed in any position of limiting equilibrium if
c<asinasin2A/sin2a. [S.]
37. Two rough planes which intersect in a horizontal straight line
are inclined at angles a and /} to the horizon. A cylindrical ruler whose
length is 2a and diameter 26 rests with one end on each plane, the axis
of the ruler being at right angles to the line of intersection of the planes.
Shew that, if the ruler be in limiting equilibrium with its axis inclined
at an angle 8 to the horizon,
cos (8 + <j>) cot (a + A) — cos (8 — <f>) cot (/3 — A) = 2 sin 6 cos </>,
where tan A is the coefficient of friction for each end and cot <f> = a/b.
[S.]
38. Two equal uniform rods AB, BG are freely hinged at B, C rests
on a rough horizontal plane and A is attached to a point above it.
When C is as far as possible from A for equilibrium, AB, BO make
angles a, ft with the vertical. Prove that the coefficient of friction
between the rod at C and the plane is 2/(tan a — 3 tan ft). [S.
39. Two uniform beams AB, AC of the same length are smoothly
hinged together at A and placed standing in a vertical plane with the
ends B . C o n a rough horizontal plane. Shew that, if the weight of one
is double that of the other, the least value of the coefficient of friction
necessary for equilibrium is £ tan %(BAC).
If the coefficient of friction has this value, at which end is the friction
limiting? [S.]
40. Two ladders of equal length but unequal weights, hinged to-
gether, form a step-ladder, the weights of the two parts being Wt and
W2 respectively; and equilibrium is maintained by friction between the
158 FRICTION [IX
ladders and the ground, the coefficient of friction being [i. Find the
inclination of the ladders at which slipping is on the point of taking
place; and calculate the action at the hinge when this is the case.
Assume the centre of gravity of each ladder to be at its middle point.
[SJ
41. Two equal uniform ladders are j ointed at one end and stand with
the other ends on a rough horizontal plane. A man whose weight is
equal to that of one of the ladders ascends one of them. Prove that the
other will slip first.
If it begins to slip when he has ascended a distance x, prove that the
coefficient of friction is (a + x) tan cc/(2a + x),a being the length of each
ladder, and a the angle each makes with the vertical. [S.]
42. A uniform sphere of weight W rests on a horizontal plane touch-
ing it at G. A uniform beam A B of weight X has its end A on the plane
and is a tangent to the sphere at B, ABC being a vertical plane. The
ratios of the tangential to the normal reactions at A, B, O in equili-
brium are a, /?, y. Shew that fi is greater than a and than y, and that
a > y i f W>X{1-cos BAG). [S.]
43. On the top of a fixed rough sphere of radius r rests a thin uniform
plank, and a man stands on the plank at this point. Shew that he can
walk slowly a distance (n + 1) re along the plank without its slipping off
the sphere, if the weight of the plank is n times that of the man, and e
is the angle of friction between the plank and cylinder. [S.]
44. A railway truck is at rest on an incline of slope a with the lower
pair of wheels locked. Shew that the coefficient of friction p between the
wheels and the rails must not be less than (a + b)/{h + b cot a), where h
is the distance of the centre of gravity of the truck from the plane of
the rails, and a, b are the distances of the centre of gravity from the
lower and upper axles measured parallel to the incline. [S.]
46. A light string, supporting two weights w and w', is placed over
a wheel (radius a) which can turn round a fixed rough axle (radius 6,
friction coefficient /x). There being no slipping of the string on the
wheel, shew that the wheel will just begin to rotate round the axle if
(w — w')a = (w + w'+ IF) 6 sin A,
where fi = tan A and W is the weight of the wheel. [S.]
46. A circular axle, of radius a, fits in a bearing whose radius is
slightly larger than a; shew that when the axle is just about to turn in
the bearing, the reaction between the axle and bearing will touch a
concentric circle of radius a sin A, where A is the angle of friction.
A uniform bar PQ, of weight W, rests in a horizontal position, being
free to turn about a horizontal axle, like the above, whose centre is at
a distance 6 from P, the nearer end of the bar. At P is hung a weight W
 EXAMPLES 159
which would maintain equilibrium in the absence of friction; prove
that if an additional weight X just starts P to move downwards
X=(W+W')(asinX)/(b-aainX). [S.]
47. A tripod, formed of three equal uniform rods freely hinged to-
gether at one end, rests with its feet at equal distances apart on a rough
horizontal plane for which the coefficient of friction is p. If the foot of
each rod is on the point of slipping, find the inclination of each rod to
the vertical. [S.]
48. The lower ends of three equal uniform rods, each of weight W,
rest upon the ground at the angular points of an equilateral triangle,
and their upper ends are loosely tied together at a point, from which
also a weight W is suspended. Shew that the coefficient of friction of
1 3W + 2W
the ground must not be less than s . Q W w / .tang, where 8 is the
inclination of each rod to the vertical. [S.]
49. Three uniform rods, VA, VB, VC, of total weight W, are con-
nected at V by a smooth weightless joint, and rest in the form of a
tripod on a rough horizontal table with their other ends at A, B, G,
points on the table. The projection of V on the table is O, and the angles
BOO, COA, AOB are respectively a, fi, y, also the rods make angles
8, ^, i/r respectively with the vertical. Prove that the frictions Flt F2,
F3 at A, B, C respectively are given by

JJ^=JA.=J^=1 K rci
L
sina sin;8 siny 2 sin a cot 8 + sin /} cot <f> + sin y cot xjt' 'J

ANSWERS
1
1. Making an angle cos" (P/F) with P, where F is the limiting
friction. 2. W sin (a + A) making an angle A with the plane.
8. 2A. 12. 1/V3. 19. At any inclination.
32. £/JLW ( 1 + ~7K) > where W is the weight. On the line bisecting AD
and BO at a distance \AB outside BO. 33. \^.wB.
39. The end of the lighter beam. 40. tan" 1 {(Wx + 3 W2)j( Wx + Wa)}
to the vertical, if Wt>W2. Horizontal ^ ( ^ + 3 ^ ) , vertical
TF2). 47. tan- 1 (2 /i ).
 Chapter X
CENTRES OF GRAVITY
10*1. The centre of gravity or centre of mass of a body was
defined in 4*51 as the centre of the parallel forces which repre-
sent the weights of the particles of which the body is composed;
and it follows from 4*53 that if m1, rn2, m 3 , ... are numbers
proportional to the weights, or to the masses, of a system of
particles situated at the points
(xi> Vi> zi)> (*2» V i , z 2 )> fe. Vz> za)> •••»
using rectangular axes, then the co-ordinates x, y, z of the
centre of gravity of the particles are given by the formulae
etc _

or x
10*11. Further, considering a scattered system of particles of masses
mx, m 2 , m3,... at points A1,A2,A3, ..., the weights of the particles are
proportional to the masses, and on a certain scale can be represented
by «!-!, m2, m8 The resultant of the parallel forces m^&\Ax and w2
at A2 is a force mx + m2 acting at a point B1 on A±At such that by 4'3

Then the resultant of m1 + ma at Bx and m3 at A3 is m1 + m2 4- m%

acting at a point B% on BtA3 such that

and so on.
It follows that what we have denned as the centre of gravity of the
system of particles is also what we defined in 2'6 as the centroid of the
points At, At, As, ... for multiples m1, ma, m3, ....
10*12. The analytical formulae « = 2ma;/Sm, etc., which
determine the centre of gravity of a set of particles, may be
used in a wider sense to determine the centre of gravity of a
set of bodies of finite size, provided that the typical 'm'
denotes the mass of such a body and the corresponding x, y, z
denote the known co-ordinates of the centre of gravity of that
body. This follows from the fact that the grouping together of
10-1—10-21] CENTRES OF GRAVITY 161
sets of particles to form finite bodies, as an intermediate step
in the process of finding the centre of the parallel forces which
represent the weights of all the particles, cannot affect the
position of that centre.
10*13. It is evident that applications of the standard
formulae of 10*1 are possible in which one or more of the wi's
may be negative. Such a case arises when a body contains a
cavity and is such that if the cavity werefilledup, the mass and
centre of gravity of the whole would be known. We may then
find the centre of gravity of the body containing the cavity by
considering the weight of the whole body as a positive force
from which has to be subtracted the weight of the matter that
fills the cavity, and this amounts to treating the m of the
cavity as a negative number in the formula.
10*2. Determination of Centres of Gravity. In a large number
of cases the positions of centres of gravity can be determined by simple
geometrical considerations.
A uniform thin straight rod. Let AB be the rod and O its middle
point. The whole rod can be regarded as composed of pairs of equal
elements equidistant from O; the centre of gravity of each such pair as
at P, Q in thefigureis then at the middle point of PQ, i.e. at O. There-
fore the centre of gravity of the whole rod is at its middle point.

A P G Q B

In what follows when we refer to the centre of gravity of a geo-

metrical figure, area or surface, a material sheet of small uniform thick-
ness with the specified boundary is to be understood.
10*21. A parallelogram. Let ABCD be the parallelogram. By
drawing a series of lines parallel to BC the whole area can be divided
into a series of narrow strips of
uniform width such that the A \K B
centre of gravity of each strip is, V \~
by the same argument as in 10'2, \ \
at its middle point. Therefore the E\ G\
centre of gravity of the whole \
figure lies on the line EF which g £^
joins the middle points of BO and **
AD, and passes through the middle points of all the strips parallel to
BO. By a similar argument, the centre of gravity lies on the line KL
which joins the middle points of AB and CD. Hence the intersection
of EF and KL is the centre of gravity.
162 CENTRES OF GRAVITY [x
10-22. A triangle. Let^BCbethetriangle. The area can be divided
into narrow strips by lines parallel to BC. Let be be any one of these
lines, let D be the middle point of BC and
let AD cut be in d. Then since be is parallel
to BC, the triangles Abd and ABD are
similar, as are Adc and ADC.
Hence, by using the similarity of the
triangles and the fact that BD=DC, we
have

so that bd — dc.
B D C
Therefore the centres of gravity of all
narrow strips parallel to BC lie on AD (the median). Therefore the
centre of gravity of the triangle lies on AD. Similarly, if E is the
middle point of AC, the centre of gravity lies on the median BE.
Therefore it is a t O where AD and BE intersect.
Again, join DE, then since DE is parallel to BA, therefore AQB,
DOE are similar triangles, as are ABC and EDC.
Therefore A G: A B = OD: DE,
or AG:GD = AB:DE-BC:DC
= 2:1.
Therefore O is at a point of trisection of the median AD. Similarly
BG:GE=2:1.
Cor. If F is the middle point of AB, G also lies on CF; but there is
only one centre of gravity, therefore the medians AD, BE, CF are
concurrent.
10*23. The centre of gravity of a triangle coincides with that of three
particlea of the same weight placed at its corners.
For, if we place particles of weight watA, B and C, the weights at
B and C have a resultant 2w at D, and the resultant of %w at D and w at
A divides DA at G so that 2DG= GA, i.e. at the point of trisection
which is the centre of gravity of the triangle.

10"24. A quadrilateral. There is no such simple method of con-

structing the centre of gravity of a quadrilateral as there is for a tri-
angle, but the following theorem is instruc-
tive as a method of procedure in a certain A g
class of problems.
The centre of gravity of a quadrilateral
coincides with that of four particles of equal
mass placed at the four corners together with
a fifth particle of equal but negative mass
placed at the intersection of the diagonals.
Let ABCD be the quadrilateral and O
the intersection of its diagonals.
10-22-10-25] TETRAHEDRON 163
Let m and m' denote the masses of the triangles ABD, CBD. Since
these triangles have a common base BD, therefore
m:m'=A0:0C.
Then by 10*23 above, the triangle ABD may be replaced by particles
of mass \m at each point A, B, D; and in like manner the triangle CBD
by particles of mass \m' at each point C, B, D.
Hence the centre of gravity of the quadrilateral coincides with that
of four particles: \m at A, \{m + m') at B, \m' at C and J (m + mf) at D.
Further, we may add particles of mass \mf at A and \m at C, so as to
make the four particles at A, B, C, D have the same mass £ (m +mf), if
we counterbalance the added masses by placing — £(m+m') at the
centre of gravity of \mf at A and Jm at G. But, since
m:m' = A0:0G,
therefore 0 is this centre of gravity, and we have therefore proved that
the centre of gravity of the quadrilateral coincides with that of four
particles each of mass \ (m + mf) &tA,B,C,D together with a particle
of mass — %(m+mf) at 0.
It will be noticed that though we have proved the proposition we
have not found the centre of gravity of the quadrilateral.

10*25. A tetrahedron. Let ABCD be the tetrahedron and H the

centre of gravity of the base
BCD,i.e. the pointwhich divides A
the median BE so that
BH=2HE.
Suppose the tetrahedron to be
divided into thin slices by planes
parallel to the plane BCD. Let
bed be any one of these sections
cutting AE in e and AH in h.
Then since AE is a median of
the triangle A CD and cd is paral-
lel to CD, therefore ce = ed and
be is a median of the triangle bed.
Again the parallel planes bed
and BCD are cut by the plane ABE in parallel lines be and BE so
that Abh and ABB. are similar triangles as are Ahe and AHE.
Therefore ^TT TT A
bh:hA = BH: HA
= 2HE:HA
= 2he:hA,
so that bh = 2he, and h is the centre of gravity of the triangle bed.
Similarly the centres of gravity of all thin slices parallel to BCD lie
on AH, and therefore the centre of gravity of the tetrahedron lies on
AH.
164 CENTRES OF GRAVITY
By a similar proof, if F is the centre of gravity of the triangle AOD,
the centre of gravity of the tetrahedron lies in BF. It is therefore the
point G in which AH and BF intersect.
Join HF. Then since
BH:HE=2:1=AF:FE,
therefore HF is parallel to AB and the triangle AOB, HOF are similar,
as are ABE and FHE.
Hence, by similar triangles,
AG:AB=GH:FH,
or AG:GH=AB:FH=BE:HE
= 3:1,
or GH = \AH;
i.e. the centre of gravity of the tetrahedron lies on the line joining the
centre of gravity of any face to the opposite vertex and divides it in
the ratio 1:3.
It is easy to verify that the centre of gravity of the tetrahedron
coincides with that of four particles of equal mass placed at the

10*26. A pyramid on a plane base. Let V be the vertex and

ABODE the base. By joining one corner of the base to each of the
others the base can be divided
up into triangles and the pyra-
mid into tetrahedra, and the
distance of the centre of gravity
of each tetrahedron from the
base is one-quarter of the height
of the pyramid j therefore the
centre of gravity of the pyramid
must be at a distance from the
base equal to one-quarter of the
height of the pyramid.
Again the pyramid can be
divided into thin slices by planes
parallel to the base, making
sections, such as abode, which
can easily be proved to be
similar to ABCDE; and if H is the centre of gravity of the base
ABODE, it follows from the similarity of the figures as in 10*25 that
VH cuts the plane abcde in a point h which is the centre of gravity of the
section abcde. The centre of gravity G of the pyramid therefore lies on
VH and, since its distance from the base of the pyramid is one-quarter
of the height, therefore
H G=
10-25-10-31] INTEGRATION 165
10*27. A solid cone with a plane base. A cone may be regarded
as a limiting form of a pyramid when the
number of sides of its base has been increased
and their length decreased without limit so
that the base is a continuous curve. Then
it follows from 10-26 that, if V is the vertex
and H the centre of gravity of the base, the
centre of gravity O of the cone lies on VH
and is such that

10*3. Centres of Gravity by Integration. The analytical

formulae aj = 2mx/2m, etc. may be adapted for finding the
centre of gravity of a continuous distribution of matter form-
ing a finite body of a given shape. We regard the body as sub-
divided into a large number of small parts of such shape that
the position of the centre of gravity of each part is known to a
first approximation. The formulae will then give approximate
values to the co-ordinates x, y, z of the centre of gravity of the
body; and the larger the number of subdivisions the more
accurate in general will be the result. Unless, however, we can
evaluate the sums Sma;, Hmy, Sm, the formulae are useless;
but if the number of subdivisions is increased the magnitudes
of the parts decreasing without limit the sums Smx, etc. be-
come definite integrals, which in many cases can be evaluated,
and, because in the integrals the small parts tend to vanishing,
what was in the first stage approximative now becomes exact.
Hence, if we use dm to denote an element of mass of a body and
x, y, z are first approximations to the co-ordinates of the centre
of gravity of dm, the formulae
_ Ixdm, _ [ydm _ \zdm
x=
~]dm' y=~]dm' Z = ~]dm'
where the integrations are taken through the body, determine
its centre of gravity.
The integrations may be along curves, or over surfaces or
throughout volumes according to the form of the distribution
of matter.
10-31. Curves. If matter be distributed along a plane curve
so that p is the mass per unit length at a point (x, y), then if ds
166 CENTRES OF GRAVITY
denotes an element of arc at (a;, y) we may write dm = pda, and,
with sufficient accuracy for our pur-
pose, x, y may be the coordinates of « Jis
any point of da, and the centre of
gravity of the whole distribution is
given by
-Jzpda -Jypda _

In the case of a uniform distribution, p is constant and

-Jyds
y
~ ids'
10*311. Circular Arc. Let A CB be a circular arc of radius a which
subtends an angle 2a at its centre O.
Take the axis of x along 00, where C is the middle point of the arc,
and a perpendicular axis Oy.
If P be any point on the arc and OPmakes
an angle 8 with Ox, the co-ordinates of P
are a; = acos# and y = asia.8; and, if PP' is
a small element of arc, we may write P'

PP' = ad6.
Therefore
acos 6. add

add
/:

and J
I'
y= —
asinB.adO
•=o.
I" add

Hence the centre of gravity O lies on the middle radius 00 (as is

obvious by symmetry) and OQ= asm a

10*312. Example. If O is the pole of the lemniscater^ — a2 cos 2$ and

O is the centre of gravity of any arc
PQofthecurve,provethatOObisects
\e angle POQ.
Let RR' be an element of arc ds,
where R is the point (r, 8).
Then, from the equation of the
curve, rdr=-a*ain.28d8.
10-31-10-32] AREAS 167
Also

Then, if POx=a. and QOx = p,

reosfl. — dO . „
i r _sin(8—sina
Jos
iaT
["rsin.B. — dd
_ }_a r _ cos a — cos /?
and
Ids /"0 a 2 , „ [PdO *
' —do I —
r Jar
C 3
Therefore I ? " —Csina
tan GOx =a;I = smjS O3
^=tani(a+i3);
zv
^"
a; smjS — sina
whence (?Oa; = £ (POa
or OG bisects the angle POQ.
10*32. Areas and Surface Distributions. In the case of
matter spread over a surface, if p is the mass per unit area at
a point (x, y, z) we may write dm = pdS, where dS is an element
of area of the surface and the formulae become
7Z
ipdS '
or, if the density be uniform so that p is constant,
-, etc.

The choice of an element dS depends on the nature of the

problem.
y
(i) For example to find the centre of
gravity of a plane area bounded by a curve
y=f{x), the co-ordinate axes and an or-
dinate x — a.
If P be a point (x, y) on the curve, a
convenient form for dS is a narrow strip o a x
of area ydx, and the co-ordinates of its centre of gravity,
168 CENTRES OF GRAVITY
with sufficient accuracy for our purpose, are (x, \y), so
that
xydx

ydx ydx
Jo Jo
where in the subjects of integration we make use of the equa-
tion of the curve y=f(x).
(ii) Sectorial area bounded by a curve r=f(8) and two radii
i
If P be a point (r, 6) on the
curve, it is convenient to take
for dS a narrow sector POP'
whose area is %rzd9 to the first
order of small quantities, and
the co-ordinates of its centre of
gravity, with sufficient accuracy
for our purpose, are
x=%rco&9, i/=frsin0.
CP
_JxdS §rcos0.fr2<Z0
Therefore Ja
X
~~W"
J! r 3 cos0d0

fJ a
rH9

ra sin 0 d6
and similarly

where in the subjects of integration we make use of the equa-

tion of the curve r=f(8).
10*321. Circular Sector. Taking a sector of radius a and angle
10-32-10-33] AKEAS 169
2a, with the figure of 10*311, we have r — a, so that the results obtained
in 10*32 (ii) become
__2 J -a _ 2 sin a

sin 6d6
- 2 -

r
and = 0.
dd

10-322. Semicircular Area

j—' and Quadrant of a Circle. Fora

semicircle a = Aw, so that .

<»•£•
Also the middle radius OC divides the semicircle
into two quadrants AOG and BOG whose centres
of gravity O1, Q% must be such that Ot O2 passes
through O and is parallel to A B. Hence the distance
4a
of Ot from OA is ^— and by symmetry its distance
from 00 is the same. So the centre of gravity of a
4a
quadrant of a circle of radius a is at a distance =- B
from each of its bounding radii. The same result is of course obtained
from putting a = \ir in the result of 10*321.
10*33. E x a m p l e s , (i) Find the co-ordinates of the centre of gravity of
a quadrant of the area of the curve x^ + y$ = a*,
bounded by the axes of co-ordinates. y
Let OA B be the quadrant whose centre
of gravity it is required to find.
Using 10*32 (i) we have
[a fa
I xydx
= _i_o -. J o
fa — /a

I ydx j ydx
J 0
Then, byJ substituting
o for y from the equa- Q
tion of the curve,
f f
I ydx— I {w—x'ydx,
Jo Jo
and by putting x = a sin 8 0
= 3o / ^ sin 0 cos Bd6= 3a2 [ U (cos4 6 - cos6 6) dB
a 2 4

Jo Jo
I ? 1\ ? 32
* v. Elementary Calculus, 6-6.
170 CENTRES OF GRAVITY
a [iir
/
xy dx = 3a s I sin 5 9 cos4 9 dd
o Jo
= - 3a s ( *" (1 - cos2fl)2cos* Bd (cos 0)
Jo

rro. r - 256O
Therefore x= .
We can calculate y in the same way, but since the equation of the
curve is symmetrical in x and y, it is evident that y=x.
(ii) Find the centre of gravity of a loop of the lemniscate r 2 = a2 cos 20.
The values of 0for whichr vanishes
are ±£w. ^^ JX p/
Hence by using 10*32 (ii) we have y^ ^-~^~^~^Z-\P

Now the denominator

= /-!-

and the numerator =1 \a* cos% 29cos 6d9,

J —if
or, by putting \Z2ainB = sin(l>, so that ->/2cos9d9=cos<j>d^,
1 3 [lv
the numerator = .a I cos* <f>d<j>
" V * 7 —i"
1 3 17a8
8V2'
and
It is evident by symmetry, or can be proved in like manner, that y = 0.
(iii) Find the centre of gravity of a sector
of a circle in which the surface density
varies as the distance from the centre.
Let a be the radius and 2a the angle
of the sector,4OB, and let p — Xr denote
the surface density, or mass per unit
area, at distance r from the centre 0.
Take the bisector of the angle AOB
as axis of x. By symmetry the centre of
gravity lies on this line.
In applying the formula

jpdS '
10-33-10-341] VOLUMES OF REVOLUTION 171
the choice of the element of area dS is limited by the consideration that
the value of p must be the same at all points of the element when its area
tends to vanish. But p has constant values along circular arcs of centre
0; therefore we take for dS a narrow strip bounded by circles of radii
r and r+dr, so that dS — 2ardr. Such a narrow strip may be regarded
as a circular arc of radius r, so that the x co-ordinate of its centre of
gravity is r 8 m a (10-311).
By substituting these values in the formula for x, we get
'rsina . „ ,
.Xr.2ardr .
_ ,„ a 3 asm a
X — =: — a

I Xr.2ardr
Jo
10-34. Students who are familiar with double and triple
integrals will recognize at once that the formulae of 10-32 for
plane surface distributions of matter can be expressed more
generally by the forms
Jlxpdxdy
x
~ ~ft~TZ~aZ > V ~

and in polar co-ordinates

__jjrcosd.prdddr _._ JJV sin B.prdddr
X
~ Hprdddr ' y~
Also that there are similar formulae in triple integrals for
finding the centres of gravity of volume distributions.
10-341. Volumes of Revolution.
To find the centre of gravity of a volume of revolution, i.e. a
solid whose boundary is obtained by revolving a given curve
y =f{x) round the axis ofx, cut off between plane ends x = a,x=b.
Since sections at right angles to the axis of x are circles, the
solid may be regarded as built up of thin discs of area iry* and
thickness dx, and the a;-co-ordinate of the centre of gravity is
given by
f6
I x.iry^dx
CC — -J rb
7S*"*"* a

Ja
and by symmetry the centre of gravity lies on the axis.
172 CENTRES OP GRAVITY [X
10*342. Find the centre of gravity of a segment of height h of a sphere
of radius a.
Let O be the centre of the sphere
and Ox an axis at right angles to
the base of the segment meeting it
in D and the curved surface in O, so
that DG=h and 0D = a-h.
Taking slices of the segment
parallel to its base and applying the
formula of 10*341,
I[" x.vyidx

But x, y are co-ordinates of a point on a circle of radius a, viz. the

section of the sphere by the plane of the paper; so that y3 = a2 — x2, and

_3(2a-fe) a
5 4 Za-h '
f
(cfl-x^dx
J a-h
Cor. Putting h = a,we find that for a hemisphere, x = | a .
10"343. Solid Octant of a Sphere. Con-
sider an octant OABC of radius a. Four such
octants, as in the figure, make up a hemisphere
whose centre of gravity G is on the radius OA
at a distance fa from its base. But the centres
of gravity of the four octants are all at the
same distance from the plane BGB'C and the
plane containing them must pass through G
the centre of gravity of the four. Hence and by
symmetry the centre of gravity of a solid octant
of the sphere is at a distance fa from each of its
plane faces.

10*35. Centre of Gravity of a Zone of the Surface of a

Sphere. Let X'OX be a diameter of a circle and PM, QN ordinates
to this diameter at neighbouring points P, Q on the circle, and let these
ordinates meet a tangent A'A parallel to X'X in P', Q'.
Now let the figure revolve round X'X. The circle generates the
surface of a sphere and the line A'A traces out a cylinder which circum-
scribes the sphere. The ordinates MPP', NQQ' will trace out planes at
right angles to X'X, between which planes lie a slice of the sphere and
the corresponding slice of the cylinder.
10-342-10-36] AREAS ON A SPHERE 173
The area of the narrow zone of the sphere traced out by the arc PQ is
io
2TTMP.PQ, and the area of the corresponding zone on the circum-
scribing cylinder is
2nMP'.P'Q'.
But
P'Q
PQ
MP MP
OP' Mh"
therefore
MP'.P'Q'=MP.PQ.
Therefore the areas of correspond-
ing narrow zones of the sphere and
the circumscribing cylinder are equal, and by addition the same
must be true for zones of any width. Therefore the area of the
surface of a sphere intercepted between parallel planes is equal to the
circumference of the sphere multiplied by the distance between the
planes.
Now to find the centre of gravity of such a zone of a sphere; let it be
divided into narrow zones by a large number of parallel planes at equal
distances apart, then the areas intercepted between consecutive planes
are all equal and their centres of gravity are uniformly distributed along
the axis offigure,so the problem is that offindingthe centre of gravity
of a uniform distribution of equal particles in the same straight line,
and it is midway between the end particles, i.e at the point on the
axis of the zone midway between its bounding planes.
Cor. i. The centre of gravity of the surface of a segment of a sphere
is at a distance from the centre of its base equal to half its height.
Cor. ii. The centre of gravity of a thin hemispherical shell is at the
middle point of its central radius.
Cor. iii. The centre of gravity of the curved surface of an octant of a
sphere is at a distance equal to half the radius from each of the three
perpendicular planes which bound the octant.
10*36. Areas on a Sphere. Let dS be a small element of area on
the surface of a sphere of radius a, dU its
projection on a plane through the centre O
and 2 its distance from this plane.
Then dll is equal to dS multiplied by
the cosine of the angle between dll and
the tangent plane at a point of dS, i.e.
dU = -dS.
a
Now if dS is a part of a finite area £ on the surface of the sphere,
174 CENTRES OF GRAVITY [x
we have for the centre of gravity of S
._SzdS_SadTI
Z
~ fdS ~ fdS
z II
a=S>
where II is the projection of S.
Cor. We obtain the result of 10-35, Cor. (it) at once by taking S to
be the surface of a hemisphere of area 27ra2, its projection IT being a
circle of radius rra2.
10*4. Applications of Orthogonal Projection. Among simple
properties offiguresobtained by orthogonal projection are the following:
(i) parallel lines project into parallel lines;
(ii) the ratio of two segments of the same line or of parallel lines is
unaltered by orthogonal projection;
(iii) the centre of gravity of a plane area projects orthogonally into
the centre of gravity of the projection.
The truth of the first two statements is evident from the figure in
which FAB and KCD are parallel lines on an inclined plane FEML

and Fab, Kcd are their projections on a horizontal plane FemL. The
projections Fb and Kd are parallel, for if they had a common point it
would be the projection of a common point of FB and KD. Also since
FB and KD are parallel they are equally inclined to the parallel lines
Fb and Kd, and then by similar figures
ab:cd-AB:CD.
To prove (iii), let S be any area on a plane inclined at an angle a to a
horizontal plane, and II the projection of S on the horizontal plane,
fig. (ii).
Let S be divided into small elements by drawing lines of greatest
slope and lines parallel to the intersections of the planes. It follows that,
if dll is the projection of an element dS, then dH = cos a. dS.
10-36-10-41] ORTHOGONAL PROJECTION 175
Also if x, y, z are the co-ordinates of the centre of gravity of S referred
to the horizontal axes Ox, Oy and a vertical axis as shewn in the figure,
and x, y, « are co-ordinates of a point of dS, then
- _ Sx d8 _ Sx sec a dll
x
~~]d8'~ Jsecadll
fxdll , . ., . . Jt/dII
= BdaiIBtaJ
"pT y Vs -jW

But x, y, 0 are co-ordinates of a point of dli, so that the last two

fractions are also the x, y co-ordinates of the centre of gravity of II and
so the centre of gravity of S projects into the centre of gravity of II.
10'41. Quadrant of an Ellipse. A quadrant of a circle can be pro-
jected orthogonally into a quadrant of an ellipse. Thus if OAB,fig.(i),
be a quadrant of a circle and it be projected on to a plane through OA,
all lines parallel to OA are unaltered in length, and all lines parallel to
OB are reduced in the same ratio.

O M A
(i)
Hence, if OM is an ordinate to OA from the centre of gravity, and
B' and 0' are the projections of B and O, we have
MG'_MG_ 4
~0W ~0B:~8S'
176 CENTRES OF GRAVITY [X
Therefore if 0A = a, OB' = b are the semiaxes of the elliptic quadrant
OAB' and x, y the co-ordinates of its centre of gravity O' referred to
OA, OB' as axes, we have
a b 3TT'

Again, in fig. (ii), by projecting the quadrant OAB of a circle on to a

plane which cuts the plane of the quadrant in any radius 00, we get a
quadrant OA'B' of an ellipse bounded by conjugate radii OA', OB',
and the ordinate MO which is parallel to OB projects into a line M'O'
parallel to OB', and
OM'OM 4

by similar triangles and 10*322; while

M'G'_MG_ 4
OB' ~ OB~3TT
(10-4 (ii) and 10'322).
Hence if OA' = a' and OB' = b' are the lengths of conjugate radii and
x', y' are oblique co-ordinates of O' parallel to the conjugate radii

10*5. Theorems of Pappus.

(i) / / a plane curve rotates through any angle about an axis in
its plane which does not cross it, the area of the surface traced out
by the curve is equal to the length of the curve multiplied by the
length of the path of its centre of gravity.
For if ds denotes an element of arc of the curve at a distance
y from the axis of rotation and 6 the angle
through which it turns, the elements traces y \ ds
out a narrow ribbon of length yd and area
y6ds. Therefore the area of the surface
traced out by the whole curve is represented

= 6y x s,
where s is the length of the curve and y is
the distance of its centre of gravity from the axis, and this
proves the theorem.
(ii) / / a plane area rotates through any angle about an axis in
its plane which does not cross it, the volume traced out is equal to
the area multiplied by the length of the path of its centre of gravity.
10-41-10-6] THEOREMS OF PAPPUS 177
For if dS denotes an element of the area at a distance y from
the axis of rotation and 9 the angle
through which it turns, the element dS
traces out a narrow tube of length yd,
cross section dS and volume yddS.
Therefore the volume traced out by the
whole surface is represented by

= 6yxS,
where 8 is the given area and y is the distance of its centre of
gravity from the axis.
10*51. Examples, (i) Tofindthe area of the surf ace and the volume
of an anchor ring obtained by rotating a circle of radius a about an axis
in its plane at a distance cfrom its centre, where oa.
From 10*5 (i) the area of the surface of the anchor ring

and from 10*5 (ii) the volume

(ii) Use the theorems of Pappus to find the centre of gravity of

(a) a semicircular arc,
(b) a semicircular area.
In each case let id denote the distance of the centre of gravity from
the centre, and a the radius of the given circle, and let a semicircular
arc and area make a complete revolution about its diameter.
In both cases the figure traced out in a complete revolution is a
sphere. Therefore
(a) %-nx x «z=area of sphere = 4n-a2,
so that x =—;
and (6) Ivx x \ 17a = volume of sphere = | iras,
2

so that x = =-.
on
10*6. Lagrange's Formula.
/ / G be the centre of gravity of a system of particles of masses
m1)m2,m3,... situated at the points A1,A2,A3,... and 0 be any
other point, then
178 CENTRES OP GRAVITY [X
For any point A, let ON be the projection of OA on the line
00.
Then
2m OA2 = 2m (OA2 + 002 + 200. ON)
o G
But if we take G as an origin and 00 produced as an axis of #,
so that ON is the x of the point A, then
2m ON = 2m a; = a 2m = 0,
since the centre of gravity is the origin: therefore
2m OA* = 2m OA2 + 002T,m.
This theorem is only a slightly modified form of the theorem
of parallel axes in the theory of moments of inertia*; it is useful
in metrical geometry.
10'61. E x a m p l e . A uniform triangular lamina ABC of weight W is
suspended from a fixed point by strings of
lengths Zx, Z2, l3 attached to its corners A, B, C.
Shew that the tensions in the strings are
WIJ{3 (Zj.2 + Zj2 + Z32) - a 2 - 6 2 - c 2 } 1
and similar expressions.
Let Tlt T%, Ta denote the tensions in the
strings OA, OB, OC. Since their resultant is
W acting in the vertical OG, where G is the
centre of gravity of the lamina, and G is also
the centre of gravity of particles of equal mass
at A, B, C, therefore, by 3*43,

K
_ ''
Butbyl0'6 i2 +
SOG

and, by completing a parallelogram of which BG, GC are adjacent

sides and observing that its diagonals are of length a and AG,

similarly
and
so that
Therefore 90 (?2 = 3 (Zx2 + Z22 + Z32) - a2 - 62 - c2,
and the required result follows from (1).

* v. Dynamics, 13'2.
10-6, 10-7] EXAMPLES 179
10*7. Miscellaneous Examples.
(i) The co-ordinates of the extremities of the bounding radii of a sector
of the ellipse x^/a* + y2/b* = 1 are xlt ylt and xit y2. Prove that the co-
ordinates of its centre of gravity are
&± y*-yi
3 6 * <j>
where abs,va.(fi=:x1yi—xiy1. [S.]
Let OPQ be the sector of the ellipse, OA, OB its semiaxes of lengths
a, b, and P', Q', B' the points on the auxiliary
B
circle corresponding to P, Q, B.
Then the ellipticfiguremay be regarded
as an orthogonal projection of the circular B
figure in which every ordinate has been
reduced in the ratio a: b; and if G' is the
centre of gravity of the circular sector
OP'Q'and G that of the elliptic sector OPQ,
G lies on the ordinate G'L and is such that
GL:G'L = b:a.
O
Now if a and a + ^ are the eccentric angles
of the points P and Q, i.e. the angles P'OA, Q'OA, the angle P'OQ' = <f,
and
0G' = §aS^& (10-321).
9
Also xx y2 — x%y± = ab {cos a sin (a + ^) — cos (a + <j>) sin a}
= aft sin ^;
identifying P'OQ' with the <j> of the question.
Then the co-ordinates of G are given by
x=0L = 0G' c
4 sin £$cos(a + £^)
= 3°
2 sin (oc -J- ©) — sin ot
3° 9
2a
"36

and y= = -aG'
L = - O(?'sin(a-f-\<j>)

_ 4 . sin %<j> sin (a + %</>)

2 1 cos a — cos (a + ^)
"3° 9

3a
 180 CENTRES OB1 GRAVITY
(ii) An isosceles triangular lamina is such that its mass per unit area
at every point is proportional to the product of the distances of the point
from the equal sides of the triangle. Prove that the distance of the centre of
gravity from, the vertex is four-fifths of the altitude.
Let OAB be the triangle in which
OA = OB, let 0 C = a be the altitude and
let the angle AOB be 2a. y
Take the axis of x along OG. Let P be
the point (a;, y) and PM, PN perpen-
dicular to O A, OB. Then it is easy to see Q
that
PM = x sin a — y cos a
and PN = x sin a + y cos a,
so that the mass per unit area is propor-
tional to xt sin 2 a_y2 COS2 a -

By symmetry the centre of gravity lies on Ox, and its distance from
O is given by
a fx tan a
a; (a;2 sin2 a — y* cos2 a) dx dy
/ 0J -xivai
(a;2 sin2 a — y2 cos2 ctjdxdy
-x tan a

(2x* tan a sin2 a — f x* tan 3 a cos 2 a) dx

t
i
(2a;3 tan a sin2 a — fa;3 tan 3 a cos2 a) dx

x*dx

x3dx
Jo
(iii) The moss per unit area at any point of a lamina bounded by a
quadrant of an ellipse and its axes is proportional to the nth power of the
linear dimensions of the similar and similarly situated ellipse which
passes through the point. Find the centre of gravity.
Let OAB be the lamina, and let
B

be the equation of the boundary ellipse J*

y
AB.
The equation of a similar ellipse PQ is
then
*** • if i o p P'

The area of the quadrant OPQ is \irka'\ and by 10-41 the x co-
10-7] EXAMPLES 181
4a'
ordinate of the centre of gravity of this quadrantal area is -=—, so that
3
the a;-moment of this quadrantal area is $ka' .
Now consider a narrow strip of the lamina bounded by the elliptic
quadrant PQ of semiaxes a', ha' and a neighbouring quadrant P'Q' of
semiaxes a' + da', h (a' + da'). The area of this strip is, to the first order,
the differential of the area OPQ, i.e. ^nlca'da', but the mass per unit
area is constant over this narrow strip and proportional to a'n, say
Xa'n, so that the mass of the strip PQQ'P' is £ nk\a'n+1 da', and the mass
of the lamina OAB

Again the z-moment of the area of strip PQQ'P' is, to the first order,
the differential of the aj-moment of the area OPQ, i.e. handa', therefore
the a-moment of the mass of the strip PQQ'P' is h\a'n+ida', and the
a;-moment of the mass of the lamina OAB
fa
Jo
Hence, by dividing the ^-moment by the mass, we find that
2(n + 2)

, . ., , . 2(n + 2 ) ,
and similarly y = ' , -~ ka.
71 \fb ~T O)

EXAMPLES
1. From a circular area of radius a a smaller circular area one
quarter the size is so stamped out that its centre bisects a radius of the
larger circle. Find the centre of gravity of the remainder. [S.]
2. If the c.G. of a quadrilateral is the same as that of four equal
particles placed at its angular points, shew that the quadrilateral must
be a parallelogram. [C]
3. Three rods of unequal length are joined together to form a triangle
ABO. If the masses are equal, prove that the centre of gravity coin-
cides with that of the area. If the masses of the sides a, b, c are pro-
portional to6 + c — a, c + a — b,a + b — c, prove that the centre of gravity
is the centre of the inscribed circle. [I.]
4. ABCD is a heavy uniform square plate, the portion CBH is
removed, where if is a point in AB. The remainder is placed with its
plane vertical and AH in contact with a smooth horizontal plane; shew
that equilibrium will be impossible unless AH\AB is greater than
[S.]
5. A triangle of uniform rods of different densities has its centre of
 182 CENTRES OP GRAVITY [X
gravity at the centre of its circumscribing circle. Prove that the den-
sities are proportional to the secants of the opposite angles. [S.]
6. Prove that the centre of gravity of the perimeter of a triangle is
the centre of the inscribed circle of the triangle formed by joining the
middle points of the sides.
If the sides are of different densities and the centre of gravity of the
perimeter is the nine-points centre of the triangle, then the densities
of the sides are proportional to the cosines of the opposite angles. [S.]
7. Particles, whose masses are proportional to Bin A cos (B—C),
sinBeos(O-A), sin C cos (A — B), are placed at the vertices of a
triangle ABO. Prove that their centre of gravity lies at the centre of
the nine-points circle of the triangle. [S.]
8. Three wires of uniform but different densities form a triangle
ABC. If the centre of gravity is at the centre of the inscribed circle,
shew that the masses of the wires are in the ratios
.A ,B .0 ro.
cot-grcot-gtcotg. [S.]
9. A uniform solid consists of a cone and a hemisphere fastened
together so that their plane faces coincide, the diameter of the hemi-
sphere being equal to that of the base of the cone. Shew that if the
height of the cone does not exceed V3 times the radius of the base, the
solid will always move to an upright position if placed with the surface
of the hemisphere on a horizontal plane. [C]
10. Find the centre of mass of a rod whose density varies as the
square of the distance from one end. [S.]
11. Prove that the centroid of the surface of a tetrahedron is the
centre of the sphere inscribed in the tetrahedron formed by joining the
centroids of the faces. [S.]
12. A body is bounded by two spheres (radii a, b) which touch intern-
ally. Find the centre of gravity and shew that in the limiting case when
a = b it divides the common diameter in the ratio of 2:1. [S.]
13. A piece of wire of given length is bent into the form of a circular
quadrant and its two bounding radii; find the centre of gravity of the
whole. [C]
14. Two tangents are drawn to a circle of radius a subtending an
angle 2a at the centre. Prove that the centre of gravity of the figure
bounded by the tangents and the smaller arc between them is at a
distance from the centre equal to
a tan2 a. sin a „
Lb<J
3" tana-a *
15. ABCD is a quadrilateral, AC and BD intersect in E, pomts F
and Q are taken on AC and BC, such that AF and BO are equal to
EC and ED respectively; prove t h a t the centres of gravity of the
triangles FBD and OAC coincide. [I.]
 EXAMPLES 183
16. A BCD is a uniform plane quadrilateral lamina, whose dia-
gonals intersect in E. If the point H divides AC in the ratio
AC+EC-.AC+EA,
and K divides BD in the ratio
BD + ED-.BD + EB,
shew that the centre of gravity of the lamina bisects HK. [S.]
17. Weights P, Q, R at the angular points of a triangle ABO have
their centre of gravity at the centre of the inscribed circle. Prove that,
p and r being the radii of the circumscribed and inscribed circles,
2pr(P+Q + R)*=:QR.a* + RP.b* + PQ.c\ [S.]
18. A wire in the form of a circular arc AB of line density w, with a
length I of string of the same line density hanging vertically from A,
rests in equilibrium in contact with a fixed smooth horizontal peg.
Prove that I is equal to A T, where T is the point in which the horizontal
plane through B meets the tangent to the arc at A. [I.]
19. A frustum is cut from a right circular cone of semivertical angle
a by planes at right angles to the axis and at distance h and h' from the
vertex. Prove that the distance from the vertex of the centre of gravity
of the whole surface of the frustum is
1 ft3 (2 sec a+ 3 tan a)-ft' 3 (2 sec a - 3 tan a) rg ,
3' ft2 (sec a + tan a.) — h'2 (sec a — tan a)
20. Two pieces of the same uniform wire, one a semi-circle and the
other an arc of a circle subtending an angle 20 (< n) at its centre, are
placed so as to form a crescent. If 9 is such that the centre of gravity is
on the inner arc, shew that it is given by the equation

21. Two parallel chords of a circle of radius a subtend angles 2a, 2yS
at the centre. Shew that the centre of gravity of the part of the disc
between the chords is at a distance ^j ^from the centre,
where A is the area of the figure. [S.]
22. A solid homogeneous hemisphere of radius a and weight W rests
with its curved surface on a fixed horizontal plane, and a particle of
weight W is placed on it at a distance x from the centre; prove that, in
the position of equilibrium, the friction exerted between the particle
and the hemisphere is equal to
8xW'*(9a*W* + 64x*W'*)-i. [S.]
23. A thin hemispherical bowl of weight W contains a weight W of
water and rests on a rough inclined plane of inclination a. Shew that
the plane of the top of the bowl makes an angle <f> with the horizontal
given by Wsin<j> = 2(W+W')ainu. [S.]
184 CENTRES OF GRAVITY [X
24. If a line is drawn cutting two sides AB and AC of a given uni-
form triangular plate in P and Q so that the area BPQC is constant,
prove that the locus of the centre of mass of this area is an arc of a
hyperbola. [S.]
25. A frustum of a uniform right circular cone whose semivertical
angle is a, is made by cutting off 1/nth of the axis; prove that the frustum
will rest with a slant side on a horizontal plane if
, , 13w 4 -4n 3 +l ra,
tan a < [S ]
4 n»-l '
26. A sphere of radius a is inscribed in a cone of vertical angle 2a.
Shew that the distance between the centre of the sphere and the centre
of gravity of that part of the volume of the cone which lies between its
vertex and the nearer part of the surface of the sphere is
a(l-f sina)2
[S.]
4 sin a
27. Find the co-ordinates (referred to the axes of the curve) of the
centre of gravity of the smaller portion of an elliptic lamina cut off by
a chord joining an end of one axis to an end of the other. [S.]
28. A frustum of a cone (vertical angle 120°) is cut off by two spheres
whose centres are at the vertex of the cone, the radius of one being
twice that of the other. If the density vary as the distance from the
vertex, shew that the centre of mass of the frustum divides its axis in
the ratio 6:19. [S.]
29. Having given two triangles ABC, A'B'C in the same plane,
prove that three masses can be found whose ratios have unique values,
such that if they are placed at the points A, B, C or at the points A',
B', C their centroid is at the same point in either case; one of the
masses may be negative. [S.]
30. A portion of a uniform thin spherical shell is bounded by a plane
circular rim. When the shell is freely suspended from a point in the rim,
the radius through the lowest point of the rim is horizontal; shew that
the plane bounding the shell is at a distance from the centre of the
sphere equal to two-thirds of the radius. [S.]
31. ABOD is a tetrahedron: a point B' is taken in AB so that
AB' = %%AB, and points 0", D' are taken in a similar manner in AC,
AD. Shew that, if the tetrahedron be divided into two parts by a plane
parallel to BCD and equidistant from it and the point A, then the
centre of mass of the larger part coincides with that of equal masses at
B', C, D'. [S.]
32. The lengths of the parallel edges of a homogeneous prismatic
solid are represented by the numbers 2, 1, 1 respectively. Prove that
the distances of the centre of gravity of the solid and of the longer edge
from the face containing the equal edges are in the ratio 3:8. [S.]
 EXAMPLES 185
33. A solid sector of semiangle a is cut from a solid sphere and re-
placed with its edge in the same position as before but with its vertex
outwards; shew that the centre of gravity of the resulting system will
coincide with the centre of the sphere if cos a = £. [S.]
34. If O is the centre of gravity of a solid cone on an elliptic base and
P is any point in the circumference of the base, then the generating line
of the cone cut by PO produced is divided in the ratio of 3 to 2. [S.]
35. A leaning tower of n equal coins is constructed on a horizontal
table, so that the centres of gravity of all the coins lie in one straight
line; find the greatest inclination of this line to the vertical. [C]
36. The thickness of a thin circular homogeneous plate at any point
is proportional to the distance of the point from a tangent to its peri-
meter. Find the volume of the plate and the position of its C.G., taking
a to be its radius and t to be its thickness at the centre. [S.]
37. If the density of any point of an arc of a uniform circular wire
varies as its distance from the central radius, prove that the centre of
mass is at a point on the central radius, midway between the chord and
the arc. [S.]
38. An isosceles triangular lamina is such that its mass per unit area
at every point is proportional to the sum of the distances of the point
from the equal sides of the triangle. Prove that the distance of the
centre of gravity from the vertex is three-fourths of the altitude.
39. Find the centre of gravity of semicircular plate of radius a
whose mass per unit area at any point varies as V(a2 ~ r2)> where r is the
distance of the point from the centre. [S.]
40. The density at any point in a sector of a circle varies as the
distance from the centre. Find the centre of gravity of the sector.
[S.]
41. The mass per unit length at any point of a straight beam of
length I is p (1 + x/l), where x is the distance of the point from one end
of the beam. Find (i) the mass of the beam, (ii) the position of its centre
of mass.
If the beam rests horizontally on supports at its ends, prove that the
bending moment is greatest at a distance (£\/21 — 1)1 from one end.
[I-]
42. Find the centre of gravity of a plate in the form of a quadrant
A OB of an ellipse, the thickness at any point of the plate varying as the
product of the distances of the point from OA and OB. [I.]
43. An elliptic section is made of a solid circular cone. If O be the
centre of gravity of the part between this section and the vertex V, and
if OM be perpendicular to the axis, prove that
VM=|(a + a') cos a, and MG=%(a—o')sina,
186 CENTRES OP GRAVITY [X
where 2a is the vertical angle of the cone, and a and a' are the longest
and shortest generators.
If H be the centre of gravity of the curved surface, and HN be per-
pendicular to the axis, prove also that

NH=l(.a-a')ainoL. [I.]
44. A light equilateral triangular frame is loaded at the corners with
weights w1, w2, iv3 and suspended from a fixed point by strings of
lengths Zj, l2, l3 attached to its corners. Prove that the tensions T1,T1,
T3 in the strings are given by
wjt _w2l2 _w3l3
Tx~ Ts ~ T3
w ^ 2 + W2?22 + V>3132) - {tUiWs+W3W1+W1W2) O2}*

where a is the length of a side of the triangle. [S.]

ANSWERS
1. \a from centre.
10. | length of rod from given end.
12. (a3 + a2 6 + a&2 + 68)/(o2 + ab + 62) from the point of contact.
13. 4-\/2o/(ir +8) from the centre.
27. 2a/3(ir-2), 26/3 (77-2).
35. tan" 1 {a/(n — 1) 6}, where a is the radius and 26 the thickness of a
coin.
36. na2t; Ja from centre.
39. f a from centre. 40. -j from centre.
41. fpZ; x = $l. 42. x =
 Chapter XI
W O R K AND E N E R G Y

11 • 1. The work done by a force when its point of application

undergoes a small displacement is defined to be the product of
the force and the orthogonal projection of the displacement
upon the line of action of the force.

Thus if the point of application A of a force P receives a

small displacement to A' where A A' makes an angle 6 with
the direction of the force P, and AN is the orthogonal projec-
tion of A A' upon the line of action of P, the work done by the
force P in the displacement AA' is defined to be P.AN or
P. A A' cos 6, and is positive, zero or negative according as 6 is
an acute angle, a right angle or an obtuse angle.
11*11. Scalar Products. If A, B are any two vectors inclined to one
another at an angle 9, then A B cos 8 is de-
fined to be the scalar product of the vectors
A, B and is variously denoted by (AB), or
A . B or simply AB.
This product of the magnitudes of the
vectors multiplied by the cosine of the angle ^
between their positive directions is clearly not
a directed quantity but simply a scalar magnitude. It is therefore called
the scalar product of the vectors to distinguish it from another product,
with which we are not now concerned, known as the vector product of
the two vectors.
We observe that the work done by a force in a small displacement is
the scalar product of the force and the displacement, and that work is
not a vector.

11*12. When any number of forces act upon a particle which

undergoes a small displacement the algebraical sum of the work
188 WORK AND ENERGY [XI
done by the separate forces is equal to the work done by their
resultant.
Let 0 be the particle which is displaced
to 0'. Let P be anyone of the forces and
R the resultant force, and let 6, <f> be the
angles which they make with 00'.
Then the algebraical sum of the work
done by the forces of which P is the
= SP.0O'cos0
= 00'. £ P cos 6
= 00'. R cos <f>
= R.OO'cos<f>
— work done by the resultant force.
11-13. If in 11-12 the forces are P 1 ; P 2 , P 3 , ..., and dplt
dp2, dp3,... denote the projections of 00' on the lines of action
of the forces, and dr denotes the projection of 00' on the line
of action of R, then what we have proved is that
ZPdp = Rdr.
Thus, in particular, if X, T , Z denote the rectangular com-
ponents of a resultant force R acting on a particle which
receives any small displacement whose projections on the co-
ordinate axes and on R are dx, dy, dz, and dr, we have
Xdx + Ydy + Zdz = Rdr.
11-14. Displacement along a Curve. Conservative Field
of Force. Now consider the displacement
of a particle along a continuous curve AB
under the action of a given field of force;
meaning by the latter term that at what-
ever point the particle may be situated
there is a definite force acting upon it
known in magnitude and direction. Let R
denote this force when the particle is at
a point Q on the curve; let 8 be the angle at Q between R and
the curve and let QQ' be an infinitesimal element ds of the
curve. Then Rcoadds denotes the work done in the displace-
ment QQ', and the work done as the particle moves along
11-12—11-15] CONSERVATIVE FIELD 189
s
c
the curve from A to B may be denoted by R cos dds or by
JA
rB
Rdr, where dr denotes the projection of ds upon the line of
JA
action of R, and this1 by 11-13 is equivalent to
A
•B
If the integrand Xdx+ Ydy + Zdz
(Xdx+Ydy is an exact differential
+ Zdz).
dW of some single-valued function of position W, then the
integral is [F"] s 4 , i.e. the excess of the value of W at B over its
value at A; and the value of the integral does not depend upon
the particular path by which the particle has been moved from
A to B but is the same for all paths. In such a case W is called
the work function and the field of force is said to be a con-
servative field. To emphasize this definition we repeat that
afield of force is conservative if the work done by the forces of the
field in the transit of a particle from one point of the field to
another depends only on the positions of these points and not upon
the particular path of transit.
From 11-13 it appears that the differential of the work
function can be expressed in the form ~LPdp, i.e. the sum of
each force multiplied by the projection of the displacement of
its point of application upon its line of action, and that when
this expression HP dp is an exact differential of a single-valued
function then the forces constitute a conservative system.
11-15. Examples, (i) The Earth's gravitationalfieldis conservative.
Let a particle move along a curve AB
in a vertical plane under the action of no y
force but its weight w. Taking horizontal
and vertical axes Ox, Oy in the plane of
the curve, the force components in any
position are X = 0, Y = — w; therefore the
work done as the particle moves from
Ato B O~
= j(Xdx+Ydy)
fb
= I -wdy
Ja
=w(a—b),
where a, b are the ordinates of the points A, B,
190 WORK AND ENERGY [XI
Hence the work done is independent of the actual path and the field
of force is conservative.
(ii) A field offorce in which the forces at any point are directed towards
fixed points and are proportional to some powers of the distances from these
points is conservative.
Let P be any point in the field and r^, r 2 , ra,... its distances from the
fixed points Olt O a , O8
Let the forces acting along POt, PO%,
PO3,... be Xxrini> Krt"2y ^3r3"s> •••; then the
work done by these forces in a small displace-
ment of P whose projections on OjP, 0%P,
O3P, ... are dr,, dr., dr3, ... is"

which is the differential of the function

Therefore the field of force is conservative.

11*16. When a particle moves round a closed path in a conservative
field of force, the work function obtained by integrating round the path
returns to its initial value on completing the circuit so that the total
work done is zero. This would not necessarily be the case if the function
were not single-valued. For example, if we had

this is an exact differential, but of a many-valued function; it is, in fact,

in plane polar co-ordinates d8, and for any closed path which includes
the origin it increases by 2w in making the circuit of the path.
The gravitationalfieldsthat occur in nature are conservative. If this
were not so, it would be possible to accumulate unlimited amounts of
work by merely allowing particles to describe closed paths.
11*2. Virtual Work. From 11*12 an immediate deduction
can be made, viz. (i) / / a system of forces acting on a particle is
in equilibrium and the particle undergoes a small displacement,
the algebraical sum of the work done by the forces is zero. This
follows from the fact that the work done by the forces is equal
to the work done by their resultant, and this is zero because
they have a zero resultant.
There is a converse to this proposition which can be proved
11-15—11*21] VIRTUAL WORK 191
in the same way, viz. (ii) / / the algebraical sum of the work done
by a system of forces acting on a particle is zero for all small dis-
placements of the particle, then the forces are in equilibrium.
For, from 11*12 if the forces have a resultant, the work done
by the resultant in any displacement is equal to the algebraical
sum of the work done by the forces, i.e. equal to zero. But the
work done by a force can only be zero if either the force is zero,
or the displacement is at right angles-to the direction of the
force. So if the work done is zero, no matter what be the direc-
tion of the displacement, it follows that the resultant force
must be zero and the given forces are in equilibrium.
These propositions are known as the Principle of Virtual
Work and its Converse for forces acting upon a single particle.
The word virtual is used to qualify work, because we are
dealing with statical systems in which there is no motion, and
the displacements are hypothetical or virtual, so that the
work calculated is the amount of work that would be done if
the displacements were actually made.
11*21. Examples, (i) A particle of weight W is supported on a smooth
plane inclined at an angle a to the horizontal by a force P which makes an
angle 8 with the plane.
Let R be the normal reaction of the plane on
the particle. Consider a small displacement of
the particle up the plane from A to A'. Since
A A' is at right angles to the direction of R,
therefore no work is done by R; but P does work
P.AA'oosQ and W does work — W.AA'sia.a., ^ t t \ w
and in equilibrium the total work done is zero,
therefore P. AA'cos 0-W.AA'ain*=0,
or Pcos8= TFsina.
The application of the principle of virtual work in this case gives the
same result as would have been obtained by resolving in the direction
AA', as we might have anticipated from the nature of the proof of the
principle. In fact in this case the principle of virtual work gives us no
more information than we already possess.
Let us next consider an application of the principle to a case in which
two particles are concerned.
(ii) Two small rings can slide on a smooth parabolic wire with its axis
vertical and vertex upwards, being connected by afinestring which passes
over a smoothpeg at thefocus. Shew that if the rings can rest in any position
in which the string is taut they must be of equal weight.
192 WOKK AND ENERGY [XI
Let W, W be the weights of the rings at P and Q. Let S be the focus.
Each ring is supported by the tension of the strmg and the pressure
of the wire, so that if the tensions of the two portions of the string PS,
QS are equal there will be equilibrium. Consider a small displacement

in which the rings move to P' and Q'. The normal reactions being at
right angles to the displacements do no work. The work done by W at
P is W. MM', where M, M' are the projections of P, P' on the axis; and
the work done by T at P is — T.Pp, where Pp is the projection of PP'
on SP. Therefore, by applying the principle of virtual work to the
ring at P, we get
W.MM'-T.Pp = 0 (1).
Similarly by considering the ring at Q we get
-W'.N'N+T.Qq = 0 (2).
Now by hypothesis P'S+SQ' = PS + SQ, and, to the first order of
small quantities, P'S=pS and SQ' = Sq, so that Pp = Qq, and by
adding (1) and (2) we get
W.MM'-W'.N'N=0 (3).
But, if the axis meets the directrix in X, we have
MM'=XM' - XM=SP' - SP

=XN-XN'=N'N.
Therefore W=W.
11*22. In the preceding example we found it necessary to introduce
the work done by the mutual action and reaction between the particles,
viz. the pull of each on the string connecting them, and we found that
the tensions at opposite ends of the string contributed equal and
opposite amounts to the total work done.
We now observe that in this case also our applications of the prin-
ciple of virtual work to the rings separately, as expressed in equations
(1) and (2), gave us no more information than we should have got by
11-21—11-23] VIRTUAL WORK 193
resolving along the tangents at P and Q; for, as appeared later,
MM' = Pp and NN'= Qq. But if we had chosen to write down in one
equation the total work done by all the forces acting on the two rings
and at the same time had been able to foresee that the work done by
the mutual action and reaction along the string would cancel out,
we should have written down at once equation (3), and the problem
would have been solved without any reference to the tension of the
string.

11*23. Our object is now to extend the Principle of Virtual

Work, proved for a single particle in 11-2, and to make it
applicable, in a much more general way, to a body or system
of bodies in equilibrium under the action of given external
forces. In 11*22 we pointed out how we could have curtailed
the work in 11*21 Ex. (ii) if we could have omitted the work
done by a mutual action and reaction, and it is only because it
is possible to omit from the work sum in the general case all
the work done by internal actions and reactions and by forces
of constraint (subject to certain limitations) that the principle
of virtual work, which we are about to establish, has any
practical value.
As the next step therefore in establishing the principle, we
will shew that there are a number of cases in which mutual
actions and reactions and forces of constraint either do equal
and opposite amounts of work, or do no work, in a small
displacement.
But we must first explain that the virtual displacements
with which we are concerned are to be regarded as small
quantities 'of the first order of smalhiess', and we shall con-
sider that forces do 'no work' when the algebraical sum of the
work done by them in terms of the displacements is 'of the
second or higher order of smalhiess'.
(i) The work done by the mutual action and reaction between
two particles is zero if their distance apart is invariable.
Let F be the force acting between
the particles A, B which undergo a A; -— —|
small displacement to A', B' so that I j
A'B' = AB. Let d be the small angle A M F F
&""
between A'B' and AB, and let M, N be the projections of
RS 13
194 WOKK AND ENERGY [XI
A', B' upon AB. The work done by the two forces F
= F.AM-F.BN
= F(AB-MN) = F(AB-A'B'cos6)
= F.AB(£02-higher powers of 6)
= 0, to the first order of the small displacement 6.
(ii) The work done by the tensions at the ends of an inextensible
string in contact with smooth surfaces is zero for a small dis-
placement.
Assuming that the string is acted upon by no forces save
tensions at its ends and the normal pressures of smooth sur-
faces, if we consider the equilibrium of any small element of

T A'

the string the only forces upon it in the directions of its length
are the tensions at its ends, so these must be equal and opposite
and the tension is therefore constant throughout the whole
length of the string.
Let APQB represent the string in the equilibrium position,
and suppose that the ends A, B are displaced to A', B', where
APA', BQB' are small angles 6, <j>. Since the string is in-
extensible A'P+QB' = AP+ QB.
If T denotes the tension at either end and M, N are the
projections of A', B' on PA, BQ, the total work done
= T.AM-T.BN
= T(PM-PA)-T(BQ-NQ)
= T (PA' cos 6- PA) - T (BQ - B'Q cos <f>)
= T (PA' -PA-BQ + B'Q), to the first order of 6, <£
= 0.
(iii) The reaction of a smooth surface upon a body which slides
on the surface does no work ; because the reaction is always along
the normal and therefore at right angles to the direction of the
displacement of its point of action.
11-23, 11-3] VIRTUAL WORK 195
(iv) The reaction of a fixed surface upon a body which rolls
on the surface does no work.
Let A be the point of contact, and after the moving body
has rolled through a small angle d let
B be the point of contact and A' the
new position of the point A on the
rolling body.
The work done is proportional to
AA', and AA' is of order AB.0.
Now 6 is the sum of the angles "E
which the tangents a,tA and^l'make
with the tangent at B, therefore
AB A'B
v= 1 —>
p p
where p, p' are the radii of curvature of the curves EAF, GAD;
and A'B — AB, therefore AB is of order 9 and AA' is of order
82, and the work done by the reaction at A is therefore neg-
ligible.
(v) The mutual reaction between two bodies which roll upon
one another does no work in a small displacement.
In this case both bodies are displaced. Referring to the last
figure, we may suppose the total displacement to be performed
in two stages. Firstly, keeping the bodies fixed relatively to
one another, let the body EAF move into its final position.
During this part of the displacement the mutual action and
reaction at A do equal and opposite amounts of work, because
thepoints-4 on both the bodies are equally displaced. Secondly,
keeping the body EAF fixed, we may allow the body CAD to
roll into its final position, then, since the whole displacement
is small, by (iv) no work is done by the reaction at A on CAD,
and therefore the total work done by the reactions remains
zero.
11*3. The Principle of Virtual Work for a System of
Bodies. A body or a system of bodies may be regarded as an
aggregate of material particles held together by forces of
cohesion together with such mutual actions and reactions as
13-2
196 WORK AND ENERGY [XI
-
we have classed in 3 6 as 'internal forces'; and in the most
general case the possible displacements of such a body or
system may be restricted by certain constraints of the type
described in 3*7 and to which reference is made in 6*2. The
theorem proved in 11*2, viz. that if a system of forces acting
on a particle is in equilibrium and the particle undergoes a
small displacement the algebraical sum of the work done by
the forces is zero, is true for the forces acting upon all the
particles that go to make up a body or system of bodies, so that
by addition it follows that in any small displacement of a body
or system of bodies in equilibrmm the total work done by all
the forces acting upon all the separate particles which compose
the body or system is zero.
Now in the case of a single rigid body in equilibrium under
the action of given external forces, there are no internal forces
save such as come under the category of mutual action and
reaction between particles whose distance apart is invariable,
and by 11*23 (i) such forces contribute nothing to the total
work done in a small displacement.
Therefore when a single rigid body in equilibrium under the
action of given external forces undergoes a small displacement
the algebraical sum of the work done by the external forces is zero.
Further, for a system of bodies in equilibrium under the
action of given external forces and certain forces of constraint
including the mutual actions and reactions between the bodies,
we have seen in 11-23 (iii), (iv) and (v) certain types of relative
displacement of bodies in which constraining forces or mutual
actions and reactions do no work; consequently, if we restrict
the small displacement of the system so that it only involves
relative displacements of the types stated, then the total work
done by all mutual actions and reactions and forces of con-
straint will amount to zero; but the total work done by all the
forces acting upon the separate particles of the system is zero,
and as the forces consist of the given external forces together
with mutual actions and reactions and forces of constraint and,
as we have just seen that the latter together contribute nothing
to the work sum, therefore the total work done by the external
forces alone must be zero.
11-3, 11-31] VIRTUAL WORK 197
We have thus arrived at the most general form of the
Principle of Virtual Work, which we may state as follows:
/ / for a system of bodies in equilibrium under the action of
given external forces and subject to certain constraints it is pos-
sible to make a displacement such that the constraining forces do
no work, then, for any such displacement, the algebraical sum of
the work done by the external forces alone is zero, or of a higher
order of smallness than the first in terms of the displacements.

11*31. The Converse Theorem. In 11*2 we established a

converse theorem for forces acting on a single particle, viz.
that if the algebraical sum of the work done by the forces is
zero for all small displacements of the particle, then the par-
ticle is in equilibrium. The converse of the general theorem is:
/ / a system of bodies is subject to the action of given external
forces which are such that, for all small displacements in which
the forces of constraint do no work, the algebraical sum of the
work done by the external forces is zero, then the system of bodies
is in equilibrium.
It is clear that this cannot be deduced by any process of
summing the results for separate particles, because the data do
not include any information about the forces which act upon
separate particles; and it is not easy to give a satisfactory
direct proof.
The following argument is open to the criticism that the
assumption made is one of which it would be difficult to
formulate an explicit proof:
Given that the algebraical sum of the work done by the ex-
ternal forces vanishes for all small displacements in which the
forces of constraint do no work, to prove that the system on
which the forces act is in equilibrium. It if were not in equili-
brium it would move. Suppose that the system is surrounded
by such frictionless constraints as will make the actual motion
the only possible one, then we assume that this motion could
be prevented by the application of a single force F, say, suit-
ably applied at a point A. Let this force be applied, so that the
system is now in equilibrium under the action of F and the
given external forces and the smooth constraints. Apply the
198 WORK AND ENERGY [XI
principle of virtual work taking a small displacement in the
direction of the hypothetical motion. This will displace A to
A' through a small distance — dfin the direction of F, and give
an equation of virtual work
— F.df+ the work done by the external forces = 0.
But by hypothesis the work done by the external forces alone
is zero, and dfis not zero, therefore F is zero and the system is
in equilibrium.
Later in this chapter (11'5) we shall shew that the vanishing of the
algebraical sum of the work done by the external forces acting on a
single rigid body in arbitrarily chosen small displacements leads to the
same equations of equilibrium as are obtained by resolving and taking
moments, and that fact constitutes a justification for assuming the truth
of the converse proposition as enunciated above for a single rigid body.
11*32. A special case. The following simple proof of the principle of
virtual work and its converse for coplanar forces acting upon a single
rigid body is due to the late Dr W. H. Besant.
Let P be any one of the forces acting at a point A. Let the body
receive a small displacement so that the
plane in which the forces act remains un-
changed. Such a displacement must con-
sist in a rotation of the body in this plane
about an instantaneous centre* I through
a small angle 6, whereby A is displaced to
A'. Then if A'M and IY are at right
angles to the line of action of P, the work done by all the forces]
= S P . A M = HP. AA' cos MAA'
= SP. 1A. 6 cos AIY

= 8x algebraical sum of the moments of the forces about I

= 0, since the forces are in equilibrium.
Conversely, if the algebraical sum of the work done by the forces is
zero, no matter about what point in the plane the rotation of the body
takes place, it follows that the algebraical sum of the moments of the
forces about each such point is zero and the forces must be in equili-
brium.
11-33. Applications. The most common direct application
of the principle is to determine the position of a body or system
of bodies in equilibrium under the action of given external
forces, when the body or system has one or more degrees of
* Dynamics, 5-41.
11-31—11-33] APPLICATIONS 199
freedom, such that small displacements are possible in which
the constraining forces do no work.
In a large number of problems the only external forces are
due to gravity, and the work done by the weights of the parts
of a system in a small displacement is equal to the work done
by the resultant weight acting at the centre of gravity of the
system. But, by the principle of virtual work, no work is done
by the external forces in a small displacement from a position
of equihbrium, and this requires that, to the first order of small
quantities, a small displacement of the system shall not pro-
duce any vertical displacement of its centre of gravity.
Hence, if z1} z2, z3,... denote the heights above a fixed level
of the centres of gravity of the bodies of weights w1>w2,w3, ...
which are supported in a position of equilibrium, and z denotes
the height of the centre of gravity of the whole, the principle of
virtual work requires that z shall be constant for small dis-
placements, or dz = O (1).
But w1z1 + w2z2 + w3z3+... = (w1+wi + w3+...)z,
therefore the condition may also be expressed in the form
w1dzl + w2dz2 + w3dz3 +... = 0 (2).
The position of the system is determined by one or more
variables or co-ordinates of position 6, <f>, if>, ..., according to
the number of degrees of freedom to be considered, so that
<h> Z2> Z3> • • • a r e expressible in terms of 6, (f>, </>,... and dzx, dzz,
dz3, ... are expressible in terms of dd, d<f>, dip, ....
After substitution, equation (2) therefore takes the form
®d6+Qd#+y¥d$ + ...=>0 (3),
where 0, <£>, T , . . . are functions of 6, <f>, tfi, ..., and the number
of these independent variables is the same as the number of
degrees of freedom.
Since the variables are independent it is possible to choose
a displacement in which, say, 6 alone is varied while all the
other variables (f>,ifi,... are kept constant; i.e. in (3) we may put
d4 = d<p=... = O, while cZ0#O,
so that from (3) we get 0 = 0.
Similarly O = 0, T = 0, etc.
200 WORK AND ENERGY [XI
and in this way we get as many equations as there are variables
and their solution determines the values of 8, $, ip, ... in the
position of equilibrium.
Reverting to the original form of the condition as formulated
in (1), we may also express it by saying that the height of the
centre of gravity above (or depth below) a fixed level is a
maximum or a minimum in a position of equilibrium. Also,
since in geometry positions of symmetry are in general posi-
tions in which the variables that define the positions have
maximum or minimum values, it is reasonable to expect that
when a system of bodies and the forces acting upon them are
capable of a symmetrical configuration it will be a position of
equilibrium if there is one.
11-34. Examples with one degree of freedom, (i) A rod lies in
equilibrium with its ends on two smooth planes inclined at angles a, (8 to
the horizontal, the planes intersecting in a horizontal line. Shew that the
inclination of the rod to the horizontal is

tan-if^r^ [S.]
2 sin asm 0
It is evident that equilibrium is only possible when the vertical
plane through the rod intersects the planes in lines of greatest slope.
The rod has then one degree of
freedom if it remains in contact
•with the planes, and its position is
determined by its inclination 6 to
the horizontal.
Let AB be the rod of length 21,
and O its centre of gravity at a
height z above the intersection G
of the planes.
Then z = lsind + ACsina.
Zsinasin(j8-0)/sin(a

and tfa = Zcos6>ri6>-2ZsinaCO3,(ft~6J.d6.

sin(oc+£)
But in equilibrium dz — 0,
therefore sin (a + /J) cos 6 — 2 sin a cos (/? — 8) = 0,
. . . . , „ sin(fi —a)
which gives t a n 6 — n . .—-=.
' 2 sin a sin jS
(ii) A heavy ring of radius r is supported by two thin smooth rods which
pass through it and intersect at an angle 2a, the plane of the rods making an
11-33, 11-34] ONE DEGREE OP FREEDOM 201
angle 8 with the horizontal. Prove that the depth of the centre of the ring
below the point of intersection of the rods is
[S.]
Let OA, OB be the rods, G the centre
of the ring which touches the rods at D
and E. Let the vertical through G meet
the chord DE in F and the horizontal
plane through O in H. Then assuming B
the position of equilibrium to be one in
which OD= OE and the plane of the ring
is vertical, the angle
DOF-FOE=a, and FOH=B.
Also, if FO = x, the depth of G below the level of 0 is

, , . . a:<fo;tan2a
therefore
dz = dx sin 8 77-5 JT—2—\ •
V(r x tan a
)
In equilibrium this expression vanishes, therefore
sin B-\/(r2 — x2 tan 2 a) = * tan 2 a,
2 2 2 2 2 2
tan 2tan
r -a; a a a:sin tan8 a tan 2 ar+2 sin2 B'
/s
Hence -+tana)
•v/(tan2 a + sin2 j3) \tan a
/ 2 2
= r v (l+cot asin j3).
(iii) A solid hemisphere is supported by a string fixed to a point on its
rim and to a point on a smooth vertical wall with which the curved surface
of the hemisphere is in contact. If 8, <j> are the inclinations of the string and
the plane base of the hemisphere to the vertical, prove that

Let 0 be the point of suspension, AB the base of the hemisphere.C

its centre and G its centre of gravity, OA the string
and D the point of contact of the hemisphere and
the wall. O
Let OA = l, and OA=a, then GG=fa, and the
depth of G below the fixed point 0 is

so t h a t d z = —Isin6d8+a(%cos<j> — sin<f>)d<f>,
and this must vanish in the equilibrium position, D
Now the variables 8, (j> are not independent be-
cause the radius CD is the horizontal projection of
CA and AO, i.e.

so that aooa<f>d<j>+loos8d9 = 0;
202 WORK AND ENERGY [XI
but we also have
— I s i n 8d8 + a (f c o s <j> — s i n <j>)dcf> — 0,
and, eliminating the ratio dd:dtf>. we find that
tan <f>^= f + tan 8.

11'35. The following examples will serve to illustrate a

method of using the principle of virtual work to determine an
unknown force or couple necessary to maintain a given posi-
tion of equilibrium:
(i) A rod AB is moveable abovt a pivot at A, and to B is attached a
string whose other end is tied to a ring. The ring slides along a smooth
horizontal wire passing through A; prove by the principle of work that the
horizontal force necessary to keep the ring at rest is -~—i—-. ^ - , where
u sin (cc "I" p)
W ts the weight of the rod and a, ft the inclinations of the rod and string to
the horizontal. [S.]
Let C be the ring and P the horizontal force.
Let AB = a, BO=l and A0 = x. A C
Consider a small displacement in which
a and I remain invariable while the angles
change. The equation of virtual work is
Pdx+Wd(iasina.) = 0 ...(1).
But a:
so that (1) is equivalent to
— P(asinad<x. + lsinpdp) + iWacosoi.da.=:0 (2)
Now a and ]8 are not independent but connected by the relation

so that acosari!a = Zcos/?<2/? (3).

Then by eliminating the ratio da: dp from (2) and (3) we get
P (tan a+tan 0) = \W,
or P=TFcosacos,8/2sin(a
(ii) Example to illustrate work done by a couple.
Let P, P be the forces of the couple acting at the end of the arm AB.
It is clear that the two forces will do equal
and opposite amounts of work in any
general translational displacement of the
arm A B that does not involve rotation. Also
no work wiE be done in displacement of the
arm AB in. & plane at right angles to the p
direction of the forces. Hence work is only
done when the arm rotates in the plane of the couple. Also it is im-
11-34—11-36] TWO DEGREES OP FREEDOM 203
material about what point the arm rotates; because by a translation,
involving no work, any chosen point of the arm can be brought into
coincidence with any point in space. Consider then a rotation of the
arm AB about A through a small angle dd; B receives a small displace-
ment BB' = AB.dd and the work done by the couple
= P.AB.de
= the moment of the couple x the angle turned through.
By addition the same result is true for displacement through an angle
of any magnitude.
A uniform square lamina of side a and weight W is suspended from
four points in a horizontal plane by equal inextensible vertical strings of
length I attached to the corners of the square. Shew that the couple that
would be necessary to hold the lamina in a position in which it has been
turned through an angle 6 from the former position is
TFa2sin0/2V{Z2-a2(l--cos0)}.
Let 0 be the centre of the lamina
A BCD before the couple is applied
and AP, BQ, CB, DS the strings. Rr
Suppose that O', A' are the displaced
positions of 0, A when a couple L is
acting on the lamina. Let O'M be the
perpendicular from O' to AP. Then
A'0'M=6, A'P = AP = l,
and O'M -O'A'=OA- aj-y/2.
If 00'= z, the equation of virtual
work for a further small displacement
is
Ld8-Wdz = 0.
But A'P2=A'M*+MP\
or i = 2a2s:inai0+(Z-z)2.
a

Whence we get O = a 2 sin0d0-2(Z-z)efe,

L W
so that
a2sin0 2(Z-z)'
or L = Waz s m e/2 ^ 2 _ a 2 (1 - cos 0)}.
11"36. Examples with two degrees of freedom, (i) Weights
wx, w2 are fastened to a light inextensible string ABC at the points B, C,
the end A being fixed. Prove that, if a horizontal force P is applied at C
and in equilibrium AB, BC are inclined at angles 9, <j> to the vertical, then
P = {w1+w2)taa6=w2ta,n<j).
Let AB = a and BO=6.
Here there are two independent variables 0, <j>. A is the only fixed
point, and the horizontal distance of C from the vertical through A is
osin0+6sin^,
204 WORK AND ENERGY [XI
so that the work done by P in a small displacement is
Pd (a sin 8 + b sin <j>),
and the equation of virtual work is
Pd (a sin 8 + b sin <j>) + wx d (a cos 0)
+ w2d(a cos 9 + b cos <j>) = 0,
or (PacoaB—wxa aia 8 —wza sin 8) d8
+ (Pb cos <j> — w2b sin <j>) d<f> = 0.
It will be observed that this equation is
of the type (3) of 11'33, and that since 8
and <f> are independent so are dd and d<j>;
i.e. the small values of d8 and dcf> may be
chosen arbitrarily, so that we may take
d</> — 0 a n d d8=£0, giving
P = {w1 + w2) tan 8,
or d8—0 and d<j>=£0, giving W-2
P
(ii) If two equal uniform rods connected by a smooth hinge are placed
over the circumference of a smooth vertical circle, apply the principle of
virtual work to shew that in the position of equilibrium the inclination of
either rod to the vertical is given by
cosec2 8 cot 8 = -,
where 21 and 2a are the length of a rod and the diameter of the circle
respectively. [S.]
If we assume that the position of equilibrium is necessarily a sym-
metrical one, then there is only one
degree of freedom; the point of inter-
section A of the rods AB, AC being
free to move up and down the vertical
through the centre of the circle. But
we can prove that a position of equi-
librium must be a symmetrical one
by allowing that two degrees of free-
dom are possible, and taking 28 to
denote the angle between the rods
and tf> to denote the inclination of
OA to the vertical.
Let O, H be the centres of gravity
of the rods AB, AC and D, E the
points of contact of the rods with the circle.
Then AD—AE = acot 8, and the height of G above O is
OB sin (8 + <f>) -DGcos(8 + <f>),
or asin(0+<£) — (I — a cot 0) cos (0 + ^).
Similarly the height of H above O is
asin(0— .£) — (I — a cot 0) cos (0 — $).
11-36, 11-4] UNKNOWN REACTIONS 205
The height above 0 of the centres of gravity of the two rods is half
of the sum of the heights of O and H, i.e.
z = a s i n 0 c o s ^ — (I—a cot 6) cos 6 cos <f>
= a ooseo 8 cos <f> — I cos 6 cos <f>,
and in equilibrium dz = 0,
therefore
— (a cosec 8 — 1 cos 8) sin j> d<f>
— (a cosec 8 cot 6 —I am 8) cos ^d0 =0 (1).
Now 8 and <j>, and therefore also dd and d<f>, are independent. If we
put dd = O and d<j> ^ 0, we must have
(a cosec 8 — I cos 6) sin (j> = 0 (2);
and one solution is (j> = 0, which gives a position of symmetry.
Then putting d<f> = 0 and dd^0 in (1), we get
(a cosec 6 cot 0 — I sin 9) cos <j> = 0 (3),
and in the symmetrical position this gives
cosec 2 tfcot0 = - (4).
The alternative solution of (2) is
(5).
V

But we have to combine this with (3), and this implies that either
(f, = £ 7T or cosec2 8 cot 8 = I/a.
Now it is easy to see that, if the angle YOE were greater than JTT,
contact between the rod AC and the circle would be broken; so that
we may reject <j> = \TT as impossible, and there remains the combination
of (5) and cosec2 8 cot 8—-.
(X
This leads to 6 = J rt and I = 2a, and refers to a special case in which a
possible position of equilibrium is with one rod horizontal and the
figure ADOE a square; and this is only possible when l = 2a.
Hence in general the position of equilibrium must be a symmetrical
one in which the inclination of the rods to the vertical is given by (4).
li"4. Application of the Principle of Virtual Work to
the determination of Unknown Reactions. The second class
of problem to which the principle of virtual work is applicable is
one in which the equilibrium configuration of a system under
the action of given external forces is known and it is required
to determine one or more of the internal actions and reactions.
In establishing the principle in its general form in 11*3 we im-
posed a restriction on the displacements for which the alge-
braical sum of the work done by the external forces alone is
206 WORK AND ENERGY [XI
zero, viz. that any such displacement must be one in which the
constraining forces do no work; but it is evident from the proof
of the principles that, if the displacement chosen be such that
some one constraining force does work, the theorem would be
valid for that displacement also provided that we add to the
work done by the external forces the work done by this par-
ticular constraining force. Then equating to zero the total
work done we have an equation in which this particular con-
straining force appears as the only unknown quantity and is
therefore determined in this way.
A similar argument applies to an unknown action and re-
action between two particles provided the distance between
them does not remain constant during the displacement, for
in this case they make a contribution to the total work done
and appear as an unknown quantity in the equation of virtual
work.
The method will easily be understood after the working of a
few examples.
11*41. In many cases it is required to find the tension or thrust in a
light rod joining two points A, B of a framework.
If the rod is rigid, it is not possible to make a displacement which
alters the distance A B. But if T is the tension in the rod, the equations
of equilibrium of the surrounding bodies are unaltered if we replace the
rod by a force T acting at A along AB and a force T acting at B along
BA. We can now make a displacement of the system which alters the
distance AB, say, increases it from I to l + dl.
Then if A', B' denote the displaced positions of A and B and M, N
their projections on AB, the work
done by the tension
A
= T.AM-T.BN '
= T(AB-MN)
= T(AB-A'B'), M T T B N
to the first order of the small angle between A'B' and AB,
= - Tdl.
Similarly the work done by a thrust T in a virtual extension of a rod
from length I to length l + dl is + Tdl.
In this way we see that it is not necessary to consider the separate
virtual displacements of the ends of the rod, but merely its increment
in length.
ll - 42. Examples, (i) AquadrilateralABCDformed'pffouruniform
rods freely jointed to each other at their ends, the rods AB, AD being equal
11-4^11-42] UNKNOWN REACTIONS 207
and also the rods BC, CD, is freely suspended from the joint A. A string
joins AtoO and is such that ABC is a right angle. Apply the principle of
virtual work to shew that the tension of the string is (w+w') sina 6 + w',
where w is the weight of an upper rod and w'ofa lower rod and 29 is equal
to the angle BAD. [SJ
Let AB=AD = 2a, A
and CB=CD = 2b.
Let the angle BCD = 2<j>.
In the given position of equilibrium <$>
is the complement of 6, ABC being a right
angle; but if at the outset we write J tx — 9
instead of ^, this implies a relation
b = atetn9
and it would be impossible to make an w'
increment in 9 without altering b or a.
This would introduce into the equation of
virtual work the unknown stresses in the
sides of the quadrilateral with which we are not concerned. We
therefore use an independent symbol for the angle ACB, and imagine
a displacement which alters the angles of the figure and the length
AC but not the lengths of the sides of the quadrilateral.
The equation of virtual work is then
- Td (2a cos 9 + 26 cos <f>)
+ 2wd (a cos 9) + 2w'd (2a cos 9 + b cos <j.) = 0,

-wasm9d8-w'(2asm.9d9+bein^>d<l>) = 0.
But 9 and <j> are connected by
a sin 9 = b sin <j>,
so that a cos 9d9=b cos (j>d<]>;
and, dividing the equation of virtual work by either of these equal
expressions, we get
T (tan 9+tan <f>) = w tan 9+w' (2 tan 8 + tan <j>).
Now put cot 8 instead of tan <f>, and we find that
T = (w + w') sin2 9 + w'.
(ii) A smoothly jointed framework of light rods forms a quadrilateral
ABCD. The middle points P, Q of an opposite pair of rods are connected
by a string in a state of tension T, and the middle points It, S of the other
pair by a light rod in state of thrust X; shew by the method of virtual work
thatT/PQ = X/BS. [S.]
The equation of virtual work for any small displacement which
alters the angles but not the lengths of the sides of the quadrilateral
ABC is
ABCD i (1),
in accordance with the explanation given in 11*41.
208 WORK AND ENERGY [XI
It is necessary therefore to find a relation between PQ and RS.
PR and QS are both parallel to BD, and
PS and R Q are both parallel to AC, there -
fore PRQS is a parallelogram and there- D
fore PQ and RS bisect one another in 0.
Then since OQ is a median of the
triangle BOG,

similarly
CO2 + OD2 = %RS2 + \CD2.
Therefore 2 (BO2 - OD2) = PQ2 - RS2 + BC2 - CD2.
Similarly 2(BO2-OD*) = RS2-PQ2 + AB2-DA2;
so that by subtraction

Now in our hypothetical displacement the lengths of PQ, RS are

altered but not the lengths of the sides of the quadrilateral, therefore
PQ.d(PQ)-RS.d{RS) =0 (2);
and from (1) and (2) it follows that
T/PQ=X/RS.
(iii) ABGDEF is a regular hexagon formed of light rods smoothly
jointed at their ends with a diagonal rod AD. Four equal forces P ad
inwards at the middle points of the rods AB, CD, DE, FA and at right
angles to the respective sides. Find the stress in the diagonal AD and state
whether it is a tension or a thrust.
Let O, H, K, L, M be the middle points of AD, AB, CD, DE, FA.
Let T be the stress in AD assumed to be a thrust.
We must not at the outset as-
sume that the hexagon is regular,
for we want to be able to make a
displacement which will alter the
length of .AD but leave the lengths
of the sides of the hexagon un-
altered. Hence we suppose that
the figure is symmetrical about
AD and denote the angles BAD,
FAD, CDA and EDA by 0.
Then if 2a denotes the length of
a side of the hexagon, the length
of AD is 2a (1 + 2 cos 9), and the work done by the thrust T in a small
displacement which alters the length of AD is

or -iaTsinddd.
For the sake of simplicity we can suppose that the displacement
does not alter the position of the point 0, but merely alters the position
11-42, 11-5] CONDITIONS OF EQUILIBRIUM 209
of A and D in the line AD, with a consequent change in 8 and in the
positions of the points H, K, L, M. By symmetry the forces P at these
four points will contribute equal amounts of work, so that it will suffice
to determine the work done by P at H.
Take an axis Ox along OA and a perpendicular axis Oy as in the
figure. The co-ordinates of H are
x = a + acos8, y = asind,
and the components of P at H are
X=-Psin0, F=-Pcos0.
The work done by P in a small displacement is therefore
Xdx+Ydy=+Pa sin2 Odd-Pa cos2 8d9
= -Pacos28de.
Hence the equation of virtual work is
- 4aT sin 6d6- iPaoos 26d6-O.
And for a regular hexagon 0 = JTT, SO that
T = Pfr/3,
and since the sign of T is positive the stress is a thrust.

11*5. Deduction of the Conditions of Equilibrium of a

Rigid Body from the Principle of Virtual Work.
(i) When the forces are coplanar. The most general small dis-
placement of the body in the plane of the forces consists of a
translation and a rotation.
Take a set of rectangular axes Ox, Oy in the plane of the
forces. Let A be the point of the body
at which a force P is applied, and let
Px, Py denote the components of P
parallel to the axes and x, y the co-ordin-
ates of A, so that OM = x and MA = y.
Consider the effect on the diagram of a M
small rotation of the body about 0 through an angle 86 in the
sense from Ox towards Oy, the axes remaining fixed in space.
The displacement of A relative to M would be y89 parallel
to MO, and the displacement of M would be x89 along MA.
Hence the rotation decreases the x co-ordinate of A bjy89 and
increases the y co-ordinate by x 89 in reference to the original
position of the axes.
If in addition the body undergoes a small translation whose
RS 14
 210 WORK AND ENERGY [XI
components parallel to the axes are Sx and By, then the total
displacements of the point A parallel to the axes are
dx = Bx — yBd and dy = Sy + xB6.
Hence, if the body is in equilibrium under the action of a set
of forces of which P is a type, we have an equation of virtual

or
But the displacements Sx, hy, Sd are independent and arbitrary,
therefore XPx = 0, S P v = 0
and
These equations represent the ordinary conditions of equili-
brium, viz. that the algebraical sums of the resolved parts of
the forces in any two directions at right angles are zero, and
the algebraical sum of the moments of the forces about a point
in the plane is zero.
(ii) When the forces are not coplanar. We now take a set of three rect-
angular axes Ox, Oy, Oz. Let A be the point of the body at which a
force P is applied whose components parallel to the axes are Px, Py, Pz.
Let x, y, z be the co-ordinates of A,
where in the figure 0L = x, LM = y
and MA = z. The most general small
rotational displacement of the body
is compounded of
a small rotation 80 about Ox,
a small rotation 8^ about Oy,
and a small rotation Si/r about Oz,
in the senses indicated by the arrows.
Considering the effect on the diagram of these rotations, the dis-
placements of A relative to M would be
zZB parallel to yO and z%j> parallel to Ox;
the displacements of M relative to L would be
ySi/i parallel to xO and ySd parallel to Oz;
and the displacements of L relative to 0 would be
xh(j> parallel to zO and xSip parallel to Oy.
Hence the combined effect of the rotations is to increase the x, y and
z co-ordinates of A in reference to the original position of the axes by
zS^-2/Si/r, xS<p-zSd and yhd-xh<j>
respectively.
11-5, 11-6] POTENTIAL ENERGY 211
If in addition the body undergoes a small translation whose com-
ponents parallel to the axes are 8x, 8y and Sz, then the total displace-
ments of the point A parallel to the axes are given by

and dz = 8z + y88-xS<t>.
Hence, if the body is in equilibrium under the action of a set of forces
of which P is a type, we have an equation of virtual work
2 (Pxdx + Pydy + Pzdz) = 0,
or Sx.'2Px+8y.'LPv + 8z.'ZP,
+ 80. 2 (yP, - zPy) + 8<j>. S (zPx - xPz) + S</r. S (xPy - yP x) = 0.
But the displacements Sx, 8y, 8z, 89, 8<f>, 8>ji are independent and
arbitrary, therefore

and S(yP,-zP v ) = 0, S(zP x -xP z ) = 0,

These equations express the facts that the algebraical sums of the
resolved parts of the forces parallel to the axes and of the moments of
the forces about the axes are separately zero. (See 14'2.)
We have thus demonstrated that the assumption of the truth of the
principle of virtual work leads to the ordinary equations of equili-
brium; and we have also demonstrated the converse theorem for forces
acting on a rigid body, because we have shewn that the vanishing of the
work sum for all displacements implies that the conditions of equili-
brium are satisfied.
11'6. Potential Energy. The potential energy of a body or
system of bodies in a given configuration in a given field of
force is the work that would be done by the forces acting on the
body or system of bodies if the body or system moved from
the given configuration to some standard configuration.
For example, considering the Earth's gravitational field,
the weights of bodies are forces which are capable of doing
work as the bodies descend from higher to lower levels. If we
choose to regard the floor of the room as the level of zero
potential energy, then a body at a higher level possesses an
amount of potential energy measured by the work which the
weight of the body would do if the body descended to the floor,
i.e. measured by the weight of the body multiplied by the
height of its centre of gravity above the floor.
It is not the absolute amount of potential energy possessed
by a body that is important, but the change in its potential
14-2
212 WORK AND ENERGY [XI
energy that would result from a given displacement. In the
case just considered we might equally well choose the ceiling
of the room as the level of zero potential energy and any body
below the ceiling would then possess negative potential energy,
equal to the work that would have to be done against its
weight to raise it to the ceiling.
It follows from the definition of potential energy that when
the forces acting on a system of bodies do positive work the
potential energy of the system decreases by the amount of
work done.
11*61. In 11*14 we denned the work function W or
f(Xdx+Ydy + Zdz)
for forces acting on a particle, and a similar definition holds
for the work function of a system of forces acting on a body or
system of bodies, and by analogy from 11*33 it follows that if
d,<j>,ip,... are a set of variables equal in number to the degrees
of freedom of a system and defining its position and W is the
work function, then dW can be expressed in the form

this expression representing the work done in a small dis-

placement of a general type.
But this work done would represent the corresponding loss
of potential energy, so that, if F denotes the potential energy,
-dV=®de+Q>d<t>+Ydili + (1).
But when V is a function of 8, <£, ip, ..., we have

Hence, since dd, d<f>, d^, ... are independent, by comparing

(1) and (2), we have
Q-JZ o»=-^ T=-^ (3)
Now just as in the expression
S(Xdx+Ydy + Zdz)
for the work function X, Y, Z denote components of force,
so, in the expression (1), 0, O, Y, ... may be called generalized
11-6-11-621] ENERGY TEST OF STABILITY 213
components of force. In particular for a displacement which
consists of an increment dB in 8, while the other variables
(f>, tp, ... are unaltered, the work done, or decrease in potential
energy is from (1) -dV=Qdd,
so that we may call 0 the 'force tending to increase 6\ and
from (3) we see that this force is — dVjdO.
If 6 denotes a length, then 0 will be a force in the strict
meaning of the word; but if 0 denotes an angle, then 0 will be
a couple tending to increase the angle.
11-62. The Energy Test of Stability. The dynamical prin-
ciple of conservation of energy is that the sum of the potential
and kinetic energies of a dynamical system is constant. It
follows from this principle that whenever a system begins to
move, since in the initial stages of the motion it acquires
kinetic energy, therefore it loses potential energy.
This fact leads to a simple proof of the theorem that
Positions of maximum potential energy are positions of un-
stable equilibrium, and positions of minimum potential energy
are positions of stable equilibrium.
Firstly, if the potential energy is stationary in value for
small displacements of a system from a given position, no
work is done in a small displacement, and by the converse of
the principle of virtual work the system is in equilibrium.
Secondly, if the system is slightly displaced from a position
of maximum potential energy and then set free, since it moves
so that the potential energy decreases, therefore it moves
farther away from the position of maximum potential energy,
which must therefore be an unstable position.
If, on the other hand, the system is slightly displaced from
a position of minimum potential energy and then set free, since
it moves so that the potential energy decreases, therefore in
this case it will move back towards the position of minimum
potential energy, which must therefore be a stable position.
11*621. A simple illustration of the theorem of 11*62 is the case of a
sphere free to move on a corrugated surface.
The weight of the sphere is the only force which does work in a dis-
placement. Unstable positions are at the tops of the ridges as at A or C,
214 WORK AND ENERGY [XI
where the potential energy measured by the height above a fixed level
has maximum values, and stable positions are at the bottoms of the
troughs as at B or D, where the potential energy has minimum values.

li*63. Applications. When the only external forces acting

on a system are the weights of the bodies which compose it,
the potential energy is measured by the whole weight multi-
plied by the height of the centre of gravity above a fixed level.
This height z is expressible in terms of one or more variables
according to the number of degrees of freedom of the system,
and the positions of equilibrium are determined and their
stability investigated by finding the values of the variables
which give to z maximum and minimum values according to
the ordinary methods of the differential calculus.
l l - 6 4 . E x a m p l e s , (i) Two equal particles are connected by a light
string which is slung over the top of a smooth vertical circle; verify that the-
position of equilibrium is unstable. (It may be supposed that both particles
rest on the circle, so that the length of the string is less than one-half of the
circumference of the circle.) [S.]
Let the string be of length 2aa, where a is the radius of the circle, and
let the radii drawn to the particles make angles 8, 2a — 6 with the
vertical.
The height of the centre of gravity of the two
particles is given by
2z = a cos 8 + a cos (2a — 8),
Therefore
35= — a sin 0 + a sin (2a— 8),
do
d?z
and -JQ2= ( )

dz d2z
The symmetrical position in which 8 = a is one in which -=-. = 0 and ^
is negative, i.e. a position in which a is a maximum, i.e. an unstable
position of equilibrium.
(ii) A solid circular cone whose height is h and semi-vertical angle a. is
placed vertex downwards in a smooth circular hole whose radius is a, cut
11-621—11-64] ENERGY TEST OF STABILITY 215
in a horizontal table. Investigate the stability of possible positions of
equilibrium. [S.]
Let A, B be the points at which the cone touches the table when its
axis makes an angle 0 with the vertical. Let O be the vertex, G the
centre of gravity and M, N the projections of 0 and G on the plane of
the hole.
Then z = NG=OGcos8-OM.
But
OM.AB = 2 (area OAB) = OA.OB sin 2a
= OM 2 sec (0 + a) sec ( 0 - a) sin 2a;
therefore
MB
OM = 2a cos (0 + a) cos (0— a) cosec 2a
= a (cos 20 4- cos 2a) cosec 2a,
and
z = fftcos 0 — a (cos 20 + cos 2a) cosec 2a.

Hence j | = - |ft sin 0 +2a sin 20 cosec 2a (1),

dfiz
and .(2).
~rm.= — fftcos(? + 4acos20cosec2a
From (1) it follows that dzjdB vanishes for 0 = 0 and for
. 3h .
cos 0 = s - sin a cos a.
oa
The value 0 = 0 refers to the symmetrical position of equilibrium, and
for this value of 0 we have from (2)
dH 3h .
-T^ = —27 + 4a cosec 2a.
This is positive if 8a > 3h sin a cos a and in this case z is a minimum and
the symmetrical position is stable. In this case also the expression
3h .
cos o0 = 5- sin a cos a
oa
gives real values for 0 on either side of the vertical, so that there are
unsymmetrical positions of equilibrium. Further, when we substitute
this value for cos 0 in (2), we find that
dH
— -4 ( sin2 a cos2 a •11 cosec 2 a ;
V64a2
and, if 8a > 3h sin a cos a, this expression for d2z/d62 is negative, so that
z is a maximum and the unsymmetrical positions are unstable. This
fact could have been foreseen because maximum and minimum values
of a function occur alternately, so that if the central symmetrical posi-
tion is stable, the unsymmetrical positions of equilibrium on either
side of the central one must be unstable.
216 WORK AND ENERGY [XI
If on the other hand 8a < 3h sin a cos a, we find that the symmetrical
position is unstable because d2z/d62 is then negative for 6 = 0; and in
this case cos 6 = 5— sin a cos a does not give real values for 8, so that there
are no unsymmetrical positions of equilibrium.
Finally, when 8a = 3h sin a cos a, then we find that
dz
-j^= — 4asin0(l — cos 9) cosec 2a,
and this vanishes only for the symmetrical position in which 0 = 0 .
Differentiating again, we find that
d2z
-j-^ = —4a cosec 2a (cos 8 — cos 26),
d?z
-y^s = 4a cosec 2a (sin 8 — 2 sin 26),
d*z =
and -j2>T 4ct cosec 2<x (cos 9 — 4 cos 20).
etc/
Thus, for 8 = 0, the first derivative which does not vanish is of even
order (the fourth) and negative, so that z is a maximum and the equili-
brium is in this case also unstable.
(iii) OneendA of a uniform rod A B of weight Wand length I is smoothly
hinged at a fixed point, while B is tied to a light string which passes over a
small smooth pulley a distance a vertically above A. and carries a weight
W/i. Ifl<a<2l, shew that the system is in stable equilibrium when AB
is vertically upwards, and that there is also a configuration of equilibrium
in which the rod is inclined at a certain angle to the vertical. [S.]
Let C be the pulley and 6 the length of the string,
so that the portion of it hanging vertically is
6 - BC=b —s/(aa +1* - 2al cos 6)
when the rod AB makes an angle 8 with the
vertical A G.
The height z of the centre of gravity is given by

Therefore
_ dz .. . . , al sin 8
5 de~ 2lsw6+
, _d2z o7 . alcosd
and 5 — =— 2icos0H r,-
d6* •y/(ai + l*-2alcoa6) (a2 +1 2 - 2aZ cos 6)*
I t follows that dzjdd vanishes for 8=0, i.e. when AB is vertically
upwards, and that for this value of 8
11-64] ENERGY TEST OF STABILITY 217

Also if I < a, this expression = — - _ , ', and is therefore positive if

also a < 21, so that z is a minimum in the vertical position and it is a
stable position of equilibrium.
dzjdd also vanishes when cos0=(3aa + 4Z2)/8aZ which gives a real
value for 0 when I < a < 21, so that there is also a configuration of equili-
brium in which the rod is inclined to the vertical.
(iv) A uniform rectangular heavy plate is suspended from a smooth
peg fixed in a vertical wall, by means of a string {length 2a) the extremities
of which are attached to two fixed points {distant 2c apart) placed sym-
metrically on one of the edges of the plate. The dimensions of the plate are
knoivn; shew that under certain circumstances there are three positions of
equilibrium of which two are stable. [S.]
Let HKLM be the plate, O the middle point of the side HM to which
the string is attached at S and S'. Let P be the peg, then since
SP + PS' = 2a,
therefore P lies on an ellipse of foci S, S' and major axis of length 2a.
Since the peg is smooth the tensions in PS, PS' must be equal and
therefore PS, PS' are equally inclined
to the vertical. But the normal to the
ellipse bisects the angle SPS', therefore
the normal at P is vertical, and if the
normal meets the radius conjugate to
CP in F then OF is horizontal.
Let O be the centre of the plate and
OC = h.
For a geometrical solution the pos-
sible positions of equilibrium could be
found by expressing the fact that the
vertical through P passes through O,
or that P must be the foot of a normal
drawn from O to the ellipse. But in
order to discuss the stability we apply the energy test thus:
If the side HM makes an angle 0 with the vertical
PF* = a2 cos2 6+(a2 - c2) sin2 0
=a*-c* sin2 0,
and the depth of O below the level of the peg is
z = h sin 6+V(a2 -
__ . dz , c2sin0cos0
Therefore T= = h cosa6 — -

, d*z , . . c2cos20 c4sin20cos20

and — = — hsm.8—--v/(a 2 -c 2 sin 2 0) (a2 - c2 sin2 0)t
218 WORK AND E N E R G Y [XI
Hence dzjdd vanishes for 9 = \TX, i.e. t h e symmetrical position, a n d
also for sin 9 — — ~. a ,_ . If h < cij^/(ai — e2) this gives real values for
c y (c + ti )
8, in which case there are three positions of equilibrium. For 9 = \TT, we
d2z c2 c2
have
d8* = -h+V{a*c*); and
' ^ y
tive a n d t h e depth z is a minimum so t h a t t h e potential energy is a
maximum and the equilibrium is unstable, and it is easily verified t h a t
with t h e same conditions sin 8 = —yph,—r^ makes dizjd92 negative so
c \ / ( c +h2) ' 6
that the two unsymmetrical positions of equilibrium are stable.
c2
If, however, h> .. 2 ^r, the symmetrical position is the only
position of equilibrium and since d2 zjd92 is now negative, z is a maxi-
mum, the potential energy a minimum and the equilibrium stable.
c2
There remains the case in which h = ., „ ^ . When this relation
holds, then dz\d9 only vanishes for 8 = \TT and we find for the value of
the first derivative which does not vanish
diz/d9i= -
This implies that z is a maximum and the potential energy a minimum,
so that the symmetrical position is the only position of equilibrium and
it is stable.
11*65. When a position of equilibrium is given and its
stability is to be discussed, it is sufl&cient to find the change in
the potential energy of the system in a slightly displaced
position by any direct method. For example.
A uniform rectangular board of weight W is free to turn about one of its
edges, and is supported in a horizontal position by a cord which is attached
at one end to the middle point of the opposite edge of the board, passes
vertically upwards from that point over a smooth fixed pulley of any
radius, and carries a weight \W at the other end. Prove that, if the board
turns through a small angle 9, the potential energy is increased by Wa2 8il&h
approximately, where 2a is the breadth of the board, and h the height of the
centre of the pulley above the horizontal plane passing through the fixed
edge. Hence shew that the horizontal position is one of stable equilibrium.
[T.]
The two figures shew the board AB when horizontal and when
turned through an angle 8 into the position AB'. In both figures
BC = h, and in the second figure B'L is at right angles to BC.
The gain in potential energy due to the displacement is
Wasva.9-\W(B0-B'G).
But B'C=(LC*+B'L2)i, and B'L = 2a(l-cos8),
so that B'C = {(h - 2a sin 0)2 + 4a2 (1 - cos 0)2}*;
11*64—11-7] ROCKING CYLINDERS 219
and, since we require a result correct to the fourth power of 6, we
substitute 8 — -^ for sin 8, and -=- for 1 —cos 8, giving
O it

powers of 0.
o
Therefore the gain in potential energy

1 Wa* ., . . .
-j —=-—tf4approximately.
4 lb

w
Since this expression is positive for both positive and negative values
of 8, it follows that every small displacement causes an increase in
potential energy, so that the horizontal position is one of minimum
potential energy and stable equilibrium.

11*7. Rocking Cylinders and Spheres. Let the figure re-

present the cross sections of two
cylinders whose axes are hori-
zontal, one of which of centre of
gravity 0 is free to roll on the
rough surface of the other which is
fixed.
In a position of equilibrium 0 is
vertically above the point of con-
tact A. Let AO = h and let the
common normal at A make an angle
a with the vertical. Consider a
small displacement in which the
upper curve rolls on the lower bringing the point B on the
upper into contact with B' on the lower.
220 WORK AND ENERGY [XI
Let the arc AB = AB' = ds, and let the tangents at B and B'
make angles d<f> and d(j>' with the common tangent at A. Then
the angle through which the upper body rolls is d<j> + d<f>', or
ds I - + -71, where p, a' are the radii of curvature of the curves
\P PI
AB,AB' sAA.
Now by the principle of virtual work, to the first order of
the small angle turned through, 0 moves horizontally, and A
being the instantaneous centre of rotation*, the horizontal dis-
placement of 0 is
AG(d<f> + d<f>'), or hds (-+
\P P
If in the displaced position O lies between the verticals at
A and B' the weight will have a clockwise moment about the
new point of support B' which would tend to restore the upper
cylinder to its former position. Hence the equilibrium will be
stable if n j \
hds I - + —; I < AB' cos a
\P PI
.„ cos a 1 1
or if —*—>-H—>•
h p p
Similarly if —=— < - + —,, the equilibrium will be unstable;
h p p
andif
cos«_l J.
h ~P+p"
the equilibrium is said to be neutral to a first approximation,
for to this order the new position of 0 will lie in the vertical
through the new point of contact of the cylinder.
11*8. Hooke's Law. As many problems on work involve
the stretching of elastic strings, it will be convenient here to
state the rule which connects the tension of such a string with
its extension and find an expression for the potential energy
stored up in a stretched stringf.
The 'extension' of a stretched elastic string means the ratio
of the increment in length to the unstretched length. Thus if
* As in Dynamics, 5-42.
t See also Dynamics, 7-3-7-32.
11-7-11-81] HOOKE'S LAW 221
l0 is the natural or unstretched length and the stretched length
is I, then the extension is (I — lo)llo •
Hooke's Law is that the tension of the string is proportional
to the extension. If T denotes the tension and we state the law
in the form m_X(l-l0)

then A is called the modulus of elasticity of the string.

To find the work done in stretching a string from its natural
length l0 to a length I.
When the length is x the tension is -~—- and the work
done against the tension in making a small increment dx is
d

Therefore the whole work done

if T denotes the final tension.

This work is stored up in the string as potential energy.
In like manner the work done in increasing the length from
I to V x f •'
I'OJ I

7
l
0
+ +T l
0
= the increment in length x the mean of the
initial and final tensions.
11*81. Hooke's Law also gives the relation between the
tension or thrust and the extension or compression of a spiral
spring.
EXAMPLES
1. A smooth circular cylinder of radius 6 is fixed parallel to a smooth
vertical wall with its axis at a distance c from the wall. A smooth uni-
form heavy rod of length 2a rests on the cylinder with one end on the
222 WORK AND ENERGY [XI
wall and in a plane perpendicular to the wall, shew that its inclination
d to the horizontal is given by
acos30 + 6sin0 = c. [S.]
2. Two equal light rods A OB, COD, freely jointed at 0, their middle
point, are at rest in a vertical plane with their ends B, C on a smooth
horizontal table. A string to the ends of which equal weights are at-
tached passes over A and D. Shew that in the position of equilibrium
the angle between the rods is tan" 1 jj. [I.]
3. A smooth ring is fixed above a smooth plane inclined at an angle
a to the horizon. Prove that the rod of least length which passes
through the ring and can rest in equilibrium with one end on the plane
makes an angle <f> with the horizontal given by sin (a + 2cf>) = 3 sin a. [S..]
4. A rectangular lamina A B CD rests with the sides A B, AD on two
smooth pegs in a horizontal line; prove that, if the distance between the
pegs is half a diagonal of the rectangle, AB, AD bisect the angles
between AC and the horizon. [8.]
5. A uniform square lamina rests in equilibrium in a vertical plane
under gravity with two of its sides in contact with smooth pegs in the
same horizontal line at a distance c apart. Shew that the angle 0 made
by a side of the square with the horizontal in a non-symmetrical posi-
tion of equilibrium is given by
c (sin 0 + cos 0) — a,
2a being the length of a side of the square. [S.]
6. Shew by the Principle of Virtual Work that if any number of
forces P, Q, R, etc. act on a particle O and p is their resultant; and if a
transversal be drawn cutting the lines of action of the forces in L, M,
N, etc., and the line of action of the resultant in a,

7. A rod, whose centre of gravity divides it into lengths a and b, is

fitted at its ends with rings which can slide on two smooth wires fixed
in a vertical plane each at an angle a to the vertical; prove that in
equilibrium the depth of the centre of gravity below the intersection
of the wires is equal to
£coseca(a2 + 62 + 2a&cos2a)£. [C]
8. Four equal uniform rods, of length a + 6, are hinged at A, B, C,D,
and are suspended by two strings of equal length, so that the diagonal
AC is vertical and A is the highest point. The strings are attached to
two pegs in the same horizontal line and to two points in AB, AD at
distance a from A. Prove that if the rods rest in the form of a square
the inclination of the strings to the horizontal is tan" 1 (a/6). [S.]
9. A Toggle joint consists of two links, AB, BC, each 1 ft. long,
hinged together at B: A is fixed and C can slide without friction along
 EXAMPLES 223
the fixed line AC. If a force P be applied at B to AB, at right angles to
it, determine the force which applied at C will maintain equilibrium
when the angle ABC is 120°. [I.]
10. A pentagon is formed of five equal uniform rods smoothly
jointed at their extremities. It hangs with the two upper rods in contact
with smooth pegs in the same horizontal line and the lowest rod hori-
zontal. Shew that, if in equilibrium the pentagon is regular, the pegs
must divide the rods in the ratio 2 + ^ 5 : 3 . [S.]
11. Two equal rods AB, BC each of length I have a smooth joint at
B and the ends A, C carry smooth light rings which can slide on the
upper arc of an elliptic wire of axes 2a, 26; the major axis (2a) is vertical
and the plane of the wire is also vertical. Shew that if 62/2a < I < b, and
the system is in equilibrium, each rod is either vertical or inclined to
the vertical at an angle 6, where
64 = Z2(4a2cos20 + 62sin20). [S.]
12. Two smooth wires OA and OB are fixed in a vertical plane and
are inclined at equal angles a to the vertical. Two equal heavy uniform
rods HC and KC are smoothly hinged at C and the ends H and K are
constrained to slide along OA and OB respectively by means of small
smooth weightless rings; the weight of each rod is W, and an additional
weight W is attached at C. Shew that in the symmetrical position of
equilibrium each rod is inclined at an angle 8 to the vertical given by
tan 0 = f cot a. [S.]
13. Two particles of masses m and mf are connected by a string of
length I resting on a smooth cycloid with its vertex upwards, and base
horizontal. Prove that in the equilibrium position the distance of the
particle m from the vertex measured along the arc is m'lftm, + tn'). [C]
14. Within a smooth fixed elliptic cylinder, whose cross section is an
ellipse having its major axis horizontal and equal to twice the vertical
minor axis 26, is placed a pair of compasses formed of two equal uniform
rods, each of length 26, j ointed together. Shew that any position of the
compasses will be one of equilibrium provided that the feet of the com-
passes meet the cylinder at points on the same level and their plane is
perpendicular to the axis of the cylinder. [S.]
15. A rhombus ABCD is formed of four equal uniform rods freely
jointed together and suspended from the point A; it is kept in position
by a light rod joining the mid-points of BC and CD; prove that if T be
the thrust in this rod and W the weight of the rhombus T = W tan \A.
[S.]
16. Four equal uniform rods of weight W are freely jointed so as to
form a square ABCD which is suspended from A and is prevented from
collapsing by an inextensible string joining the middle points of AB
and BC. Prove that the tension of the string is 4TF and find the
magnitude and direction of the reaction at B. [S.]
224 WORK AND ENERGY [XI
17. ABCD is a rhombus of freely jointed rods lying fiat on a smooth
table and P, Q are the middle points of AB, AD. Prove that if the
system is held in equilibrium by tight strings joining P to Q and A to C,
the tensions in these strings are in the ratio of 2BD to AC. [S.]
18. ABCD is a rhombus formed of light rods loosely jointed to-
gether; OB, OD are two equal rods jointed at 0, B and D. If 0 is
connected to A and C by strings in tension, prove that the tension of
each string is inversely proportional to its length. [S.]
19. Four equal heavy rods freely jointed at the ends form a square
ABCD, which when balanced over its lowest point A is kept from
changing its form by a light string joining B, D. The mass of each rod
being m, find, by the method of virtual work or otherwise, the tension
of the string when a weight of mass M is attached to C. [S.]
20. Four uniform rods are jointed to form a rectangle ABCD. AB is
fixed in a vertical position with A uppermost, and the rectangle is kept
in shape by a string joining AC. Find the tension of the string. [S.]
21. Three equal rods AB, BC, CD are jointed together smoothly at
B and C whilst A and D are fixed at a distance apart equal to the length
of either rod. B and C are fastened by strings to points E and F in the
same straight line with AD such that EA = AD = DF. Shew that the
tensions in these strings are proportional to their lengths. [S.]
22. A framework of n equal weightless rods freely jointed together
is maintained in the form of a regular polygon by equal forces P applied
at the middle point of each rod at right angles to it. Prove that the
tension or thrust in each rod is £ P c o t - . [S.]
23. A regular octagon ABCDEFOH is formed of eight equal heavy
rods jointed together; the rod AB is fixed horizontally and the frame-
work hangs from it, the regular octagonal form being maintained by
two weightless rods joining C and H, D and O. Shew that the reactions
alongtherods CH andDG&re 3 IF and W respectively, where TF=weight
of one of the heavy rods. [S.]
24. A regular octahedron formed of twelve equal rods of weight w
freely jointed is suspended from one corner. Prove that the thrust in
each horizontal rod is 3w/V2. [S.]
25. A regular pentagon ABODE is formed of five uniform heavy
rods each of weight W freely jointed at their extremities. It is freely
suspended from A and is maintained in its regular pentagonal form by
a light rod joining B and E. Prove that the stress in this rod is W cot 18°.
[S.]
26. A triangular lamina ABC rests with its plane vertical, and with
the sides AB, AC supported by smooth pegs D, E in a horizontal line.
Prove that, if AD=p, AE=q, then
q%) = Q. [I.]
 EXAMPLES 225
27. A square ABCD formed of light rods, loosely jointed, has the
side AB fixed. The middle points of AB, BC are joined by a string
which is kept taut by a force P acting at the middle point of AD
parallel to AB. Shew that the tension of the string is equal te P\/2.
[S.]
28. A rhombus ABCD of loosely jointed rods is in a horizontal
plane with the rod BCfixedin position. The middle points of AD, DC
are joined by a string which is kept taut by a couple L applied to the
rod AB. Prove that the tension of the string is 2L/AB cos \ABC. [I.]
29. A light rhombus formed of rods smoothly jointed at A, B, C,D
rests in a vertical plane with A vertically above C and the rods AB, AD
over smooth pegs at the same level at a distance 2c apart. B, D are
connected by a light rod so that the angle A of the rhombus is 2a.
Shew that if a weight W is hung from C, the tension in the rod BD is
W (gj- sec a cosec2 a — tan a J, a being a side of the rhombus. [S.]

30. Three rhombuses ACBD, CEDF, EGFH, each made of four

equal light rods loosely jointed, are freely connected at the extremities
of their common diagonals CD, EF. The rhombuses are not necessarily
in the same plane. Shew that, if the diagonals AB and OH are two
stretched strings, their tensions must be proportional to their lengths.
[S.]
31. A rhombus ABCD formed of four uniform freely jointed rods
each of weight W and length a rests symmetrically in a vertical plane
with AB, AD in contact with two smooth pegs in the same horizontal
plane at a distance 2c apart (the vertex A being downwards), and is
kept from collapsing by a light string BD. Prove that the tension pf
the string is 2W(asia^A-ccosec2\A)\acos\A. [S.]
32. A parallelogram ABCD is formed of uniform heavy rods freely
jointed at their extremities. AB is held fixed in a horizontal position
and the parallelogram is maintained in its form so that ADC is an
acute angle a by means of a string joining A to a point Pin. DC. Prove
that the tension of the string is W.AP cot tx/DP, where W is half the
weight of the parallelogram. [S.]
33. A string of length a forms the shorter diagonal of a rhombus
formed of four uniform rods, each of length b and weight W, hinged
together; prove that, if one of the rods is supported in a horizontal
position, the tension of the string is
2TF6- 1 (26 2 -a 2 )(46 2 -a 2 r i - [S.]
34. AB is the horizontal diameter of a circular wire of radius a
whose plane is vertical. A bead of mass M at the lowest point C can
slide on the wire and is attached to two strings which pass through
small fixed rings at A, B. To the other ends of the strings are attached
equal masses m which hang freely. Find the potential energy of the
RS 15
226 WORK AND ENERGY [XI
system when it is displaced so that the radius to the bead makes an
angle 8 with the vertical.
Shew that the equilibrium with M at C is stable if m<MV2. [S.]
35. An isosceles triangle of angle 2a rests between two smooth pegs
at the same level, distant 2c apart; prove that if h is the distance of the
centre of gravity from the vertex, and if
2c sec a.<h< 2c/sin a cos a,
then oblique positions of equilibrium exist, which are unstable.
Discuss the stability of the vertical position in case
h = 2c/sin a cos a. [C]
36. A uniform smooth rod passes through a ring at the focus of a
fixed parabola, whose axis is vertical and vertex below the focus, and
rests with one end on the parabola. Prove that the rod will be in
equilibrium if it makes with the vertical an angle 0 given by the
equation g a
COs4 =
2 2c'
where 4a is the latus rectum and 2c the length of the rod.
Investigate also the stability of the equilibrium in this position. [S]
37. Three equal uniform rods, DA, DB, DC, each of length 21 and
of weight W, are smoothly jointed together at D, and respectively pass
through three small smooth fixed rings at the corners of an equilateral
triangle, whose length of side is a and plane horizontal. Also a weight w
is attached to D. Prove that a symmetrical position of equilibrium is
possible, if \w
\w \/3l
3W<~a~~
Also prove that two symmetrical positions of equilibrium will then
exist, and investigate their stability. [S.]
38. A uniform rod A B of weight W can turn freely round one end A:
a fine cord is attached to a point C vertically above A, passes through
an eyelet fixed on the rod at the end B and carries a hanging weight w
at its other end. Prove that, in the absence of friction, the rod will be
in equilibrium when BO is equal to w.AC/(%W + w). Prove also that
this position of equilibrium is an unstable one. [S.]
39. AB is a uniform rod, of length 6a and weight W, which can turn
freely about a fixed point in its length distant 2a from A. AC and BO
are light strings of length 5a attached to a particle C of weight w.
Shew that if W is less than 2w there will be stable equilibrium with AB
inclined to the horizontal at an angle tan"1—j . [S.]

40. A uniform heavy rod of length 21 rests with its ends on a fixed
smooth parabola with axis vertical and vertex downwards (latus
rectum = 4a). Shew that if l>2a there are three positions of equili-
brium and that the horizontal position is then unstable, but that if
the only position of equilibrium is horizontal. [S.]
 EXAMPLES 227
41. Shew that a rough uniform plank resting horizontally on the
top of a circular cylinder will be in stable equilibrium if its thickness is
less than the diameter of the cylinder.
Supposing the condition satisfied, find the greatest displacement for
which the stability obtains. [C]
42. A solid sphere rests inside a fixed rough hemispherical bowl of
twice its radius. Shew that, however large a weight is attached to the
highest point of the sphere, the equilibrium is stable. [S.]
43. A solid circular cylinder of radius a and height h has one end in
the shape of a hemisphere; find the condition that it will be in stable
equilibrium when standing on that end, on a smooth horizontal plane,
with its axis vertical. [S.]
44. A uniform hemisphere rests in equilibrium with its base upwards
on top of a sphere of double its radius. Shew that the greatest weight
which can be placed at the centre of the plane face without rendering
the equilibrium unstable is one-eighth of the weight of the hemisphere.
[S.]
45. A homogeneous hemispherical shell of radius a and weight W
has a weight 2W/3 fastened to a point in the rim of the hemisphere.
Prove that if properly placed it can rest in neutral equilibrium on the
top of a fixed sphere of radius a. [S.]
46. A stiff wire in the form of a parabola rests on the ground with its
plane vertical. The centre of gravity of the wire is on the axis of the
parabola at a distance h from the vertex, and the latus rectum is 4a. Prove
that, if h > 2a, there is a position of equilibrium in which the axis makes
an angle tan - 1 ( T—«- J with the horizon. Also prove that this position
of equilibrium is stable. [S.]
47. Two equal particles repel each other according to the fifth power
of the distance, and are connected by an elastic string. Find the posi-
tion of equilibrium, and shew that it is stable if the extension of the
string is less than one-quarter of its original length. [S.]
48. A uniform elastic string has a length ax when the tension is Tx,
and a length o2 when the tension is T2 • Shew that its natural length is

and that the amount of work done in stretching it from its natural
length to a length (ax + o2) is

49. A uniform rod of weight W and length I is suspended from a

fixed point by two light elastic strings attached to its ends. If the
strings have the same modulus of elasticity, W, and are of natural
15-2
228 WORK AND BKERGY [XI
lengths I and Z/2, prove that their lengths in the position of equilibrium
are l/x and l/y, where y=l+x,

and
^
50. OA is a slightly compressible vertical rod of height h and negli-
gible mass (modulus of compressibility fi) freely pivoted at its lowest
point O. AB is a slightly extensible cord of natural length I (modulus
A). B is a point in the horizontal plane through O distant a from O
where a2 = Z2 — h2. A horizontal force P is applied at A in the direction
BO. Shew that the horizontal and vertical components of the displace-
ment of A are approximately (neglecting x2 and y2)
P fh3 , l»\ Ph*
=-» — + T ) > y- — •
51. Two small heavy rings a2\ix. connected
XJ " byajxa light elastic string can
slide without friction one on each of two fixed straight wires OA, OB,
which lie in a vertical plane through O, the highest point, and are both
inclined to the vertical at 45°. Prove that there is only one configuration
of equilibrium, and that if the weights of the rings are \ and J of the
modulus of elasticity of the string, the length of the string is twice its
natural length.
Investigate the stability of this configuration. [S.]
52. The sides of a parallelogram A BCD are four stretched extensible
strings with their ends tied to the two straight rods AC and BD which
form the diagonals. The natural lengths of AD and BC are a, and those
of AB and DC are b; the stretched length BC is r. Prove that, if the
system is in equilibrium, r satisfies the equation

where c and d are the lengths of AC and BD, and all the strings have
the same modulus of elasticity. [S.]
53. The lower ends of three identical vertical springs of length l0
and large modulus of elasticity A are attached to three fixed points at
equal distances a apart in a horizontal line. A bar of mass M is placed
across their upper ends and attached to them in a position in which its
centre of gravity is at a distance c from the middle spring. Find the
potential energy when the middle spring is compressed a distance x,
and the rod makes a small angle 6 with the horizontal, and hence find
the position of equilibrium and for what values of cja one of the end
springs becomes extended. [S.]
64. Two equal uniform rods are hinged together at one of their ends,
and the other ends are connected by a light elastic string whose natural
length is equal to the length of either rod. They are placed with the
hinge upwards, in a vertical plane, and resting on a smooth horizontal
plane. If it be assumed that the tension of an elastic string varies as
 EXAMPLES 229
the amount of stretching, and if a tension equal to the weight of either
rod would stretch the string to double its natural length, shew that in
the position of equilibrium the inclination of either rod to the horizon
is given by 4 -L+^_ rS 1
L J
cos0'

ANSWERS
9. P. 16. W vertically, %W horizontally.
19. (M + 2m)g.
20. W. A C/2A B, where W=weight of rectangle.
34. 2 ^/2mga cos \6 — Mga cos 9 + const.
35. Unstable. 36. Stable.
37. Stable or unstable according as D is above or below the triangle.
41. Stable so long as cos2 8> (thickness)/(diameter).
43. A<a(l + 1/V2). 51. Stable.
53.
 Chapter XII
FLEXIBLE CHAINS AND STRINGS
12*1. Equations of equilibrium of a chain in one plane
under the action of any given forces. Let s be the length
of the chain measured from
some fixed points, up to a vari-
able point P, and 8s the length \ y ^ f T+6T
of the small arc PQ. When the
chain is acted upon by external
forces the tension will not be A
constant. It is convenient to
assume that the tension in-
creases in the sense in which we take the arc s to increase, so
let T denote the tension at P and T + 8T denote the tension
at Q. Also let hift denote the small angle between the tangents
at P and Q.
Now when we consider the equilibrium of the element PQ
of the chain we have to take into account not only the external
forces, such as its weight, which act upon it, but also the pull
of the rest of the chain upon it; and this pull is represented by
a tension T along the tangent at P in the sense PA and a ten-
sion T + ST along the tangent at Q in the sense QB.
Let us resolve these two forces along the tangent at P in the
sense in which s increases, and along the inward normal to the
curve AB at P. We get
- T + (T + 8T) cos Sifi along the tangent
and (T + 8T) sin Stfi along the inward normal.
To the first order of small quantities these expressions
reduce to S77 along the tangent (1),
and T8ip along the inward normal (2).
These expressions are of great importance; they constitute
a convenient measurement of the reaction of the rest of a chain
12-1, 12-11] EQUATIONS OF EQUILIBRIUM 231
upon an element of itself; viz. that this reaction is compounded
of a force 8T along the tangent to the element and a force T8ip
along the inward normal.
In order to write down equations of equilibrium, our only
further requirement is a specification of the external force
system. The external forces acting upon an element PQ of the
chain will necessarily be proportional to the length 8s of the
element, and may therefore be represented by a force F8s
along the tangent at P (in the sense PQ) and a force 08s along
the inward normal at P ; i.e. F and G denote components of
force, per unit length of chain.
The equations of equilibrium of the element PQ are then
8T + F8s = O (3),
and T8I/J + G8S = 0 (4),

which may also be written

and - + G=0 (6),

P
where p is the radius of curvature of the curve formed by the
chain.

12*11. Special cases. In what follows we shall suppose that

the chain or string and the forces acting upon it lie in the same
plane unless the contrary is expressly stated.
(i) A string in contact with a smooth surface under the action
of no forces but the reaction of the surface and tensions applied
at its ends.
Here there is no tangential component of external force, so
that equation (3) is simply

or T = constant,
along the whole length of the string.
232 FLEXIBLE CHAINS AND STRINGS [XII

Again, if R denotes the reaction of the surface per unit

length of string along the outward normal, equation (4) gives

or if =—,
P
where p is the radius of curvature; so that T, being constant,
the reaction at any point varies as the curvature.
12-12. (ii) A string in limiting equilibrium in contact with a
rough surface and under the action of no forces but the reaction of
the surface and tensions applied at its ends.
Let AB be the string just about to slip from A towards B
under the action of tensions TA,
TB at its ends.
Then with the notation of 12-1,
let the tangents at A, P, B make
angles a, >p, ft with a fixed
direction.
The reaction of the surface on
an element PQ of length §s may
be represented by a force RSs
along the outward normal and a
force /JLRSS along the tangent
opposed to the direction of supping. Then equations (3) and
(4) of 12*1 become *m
r
and j rsi/j-RSs=o.
Whence we get 8T

or log 5r = /jL>p + constant,

or
But when I/J — a then T = TA, therefore

and, by division rp __ rp ef4(0-a) Q\

J l
In particular TB = TAeJ -<fi-° '> (2).
I t follows that the tension increases along the string in the
12-11—12-13] HEAVY STRING ON SMOOTH CURVE 233
sense in which it is about to slip by this exponential factor, in
which the angle is the angle turned through by the tangent to
the string as we proceed along it.
12*121. The result of 12*12 is of great practical importance and
explains for example how it is possible for a single man, by hitching a
rope round a post, to destroy the momentum of a vessel of some size
arriving at a landing stage.
Example. Find the ratio of the tensions when a rope is coiled three
times round a post and just about to slip, taking fi = 0-5.
The required ratio = e'BX •"• = e 3ir .
But log10 e37r = 3i7 x -4343
= 4-093;
therefore the ratio is as 12,400:1.
12-13. (iii) A heavy uniform string in contact with a smooth
curve in a vertical plane.
Let PQ be an element of the string AB. The string may either
be regarded as resting on the upper side of the curve, or as
pressed against the lower side of the curve by applying tensions
y

at its ends; and the normal reaction B 8s of the curve on an

element Ss of the string will be inwards or outwards in the two
cases. In the figure the latter case is taken.
Let >(i denote the angle which the tangent at P makes with a
horizontal axis Ox, and let P be the point (x, y). The arc AP is
measured upwards so that s and i/r increase together. Let w
denote the weight of unit length of the string.
Then by resolving along the tangent and normal at P we get
ST — w8ssm>p = 0,
and T Sip —wSs cos ift — R8s = 0,
or since sin ifi = dy/d8, T = wy + const (1),
T
and .(2).
R= w cos
P
234 FLEXIBLE CHAINS AND STRINGS [XII
Equation (1) determines the tension at any point when its
value is known at some one point and equation (2) then
determines the pressure at any point.
It follows from (1) that, if T1, T2 be the tensions at points
whose ordinates are y1, y2, then

or the difference between the tensions at two points is pro-

portional to the difference between their levels.

12'131. E x a m p l e . A heavy uniform string passes round a circular

cylinder whose axis is horizontal. Find the tension of the string at the
points where it is vertical, if the pressure on the cylinder vanishes at the
lowest point.
Let a be the radius of the cylinder, w
the weight of unit length of the string.
Consider a small element of length
a 89, where the radius makes an angle 9
with the downward vertical.
The forces acting on this element are
as shewn in the figure, and, by resolv-
ing along the tangent and normal, we
Set 8T = wa$8 sin 8, jp j
waS6
or T=C-wacoB6 (1);
and T 88 = Ra 89+wa 89 cos 6,
or T = Ra + waooa9 (2).
If R vanishes when 9—0, the tension at the lowest point is from (2)
T = wa, and substituting this value for T in (1) when 9 = 0, gives G = 2wa.
Therefore T = 2wa — wa cos 6,
and where the string is vertical 8 = JTT, SO that the required tension is
Iwa.

12*14. (iv) A heavy uniform string in limiting equilibrium

in contact with a rough curve in a vertical plane.
Using the notation and figure of 12-13, if the string is about
to slip from A towards B, there is an additional tangential
force on the element PQ of amount fj,R8s, due to friction, and,
by resolving along the tangent and normal at P we get
hT - fiRSs - wSs sin ifi = 0,
and
12-13—12-141] HEAVY STRING ON ROUGH CTTBVE 235
Eliminating B and dividing by S^i we get
dT
— \x.T = wp (sin tfi — p cos tp) (1),

where p is the radius of curvature of the curve at P .

This equation is integrated by multiplying both sides by
the integrating factor erv$, and leads to
Terv* =C + w$e-^p(sinip-pcos>p)dip (2).
For a curve of given intrinsic equation, p can be expressed
in terms of ip, and in some cases the integral in (2) can be found
by the aid of the standard forms

i~^smciApd^ = j (a cos aip + n sin aip),

•rffr cos a<p dip = - , (a sin aip — jj. cos aip).

12*141. Example. A heavy string occupies a quadrant of the upper

half of a rough vertical circle in a state bordering on motion. Prove that
the radius through the lower end makes an angle a with the vertical given
by tan (a — 2A) = e-i"'1, where fi = tan A is the coefficient of friction.
Let AB be the string, a the radius of the circle, PQ a small element of
the string of length a 88, where 8 is the angle
which the radius OP makes with the vertical. RaSd
Then using T, R and w as in 12*14, the forces
acting on the element PQ are as shewn in the
figure, and the equations of equilibrium are
ST-fiRaS8= -waSdsind,
and TS8—BaS6= -waSdcosd,
leading to the differential equation waS6
dT
~JTQ—IILT= — wa(sm.8
This gives on integration
a
- l)cos0-2/isin0},
or putting tan A for fi
=CeuB + wacoa(8-2\).
Now T vanishes at both ends of the string, i.e. at A where

and at B where 8 = a.
Therefore 0= ^ + wa sin (a - 2A),
and 0=
so that, by division, tan (a — 2A) =
236 FLEXIBLE CHAINS AND STRINGS [XII

12*2. The Common Catenary. This is the curve in which

a uniform chain or a 'perfectly flexible' string hangs when
freely suspended from two fixed points.
At the lowest point 0 of the curve the tension is horizontal
and denoted by To.

O N M G
Let s be the length of the arc measured from C to any point
P of the chain. Let T be the tension at P and >fi its inclination
to the horizontal.
If w denotes the weight of a unit of length of the chain, the
weight of the portion CP is ws, and the portion CP is in equili-
brium under the action of three forces, viz. its weight and the
tensions To, T at its ends.
Therefore, by resolving horizontally and vertically, we get
Tcost=T0. (1),
and T8ini[i = ws (2).
It is convenient to introduce another constant and write
T0 = wc in (1); then by division we obtain the equation
5 = ctam/r (3),
which is the intrinsic equation of the catenary.
Since c is the only constant in the equation, it is called the
parameter of the catenary.
The Cartesian equation is easily deduced thus:
dy
By differentiating s = c tan tfi = c
dx*
ds_ d2y
we obtain C

-T- I t = C
or 'dx%'
12-2, 12-21] COMMON CATENARY 237
This gives on integration
c sinh"1 -f- = x + A,
ax
where A is a constant of integration.
The positions of the axes of co-ordinates are not yet fixed.
Let us take the vertical through the lowest point of the chain
for axis of y; then dyjdx = 0 when x = 0, so that A = 0 and
^ = sinh (4).
dx c
Integrating again we get
, x _
y = c cosh - + B,
c
where B is a constant of integration.
Now take the origin at a depth c below the lowest point of
the chain, and we have y = c when x = 0, so that B = 0, and
or
y = ccosh- (5)
c
is the equation of the catenary; and the axis Ox is called its
directrix. The curve is clearly symmetrical about its lowest
point.
12*21. Geometrical Properties of the Common Catenary.
Since s ^
and from (4) -j- = sinh-,

therefore s = csinh- (6).

c
Again if PM or y be the ordinate drawn to the directrix and MK be
the perpendicular from M to the tangent at P, we have, from (5)
«2 = c2 cosh2 - = c2 + c2 sinh2 - ,
c c
therefore, from (6) yi = c2 + si (7).
But s = ctan^r, therefore y = cseci/i (8).
Again, in the triangle PKM,
MK=MPOOBI/,

= o,
and PiC=lOftan</.=ctan^=«;
i.e. PK is equal to the arc PC of the curve.
238 FLEXIBLE CHAINS AND STEINGS [XII
It follows that if a string were wrapt round a material curve in the
form of a catenary and then cut at the vertex C and gradually unwrapt,
the end of the string would trace out the locus of K in the figure; in
other words the locus of K is an involute of the catenary. It is called
the tractrix. KP is the normal to the involute, and KM is the tangent,
and we have proved that KM, the portion of the tangent to the in-
volute cut off by the directrix, is constant and equal in length to the
parameter c of the catenary.
Further, if the normal at P to the catenary meets the directrix in G,
we have
PG - PM sec 0 = c sec2 </r
_ds_

so that the radius of curvature at any point on the catenary is numeric-

ally equal to the length of the normal intercepted between the curve
and the directrix, but they are drawn in opposite directions.
12-22. The Tension. It follows from the Cartesian equation
that all common catenaries are similar and that their relative
magnitudes depend only on the parameter c.
The tension T at any point is given by 12*2 (1); thus the
horizontal component is
T cos tfi=T0 = we,
so that T = we sec ip = wy (9),
i.e. the tension at any point is proportional to the height of the
point above the directrix.
It follows that if a chain hangs in festoons over a number
of smooth pegs that the catenaries in which it hangs all have
the same directrix and that if the chain has free ends they must
be on the directrix.
12*23. The Parameter. When a given length of uniform chain
hangs between two given points it is a
definite problem to find the parameter of the
catenary.
Thus if A B is the chain of length I and a, b
are the co-ordinates of B referred to A, and
A is the point (x, y) on a catenary of para-
meter c whose arc from A to B is of length I,
we have
2/ = ccosh-, s = csinh-,
c c
, , x+ a , . , x+a
and y + b = c cosh , s + l = c sinn ;
12-21—12-231] COMMON CATENARY 239

so that b = c (cosh cosh - ) ,

, , / . , x + a . , x\
and i = c(sinh sinh-l.
V c cj
We find that x can be eliminated by squaring and subtracting, and
that „

This equation cannot be solved explicitly, but it can be shewn that

there is only one positive value of c that will satisfy the equation*, so
that only one position of equilibrium is possible. An approximate
solution may be found by the help of tables of hyperbolic functions in
particular cases.
If the chain is nearly taut, the parameter c is large compared to the
length of the string. For if ifi, <p' are the inclinations to the horizontal at
the ends A, B, since CA = c tan ^r and CB = c tan ifi', therefore

but ifi and tfi' are nearly equal in this case, so that c must be large
compared to I.
12-231. E x a m p l e . A uniform chain of length I when tightly stretched
between two points at the same level has a sag k in the middle. Prove that
the length of the chain exceeds the distance between the points by
approximately.
If 2a be the distance between the ,— —
points and c the parameter of the "
catenary, we have
A;+ c = ccosh- (1),
c
and JZ = c s i n h - (2),
c
so that {k + c)* = ci + ili, and 2kc=\V-k2 (3);
this gives the parameter of the catenary in the form
_l* k ...
c (4)
-8k~2 '
where, since k is small, the second term may be neglected in comparison
with the first.
Since a/c is small we may expand sinh - in (2) and retain only the
first two terms, so that

Z 2o
or - = £-
* v. Eouth's Analytical Statics, vol. I, § 447.
240 FLEXIBLE CHAINS AND STBINGS [XII
A first approximation a= JZ may be substituted in the term a /3c2,
3

so that 73

and substituting for c from (4) we get

12*3. The Parabolic Chain. If particles of equal weight are so dis-

tributed along a light string hanging between two fixed points that the
horizontal distance between each pair of consecutive particles is the same,
then the particles lie on a parabola whose axis is vertical.
Let O, P, Q, R be four of the particles. Take 0 as origin and hori-
zontal and vertical axes Ox, Oy.
Let a be the horizontal distance be- y
tween each pair of consecutive particles,
and let yx, ?/2, y3 denote the ordinates of
P, Q, R and dlt 02, 83 the inclinations of
OP, PQ, QR to the vertical. Then, from
6*5, since the weights are equal,
cot 83 — cot 02 = cot 02 — cot 0X,
r 2/3-2/2 2/2-2/1. _2/i
a a a' a
Therefore (1).
Now one parabola with axis vertical can be drawn to pass through
three assigned points, say O, P, Q, for the equation
y = Ax2 + Bx (2)
represents a parabola with its axis vertical passing through O, and A
and B can be so chosen that the curve will pass through any two
assigned points.
The co-ordinates of P, Q and R are a, yt; 2a, j / 2 and 3a, y3, and the
curve (2) passes through P and Q if

and y2 = iAa2 + 2Ba.

Therefore, from (1),
2/s = 3 (2/2 - 2/i) =
shewing that the same parabola passes through R, and similarly
through all the points on the chain.
12*31. If the distances between the particles be indefinitely
diminished, we arrive at the case of a light string or chain supporting a
continuous load whose horizontal distribution is uniform; i.e. the case

Hrr Till
 12-231-12-33] SUSPENSION BRIDGE 241

of a suspension bridge in which the weights of other parts are negligible

in comparison with that of the horizontal roadway.
The vertical rods that link the roadway to the parabolic chain are
spaced at equal horizontal distances and so carry equal loads.
12-32. The parabolic form of a chain carrying a continuous load
uniformly distributed horizontally may be proved simply as follows:
Let 0 be the lowest point of the chain, and P any point (x, y) referred
to horizontal and vertical axes
through O. The weight carried by
OP is proportional to ON and may
be denoted by wx, and this force
acts through the middle point Q of
ON, since the load is uniformly
distributed horizontally. The por-
tion OP of the chain is also acted
upon by the tensions at its ends, say
T at P and To at 0. The former
depends on the position of P on the
chain, and the latter is a definite 'wx
constant. The three forces acting upon the portion OP of the chain
must meet in the point Q, and PNQ is a triangle of forces, so that
WX To
PN^WQ' o r
which represents a parabola of latus rectum 2TJW.
12*33. E x a m p l e s , (i) In a suspension bridge of 400 ft. span and
4:0 ft. dip the whole weight supported by the two chains is 2 tons per hori-
zontal foot. Find the horizontal tension in the chains and the tension at
the point of support. [T.]

200
Ji There are two parallel chains each carrying half the load. If OP
represents half of one of the chains, the load that it carries is 200 tons,
and with the notation of 12*32
QN=100, and PiV=40 so that PQ = 10^/116= 107-7.
Then by the triangle of forces

NQ QP PN'
To T 200 _
or
100 ~ 107-7 ~~l o = 5 >
m
so that 0 = 500 tons and T = 538-5 tons
RS 16
242 FLEXIBLE CHAINS AND STRINGS [XII
(ii) If a telegraph wire has a span of 75 yd. and a sag in the middle of
1 ft., shew that the tension in the wire is approximately 480 lb. weight,
when the weight of the wire is 400 lb. per mile. [T.]
Considering the equilibrium of one-half of the wire, its weight is
V * TW<J Ik. acting at a distance from the end P approximately equal
to one-quarter of the span, i.e. af.s ft.

Hence, if To is the tension at the lowest point 0, by taking moments

about P for the forces acting on the portion OP, we get
21o = ¥xi 4 A°o>< a ! £
= 47941b.
The tension in the wire at other points has a vertical component
which is negligible in comparison with the horizontal component
because the wire is nearly horizontal at all points. Hence the tension
at all points is approximately 480 lb. weight.

12*4. Chain of Variable Density or Thickness. Let p be

the density and co the area of the cross section at any point,
then the mass of a small element
of length 8s at this point is pwSs,
or mhs, where m is variable on
account of the variability of p or
of OJ or of both p and w.
Hence if To and T denote the
tensions at the lowest point C and
at a point P, where CP=s and the
tangent at P makes an angle ifs with the horizontal, by resolving
horizontally and vertically for the portion CP, we get
Tcosil>=T0
rs
and Tsmtfi= mgds;
Jo
therefore Tot&nJ>= mgds,
Jo
and, by differentiation,

ds
or m cos2 0-^=constant (1).
dtp
12-33-12-5] CATENARY OF UNIFORM STRENGTH 243
If the form of the curve be given, this equation determines
the value of m; and, if m be given, the integral of equation (1)
gives the intrinsic equation of the curve.
12*41. E x a m p l e . Find the law of density ofa chain which hangs in a
cycloid.
The intrinsic equation of a cycloid is

so that 12*4 (1) gives m cos3 $=constant,

or «ioc sec 3 <ji.
12'5. Catenary of Uniform Strength. I.e. a chain of uni-
form material, suspended from two fixed points, and such
that the area of its cross section is proportional to the tension.
It follows from this definition that the weight of an element
is proportional to the cross section so that if wBs denotes the
weight of an element 8s at a place where the tension is T, then
T = we, where c is a constant.
Hence, as in 12*4, we have

and Tsini[i= wds,

If8 1 CST 1 fs
so that tani/' = ^ - I wds = jff I —ds =C- I aectjtds,
-'oJo J-oJ o c Jo
and, by differentiation,
dip sec if;
T
ds c
dift cos ifi 1 dx
ds c c ds'
Therefore x = a// + A,
and, if the origin is taken at the lowest point, the constant A is
zero, and x = ap.
dy x
Then -=^ = t a n ^ = tan~,
dx c
so that y = clog sec-.
c
The curve is symmetrical about the lowest point, and y-> oo
as X->±^TTC, SO that there are vertical asymptotes and the
span cannot exceed ire.
16-2
244 FLEXIBLE CHAINS AND STRINGS [XII
12-6. Elastic Strings. The law relating the tension of an
elastic string to its extension was enunciated in 11*8: but in
the cases there considered it was assumed that the tension and
extension of all parts of the string were the same. In the case
of a heavy string this assumption does not hold, but Hooke's
Law is to be applied to each small element of the string.
It is clear that after an elastic string has been stretched and
taken up a position of equilibrium, the equations of equilibrium
must be of the same form as for an inextensible string, save for
the fact that the 'line density' or mass per unit length is the
unknown line density which the string possesses after being
stretched.
Thus suppose that mSs is the mass of an element 8s of the
stretched string, whose line density and length when un-
stretched are m0 and Ss0, then

and by Hooke's Law the tension of the element is

therefore m=i
and this is the additional information required in solving a
problem of an elastic string.
12"61. A heavy elastic string hangs vertically and supports a weight W.
Let OA represent the stretched string carrying a weight W
at A.
Let P be a point at a distance x from 0, PQ an element dx,
and I the length OA.
Let l0, x0, dx0 be the unstretched lengths of I, x and dx, and
let w be the weight of unit length of unstretched string.
Then by Hooke's Law the tension at P is -P
Q

But the tension at P supports the weight W and the weight

of the portion PA of the string, so that
T=W+w(lo-xo) (2).
Therefore, from (1) and (2)
12-6-12-62] ELASTIC CATENARY 245
and, by integration
(

there being no constant of integration since x and x0 vanish together.

Putting x0 —10, we get for the whole extension

l
Also, when W is zero, -^S - ^ ? .
and the string is doubled in length, i.e. l = 2l0 if A = £wJ0 = half the
weight of the string.
12*62. The Elastic Catenary. An elastic string which is uniform
when unstretched hangs between two fixed points.
Using the figure of 12'4 to represent the equilibrium position of the
string, and taking w as the weight of unit length of unstretched string,
let GP = 8 and let s0 be the unstretched length of GP so that ws0 is its
weight.
Then by resolving horizontally and vertically for OP we have
Tcoaifi=T0 — wc say (1)
and Tsin >p = ws0 (2),
so that by squaring and adding
T 2 =w a (c a +V) (3).
Also, by Hooke's Law,
( ^ (4).
XT dx we „ ...
Now -j-= cos i/r = - = , from (1);

therefore x = wc I 7 f r =wc I (m + T ) da0, from (4)

J J- J \J V

1 0
- ^-—,^ (5),
with no constant of integration if the axis of y passes through the
lowest point of the string.
Similarly ^ = sin^ = ^-°, from (2).

Therefore y = w J^ds =w f S ^ ^ + ^ds,, from (3)

+^ (6),
 246 FLEXIBLE CHAINS AND STRINGS [XII
with no constant of integration if the axis of re is at a depth c below the
lowest point of the string.
Equations (5) and (6) express the co-ordinates of a point on the curve
in terms of the parameter s 0 .
12"7. Miscellaneous Examples, (i) A rope is passed round a
framework of rectangular beams with rounded corners as shewninthe figure.
If the diameter of the rope d is small compared with the dimensions of the

framework, shew that the condition that it should not slip when a tension is
applied at A, the end C being free, is jne2ir'1 > l/8d, where I is measured as in
the figure and fi is the coefficient of friction both between the rope and the
beams and between the two parts of the rope.
[Assume that the frictions at B act along BC] [T.]
Let R be the pressure at B between the framework and the under
rope and also between the two ropes.
Let T and T' be the tensions in the
under and upper ropes at B.
The upward pressure R on an
element of the upper rope at B
must be balanced by the resolved
part of the tensions T' on either side, so that

The friction available to prevent the sliding of the lower rope in the
direction OB is then 2/j.R or -~— T', and this must exceed the tension
T in order to maintain equilibrium.
But in passing along the rope, from the lower to the upper position
at B, there are two turns in the rope at each of which the rope turns
through an angle \TT in getting at right angles to the plane of the figure
and through another \TT in getting back into the plane of thefigure,so
that the total angle turned through is 2ir, and, by 12-12, T'=
12-62, 12-7] EXAMPLES 247
Therefore the condition that the rope should not slip is

2 i
or tj.e i *>ll8d.
(ii) Prove that in order that a chain of variable density may hang in a
given curve, the weight per unit of the horizontal co-ordinate x must be
Tod%yjdx%, where To is the horizontal tension.
A chain whose ends are fixed at A and B revolves about AB with con-
stant angular velocity (gravity neglected). Prove that if it has the form of a
curve of sines having points of zero curvature at A and B, the mass of any
portion must be proportional to the orthogonal projection of that portion
on the line AB. [T.]
For the first part, using the figure of 12'4 and taking w as the weight
per unit length of the horizontal co-ordinate x, the weight of the portion
fx
OP is I wdx, so that
JO 2 7 COSl/r=r 0 ,
fx
and Tsin\jj= I wdx.
Jo
fx
Therefore Totan.tb= I wdx,
Jo

whence, by differentiation w=

For the second part let y = asinx be the curve of the chain, with A
as origin and 4 B a s axis of x.

Let P be the point (x, y) and mSs the mass of a small element at P.
This has an acceleration w2y along the ordinate PN, and the forces
producing this acceleration are the tangential and normal com-
ponents of the tension at the ends of the element, i.e. ST along the
tangent and T 8<j> along the inward normal. Hence, by resolving along
the tangent, we get mSs.^yahx^ -ST.
Also, by resolving parallel to AB for the portion CP we have
248 F L E X I B L E CHAINS AND STRINGS [XII
Therefore the mass of the portion AP is
{' J f dT To (sec*
I mds=- -5— : — T = ?
Jo J co22/sini/r to2 J y
where y = aainx, so that taxi 1(1 = ~ —a cos x,

and sec2 i/r diji == — a sin a; (fa;.

s 71 fa; To;

/
*wefe = ^ dx = ^2 .
0 6J 2 Jo W
(iii) -4n extensible string, uniform when unstretched and of length I, lies
initially unstretched in a straight line on a rough horizontal plane. The
string is then pulled at one end in the direction of its line produced, with
a gradually increasing force, so that its acceleration is always infinitely
small. Prove that when the force is F, the extension of the string is
J-F21//J.WX, where W is the weight of the string, A the coefficient of elasticity,
jtt the coefficient offriction, and F < fiW. [T.]
L N P O F

Let OL represent the string when the force is F, and let x denote the
distance of any point P from the end O at which F is applied. Let x0 be
the unstretched length of the portion OP and T the tension at P. The
weight of OP is -~ W and the friction acting on it is ~- W, so that
v V

(1).

Now since F<fi.W, there is not enough force to move the whole
string, so we assume that when the force is F a certain portion ON has
been extended while NL remains unstretched and this implies that
there is a point N at which the tension vanishes. The unstretched
length of ON is obtained from (1) by putting T = 0, which gives
(2).
Again, by Hooke's Law, the tension a t P is given by
_, . (fa: — dx0
J A
- ~~dx^~'

so that from (1) dx — dx0 + ^— xg dx0 = -r-(fa;0,

A A
and, by integration,
x x
~ o'^"a -r~xo = "T a 'o (3),
id A A

there being no constant of integration since x and x0 vanish together

at 0.
12-7] EXAMPLES 249
The whole extension when the force is F is then the value of x — x0
when for xa we take the unstretched length of ON given by (2), and we
F

F IF
2A

EXAMPLES
1. A rope is coiled round two fixed bollards as shewn in the figure,
and one end is held with a force of 60 lb. Find the greatest force which
can be applied at the other end without causing the rope to slip. Take
the coefficient of friction between the rope and the bollards to be 0-2.
[S.]
IP

2. Three equally rough pegs A, B, C of the same circular cross

section are placed at the corners of an equilateral triangle of side a, so
that BO is horizontal and A above BC. Shew that the greatest weight
which can be supported by a weight W tied to the end of a string,
which is carried once round the pegs and never completely surrounds
any peg, is TFe6'"1', jn being the coefficient of friction. [C]
3. A uniform chain rests on the upper half of a smooth vertical
circle, its ends reaching to the horizontal diameter. Prove that the
pressure per unit length at the highest point is twice the weight per
unit length.
4. A uniform chain hangs round a smooth vertical circle. Prove that
if the pressure vanishes at the lowest point, then the tension at the
highest point is three times that at the lowest point.
5. A uniform chain hangs beneath the lower half of a smooth vertical
circle. Prove that if the pressure per unit length at the lowest point is
equal to the weight per unit length the tension at the ends must be half
as much again as the tension at the lowest point.
250 FLEXIBLE CHAINS AND STRINGS [XII
6. A uniform string rests in limiting equilibrium on a rough quad-
rant of a circle with one end at the lowest point. Prove that the angle
of friction A satisfies the relation
\ IT tan A = log tan 2A.
7. A uniform string rests on a smooth cycloidal curve whose axis is
vertical and vertex upwards, the string just reaching to the cusps of
the cycloid. Prove that the pressure at any point varies as the curvature.
8. If the tangents at the points P and Q of a catenary are at right
angles, prove that the tension at the middle point of the arc PQ is equal
to the weight of a length of the string equal to half the arc PQ.
9. A uniform string of weight W rests on a rough cycloidal curve
which has its axis vertical and vertex upwards, the string extending
from the vertex to a cusp. Shew that the least horizontal force applied
at the vertex that will cause the string to slip upwards is

where /n is the coefficient of friction.

10. A uniform chain is hung up by its two ends which are on the
same level and the sag in the middle is small; shew that the terminal
tension varies inversely as the sag. {S.]
11. A chain 40 ft. long, which weighs 1 lb. per foot, hangs between'
two points on the same level and has a sag of 5 ft. Find the parameter
of the catenary and the tension at a point of support. [I.]
12. A telegraph wire, stretched between two poles at distance a feet
apart, sags n feet in the middle; prove that the tension at the ends is
approximately

13. ACB is a telegraph wire, the straight line AB being horizontal

and of length 21, and C the middle point of the wire is at a distance h
below AB. Shew that the length of the wire is approximately

14. A chain of length 20 ft. and weight 10 lb. is stretched nearly

straight between two points at different levels. Assuming that vertically
below the middle point of the chord the chain is approximately parallel
to the chord and that the tension there is 100 lb. weight, prove that the
sag measured vertically from the middle point of the chord is approxi-
mately 3 in. [S.]
15. A suspension chain carries a load uniformly distributed on a
horizontal platform. The load is £ ton per foot length of the span of
600 ft., and the height of the point of support above the lowest point
of the chain is 50 ft. Find the greatest and least tensions in the chain,
neglecting its weight. [T.]
 EXAMPLES 251
16. The span of a suspension bridge is 100 ft. and the sag at the
middle of the chain is 10 ft.: if the total load on each chain is 25 tons,
find the greatest tension in each chain and the tension at the lowest
point. [S.]
17. A telegraph wire, of length I, hangs between two posts on the
same level, at distance a apart, the small sag in the centre being b;
shew that I — a = f b2/a, approximately, and that the least tension per
unit area of the section isfawa2/b,if w is the weight per unit volume of
the wire.
If the greatest tension is to be 15,000 lb. per square inch of section,
calculate the least sag allowable in a span of 100 yd. if the wire weighs
•28 lb. per cubic inch; and prove that the increase in length due to the
sagging is about 1 ft. per mile. [C]
18. If a telegraph wire has a sag of 30 in. in a span of 100 yd.,
calculate the least tensile stress per square inch of the section, assuming
the weight to be -28 lb. per cubic inch.
What extra length of wire would be required over 10 miles, supposing
the posts uniformly spaced at 100 yd. apart and the sag to be 30 in. in
each span? [C]
19. In a suspension bridge of 600 ft. span and 60 ft. dip the whole
weight supported is 2 tons per horizontal foot run. Find the horizontal
tension in each of the two chains and the tension in each at the points
of support.
A freely hanging chain of the same span carries the same load per
foot of its own length and has the same horizontal tension as in the case
above; find its dip and the tension at the points of support.
[cosh-8 =1-337.] [T.]
20. A box-kite isflyingat a height h, with a length I of wire paid out,
and with the vertex of the catenary on the ground; prove that at the
kite tan \ i/r = hjl, and that the tension there is equal to the weight of a
length i (i2 + W)lh of wire. [C]
21. A heavy uniform string rests on a smooth catenary with its axis
vertical and its vertex upwards; prove that the pressure on the curve
at any point varies inversely as the square of the distance of that point
below the directrix. [I.]
22. Two smooth rods are situated in the same vertical plane and
make equal angles a with the downward vertical. A uniform chain
of length I hangs by weightless rings from the rods. Prove that the
distance between the rings is
?cotasinh~ 1 tana. [I.]
23. A particle of mass m is suspended from a fixed point by a light
string which is blown from the vertical by a steady horizontal wind of
uniform velocity V. Assuming that the force exerted by the wind on
each element of length 8« of the string is normal to the element and of
252 FLEXIBLE CHAINS AND STRINGS [XII
magnitude KV* SS, V being the component of V normal to the element,
shew that the string hangs in a catenary.
If the wind pressure on the particle is negligible, prove that the
depth of the particle below the point of suspension is
clog '
o J
where c = mgji<V3, and I is the length of the string. [S.]
24. Prove that if a uniform chain lie in one plane under the action of
a system of force fixed in direction and of magnitude at any point
varying as cosec2 iji, where \jt is the inclination of the chain to the fixed
direction, the chain must lie along an arc of a circle. [I.]
25. A heavy string of length 21 is hung from two fixed points A, B
in the same horizontal line, at a distance apart equal to 2a. A weight
W is attached to a certain point of the string. Shew that the para-
meters of the two catenaries in which the string hangs are the same,
and shew that if W is in the middle of the string, and its weight is
great in comparison to that of the string, the parameter c is equal to
Wa/2wVP — cfi nearly, while if on the other hand the weight of W is
small in comparison with that of the string,
a ,W ( ,o
- +,r- ^cosh—
c 2w { c
w being the weight per unit length of the string. [S.]
26. A chain hangs freely in the form of an arc of a circle. Shew that
its weight per unit length at any point varies as the square of the
secant of the angle which the radius to that point makes with the
vertical. [S.]
27. If a heavy uniform string hangs in equilibrium over two smooth
pegs, prove that the free ends must be at the same level.
If the pegs are rough and in the same horizontal line and every point
of the string is on the point of slipping in the direction of the corre-
sponding tangent, prove that the lengths of the vertical portions are
in the ratio e^7^2*': 1, where /j, is the coefficient of friction and i/r = tan" 1 ,
c
21 being the length hanging between the pegs and c the parameter of
the catenary. [I.]
28. A heavy string of variable density is hung up from two points.
Prove that if Tlt Tit T3 are the tensions at points A, B,C, where the
inclinations of the tangents to the horizon are a —fi,a, a + /J, and wx,
w2 the weights of the parts of the string AB, BC respectively,
1 1 _ 2 cos ft

—x = —3. [S.]
 EXAMPLES 253
29. In any network of strings hanging in a vertical plane, if three
uniform inextensible strings of the same line density meet in a knot
and their directions at that point make angles 8, (/>, <jt with the hori-
zontal, prove that the parameters of the catenaries are numerically as
tan <j> — tan ifi: tan ^t — tan 9: tan 0 — tan fy. [I.]
30. A ring of weight wb is attached to the middle point of a string of
length I which hangs symmetrically over two smooth pegs in the same
horizontal line, the ends of the string being vertical. If w is the weight
per unit length of the string and 2a the distance apart of the pegs,
prove the equation to determine the parameter of the catenaries in
which the string hangs, 6 + Z = e°'c{6 + \/4c 2 + 62}; and shew that the
least value of I for which equilibrium is possible occurs when
l/(c-a)*-l/c» = 4/&«. [I.]
31. One end of a uniform rough string is fastened to a point P, at a
height h above a table, and part of the string rests on the table in a
vertical plane through P. Shew that the greatest length which can lie
on the table is the smaller root of the equation

where I is the length of the string. [I.]

32. One end of a rough uniform chain of length I is fastened to a
point on a vertical wall at a height h above the ground. Shew that the
greatest distance from the wall at which the free end of the chain will
rest on level ground is given by the expression

where u
and IL is the coefficient of friction. [S.]
33. A weightless string is attached to two points in the same vertical
line at a distance h apart. The wind produces a force k sin >fi per unit
length in the direction of the normal, where tp is the angle between the
tangent to the string and the horizontal, and k is a constant. Prove
that the tension T is the same at all points of the string and that it
cannot be less than kh/ir however long the string may be. [T.]
34. A uniform chain of length I and weight W hangs between two
fixed points at the same level and a weight is suspended from its middle
point so that the total sag in the middle is h. Shew that, if P is the pull
on either point of support, the total load is

35. Prove that in the catenary of uniform strength

254 FLEXIBLE CHAINS AND STRINGS [XII
36. One end A of a uniform string of length I isfixed,the other end B
moves along a horizontal line through A. Shew that the locus of the
vertex of the catenary formed by the string referred to horizontal and
vertical axes through A is given by
{±g [C]
37. A uniform telegraph wire is made of a given material and such
a length I is stretched between two posts distant d apart as will produce
the least possible stress on the posts. Shew that Z = d/AsinhA, where A
satisfies the equation Atanh A= 1. It may be assumed that the ends of
the wire are at the same level. [C]
38. In a non-uniform string hanging under gravity the area of the
cross section at any point is inversely proportional to the tension.
Prove that the curve is an arc of a parabola with its axis vertical. [I.]
39. Find the intrinsic equation of the catenary formed by an elastic
thread which is uniform when unstretched. Shew that for a large
modulus it approximates to a common catenary, and for a small
modulus to a parabola. [I.]

ANSWERS
1. 4051b. 11. 37-5ft.; 42-51b.
15. 474-3tons, 450tons. 16. 33-66tons, 31-25tons.
17. 30in. 18. 15,1201b.; 9-9ft.
19. 750tons, 807-75tons, 126-5ft., 1003 tons.
39. s = c tan <p + -=r- {tan \ji + sec iji+log (tan </i+sec
 Chapter XIII
ELASTICITY
13*1. So far we have been dealing largely with such mathe-
matical fictions as rigid bodies, weightless rods or inexten-
sible strings, but in the present chapter we propose to bring
our investigations into closer touch with reality by showing
how to make allowance for the fact that bodies are not rigid
but undergo small changes in form when subject to the action
of force.
We shall confine our considerations to a few simple cases of
isotropic bodies. An isotropic body is such that if a sphere is
cut out of the body anywhere it possesses no directional pro-
perties of any kind, as distinct from a crystalline body or a
body of fibrous structure.
The simplest type of deformation or strain that a body can
undergo is a uniform extension, in which all elements of length
PQ in a certain direction are altered to elements of length
P'Q' such that the increment in length P'Q' — PQ is a certain
fraction e of the original length, i.e. e=(P'Q' -PQ)jPQ.
This fraction e is then called the extension.
Regarding a contraction as a negative extension it is
clearly possible for a body to be extended in more directions
than one. It is obvious, for example, that if a bar is extended
longitudinally it will in general contract laterally. We do not
propose, however, to analyse the different kinds of strain that
are possible but only to deal with some simple cases.

13-2. Extension of Bars. Young's Modulus. Within

certain limits the extension or compression of a uniform bar is
related to the tension or thrust in it by Hooke's Law (11*8), i.e.

where A is called the modulus of elasticity of the bar. It is,

 256 ELASTICITY [XIII
however, more usual to put A = Ecu, where w is the area of the
cross section of the bar and E is a constant known as Young's
modulus of elasticity for the material of which the bar is com-
posed. If we write e for the extension (I — lo)llo, we have

or the stress per unit area of the cross section is E times the
extension.
We assume as above that Hooke's Law is true, as a first
approximation at any rate, for small extensions or compres-
sions of bar or other bodies, i.e. that if equal and opposite
forces are applied longitudinally at the ends of a bar, extension
or compression will take place in accordance with this law
and that when the forces are removed the body returns to its
former state. But there is a limit in each case to the magnitude
of the displacement for which the law remains true, known as
the elastic limit or the limit of linear elasticity, and we must
assume that these limits are not exceeded. There are materials,
such as cast-iron and cement, which do not obey Hooke's
Law at all.

13*21. Variable Extension. When there are externally

applied forces the stress per unit area along a bar generally
varies from one cross section to another so that the extension is
also variable.
o P Q A

Q'

Let 0 be a fixed point on the bar OA when unstressed. Let

OP = x and PQ = Sx. Let 0, P', Q' A' be the points which
correspond to 0, P, Q, A in the strained condition, and let
OP' = x + (;, so that £ denotes the displacement of the point P.

Then OQ' = x + Sx + g + ^-Sx,

ox
so that P'Q' = (l + ^)sx.
13-2-13-23] ELASTIC ENERGY 257
Therefore the extension is given by
P'Q'-PQdg
6
PQ ~dx'
and if to is the cross section, the total stress in the bar at P is

dx
13*22. Extension of a Bar under its own weight. This is the
same problem as that of the extension of an elastic string in 12-61.
With our present notation and w for the weight of unit volume of the
bar, I for the total length and T for the tension at a point whose un-
strained distance from the upper end is x, we have

Therefore E
and E£
no constant of integration being required since £ and x vanish together.
The total increment in length of the bar is the value of £ when x=I*

13-23. Elastic Energy of a Longitudinal Strain. Consider

a unit cube of a body which is to be subject to a longitudinal
strain in the direction of one set of edges. Suppose that at any

/ /
1+e

instant these edges have increased in length to 1 + e under a

longitudinal stress of magnitude p per unit of area. Assuming
there to be no change of volume the other edges of the cube
will have contracted to 1 — e', such that (1 + e) (1 — e')2= 1, or
to the first order of e and e', e — 2e' = 0, so that e' = | e.
The stress p per unit area is connected with the extension
by the formula p = Ee, and the work done in a further small
extension of the cube is
e, or
RS
258 ELASTICITY [XIII
so that the work done in the extension from 0 to e is

2
f
or f i7e , neglecting higher powers of e.
This is a measure of the elastic energy per unit volume as-
suming that there is free (i.e. unopposed) lateral contraction.

13*3. Bending of Bars. Consider a straight bar to be

slightly bent in such a way that plane cross sections remain
plane sections but are no longer parallel but slightly inclined
to one another.

A'
A"

(i)

Let AEBF, A'E'B'F' be parallel cross sections of a bar at

a small distance 00' apart (fig. (i)). Let the bar be slightly bent
so that the planes of these sections intersect in a line CD
(fig. (ii)). Then the distance PP" between corresponding
elements of the sections in fig. (ii) will in general not be the
same as the corresponding distance PP' in fig. (i): but there
will be a pair of lines EF, E"F" parallel to CD infig.(ii) whose
distance apart has not been altered by the bending. These are
called the neutral lines. Take rectangular axes Ox, Oy in the
plane AEBF with Ox along EF; and consider a narrow strip
13-23, 13-3] BENDING OP BARS 259
PP' of the bar of cross section dxdy, which becomes PP"
when the bar is bent; P being the point (x, y). The extension
e of this strip is given by
PP"-PP'
«- pp> •
But from fig. (i) PP' = MM\ and MM' is unaltered by
bending and is equal to MM"; also
PP" PD .. .....

_y+K
~ R '
where for MD we write R, the radius of curvature of the curve
MM" or 00".
Therefore e = ~, and the stress in this strip of cross section
Jib
dxdy is By , ,
* -^dxdy.
Hence the resultant stress in the bar taken over the cross
section AEBF is
-g JJ ydxdy,
and this will vanish if the centroid of the cross section lies in
the neutral line Ox.
It follows that if a bar is unstressed (i.e. if the resultant
stress across every cross section is zero) but slightly bent by
the application of couples, the centroids of the cross sections
lie on a central line 00', which undergoes no extension in the
bending.
Consider next what are the sums of the moments about the
axes Oy and Ox of the stresses over the cross section AEBF.
The resultant moment about Oy is

•^ xydxdy,

and this will vanish if the axis Oy is a principal axis of inertia

of the cross section or simply an axis of symmetry of the cross
17-2
260 ELASTICITY [xni
section. In this case the couples necessary to maintain equili-
brium will lie entirely in the plane of the bending; the total
moment of the stresses about Ox being
E CC
-g IIyUxdy
El
or -rfr , where I is the moment of inertia of the area of the
cross section about the neutral line, and R is the radius of
curvature of the central line of the bar.
This expression EI/R measures the bending moment at any
cross section of the bar, as denned in Chapter vn.

13"31. In using the results of 13*3 it must be remembered that they

have only been obtained with certain assumptions, e.g. that plane
sections remain plane sections, and without taking account of the local
effects of concentrated loads or of points of support, and that when we
make use of these results to deduce others we must accept them with
reservations as to how far they will necessarily accord with the results
of experiment.

13-4. Applications. In Chapter vn we investigated the

method of measuring the shearing force and bending moment
at any point of a beam in an equilibrium position. We have in
13*3 obtained a measure of the bending moment in a strained
position of a beam in terms of its shape in this position and
Young's modulus for the material of which it is composed.
We shall now show how, by equating these two measures of
the bending moment, to determine the strained form of the
beam on the hypothesis that the strain is small.
We shall assume that the breadth and depth of the beam
are small compared to its length, and that the deviation from
the unstrained state is small, so that if we take an axis of x
along the unstrained direction of the beam it will be sufficient
to take d2yjdx2 as the measure of the curvature 1/J? of the bent
beam. We may also use a single symbol B to denote the
product El when the beam is of uniform cross section; the
d2y
bending moment is then B - j - | and B is called the flexural
rigidity of the beam.
13-3-13-42] BENDING OF BEAMS 261
13*41. Care must be taken about the sign of the bending moment.
Thus, if, in the figure, AB is a beam
divided at P, with the axes as shewn, -
thenfiff atP
ax B
= clockwise moment about P of
the forces acting upon PB, or
= counter-clockwise moment about P of the forces acting upon PA.
13-42. Examples, (i) A light beam of length I is supported at its ends
at the same level and carries a load W concentrated at a point at a distance
a from one end. [7*21 (i).]
Let AB be the beam,
AG=a, CB=b = l-a.

The forces of support at A and B are Wbjl and Wa/l. Taking an axis
of a; along AB and the axis of y vertically downwards at A, let P be the
point {x, y), then taking the bending moment at P, we have, for

so that B(f-t*na)=-l^z* (1),

where tan a is the gradient of the central line at A; and
B(y—o;tana)= ~ s - j - * 3 (2),
with no further constant of integration, since x and y vanish together.
Again,for a<x<I, B-~= j-(l—x),

so that B f^ + tan/3J = 2 - r - (I — x)a (3),

where f} is the acute angle which the beam makes with the horizontal
at B; and j ppa

with no further constant of integration since y vanishes when x = I.

Now there is no discontinuity at O where x = a, so that, when x = a,
(1) and (3) will give the same value for dyjdx and (2) and (4) will give
the same value for y.
262 ELASTICITY
„ . „. lWba* D, . , lWab*
Therefore B tan a — -x—=— = — B tan p + s —=—,
, „ , 1 Wba? _ , , . lTFa&3
and Ba tan a — s — j — = Bb tan /3 — s —=—;
1 TF
whence Bt
1W
and Btan^
Then the deflection at different points of the beam is given from (2)
and (4) thus
1 Wb
0<x<a, By = ±Y-{a{.a + 2b)x-x*} (5),

and a<x<l, By = g~?{6(2a + 6)(I-x)-(l-x) 3 } ...(6).

As a special case when the load is at the middle point of the beam, so
that a = b — \l, we have,

and
and the deflection at the middle point is
(ii) A light beam is clamped horizontally at one end and carries a load
W concentrated at the other end.
Let I be the length of the beam AB clamped at A.
Taking horizontal and vertical
axes through A, we have for the
bending moment at a point P (x, y) L
(•

so that B^=W{lx-\x%
no constant being necessary because, by hypothesis, dyjdx vanishes
when x=0.
Integrating again, we get

since x and y vanish together.

We find for the deflection at B, where x = I,

The result could have been predicted from the preceding example.
For if we fixed the ends of a beam of length 21 at the same level and
applied an upward force 2W at the middle point, the beam would be
horizontal in the middle and the deflection at the middle would, by
13-42, 13-43] BENDING 0 1 BEAMS 263
3
Ex. (i), be \Wl jB. But the downward force at each end would be W,
and either half of the beam might be regarded as clamped at one end
and carrying a load W at the other.
We solved this problem by taking moments about P for the forces
acting upon PB. We might obtain the same result by taking moments
about P for the forces acting upon AP; but in this case we must note
that the clamping of the beam at A means that there is a bending
moment at A as well as a supporting force W. If we represent this as a
couple M (counter-clockwise as it acts upon AB), we get by moments
about A for the whole beam
M-Wl=0.
Then, by momenta about P for the forces on AP, we have

Bpt=M-Wx,
dx*
or B ^
as before.

13*43. In the problems considered in 13*42 and in kindred

problems the equations for the deflections are linear and there-
fore solutions can be superposed and results obtained by
addition. For example, if we consider the case of a uniform
heavy beam supported at its ends we may add the deflection
to that obtained in Ex. (i) caused by a concentrated load on
a beam of negligible weight.
Uniform heavy beam, of length I with supports at the same level at its ends.

If w denotes the weight of unit length, the supporting forces are each
\wl, and for the bending moment at a point P (x, y) we have

so that B^-tanceW\vxx?-\wlx? (1),

where tan a is the gradient at x = 0.

Also B(y — a:tana) = ^jW«4 — ^wlx3 (2),
since x and y vanish together.
264 ELASTICITY [XIII
But y vanishes when x=I, therefore
(3),
and therefore, from (2),
By=^ivx(l-x)(la + lx-x2) (4).
This makes the sag at the middle point -
13*44. It is sometimes advantageous to make use of the
relations between shearing force and bending moment estab-
lished in 7*3; e.g. for a beam carrying a uniformly distributed
load w per unit length, the shearing force 8 and bending
moment M satisfy the relations
dS , dM
-j- =— w and -^———8.
dx dx

But M = Bp[, so that B^\=-8,

dx2 dx3
and B -=-^ = w, but subject to the conditions that at a free end
ax
d2y d3y
M = 0 and 8=0, i.e. -=-f = 0 and-=-^ = 0; whereas at a clamped
end there is a bending moment, but the direction being fixed
we have - / = 0 instead of 3-f2 = 0.
dx dx

13*45. The Theorem of Three Moments. Let a uniformly

loaded beam be supported at a number of points at the same
level and let MA, MB, Mo denote the bending moments at

Mc

Sc

three consecutive points of support A, B, C. Let the shearing

forces be as indicated in the diagram, taking note that there
will be a discontinuity in the shearing force at each point of
support, and that the pressure on the support B, say, will be
SB.-8B.
Let AB = l, BG = V, and let w be the load per unit length.
13-43-13-45] T H E O R E M OF THREE M O M E N T S 265
By taking moments about B and C for A B and B C respectively,
we get
l8A. + MB-MA-±wl*=0, and V8B, + Mc-MB-\wl'* = 0
(1).
Then for the bending moment at a point in AB at a distance
xfromA we have

so that B^--tan.a\ = \wx* + MAx-%8A.x* (2),

where tan a is the value of dyjdx at A; and
B(y-xta,no.) = i4Wxi+$MAxz-l8A,x3 (3),
since x and y vanish together at A.
But y also vanishes when x=I, therefore
-Bt&nx = ^wl3 + ^MAl-i8Ad2 (4),
and by substituting this value in (2) and again putting x=l,
wefindfor the value of dyjdx at B

and substituting for 8A, from (1), this gives

d
£ (5)
for the value of dy\dx at B.
In like manner by taking an origin at B and writing down
the bending moment at any point of BC, we get instead of (4)
the equation

where tan j8 is the value of dyjdx at B.

Substituting in this the value of 8B> from (1), we get

(6).
But tan/J is the dyjdx of (5), so that, by equating the two
values, we get
l(MA + 2MB) + l'(2MB + Mc) = iw(P + r3) ...(7),
which is known as 'the equation of three moments'.
266 ELASTICITY [xm
13*5. Combined Extension and Bending. The case of
combined extension and bending may be illustrated by the
following example:
A uniform beam is held inclined at any angle to the vertical in collinear
clamps which are at a distance apart equal to the natural length of the beam.
Prove thai at every inclination the forces on the clamps are equal and
vertical, and calculate the couples. [T.]
Let AB be the beam clamped at A and B, where AB ( = 1) is inclined
at an angle a to the vertical. Let w be the weight of unit length of the
unstrained beam. Let X l 7 Yx and Xa, Y2 be the components of force
along and perpendicular to the beam at A and B andM x , M2 the couples.

The weight of the beam will cause a variable extension as well as

bending. Let x + t- be the projection on AB of a portion AP whose
unstretched length is x, and let y be the distance of P from AB.
Then, if T and <S denote the tension and shearing force at P, by re-
solving for the equilibrium of the portion AP whose weight is wx, we
have
-S+Y^wxsma. (1),
and .(2).
But by 13*21 if w denotes the cross section of the beam and E is
Young's modulus, ,.
dx

therefore from (2)

dx
and by integration Ew£ + X-^x = \wx^ cos a,
with no constant of integration since £ and x vanish together. But £ is
also zero when x = I, so that
13-5, 13-6] ELASTIC ENERGY 267
Also, by resolving along A B for the whole rod, we have
X x +X a = wl cos a,
therefore X2 = XX = ^wlcoaa. (3).
d2y
Again, the bending moment at P is M = B -=-y.
dec
But -£ = ^
therefore (1) gives B -=-| + Yx = wx sin a,
dec
so that, by integration
p (4),
Mx being the value of B -yrf at x = 0;

therefore fi/+\Y±X* = \WX*WD.O.+MXX (5),

and Bydx
+ iY1x3 = -£[wxisin.a + iM1xi (6),
with no other constants of integration since y and dyjdx both vanish
with x; but they also vanish for x = l so that from (5) and (6)

and ^Y1l
Whence we find that
F x = \ wl sin a, and MI = ^ wl2 sin a,
and, by resolving at right angles to AB for the whole rod,
Yx+Y2 = wl sin a,
therefore Yi=Y1 — ^wlsin.a (7).
I t follows from (3) and (7) that the resultant forces a t A and B are
vertical and each equal to %wl.

The couple M2 is the value of B -=-|[ a t x — I, but (4) may be written

B j - f + \wlx sin a = Jira;2 sin a + Mx,

and putting x=l, we have
Af2 = Mx = T% w>Z2 sin a.

13*6. Elastic Energy. When a bar or beam undergoes a

longitudinal strain due to stretching or bending within the
elastic limits of the material of which it is composed, its elastic
energy is \E<? per unit volume (13*23).
This formula may be applied to determine the elastic energy
268 ELASTICITY [XIII
of the beam in any of the previous examples. Thus for a bent
bar (13*3) e = yjB, and the elastic energy per unit length
1 W

integrated over the cross section

For example, in 13*42 (i), taking the values of -~ which correspond

to the two parts AG, OB of the beam, we have for the whole elastic
energy
v, where b = l—a,

which reduces to 5 57 .
2
Since from (5) or (6) the sag at the point C is ^1 —^—,
WaPb
therefore the
6 HI
elastic energy is one-half of the loss of potential energy due to the
descent of the weight W from the level AB.

13*7. General Equations for a thin rod bent in one plane.

Let PQ be an element 8s of the rod AB in its equilibrium
position. Let T, 8 and M be
the tension, shearing force and g+g
bending moment at P, and
T + 8T, 8 + 8S and M + 8M
the corresponding quantities
&tQ.
Let F 8s and O 8s be the tan- A.
gential and normal compo-
nents of the externally applied forces on the element 8s. Let
81^ be the angle between the tangents at P and Q, so measured
that s and ip increase together.
Then by resolving along the tangent and inward normal at
P we have
cos 8i[>-(S+8S) sin 8i/r-|-.FSs = O,
and
- S + (8 + 88) cos ty + (T + 8T) sin 8^ + 08s = 0.
13-6-13-8] GENERAL EQUATIONS 269
In the limit as Q approaches P these become

£ ! < • <•>•
and ^- + ^+0 = 0 (2),
ds p
where p is the radius of curvature at P.
Also, by taking moments about Q,

or
Since the strained form of the rod is not known, another
equation is necessary to determine the form in addition to T, 8
and M. This equation is
M =B(---\ (4),
\P Pol
where B is the flexural rigidity and p0 is the radius of curvature
at P before the deformation. On the hypotheses of 13*3, this
equation can be established by a similar proof, the extension
e of an element at a distance y from the neutral line in this case
being,
\P Po
13*8. Euler's Strut.* Consider the possible positions of
equilibrium of a long thin rod clamped vertically at its lower
end and carrying a weight W at its upper end, on
the hypothesis of a deflection of the upper end from
the vertical.
Let the rod be of length I and suppose that the
upper end is at a distance a from the vertical through
the lower end. Then, taking axes as shown in the
figure, for the bending moment at a point P (x, y)
we have

The primitive of this equation is

y = a + F cos V( W/B) x + O sin V( WjB) x,
* L. Euler, 1757.
270 ELASTICITY [an
and if we determine the constants F, 0 so as to make y = 0
when x = 0 and y=a when x = I, the solution becomes
M_ff[T sin{V(W/B)(l-x)}
y a l
-\ Sm{V(W/B}l)

But dyjdx is to vanish for x = 0, and this requires that

cos{-v/(TF/-B)Z} = 0, and the smallest value of I for which this
is possible is ^TT\/(B/W).
The inference to be drawn from this result is that if the rod
were shorter than this critical length \TT^{B\W) it would
remain vertical and merely contract under the load W, but if
the length were increased up to the critical length an equili-
brium form is that of a portion of a sine curve of small ampli-
tude, and it can be shewn by calculating the potential energy
that the vertical position is stable when the length is less than
the critical length and that when the critical length is slightly
exceeded the stability passes from the straight to the bent
form.
13*81. The problem of a long thin rod A B (fig. (i)) of length I whose
ends are subject to equal and opposite forces W is
seen to be subj ect to the same analysis as the problem
of 13'8, if we place an origin O at the middle point of
the rod, and in the result of 13*8 substitute JZ for I,
and the rod will bend under the given forces if

Similarly, if the ends G, D are both clamped in the

same straight line (fig. (ii)) and the rod be of length I
and subject to the same terminal forces W and we
draw a sine curve touching CD at O and D and of
small amplitude, the distance between its inflexions
A, B is %l and the problem for the portion A OB of
the rod is the same as for the rod AOB of fig. (i), so
that the condition for bending in this case is

13'9. Loaded Column. When a vertical column supports a load

the stress at every point of a cross section of the column ought to be of
the same sign—for, if the stress were a compression over part of the
cross section and a tension over the remaining part, the result might
be a fracture of the column.
In 13*3fig.(ii) EOF is the neutral line of a cross section of a bar and
the parts of the bar above the neutral line are extended and in a state
13-8, 13-9] LOADED COLUMN 271
of tension while the parts below are compressed and in a state of thrust;
and it is clear that in order that the stress may have the same sign at
all points of a cross section the neutral line must fall outside the
cross section.
The position of the neutral line depends on the state of stress and
this depends on the line of action of the resultant load.
Now the stress across an element dxdy of the cross section is ~ dxdy
(13*3), giving a resultant ^j jydxdy = ~ y, where a> is the area of the
cross section and y is the distance of its centroid from the neutral line.
And the line of action of the resultant stress is at a distance ?j from the
neutral line obtained by taking moments about the neutral line, thus
Eco.._ Eff_,^^_ P
or n = -^,
where k is the radius of gyration of the cross section about the neutral
line.
Now let Ox be the neutral line when
the load W acts in the line LG cutting w
the cross section in C; and let O be the
centroid of the cross section.
Then W is equal to the resultant stress
-^s~y> OO is y and OC is ij; and if the
G
shape of the cross section is known, the / ^ — S--^ V
dataarenowsufncienttodeterminewith- /x
in what area the point O must lie in
order that Ox may fall outside the cross section.
Thus if the section is a circle of radius a, by the theorem of parallel

and Ox will fall outside the circle if y > a. But

OC=i]— ^=y+-7-^, so that (?C=-p.,
and therefore we shall have y > a if OO < Ja.
This means that in order that the stress in a circular column may
nowhere be a tension, the line of action of the load must cut the cross
section within a concentric circle of radius Jo.

EXAMPLES
1. A uniform beam of length 6a rests symmetrically upon two sup-
ports distant 2a apart. Find the elevation of the highest point above
the horizontal line joining the supports. [I.]
2. A uniform horizontal beam, of length I and negligible weight,
is supported at the ends, and carries a weight W at the middle
272 ELASTICITY
point. Find the deflection at this point, and prove that the elastic
WH3
energy is g g ^ . [I.]

3. Prove that the elastic energy of a uniform beam of weight W and

length I supported at its ends at the same level is
4. Prove that the elastic energy of a uniform rod of weight W and
length I hanging from one end is $W*l/Ew.
5. Find the bending moment at any point of a uniform heavy beam
of length 2a supported at its middle point; find also the depression of
the ends and shew that it is increased if a weight W be suspended from
1 Wa3
each end by an amount 5 —=- , where B is the flexural constant. [I.]
o JD

6. Prove that the deflection of the middle point of a uniform homo-

geneous elastic beam of length 2a supported at two points distant 6
from the middle point and lying in the same horizontal line is

where w is the weight per unit length of the beam and B the flexural
rigidity. [I.]
7. A uniform slightly flexible rod AB, of length 2a, is supported at
its two ends, and also at its middle point M; the end supports are in a
horizontal line and the middle support at a small depth 8 below this
line. Shew that the reaction at M is decreased by an amount 6EIS/a3
as compared with its value if AMB were horizontal. [I.]
8. Prove that, if a heavy uniform beam of length I is supported at
two points at the same level, at one end and at a distance § I from
that end, then the beam is horizontal at the latter point.
9. A slightly flexible rod of length 2a has one end clamped hori-
zontally; a support is placed under the middle point of the rod so that
the free end is in the same horizontal line as the fixed end. Prove that
11 Wa3
the height of the middle point above the ends is . . „ , where B is the
flexural constant and W is the weight of the rod.
Shew also that the pressure on the support is f W. [I.]
10. A uniform thin beam of length I is clamped at one extremity A
so that the tangent at it is horizontal; and the other extremity B rests
on a fixed support in the same horizontal line with A. Shew that there
is a point of inflexion at a distance £Z from A, and that the tangent is
horizontal at a distance
(15-V33)Z/16
from A; and that the pressure on B is f the total weight of the beam.
 EXAMPLES 273
11. If a rod be clamped horizontally at each end at the same level
and the middle point be pulled upwards by a force through a distance
8 above the level of the ends, prove that the magnitude of the force is
2£B& W
—3—I- -p and that the bending couples at the ends are equal to
ft RX
—a -£iWa, where 2a is the length of the rod, W its weight, and B the
nexural rigidity. [I.]
12. Shew that, with the notation of 13*42 (i), if the ends of the beam
AB are clamped horizontally, the supporting forces at A and B are

and the bending moments at A and B are

Wab*/l* and Wa*bjl\
13. A light beam is supported at its ends at the same level. Prove
that the deflection at a point P when there is a load at Q is equal to
the deflection at Q when there is an equal load at P .
14. Prove the same theorem when the ends of the beam are clamped
in the same horizontal line.
15. A uniform beam rests with its ends and middle point supported
upon three rigid props at the same level. Prove that the pressure on
the middle prop is f of the weight of the beam.
Prove further, that if the weight of the beam, instead of being uni-
formly distributed, is concentrated in two equal weights at the middle
points of the two spans, the pressure on the middle prop is JJ of the
weight. [T.]
16. A thin uniform elastic beam of length I, which is straight when
unstrained, is supported at two points which are at the same level and
are equidistant from the centre of the beam. How far apart must the
supports be placed in order that the height of the middle of the beam
above this level may be as great as possible ? [T.]
17. A uniform beam of length 21 is supported at the ends: the deflec-
tion at the centre is y1 when a load rests on the centre, and is j/ 2 when
the same load rests on the middle of another shorter beam (of length 2a)
whose ends are supported by the former at points equidistant from the
middle. Prove that

18. A light beam PQ, of span 2a, rests on two supports at P and Q
in the same horizontal line. It is loaded so that the weight of the load
per unit length at any point X varies as the square of PX. Prove that
the deflection at the middle point of the beam is 89wa4/(1440B), where
B is theflexuralrigidity of the beam and w is the weight of the load per
unit length at Q. [T.]
RS l8
274 ELASTICITY [
19. A uniform beam ABG(AB = BG) is supported at A, B, 0 so
that A, B, C are horizontal and is loaded at D, E, the midpoints of
AB, BC respectively, with loads in the ratio 1:3. The weight of the
beam is neglected. Prove that the beam is horizontal at A. [T.]
20. A spring-board consists of a heavy uniform beam of length
a+b clamped horizontally at one end and with a support of small
height h at a distance a from this end. If a weight W is placed on the
free end, shew that the difference of level of that end before and after
the weight is placed upon it is Wb* (3a + 46)/12B, where B is the flexural
rigidity of the beam. [T.]
21. AD is a heavy, uniform beam resting on supports at its ends and
at points B, C in the horizontal line AD; AB — 2a, BO —a, 0D = a.
Find the ratios of the pressures on the four supports. [I.]
22. Shew that, if a thin rod of length I and flexural rigidity B is set
up vertically with its lowest point fixed and the rod passing through a
smooth ring fixed at a vertical height $1 above the lowest point and
carrying at the top a weight W, then the least value of I for which the
rod bends under the load is the smallest root of the equation tan $nl=nl,
where
n*=W/B. [T.]
23. A light uniform bar of length I is placed in a vertical position,
the lower end being clamped and the upper end free. A force, whose
horizontal and downward vertical components are F and W, is applied
at the centroid of the cross section of the upper end so as to cause
bending in the plane containing the axis of symmetry of a cross section
and the axis of the bar, W being smaller than the vertical force for
which the vertical position of the bar is unstable. Assuming the
inclination of the bar to the vertical to be everywhere small, shew that
the deflection of the free end is -^ ( r — 1), where n2 = W/B and
B is the flexural rigidity.
If the cross section of the bar is a circle of radius a, shew that, in
order that the normal stress across a cross section should nowhere be
a tension, ni
W tan.nl A

ANSWERS
w denotes weight per unit length
2. W13I^B. 5. w
16. J(9-V69)Z. 21. 75:186:64:43.
 Chapter XIV
FORCES IN THREE DIMENSIONS

14*1. A system of forces acting on a rigid body can be reduced

in general to a force acting at a specified point of the body and
a couple.
Take a set of rectangular axes with the specified point 0 as
origin and let x, y, z be the co-ordinates of the point of applica-
tion .4 of a typical force P, and let the components of P parallel

'P.
M

to the axes be Px, Py, Pz. Let M be the projection of A on the

plane xy and L its projection on Ox. At L and at 0 introduce
pairs of equal and opposite forces equal and parallel to Pz.
These do not affect the resultant of the system. There are now
five forces of magnitude P z and they constitute a force Pz
along Oz, a couple of forces Pz with an arm ML, i.e. a couple
of moment yPz in or parallel to the plane yOz in the sense from
Oy to Oz, and a couple of forces Pz with an arm OL, i.e. a
couple of moment xPz in the plane zOx in the sense from Ox
to Oz.
Hence, adopting the convention of signs for couples ex-
plained in 4*641, it appears that the force Pz acting at A is
equivalent to an equal and parallel force at 0 together with a
couple whose axis is yPz along Ox and a couple whose axis is
— xPs along Oy.
18-2
276 FORCES IN THREE DIMENSIONS [XIV

By similar steps the forces Px, Py acting at A can be trans-

ferred to act along Ox and Oy, the necessary couples being
zPx along Oy and —yPx along Oz,
and xPy along Oz and — zPy along Ox.
Hence the force P acting at A is equivalent to component
forces Px, Py, Pz acting along the axes at 0, and couples of
moment yPz-zPy with axis along Ox,
zPx — xPz with axis along Oy,
x Py — y Px with axis along Oz.
All the forces of the given system can be treated in this way,
so that the given system is equivalent to a force B acting at 0
whose components are

together with a couple 0 whose components are

L = X(yPz-zPy), M = *(zPx-xPz), N=Z(xPy-yPx).
14*11. I t is clear that X, Y, Z are the algebraical sums of
the resolved parts of the given forces in the directions of the
co-ordinate axes, and by reference to 4-21 it is also clear that
L, M, N are the algebraical sums of the moments of the given
forces about the co-ordinate axes.
Also B is the resultant of the given forces moved parallel to
themselves to act at the point 0 and is clearly independent of
the position of the point 0. Hence B2 or X2 + Y2 + Z2 is said
to be an invariant of the system of forces.
The values of 0 and its components depend on the choice
of origin. For if the system were transferred to a point 0' whose
co-ordinates are £, r\, £ referred to the chosen axes through O,
the only difference would be in the lengths of the arms of the
couples. Instead of x, y, z we should have x — g, y - r\, z — £, so
that the components of the resultant couple 0' would be
L', M', N', where

and similarly M' = M-t,X + £Z,

and N' = N
14-1-14-21] CONDITIONS OP EQUILIBRIUM 277
It follows that
L'X + M' Y + N'Z = LX + MY+NZ.
14*2. Conditions of Equilibrium. It is necessary for equili-
brium that there shall be neither resultant force nor resultant
couple; i.e. that i? = 0 and (2 = 0
or that X2+Y2 + Z2 = 0 and i 2 + M2 + N2 = 0.
These conditions can only be satisfied if
X=0, 7 = 0, Z = 0 and L = 0, M = 0, N = 0.
Since the choice of axes is arbitrary, this means that the alge-
braical sum of the resolved parts of the forces in every direction
must be zero and the algebraical sum of the moments of the
forces about every straight line must be zero.
To ensure the vanishing of R it will be sufficient to know
that the sums of the resolved parts of the forces in any three
directions not all parallel to the same plane are zero. For if
OA, OB, OG are lines in the given directions, the fact that R
has no component in direction OA means that R is either zero
or at right angles to OA, and it cannot be at right angles to
each of three non-coplanar directions, therefore it is zero.
Similarly, since couples are compounded by the vector law,
to ensure the vanishing of G it is sufficient to know that the
algebraical sums of the moments of the forces about any three
lines meeting in 0 and not coplanar are zero.
14 - 21. Special cases.
(1) When R is zero and G is not zero. The system reduces to
a couple G. The vector which represents the axis of this couple
may be placed at any point of
space. The algebraical sum of the
moments of the forces about any
line parallel to the axis of the couple / I6>^~^G C0S 8
is G, and the algebraical sum of the
moments of the forces about a line inclined at an angle 6 to
the axis of the couple is G cos 6.
(2) When G is zero. The system reduces to a single resultant
force R.
278 FORCES IN THREE DIMENSIONS [XIV

14*3. Poinsot's Central Axis. The Wrench. When neither

R nor 0 is zero.
Let the axis of the couple 0 make an angle 6 with the direc-
tion of R at the point 0 as infig.(i). G can then be resolved
into a couple Ocosd with axis along R and 6?sin0 with axis
perpendicular to R as infig.(ii), in which the vector GsinO is
supposed to be at right angles to the plane of the paper. The
couple G sin 6 is then represented by two forces of magnitude
R at a distance G sin 8/R apart, so placed that one of them
balances the given force ii! at 0 and the other acts in a parallel
line through O', as in fig. (iii). Removing the equal and
opposite forces at 0, there remain (fig. (iv)) the force R at 0'

Gcos0 G cos 0^ ) G cos 0

O O'
Gsln 6

(i) (ii) (iii) (iv)

and a couple G cos 9 whose axis is in the direction of R, so

that the system has been reduced to a single force R at 0' and
a couple whose axis coincides in direction with the force, or a
force along a certain line and a couple in a plane perpendicular
to the line. Such a combination of force and couple is called a
wrench. The line of action of the force is called Poinsot's
central axis after L. Poinsot whofirstinvestigated the theory *.
It is also called the axis of the wrench; the pitch of the wrench
is defined to be the ratio of the moment of the couple to the
magnitude of the force, and the magnitude of the force is
called the intensity of the wrench.
14*31. Equations of the Central Axis. With the notation of
14*11 let (£, rj, t) be a point on the central axis. The system re-
duces to a force R at this point and a couple whose magnitude
we will denote by T, whose axis coincides in direction with R.
* Mtments de Statique, 1804.
14-3-14-33] RESULTANT WRENCH 279
But the direction cosines of R are X/R, YjR, ZjR and those of
F are L'/V, M'/F, N'/F, so that

V + ^ _ M ^ X + ^ Z _ ^ + , _ _ ^
or ^ - ^ -R-p
(1),
where p is the pitch of the resultant wrench.
The equations with g, -q, £ as current co-ordinates are the
equations of the central axis.
By whatever process we reduce the forces to the equivalent
wrench we must arrive at the same force R and couple F. In
14*3 the couple of the wrench is 0 cos 8; therefore

But 8 is the angle between R, whose direction cosines are as

above, and 0, whose direction cosines are LjO, MjO, NjO, so
that LX + MY + NZ
R O *
Therefore LX + MY+NZ^RF (2)
is an invariant of the system of forces.
The pitch p of the resultant wrench may also be expressed
in terms of the component forces and couples, thus
LX + MY+NZ
P
~ X*+Y* + Z* '
14'32. As stated in 14*31 there can only be one central axis for a
given system of forces. For if there were two equivalent wrenches with
different axes, then one reversed would balance the other. But the
forces R are the same for both wrenches and when one is reversed they
would constitute a couple balancing a couple in a perpendicular plane,
which is not possible.
14*33. Condition for a Single Resultant Force. When F is
zero and R is not zero, the system has a single resultant R.
The condition for this is by 14*31 (2)
LX + MY+NZ = 0 (1).
In this case, at any point (£, -q, £) on the line of action of the
single resultant, the components of couple L', M', N' vanish:
280 FORCES IN THREE DIMENSIONS [XIV
i.e. ~ " "~

and
In virtue of (1) these three equations are not independent,
but any two of them may be regarded as the equations of the
line of action of the single resultant force.
14*34. Given the resultant wrench to find the force and couple at any
point.
Let R, F be the force and couple of the wrench and a line Oz its axis
—fig- (i).
Tofindthe force and couple at a point O'; let O'O = h be perpendicular
to Oz. At 0' introduce opposite forces equal and parallel to R—fig. (ii).
.Z
R

O' ^Or

(i) (ii) (nil (iv)

One of these constitutes, with R at O, a couple of moment Rh, whose
axis is at right angles to the plane O'Oz; and the axis of the couple T
can be placed at 0'—fig. (iii). The two couples T and Rh can then be
compounded into a single couple G whose axis makes an angle 6 with
R, where G* = T2 + R*W and tan 0 = i?A/r,
and thus the wrench in the line Oz is equivalent to the force R and
couple G at O''. ^
14-341. Example. Forces P, Q, R,
P, Q, R act along the edges BC, CA,AB,
AD, BD, CD of a regular tetrahedron
ABCD. Shew that they are equivalent to
a wrench of pitch a/2\/2, where a is the
length of an edge. [I.]
It is convenient to take as the six
edges of the tetrahedron six diagonals
of the faces of a cube as in the figure.
Then, taking axes at the centre of the
cube parallel to its edges, we have

2R 2Q a
and N=-
for the length of the edge of the cube is ajy/2.
14-33-14-4] INVARIANTS 281
Then the pitch being numerically equal to

is a/2\/2.
14*4. The Invariants. The invariants of a system of forces are the
resultant force B, and the expression LX + MY + NZ which is equal
to the product of the force and couple of the resultant wrench and will
be denoted by I.
Since R is the resultant of all the forces moved parallel to themselves
to act at a point, it follows from 3*31 that for any system of forces
Plf P 2 , P 3 , ... JR2 = S

where 0rs denotes the angle between the forces P r , Ps and the latter
sum includes the products of all the forces taken in pairs.
To express I in terms of the forces, their inclinations and their dis-
tances apart:
(i) Consider first the case of two forces Plt P 2 inclined at an angle 8
at a distance h apart. Take the axis Oz along
P t and the axis Ox along the common perpen-
dicular to the lines of action.
Then
X = 0 , F=P 2 sin0, Z = P 1 + P a cos0,
£ = 0, M=-P2hcos8, N=P2hsia8. h
Therefore y.
I=LX + MY+NZ=P1Pihsin8.
(ii) A force B and a couple G whose axis makes an angle 8 with the
direction of B. Since, by 14-3, these are equivalent to a wrench whose
force is B and couple G cos 8, therefore in this case
I = BG cos 0.
(iii) Two couples. The couples can be so placed that all the four forces
are parallel and then by (i), since 0 = 0, I is zero for each pair of the
forces and therefore for the couples.
(iv) General case. By definition I is a sum of products of forces and
couples and is therefore of two dimensions in force. We may therefore
assume that I is a quadratic expression in terms of the forces and write

where the coefficients do not depend upon the magnitudes of the forces
but only on their relative positions.
To find the coefficients let us suppose that all the forces are zero
except P j and P 2 , then from (i) we have

where 7i18 denotes the shortest distance and 012 the angle between Px
and P 8 .
282 FORCES IN THREE DIMENSIONS [XIV
By comparison it follows that

and 2ala = hxi sin #12.

Similarly the other coefficients are determined and we have, for any
number of forces, I = SP f PA . sin Qn,
where the summation includes the products of all the forces taken in
pairs.
It is evident that in this sum there may be terms of opposite signs
and we need a convention for the determination of the sign of each term.
Each term may be regarded as the product of a force Pr and the
moment of a force Ps about the line of action of Pr. We must then
choose either a right-handed or a left-handed screw convention (4*641)
and take the term as positive or negative according as the vector
which represents the moment of P, about P r is in the same sense as the
force P r or in the opposite sense.
(v) Two wrenches of intensities Px, P 2 and pitches pt ,p2ata distance
h apart and inclined at an angle 8. There are two forces Px, P 2 and two
couples px Px tP^Pi, and we have the following contributions:
from the two forces P1Pihsin 8 by (i)
from the two couples 0 by (ii)
from the first force and couple
from the second force and couple
from the first force and second couple P1.p2P2cos8
from the second force and first couple P 2 .px Px cos 8
Therefore / =px Pf +p2 P 2 2 + Px P 2 {(px +p3) cos8 + h sin 8}.
14-41. Geometrical Representation of the Moment of a
Force about a Line.
Lemma. The volume of a tetrahedron is one-sixth of the product of a
pair of opposite edges, the shortest distance between them and the sine of
their inclination.
Let DABC be the tetrahedron. Complete the parallelogram ABEG,
and by taking parallel planes com-
plete the parallelepiped as shewn in
the figure.
Then the volume DABC
= f base x height
= \ volume parallelepiped
= J area BEFD x distance between
BEFD and ACGH
= \BD. BE sin DBE x shortest distance between BD and A G
. AC x shortest distance between BD and AC x sin {BD. AC).
14-4-14-5] GEOMETRICAL REPRESENTATION 283
Now let P be a force in a line AB, then by definition its moment
about a line CD is Ph sin 8, where h is the
shortest distance and 8 the angle between
the lines.
But, if V is the volume of the tetrahedron
ABCD,
V = iAB.CD.hain8;
therefore the moment of P about CD h
6F.P
'AB.CD'
The choice of the lengths AB, CD is arbitrary. If we like to take
them of unit length, then the moment of P about CD will be measured
by 6P times the volume of ABCD.
Further, if two forces P j , P 2 are represented by straight lines AB,
CD, then by 14*4 (i) the invariant / of the forces is
P1 P2hsinff, or AB.CDhsin.6,
i.e. six times the volume of the tetrahedron ABCD, with due regard to
sign as explained in 14'4 (iv).
14*42. Example. If ABCD be a tetrahedron and forces are repre-
sented by pBC, qCA, rAB, p'DA, q'DB, r'DC, shew that they are equi-
valent to a single force if pp' + qq' + rr' = 0. [I.]
By 14*33 the condition for a single resultant force is
LX + MY+NZ = 0,
or the vanishing of the invariant / . But
by 14*4 (i) two forces which intersect con-
tribute nothing to / , so that the only
contributions in this case arise from the
forces in pairs of opposite sides of the
tetrahedron; i.e. from pBC and p'DA,
from qCA and q'DB, and from rAB and
r'DC.
And from 14*41 the contribution oipBC and p'DA is
dpp' x volume DABG,
and similarly for the other pairs, so that the invariant / vanishes if

14*5. To find a wrench equivalent to two forces P, P' in-

clined at an angle 6 at a distance h apart.
(i) The intensity R of the wrench and the pitch w may be
written down at once by equating invariants. Thus

and
284 FORCES IN THREE DIMENSIONS [XIV
(ii) Or we may proceed thus:
Let the shortest distance between the forces be AA'. Determine an
angle a such that Psina = P / sin(0-a) (1).
Then parallel lines Az, A'z' may be drawn making angles a and 0 — a.
with P and P ' . Resolve P into P cos a and P sin a, and P ' into
P ' cos (0 — a) and P ' sin (0 — a.). Two of these forces P cos a, P ' cos (# — a)
act in parallel lines Az, A'z' and the other two are by (1) equal and
opposite parallel forces in a plane perpendicular to Az and A'z'.

P'cos(e-a)

'A'
Psln a
Also since Psina = P'sin(0-a),
t h e r e f o r esin a cos a 1
P W - p + P'c O s0-
and the moment of the couple is Pwsina or PP'h sin 0/B as above.
The position of the wrench is found from the fact that B is the
resultant of the parallel forces Pcosa and P'cos(0—a), and therefore
the axis of the wrench divides AA' in the ratio
P'cos(0-a):Peosa,
or P' (cos 0 cos a + sin 0 sin a): Pcosa,
i.e. P' cos 0 (P + P ' cos 0) + P' 2 sin2 0: P (P + P' cos 0),
or P ' ( P ' + Pcos 0) : P ( P + P'cos 0).
14*51. Tofindthe resultant of two given wrenches (P, p), (P't p') at a
distanceftapart and inclined at an angle 0.
By equating invariants we get for the intensity B and pitch w of the
equivalent wrench R2 _ p 2 + p>i + 2pp' c o g e>
and
Further, if A A' be the shortest distance be-
tween the axes of the given wrenches neither
wrench has any moment about AA', so that
the resultant wrench can have no moment
about A A' and its axis must therefore intersect
AA' at right angles.
The direction of B is determined by the angle
a of 14*5 (1), and its distance from A say is A'
obtained by equating moments about a line Ay
perpendicular to AA' and Az. Thus

which by 14*5 (2) reduces to

B2.OA = PP'(p-p')sia6+P'(Pcos8+P')h.
14-5-14-63] NTTL POINTS AND PLANES 285
14*6. Nul Points, Lines and Planes. In relation to a given
system of forces a nul line is a line about which the sum of the
moments of the forces is zero. Since a system of forces can be
reduced to a single force R acting at a chosen point 0 together
with a couple G, therefore a line through 0 is a nul line if, and
only if, it lies in the plane of the couple 0. This plane is called
the nul plane of the point 0 and 0 is called the nul point of the
plane. In other words the nul plane of the point 0 is the plane
of the principal couple 0 at the point 0.

14*61. Consider a line in the nul plane of 0 but not passing through
O. If it were a nul line the moment of R about it would be zero, which
requires either that R is zero or that R itself lies in the nul plane. In the
one case the system is equivalent to a couple and in the other to a single
force in the nul plane of the point 0. Also when R lies in the nul plane
of the point 0 every line in the plane is a nul line and every point of the
plane is a nul point. And in both the cases included in this paragraph
the invariant / of the system is zero.

14*62. Prom the definitions of 14*6 it is evident that the

nul planes of all points on a nul line contain the line, so that in
general as a point travels along a nul line the nul plane of the
point turns round the line.
It follows that if the nul plane of a point A passes through B
then the nul plane of B passes through A, for the line AB is a
nul line in both the nul planes.

14*63. The equation of the nul plane of a point (f, ij, £). I n
the notation of 14*11 it is the plane of the couple (?', and the
direction cosines of the normal to the plane are proportional to
L', M', N'. Therefore the equation of the plane is

or, giving L', M', N' their values,

which reduces to
+ {M-lX + SZ)y + (N-iT + 7iX)z
...(1).
286 FORCES IN THREE DIMENSIONS [XIV
14-64. To find the nul point of a given plane.
Take any two points on the given plane and as in 14-63
write down the equations of their nul planes. These planes
intersect the given plane in the point required.
14-641. Example. Find the nul point of the plane x + y + z = 0,/or the
force system (X, Y, Z, L, M, N).
The points (0, 0, 0) and (1, — 1, 0) both lie on the given plane
x+y +z= 0 (1).
Using 14-63 (1), the nul plane of (0, 0, 0) is
Lx + My + Nz = 0 (2),
and the nul plane of (1, — 1, 0) is
(L + Z)x+(M + Z)y + (N-Y-X)z = L-M (3).
From (1) and (2) we have

N-M~ L-N~M-L'
and from (3) these also
L-M

1
~X+Y+Z'
These relations give the co-ordinates of the required nul point.
14-7. Conjugate Forces. A system of forces can in general
be reduced to two forces one of which acts in a given straight line.
Let the system be equivalent to a force R acting at any
point O in the given straight line OA together with a couple O,
fig. (i). The plane through OA and R cuts the plane of the

couple G in another line OB through 0. Let R be resolved into

oblique components P, P' in the lines OA, OB, fig. (ii). The
couple 6 may then be represented by two equal and opposite
parallel forces of magnitude P' so placed that one of them
balances the force P' along OB, fig. (iii). The system has then
14-64-14-73] CONJUGATE FORCES 287
been reduced to a force P along OA and a force P' acting in the
nul plane of the point 0, fig. (iv). The magnitudes of the forces
are connected by the relations
P _ P' _ R
sinBOR~ sinAOR~ sinAOB'
The forces P and P' are called conjugate forces and the lines
in which they act are called conjugate lines with respect to the
given system of forces.
Since the conjugate of OA lies in the nul plane of 0, therefore
as the point 0 moves along OA its nul plane turns round this
conjugate line.
14*71. Exceptional Cases.
(i) B = 0. (a) The line OA may be inclined to the plane of the
couple G; then the force P in OA is zero and the conjugate force P ' is at
an infinite distance and also zero. (/?) OA may lie in the plane of the
couple; the force P is then one force of the couple and its conjugate is
the equal and opposite parallel force, the distance between them
depending on the magnitude of P which may be chosen arbitrarily.
(ii) B lies in the nul plane of 0. In this case the system reduces to a
single force B. (a) If OA intersects the single force B, then any line
through 0 in the plane of OA and B may be taken as the conjugate line,
and B may be resolved along OA and the chosen direction. (/J) If OA
does not intersect the single force B, then the force P in OA is zero
and the conjugate is simply the force B itself.
(iii) OA is a nul line. I.e. OA lies in the nul plane of 0. In fig. (i) of
14-7 BOA is then a straight line in the nul plane, and the components
of B in directions OA, OB are both infinite, and the line OA is its own
conjugate.
14-72. To shew that the conjugates of all straight lines in the nul plane
of a point 0 must pass through 0.
Let I be a line in the nul plane of 0 and m its conjugate. Any line OK
in the nul plane of 0 is a nul line, but it meets I because I lies in the
plane, and the conjugate forces in I and m have a zero moment about it,
therefore it also meets m (rejecting the possibility of m being parallel to
OK because of the arbitrariness of the direction of the latter). Therefore
m meets every line through 0 in the nul plane; but m cannot lie in the
nul plane (unless the system is coplanar, a case we need not consider),
therefore m passes through 0.

14*73. To find the equations of the conjugate of a line whose

equations are given.
288 FORCES IN THREE DIMENSIONS [XIV
It is only necessary to choose any two points on the given
line and write down the equations of their nul planes as in
14*63. Then since these nul planes intersect in the conjugate
of the given line their equations are the equations of the
conjugate line.
14*74. Examples, (i) Find the equations of the line conjugate to the
line x=y = zwith regard to the force system (X, Y,Z,L,M,N). Findalso
the magnitudes of the forces in the line and its conjugate.
The points (0, 0, 0), (1,1,1) both lie on the given line, and by 14*63 (1)
the equations of their nul planes are
Lx + My+Nz = 0 (1),
and

Subtracting the first from the second, we have

(Y-Z)x + (Z-X)y+(X-Y)z = L + M+N (2),
and this with thefirstequation may be taken as the required equations
of the conjugate line.
If P, P' denote the forces, the direction cosines of P are proportional
to 1, 1, 1 and of P' are A, fi, v, where
X:li:v=M(X-Y)-N(Z-X)
:N(Y-Z)-L(X-Y):L(Z-X)-M(Y-Z).
Then

so that (X-^P'^X-Y.
This reduces to

and then we find that

(L+M+N) P = (LX + M Y+NZ) V3.
(ii) A system of forces is reduced to two forces one of which acts at a
fixed point and lies in afixedplane through the point. Prove that the other
passes through anotherfixedpoint and lies in afixedplane.
Let P, P ' be the two forces of which P acts at a fixed point O and
lies in afixedplane A. Since O is a point on P, the conjugate P' lies in
the nul plane of O, i.e. afixedplane. Again let O' be the nul point of the
plane A; then by 14*72 since P lies in the nul plane of 0' its conjugate
P' passes through O', i.e. through afixedpoint.

14*75. The relation of conjugate lines to the central axis. Let

P, P' be a pair of conjugate forces and AA' the shortest dis-
tance between them. The central axis must intersect AA', for
otherwise the system would have a moment about AA'; and
14-73-14-81] FOUR FORCES 289
since P, P' are both at right angles to AA', therefore B must
also be at right angles to AA'. Let Oz be the central axis
cutting A A' in 0 and making
angles a, a' with P, P'. Let B, V
be the force and couple of the
wrench along Oz. Let AO = a,
OA' = a! and A A' = h.
Then B may be replaced by
forces Ba'/h at A and Bajh at
A', while Fisequivalentto forces
F/A at both A and A'. The forces at A must compound into P
and those at A' into P', so that
i?2a'2 + r 2 = P 2 ^ 2 and i?2a2 + r 2 = P'2A2.
Also tan a == YjRa' and tan a' = YjBa,
or ta,n<x.=pja' a n d tsLU.cn.' =p/a,
where p is the pitch of the resultant wrench.
14*8. Equilibrium of Four Forces.
If four non-intersecting forces are in equilibrium they act along
generators of the same system of a hyperboloid.
Three non-intersecting straight lines determine a hyper-
boloid on which the three lines are generators of the same
system. Consider the hyperboloid determined by the lines of
action of three of the given forces. The algebraical sum of the
moments about any generator of the opposite system is zero,
but the three chosen forces intersect all generators of the
opposite system, therefore the moment of the fourth force
about every generator of the opposite system is zero; there-
fore it intersects all these generators and itself acts along a
generator of the original system.
14-81. Forces acting along generators of the same system of a hyper-
boloid are in equilibrium if they would be in equilibrium when acting at a
point in the same directions.
The equations of a generator of the hyperboloid

through the point (a cos 9, b sin 6, 0) are

x — a cos 6 _y — b sin 8 _ z
a sin 6 ~ — b cos d ~ c'
290 FORCES IN THREE DIMENSIONS [XIV
Let a force P act in this lines—say at the point (acosfl, 6 sin 8, 0).
Then if X, Y, Z, L, M, N are the corresponding components of force
and couple at the origin
X Y Z

be
and L = bsin8.Z = — X,
a
M=-acos8.Z=™Y,

N=acos8.Y-bain6.X= -~Z.
c
For any number of such forces acting along generators of the same
system the conditions of equilibrium are
2X=0, S F = 0 , XZ=0 and SL = 0, SJfcf = O, SJV = O,
and it follows that if the first three are satisfied so also are the second
three and hence the required result.
14'82. E x a m p l e s , (i) Shew that any system of forces acting along
generators of the same system of a hyperboloid of one sheet can be reduced
to two forces, one acting along a given generator of the same system and the
other along some other generator.
The system can be reduced to a force P along a given generator and
a conjugate force P'. If we now reverse P and P ' and combine them
with the given system we get a system in equilibrium. Therefore the
sum of the moments about any generator of the opposite system is zero.
But this generator of the opposite system intersects all the given forces
and the force — P, therefore it must also intersect — P'. Hence P'must
intersect all generators of the opposite system and it therefore acts
along a generator of the same system.
(ii) Prove that any system of forces acting on a rigid body can be reduced
to two, one of which acts in a given plane and the other is perpendicular to it.
Also that (a) the shortest distance between the lines of action of the forces
cuts the central axis; (b) if the point of intersection be fixed and the plane
revolves about a line through the point perpendicular to the central axis, the
lines of action of the forces generate a hyperbolic paraboloid; (c) the mini-
mum value of the shortest distance is twice the pitch of the resultant
wrench. [T.]
Let the axis of the equivalent wrench be taken as axis of z and its
intersection with the given plane as origin O. Choose axes of a; and y so
that the given plane is z — x tan 6. Then the resultant force R and couple
F can be resolved into R sin 8, V sin 8, R cos 8, V cos 8, in directions OA,
OB whereof the former is in the given plane and the latter normal to it.
The couple Y sin 8 and the force R cos 8 compound into a single force
Rcos 8 parallel to OB and cutting Oy at E, where OE = b is such that
T sin 8 = bR cos0.
14-81, 14-82] EXAMPLES 291
Similarly the couple Fcos 8 and the force jRsin 8 compound into a
single force R sin 8 cutting yO produced at F, where OF = c is such that
F cos 8=cR sin 8.

Rcostf

The system is thus reduced to a force R sin 8 at F in the given plane

and a force J? cos 8 at E perpendicular to it, and the line EF is perpen-
dicular to both the forces and intersects the axis of the wrench.
Now suppose that the given plane revolves about Oy so that 8 is
variable.
The equations of the line of action of R cos 8 at E are
z— — a;cot0,
•p
and y =b=

so that the locus generated is yz = — p x, which is a hyperbolic para-

boloid. Similarly the line of action of R sin 8 at F lies on the same
surface.
Again, the shortest distance between the forces is
V
Jtt
= 2T/Rshx28,
the least value of which is 2T/R.

EXAMPLES
1. Two smooth planes intersect in a horizontal line and are inclined
at the same angle a to the vertical. A uniform rod of weight W and
length 2a is placed between them in a horizontal position making an
angle 8 with their line of intersection. Shew that the horizontal couple
required to maintain equilibrium is Wa cos 6 cot a. [I.]
2. Three smooth spheres of weight w and radius r rest symmetrically
within a sphere of radius Sr and another smooth sphere of weight Zw
292 FORCES IN THREE DIMENSIONS [XIV
is placed on the top of them. If the equilibrium is not disturbed, prove
that the radius of the upper sphere cannot be less than (T/2 — l)r. [S.]
3. Forces act along the sides BA, BC, DA, DC of a tetrahedron,
each force being inversely proportional to the side along which it acts;
prove that if the four forces give a single resultant
AD.BC=AB.GD. [S.]
4. Three rods OA, OB, OC, each of length I and of equal weight, are
smoothly jointed together at 0 and are placed symmetrically over a
smooth sphere of radius a, the joint 0 being vertically above the centre
of the sphere, and the rods resting against its surface. Prove that, if
\/2l=Sa, the rods, when in equilibrium, will be mutually at right
angles to one another. [S.]
5. Four rods, each of length 2a and weight W, are smoothly jointed
together and are placed over a smooth sphere of radius r (r > a) so as to
be in equilibrium in the form of a horizontal square in contact with the
sphere. Prove that the stress at each joint is equal to

g
a2)}*'
6. OABC is a tetrahedron formed by loosely jointing together six
rods. A point X on OA is connected by a string, whose tension is T, to
a point F o n BC. Shew that the action of the hinge at A on the rod
AC lies along AC and that its magnitude is
OXBYAC
OABCXY1' L
^J
7. A heavy sphere rests on three smooth pegs A, B, C in a horizontal
plane. Prove that the pressures on the pegs are proportional to sin 2A,
sin25, sin 20. [S.]
8. If A, B, C, D, A', B', C, D' be eight points in space so situated
that ABCD, A'B'C'D' are squares, then the condition of equilibrium of
forces represented in all respects by AA', B'B, CC, D'D, when acting
on a rigid body, is that the plane ABGD should be parallel to the plane
A'B'C'D'. [I.]
9. A weightless rod of length 21 rests in a given horizontal position
with its ends on the curved surfaces of two horizontal smooth circular
cylinders, each of radius a, which have their axes parallel and at a
distance 2c. The rod is acted on at its centre by a given force P and a
couple. Find the couple when there is equilibrium and prove that the
magnitude of the couple is least when P acts vertically, provided that

where <f> is the angle between the rod and the axes of the cylinders. [T.]
 EXAMPLES 293
10. A heavy uniform ellipsoid is placed on three smooth pegs in the
same horizontal plane, so that the pegs are at the ends of conjugate
radii. Prove that the ellipsoid is in equilibrium, and that the pressures
on the pegs are proportional to the areas of the conjugate central
sections. [C]
11. Wrenches of the same pitch p act along the edges of a regular
tetrahedron A BOD of side a. If the intensities of the wrenches along
AB, DC are the same, and also those along BO, DA, and DB, GA,
prove that the pitch of the equivalent wrench is p + a/2-y/2. [I.]
12. ABGD is a regular tetrahedron. Wrenches of equal intensities
act along BO, CA, AB, AD, BD, CD. Prove that the pitch of the
resultant wrench is$(p + q+r) + ia-\/2, where p, q, r are the pitches of
the wrenches which act along AD, BC; BD, A O; CD, AB and a is the
length of an edge. [T.]
13. Prove that if equal forces act along the edges BC, CA,AB, DA,
DB, DO of a regular tetrahedron, the central axis is the perpendicular
from D to the plane ABO and the pitch of the equivalent wrench is
H~7O a> where a is an edge of the tetrahedron.

14. Shew that the following system of forces reduces to a single

force:
Force 5\/2 acting through (0, 0, J) in direction (3:4: — 5),
Force 3-^/H acting through (£, 0, 0) in direction (5:5: — 7),
Force -y/H acting through (0, £, 0) in direction ( — 3:1:1),
the system being referred to rectangular axes. [I.]
15. A given system of forces is replaced by a force at a point P of a
given plane and a couple. A line PP' is drawn parallel to the axis of
the couple and proportional to its moment Shew that the locus of P'
is a plane. [I.]
16. Forces X, Y, Z act along the three lines given by the equations
y = 0, z = c; z = 0, x — a; x=0, y — b;
prove that pitch of the equivalent wrench is

If the wrench reduces to a single force, shew that the line of action
of the force must lie on the hyperboloid
(x-a)(y-b)(z-c)-xyz = Q. [T.]
17. A given force acts along the axis of x and another given force
along a generator of the cylinder xi + y* = a*; prove that the locus of
the central axis is an elliptic cylinder. [I.]
18. A given system of forces is equivalent to a wrench T of intensity
22 and pitch p: when an additional force B is added to the system the
pitch of the resulting wrench F' is pj2. Prove that this additional force
 294 FORCES IN THREE DIMENSIONS [XTV
intersects the axis of T and if the angle between them is 2<f>, different
from zero, then the axes of T and I " are a t a distance kp tan <f> and
inclined a t an angle <f>. [I.]
19. OA1AiA3 ..., O'' Ax'A2'' A3... are two straight lines in space and
forces X1.A1A1', X^.A^A/, X3.A3A3'... act along AJAJ", A%A2',
A3A3'..., respectively. Prove that they will be equivalent to a single
force if E\1.\t.A1As.A1'At' = 0; and to a single couple if 2A1 = 0,
SAj. OAX = 0, and SAj. O'AX' = 0. [I.]
20. A given system of forces is equivalent to a force and couple such
that the angle between the axis of the couple and the line of action of
the force is given. Find the locus of the line of action of the force. [I.]
21. Prove that a wrench is replaceable by forces on two lines of
which one may be arbitrary and shew that if one line is at a given dis-
tance from the axis of the wrench the other is at a given inclination to
the axis. [I.]
22. A system of forces can be reduced to a force B acting along a
certain line I, and a couple G acting in a plane perpendicular to I.
If the system is also reduced to two forces P and Q, such that the
shortest distances between their lines of action and I are a, b, respec-
tively, shew that
[I.]
23. Prove that if three wrenches are in equilibrium their pitches,
distances apart and inclinations are connected by the relations
Pi ~ «23 C 0 t #23 =Pi~ Z
31 C 0 t #31 = Pa~ Z
12 O o t #12 • E1-]
24. Two wrenches of pitches p, q, whose axes are at a distance 2a
from each other, have a resultant wrench of pitch m, whose axis inter-
sects the shortest distance between the axes of the given wrenches at a
distance £ from its middle point. Prove that the angle between the
axes of the given wrenches is equal to
t a n _ ! £(p-q)-a(2w-p-q)
2 - LA-J

25. Two wrenches have pitches p1, p2 and their axes intersect
perpendicularly: the resultant wrench has pitch p and an axis which
makes an angle 8 with the axis of the first wrench and is distant h from
that axis. Shew that
P—Picos2 8+Pn s i n 2 ^'
h = (p1~pi) sine cosfi. [T.]
26. The axes of two wrenches are along the axes of a; and y, which are
at right angles. The force and couple constituting the first are B1 and
Ox, and the second i?2 and O2. Shew that the resultant wrench consists
of a force (B^ + B^i and a couple {B1G1 + BiG2)j(R1!i + R^i, and
find its axis. [T.]
 EXAMPLES 295
27. The axes of two wrenches are at right angles and the shortest
distance between them is 2a. Prove that the axis of the resultant
wrench divides the shortest distance in the ratio
Q{2aQ + (p-q)P}:P{2aP-(p-q)Q},
where P and Q are the respective intensities of the wrenches, andp and
q are the pitches. [C]
28. Shew that two non-intersecting wrenches of pitches p, p' at
right angles to one another can be replaced by two intersecting wrenches
of pitches p, p' whose axes lie in a plane bisecting the shortest distance
between the first two and also bisect the angles between them, and
such that
2c" + 2c (p -p') = (p - P) (p' - P') + (p - P') (p' - P),
where 2c is the shortest distance between the first pair of wrenches. [I.]
29. Three forces act along given lines, no two of which meet or are
parallel. Prove that if and only if the given lines are parallel to a plane,
the magnitudes of the forces can be so chosen that the system is
equivalent to a couple; and prove that the axis of the couple is in a fixed
direction. [I.]
30. Prove that a wrench of intensity B and pitch arc may be replaced
by two forces inclined at an angle 28 to each other, the shortest distance
between them being 2c and their magnitudes given by
\R {V(l + & tan 0)+- v / (l-nr cot 6)}. [T.]
31. Shew that if two conjugate lines meet any plane in A, A', then
A A' passes through the nul point of the plane: also shew that if the
lines are the common conjugate lines of two force systems their inter-
sections with any plane and the nul points of the plane with regard to
the systems form a range whose cr^oss ratio is the same for all planes.
[T.]
32. Prove that the co-ordinates of the nul point of the plane
lx + my+nz= 1, with respect to the system (X, Y, Z, L, M, N), are
given by the equations
x _ V _ z _ 1 „,

33. A system of forces is such that every tangent to the curve given
by x = a cos 9, y = a sin 6, z = bd is a nul line. Find the central axis and
prove that the pitch of the equivalent wrench is a2jb. [I.]
34. A sphere is described with its centre on the central axis of a
system of forces. Shew that the nul points of tangent planes to the
sphere lie on a hyperboloid of revolution. [I.]
296 EORCES IN THREE DIMENSIONS [XIV
35. Two equal forces act along generators of the same system of the
x* + v2 z2
hyperboloid ^ — r^ = 1, and cut the plane z = 0 at the extremities of
perpendicular diameters of the circle xi + yi = a?: shew that the pitch
a2fc
of the equivalent wrench is „ . , . . [I.]

36. Shew how to find the generators of the hyperboloids

which are nul lines with respect to a given system of forces (X, Y, Z,
L, M, N), and shew that these generators are parallel to one or other
of the planes
(^ Yy-~Zz\ [T.]

37. Forces P, Q, R act along any three mutually perpendicular

generators of the same system of the surface x2 + y2 — 2 (z2 + a2), the
positive direction of the forces being towards the same side of the
plane xy. Prove that the pitch of the equivalent wrench is
[T.]
38. Any number of wrenches of the same pitchp act along generators
of the same system of the hyperboloid x2ja2 + y2jb'i — z2/c2= 1. Prove
that they will be reducible to a single resultant provided their central
axis is paraEel to a generator of the cone

ANSWERS
20. A circular cylinder. 33. Axis of z.

CAMBRIDGE: PBINTBD BY W. LEWIS, M.A., AT THE UNIVERSITY PRESS