Undergraduate Lecture Notes in Physics

Editorial Board
Neil Ashby
University of Colorado, Boulder, Colorado, USA

William Brantley
Department of Physics, Furman University, Greenville, South Carolina, USA

Matthew Deady
Physics Program, Bard College, Annandale-on-Hudson, New York, USA

Michael Fowler
Dept of Physics, Univ of Virginia, Charlottesville, Virginia, USA

Morten Hjorth-Jensen
Dept. of Physics, University of Oslo, Oslo, Norway

Michael Inglis
Earth &Space Sci, Smithtown Sci Bld, SUNY Suffolk County Community
College, Long Island, New York, USA

Heinz Klose
Humboldt University, Oldenburg, Niedersachsen, Germany

Helmy Sherif
Department of Physics, University of Alberta, Edmonton, Alberta, Canada

Undergraduate Lecture Notes in Physics (ULNP) publishes authoritative texts

covering topics throughout pure and applied physics. Each title in the series is
suitable as a basis for undergraduate instruction, typically containing practice
problems, worked examples, chapter summaries, and suggestions for further
reading.
ULNP titles must provide at least one of the following:
 An exceptionally clear and concise treatment of a standard undergraduate
subject.
A solid undergraduate-level introduction to a graduate, advanced, or non-
standard subject.
A novel perspective or an unusual approach to teaching a subject.
ULNP especially encourages new, original, and idiosyncratic approaches to
physics teaching at the undergraduate level.
The purpose of ULNP is to provide intriguing, absorbing books that will
continue to be the reader's preferred reference throughout their academic career.
Series editors
Neil Ashby
Professor, University of Colorado, Boulder, CO, USA
William Brantley
Professor, Department of Physics, Furman University, Greenville, SC, USA
Matthew Deady
Professor, Bard College Physics Program, Annandale-on-Hudson, NY, USA
Michael Fowler
Professor, Department of Physics, University of Virginia, Charlottesville,
VA, USA
Morten Hjorth-Jensen
Professor, University of Oslo, Oslo, Norway
Michael Inglis
Professor, SUNY Suffolk County Community College, Long Island, NY,
USA
Heinz Klose
Professor Emeritus, Humboldt University Berlin, Berlin, Germany
Helmy Sherif
Professor, Department of Physics, University of Alberta, Edmonton, Alberta,
Canada
More information about this series at http://www.springer.com/series/8917
Alessandro Bettini

A Course in Classical Physics 1—

Mechanics
1st ed. 2016
Alessandro Bettini
Dipartimento di Fisica e Astronomia, Università di Padova, Padova, Italy

ISSN 2192-4791 e-ISSN 2192-4805

ISBN 978-3-319-29256-4 e-ISBN 978-3-319-29257-1

DOI 10.1007/978-3-319-29257-1

Library of Congress Control Number: 2016934941

© Springer International Publishing Switzerland 2016

The text is primarily based on the book "Meccanica e termodinamica", A.

Bettini, Zanichelli, 1995. The author own all rights in the former publications.

This work is subject to copyright. All rights are reserved by the Publisher,
whether the whole or part of the material is concerned, specifically the rights of
translation, reprinting, reuse of illustrations, recitation, broadcasting,
reproduction on microfilms or in any other physical way, and transmission or
information storage and retrieval, electronic adaptation, computer software, or
by similar or dissimilar methodology now known or hereafter developed.

The use of general descriptive names, registered names, trademarks, service

marks, etc. in this publication does not imply, even in the absence of a specific
statement, that such names are exempt from the relevant protective laws and
regulations and therefore free for general use.

The publisher, the authors and the editors are safe to assume that the advice and
information in this book are believed to be true and accurate at the date of
publication. Neither the publisher nor the authors or the editors give a warranty,
express or implied, with respect to the material contained herein or for any errors
or omissions that may have been made.

Printed on acid-free paper

This Springer imprint is published by Springer Nature The registered company is

Springer International Publishing AG Switzerland
Preface
This is the first in a series of four volumes, all written at an elementary calculus
level. The complete course covers the most important areas of classical physics
such as mechanics, thermodynamics, statistical mechanics, electromagnetism,
waves, and optics. The volumes result from a translation, an in depth revision
and update of the Italian version published by Decibel-Zanichelli. This first
volume deals with classical mechanics, including an introduction to relativity.
The laws of Physics, and more in general of Nature, are written in the
language of mathematics. The reader is assumed to know already the basic
concepts of calculus: functions, limits, and the differentiation and integration
operations. We shall however, without mathematical rigor, give the necessary
information on vectors and matrices.
Physics is an experimental science, meaning that it is based on the
experimental method, which was developed by Galileo Galilei in the seventeenth
century. He taught us, in particular, that to try to understand a phenomenon one
must simplify as much as possible the relevant working conditions,
understanding which of the aspects are secondary and eliminating them as far as
possible. The understanding process is not immediate, but rather it proceeds by
trial and error, in a series of experiments, which might lead, with a bit of fortune
and a lot of thinking, to discover the governing laws. Induction of the physics
laws process goes back from the observed effects to their causes, and, as such,
cannot be purely logic. Once a physical law is found, it is necessary to consider
all its possible consequences. This is now a deductive process, which is logical
and similar to the mathematical one. Each of the consequences, the predictions,
of the law must then be experimentally verified. If only one prediction is found
to be false by the experiment, even if thousands of them had been found true, it
is enough to prove that the law is false. This implies that we can never be
completely sure that a law is true; indeed the number of its possible predictions
does not have limits, and in any historical moment not all of them have been
controlled. However, this is the price we must pay in choosing the experimental
method, which has allowed humankind to advance in the past four centuries
much more than in all the preceding millennia.
Classical Mechanics is one of the big intellectual constructions of Physics.
Its laws are well established as well as their limits of validity. Consequently, it
can be exposed in an axiomatic way, as a chapter of mathematics. We can start
from a set of propositions whose axioms are assumed to be true by definition,
and deduce from them a number of theorems using only logics, as from the
Euclid postulates the Euclidean geometry theorems are deduced.
We shall not follow this path. The reason is that, while it allows a shorter and
quicker treatment and is also logically more satisfactory for somebody, it also
hides the inductive trial-and-error historical process through which the postulates
and the general laws have been discovered. These are arrival, rather than starting
points. This path has been complex, laborious, and highly nonlinear. Errors have
been made, hypotheses have been advanced that turned out to be false, but
finally the laws were discovered. The knowledge of at least a few of the most
important aspects of this process is indispensable to develop the mental
capabilities that are necessary to anybody contributing to the progress of natural
sciences, whether they pursue applications or teach them. This is one of the
reasons for which we shall read and discuss several pages of the two scientists
that built the foundations of physics, Galileo Galilei and Isaac Newton. A second
reason is that reading the geniuses is always an enlightening experience.
The Galilei and Newton mechanics that we shall discuss in this book is a
coherent set of laws able to describe a great number of physical phenomena.
These laws, however, have a limited validity. One type of limitations does not
have a fundamental nature. Some of the laws, as for example the laws of friction
or the elastic force are, consciously we can say, approximate. In other words,
they provide a description that we know to be valid only in a first approximation
and provided that the values of certain quantities are within some definite
intervals (for example, for the elastic force, for not too large strains). We shall
always clearly state those limits.
The limits of the second type are of a fundamental nature. A first limit of the
Galilei-Newton laws is met when the velocities are very high, high enough to get
close to the speed of light.
The latter is so high, 300,000 km/s, that the speeds of all objects of a
common experience, planets included, are extremely small in comparison.
However, we can reach velocities close to that of light in experiments with
microscopic particles, like atomic nuclei and electrons. In the universe there are
double stars and double black-holes, which are extremely dense and rotate about
each other at very high speeds, close to the speed of light. We observe that in
these conditions the predictions of Newtonian mechanics are in contradiction
with experience. Newtonian mechanics is an approximation valid at velocities
substantially smaller than the speed of light. The theory that generalizes
Newtonian mechanics, including high-speed phenomena, is relativistic
mechanics, which was developed between the end of the nineteenth and the
beginning of the twentieth centuries by, principally, Hendrik Lorentz, Henry
Poincaré, and Albert Einstein. We discuss the basic elements of relativistic
mechanics in Chap. 6 . They are not necessary for understanding of the
following ones.
Newton or relativistic mechanics, depending on the velocities of the
problem, is called classical mechanics. However, not even this is true in every
circumstance; the laws of classical mechanics do not describe correctly the very
small-scale phenomena, like vibrations and rotations of molecules, those of the
electrons inside atoms, the nuclear, and subnuclear phenomena. As a matter of
fact, the bodies at these microscopic scales behave in a completely different way
than those of everyday experience. The theory able to describe all the known
phenomena both at small and large scale is quantum physics. Its limit for large
scales is classical mechanics. The study of quantum physics not only requires
mathematical instruments more advanced than classical physics, but, even more
importantly, cannot be profitably studied without an in-depth knowledge of
classical physics. Consequently, this course is limited to classical physics. We
shall however warn the reader of the limit of validity, whenever necessary.
In this book we deal with the mechanics of a material point and of extended
bodies, in particular of the rigid ones. The mechanics of fluids will be one of the
objects of study in the second volume, together with their thermal properties.
Mechanical oscillations are treated here only in their most elementary aspects. A
deeper discussion will be given, together with electric oscillations in the fourth
volume.
We start the first chapter with introductory elements: the measurement of
physical quantities, the measurement units and their internationally adopted
system, the International System, reference frames, and basic concepts on
vectors and matrices. The second part of the first chapter deals with kinematics,
which is the mathematical description of motion, without reference to its causes.
The second chapter contains the fundamental laws of the material point (the
simplest body) and the basic concepts of mass (both the inertial and the
gravitational masses), of force, of momentum, of moment of a force, and of
angular momentum. We introduce also the concepts of work of a force, of
energy, of power, and the energy conservation principle. We work on these
arguments mainly considering the two most usual examples of force, weight and
friction. At this point we have acquired the basic laws of mechanics.
Historically, these are the result of the work of G. Galilei and I. Newton. It is
important to have some knowledge of how these great authors came to establish
the laws of mechanics. For this purpose a few of their fundamental pages,
describing experiments and mathematical arguments, are reproduced and
discussed. The reader will see also how both authors expose the concepts in a
scientific superb language.
 The third chapter describes the different forces, gives their mathematical
expressions, and discusses their limits of validity. We discuss important
examples of motion, in particular the circular and the oscillatory ones. We know
now that the different forces that we see in nature and that look at first sight very
different can be reduced to a very limited number of fundamental forces. The
forces present at macroscopic level, the level of classical mechanics, are
different manifestations of two basic ones: the gravitational and the
electromagnetic forces. The latter will be studied in the third volume of this
course, the former in the fourth chapter of this first book. As a matter of fact the
study of gravitation has enormous historic and cultural importance. It underlines
our comprehension of the universe in which we live. For this reason we recall
the most important steps in the historical development of the universal
gravitation theory.
The description of any motion depends on the frame to which it is referred.
In particular it is different in two frames moving one relative to the other. The
study of this issue is the object of the fifth chapter, in the limit of velocities
much smaller than that of light. We shall meet with the extremely important
principle of relativity, a universally valid principle established already by
Galilei. The relations between reference frames at speed comparable to that of
light and the critical analysis of the concepts of time and space intervals leading
to the relativistic mechanics are dealt with in Chap. 6 .
In the last two chapters we study the mechanics of extended bodies. We start
Chap. 7 with systems made of only two different material points. We show that
in any case in which a force acts on a body, this is due to another body, which in
turn is acted upon by a force due to the first one. In other words the forces are
always due to the interaction between bodies. Having studied the issue on two-
body systems, we proceed in the second part of the chapter with the study of
material systems in full generality, finding the fundamental laws of their motion.
In the last chapter we study the principal aspects of the motion of particular, and
importantly, material systems, namely rigid bodies. Their motion is described by
well-defined differential equations. Their solution is an important mathematical
problem, which is however outside the scope of this course.
Each chapter of the book starts with a brief introduction to a scope that will
give to the reader a preliminary idea of the arguments he/she will find. There is
no need to fully understand these introductions, at the first reading, as all the
arguments are fully developed in the following pages.
At the end of each chapter the reader will find a number of queries on which
to check his/her level of understanding of the arguments of the chapter. The
difficulty of the queries is variable; some of them are very simple, some more
complex, a few are true numerical exercises. On the other hand, the book does
not contain a sequence of full exercises, considering the existence of very good
textbooks dedicated specifically to that.
The answers to a large majority of the queries are included. However, the
solution of numerical exercises (without looking at the answers) is mental
gymnastics that is absolutely necessary for understanding the subject. Only the
effort to apply what has been learned to specific cases allows us to master them
completely. The reader should be conscious of the fact that the solution of
numerical exercises requires mental mechanisms different from those engaged in
understanding a text. The latter, indeed, has been already organized by the
author; solving a problem requires much more active initiative from the student.
This is just the type of initiative, a creative activity that is needed, for advancing
scientific knowledge and its technical applications as well. Consequently, the
student should work on exercises alone, without looking at the solutions in the
book. Even failed attempts to autonomously reach the solution, provided they
are undertaken with sufficient persistence, give important returns, because they
develop processing skills. If after several failed attempts the solution has not yet
been reached, it is a better practice to momentarily abandon the exercise, rather
than looking at the solution, going to another one, and coming back later.
The following working scheme is methodologically advisable:

1. Examine at depth the conditions posed by the problem. If it is possible, make

a drawing containing the essential elements.

2. Solve the problem using letters, not numbers, in the formulas, then develop
them until the requested quantities are expressed in terms of the known ones.
Only then should you put numbers in the formulas.

3. Confirm the correctness of the physical dimensions (see Sect. 1.3).

4. When necessary transform all the data into the same system of units
(preferably SI, see Sect. 1.2 ). Use scientific notation, for example 2.5 × 10 3
rather than 2500, 2.5 × 10 −3 rather than 0.0025. In general two or three
significant figures are enough.
5. Once you have the final result, always verify if it is reasonable. For example
the mass of a molecule cannot turn out to be 30 mg, the speed of a bullet
cannot be 10 6 m/s, the distance between two towns cannot be 25 mm, etc.

Acknowledgments
The pages from Isaac Newton’s, Phylosophyae Naturalis Principia are from the
English translation from Latin by Andrew Motte (1729) modernized by the
author.
The pages from G. Galilei’s Dialogue concerning two chief world systems
are a translation into English by the author from the Edizione Nazionale delle
Opere, edited by Antonio Favaro; Florence, tip. Barbèra, 1890–1909.
The pages from G. Galilei’s Dialogues and mathematical demonstrations
concerning two new sciences are adapted from the English translation from
Italian and Latin by Henry Crew and Alfonso de Salvio; McMillan 1914.
Figure 4.18 is from the National Aeronautics and Space Administration at
http://www.compadre.org/Informal/images/features/Jupitmoons12-20-072.jpg
Figure 4.21 is from the European Space Agency at http://www.esa.int/var/
esa/storage/images/esa_multimedia/images/2007/05/globular_cluster_ngc_
28082/9535369-4-eng-GB/Globular_Cluster_NGC_2808.jpg
Figure 4.22 is from the National Aeronautics and Space Administration at
http://hubblesite.org/newscenter/archive/releases/2007/41/image/a//
Symbols and Units
Table 1 Symbols for the principal quantities

Acceleration a,as
Angular acceleration α,α
Angular frequency ω
Angular momentum l,L
Density (mass) ρ
Dynamic friction coefficient μd
Force F
Frequency ν
Gravitational field G
Gravitational mass mg
Gravity acceleration g
Impulse i
Inertia radius ρ
Inertial mass mi
Kinetic energy UK
Mass m,M
Moment of a force τ
Moment of inertia about a -axis Ia
Momentum p
Newton constant GN
Normal constraint reaction N
Period T
Plane angle θ
Polar angle θ,ϕ
Polar coordinates (space) ρ,θ,ϕ
Position vector r
Potential ϕ
Potential energy Up
Power w
Pressure p
Reduced mass μ
Spring constant k
Static friction coefficient μs
Time t
Tension T
Total angular momentum L tot
Total (mechanical) energy U tot
Total moment M
Total momentum P
Young module E
Weight Fw
Work W
Mean value, of x <x>
Angular velocity ω,Ω
Velocity of light (in vacuum) c
Velocity v,υ
Velocity divided by light velocity β
Unit vector of v uυ
Unit vectors of the axes i,j,k
Volume V

Table 2 Base units in the SI

Quantity Unit Symbol
Length metre/meter m
Mass kilogram kg
Time second s
Current intensity ampere A
Thermodynamic temperature kelvin K
Amount of substance mole mol
Luminous intensity candela cd

Table 3 Decimal multiples and submultiples of the units

Factor Prefix Symbol Factor Prefix Symbol
10 24 yotta Y 10 −1 deci d
10 21 zetta Z 10 −2 centi c
10 18 exa E 10 −3 milli m
10 15 peta P 10 −6 micro µ
10 12 tera T 10 −9 nano n
10 9 giga G 10 −12 pico p
mega M femto f
10 6 10 −15
10 3 kilo k 10 −18 atto a

10 2 hecto h 10 −21 zepto z

10 deka da 10 −24 yocto y

Table 4. Fundamental constants

Quantity Symb. Value Uncertainty
Speed of light in vacuum c 299 792 458 m s −1 Definition
Newton constant GN 6.67384(80) × 10 −11 m 3 kg −1 s _2 120 ppm
Astronomical unit a.u. 149 597 870 700 Definition
Avogadro number NA 6.022 1415(10) × 10 23 mole −1 170 ppb

Table 5. Solar planets orbits

Planet Mean distance from sun Sidereal period (tropical Angle with Eccentricity
(a.u.) year) ecliptic
Mercury 0.387099 0.24085 7°00'14'' 0.2056
Venus 0.723332 0.61521 3°23'39'' 0.0068
Earth 1 1.00004 0 0.0167
Mars 1.523691 1.88089 1°50'59'' 0.0934
Jupiter 5.202803 11.86223 1°18'19'' 0.0484
Uranus 19.181945 84.01308 0°46'23'' 0.0472
Neptune 30.057767 164.79405 1°46'26'' 0.0086
Pluto 39.51774 248.4302 17°08'38'' 0.2486

Tropical year = time interval between two consecutive passages of the sun at the
spring equinox
Table 6. Data on some bodies of the solar system
Body Mean radius Radius Mass Mean density (kg/m 3 )
(Mm) (Earth radiuses) (Earth masses)
Mercury 2.44 0.38 0.055 5430
Venus 6.05 0.95 0.815 5250
Earth 6.37 1 1 5520
Moon 1.74 0.27 0.012 3360
Mars 3.38 0.53 0.108 3930
Jupiter 71.49 11.19 317.9 1330
Saturn 60.27 9.46 95.18 710
Uranus 25.56 3.98 14.54 1240
Neptune 24.76 3.81 17.13 1670
Pluto 1.12 0.176 0.0026 1990
Sun 696 109.3 330,000 1400

Table 7. Greek alphabet

alpha α Α iota ι Ι rho ρ Ρ

beta β Β kappa κ Κ sigma σ, ς Σ
gamma γ Γ lambda λ Λ tau τ Τ
delta δ Δ mu μ Μ upsilon υ Υ,
epsilon ε Ε nu ν Ν phi ϕ, φ Φ
zeta ζ Ζ xi ξ Ξ chi χ Χ
eta η Η omicron ο Ο psi ψ Ψ
theta θ, ϑ Θ pi π Π omega ω Ω
Contents
1 Space, Time and Motion

1.1 Measurement of Physical Quantities

1.2 The International System (SI)

1.3 Space and Time

1.4 Vectors

1.5 Operations with Vectors

1.6 Scalar Product of Two Vectors

1.7 Vector Product of Two Vectors

1.8 Bound Vectors, Moment, Couple

1.9 Matrices

1.10 Velocity

1.11 Angular Velocity

1.12 Acceleration

1.13 Time Derivative of a Vector

1.14 Motion on the Plane

1.15 From Acceleration to Motion

1.16 Free Fall Motion

1.17 Scalars, Pseudoscalars, Vectors and Pseudovectors

Problems
2 Dynamics of a Material Point

2.1 Force, Operational Definition

2.2 Force Is a Vector

2.3 The Law of Inertia

2.4 The Newton Laws of Motion

2.5 Weight

2.6 Examples

2.7 Curvilinear Motion

2.8 Angular Momentum and Moment of a Force

2.9 The Simple Pendulum

2.10 The Work of a Force. The Kinetic Energy Theorem

2.11 Calculating Work

2.12 An Experiment of Galilei on Energy Conservation

2.13 Conservative Forces

2.14 Energy Conservation

2.15 A Theorem Concerning Central Forces

2.16 Power

Problems

3 The Forces

3.1 Elastic Force

3.2 Harmonic Motion

 3.3 Intermolecular Forces

3.4 Contact Forces. Constraint Forces

3.5 Friction

3.6 Viscous Drag

3.7 Air Drag and Independence of Motions

3.8 Damped Oscillator

3.9 Forced Oscillator. Resonance

3.10 Energy Diagrams in One Dimension

3.11 Energy Diagrams for Relevant Forces

Problems

4 Gravitation

4.1 The Orbits of the Planets

4.2 The Periods of the Planets and the Radii of Their Orbits

4.3 The Kepler Laws

4.4 The Newton Law

4.5 The Moon and the Apple

4.6 The Gravitational Force of the Homogeneous Sphere

4.7 Measuring the Newton Constant

4.8 The Gravitational Field

4.9 Galilei and the Jovian System

4.10 Galaxies, Clusters and Something Else

 4.11 Elliptic Orbits

4.12 The Newton Solution

4.13 The Constants of Motion

Problems

5 Relative Motions

5.1 Covariance of the Physical Laws Under Rotations and Translations

5.2 Uniform Relative Translation. Relativity Principle

5.3 Non-uniform Translation. Pseudo Forces

5.4 Rotation and Translation. Pseudo Forces

5.5 Motion in a Rotating Frame

5.6 The Inertial Frame

5.7 Earth, as a Non-inertial Frame

5.8 The Eötvös Experiment

Problems

6 Relativity

6.1 Does an Absolute Reference Frame Exist?

6.2 The Michelson and Morley Experiment

6.3 The Lorentz Transformations

6.4 Criticism of Simultaneity

6.5 Dilation of Time Intervals

6.6 Contraction of Distances

 6.7 Addition of Velocities

6.8 Space-Time

6.9 Momentum, Energy and Mass

6.10 Mass, Momentum and Energy for a System of Particles

6.11 Force and acceleration

6.12 Lorentz Covariance of the Physics Laws

6.13 What Is Equal and What Is Different

Problems

7 Extended Systems

7.1 Interaction Energy

7.2 Centre of Mass and Reduced Mass

7.3 Double Stars

7.4 Tides

7.5 Impulse and Momentum

7.6 The Action-Reaction Law

7.7 Action, Reaction and Linear Momentum Conservation

7.8 Systems of Particles

7.9 The Center of Mass

7.10 Linear Momentum Conservation

7.11 Continuous Systems

7.12 Angular Momentum

 7.13 Angular Momentum Conservation

7.14 Energy of a Mechanical System

7.15 Center of Mass Reference Frame

7.16 The König Theorems

7.17 Elastic Collisions

7.18 Inelastic Collisions

Problems

8 Rigid Bodies

8.1 Rigid Bodies and Their Movements

8.2 Applied Forces

8.3 Equilibrium of the Rigid Bodies

8.4 Rotation About a Fixed Axis

8.5 Conservation Angular Momentum About an Axis

8.6 Work and Kinetic Energy

8.7 Calculating Inertia Moments

8.8 Theorems on the Moments of Inertia

8.9 Torsion Balance

8.10 Composite Pendulum

8.11 Dumbbell

8.12 Angular Momentum About a Fixed Pole

8.13 Kinetic Energy

 8.14 Rotation About a Fixed Axis. Forces on the Supports

8.15 Rolling Motion

8.16 Rolling on an Inclined Plane

8.17 Gyroscopes

8.18 Collisions Between Material Systems

8.19 The Virtual Works Principle

Problems

Solutions

Index
© Springer International Publishing Switzerland 2016
Alessandro Bettini, A Course in Classical Physics 1—Mechanics, Undergraduate Lecture Notes in Physics,
DOI 10.1007/978-3-319-29257-1_1

1. Space, Time and Motion

Alessandro Bettini1
(1) Dipartimento di Fisica e Astronomia, Università di Padova, Padova, Italy

Alessandro Bettini
Email: alessandro.bettini@pd.infn.it

Physics is an experimental science that gives a quantitative, mathematical

description of natural phenomena. This means that physical laws are
mathematical relations amongst physical quantities (such as position, velocity,
force, energy, etc.). These relations are to be considered true only if they
correspond to experience. Physical laws must always be experimentally verified.
Experiment is the sole judge of scientific truth. Consequently, any physical
quantity must be measurable, namely the set of operations that must be
performed to measure it must be defined. First, a system of units of measurement
must be defined. We shall see in the first three sections how this is done. The
choice of units is a priori arbitrary; the physical laws depend on Nature, not on
our choices, In practice, however, having standardized choices is extremely
important to make the results understandable to everybody. International
agreements have defined the system of units to be named, in French, Système
International (International System).
Some of the physical quantities, like mass and temperature, are represented
by a single number and are called scalar. Other, like velocity and force, are more
complex; specifying how big they are is not sufficient, also their direction must
be given. Mathematically, an ordered set of real numbers represents them; they
are vector quantities. We shall study in the sections from Sects. 1.5 to 1.8 the
elementary mathematical properties of vectors and of the operations (sum,
difference, products) amongst them. In Sect. 1.9 we shall introduce some
elements that will be useful in the following on another mathematical object,
matrices.
In the second part of the chapter we shall move to physics, dealing with the
kinematics of the point-like particle, namely the study of its motion,
independently of its causes. We shall introduce the concepts of velocity, angular
velocity and acceleration. These are vector quantities, in general depending on
time. Section 1.13 is again of mathematical type, presenting a formula that will
be very useful in the following, the time derivative of a vector.
A few types of motion are particularly important: circular motion, studied in
Sects. 1.11 and 1.12, motion on a plane in Sect. 1.14 and free fall of weights in
Sect. 1.16.

1.1 Measurement of Physical Quantities

Physics gives a quantitative description of natural phenomena (or, better, of the
known part of them). Measurement of the relevant physical quantities leads to
discovery of the physical laws, which are mathematical relations amongst those
quantities (for example, the law of the free fall, the Kepler laws, etc.).
All natural phenomena take place in space and have a temporal duration;
some of them happen before, others afterwards. Consequently, space and time
are fundamental concepts. Physical objects are characterized by quantities like
length, area, volume, color, hardness, mass, temperature, etc. All these concepts
result from a common experience and are present in common language.
However, Physics must give to each quantity a rigorous definition, in order to be
able to give it numerical values. In this definition process, the concept may
become rather different from a common language.
Consider for example the length of an object or the distance between two
places. If we want to designate a number we must first define a unit of length.
Indeed we say: “That bar is 5 m long”, or, if we are in England: “That city is
20 miles away”. The measure of the length of an object is the ratio between its
length and the length of another object we have chosen as unit. “A bar is 5 m
long,” means that its length is equal to that of 5 one-meter long rules in a line.
“The mass of a body is 8 kg,” means that it is equal to that of eight bodies of
1 kg together.
The measurement of physical quantity is the ratio between that quantity and
its measurement unit.
The measurement operation allows associating to each physical quantity a
number. The symbols that appear in the physical laws representing the various
physical quantities are just these numbers. For example, when we write F = ma
we mean that the ratio between the force we are considering and the force taken
as unit, is equal to the ratio of the mass of the object and the mass of the object
taken as unit, times the ratio between the designated acceleration and the unit
acceleration.
Every physical quantity must be measurable and its definition must be
precise and rigorous. The operational definition is the most effective way to
define a physical quantity. This is defined as the set of operations needed to
measure that quantity.
This procedure has two important implications. The first implication is that
quantities that are not, even in principle, measurable are not physical quantities.
This does not imply that such quantities cannot be used. Indeed they are often
useful in the mathematical developments of a theory. Any theory, however, if it
has to be a physical theory, must lead to predictions that are experimentally
testable. The experimentally testable predictions are mathematical relations
amongst physical quantities, meaning measurable ones. The auxiliary, non-
measurable, quantities, should not appear in the final theoretical expression.
The second implication is that scale matters: quantities may be small or
large. Consider for example the length. If we want to measure distances, say,
from millimeters to kilometers we can use graduated bars or rules (like
yardsticks, measuring tapes, calipers, gauges, etc.). If we need to measure
distances of tens or hundreds of kilometers, as for example between two
mountaintops or the height of Mount Everest, the procedure is very different and
we must perform triangulations. If the distances are very much larger, as those of
the galaxies, the procedures change completely again. And different procedures
are required to measure small distances such as the diameter of an atom or of an
atomic nucleus. In every range of orders of magnitude, the set of procedures to
measure a length is different. To be rigorous we would need to talk of many
different lengths. This would lead to a terrible confusion. Fortunately, we
experimentally verify that, in the large intervals in which two or more methods
work contemporarily, the results are equal, and we can define a single length
concept. However, the above arguments tell us to be careful. Suppose that a
physical law is well experimentally verified for objects of sizes between meters
and kilometers. We tend to think the same law to be valid also for objects much
smaller and much larger than that. But we have no guarantee that the
extrapolation is true. On that, as always, only experiment can judge. For example
the Newton laws valid at the speeds of ordinary experience are no longer valid at
speeds comparable to the speed of light. The laws of classical mechanics are
simply not valid for atoms and smaller objects.
Let us go back to the measuring operation. For each quantity we need a
measurement unit. The choice is in principle arbitrary but is far from being so in
practice. If every Country, for example, would choose a different unit for lengths
or areas, the exchanges, not only the scientific ones, but also the commercial
ones, would be extremely complex. The units must be standardized. The issue is
so important that both units and procedures are made compulsory by law in the
majority of Countries.

1.2 The International System (SI)

The modern international standardization of units started with the French
Revolution. In 1791 the Decimal Metric System was officially announced, but it
took almost a century for its substantial diffusion and acceptance (and, most
important, Napoleon to impose it; in Britain, where his reach was insufficient,
the Imperial System is still used, as it is also in the USA). In May 1875, at the
“Metre convention”, the representatives of 17 Nations signed an international
treaty in Paris. National and international laboratories were created with the
mission to develop measurement standards and procedures. This is a very
important sector of physics, know as metrology .
International Organizations were created to foster international
standardization of weights and measures in the world. The International
Conference of Weights and Measures, CGPM for brief using the initials in
French, which meets every several years, is the main decision-making body. It
decides on the evolution of the internationally adopted system of units, named in
French Sistème International or SI for brief. In 1971 the European Community
issued a directive to the member states for the legal adoption of the SI.
There are two classes of units: base units and derived units . The base units
are given by definition. Each derived unit is obtained using a physics law,
namely a mathematical expression that links it to quantities of the basic units.
The choice of the basic units, and even their number, is, from a logic point of
view, arbitrary. The choices are based on convenience, taking quantities for
which measurements can be as much as possible precise and reproducible.
Let us consider an example. Take the “physical laws”: (1) the area S of a
rectangle of sides of lengths a and b is proportional to the product of the lengths,
(2) the area A of a circle is proportional to the square of the length R of its
radius, (3) the space s covered by a body moving in absence of any force is
proportional to the time t employed and to its velocity υ. The mathematical
expression for these laws would be
 (1.1)
where k, k′ and k″ are purely numerical constants. They depend on the choice
of measurement units. We might take both length and area as base quantities and
as units the meter and square foot respectively. The k and k′ would then have
definite values. Our measuring system would be simpler taking length as the
base unit, say 1 m, and the area as derived. Still, however, some arbitrariness
remains. For example, we can choose the units in such a way to have k = 1 or,
differently, to have k′ = 1. In the first option the unitary area is the square of
1 m side, in the second one it is the circle of 1 m radius. The second choice gives
k = 1/π, and appears funny. The choice k = 1 appears to be the obvious one, and
is the universally used one, but, in principle, it is not necessary.
Similarly, in the third equation we make k″ = 1 by choosing as measuring
unit of velocity the velocity of a body that covers the unit length in the unit of
time.
As already mentioned, the internationally accepted system of units is the SI.
It is the easiest to use and the most rational one. In the SI the base units are
seven: length, mass, time, electric current intensity, thermodynamic temperature,
amount of substance, luminous intensity. For each of them, the name of the unit
(e.g. “meter”) and its symbol (e.g. “m”) are fixed, as in Table 1.1. Notice that the
initial of the name of a unit is always lower case, including when it is the name
of a scientist (e.g. “ampere”). Most important, the SI gives a precise and clear
definition for each unit. Notice that these may change with time, as a
consequence of the progress of metrology, after formal approval by the CGPM.
We shall give here the definitions of the first three units, which are the only ones
needed in this textbook. The other ones will be defined in the other volumes of
the series, when needed.
Table 1.1 The base quantities, their units and symbols
Quantity Unit Symbol
Length metre/meter m
Mass kilogram kg
Time second s
Current intensity ampere A
Thermodynamic temperature kelvin K
Amount of substance mole mol
Luminous intensity candela cd

The metre ( meter is the distance travelled by light in vacuum in a time

interval of 1/229 792 458 of a second.
 The kilogram is the mass of the international prototype kilogram (located
in the Pavillon de Breteuil at Sèvres).
The second is the duration of 9192631770 periods of the radiation
corresponding to the transition between the two hyperfine levels of the
ground state of the Cesium 133 atom.
The SI defines the names and symbols of all the derived units. We shall
introduce them when we meet them for the first time. The SI further defines
names and symbols of multiples and submultiples of the units . This is done in
steps in general of three orders of magnitudes, of one order for the first three, as
in Table 1.2. With the exception of da, h, and k, all multiple prefix symbols are
upper case; all submultiple prefix symbols are lower case letters.
Table 1.2 Decimal multiples and submultiples
Factor Prefix Symbol Factor Prefix Symbol
1024 yotta Y 10−1 deci d
1021 zetta Z 10−2 centi c

1018 exa E 10−3 milli m

1015 peta P 10−6 micro µ

1012 tera T 10−9 nano n

109 giga G 10−12 pico p

106 mega M 10−15 femto f

103 kilo k 10−18 atto a

102 hecto h 10−21 zepto z

10 deka da 10−24 yocto y

The derived measurement units are defined, as mentioned, using a physical

law in order to have a definition as simple as possible. Hence, the unit for areas
is the square of 1 m side, the unit of volume is the cube of 1 m side, the unit of
velocity is the velocity of a body travelling 1 m in one second, etc.
The (mean) acceleration is the change of velocity Δυ divided by the time
interval Δt in which that change happens, namely . The acceleration
unit is the acceleration of a body, the velocity of which varies by a unit (1 m/s or
1 m s−1) in the unit of time (1 s). It is consequently the meter per second per
second (m/s2 or m s−2).
Let us now observe, as an example, that all the plane figures, triangles,
rectangles, circles etc. are expressed as a numerical factor times the product of
two lengths. Namely, all areas have a physical dimension of length squared. If
we change the unit of length, for example from meter to centimeter, the
measures of all the areas change by the same factor: 1002 in the example. The
physical dimensions of velocity are length divided by time, of acceleration of
length divided by time squared, etc. The corresponding mathematical
expressions are called dimensional equations which are of the type

(1.2)
Dimensional equations are very useful in practice. Consider any relationship
amongst physical quantities, for example F = ma or A + B = C. All the terms
must have the same physical dimensions. Otherwise, a change of units will cause
the different terms to change in different ways; the validity of the relation would
depend on the choice of units, which is arbitrary. This is the so-called
homogeneity principle . It is very useful to check analytical expressions obtained
with more or less complex calculations. If we find that some of the terms have
different dimensions, we must conclude that we have made some mistake.
Notice that there are also physical quantities having nil dimensions, namely
[L 0 T 0 M 0], they are pure numbers. An important example is the angle. In
radians (rd) it is the ratio between the arc of a circumference and its radius. If we
change the unit of length, the ratio between two of them does not change.
Finally notice that a physical law may contain mathematical functions, for
example . These expressions make sense only if
both the functions themselves (x, y, z) and their arguments (α, β, γ) have no
physical dimensions. All of them must be pure numbers.

1.3 Space and Time

Our study begins with the study of the motion of bodies. Motion of a body
means that its position in space varies in time . The notion of motion is relative:
a passenger in a plane sitting in his chair has a fixed position relative to the
plane, but moves at, say 800 km/h relative to a person standing on earth. The
latter moves at 800 km/h relative to the passenger, in the opposite direction. To
describe the motion we then need a reference frame .
We normally live standing on earth and such are the laboratories in which we
do our experiments. Let us then start by choosing a reference frame fixed on the
earth. The possible choices are still infinite.
The position of a body is defined when we know were it is. The simplest
case is when we deal with a particle, a body that is so small that it can be
considered point-like. It is called a material point . Let us see how we can define
the position of a material point. For an extended body the positions of all its
points should be similarly defined.
To know the position of a point in space we need three numbers, one for
each of its dimensions. To define its position on a given surface, two numbers
are needed (as for example longitude and latitude on the earth surface). To know
the position on a given curve, one number is needed.
Let us start by considering a point P that can move only on a straight line
(see Fig. 1.1a). To define its position: (1) we choose one of the two directions
and call it positive, (2) we choose a point on the line and call it the origin of the
co-ordinates (O in Fig. 1.1a), (3) we choose a unit length. The oriented line, with
an origin and a measuring unit is called a co-ordinate axis . The position of the
generic point P is given by a real number, called the co-ordinate of the point (x
in the figure), which is the distance of P from O, taken as positive if P is on the
right of O, negative if it is on the left.

Fig. 1.1 Orthogonal co-ordinate frames. a One dimension, b two dimensions, c three dimensions

Let us now assume that point P can move on a plane (Fig. 1.1b). We now
need two co-ordinate axes, which should not be parallel. It is usually convenient
to take them perpendicular, the origin at the point in which they cross and the
same unit length for both (none of these choices is compulsory, they are just
generally the most convenient). The position of P is given by its two co-
ordinates, which is an ordered pair of real numbers (x, y).
Consider now a point in space. The reference frame shown in Fig. 1.1c is
called a Cartesian rectangular right-handed frame, after René Descartes (1596–
1650). It is made of three co-ordinate axes, called x, y and z. They cross in a
single point, the origin of the frame. All the angles between the (three) pairs of
axes are right. The length units on the three axes are equal. Finally we must
choose positive orientations of the axes. There are two basic possibilities. Let us
assume that we have already defined the positive directions of x and y. We have
two possible choices for the positive direction of z. Figure 1.1c shows one of
them; an observer standing with his feet on the xy plane lying along the z axis
and looking down, willing to move the x axis on the y axis by a 90° rotation, sees
this rotation happening anticlockwise. The second possibility is the opposite sign
of z. The two frames are called right-handed and left-handed respectively.
Now consider the inversion of the axes. If we start from a right-handed frame
and invert one axis, that is a mirror reflection and we get a left-ended frame. The
same happens if we invert all three axes. The inversion of two axes gives, on the
contrary, the same result as a rotation of 180° around the third axis: the initial
and final frame have the same “handness ”.
To define the reference frame we have made a series of choices, which we
recall:

(1) choice of the origin

(2) choice of the directions of the axes

(3) choice of the positive directions (left-handed or right-handed)

(4) choice of the units.

While each of these choices is arbitrary, we can ask whether there is any
privileged choice, or if there is one that is better posed, are the physics laws
independent of these choices? The answers cannot come from logics or
mathematics, but only from an experiment. Let us consider each of them.

(1) Are the physics laws independent of the origin of the axes? To check the
point, let us build two identical apparatuses. Let each of them contain
inclined planes with balls rolling on them, pendulums, flywheels, gears, etc.,
all identical. We position the two apparatuses in two different locations. We
prepare them to be in exactly the same initial state: the pendulums are out of
 equilibrium at the same distance, the spheres are at the same heights on the
inclined planes, the gears and the flywheels are in the same positions. We let
them go contemporarily and observe their evolutions. Do the two systems
evolve in the same way? Do they assume the same configurations at the
same times? As a matter of fact the answer is not always yes. However,
every time some difference is noticed, it is possible to identify the reason for
that in some physical condition that is different in the two locations. For
example, the gravitational acceleration might be a bit different in the two
sites and consequently the periods of the pendulums are a bit different too.
In any case, experiments show that, once all the local effects are eliminated,
or accounted for, the apparatuses evolve in the same manner, i.e. going
through the same configurations at the same instants.
The very important conclusion is: The physical laws are independent on
position. In other words all positions are equivalent, or space is
homogeneous . Let us repeat that this is an experimental conclusion. No
experiment up to now has found it wrong. One can state that the physical
laws are invariant, meaning that they do not vary, under space translations.

(2) Are the physics laws independent of directions of the axes? We now take
our two identical apparatuses and rotate one to the other. For example, in
one case the z-axis is vertical, in the other is at 45° with the vertical. Do the
two systems evolve through the same states? Certainly not! Indeed, for
example, pendulums oscillate around a vertical axis in one case, around an
inclined one in the other. In this case a privileged direction exists, the
direction of weight. But, think a moment. If we were far from earth, or in
absence of weight, the privileged direction would not exist. That direction is
not a property of the space, but is the “local” effect of a body, the earth. In
other words, if we want to compare the two experiments in the same
conditions, we should also rotate the earth in the second case. If all the
external conditions are properly taken into account, all the experiments show
that the physical laws are independent on the directions of the axes. In other
words, no privileged direction exists, or, space is isotropic . Still in other
words, physics laws are invariant under rotations.

(3) Are the physics laws, independent of the orientation, left-handed or right-
handed? Experiments have shown that all physics laws at the macroscopic
 level are independent of the choice. But this is no truer at a microscopic
level. A class of radioactive phenomena, like beta decays, is due to a
fundamental force called weak interaction. Its laws distinguish between the
left and right cases. Namely, not all the physics laws are invariant under
inversion of the axes.

(4) Are the physics laws independent of the scale of length? This time we build
two apparatuses that are identical but for having all their dimensions
different, scaled by the same factor. Do the two evolve in the same manner?
The answer was discovered by Galileo Galilei (Italy, 1564–1642) and is NO.

Consider for example a beam made of a certain homogeneous material. The

beam has a certain length, and its cross section, which we assume to be
rectangular, has a certain width and a certain height. We lay it on two supports
near to its extremes on a horizontal plane. Suppose the beam to be in
equilibrium. We now take a beam geometrically similar to the first one but ten
times longer, ten times wider, ten times higher. Again we lay it on two supports
near to its extremes. We observe that the beam breaks down in its middle point.
The reason is the following. The weight of the beam is a force applied in its
middle point directed downward. The weight tends to break the beam, the
cohesion forces between molecules tend to keep it together. The weight, which is
proportional to the volume, is for the second beam one thousand times larger
than for the first one. The resistance to fracture is proportional to the area of the
cross section and for the second beam is one hundred times larger than for the
first one. Consequently, above a certain dimension the beam breaks down under
its own weight. For the same reason the animals cannot be too big. The bones of
the legs of a hypothetical horse ten times bigger than the real ones would break
under their own weights. We know now that the fundamental reason for that is
that substances are made of molecules and atoms, which have a definite size.
Certainly we cannot build one of the above-considered apparatuses so small to
be made of a few molecules.
As another example, consider the heavenly bodies. Stars and the largest
planet emit light, the smaller planet, like earth, do not. Only if the size, hence the
mass of the body, is large enough, the pressure and temperature in its core,
which are due to the action of the gravitational forces between its parts, are large
enough to fire the thermonuclear fusion reactions that produce light.
 In conclusion: The physics laws are NOT invariant under changes of scale.
We now come back to mathematics of reference frames. In the following we
shall need to use another type of, equivalent, co-ordinates, the spherical polar
coordinates . Figure 1.2a shows such co-ordinates on the plane, Fig. 1.2b in
space, together with the orthogonal coordinates in both cases. On the plane, the
two polar co-ordinates of the generic point P are its distance from the origin ρ,
which is a non-negative number, called a radius , and the angle ϕ, between the x-
axis and the segment OP, called an azimuth . It is measured in anticlockwise
direction and varies between 0 and 2π, namely

Fig. 1.2 Polar co-ordinates. a Two dimensions, b three dimensions

(1.3)
We can easily see from the figure that the relations between polar and
rectangular co-ordinates are
(1.4)
and the inverse ones

(1.5)
Figure 1.2b shows polar co-ordinates in three dimensions. The first co-
ordinate of the generic point P is again its distance r from the origin (radius), the
second co-ordinate is the angle ϕ between the plane through the z and P and the
plane xz (azimuth), the third co-ordinate is the angle θ between the segment OP
and the z axis ( zenith angle). Again r is a non-negative number. The angle θ
varies from 0 to π, covering in such a way the semi-plane shown in the figure.
This semi-plane rotates around z when ϕ varies between 0 and 2π. Hence
(1.6)
The relations with the orthogonal co-ordinates are
 (1.7)
and the inverse ones

(1.8)
If the point P is on the xy plane, namely if θ = 0, Eq. (1.8) become

which are equal to Eq. (1.4).

To know the motion of a body we need to know its position in different time
instants. Consequently we most measure the time. More precisely, we measure
intervals of time, rather than an absolute time. In practice we choose a certain
instant and define it as the origin of times, for which t = 0. We next choose a
time interval and define it as the unit of time. In the SI it is the second. In
principle we should also choose one of the two directions as positive, but the
choice is obvious. It is, we can say, imposed by Nature: the positive direction of
time is from past to future. Consequently, the time of an event is negative if it
happened before t = 0, positive if after that.
We now ask: is the origin of times arbitrary? As always we must apply to the
experiment. Let us go back to one of our experimental apparatuses and let us
repeat the experiment starting from the same initial state, for example in the
morning, then in the afternoon, and again in the night, etc. For each trial we take
the origin of time as the initial instant. We observe that, once all the spurious
elements are taken care of (e.g. light in the day, dark in the night) all the
experiments evolve in the same way. The origin of times is arbitrary, time is
homogeneous. The physics laws are invariant under translations in time. In
addition, similarly to space, no fundamental time interval exists.
We have said that the choice of the positive direction of time is imposed by
Nature. Several books have been written on this issue, the “arrow of time”. We
shall not enter in this discussion. We only state here that in the purely
mechanical phenomena no arrow of time exists. Suppose we hit a billiard ball
and shoot a movie of its motion hitting other balls, the walls, etc. If we now play
the movie backwards we observe a perfectively legitimate evolution. We cannot
know if it is backwards of forwards. But, wait; this is not true forever. Indeed, if
the movie is long enough we see that, when plaid forwards, the speeds of the
balls gradually slow down and finally they stop. If it is plaid backwards, the balls
are initially steady and start moving alone. The natural arrow of time is the one
in which the kinetic energy diminishes. When we study thermodynamics in the
second volume of this course we shall see how it explains the arrow of time.
1.4 Vectors
Many physical quantities, such as temperature and atmospheric pressure, are
represented by a single number. This is not the case of other ones, such as
velocity, acceleration, force, etc. For example to know the velocity of a car is not
sufficient to know how fast it moves (that is the number we read on the
speedometer), but also in which direction (towards South, North or other).
Another example is a displacement in space. To know it we need to know how
long it is and in which direction it happens. These physical quantities are
represented by vectors .
A vector is a mathematical entity. To define it, let us start considering line
segments. A segment is called oriented , if one of its two senses is chosen as
positive. Two oriented segments are said to be equipollent if they have the same
length, the same direction and the same sense. A vector is the class of all the
oriented segments equipollent to a given one. It is graphically represented with
an arrow. It is characterized by the length, called magnitude , the direction and
the sense. Differently from the oriented segment, it is not characterized by its
position. The velocities of two cars moving at 100 km/h heading West, one near
Paris, one near London are the same.
Once a reference frame is chosen, we can represent a vector by an ordered
triple of real numbers, which are its components in that reference. However, an
ordered triple of real numbers is not necessarily a vector. To be so the following
important property must be satisfied. Indeed, if we change the reference frame,
for example rotating the axes, the components of the vector change, but the
vector does not. Vector is a definite object; its components are the way to see it
in one or another frame. To satisfy these properties, the vector components,
namely the ordered triples in the two frames, must be connected by well-defined
relations, which we shall now find.
Figure 1.3 shows a reference frame and a point P of co-ordinates x, y, z.
Consider the oriented segment from the origin O to P and the corresponding
vector r (namely the class of equipollent oriented segments). It is called a
position vector , and is, we can say, the prototype of all vectors. Its components
in the (Cartesian) reference frame we are considering are simply the coordinates
of P, i.e. the ordered triple (x, y, z). Let us now take another reference with the
same origin and axes rotated by an angle θ. The point P does not move and r
does not change. But its components (x′, y′, z′) are different. The general relation
between the two triplets is rather complex. For simplicity we shall consider two
frames with the same origin and the same z-axis, as shown in Fig. 1.4.
Fig. 1.3 The orthogonal co-ordinates and the position vector

Fig. 1.4 A rotation of a Cartesian reference frame around the common z axis

Let us consider the point P in the figure of co-ordinates (x, y) in one frame,
(x′, y′) in the other. We must express x′ and y′ as functions of x, y and θ. One
relation is obvious, z′ = z. In practice, we are reduced to two dimensions.
We now draw perpendiculars from P to all the axes. We also draw the
segment AB perpendicular to PQ. The figure shows that x′ is the sum of two
lengths along the x′ axis and y′ the difference of two lengths along AB. We
obtain

(1.9)

where, to be complete, we included also the third co-ordinate. Notice that

these relations are both the relations between the components of the position
vector in the two frames and the relations between the co-ordinates in the two
frames. As a matter if fact they analytically define the rotation of the axes.
 We now state that a vector is an ordered triple of real numbers that under
rotations of the reference frame transforms (changes) in the same way as the
triple representing the position vector, namely as co-ordinates.
Figure 1.5a represents, in a plane for simplicity, a generic vector A, which
we can think as of drawn starting from the origin, because all the equipollent
segments are the same vector, and its components in the two frames.

Fig. 1.5 Components of the vector A in two frames different for a a rotation b a translation

By definition, the relations amongst its components are equal to Eq. (1.9),
namely

(1.10)

The inverse relations, namely the expressions of (A′ x , A′ y , A′ z ) as

functions of (A x , A y , A z ) and θ can be obtained in two ways: inverting the
system (1.10) or, which is simpler, thinking that the first reference is obtained
from the second by a rotation of an angle –θ. Consequently we have

(1.11)

We have considered two frames differing for a rotation of the axes, with a
common origin. Consider now two frames differing for a translation, namely
with parallel axes and different origins, as shown in Fig. 1.5b, again for
simplicity in a plane. We see that the components of the vector A in the two
frames are equal.

1.5 Operations with Vectors

A quantity represented by a number, like temperature or pressure, is called a
scalar . Scalars are invariant under rotations of the axes. In two reference frames
rotated one to the other a scalar has the same value. Notice that not every
quantity is scalar. For example the x component of a vector is not, because it
changes under rotations.
We shall represent a vector with its components in a given frame with A = (A
x , A y , A z ).
Given the vector A and the scalar k their product is the vector k A = (kA x ,
kA y , kA z ). Namely the components of k A are k times those of A. To be sure,
we must verify that the just given definition agrees with the definition of vector.
Indeed, it is immediate to check that the oriented triple (kA x , kA y , kA z )
transforms like a vector.
Geometrically, k A is the vector with the same direction as A, the magnitude
|k| times the one of A and the sense of A or opposite depending on k being
positive or negative respectively.
The product of A times the reciprocal of its magnitude is a vector with the
direction of A and unitary magnitude. A vector of unitary magnitude is called a
unit vector or versor. We shall use the symbol u A for the unit vector of A.
The product of the vector A and –1 is called the opposite of A. It has the
same magnitude and direction of A and opposite sense.
Consider now two vectors A and B, which in a given reference frame have
the components (A x , A y , A z ) and (B x , B y , B z ) respectively. Consider the
triple of numbers that are the sums of the homologous components of A and B,
namely (A x + B x , A y + B y , A z + B z ). Is it a vector ? Let us check. Knowing
that A are B vectors we know that

By summing member to member we have

We see that the answer is positive. We can then define as the vector sum of
two vectors the vector with components equal to the sums of their homologous
components. Notice that the just found properties are immediate consequences
of the component transformations being linear operations.
 It is immediate to verify that the sum of vectors has the usual properties of
the sum, namely commutative
(1.12)
and associative
(1.13)
Figure 1.6 shows the geometric meaning of the vector sum. In Fig. 1.6a the
sum is made putting the tail of B on the head of A; the sum is the vector from the
tail of A to the head of B, as one immediately understands thinking to the
components. For the commutative property, we might have done vice versa,
namely start from B and putting the tail of A on the head of B. We should have
reached the same point.

Fig. 1.6 The sum of two vectors

Figure 1.6b shows an equivalent way to sum, the parallelogram rule . We put
both vectors with the tails in the same point and we draw the parallelogram
having them as sides.
The vector difference between the two vectors A and B is the vector of
components equal to the differences between the homologous components or,
equivalently, the sum of A and –B. The geometrical meaning is shown in
Fig. 1.7.

Fig. 1.7 The difference between two vectors

The properties of vector sums, or composition, which we have just discussed

looks to be obvious, but they are not. Indeed they are valid if the space is flat,
not if it has any curvature. To make things simpler, consider two dimensions. A
plane surface is flat, but not a spherical one or a saddle shaped one. As a matter
of fact the surface on which we live, the surface of the earth, is flat only if we
consider distances substantially smaller than the earth radius, which has a mean
value R = 6 371 km, and only in a first approximation.
Let us consider the following example of vector addition. Consider a vector
with the tail in A at 45° in latitude and 0° in longitude and the head B on the
same meridian at 46° latitude. The length of one degree along a meridian is
everywhere 10 000 km/90 = 111 km. The second vector has the tail in B and the
head on the same parallel 100 km towards West, say in C. Now we commute the
operations. We start with a vector 100 km long from A to, say, D on its parallel
at 100 km to West. Then we add a 111 m long vector to the North with tail in D
and head, say, in C′. Will C′ be equal to C? The answer is NO. This is because
the distance between two meridians along a parallel is different at different
latitudes. Indeed the radius of the parallel at the latitude λ is r(λ) = R cos λ, that is
7071 km at 45° and 6947 km at 46°, which is 1.8 % shorter. Consequently, C′ is
1.8 km West of C.
QUESTION Q 1.2. Repeat the calculation with the same vector lengths starting
at 65° latitude.
QUESTION Q 1.1. Repeat the calculation with vector lengths of 1 km starting
at 45° latitude.
The question whether or not space has a curvature should be answered
experimentally, and experiments show this being the case. In particular, the
measurement of the mean curvature of the Universe over cosmological distances
is one of the important objects of contemporary cosmology. All measurements
are compatible with zero mean curvature, within their uncertainties. However,
we should mention that space curvature exists in another context. General
relativity describes local gravitational effects in terms of a modification of
geometry in the space surrounding a massive object. Its predictions are
confirmed by observations. We shall not deal with this topic in this book.

1.6 Scalar Product of Two Vectors

There two ways to take the product of two vectors, called dot product and cross
product respectively. We start here with the former.
Consider the two vectors A and B. Their dot product is indicated with a dot
between them, namely . In a given reference frame the dot product is, by
definition, the sum of the products of the homologous components
(1.14)
 The dot product has the important property to be scalar, namely invariant
under rotations of the axes. It is consequently also called a scalar product . Let
us show the property, namely that

For simplicity, let us consider only a rotation around the z-axis. The
components of A in the rotated frame as functions of its components in the
starting one are given by Eq. (1.10) and similarly for B. We can write

We see that the product is invariant.

It is easy to show that both the commutative and distributive properties are
valid for the dot product.
(1.15)
We shall see now the geometric meaning of the scalar product. We can profit
from it being invariant to choose convenient axes. We take x in the direction of
A and y in the plane defined by A and B (Fig. 1.18a). If θ is the angle between
the vectors, the components are A = (A, 0, 0) and . Their dot
product is then
(1.16)
In words, the scalar product of two vectors is the product of their magnitudes
times the cosine of the angle between them. There are also two other
interpretations that may be useful. The scalar product is the product of the
magnitude of the first vector times the projection of the second vector on the first
one (Fig. 1.8b), or, the same with inverted roles (Fig. 1.8c).

Fig. 1.8 Geometric meanings of scalar product

 The dot product is zero if the vectors are perpendicular, positive if the angle
is acute, and negative if obtuse.
A particular and interesting case is the product of a vector by itself
(1.17)
By definition the square of a vector is the dot product of the vector times
itself and is equal to the square of its magnitude and also to the sum of the
squares of its components. The latter property is an immediate consequence of
the Pythagorean theorem. It is also called the norm of the vector . The norm is
obviously the same in any reference.
Figure 1.9 shows a Cartesian reference frame in which three important
vectors are drawn, the unit vectors of the coordinate axes, i, j and k. They have
unit magnitude and are mutually normal. Consequently

Fig. 1.9 The unit vectors of the Cartesian axes

(1.18)
The components of any vector can be written in terms of the three unit
vectors. Indeed, the x component of the vector A is its dot product with i,
because the magnitude of the latter is 1, and similarly for the other components.
We then can write the vector as
(1.19)
namely as the sum of three vectors having the directions of the axes. These
are called the vector components .
In particular the position vector can be written as
(1.20)

1.7 Vector Product of Two Vectors

Given the two vectors A = (A x , A y , A z ) and B = (B x , B y , B z ), their cross
product is defined as the ordered triple of real numbers

(1.21)
We now show that the cross product transforms as a vector under rotations of
the axes and is also called the vector product . We show that for the x′
component, the demonstration for the other two are exactly the same.

The vector product is not commutative and the order of the factors matters.
We have immediately from the definition that
(1.22)
Inverting the order of the factors the product changes sign. The property is
called anticommutative.
It is easy to see that the vector product is distributive to the sum
(1.23)
We now see the geometric meaning of the cross product using the same
frame as in the previous section. We draw the two vectors as starting from the
same point and take the x axis in the direction and sense of A, the y axis in the
plane of the two vectors and the z axis to complete the right-handed reference
(Fig. 1.10). The components are A = (A, 0, 0) and . The
cross product has only the z component different from zero

Fig. 1.10 The vector product of two vectors

(1.24)
Hence, the cross product is in the positive direction of the z-axis if θ is in one
of the first two quadrants (Fig. 1.10a), in the negative one if in the third and
fourth ones (Fig. 1.10b).
In conclusion, the geometric meaning of the vector product, independently of
the reference frame, is the following. Its magnitude is equal to the area of the
parallelogram having the two vectors as sides. Alternatively, we can also say that
its magnitude is the magnitude of the first (A) times the projection of the second
on the normal to the first (B sin θ) or vice versa. The direction of the product is
perpendicular to the plane of the two vectors. Its sense is the one seeing the first
factor going to the second through the smaller angle in anticlockwise direction.
Notice that we have followed here the same convention we used to define the
positive direction of the z-axis. In a left-handed frame, the sense of the vector
product would have changed too.
The cross product is zero if one of the vectors is zero or if the two are
parallel. In particular the product of a vector times itself is zero.
Each of the unit vectors of the axes is the cross product of the other two
(1.25)
The expressions of this type can be easier remembered thinking that each of
them is obtained from the previous one by cyclic permutation. We now define
the scalar triple product of three vectors , in the order A, B and C. It is the dot
product of the first vectors times the cross product of the second times the third:

(1.26)
To see the geometrical meaning, we take the three vectors starting from the
same point as in Fig. 1.11.

Fig. 1.11 The scalar triple product

We can consider them as the sides of a parallelepiped. As we know the

magnitude of B × C is equal to the area of the parallelogram having the two
vectors as sides, which is a face of the parallelepiped. Its direction is the normal
to that plane and the positive sense is the one that sees B going to C, rotating
through the smaller angle, in anticlockwise direction. Let assume that A lies on
the same side of the plane made by B and C as B × C. The dot product of A
times B × C is the product of the projection of A on the direction of B × C hence
on the direction perpendicular to the plane of B and C times the magnitude of
B × C. But this projection is just the height h of the parallelepiped. In conclusion
the triple product is equal to the volume of the parallelepiped having the three
vectors as sides. In this case we are considering that this is true in absolute value
and sign. It is the opposite of this volume if case A lies on the opposite side of
the plane made by B and C that B × C.
The following properties are immediately demonstrated: the triple scalar
product is zero if the three vectors are coplanar, hence, in particular, if two or
three are parallel. The triple product does not vary if the factors are circularly
permuted
(1.27)
Obviously also
(1.28)
A second triple product is the triple vector product , which is the cross
product of the first vector times the cross product of the second and third ones.
By direct verification one shows that
(1.29)

1.8 Bound Vectors, Moment, Couple

The forces are vectors. However, to completely characterize a force we need also
to know its application point . If we push an object with our finger, we not only
exert on it an action of a certain intensity and in a certain direction, but also we
do that in a certain point. If we change that point, the effect of the force would
change. A vector with an associated application point is called a bound vector .
The line with the direction of the force through the application point is called the
line of action .
Figure 1.12 shows the vector A and its point of application P. It may be a
force for example. We arbitrarily choose a point Ω, which we call the pole . The
moment of A about Ω is defined as the vector product of the vector leading from
the pole to the application point of A, namely
Fig. 1.12 The moment of vector A about the pole Ω

(1.30)
Let us see its geometrical meaning. The direction of the moment of the
vector A is perpendicular to the plane defined by the segment ΩP and A. To see
its positive direction we imagine A to be a force and ΩP a rigid bar. If we see
the force turning the bar in an anticlockwise direction, we are on the positive
side of the moment. The magnitude of the moment is given by the product of
magnitude of the distance (h in the figure) of the pole Ω from the action line of
A. In particular, if Ω lies on the action line the moment is zero.
The importance of the moments will be clear when we study the mechanics
of the extended bodies in Chap. 7. We now consider a simple and particularly
important case, the couple of vectors. A couple is a pair of bound vectors equal
in magnitude in equal and opposite direction. The distance between the two
action lines is called the arm of the couple .
A very important property of the couple is that their moment is independent
of the pole. This may be called the moment of the couple or a couple torque .
The two terms are synonymous.
Consider for simplicity the pole Ω lying in the plane of the couple, as in
Fig. 1.13 (but the argument is valid in general). The two vectors are A and –A. P
1 and P 2 the application points respectively. The total moment, i.e. the sum of
the two moments about Ω is

Fig. 1.13 A couple of bound vectors

which is independent of the pole. We can also see that the magnitude of the
couple moment (or torque) is the product of the magnitude A of the vectors times
the arm d of the couple, namely
(1.31)
Its direction is perpendicular to the plane of the couple, positive on the side
seeing the couple rotate in an anticlockwise direction.

1.9 Matrices
Matrices are properly studied in mathematics courses. In this textbook only a
few simple concepts and definitions will be needed and are recalled here.
A matrix A is an array of numbers ordered in rows and columns, say M lines
and N columns

(1.32)

The matrix is said to be square if the numbers of rows and column are equal;
this number is called the order of the matrix . The generic element of the matrix
is a ij where the first index i (i = 1, …, M) is the row index, the second j (j = 1,
…, N) the column index.
Matrices with the same numbers of rows and column can be added. The sum
S = A + B of such matrices A and B is the matrix having as elements the sums of
the corresponding elements of A and B, namely s ij = a ij + b ij .
If the number of columns of the matrix A is equal to the number of rows of
matrix B the product P = A B is defined as follows. Be M the number of rows
and N the number of columns of A, N the number of rows and L the number of
columns of B. The product matrix P has M rows and L columns and its generic
element is

(1.33)
We can use the concept of matrix product to re-write Eq. (1.10) for the
transformation of a vector between two reference frames in compact form:
 (1.34)

We see that vectors are represented by a matrix with one column and three
rows, while the rotation is represented by a three-by-three matrix.
Continuing with the definitions, the minor A ij of the generic element a ij is
defined as the matrix one obtains from A suppressing row i and column j (i.e. the
row and the column to which the element we are considering belongs).
For square matrices, say A, the determinant can be defined. It is a number,
indicated with ||A|| or with det A. The definition is recurrent. If the order of the
matrix is one, its determinant is its only element. If the order is two,

(1.35)
If the matrix order is three or larger, one starts choosing a row (or a column).
It can be shown that the choice is arbitrary. We then choose the first row. Then
we multiply each element of the row times the determinant of its minor, keeping
it as it is, if the sum of the indices is even (11, 13, 15,…), changing its sign, if it
is odd (12, 14, 16,…). Finally we sum all these numbers. The determinant of the
3×3 matrix

is

(1.36)

It is easy to show that if two (or more) rows or two columns are equal, or
simply proportional, the determinant is null. It is also shown that the
determinants of two matrices differing only for the exchange of two contiguous
rows or two contiguous columns are equal and opposite.
The scalar triple product of three vectors, say A, B and C, can be usefully
expressed as the determinant of a 3 × 3 matrix of their components
 (1.37)

that is Eq. (1.26). The just mentioned properties of the determinant

correspond to the known properties of the triple product: it is null if two factors
are equal or parallel, i.e. with proportional components; inverting two factors the
triple product changes sign.
Finally, also the vector product of two vectors can be written formally as the
determinant of the matrix having in the first row the unit vectors of the axes, and
second and third rows the components of the two vectors in the order. Indeed

(1.38)

that is Eq. (1.21).

1.10 Velocity
We shall now study the motion of the simplest body, the material point or
particle . This is the case when its dimensions are small compared to the
distances from other objects. This is clearly an idealization but it works often in
practice. For example the planets are certainly not point-like, however in the
mathematical description of their motions around the sun they can be considered
as such in a good approximation, as long as we do not consider the rotations
about their axes, or the variations of the directions of those axes, or the tides on
their surfaces. A ship can be considered a point when she is far from shore, but
when she enters a harbor her dimension must be precisely known.
As we have already stated, the motion has to be studied in a given reference
frame. The particle describes in its motion a curve, which is called the trajectory
, as shown in Fig. 1.14a. The position vector is a function of time r(t) or, in other
words, the co-ordinates are three functions of time x(t), y(t), z(t). If we know
these functions we completely know the motion of the particle. We say that the
system has three degrees of freedom.
Fig. 1.14 a The trajectory of a particle, b the velocity

Let us consider the position vector at the instant of time t, r(t) as represented
in Fig. 1.14a and an immediately following instant t + Δt, r(t + Δt), where Δt is a
short time interval. In this time interval the particle has moved by Δ s, which is a
step in the space having a magnitude and a direction, namely it is a vector.
Looking at the figure one immediately sees that Δ s is equal to the difference
between the two vectors r(t + Δt) and r(t). This is the variation of the vector r in
the time interval Δt. Hence
(1.39)
The average velocity in the time interval Δt is the vector obtained by dividing
the displacement by the time interval in which it happens:

(1.40)
or, for the components

(1.41)
Velocity is the limit for of the average velocity, namely

(1.42)
In words, the velocity is the time derivative of the position vector. Its
components are the derivatives of the coordinates

(1.43)
In the limit the direction of Δ s becomes tangent to the trajectory, in
every point of the trajectory the direction of velocity is that of the tangent in that
point (Fig. 1.14b).
The physical dimensions of velocity are those of a length divided by a time;
the unit is consequently the meter per second (m/s or ms−1).
The motion is said to be uniform, if the magnitude of velocity does not vary
in time. In a uniform motion however, the velocity is not necessarily constant,
because its direction may vary. The direction of velocity does not vary if the
motion is rectilinear. Hence a motion with constant velocity is rectilinear
uniform .

Example E 1.1
The motion of a particle is known when its three co-ordinates as functions of
time are known. Consider the motion given by the equations

where a and b are constants.

The co-ordinates y and z are always zero. Consequently the motion is along
the x-axis, hence rectilinear. In the initial instant (t = 0) the particle is in the
position x(0) = c. It is called the initial position. As time varies the position
varies in proportion, b being the proportionality constant. The particle moves in
the positive x direction (increasing x) if b > 0, in the negative one (decreasing x)
if b < 0. The velocity has only one component different from zero,
Hence, the motion is also uniform.

Example E 1.2
Consider the motion given by the equations

Now the motion takes place in the xy plane, because the z co-ordinate is
always zero. The initial position is

In order to find the equation of the trajectory we may take the ratio of the
distances travelled in the same time along y and x. We find

which is a constant. This means that the trajectory is the straight line through
the point (c1, c2) and making with x-axis the angle arctan (b 2/b 1).
Hence, the motion is rectilinear as shown in Fig. 1.15.
Fig. 1.15 Geometry of the motion of E 1.2

The components of velocity in the plane of the motion are and

The velocity vector is then v = (b 1, b 2, 0), having the same
direction as the (rectilinear) trajectory. It is consequently rectilinear uniform.

Example E 1.3
Consider the motion
(1.44)
and let us calculate the velocity. There is only one non-zero component, namely
The velocity is not constant but increases (decreases) linearly
with time if b > 0 (b < 0). The motion is rectilinear but not uniform.

As we have already seen, the motion of the bodies is always relative to the
assumed reference frame. Consequently also the velocity is relative to the frame.
In Chap. 5 we shall study in detail the relations between the kinematic quantities
(position, velocity, acceleration, etc.) in different frames in relative motion. We
anticipate here a simple concept, the relative velocity .
The velocity of a body relative to another one is the vector difference
between their two velocities. Indeed, let r 1 be the position vector of the first
body and r 2 that of the second. The position of the second body relative to the
first is the vector . The time derivative of this vector is the velocity of
2 relative to 1, which is the velocity of 2 seen by an observer travelling with 1.
Calling it v 12 we have
(1.45)
The velocity of a passenger walking on the deck of a ship relative to the
vessel is the difference between the velocity vectors of the passenger and of the
ship relative to the sea.
 Notice that the position of 1 relative to 2 is the opposite of the position of 2
relative to 1. The same is true for the velocities.

Example E 1.4
Consider two ships, A and B, which at a certain instant are in the position shown
in Fig. 1.16. Their velocities are v 1 and v 2 respectively. The two courses
intercept in the point o P. Will the ships collide in P if they move with constant
velocities?

Fig. 1.16 Motion relative to the sea and of one sheep relative to the other

The answer is immediate in a frame fixed with one of the two vessels, for
example with A as in Fig. 1.16b. In this frame, all the relevant velocities,
including that of the sea, are obtained from those relative to the sea by
subtracting v 1. Hence A does not move (by definition) and B moves with
velocity v 2 − v 1. The vector R leading from A to B is the same in the two
frames (they differ by a translation). Ship B, as seen by A, moves on the course
shown in the figure. Hence the minimum distance she will pass from A is AC,
namely the distance of A from the straight line B travels. In conclusion, they will
pass close but will not collide.
Notice on purpose that a passenger A sees B moving sideway, not in the
direction of bow. Indeed, we have a strange impression when we cross closely
another ship, particularly offshore, when any reference to ground is missing. She
looks to be travelling in a not “natural” direction.

1.11 Angular Velocity

An important motion is the circular one, in which the trajectory is a circle. Let R
be its radius. It is always convenient to choose the reference frame taking profit
from the symmetry of the problem, if any is present. We take the origin in the
center of the circle and the z-axis perpendicular to its plane. The motion is then
in the xy plane as shown in Fig. 1.17a.

Fig. 1.17 a The circular motion, b an infinitesimal movement, c angular velocity ω

We further choose the origin of time in the moment in which the point
crosses the positive x-axis. Let ϕ(t) be the angle between the position vector and
the x axis at time t, taken as positive in anticlockwise direction and let s(t) be the
length of the arc subtended by ϕ(t), taken with the same sign as ϕ, namely s(t) =
R ϕ(t). Let d s be the infinitesimal movement in dt (Fig. 1.17b). The infinitesimal
changes of s and ϕ are linked by the relation ds = R dϕ, where in our notation ds
is the magnitude of d s if the motion is anticlockwise (as in Fig. 1.17), and is
opposite if clockwise. The angular velocity measures the rate of change of the
angle. We then consider the time derivative

(1.46)
This quantity has magnitude and a sign, depending on the sense of rotation.
In fact, it is the z component of the angular velocity, which is a vector. Its
magnitude is the absolute value of Eq. (1.46), its direction is perpendicular to the
plane of the motion, taken positive on the side seeing the motion is
anticlockwise. This is the z-axis in Fig. 1.17c.
The physical dimensions of the angular velocity are the inverse of time; its
unit is radians per second (rad/s)
In a circular motion, the magnitudes of velocity υ = |ds|/dt and the magnitude
of the angular velocity ω are related by
(1.47)
The relation between the corresponding vectors, as immediately seen from
Fig. 1.17c is
(1.48)
Let us consider the case in which the magnitude υ of the velocity is constant.
The motion is circular and uniform, the arcs and the corresponding angles are
proportional to the times taken to travel them, namely
(where, as usual the sign is positive if the direction is anticlockwise and vice
versa). Hence we have the equations of motion in polar co-ordinates:

(1.49)
The equations of motion in Cartesian co-ordinates are
(1.50)
As an exercise we can check that the trajectory is indeed a circle. Taking the
squares of the members and summing we have which is
the equation of a circumference. Notice that the two Cartesian co-ordinates x and
y are not independent but if we know one we know also the other. In fact the
particle is bound to travel onto a prefixed trajectory. The system has one degree
of freedom. This is evident in polar co-ordinates, Eq. (1.49). Two of them are
constant.
We now express the Cartesian components of velocity

(1.51)
The components of the velocity vector change in time: when the particle
moves on the circle its direction continuously varies even if its magnitude is
constant. Indeed, the magnitude is

(1.52)
which is a constant.
As a further exercise, let us check that the velocity is always tangent to the
trajectory, i.e., perpendicular to the position vector everywhere. To see that we
take their scalar product and get

We now make the following observation that will be useful in the following.
In the case we have noted that we have two vectors: the position vector and the
velocity. The x and y components of the first vector are proportional to the
cosine (Eq. 1.50) and the sine of the angular co-ordinate respectively, those of
the second to the opposit of its sine and to its cosine respectively (Eq. 1.51).
When this happens the two vectors are perpendicular.
Both the co-ordinates and the components of velocity are proportional to the
circular functions cosωt or sinωt, which are periodic. In fact the motion is
periodic , meaning that if position and velocity have some values in the instant t
they have again the same values at the instants t + T, t + 2T, etc., for every t. The
time T is called the period of the motion. It is inversely proportional to the
angular velocity

(1.53)

1.12 Acceleration
The motion of a body in which the velocity varies with time in magnitude or
direction is called accelerated. If the change of velocity in the time interval Δt is
Δ v, the average acceleration in that time interval is the ratio

(1.54)
The instantaneous acceleration at time t is its limit for , namely the
time derivative of the velocity

(1.55)
In the particular case of the rectilinear motion, when the direction of the
velocity is constant, the acceleration direction is also on the line and its
magnitude and sign are

(1.56)

Example E 1.5
Consider again the motion of Example E 1.3, namely

The motion is along the x-axis with velocity . The x component of

the acceleration, the only different from zero, is then . The
acceleration is constant in magnitude and direction. Such motions are called
uniformly accelerated .

We now consider a uniform circular motion in which the velocity vector has a
constant magnitude and varies in direction with constant angular velocity. In
order to find the acceleration, consider the auxiliary diagram of Fig. 1.18a (we
assume an anticlockwise rotation direction). The axes of the figure are the x and
y components of the velocity vector that we think of as having its tail in the
origin. It is analogous to the position vector in the xy plane. The analogy is
complete because both vectors rotate with constant angular velocity ω. In other
words, the head of the velocity vector A describes a circularly uniform motion in
the velocity plane, having a radius equal to its magnitude υ.

Fig. 1.18 Uniform circular motion

Clearly the “velocity” of point A is just the acceleration of particle P because

the displacement of A in the time interval dt is d v and consequently its
“velocity” is . This vector is tangent to the circle and consequently
perpendicular to the velocity (Fig. 1.18a). More precisely, the direction of
acceleration is obtained from that of the velocity by a rotation of 90° in an
anticlockwise direction. Going back to the representation of the motion in the xy
plane in Fig. 1.18b, the acceleration, which we just saw to be at 90° from the
velocity anticlockwise, is radial directed towards the center. It is then called
centripetal acceleration .
We immediately find the magnitude of the acceleration. We denote by α the
angle between the vector v and the abscissa axis in Fig. 1.18a, and dα its
variation in the time dt. Considering that the vector rotates with constant angular
velocity ω, we have dα = ω dt. On the other hand the change in velocity is
and we get the important relation

(1.57)
Summing up, if the velocity varies only in magnitude, the acceleration is
parallel to velocity, if the velocity varies only in direction, the acceleration is
perpendicular to the velocity, directed towards the center of the trajectory. We
shall see in Sect. 1.14 that in the general case in which both magnitude and
direction of velocity vary, acceleration has two components one parallel and one
perpendicular to velocity.

1.13 Time Derivative of a Vector

In the study of uniform circular motion we have dealt with the position vector r
and the velocity v. Both are constant in magnitude and vary in direction with
time, rotating at the angular velocity ω. We have seen that the magnitudes of
their time derivatives are respectively ωr and ωυ, namely, in both cases the
magnitude of the vector times ω. In both cases the direction of the derivative
vector is at 90° forward to the original vector . The result is valid also if the
angular velocity is not constant. Indeed, we did not use this assumption.
We now generalize the argument as follows, with reference to Fig. 1.19.
Consider the vector function of time A(t), constant in magnitude, varying only in
direction. At the generic time instant the vector rotates with angular velocity ω,
not necessarily constant. Let u p be the unit vector rotated by π/2 relative to A in
the direction of the rotation. The time derivative of A is

Fig. 1.19 A rotating vector and its time derivative

(1.58)
or better

(1.59)
This important formula that we shall use often in the following is due to
Siméon-Denis Poisson (1781–1842) and is called a Poisson formula . It is valid
if the magnitude A is constant.
In the general case in which the vector A varies both in direction and
magnitude, its time derivative is immediately obtained by writing the vector as
the product of its magnitude and its unitary vector
 But the vector u A is constant in magnitude, being unitary, and we can use the
Poisson formula for its derivative. We get

(1.60)
which is an important result that we shall use often in the following.

1.14 Motion on the Plane

We now consider a general motion in a plane. We indicate with u t the unit
vector tangent to the trajectory in its generic point in the direction of the velocity
in that point. In general u t varies in time. Figure 1.20 shows the situation in two
consecutive instants.

Fig. 1.20 a The acceleration vector in two different points of the trajectory; b the osculating circle

In every instant, i.e. in every point of the trajectory, in general the velocity is
different. We indicate with u n the unit vector normal to the trajectory. Its
positive direction is the direction obtained by rotating u t by 90° in the direction
of the instantaneous rotation of the velocity vector. This geometrically means
that u n is directed towards the curvature center. The latter may lie on the left or
the right of the trajectory depending on the case. To obtain the acceleration we
take the derivative of the velocity expressed as the product of magnitude times
unit vector, v = υ u t .

(1.61)
As anticipated, the acceleration has two components. One is tangent to the
trajectory and equal to the time derivative of the magnitude of velocity. It is null
if the motion is uniform, positive if it is accelerated, negative if decelerated. The
other component is normal to the trajectory in any case towards the “interior” of
the curve. It is zero when the direction of the velocity does not vary, even if
instantaneously, as in the flex points of the trajectory.
We can express the normal component of the acceleration in terms of the
curvature radius of the trajectory in the point P under consideration.
Figure 1.20b shows the situation. Consider all the circles tangent to the curve in
P having radiuses between 0 and infinity. One of these gives locally the best
approximation of the curve. It is called an osculating circle , from the Latin word
osculum, meaning kiss. Its radius R is called the curvature radius of the curve in
the point P. Its reciprocal is the curvature . In an inflexion point the curvature
radius is infinite and the curvature is null.
Now we can approximate the small curve segment around P with the arc of
the osculating circle and think of the point as moving on that arc with angular
velocity ω = υ/R. In conclusion, the two components of the acceleration are

(1.62)
We see that the normal component of the acceleration is proportional to the
curvature and to the square of the velocity.

1.15 From Acceleration to Motion

Figure 1.21 represents the trajectory of a material point P in a given Cartesian
reference frame, its position vector r(t), its velocity v(t) and its acceleration a(t),
that are all functions of time.

Fig. 1.21 Trajectory, position vector, velocity and acceleration

We recall their expressions

 (1.63)

(1.64)

(1.65)

(1.66)
In words, the velocity is the time derivative of the position vector and the
acceleration is the time derivative of the velocity or the second time derivative of
the position vector. We shall see in the next chapter that acceleration is
proportional to the force.
We consider now the inverse problem, namely to find the velocity and the
law of motion once the acceleration a(t) is given. As the velocity is the time
derivative of the position vector, the latter is given by the integral of the velocity
on time from the initial instant t 0 to the time t considered, namely

In general, we want to know the position of P at the time t and rewrite the
expression as

(1.67)

We see that knowledge of the velocity v(t) is not sufficient. We need also to
know the position of the body at a certain instant t 0. This instant can be any, but
generally we know how the motion began, namely we know the initial position .
It is customary to choose that instant as the origin and t 0 = 0.
To a question like “A car has been travelling at a constant speed of
100 km/h. Where is it after 2 h?” We can only answer it has travelled 200 km.
We can know its position only if we know from were it started.
Equation (1.67) corresponds to three integrals
 (1.68)

If we want to know the velocity for a given acceleration, we proceed in the

same manner by integrating

(1.69)

or, in terms of the components

(1.70)

Again, we need to determine the integration constants, namely the velocity at

a certain instant, which is usually the initial one.
Once the velocity is known we need to integrate again to have the law of
motion. For that we need to know both the initial position and the initial
velocity.

1.16 Free Fall Motion

The study of the free fall motion of bodies near the surface of earth is an
important example of the use of the just developed formalism. With “free fall ”
we mean an idealized situation in which the air resistance can be neglected and
the bodies move only under the action of gravity We anticipate that under these
conditions the vertical and horizontal motions are independent from one another,
as we shall study in Sect. 3.7, and that any free body moves with a constant
acceleration, g, which is vertically directed downwards and has a magnitude
(approximately) g = 9.8 m/s2. We choose a reference frame with the z-axis
vertical upward, and the x and y-axes in a horizontal plane, for example the
ground. The acceleration of the body, that we shall consider point-like, P, has
the components
(1.71)
The motion of the body depends on the initial conditions. If for example we
drop the body from a certain height with null velocity it will move vertically
down with uniform acceleration. If we launch it vertically upwards it will
gradually slow down, stop and then fall down. If we launch it at an angle with
the horizontal it will describe a curved trajectory, etc. Let us study these
motions.
Let us start from the simplest case. We drop the body at the height h above
ground with null velocity at t = 0. The initial conditions are

The x component of the velocity at the generic time t is

The x component of the velocity is identically zero (i.e. is zero at every

instant of time) because the x components of both acceleration and initial
velocity are zero. A similar argument leads immediately to conclude that also
x(t) = 0. The same is true for the y components of velocity and position vectors.
Notice that the initial conditions x(0) = 0 and y(0) = 0 depend on the reference
frame. Its origin has been chosen in such a way to have the point from which we
drop the particle on the z-axis. A different choice would have led to the initial
conditions, say, x(0) = a, y(0) = b. The two co-ordinates as functions of time
would have been x(t) = a, y(t) = b. The motion obviously is the same.
We have found that the motion is along the z-axis. As the acceleration is
constant, it is uniformly accelerated (acceleration may have both signs, if we
want to be specific we can say accelerated, if the acceleration is positive, delayed
if it is negative). Let us now find the velocity in the z direction.

(1.72)

Velocity is always negative. Indeed the body moves always in the z direction
we have chosen as negative. We now integrate once more to find the position as
a function of time
 (1.73)

which is the law of the motion. Knowing completely the motion, we can look
for interesting properties, for example the time taken to reach the ground. This is
the instant in which z = 0, hence and the velocity in that instant

(1.74)
Consider now the same initial conditions with the difference that the initial
velocity has a nonzero vertical value υ 0. With the same arguments as before, we
obtain
(1.75)

(1.76)
We should now distinguish the two cases of positive (downwards) and
negative (upwards) initial velocity.
If υ 0 < 0, the velocity is always negative. To find the instant t in which the
body is at the height z we solve Eq. (1.76), obtaining

We have two solutions because Eq. (1.76) is of second degree in t. However,

in the case we are considering, one of them, the one with the negative sign, is
always negative and consequently does not have physical meaning. We must
choose the solution with a positive sign, because the motion starts at t = 0.
The time of arrival at ground, the duration of the fall, is the time at which
z = 0, namely

which is shorter than in the case of null initial velocity. Obviously the
expressions found in the latter case are particular cases.
If υ 0 > 0, from Eq. (1.75) we see that the velocity is positive, namely
upwards, for a while, but it diminishes with increasing time. It is zero in the
instant t m = υ 0/g, and negative in later times. Indeed, the body reaches the
maximum height at t m , namely (see Fig. 1.22a). In this
case both roots for t(z) have physical meanings provided t ≥ 0. Indeed, the body
goes twice through the same height, if it is z ≥ h, first going up later going down.
If z < h one solution is negative and again does not have physical meaning.

Fig. 1.22 Free fall trajectories with initial velocity a vertical upward, b at an angle α with the horizontal

Why may it happen that a mathematical solution should be discarded on

physical grounds? The reason is that the equations stating the “initial conditions”
do not give information on the system before the “initial” instant. In this case the
body stood still, say in our hand. But it would have been possible that it was
moving upwards in such a way as to reach z = h at t = 0 with velocity equal to υ
0. The discarded solution would have made sense.
We now suppose that the initial position is again at the height h above
ground, but that the velocity v 0 is at an angle α with the horizontal. This is what
happens when shooting with a cannon from the top of a tower. We choose the z
vertical upwards as before and the x horizontal in the plane of z and of the initial
velocity. The initial conditions are

The motion is in the plane xz, as show in Fig. 1.22b. We find, as usual, the
velocity using Eq. (1.69) and the initial conditions.
(1.77)
We see that the horizontal, x, component of the velocity is constant and equal
to its initial value and that the vertical one, z, decreases linearly in time, exactly
as in the case we have considered.
We integrate once more and use the initial conditions to obtain the law of
motion, finding
 (1.78)
or

(1.79)
We now know completely the motion. If, for example, we want to know the
shape of the trajectory we must eliminate t from the equations for the co-
ordinates. From the first one we have , which, substituted in the
second equation, gives

(1.80)
which is the equation of a parabola. The distance x f at which the body
touches the ground, namely the range of the weapon, is the value of x
corresponding to z = 0. We then put this value in Eq. (1.80) and solve for x. We
find

(1.81)
The negative root solution is for t < 0 and corresponds to the intersection of
the parabola on the left of the tower. It is shown dotted in Fig. 1.22b and should
be discarded. The positive root is the solution for which we searched.
We now find the duration of the shot, which is the time t f at which the body
touches ground. With x = x f the first of the (1.79) solved for t gives

(1.82)
We now find the maximum height z m reached by the body. This can be done
in different ways. One is noticing that this is the height at which υ z = 0. From
Eq. (1.78) we see that a happening at , which was substituted in the
second Eq. (1.79), gives

The same result can be reached finding the maximum of the second
Eq. (1.79).
It is interesting to consider the special case α = 0. We want the time t f taken
by the bullet to reach ground. Equation (1.79) become
 The bullet hits the ground in the instant , which, as we see, is
independent of υ 0. This implies that for whatever initial velocity, even if
enormous, the time taken to fall from the height h is always the same and is then
equal to the free vertical (the special case υ 0 = 0). In other words, the vertical
and horizontal motions are independent.
The law of independence of (the components of) motion was discovered by
G. Galilei . In the “Dialogue concerning the two Chief World Systems” he writes
(translation by the author):

…suppose having on the top of a tower a horizontally arranged culverin (a

relatively light cannon) and firing point-blank shots, namely parallel to the
horizon; then for little or much gunpowder charge given to it, such that the
cannonball would fall at a distance of either one thousands arms, or four
thousand, or six thousand, or ten thousand, etc., all these shots would take
place in times equal to each other, and each equal to the time the ball would
take to fall from the cannon’s mouth to earth, when dropped, without any
other impulse, for a simple vertical fall. Indeed it looks really wonderful
that in the same short time of the vertical fall from a height, for example, of
one hundred arms, could the same ball travel either four hundred, or one
thousand, or four thousand, or even ten thousand arms, in such a way that in
all the point-blank (horizontal) shots it would be in the air for equal times.

A little later Galilei specifies that that would be true

…when there were no accidental impediments by the air…

1.17 Scalars, Pseudoscalars, Vectors and Pseudovectors

In Sect. 1.4 we have defined the vector as an ordered triple of real numbers that
under rotations of the reference frame transforms in the same way as the triplet
representing the position vector.
In Sect. 1.6 we have met a scalar quantity, the dot product of two vectors.
We have seen that it is the same in two reference frames differing for a rotation
of the axes. Indeed, in general, a quantity is, by definition, a scalar if it is
invariant under change of the reference frame. For example, the x component of
a vector is a single number but is not, properly speaking, a scalar, because it is
not invariant under rotations of the axes.
Hence, both vector and scalar properties are expressed in terms of
transformations between reference frames. We shall now consider the behaviors
of these quantities under the inversion of the axes. It is called parity operation. It
leads from a, say, left-handed frame to a right-handed one.
We now consider the transformation properties of physical quantities.
A quantity can be scalar or pseudoscalar . Both are invariant under rotations
but the former is invariant under parity operation, the latter changes sign, while
keeping its absolute value.
The dot product of two vectors is a scalar; the “scalar” triple product is a
pseudoscalar. This is immediately evident considering that under inversion of
the axes all the three vector factors change sign.
A quantity can be a vector or a pseudovector (also called an axial vector ).
Both transform in the same way under rotations, but the components of the
former change sign under inversion of the axes, as the position vector does,
while the components of the latter do not change sign.
The cross product of two vectors is a pseudovector, because both the vector
factors change sign and their product does not. We met both types of physical
quantities. Position vector, velocity and acceleration are (proper) vectors;
angular velocity and moment of a vector are pseudovectors.
This type of properties of the physical quantities belong to a class generically
called symmetry properties .

1.18 Problems
1.1. The vector V varies by Δ V, its absolute value varies by ΔV in the time
interval Δt. (a) Can ΔV be larger than the magnitude of the variation,
namely |Δ V|? Can they be equal?

1.2. The vector V changes its verse. Express Δ V, ΔV and |Δ V|?

1.3. At the instant t 1 the velocity of a body is, with certain units, v 1 = (1, 3, 2),
at time t 2 is v 2 = (5, 3, 5). Find: (a) The variation of the velocity Δ v, (b)
the magnitude of the variation of the velocity |Δ v| and (c) the variation of
the magnitude of velocity Δυ.
1.4. A particle travels on a circle with velocity υ constant in magnitude. After a
complete turn, (a) which is the mean value of υ? (b) which is the mean
velocity <v>?

1.5. A particle moves with a position vector, in the given frame,

. Find: (a) velocity and acceleration as functions of
time, (b) the velocity at t = 2 s.

1.6. A point moves uniformly on a plane curve trajectory with velocity υ. The
magnitude of acceleration on a certain point of the trajectory is a. What is
the curvature radius in that point?

1.7. The position vector of a point is . (a) Find the

velocity and acceleration vectors and their magnitudes. (b) Express the
scalar product of r and v. What does the result mean? (c) Express the scalar
product of r and a. (d) Find the trajectory of the point. (e) How would the
motion change changing the sign of y(t)?

1.8. A cyclist travels at 10 km/h heading north. Wind blows with a speed
(relative to ground) of 6 km/h from a direction between N and E. To the
cyclist the wind appears to come from the direction at 15° from North to
East. (a) Find the speed of the wind relative to the cyclist and the direction
of the wind, relative to ground. When the cyclist goes back, which are
velocity and apparent direction of the wind (wind did not vary).

1.9. We are on a ship travelling at 10 kn heading east. We see another ship,

which we know moves at 20 kn to North, 6 miles distant in the South
 direction. What is the minimum distance the two ships will be (without
changing their courses)? After how much time? Refer to Fig. 1.16. N.B. On
the sea distances are measured in nautical miles and velocities in knots
(1 kn = 1 mile/h). Assume for the mile the round figure of 1800 m.

1.10. Consider a flat platform rotating with angular velocity ω 1 = K t 2 k where

k is the unit vector of the z-axis directed vertically upwards. A body on the
platform rotates with angular velocity, relative to it, ω 2 = 2 Kt 2 i (the x
axis is horizontal). K = 1 rad/s3. (a) Find the direction of the body relative
to the ground. (b) Find the angle ϕ of which the body has rotated relative
to ground at t = 3 s. (c) Does the magnitude of the resultant angular
velocity vary in time? And its direction?

1.11. An airplane is flying at constant velocity V, of horizontal direction and

magnitude V = 100 m/s at the height h = 5000 m. A (super)cannon on
earth shoots against it a ball at the moment in which the plane is just above
the weapon (Fig. 1.23). The velocity of the ball is υ 0 = 500 m/s.
Neglecting the presence of air, find: (a) the angle α at which we must
shoot to hit the plane; (b) the time of the collision (which of the two
solution should be chosen?); (c) how much did the plane travel up to this
moment.

Fig. 1.23 The plane and the cannon of problem 11

1.12.
The wheel shown in Fig. 1.24 rotates without slipping. Its axis moves forward at
the velocity v. Find the velocities (namely their components on the two co-
ordinate axes) of the points A, B, C.

Fig. 1.24 The wheel of problem 12

© Springer International Publishing Switzerland 2016
Alessandro Bettini, A Course in Classical Physics 1—Mechanics, Undergraduate Lecture Notes in Physics,
DOI 10.1007/978-3-319-29257-1_2

2. Dynamics of a Material Point

Alessandro Bettini1
(1) Dipartimento di Fisica e Astronomia, Università di Padova, Padova, Italy

Alessandro Bettini
Email: alessandro.bettini@pd.infn.it

In this chapter we study the dynamics of a material point, namely the laws
governing motion by its causes, which are the forces. We shall then start by
defining and discussing the concept of force. The experimental method was
introduced by Galileo Galilei at the end of the XVI century. He also discovered
part of the laws of mechanics. The complete theory of mechanics was built by
Isaac Newton, who published in 1686 the “Philosophiae Naturalis Principia
Mathematica”, known generally as simply “Principia”.
The law of inertia was discovered by Galilei and assumed by Newton as the
first law of mechanics. It will be studied in Sect. 2.3. The law states that a body
in absence of forces acting on it moves naturally with constant velocity in a
straight line, a rectilinear uniform motion. The second law was also discovered
by Galilei and precisely formulated by Newton. It states that the rate of change
of the momentum, a vector that we shall define, namely its time derivative, is
equal to the force acting on the body. In an equivalent manner the acceleration is
proportional to the force. This is the subject of Sect. 2.4. In the same section we
shall discuss Newton’s third law, the action-reaction law.
There are several types of force in Nature, as we shall see in the next chapter.
In this one, however, in Sect. 2.5, we shall talk of weight, the force acting on all
the bodies near the surface of the earth. A few examples will be discussed in
Sects. 2.6 and 2.7.
In Sect. 2.8 we introduce two of the fundamental mechanical quantities
(beyond momentum, or quantity of motion, already introduced in Sect. 2.4), the
angular momentum and the moment of a force.
In Sect. 2.9 we shall study a simple but very important system, the pendulum
and its harmonic motion. We shall also see how two concepts of mass, the
inertial and the gravitational mass, are in fact only one.
After having introduced the concept of work made by a force and shown the
theorem of energy conservation in Sect. 2.10, we shall describe an interesting
experiment by Galilei. It establishes that the work done on a body by the weight
force depends only on the difference between initial and final heights, not on the
particular path followed. In modern language the experiment established that the
weight force is conservative. This very important concept will be defined in
Sect. 2.13. We then demonstrate the energy conservation theorem. Energy
conservation is a fundamental law of all physics. We shall deal in this book only
with mechanical energy, in its kinetic and potential forms, but we warn the
reader that other important forms of energy exist, in particular thermal energy, as
we shall discuss in the second volume of this course when dealing with
thermodynamics.
The historical process leading to a precise definition of the concept of energy
and to the establishment of the law of energy conservation took more than two
centuries. Starting with Galilei, it came to maturity around mid XIX century,
with the experiments of Mayer and Joule and enunciation of the energy
conservation law by Mayer and Helmholtz. We shall give some hints in
Sect. 2.14.
In Sect. 2.15 we shall discuss a particular type of force, the central forces.
The gravitational attraction of the sun on a planet is an important example of this
category.
In the last paragraph we introduce the concept of power, which is the work
done by a force per unit of time.

2.1 Force, Operational Definition

The primitive concept of force is linked to muscular strain. If we lift a weight,
push an object, we must exert a force with our hands and arms and we feel
strain. Since ancient times humans developed simple mechanical devices to exert
forces or amplify the muscular effect. The string of an archer’s drawn bow exerts
a force on the arrow, throwing it in the air; a lever can be used to lift big weights,
etc. However, in physics the concept must be quantitative. For that, we must
define force accurately enough to be able to measure it. This means that we must
be able to compare two forces and establish when they are equal, when one is
twice the other, etc. In other words we must be able to determine the ratio
between two different forces.
A direct method to compare two forces is based on the lever rule , which was
discovered by Archimedes of Syracuse (287–212 BC) more then two thousand
years ago. The rule states that two equal forces balance when applied at equal
distances on two sides of the pivot (Fig. 2.1a) and that two different forces F 1
and F 2 balance when applied at distances from pivot (l 1 and l 2 respectively)
inversely proportional to the forces (Fig. 2.1b), i.e. such as

Fig. 2.1 Comparison of two forces

(2.1)
The first statement can be proven simply with symmetry arguments. If the
two forces are equal and the two arms are equal, the system is symmetric. How
could it choose on which side to bend? The second statement on the contrary,
namely the validity of Eq. (2.1), must be experimentally verified.
We know that a spring exerts a force when compressed or stretched relative
to its natural length; we feel the muscular strain when we compress or pull it. We
build a certain number of springs as equal as possible to each other. We can then
verify that they exert equal forces when compressed (or stretched) in the same
measure by applying those forces at equal distances from the pivot of a lever as
in Fig. 2.1a. We can now define as unitary the force expanded of a specific
length (N.B.: this is not the official definition).
We can then define the multiples of the unit force. If for example, we want a
force of three units, we put three of our springs in parallel. We can
experimentally verify the lever rule Eq. (2.1) as shown in Fig. 2.2b with different
combinations of unit forces. Once we have stated that, we can use it to measure
forces. As a matter of fact the method has been used in steelyards since very
ancient times and is still used now in fruit or other goods markets to weigh a
wide variety of goods. The weight to be measured is compared with the weight
of a standard object seeking for equilibrium by changing the length of the lever
arm of the latter.

Fig. 2.2 The dynamometer

In the operational definition of the force we have just chosen, we did not
make any hypothesis on the relation between the force exerted by the spring and
its length. However, this definition is not simple to use in practice. A handier
device is the dynamometer (from the Greek dynami for force and metro for
measure).
The dynamometer, shown schematically in Fig. 2.2, is made of a spring fixed
at one extreme on a wood, or other material, plate and with a ring at its other
extreme. The force to be measured is applied to the ring. A pointer moving on a
scale gives a measurement of the dilation of the spring. Once we have built the
device we must calibrate it. With the above described procedure we have built a
number of springs, multiple and submultiples of the unit. We apply each of them
to the ring and mark the position of the pointer on the table. In this way we build
a scale on which we will read the values of unknown forces. In practice, we find
that the scale is linear, namely the stretch is proportional to the applied force, if
the stretch is not too large. However, this property is comfortable, but not
necessary.
The method we have described is used in practice, but does not allow a
precise definition of force. In the SI the unit of force is a derived one, It is the
force imparting the unit acceleration (1 m/s2) to the unit mass (1 kg). It is called
newton (N). To have an idea of the order of magnitude, think that the weight of
one liter of water, 1 kg, is about 9.8 N. In other words one Newton is about the
weight of the water filling a glass.

2.2 Force Is a Vector

In giving the operational definition of force in the previous section we have
implicitly assumed, and we did that by definition, that two equal and opposite
forces when applied to a point do not cause acceleration. Namely, the two forces
are in equilibrium. Clearly, a force not only has a magnitude but also a direction.
We can exert a force on a body applying one of our springs and pulling in
different directions. We are led to think that force is a vector quantity. However,
the conclusion cannot be reached by logic, rather it needs experimental
verification. To be a vector, a quantity not only should have a magnitude and a
direction, but also satisfy the rule of addition of vectors.
The experience with three forces was originally devised by Pierre Varignon
(1654–1722), a contemporary of Newton. Its device is shown in Fig. 2.3. In the
plane of the figure, which is vertical, three pulleys are fixed. The three weights
of masses m 1, m 2 and m 3, act by means of wires, drawn in the figure, joined in
the point O. The forces exerted by the wires have magnitudes proportional to the
weights and the directions of the wires. Once we have joined the three wires in O
and let the system alone, the system moves until it reaches its equilibrium
configuration, the one represented in the figure. We know the values of the
weights, say F 1, F 2 and F 3, and measure the angles θ 1 and θ 2. We find that
the following relations are satisfied:

or

Fig. 2.3 Varignon experiment showing the composition of forces

The Varignon experiment and similar ones made afterwards verify the vector
character of the force. The most precise tests, however, are indirect and come
from the agreement of the experimental data with the predictions made under
this hypothesis in the most different conditions.
Once we have established that forces add as vectors, we define as the
resultant of the set of forces F 1, F 2, F 3, … and their vector sum
(2.2)
Let us now think of some forces that we know from our everyday
experience. We can distinguish two types. The just considered forces exerted by
a spring, the force a table exerts on an object it supports, the force we exert with
our hand pushing an object, are each exerted by contact. A body, the spring, the
plane of the table and the hand each apply force to the object touching it. The
everyday example of the second type of force is weight. Weight is the force with
which earth gravitationally attracts all bodies. It is directed vertically down,
towards the center of earth. This force is exerted at a distance i.e., it does not
need contact.

2.3 The Law of Inertia

One of the most revolutionary discoveries of Galilei was the establishment of the
behavior of a body not subject to forces. The problem lies in the fact that in
practice it is impossible to eliminate all the forces. Weight is always present on
earth. It cannot be eliminated, but it can be balanced. If we put a body on a
horizontal plane, the latter will exert on the body a force equal and opposite to its
weight. However, when the body moves, frictional forces due both to the contact
between the surfaces of the plane and the body and the air are present. The effect
of these “passive” forces is much more difficult to control and was not known
before Galilei.
Consider the following experiment. We put a bronze sphere on a horizontal
plane. We then give it a push. That is, we apply a force for a brief time interval,
giving it a certain initial velocity on the plane. We observe the sphere’s motion
and see that its velocity gradually decreases and finally stops. To have the sphere
moving at constant velocity we need to apply a force continuously. The
conclusion seems to be that, when not acted on by forces, a body stands still. If it
moves at constant velocity it is acted upon by a force proportional to its velocity.
We now know that the conclusion, thought to be true for centuries, is actually
false.
Galilei’s argument can be summarized as follows. The fact that, when we
apply a force to a body and then we cease to apply it, the body slows down and
finally stops is obviously true. But the cause is not the absence of acting forces.
On the contrary, the cause is the presence of forces that we do not apply, we do
not see, yet exist (they are called passive) and we are unable to avoid, like
friction and air drag,
Galilei could not prove his statement experimentally by eliminating all the
passive resistive forces. He observed however that, when launching a solid
polished sphere of brass or ivory on a horizontal guide, the distance travelled by
the sphere before coming to rest was longer and longer when the surfaces of the
guide and the spheres were smoother and smoother. Mentally going to the limit
of infinite smoothness, he concluded that in those conditions the sphere would
never stop, but would continue to move forever with the same velocity.
The conclusion is the law of inertia . In the words of Newton

Every body preservs in its state of rest, or of uniform motion in a right line,
unless it is compelled to change that state by impressed forces.

The law of inertia is not however valid in just any circumstance. Whether it
is valid or not depends on the reference frame . Up to now we have made
experiments in a reference fixed to earth. We now suppose that we want to build
a laboratory on a carriage moving on straight rails at constant velocity, relative
to earth. In our laboratory we have a smooth horizontal plane. We lay a bronze
sphere on the table and observe that, as expected, it remains still. However,
suddenly the sphere moves, accelerates and moves quickly forward, without any
visible force acting on it. What did happen? It happened that the carriage
suddenly started to slow down till coming to rest. Even if our laboratory is
closed with no window to look out, we know that the carriage decelerates
because we also experience a mysterious force pushing us forwards.
An observer on earth, namely in the frame we had been considering above,
easily interprets the phenomenon. The sphere is free to move horizontally, the
table being smooth. A force acted upon by brakes on its reels has slowed the
carriage down. This force, however, does not act on the sphere, because the
support plane is smooth. The resultant of the forces on the sphere is null. For the
law of inertia it will continue in its motion with constant velocity. This is relative
to the ground. But the observer on the carriage, which slows down relative to the
ground, sees the sphere accelerating to reach the velocity that the carriage had
before braking.
A reference frame in which the law of inertia is valid is called an inertial
frame . We shall see that inertial frames have a privileged role in mechanics, and
more generally in physics.
 More precisely, the law of inertia can be stated as: Reference frames do exist
in which every body not subject to force indefinitely remains in its state of rest or
uniform rectilinear motion.
One might think that the law of inertia is a consequence of our definition of
inertial frame, in other words that the argument is circular. But this is not true.
Indeed, we can give arbitrarily any definition we like, but we can never establish
by definition a law of nature, namely how she behaves. The existence of inertial
frames is a law of nature not a definition by men.
We further observe that we have considered inertial any reference stationary
on earth. The conclusion comes from the fact that, while doing experiments in
such laboratories, we never observe objects suddenly moving when no force acts
on them, nor do we feel as though we are being pushed in one direction or
another. However, the conclusion is valid only in a first approximation. Accurate
measurements show that frames that are stationary on earth are not exactly
inertial. This is due to the fact that earth moves around the sun and rotates on its
axis. We shall come back to that in Chap. 4. For the moment it will be enough to
know that stationary reference frames on earth are close enough to be inertial for
the vast majority of measurements carried out in laboratories and, on the other
hand, procedures exist to define inertial reference systems with all the requested
precision in case this is needed.

2.4 The Newton Laws of Motion

In the Principia , Newton begins by stating, as axioms induced from the
experiments, the three fundamental laws from which the description of all the
mechanical phenomena, both on earth and in the Universe can be deduced. The
first law is the law of inertia we already discussed. The causes of any change of
the state of rest or rectilinear uniform motion of a body are to be searched for in
the bodies around it. For example the racket that hits it changes the state of
motion of a tennis ball, the state of the compass needle is changed by the
presence of a magnet, etc. The same hit imparted with a racket to a tennis or
ping-pong ball produces different accelerations in the two bodies. By the term
inertial mass we mean the characteristic of a given object that makes it more or
less resistant to changing its state of motion under the action of a given force.
Galilei had already proven with his experiments that a body under the action of a
constant force, its weight or a component of its weight, moves with a constant
acceleration in the direction of the force.
Let us study the phenomenon quantitatively. We have already built springs
producing forces of different magnitudes. We have performed an analogous
procedure for mass. We have built a number of blocks of the same material
making them as equal as possible to each other. We can say that one block has
unit inertial mass, two blocks inertial mass equal to two, etc.
We have also prepared a horizontal plane, the function of which is to
equilibrate the weights of our blocks. In our experiments we shall put the blocks
in motion sliding on the plane and we want to reduce as much as possible the
friction forces between the plane and the blocks. We prepare the surface of the
plane as smooth as possible. We can also play the following trick. We can build
the blocks with a cavity inside and a series of holes between the cavity and the
lower face. We fill the cavity with dry ice (frozen CO2), which will sublimate
pushing CO2 gas through the holes. The thin layer of gas between the block and
plane surfaces reduces friction to negligible values.

1. We attach one of our springs to one block, we give it a certain deformation,

stretch or compression (Fig. 2.4a). We observe that the body moves with
constant acceleration, say a 0, in the direction of the force, as long as we keep
constant the force (i.e. the deformation)

Fig. 2.4 Simple experiments to study the relation between force, acceleration and inertial mass

2. We attach two springs (Fig. 2.4b) to the block and give them the same
deformation as in the first experiment. We observe the body moving again
with constant acceleration in the direction of the force. The acceleration is
twice as large, 2 a 0.
3. We fix two blocks one on top of the other and attach one spring to which we
give once more the same deformation. The acceleration is now one half as in
the first experiment, a 0/2 (Fig. 2.4c).

Continuing with similar experiments changing the force on a body or the

inertial mass, we come to the conclusion that its acceleration a is proportional to
the force F and inversely proportional to its inertial mass m i and we write

We can do better, because we have found that acceleration and force, which
are two vectors, have the same direction. The second law states that

(2.3)
This is the form that is more often expressed. However, Newton stated it as

A change of motion is proportional to the motive force impressed, and takes

place in the direction of the right line in which the force is impressed.

The quantity called by Newton “motion” is a fundamental vector quantity, p,

now called quantity of motion , or momentum (sometimes linear momentum ). It
is the velocity times the inertial mass

(2.4)
Two bodies of different masses can have the same quantity of motion if their
velocities are in the inverse proportion of the masses. The second Newton law is

(2.5)
In words, the rate of change of the momentum of a material point is equal to
the force acting on it. Considering that m i is a constant, and using Eq. (2.4) we
have

(2.6)
As for the law of inertia, the second law is not valid in every reference
frame. Recall the example of the sphere in a laboratory on a carriage that starts
suddenly to accelerate without any force being acting. Like the first law, the
second Newton law is valid only in inertial frames.
Equation (2.3) says that acceleration has the same direction as the relevant
force. This may appear to be obvious but it is not true in every circumstance.
The equation also says that the acceleration due to a given force acting on a
given body is independent of the velocity of the body. Experiments show that
both of these, while true at common experience velocities, are not so for
velocities close to the speed of light. In these conditions, called relativistic,
Eq. (2.3) fails. However, even in these high velocities regimes, Eq. (2.5) remains
valid, namely, as Newton stated, the force and the time derivative of momentum
are equal. What needs to be changed is the relation between momentum and
velocity.
We shall study relativistic mechanics in Chap. 6; we anticipate that in a
relativistic regime, the concept of inertial mass remains exactly the same. Mass
is a constant, independent of velocity, characteristic of the body. The concept of
momentum however must be made more general. Its expression is
(2.7)
where γ(υ) is a function of velocity, called the Lorentz factor , after Hendrik
Lorentz (1853–1928), one of the fathers of relativistic mechanics. Its value is
very close to 1 up to velocities close to that of light, c ≈ 3 × 108 m/s, but
increases very rapidly when υ approaches c.1
For comparison, the speed of the earth relative to the sun is about 3 × 104
m/s, 10−4 of the speed of light, the speeds of the stars relative to their galaxies,
including our sun, are an order of magnitude larger, but still 10−3 of the speed of
light. For the latter, the Lorentz factor differs from 1 only by 0.5 × 10−6.
A second limit of validity of the Newton laws is at very small dimensions.
Indeed, classical physics ceases to be valid and must be modified in quantum
physics, at atomic scales. These however are very small compared to the objects
of everyday experience, e.g., atomic radiuses are typically 30–300 pm.
The Newton law gives the acceleration once the forces are known.
Consequently, in the analysis of any motion we deal with the position vector, the
velocity, which is its first time derivative, and the acceleration, its second time
derivative. We do not need higher derivatives. For these reasons we did not go
beyond the second derivative of the position vector when we studied kinematics.
We recall on purpose that to know the motion of a particle we need to know not
only the acting forces, but also the initial position and velocity.
Let us now look at another aspect. The second law can be used in three main
ways:
1. If we know the inertial mass of a body and all the forces acting on it, and the
initial conditions, we can calculate its motion

2. If we know the motion of a body and its inertial mass, we can infer the forces
acting on it.
Distinguishing the two points of view is not as trivial as it may look. The
first point of view is deductive. The laws of mechanics are used to calculate
the motion of bodies in all possible circumstances. In this way physicists and
engineers design mechanical devices and engines. The second point of view
is inductive and is the point of view taken to make progress in physics. The
challenge of the physics research is to understand from the study of motion
the fundamental nature of the forces that cause it. This is the way followed by
Newton to discover universal gravitation from study of the motions of
heavenly bodies. This is the way in which Ernest Rutherford (1871–1937)
discovered the atomic nucleus in 1911 when studying the scattering of
energetic alpha particle by a thin gold sheet. This is the way followed today
to study the properties of atomic nuclei and elementary particles.
We can state that the success of the Newton law is just as follows. It
substantially tells us: if you see a body that does not move in a uniform
rectilinear motion, a force should act. Search for it and search for the physical
agent to which it is due. You will find a force, the mathematical expression of
which will be simple and, as a consequence, you will be able to lay down a
simple theory. From this point of view the Newton law is a research program.
We shall see in Chap. 3 that, indeed, the various forces of nature have simple
expressions in terms of the co-ordinates and characteristics of the system. The
program is successful.

3. A third possibility is that, if we know both forces and motion we can deduce
the inertial mass of the body. To know the mass of the proton for example,
we can measure how its momentum and energy vary under the action of a
known force.

The law of composition of forces . If more than one force act at the same time
on the material point we are discussing, their effect is the same as if only one
force were acting, equal to the resultant of those forces. Consider for example
that two forces are applied as in Fig. 2.5. The first spring exerts the force F 1 in
the x direction. When acting alone it produces the acceleration F 1/m i along x.
The second spring exerts the force F 2 in the y direction. When acting alone it
produces the acceleration F 2/m i along y. To know what happens if the two
forces act contemporarily is something that cannot be found by logic, rather it
has to be found experimentally. Indeed, what experiments show is that the
acceleration is just what one calculates assuming that only one force were acting,
equal to the resultant F of F 1 and F 2. In other words, the observed acceleration
is a = F/m i .

Fig. 2.5 Two forces acting at the same time

The third Newton law is the law of action-reaction .

If a body exerts a force (an action) on a second body, the second always
exerts on the first a force (a reaction) that is equal and opposite on the same line
of action.
Given its importance, we reproduce how it is stated, in an equivalent manner,
by Newton .

To every action there is always opposed an equal reaction: or, the mutual
actions of two bodies upon each other are always equal, and directed to
contrary parts.

Newton gives then a few examples.

Whatever draws or presses another is as much drawn or pressed by that

other. If you press a stone with your finger, the finger is also pressed by the
stone. If a horse draws a stone tied to a rope, the horse (if I may so say) will
be equally drawn back towards the stone; for the distended rope, by the
 same endeavor to relax or unbend itself will draw the horse as much
towards the stone as it does the stone towards the horse and will obstruct
the progress of the one as much as it advances that of the other.

We notice that, differently from the first two, the third law deals with two,
rather than one, bodies. It tells us that isolated forces (actions) do not exist, only
interactions do exist.
Pay attention to the fact that action and reactions are applied in different
points, one on one body, the other on the other body. If we push a stone with a
finger, the action of the finger is applied in a point of the stone; the reaction of
the stone is on the tip of our finger. The force exerted by the horse drawing the
stone is exerted on the stone through the rope, the reaction acts, again through
the rope, in the point of the horse at the end of the rope. Every object whether it
is falling or laying on a support, weighs, meaning that the weight force is applied
on it. Weight is the force with which the earth attracts all bodies. As a reaction,
each body attracts the earth with an equal and opposite force. The reaction is
applied to a point of the earth, its center.
The action-reaction principle, as all physical laws, must be experimentally
verified. Direct verifications are based on the fact that in a collision between two
bodies the total quantity of motion, namely the vector sum of the two, is
conserved, meaning that its values before and after the collision are equal (while
each of the two vary).
The vectors we have met so far, position vector, velocity and acceleration
depend, as we have seen, on the reference frame. On the contrary, force does
not.

2.5 Weight
We know from every day experience that all the bodies on earth are subject to a
force, vertically directed downwards, called the weight . We can measure the
weight of a body, for example, attaching it to a dynamometer vertically
positioned and reading on its scale the position of the pointer, namely the stretch
of the spring. If we repeat the measurement in different points of our laboratory
we find that it does not vary. However, if we repeat the measurement at much
larger distances, for example at the Equator and at 45° latitude, or at different
altitudes, for example at the sea level and at 2000 m altitude, we notice small
differences (of the order of a few per mille) between them. As we shall discuss
in Sect. 5.7, these small variations are due to the rotation of the Erath. Apart
from these small corrections, the weight is the gravitational attraction exerted by
the earth on the body. This is universal; it is the same force with which the earth
attracts the moon. We shall discuss this fundamental force in Chap. 4. We
anticipate that the gravitational attraction decreases as the reciprocal of the
distance squared. This is one of the reasons (the other is the rotation motion of
earth) why the weight of an object is a bit smaller on a mountain than at the sea
level.
Different objects, in the same place, may have different weights. This means
that the force with which earth attracts a body depends on a characteristic of the
body. We state that the gravitational force on a body is proportional to its
gravitational mass, which we denote with m g . This is similar to the electric
attraction. A charged body A at a certain distance from another body that is also
charged, is subject to an electrical force. If in the place of A we put a body B
with twice the charge, the force on it is double. Hence, the electric force on a
body is proportional to its electric charge. In a similar way two massive bodies,
for example two spheres, at a certain distance attract with the gravitational force
that is proportional to the gravitational mass of each of them. This force, if
between two objects of every day life is quite small, but can be measured with
very delicate experiments, as we shall see in Sect. 4.7, but is large between
Heavenly bodies. Considering that the gravitational mass is for the gravitational
force the analogous of the electric charge for the electric forces, we might call it
gravitational charge, but we shall soon see the reason why we call it mass.
The weight force F W acting on a body of gravitational mass m g is then
(2.8)
The vector quantity g does depend on the location, but in a given site it is
equal for all bodies. If r is the position vector, the vector g(r) is the gravitational
force at r per unit gravitational mass. It is called gravity acceleration. We shall
see soon the reason for the name. We notice that the gravitational mass being a
characteristic of a body is the same in any point, differently from its weight. If
we measure the weights of two bodies in different points on the earth we find
that each of them varies a bit, as already mentioned, but the ratio of the two
remains rigorously equal. Even if we should do this experiment on the moon.
Operationally, the gravitational mass is the physical quantity measured by a
balance. A balance, see Fig. 2.6, consists of a lever with pivot in O and two pans,
which we shall consider, to make it simple, exactly at the same distance on the
two sides of O. The balance compares the weights of the two objects on its pans.
If they are equal the balance is in equilibrium. We have seen that, by definition,
the weights of different objects in the same place are proportional to their
gravitational mass. We can then the state that two objects have the same
gravitational mass when, put on the pans of the balance, they are in equilibrium.

Fig. 2.6 Comparing the weights of two equal masses

We now need a body having unit mass by definition. We put it on a pan.

Another body has gravitational mass equal to one when, put on the other pan it is
in equilibrium. A body has gravitational mass equal to 2, if put on a pan is in
equilibrium with two of the unit masses on the other, etc.
Gravitational mass and inertial mass are two different properties of every
body. The former is a measure of the strength of the gravitational attraction to
which it is subject, the latter of how difficult it is to modify its quantity of
motion. However, we know from every day experience, that heavier bodies are
also more difficult to accelerate because they are more inert. To search for a
mathematical relation, suppose to observe the free fall of two different bodies.
Their inertial masses are m 1i and m 2i and their gravitational masses m 1g and m
2g . The weight of the first is, F 1w = m 1g g, the weight of the second F 2w = m 2g
g. Calling a 1 and a 2 the two accelerations, we have:

which can be written as

(2.9)
We see that the free fall accelerations of different bodies in the same place
are proportional to the ratios of their gravitational and inertial mass.
Consequently, if this ratio is equal for all the bodies, light or heavy, all of them
fall with the same acceleration. This fundamental property was experimentally
shown to be true by G. Galilei .
It is often told that Galilei dropped contemporarily two balls, one made of
lead, one of wood, from the Pisa tower and that he observed them reaching
ground at the same instant, showing in this way that they fall with the same
acceleration. The experiment was absolutely success and spectacularly carried
out in 1971 by the NASA Apollo 15 astronaut D. Scott dropping a hammer and a
feather on the moon. As a matter of fact Galilei never mentions having made his
fundamental experiments in such a way. He new very well that it could not
work, both for the perturbing effects of the atmosphere and due to the smallness
of the fall times, a fact that did not allow him precise measurements. His very
precise experiments were done with reduced, to say so, weight forces, with
spheres on inclined planes and with pendulums. We shall discuss this in
Sect. 2.9.
We can conclude that the free fall accelerations of all bodies in a given place
are equal, action of the atmosphere apart. The ratio between gravitational and
inertial mass is a universal constant, the same for all bodies. The value of the
constant is arbitrary, because depends on the choice of the two units. Clearly, the
most convenient choice is to have the ratio equal to one. With this choice
gravitational and inertial mass are not only proportional, they are equal. The unit
of both is the kilogram. From now on we shall indicate with the same symbol,
for example m without any subscript, both quantities.

2.6 Examples
In this section we study a number of examples of application of the Newton
laws. A good way to proceed is the following.
The first step is to identify all the bodies present in the problem. Next we
identify for each of them all the forces acting on it. To do that it is convenient to
wrap it, ideally in an envelope, in order to identify all the forces acting on the
body from its exterior. To this aim it is often useful to draw each object
separately, in its ideal envelope, and the acting forces and write down for each of
them its type and its agent (for example: weight due to earth, normal force due to
the constraint, friction due to the supporting surface). If the problem contains
more than one body, we must identify the action and reaction pairs, and the
bodies on which they act. Once all the forces are identified we must calculate the
resultants on each of the bodies. To do that we choose a reference frame. The
choice should be guided by any symmetry the problem might have. We must
then calculate the Cartesian components of the resultant by summing the
correspondent components of all the forces. The components divided by the
mass of the body are the three components of the acceleration of the body. From
the acceleration we find the law of motion with the procedures we studied in
Sects. 1.15 and 1.16.

Example E 2.1.
Place a block on a horizontal frictionless surface horizontally drawn by a rope.
 Frictionless means a physical surface that does not exert forces parallel to it.
It is an idealization. Friction always exists, but we can reduce it, for example
with the dry ice trick of Sect. 2.4. We attach a rope to the block and draw it
horizontally with the force F r . The situation is shown in Fig. 2.7.

Fig. 2.7 N normal constraint force, F r force exerted by the rope, F w weight, due to earth

Knowing F r and the mass m of the block we want to know its motion,
considering it as a point. We draw the body in its ideal envelope. We identify the
forces acting through the surface: (1) the weight of the block F w , due to earth,
vertically directed downwards, (2) the constraint force exerted by the plane. As
we have assumed it to be frictionless the force is normal to the surface, upwards
and we call it N, (3) the force (tension) exerted by the rope, F r . We have drawn
all of that in Fig. 2.7b. As we are considering the block as a material point, all
the forces are applied in the same point. One of the forces, N, is not given. This
is always the case of constraint forces. The body cannot penetrate the support
plane because the molecules of the body and the plane repel each other. We
know that the body has no vertical acceleration. We infer that the support
develops the force that is exactly what is needed to keep it steady. We will find it
by solving the equations.
All the forces of the problem lay in the same vertical plane. It is then
convenient to choose a reference frame with one axis, say z, vertical upwards
and a second one, say x, horizontal to the right in the figure. We do not need the
third axis because there are neither forces nor motion in that direction. We now
write the second Newton law and its two components

We conclude that the normal force exerted by the support plane has
magnitude equal to the weight. Both forces are vertical and have opposite
direction; hence their resultant is zero. The resultant of the forces is the tension
of the rope, which causes a uniformly accelerated motion in the x direction.
Example E 2.2
A block moving on a horizontal frictionless surface drawn by a rope at an angle
with the horizontal.
The situation is the same as in the previous example, but for the rope now
pulling at an angle θ with the horizontal (see Fig. 2.8a). However, we still
assume that the motion is on the plane, namely that there is no vertical
acceleration. The forces are the same, but F r has different components. We have

Fig. 2.8 N normal constraint force, F r force exerted by the rope, F w weight, due to earth

The equation for the z components gives again the normal constraint force,
N = F w –F r sin θ. If θ > 0 as in the figure, N is smaller than in the previous
example because the rope helps in sustaining the block, the opposite if θ < 0.
The second equation gives horizontal acceleration.
Notice that a physical limitation of this analysis exists. The normal force
cannot be negative, because the support plane cannot attract the body (there is no
glue). Hence, if F r sin θ > F w , the assumed conditions cannot be satisfied.
Clearly, in this situation the block is lifted up and its acceleration has a vertical
component.

Example E 2.3
Block on an inclined frictionless surface.
There are two forces acting on the body (Fig. 2.9), the weight F w and the
constraint force N perpendicular to the support plane, which is now inclined. The
convenient choice of the axes is to take z perpendicular to the plane and x along
the plane, downwards. Clearly, the body will slide accelerating downwards,
namely in the x direction we have chosen.
Fig. 2.9 A block on a frictionless incline

The Newton equation and its components are

The z component gives us the normal force N = F w cos α. The x component

gives the acceleration (a = a x ). Recalling that F w = mg, we have that the notion
on an inclined frictionless plane is uniformly accelerated with acceleration
(2.10)
We see that the motion on an incline is completely similar to the motion of
free fall, as long as we can neglect the resistive forces. The difference is that the
acceleration is smaller on the incline by a factor sin α. We can reduce
acceleration by reducing the slope of the plane. If the motion starts from rest
from the origin, the law of motion is obtained by integrating twice Eq. (2.10),
obtaining

(2.11)
In words: the distances travelled are proportional to the squares of the times
taken to travel them.
The incline allows us to slow down the free fall motion and to study its laws
over longer times, which can be measured with better precision.
As mentioned in Sect. 2.5 this is one of the great discoveries of Galilei . He
did not have a modern chronometer, but invented an ingenious water
chronometer , with which he was able to measure the times of the motion, a few
seconds long, with a precision better than 0.1 s. He describes his experiments in
the book “Dialogues and mathematical demonstrations concerning two new
sciences” or “Two new sciences ” published in 1638. He writes:

A piece of wooden molding or scantling, about 12 cubits long, half a cubit

wide, and three finger-breadths thick, was taken; on its edge was cut a
channel a little more than one finger in breadth; having made this groove
 very straight, smooth, and polished, and having lined it with parchment,
also as smooth and polished as possible, we rolled along it a hard, smooth,
and very round bronze ball. Having placed this board in a sloping position,
by lifting one end some one or two cubits above the other, we rolled the
ball, as I was just saying, along the channel, noting, in a manner presently
to be described, the time required to make the descent. We repeated this
experiment more than once in order to measure the time with an accuracy
such that the deviation between two observations never exceeded one-tenth
of a pulse-beat. Having performed this operation and having assured
ourselves of its reliability, we now rolled the ball only one-quarter the
length of the channel; and having measured the time of its descent, we
found it precisely one-half of the former. Next we tried other distances,
comparing the time for the whole length with that for the half, or with that
for two-thirds, or three-fourths, or indeed for any fraction; in such
experiments, repeated a full hundred times, we always found that the spaces
traversed were to each other as the squares of the times, and this was true
for all inclinations of the plane, i.e., of the channel, along which we rolled
the ball. We also observed that the times of descent, for various inclinations
of the plane, bore to one another precisely that ratio which, as we shall see
later, the Author had predicted and demonstrated for them.
For the measurement of time, we employed a large vessel of water
placed in an elevated position; to the bottom of this vessel was soldered a
pipe of small diameter giving a thin jet of water, which we collected in a
small glass during the time of each descent, whether for the whole length of
the channel or for a part of its length; the water thus collected was weighed,
after each descent, on a very accurate balance; the differences and ratios of
these weights gave us the differences and ratios of the times, and this with
such accuracy that although the operation was repeated many, many times,
there was no appreciable discrepancy in the results.

Example E 2.4
A block at rest in a lift.
A block of mass m lies in a lift on a horizontal pan of a balance, one of those,
for example, that are used to weigh people. What is the apparent weight of the
block when the lift accelerates up or down?
As usual we imagine the block in an ideal envelope (Fig. 2.10). Two forces
act on it, the weight F w vertical down, and the normal constraint of the pan N
upwards. The balance measures the reaction to N, namely the force on it, which
is –N. Hence, N is the apparent weight of the block.
Fig. 2.10 A block in an accelerating lift

If the lift moves with acceleration a upward, the unknown N is given by the
Newton law Hence, the apparent weight is ,
which is larger than the true weight. If the lift accelerates downwards, the
apparent weight is , smaller than the real one. Notice that if the
acceleration downwards is g the apparent weight is null. Indeed, the block is
falling with the same acceleration of the lift.
If the lift moves uniformly both upwards and downwards the apparent
weight is equal to the real one, as if it were standing. We feel an increase of our
weight either if the lift accelerates going up or if it decelerates going down. In
both cases its acceleration is upwards. Similarly we feel a decrease of our weight
when the lift slows down going up or accelerates going down.
Tension of the ropes and wires. In some of the examples we made we have
used a stretched rope or wire to apply a force in a point of a body. This force is
equal to the tension of the wire. We generally assume the wire to be inextensible,
meaning that its length does not vary whichever the tension may be, and
perfectly flexible, meaning that the tension is always parallel to the wire, and of
negligible mass. Once more, these are idealizations.
Let us clarify the concept of tension. Consider a wire, stretched and steady as
in Fig. 2.11a. We mentally isolate a small segment, enlarged in Fig. 2.11b. Two
forces act on the segment (neglecting the weight), applied to its extremes and
due to the contiguous elements of the wire. These are the tension forces. As the
wire is at rest, the two forces are equal and opposite. Consequently, the tension
is the same in every section of the wire.
Fig. 2.11 a The tension forces on a wire and, b on a segment

Each of the extremes of the wire is not in contact with another element. As it
does not accelerate, a force must act on it from outside equal in magnitude to the
tension and directed outwards, as in Fig, 2.11a. The forces on the extremes are
equal and opposite and have the magnitude of the tension.
Consider now the case in which the wire moves. As an example, suppose
that one extreme is fixed to a block of mass M lying on a horizontal plane of
negligible friction. We draw the block applying to the free extreme of the wire a
force F 1 obtaining an acceleration a, as shown in Fig, 2.12a. We want to
understand under which conditions we really can neglect the mass of the wire.
To do that, let us start assuming the mass of the wire to be m.

Fig. 2.12 a Accelerated motion of a block drawn by a rope, b N normal constraint force, M g weight due
to earth, F 2 force due to the wire, c T 2 force on the wire due to the block, F 1 force pulling the wire

We are now dealing with two bodies, the block and the wire. We ideally
isolate each of them and draw the force diagrams on each of them, in Fig. 2.12b,
c.
We next identify the action reaction pairs. There is one such pair, consisting
of the forces F 2 applied to the block and T 2 applied to the left extreme of the
wire. They are equal and opposite. The force F 1 applied to the right extreme of
the wire is its tension and we can call it T 1. The Newton equations for the two
bodies are

hence, for the magnitudes, and . We see that the

tensions at the two extremes are different. Indeed T 1 > T 2 because T 1 must
accelerate wire and block, T 2 only the block. Let us consider their ratio

which becomes unity for m/M → 0. We can then state that the tensions at the
extremes can be considered equal if the mass of the wire is negligible compared
to the mass of the block. When we speak of massless ropes or wires we mean of
negligible mass compared to the masses of the other objects.
Notice that we can arrange a stretched wire, or rope, to have forces at its
extremes of equal magnitude but different directions, by using pulleys. We did
so already discussing the Varignon experiment (Fig. 2.3). Notice that in these
cases, if the motion is accelerated, the magnitudes of the tensions at the extremes
can be considered equal only if also the mass of the pulley is negligible and if it
can rotate with negligible friction on the pivot (Fig. 2.13).

Fig. 2.13 With a pulley, the direction of the force exerted by a wire can be changed

Example E 2.5
Two blocks linked by a rope of negligible mass.
Figure 2.14a shows two blocks of masses m 1 and m 2 lying on a horizontal
frictionless plane, connected by an inextensible wire of negligible mass. To the
second block, at the right, a horizontal force F is applied. The motion is on the
support plane. To know it, we do not need to analyze the vertical forces, which
have zero resultants (Fig. 2.14a, b, c, d).

Fig. 2.14 a Two blocks connected by a wire, b force on m 1, c forces on the wire, d forces on m 2

We start by considering the whole system, thinking of it as a unique ideal

envelope. The only force acting on this surface is F. Hence we have
. which gives an acceleration a equal for the two bodies.
We now isolate each of the bodies. The block on the left (Fig. 2.14b) is
attached to an extreme of the wire. This exerts on the block the horizontal force
F 1. For the action-reaction law the block exerts on the extreme of the wire an
equal and opposite force, which is the tension of the wire at that extreme (F 1 = –
T 1). Two other forces act on the block, the external force F and the force F 2
due to the right extreme of the wire (Fig. 2.14d). Again, for the action-reaction
law the block exerts, on the right extreme of the wire, a force equal and opposite
to F 2 that is the tension T 2 at that extreme (F 2 = – T 2). As we have discussed
above, the magnitude of the tension is the same in all points of the wire. Taking
into account the directions we have T 1 = – T 2 (Fig. 2.14c). Calling T the
magnitude of the tension we can write the Newton equations as
.
The sum of the two equations gives the acceleration of the system a = F/(m
1 + m 2). If we want the value for tension, we substitute a in the first equation
obtaining

We see, in particular, that T < F, namely the tension is smaller than the force
with which we pull.

2.7 Curvilinear Motion

In the previous section we have studied a few examples in which the forces were
known, a part of the constraint ones, and the motion that had to be found. In this
action we shall consider the inverse problem, namely, the motion of a material
point being known, find the resultant of the forces. The singular forces, in case
more than one is present, cannot be found, because systems of forces with the
same resultant produce the same motion in the case of material points.
Circular uniform motion
Consider the motion of a material point P with mass m constrained to move
on a circumference of radius R. Suppose the motion to be uniform, namely the
magnitude of its speed υ to be constant, as in Fig. 2.15. The motion is however
accelerated, because the direction of the velocity varies. As we already found,
the acceleration has a constant magnitude (Eq. 1.57) and is in every
point directed to the center (centripetal acceleration). This acceleration must be
given by a force of magnitude
Fig. 2.15 Circular motion, a uniform, b increasing velocity, c decreasing velocity

(2.12)
The corresponding force has the same direction as the acceleration and is
called centripetal force . The adjective “centripetal, from the Latin “petere” for
“point towards”, recalls only its direction but does not specify at all its nature. It
may be the tension of a wire, the normal force of a circular guide, the
gravitational force of the earth on the moon, etc. We shall discuss a few
examples in Sect. 3.4.
Variable speed motion .
If the magnitude of the velocity of a particle moving on a circle varies, its
acceleration has two components. One component, a n , is perpendicular to the
trajectory, or, the latter being circular, directed to the center. It is again the
variation of the direction of the velocity, namely the just discussed centripetal
acceleration of value υ 2/R where υ, we must now specify, is the instantaneous
velocity. The second component, a t , is in the direction of the motion, i.e.
tangent to the trajectory and expresses the variation in time of the magnitude of
the velocity. We have

(2.13)
The acceleration vector, and the force, is directed at an angle with the radius
that is forward if the velocity is increasing (Fig. 2.15b), backward if it is
decreasing (Fig. 2.15c). The magnitude of the force is
As an example, consider a block lying on the platform of a merry go round,
which is initially still. When the platform starts moving, gradually increasing its
angular velocity, the acceleration of the block has two components, one
centripetal and one tangential. The corresponding force, equal to the mass of the
ball times this acceleration, is given by the friction on the platform. If the latter
is not enough, the block slides towards the periphery of the platform.
 As a second example consider the launch of the hammer. The athlete acting
on the rope he holds in his hands puts the hammer in rotation with increasing
speed. The force on the hammer must be adequate to keep it on a circular orbit
(component mυ 2/R towards the center) and makes its speed increase (a
component in the direction of the motion). The rope must then be directed
forward, as in Fig. 2.15b
General plane motion .
We consider now a material point of mass m moving on a plane trajectory of
arbitrary shape with velocity not necessarily constant in magnitude. We have
already studied the kinematics of the problem in Sect. 1.14. Even in this case, the
acceleration has two components, a tangential and a normal one, as in Eq. (1.62).
They are given by Eq. (2.13).
The only difference from the circular case is that now R is the local curvature
radius, which is not fixed but varies along the trajectory. The second Newton
law tells us that the resultant of the forces acting on the point must be its
acceleration times its mass.
If we know only the trajectory, but nothing of the velocity, we can still say
that in every point of the trajectory in which the curvature is not zero, the
resultant of the forces must be directed on the side of the curvature center,
pointing forward from it (Fig. 2.16a) or backwards (Fig. 2.16b) depending on
whether the motion is accelerated or delayed respectively.

Fig. 2.16 General plane motion. a Increasing speed, b decreasing speed

2.8 Angular Momentum and Moment of a Force

Consider a material point P moving in an inertial frame as shown in Fig. 2.17.
Let p = m v be its momentum and r its position vector. Consider a generic point
Ω, which may be at rest or moving relative to the frame. We shall now introduce
the concepts of angular momentum and moment of a force about the pole Ω.
Fig. 2.17 The vectors relevant for angular momentum

We have already defined the moment of a bound vector in Sect. 1.8. The
angular momentum is the moment of the linear momentum, considering it, for
this purpose, as applied to the material point, as shown in Fig. 2.17.
Hence, the angular momentum of the point P about the pole Ω is the vector
product of the vector from Ω to P and its quantity of motion (or momentum).
(2.14)
Consider the force F applied to P. The moment of the force about the pole Ω
is the vector product of the vector from Ω to P and F
(2.15)
Remember that the order of the factors matters in cross products. Notice also
that the moments change if the reference frame changes.
Let us now see how the angular momentum changes in time. For that, we
take the time derivative of Eq. (2.14) using the rule of the derivative of products,
paying attention to the order of the factors

(2.16)
To find the derivative of the vector we notice that it is the difference of
two vectors, both varying with time, . Deriving we have.

The meaning of this expression is clear: the derivative of a vector joining

two moving points is the relative velocity of those points. We substitute this
expression in Eq. (2.16) and also notice that the derivative of the momentum is
equal to the resultant F of the forces acting on P, because the frame is inertial.
We get

The first term in the second member is zero, being the cross product of two
parallel vectors; the last term is the moment of the resultant about the pole τ Ω.
In conclusion

(2.17)
This is a very important equation that we shall use often in the following. It
becomes particularly simple if we choose a stationary pole in the reference
frame. The equation becomes

(2.18)
In words the equation is called the angular momentum theorem for a material
point: the time derivative of the angular momentum of a material point about a
pole fixed in an inertial reference frame is equal to the moment of the resultant
of the forces acting on it about the same pole.
Notice that if the body is extended, as we shall discuss in the following
chapter, the different forces acting on it, say f 1, f 2,…, may be applied in
different points and the moment of their resultant F = f 1 + f 2 + ···,
is in general different from the vector sum of their moments. In the case under
study however, all the forces are applied in P and

The resultant of the moments is equal to the moment of the resultant of the
forces. We stress that this is true only if all the forces are applied at the same
point.

2.9 The Simple Pendulum

The pendulum is a material point constrained to move on an arc of a
circumference. It can be simply made by fixing a thin wire to a small sphere on
an extreme and to a fixed point on the other, which we call Ω. The length l of the
wire, or better the distance between the fixed point and the center of the sphere,
is called the length of the pendulum . If we take the pendulum away from its
equilibrium position O and abandon it with zero velocity, the body moves
towards O under the action of two forces, the weight, directed vertically down,
and the tension of the wire (T), directed as the wire. The acceleration has the
direction of the resultant of these two forces. Consequently it is always in the
plane defined by the wire and the vertical. If the initial velocity is zero, the
motion is on the plane. As the distance from Ω is kept fixed by the wire, which
we assume inextensible, the trajectory is an arc of a circle of radius l.
As shown in Fig. 2.18, we take a reference system with the origin O in the
rest position of the pendulum, the y-axis vertical upwards, the x-axis horizontally
in the plane of motion and z such as to complete the triplet. The z-axis is normal
to the figure towards the observer. We call θ the angle between the wire and the
vertical, taking it positive if seen anticlockwise by the observer.

Fig. 2.18 The simple pendulum

Historically, as we have already mentioned, the study of the motion of

pendulums, with their periodic motion, made a fundamental contribution to the
development of mechanics. Galilei discovered two important properties. The
first one is the isochronism of small oscillations; if the amplitude is not too large
(we shall be more precise in the following), the oscillation period is independent
of the amplitude. This property allowed building of precise clocks. The second
property is even more important; the oscillation periods of pendulums of the
same lengths and different masses are identical. This proves, as we shall now
see, that gravitational mass and inertial mass are equal. The property was later
called equivalence principle and is at the basis of general relativity .
In our demonstration, we start by assuming that the two masses might be
different. We call m i the inertial mass of the pendulum, namely the
proportionality constant between acceleration and force, and m g its gravitational
mass, the constant that appears in the weight, which is then m g g.
The tension is a constraint force, due to the wire, which we assume to be
perfectly flexible and inextensible. The constraint develops a force, in general
unknown a priori, automatically adjusted to make the motion happen, in our
case, at a fixed distance from Ω. We do not know the intensity of the wire
tension T, but we know its direction, which is along the wire.
In our study of the motion we shall use the angular momentum theorem. We
choose the pole in the suspension point Ω, for reasons that will become clear
soon. We use Eq. (2.18) with

We now see the reason for our choice of pole. The first term is always zero,
being the vector product of two parallel vectors. Consequently we do not need to
know the intensity of the tension. We have
(2.19)
The angular momentum about the same pole is
(2.20)
where the mass is the inertial one. Equation (2.18) gives

(2.21)
All the vectors in these equations, in any position of the pendulum, belong to
the plane xy. Both vector products are consequently in z direction. The equation
has only the z component. The z component of is . The
velocity is always perpendicular to . As a consequence the z component of
is simply , where . So, we have

And finally we can write

(2.22)
This is a differential equation, whose unknown is a function of time θ(t).
Once it is solved, we know the motion of the pendulum, because if we know θ,
we know its position. Equation (2.22) cannot be solved analytically. However, if
the oscillations are “small”, we can approximate the sine with its argument and
the equation becomes
 (2.23)
This is a well-known differential linear equation with constant coefficients,
which we shall meet several times. We leave its study to calculus courses and
directly give the general solution, which is
(2.24)
where

(2.25)
is called proper angular frequency . As one sees, it depends only on the
characteristics of the pendulum, including its weight.
The reader can easily verify, with two derivatives, that this expression indeed
satisfies Eq. (2.23), for whatever values of the constants θ 0 and ϕ. These
constants do not depend on the characteristics of the pendulum but on how the
motion has started. They should be found in each case on the basis of two initial
conditions. We can use the position and velocity at the starting time that we shall
take as t = 0. We immediately see that

The initial velocity being zero, the second equation gives ϕ = 0 (θ 0 = 0 is

also a solution of Eq. (2.24) but identically null). The first condition says that θ 0
is just the initial angle, the angle at which we have let the pendulum go. In
conclusion the motion of the pendulum is described by the equation
(2.26)
The motion is periodic , meaning that, for any instant of time t we can
consider, both the position and the velocity become the same after a certain time
interval T, called the period , namely at the instant t + T. From Eq. (2.26) we
immediately see that the period is

(2.27)
where we used Eq. (2.25).
The motion is represented in Fig. 2.19. This is the most common periodic
motion in Nature. It is called harmonic motion . In the next chapter we shall
study it in depth.
Fig. 2.19 Angular harmonic motion

We now make an important observation on the expressions of proper angular

frequency and period. Angular frequency and period depend on the length of the
pendulum, but not on the oscillation amplitude : two pendulums of the same
length (and in the same place, hence at the same g) are isochronous . On the
other hand, angular frequency and period depend on the ratio between
gravitational and inertial masses . If this ratio is the same for all bodies,
independently of the substance they are made of and of their position, that ratio
is a constant and angular frequency will be independent of the mass of the
pendulum. If we want to experimentally test if gravitational and inertial masses
are proportional or not, we can test whether pendulums of the same length, and
different masses or made of different substances, do oscillate or not with the
same period.
Galilei noticed that this method is much more accurate than others he knew.
In principle, one could think to drop two spheres, e.g. one of wood and one of
lead, from the top of a tower and check if they the reach the ground
simultaneously. However, Galilei never mentions having done such an
experiment, from the leaning tower of Pisa. This is a legend without any
historical support. Indeed, Galilei observed, and wrote, that that method is not
accurate enough, because is too fast and at high speeds the air resistance
noticeably perturbs observations. Galilei used inclined slopes, as we have
discussed, to slow down the motion, reduce the air drag, and increase the relative
measurement accuracy due to the longer times to be measured. The use of
pendulums allows even better accuracy. He used pairs of pendulums made of
different materials and of exactly the same lengths, took them out of equilibrium
and let them go at the same time. He found that they keep oscillating in phase for
hundreds of periods. The air drag did act more effectively on the lighter
pendulum gradually reducing its amplitude more than that of the heavier one.
However this did not matter because the period is independent of amplitude.
G. Galilei describes accurately his progress toward a more precise
experiment, gradually eliminating the spurious effects and the sources of errors
in the “Dialogs concerning two new sciences ” (1638) (translation from Italian
by Henry Crew and Alfonso de Salvio). He established the proportionality of
inertial and gravitational mass with an uncertainty of 2–3 × 10−3.

The experiment made to ascertain whether two bodies, differing greatly in

weight will fall from a given height with the same speed, offers some
difficulty; because, if the height is considerable, the retarding effect of the
medium, … will be greater in the case of the small momentum of the very
light body than in the case of the great force of the heavy body; so that, in a
long distance, the light body will be left behind; if the height be small, one
may well doubt whether there is any difference; and if there be a difference
it will be inappreciable. It occurred to me therefore to repeat many times the
fall through a small height in such a way that I might accumulate all those
small intervals of time that elapse between the arrival of the heavy and light
bodies respectively at their common terminus, so that this sum makes an
interval of time which is not only observable, but easily observable. In
order to employ the slowest speeds possible and thus reduce the change
which the resisting medium produces upon the simple effect of gravity it
occurred to me to allow the bodies to fall along a plane slightly inclined to
the horizontal. For in such a plane, just as well as in a vertical plane, one
may discover how bodies of different weight behave: and besides this, I
also wished to rid myself of the resistance which might arise from contact
of the moving body with the aforesaid inclined plane. Accordingly I took
two balls, one of lead and one of cork, the former more than a hundred
times heavier than the latter, and suspended them by means of two equal
fine threads, each four or five cubits long. Pulling each ball aside from the
perpendicular, I let them go at the same instant, and they, falling along the
circumferences of circles having these equal strings for semi-diameters,
passed beyond the perpendicular and returned along the same path. This
free vibration repeated a hundred times showed clearly that the heavy body
maintains so nearly the period of the light body that neither in a hundred
swings nor even in a thousand will the former anticipate the latter by as
much as a single moment, so perfectly do they keep step. We can also
observe the effect of the medium which, by the resistance which it offers to
motion, diminishes the vibration of the cork more than that of the lead, but
without altering the frequency of either.

In conclusion, Galilei experimentally demonstrated the equality of inertial

and gravitational masses with an accuracy of about one per mille, namely that
 (2.28)
Newton repeated this later on the Galilei experiments. He writes in the
“Principia ”:

It has been, now of a long time, observed by others, that all sorts of heavy
bodies (allowance being made for the inequality of retardation which they
suffer from a small power of resistance in the air) descend to the earth from
equal heights in equal times; and that equality of times we may distinguish
to a great accuracy, by the help of pendulums. I tried the thing in gold,
silver, lead, glass, sand, common salt, wood, water, and wheat. I provided
two wooden boxes, round and equal: I filled the one with wood, and
suspended an equal weight of gold (as exactly as I could) in the center of
oscillation of the other.

He concluded that:

By these experiments, in bodies of the same weight, I could manifestly have

discovered a difference of matter (i.e. inertial mass) less than the
thousandth part of the whole, had any such been.

Hence Newton confirmed what Galilei had discovered with a similar

precision of 1 × 10−3. After having found an expression of the gravitational
force, Newton did also a check of the equivalence principle , on a solar system
scale. He did that, in particular, on the system of Jupiter and its satellites. We
shall see his argument in Sect. 4.4. Here we just say that the precision was, once
more, of one per mille.
Having established the proportionality of the two types of mass, we can
make them equal by choosing their units. With this choice Eqs. (2.25) and (2.27)
become

(2.29)
To have a feeling of the orders of magnitude, we can easily calculate that a
1-m long pendulum has a period of about 2 s.
We now recall having approximated the sine of the angle with the angle (in
radiants) itself. Let us verify when the approximation is good. For example, if
θ = 30°, or 0.52 rad, its sine is sin 30° = 0.50. The relative error is (0.52–
0.50)/0.50 = 4 %, which is quite small. Even for θ = 60°, or 1.05 rad, the error is
not enormous, but already noticeable. Indeed, sin 60° = 0.87 and the
corresponding error is 20 %. These are the relative errors making the sine equal
to the angle, but the corresponding ones on the period are even smaller, as we
now shall see.
The exact Eq. (2.22), as we said, cannot be solved analytically. However, it
can be solved by successive approximations. In fact, the approximation we made
is a series expansion stopped at the first term (sin θ = θ); the next approximation
we stop at the second term (sin θ = θ − θ 3/6). The resulting expression for the
period with amplitude θ 0, calling T o the period given by Eq. (2.28), is

which, as it is seen, depends on the amplitude θ 0. The relative error made

using the usual expression of the period is . Going back to the above
examples, we find that the relative error for θ = 30° is 1.6 %, the one for θ = 60°
is 6.3 %. They are not large.
We make a last observation. If the oscillations are small, the pendulum
moves substantially on the horizontal, namely on the x-axis in Fig. 2.18. Now
x = l tan θ, that we can approximate with x = lθ. We can then conclude that the
motion, as represented by the x coordinate, has the equation
(2.30)
As expected it is a harmonic motion , of amplitude x 0.

2.10 The Work of a Force. The Kinetic Energy Theorem

In this section we introduce the concepts of work , done by a force, and kinetic
energy , of a body. The meaning of “work” in physics is rather different from its
meanings in everyday language and consequently from what intuition might
suggest. For example, holding in one hand a heavy object even if we do not
move it we still need to apply a force with our muscles and make some effort.
However, we do not perform any work, in the language of physics. In physics, a
force makes work only if its application point moves. In the example, the work
done by the force we exert on the body is positive if we raise, negative if we
lower it, but zero if we do not move it.
Consider the material point P moving in a reference frame with position
vector r, along a certain trajectory, the curve Γ. As shown in Fig. 2.20, consider
the position vector in the instants t, r(t), and immediately after t + dt, r(t + dt).
The displacement of P in the interval dt is the infinitesimal vector

Fig. 2.20 The elements to define the work of force F

(2.31)
If F is a force acting on the point, its work for the infinitesimal displacement
(2.31) is defined as
(2.32)
The finite work having been done by the force, a finite displacement of the
point, say from A to B along the trajectory Γ, is the line integral along the curve
Γ from A to B

(2.33)

where F(r) is the force in the point of position r. The line integral is the sum
of all the elementary dot products on all the elements of the curve.
Clearly, the integral does not depend only on the initial and final points A and B,
but also on the specific path taken to go from the former to the latter. Indeed, if
the path changes, also the force in the new points may change. To make this
explicit in the notation we have included both A and B and Γ in the subscripts of
W. The case in which the integral depends on the origin and the end but not on
the path is however important and will be studied in Sect. 2.13.
Notice that more forces, call them F i , may act contemporarily on the point
P, for example weight, friction, air resistance, etc. In this case, the total work
made by all the forces is equal to the sum of the works each force would do if
acting separately

(2.34)

Clearly, the elementary (meaning “infinitesimal”) displacement of the

application points of the forces d s is the same for all them. Considering that the
sum of integrals is equal to the integral of the sum, which is in our case the
resultant of the forces , we have

(2.35)

Namely, the total work made by the acting forces is equal to the work made
by their resultant . Notice, again, that this is true only if all forces are applied in
the same point.
The physical dimension of the work is those of a force times a displacement.
Its unit is the jule , with symbol J, which is the work done by the unit force, 1 N,
when its application point moves one unit of length, 1 m, in the direction of the
force. To appreciate the order of magnitude, a jule is roughly the work you do
when you raise a glass of water by 1 m.
We now prove the work -kinetic energy theorem. Being a consequence of the
second Newton law it is valid in inertial frames. Consider a material point and
the resultant R of the forces acting on it. The Newton law says

We take the scalar product with the elementary displacement d s = v dt of the

two members

Now consider the dot product . We recall that the square of a vector is
the dot product of the vector by itself, in this case . Differentiating this
expression we have

hence
The work done by R when the point moves from A to B on the given
trajectory is then

(2.36)

We then define the kinetic energy of the material point of mass m and
velocity υ as

(2.37)
which is independent of the position. The kinetic energy has the same
physical dimension as the work and is measured in jule. We finally can write
Eq. (2.36) as
(2.38)
which is the work-kinetic energy theorem. In words: when a material point
moves on a certain trajectory from A to B, the work done by the forces acting on
it is equal to the difference between the kinetic energy of the point has in B and
that it had in A.
It is sometimes useful to express kinetic energy in terms of momentum rather
than velocity, namely

(2.39)

2.11 Calculating Work

In this section we shall see two examples of calculation of works, made
respectively by weight and friction, when the application point P moves on its
trajectory from the initial position A to the final one B. We shall see that in the
former case the work depends only on the initial and final position, and not on
the path taken between them, in the latter it depends on the path too.
Starting with weight , Fig. 2.21 shows the reference frame (not necessarily
inertial) where we have chosen the z-axis to be vertical. Point P moves on the
trajectory from the position A, with the position vector r A = (x A , y A , z A ) to
the position B, with the position vector r B = (x B , y B , z B ). The figure shows
also the position vector at the generic instant t and in the immediately following
instant t + dt. The force acting on the point is its weight m g, which is equal in all
points. The elementary work done by the weight, which is vertically directed
downwards is . The total work is given by the integral
Fig. 2.21 Trajectory of the material point and its weight

(2.40)

We see that in this relevant case the work is independent of the path,
depending only on the final and initial position, even better, on their heights
only. This conclusion was experimentally proven by Galilei with a simple
experiment that we shall describe in the next section.
This is not the case of the second example, the friction force, which we shall
study in Sect. 3.5.
Suppose we have an object, say a book or a brick, lying on a table. In real
cases, the constraint does not apply to the body only the normal force, but also a
friction that is tangent to the contact surface. If we want to move the body on the
trajectory Γ in Fig. 2.22 at a constant speed, as we know from every day
experience, we need to pull it, apply a force, parallel to the plane in the direction
of the displacement. This means that the plane exerts on the body a force equal
and opposite to our pull, because the velocity is constant in magnitude and then
the resultant of the forces in the direction of the motion must be zero. Indeed, as
we shall see in Sect. 3.5, the friction force, F a , is always parallel and opposite
to the elementary displacement d s. We now calculate the work of F a .
Fig. 2.22 Calculating the work of the friction force

The elementary work is which is always negative. The

total work is given by the line integral on the trajectory

(2.41)

where s AB (Γ) is the length of the trajectory Γ between A and B. The work is
proportional to the length of the path, a quantity obviously depending on the
path.
We conclude with an observation that we shall generalize in Sect. 2.13. We
have seen that the work of the weight force for displacement A to B is W AB
= –mg(z B – z A ). Suppose now that the point goes back to A. The work of
weight is W BA = –mg(z A – z B ) = –W AB . Namely the total work of the weight
on a closed path is zero. On the other hand, the work of the friction force to go
from A to B on the curve Γ is . If we now go back on another
curve, say Γ′ in the Fig. 2.22, the work of the friction is , which
is again negative. Consequently the work of the friction on a closed path is not
zero, it is negative.

2.12 An Experiment of Galilei on Energy Conservation

One of the discoveries of G. Galilei was the fact, as we have mentioned, that the
velocity of body descending under the action of is weight only, starting from
rest, depends on the difference between the initial and final levels, and not on the
followed path.
In the “Dialogue on Two new sciences ” he states that the velocities of
bodies descending on inclines of different slopes and the same height are equal.
In his words (translations by the author):
 All contrasts and impediments removed… a heavy and perfectly round ball,
descending through the lines CA, CD, CB would reach the final points A,
D, C with the same moments

with reference to Fig. 2.23a reproduced from the book. Notice that, at the
time, Galilei was searching for and developing the laws of mechanics and that
several concepts had not yet been completely defined. In particular, impetus,
momentum, kinetic energy were not well-separated concepts.

Fig. 2.23 a Ball falling on inclines of different slopes; b the pendulum and nail experiment

However, accurate measurements of those velocities were impossible to do.

To prove the statement, he invented a simple and genial experiment, using a
pendulum and a nail. Figure 2.23b is also reproduced from his book.
Salviati, the person who in the Dialogues represents Galilei, starts with the
description:

Suppose this sheet to be a vertical wall and to have a lead ball of one or two
ounces hanging from a nail fixed in the wall, suspended to a thin wire AB,
two or three arms long, perpendicular to the horizon… and about two finger
far from the wall.

Then draw the vertical line AB and, perpendicular to it DC. Move the wire
with the ball in AC and let it go. We shall see the ball

descending first through the arc BCD, and going beyond point B as much
as, sliding on the arc BD, almost reaching the drawn horizontal CD, failing
to reach it by a very small gap, which has been taken away by the
impediments of the air and the wire; from which we can likely conclude
that the momentum (impetus) gained by the ball in B, in the descent on the
arc CB, was so much to pull it back through the similar arc BD to the same
 height.

He continues with the request to repeat the experiments several times to

check the result. Then

I want we fix in the wall, grazing the vertical AB, a nail, like in E or in F,
which should protrude out five or six fingers.

As before, the wire with the ball is moved to AC and let go. The ball will
again move on the arc CB. But, when it is in B, the wire hits the nail, forcing the
ball to move on the arc BG, having center in E.

Now, my Lords, you will see with enjoyment the ball reaching the
horizontal line in the point G, and the same to happen if the obstacle would
be lower, as in F, where the ball would go through the arc BI, always
finishing its ascent on the line CD.

Salviati concludes that the momentum acquired by a body descending from a

certain height is just what is needed to bring it back to the same height, through
whatever path. He observes that the momentum acquired in the descent on a
given arc is equal to the momentum needed to rise through the same arc. He
concludes that the momentum, and we can say also the velocity and kinetic
energy in B, is the same whether it descends through CB, or GB or IB or any arc
beginning on the horizontal DC and having its lowest point in B. On the other
hand, the fall along an arc can be thought of as the fall on an “incline” of varying
slope, proving the assumption.
The importance of the result of this experiment became clear in the following
evolution of mechanics. In his experiment the kinetic energy of the ball in B is
the same whatever the path starting from stillness from the same level. We now
know that this energy is equal to the work done by the weight force. We
conclude that the work done by the weight depends only on the difference of
level and not on the particular path followed. We have already discussed this
property in the previous section. Indeed, it is a fundamental one; it shows that
there is a quantity, the energy, which is conserved, does not change in the
motion under the action of weight. Weight is a conservative force, as we shall
now see.

2.13 Conservative Forces

In general the work of a force on a point depends on the trajectory of the point.
However, we have seen a case, the case of the weight force, in which the work
depends only on the origin A and end B and not on the trajectory between them.
Forces having these properties are said to be conservative . In the opposite case,
as for the friction, they are said to be non-conservative or dissipative .
Let r be the position vector in the chosen reference frame and F(r) be a
conservative force, a function of the position. The definition of conservative
force states that, for whatever curve Γ with origin in A and end in B,

(2.42)

where f is a function of the co-ordinates of A and of B. It is easy to show that

in this case it is always possible to find a function of the co-ordinates, which we
shall indicate with U p (r), such as
(2.43)
To show that, consider an arbitrarily chosen point o, as in Fig. 2.24. The
work from A to o on whatever path is

Fig. 2.24 Different paths

(2.44)
and similarly the work from o to B is
(2.45)
But, we can go from o to B also going from o to A and then from A to B.
Considering that work is an additive quantity we can write W oB = W oA + W AB .
Hence
(2.46)
By subtracting Eq. (2.43) from this expression we have
(2.47)
We then reach the result by putting . The function U p (r) is
the potential energy of the force F(r) and is a function of the co-ordinates only.
In conclusion the potential energy, or better its difference, is defined by the
relation
 (2.48)

In words: the difference of potential energy of the force F in the point B and
in the point A is equal to the opposite of the work done by the force when its
application point moves from A to B, following any trajectory.
The reason of the—sign, or the word “opposite”, is the following. To be
concrete, consider the weight. If we move a body of mass m from the level z A to
the higher level z B , the displacement is opposite to the force and the work
–mg(z B – z A ) is negative. The potential energy of the body is then larger when
its level is higher. The work done by the weight force is equal and opposite to
the gain of potential energy of the body. This energy can be given back as work
by the body, taking it down to the original level. The higher the body, the greater
is its potential to produce work.
We can conclude, and this is true in complete generality, by stating that the
potential energy difference between two states of a body is equal to the work we
need to do against the force acting on the body to change it from the first to the
second state.
Notice again that a potential energy can be defined for a force only if its
work is independent of the path. No potential energy exists, for example, for the
friction forces.
Notice also that only differences of potential energy can be defined, not its
absolute value. In other words, potential energy is defined up to an arbitrary
additive constant. In practice, we fix the constant choosing a reference position,
say o, in which we define the potential energy to be zero (U p (o) = 0), The
potential energy in the arbitrary point P is then

For example for the weight, we arbitrarily fix a reference level at which the
potential energy is zero by definition. This may be the ground level but some
other level too. We take that level as the origin of the vertical upward directed z-
axis and the potential energy is
(2.49)
We have stated that a force F is conservative if the work it does on a point
when it moves from position A to B is independent of the path. There are two
equivalent ways to state the same, which may be useful in certain circumstances.
1. A force is conservative if the work it does moving from A to B on any path is
equal and opposite to the work done moving from B to A on any path
(Fig. 2.25). This follows immediately from (2.48).

Fig. 2.25 The paths discussed in the text

2. The work of a conservative force on any closed path is zero.

In summary we can briefly say that the (equivalent) properties of
conservative forces are: (1) its work does not depend on the path, (2) admits a
potential energy, (3) the works going and going back are equal and opposite,
(4) the work on a closed path is zero.

2.14 Energy Conservation

Consider a material point P of mass m moving from the position A to the
position B on the trajectory Γ under the action of the (only) force F. Whether the
force is conservative or not its work is equal to the change of the kinetic energy
of the point. Denoting with U k the kinetic energy, we write
(2.50)
If, and only if, F is conservative, the same work is also the opposite of the
change of potential energy of the force
(2.51)
It immediately follows that
(2.52)
Considering that the positions A and B are arbitrary, we conclude that the
sum of the kinetic and potential energies is the same, i.e., is constant, in every
position of the motion. The sum is the total mechanical energy , say U tot of the
material point. The conclusion is so important that it is often called a “principle”.
The principle, or law, of energy conservation states that
(2.53)
If more than one force is acting on the point P and all of them are
conservative, Eq. (2.53) is still valid, provided that U p is the sum of the potential
energies of all the acting forces, or, in an equivalent manner, if it is the potential
energy of the resultant of those forces. Notice however, that the law is no longer
valid even if only one of the forces is dissipative.
In words, the law of energy conservation states that if a point moves under
the action of conservative forces only, its total mechanical energy is conserved
during its motion.
Consider now the case, which is what happens in practice, that also
dissipative forces are present. Consider for example the motion on an incline
under the actions of weight and friction. The kinetic energy theorem is still valid.
The work done by the forces for the displacement from A to B on the curve Γ,
can be written as the sum of the work of the conservative forces and that
of the dissipative ones and we have

but , and in conclusion

(2.54)
We see that, if non-conservative forces are active, the total mechanical
energy varies and its variation is equal to the work of the non-conservative
forces. The work of these forces is negative, as we saw for friction. Hence the
energy diminishes. This is the reason of the dissipative term.
The physical dimension of kinetic, potential and total energies are the same
as of the work. The measurement unit is consequently the jule .

Example E 2.1
Let us go back to the discussion made in Sect. 2.12 on the experiments by
Galilei on inclined planes . Figure 2.26 shows a body of mass m, which can fall,
starting from rest from point C, on inclines of different slopes CA or CD or
vertically on CB. Take a vertical upwards axis z, and denote by z C the height of
C (that is the height of the inclined plane).
Fig. 2.26 Fall on inclines or vertical

Consider the motion on CA. If friction is negligible the force exerted by the
constraint is normal and does not make work. The other acting force is the
weight m g.
The energy conservation principle applied to the displacement CA from C,
where the velocity is zero, to A, where z = 0, gives

(2.55)
or
(2.56)
We see that the final velocity depends only on the difference in level not on
the inclination.
If the friction is not negligible, the final energy is less than we have just
calculated. We can obtain it with Eq. (2.54) calculating the work of friction. The
latter does depend on the inclination for two reasons: the lengths of the paths are
different and the body pushes with different forces on the plane. To do the
calculation, however, we need to know something more on friction. We shall do
that in the next chapter.
We finally observe that the above arguments are valid if the body can be
considered a material point. If the body also rotates, like balls do, there is also
kinetic energy associated to the latter that should be considered. We shall discuss
this point in Sect. 8.16.

As we have just seen, in the presence of dissipative forces, the total mechanical
energy, namely the sum of kinetic and potential energy, is not conserved.
However, these are only two of many forms of energy. As a matter of fact the
law of energy conservation is one of the basic laws of physics. The law is
universally valid, without any exception, provided all the forms of energy are
included in the balance. Other forms of energy are chemical energy, thermal
energy, electric energy, nuclear energy, etc. Every time energy seems not to be
conserved, it is because we have failed to include one of its forms. The issue is
one of the main objects of thermodynamics, which will be discussed in the
second volume of this course. The historic process that led to clarification of the
concept of energy and to the establishment of the universal law of energy
conservation was very long. Starting, as we have seen, already with Galilei, the
process came to maturity only in the middle of the XIX century. It was then
established with the first law of thermodynamics, mainly by Julius von Mayer
(1814–1878) and James Prescott Joule (1818–1889). Energy is conserved also in
the presence of dissipative forces if internal thermal energy is included in
addition to macroscopic mechanical energy.

2.15 A Theorem Concerning Central Forces

A region of space in which a force that is a function only of the point, and
possibly of time, is called a force field . If the force does not depend on time, the
field is said to be stationary ; if it does not depend on the position, it is said to be
uniform .
The most common example of a uniform stationary field is weight, which is
constant in time and space (at least within the limits of a laboratory). On the
contrary, the viscous drag, the resistance of air to the motion, say, of a car or an
airplane, is an (increasing) function of speed and consequently is not a force
field.
A force field is said to be central if in every point P the force is directed as
the line between P and a fixed point, called the center of the forces . The
situation is sketched in Fig. 2.27, where C is the center of the forces.

Fig. 2.27 A central field of forces

 It is clearly convenient to choose the center of the forces as the origin of the
reference frame. We shall employ polar co-ordinates in which r = (r,θ,ϕ) is the
position vector. Let F(r) be the force under consideration. Saying that the force
is central means that the two vectors F and r are everywhere parallel. They may
have the same or opposite directions. The component of the force on the position
vector, the radial component, is its magnitude in the former case, the opposite in
the latter. This quantity may depend on the three coordinated, the two angles and
the distance r from the center. If the force depends only on r, the field is said to
have a spherical symmetry . On the other hand, a central force may be
conservative or not. We shall now prove that these two properties are correlated:
if a field of central forces has spherical symmetry, the force is conservative and,
vice versa, if a central field of force is conservative it possesses spherical
symmetry.
We start with the first statement. The radial component of the force, say F r
(r), is by hypothesis a function of the distance from the center r only. Given any
two points like A and B in Fig. 2.28, let us calculate the work done by the force
on an arbitrary curve Γ, having A as origin and B as end. We shall proof that it is
independent of the chosen curve. We indicate with d s the generic element of the
curve. The work corresponding to this elementary displacement is

Fig. 2.28 Work by a central force

(2.57)
where α is the angle between F and d s, which is also the angle between the
directions of r of d s. Hence, ds cos α is the projection of d s on the direction of
r, namely simply dr, i.e. the elementary variation of the distance from center.
N.B. Pay attention! This notation is universally employed, but is ambiguous. The
designation dr means the variation of the magnitude of the vector r, namely d|r|,
not the magnitude of the vector variation of r, namely |d r|.
 Anyway we have
(2.58)
Notice that this elementary work may be positive or negative depending on F
r and dr having the same or opposite sign. The total work on the curve Γ is

(2.59)

which is independent of the chosen curve, proving that the force is

conservative.
What we have proven is valid for whatever dependence on r. A particularly
important case is the gravitational force exerted by a mass M, which we shall
consider to be point-like, on another mass m, point-like too. We shall study the
gravitational force in Chap. 5. We anticipate here that such a force acting on m is
in any point directed towards the position of M; namely it is central. Its
magnitude is proportional to the product of the two masses and inversely to the
square of their distance r. Indicating by G N the proportionality constant, the
force is

(2.60)
where the minus sign indicates that the force is always in the direction
opposite to r, namely is attractive. The work done on a displacement from A to B
is

(2.61)

As expected, it is independent of path. We can then define the potential

energy of the gravitational force. The potential energy difference between the
point of position vector r B and the point of position vector r A is the opposite of
the work Eq. (2.61), namely

(2.62)
As always, the potential energy is defined up to an arbitrary additive
constant, namely

(2.63)
The constant is fixed choosing a point in which the potential energy is zero
by definition. In this case it is obviously convenient (but not at all necessary) to
choose this point at infinite distance, obtaining

(2.64)
This is the potential energy of a point-like mass m (the earth for example) in
the gravitational field of the point-like mass M (the sun). Notice that, in fact, this
is the energy of the pair of masses m and M (see Chap. 7).
We now prove the second of the above stated properties. We assume the
force to be central and conservative and show that its component (magnitude
with sign) on the position vector cannot depend on angles.
Let us consider for simplicity displacements on a plane. Consider a closed
path, as in Fig. 2.29, composed of two circular arcs centered on the center of
forces C, and two radial segments joining their extremes, at the angles θ 1 and θ 2
respectively. Take the radial segments of a very short length Δs. Assume by
contradiction that the magnitude of the force F would depend not only on r but
also on the angle θ. Under this hypothesis F r has different values on the two
radial sides that are at different angles, say F r1 and F r2. Let us calculate the
work of the force on this path. The contributions of the arcs are zero because on
them the force is perpendicular to displacement. The contributions of the radial
segments are –F r1Δs and F r2Δs. The total work is then , in
contradiction with the hypothesis that the force is conservative.

Fig. 2.29 The closed path used in the demonstration

2.16 Power
In physics, power is defined as the work done per unit time. For a given
delivered work, the power is larger for shorter delivery times. The simplest case
is the work done by a force, say F, on a material point, say P. Consider the
elementary displacement d s of the point, taking place between the instants t and
t + dt. The work done by the force is . The power w given by the force
the work divided by the corresponding time interval, that is

(2.65)
In words: the power delivered by the force F acting on a material point
moving at the velocity v in a given instant is equal to the dot product of the force
and the velocity of the point in that instant. If the force is a function of the
position, it must be obviously evaluated in the position of the point.
The physical dimensions of the power are those of a work divided by a time.
Its unit is the watt , after James Watt (1736–1819) One watt is the power
developed by a force delivering the work of one joule in one second
(1 W = 1 J/1 s). To have an idea of the order of magnitude, you develop about
1 W if you raise a glass of water by 1 m in one second.

2.17 Problems
2.1 A person is sitting on a chair supported by a horizontal ground. Draw the
diagrams of the forces for the person, the chair, and the earth. Describe each
of the forces, identifying the body that produces them and the body on
which they act. Identify the action reaction pairs.

2.2 A block hangs from the ceiling through a rope. A second rope is attached to
the bottom of the block. It hangs vertically and you draw it with your hands
downwards. Draw the diagrams of the forces for the block, each of the
ropes, your body, the ceiling and the earth. Describe each of the forces,
identifying the body that produce them and the body on which they act.
Identify the action reaction pairs

2.3 Fig. 2.30 represents two blocks of masses m 1 and m 2 on frictionless planes.
The plane of the first block is horizontal; the plane of the second is at an
angle θ. The two blocks are tied by a mass less inextensible wire that can
slide over a pulley without friction. (a) mentally insulate each block and
draw the force diagrams; then write three equations of motion, (b) find the
tension of the wire and the acceleration of m 2.
 Fig. 2.30 The two blocks of problem 2.3

2.4 A body of mass m = 1 kg moves in a circular uniform motion on a circle of

radius R = 0.1 m. What is the value of the centripetal force?

2.5 The system represented in Fig. 2.31 is in a vertical plane. M > m. Letting it
free, M goes down and m goes up. Neglecting the frictions, draw the
diagrams of the forces and determine the accelerations of M and of m.

Fig. 2.31 The two blocks of problem 2.5

2.6 With a hammer of mass m = 0.1 kg we beat on a nail, which is already

partially stuck in a piece of wood, with a speed of υ = 1 m/s. The nail
advances a distance of s = 1 cm. Find the force exerted by the hammer.

2.7 Two people pull a rope, each on one end, each with a force of magnitude F.
What is the tension? F or 2F? Why?
2.8 Two ropes hang from the ceiling. Two spheres of different masses hang at
the two ends. With both your hands you apply to the two spheres the same
force F, which is not necessarily in the direction of the rope. What are the
forces on each hand?

2.9 The three curves in Fig. 2.32 represent three rigid guides in a vertical plane.
Three rings of different masses slide without friction, one on each of them.
The three rings start from A at the same time with null velocity. State for
each of the following statements if it is true or false. 1. The rings reach B
contemporarily. 2. The rings reach B with velocities equal in magnitude.

Fig. 2.32 The three guides of problem 2.9 in a vertical plane

2.10 A man of mass m = 80 kg jumps from a platform at the eight h = 0.5 m

above ground. Reaching the soil he forgets to fold his legs. Fortunately the
ground is quite soft and stops the motion in a distance s = 2 cm. What is
the average force exerted on his bones during the stoppage?

2.11 Give an approximate evaluation of the height reached by a pole vaulter

athlete able to reach in his run the speed of υ = 10 m/s.

2.12 Fig. 2.33 shows three blocks of equal weight F p . The pulley is frictionless.
If we gradually increase all the weights, keeping them equal to each other,
which rope will break?
 Fig. 2.33 The system of problem 2.12

2.13 Two spheres, one with mass double that of the other, are launched upwards
with the same initial momentum p 0. If the resistance of air can be
neglected, what is the ratio of the heights they reach.

2.14 A particle of mass m = 2 kg oscillates on the x-axis. The equation of its

motion is , with x in meters and t in seconds. (a) What is
the magnitude of the force acting on the particle at time t = 0? What is the
maximum value of the force?

2.15 A twine of length l can hold a maximum tension T. It is employed to rotate

a mass m on a circle. Find the maximum velocity the body can rotate if (a)
the rotation is in a horizontal plane, (b) in a vertical plane. Draw in each
case the diagram of the forces.

Footnotes
1 The reader is warned that one can still find books and articles calling the product m i γ(υ) “relativistic
mass” and m i “rest mass”. The former in a relativistic regime increases with increasing velocity. These
concepts were introduced in the last years of the 19th century and the first ones of the 20th when
relativity theory was being developed and things were not yet completely clear. They are misleading
concepts (what varies with velocity is the Lorentz factor, not the mass, which is invariant) and should be
avoided. We shall treat relativity in Chap. 6.
© Springer International Publishing Switzerland 2016
Alessandro Bettini, A Course in Classical Physics 1—Mechanics, Undergraduate Lecture Notes in Physics,
DOI 10.1007/978-3-319-29257-1_3

3. The Forces
Alessandro Bettini1
(1) Dipartimento di Fisica e Astronomia, Università di Padova, Padova, Italy

Alessandro Bettini
Email: alessandro.bettini@pd.infn.it

In 1686, in his Preface to the First Edition of the Principia, Newton wrote
… the whole burden of philosophy seems to consist in this – from the
phenomena of motion to investigate the forces of nature, and then from
these forces to demonstrate the other phenomena.

The second law of motion states that the time derivative of the momentum of
a body is equal to the force acting on it. The law is not complete as long as the
forces are not known. As a matter of fact, the forces present in nature have
simple expressions. There are four fundamental forces: the gravitational force,
the electromagnetic force, the strong nuclear force and the weak nuclear force.
The two latter forces explain how matter behaves at a fundamental level. They
appear at nuclear and subnuclear scales, at which quantum physics is valid, and
do not directly appear in everyday macroscopic phenomena (even if, for
example, the weak force is responsible for nuclear fusion processes in the Sun
that give us light and energy). In the next chapter we shall study in some detail
the gravitational force and related phenomena. The electromagnetic force is the
object of the 3rd Volume of this course.
We have experience of several other forces. Apart from weight, which is
(mainly) due to the gravitational attraction of earth, all the other forces are
macroscopic effects of electromagnetic nature at microscopic level. Such are the
elastic force, the normal force of constraints, friction and viscous drag in a fluid,
both gas and liquid. These forces are not fundamental but are extremely
important for the study of every day phenomena.
We shall study these forces and the corresponding phenomena in this
chapter. The gravitational force will be treated in Chap. 4, fully dedicated to it.
Further study of the viscous drag will be done in the second volume of this
course.
The elastic force is met in a wide variety of circumstances. It gives rise to the
most important periodic motion, the harmonic oscillations. Harmonic
oscillations and the connected resonance phenomena, of which the mechanical
ones are prototypes, are present with very similar characteristics in all the
branches of physics, electromagnetism, optics, atomic an nuclear physics. Also,
the vast majority of the strongly interacting particles, which are called hadrons,
are extremely unstable, living only a few yoctoseconds. They are detected as
resonances. We shall study the harmonic oscillator in Sects. 3.8 and 3.9 and, at a
deeper level, in Volume 4 of this course.
In the last two sections we shall discuss the information that we can gather
on the motion of the bodies starting from the potential energy, rather then from
the force, which is possible if the forces are conservative. We shall introduce
energy diagrams in Sect. 3.10 and employ them in three important cases, elastic
force, pendulum and molecular forces, in Sect. 3.11.

3.1 Elastic Force

The solid objects of everyday experience have a definite shape. However, if
stressed by a force, or a system of forces, they deform. Consider the
geometrically simple situation of Fig. 3.1, showing a cylindrical metal bar
attached on one face to a fixed support. If we apply a force, parallel to its axis,
on the other face, the bar shortens if we push and lengthens if we pull. If the
force is perpendicular to the axis, the bar flexes.
Fig. 3.1 A metal bar and its deformations under an applied force F a

As another example consider a rubber band. If we pull on it at one end while

keeping the other end fixed, the band lengthens according to the force exerted.
When the force is removed, the ribbon returns to its natural length.
Many more examples can be cited. Their study shows that if a force is
applied to a body and then removed, the body resumes the shape it had before
being deformed, provided the deformation has not been too big. In these
conditions we speak of elastic deformation .
For larger deformations, the original size is not completely recovered, rather
some deformation is left permanently. This regime is called of plastic
deformation . The transition between elastic and plastic regimes is smooth. It
takes place at stress values that strongly depend on the chosen material. The bar
in Fig. 3.1, for example, will be permanently deformed by more feeble forces if
it is made of wax rather than of steel.
Let us study the phenomenon with reference to the metal bar of Fig. 3.1.
When it is not stressed by a force, its length is called rest length or natural
length is h. We choose a reference axis parallel to the bar directed outside with
the origin in its free end. When we apply a force F a (a for applied) the extreme
end moves and a new equilibrium state is reached. This means that the bar has
reacted by developing a force, say F, equal and opposite to F a . The
deformation, namely the difference between the actual and original shapes is, in
this, case, a change in the length of the bar. With the chosen co-ordinate, the
infinitesimal change dx, is positive for lengthening, negative shortening. The x
component of the force developed by the bar is in the positive x direction in case
of compression, negative in case of stretching. The magnitude of this force
increases with the deformation and is experimentally found to be, for not too
large deformations, proportional to the deformation x, namely
(3.1)
The constant k is called elastic constant or spring constant . It is a property
of the material characterizing its stiffness. Its physical dimensions are a force
divided by a length, its units are the newton per meter (N/m).
The proportionality between force and lengthening was experimentally
discovered by Robert Hooke (1635–1703) in 1676 and Eq. (3.1) is called
Hooke’s law . He made it public in a curious way. Initially he challenged his
colleagues with the anagram “ceiiinosssttuv”. Two years later, considering that
nobody had solved the quiz, he gave the solution: “ut tensio, sic vis” (as is the
stretch so is the force).
The Hooke’s law is very simple and very useful. However, it is not exact, but
approximate. Let us study the phenomenon more precisely. We apply to the
extreme of the bar a force of increasing and known values of intensity. At
equilibrium, these are equal to the magnitude of the force developed by the bar.
For each value we measure the deformation, both for extension (positive
deformation) and compression (negative deformation). Plotting the results in a
diagram we usually find the behavior of Fig. 3.2. For small enough values the
dependence of the force on the deformation is linear, Hook’s law holds. If the
force is too large however, the deformation, in the case of metals we are
considering, is larger than foreseen by the Hook law in compression, smaller in
extension (we shall understand the reason in Sect. 3.11). The non-linearity starts
at values smaller than those at which the deformation is permanent. The regime
in which the body returns to its original shape when the stress is removed is
called elastic, both if the deformation is linear or not. The former is called a
linear regime .

Fig. 3.2 Force versus deformation in the elastic and non-elastic regimes

The transition between linear and non-linear behavior is smooth and is found
at different values from metal to metal. It is smaller, for example, for lead than
for steel.
Suppose that we now keep increasing the force further, for example in
compression. The dependence of the deformation on the form is shown in
Fig. 3.3, curve (a). Let us now suppose that, having reached point Q, we start
decreasing the force, always measuring the deformation. We find that the
representative point in the diagram does not go back on the curve (a) but on (b).
Namely, for the same value of the force, the deformation is larger, in absolute
value, when we start from a deformed state. In particular, when the external
force, and the force of the bar with it, is back to zero, the deformation has a
value, x r , different from zero. It is called permanent deformation . We have
deformed the bar so much that we went out of the elastic regime and entered the
plastic regime .

Fig. 3.3 The elastic hysteresis

Figure 3.3 shows, for one value of the deformation, two values of the force.
In fact, the values are not only two, but a full range between a minimum and
maximum. If we perform the same process, changing the point Q at which we
invert somewhat further or somewhat sooner, the return branch is no longer (b),
but a similar one lower or higher in the diagram, but always below the curve (a).
In conclusion, the force does not depend only on the deformation but also on the
past elastic history of the body. The phenomenon is called elastic hysteresis .
For a given material we can define the elastic limit , which we indicate with
L. It is the maximum value of the deforming force (and of the force developed
by the body) divided by the section of the bar to remain in the elastic regime. It
is measured in newton per square meter (N/m2).
As all the forces that depend only on distance, as the elastic force (within the
elastic limit), are conservative. With reference to the co-ordinate in Fig. 3.1, we
now express the work W of F when the extreme of the bar moves from x 1 to x 2
in the linear regime. The work for the elementary displacement dx is, in this
regime, , hence

and we can define the potential energy function of x or, better, its difference

As always, to define its absolute value we need to choose arbitrarily a point

in which the potential energy is zero by definition. In this case it is quite
obvious, but not necessary, to chose the point x = 0 (that is zero deformation).
With this choice we have

(3.2)
This expression is valid within the linear regime. In the elastic non-linear
regime the force is still conservative and a potential energy could be defined, but
with a more complicated expression. In the plastic regime the force is dissipative
and no potential energy can be defined. Indeed, to be very rigorous, small
dissipative effects exist also in the elastic regime, but they can be neglected for
many practical purposes.
A deeper study of the elastic force shows that it is the resultant of an
enormous number of microscopic forces acting between the molecules of the
material. These are ultimately electromagnetic forces. The elastic force and the
Hook’s law are a macroscopic description of a very complex situation, which
depends on the specific microscopic structure of the matter of the body under
consideration.
Let us go back to the linear regime. If the body has a simple geometry, a
cylinder, a parallelepiped, a wire or a band we can define its length, say h, and
its section, say S. In these cases it is found with a good approximation that the
elastic constant of a body is directly proportional to its section and inversely to
its length

(3.3)
The coefficient E, which depends on the material (and its temperature), is
called Young modulus after Thomas Young (1773–1829) Using Eq. (3.1) in
absolute value, we can express the Young modulus as
 (3.4)
Namely it is the ratio of the deforming force per unit section, which is called
stress , and the deformation per unit initial length, called the strain . The stress is
a pure number, the strain and the Young module are forces per unit area and are
measured in N/m2.
It is useful to appreciate the orders of magnitude. The Young modulus values
of the metals range in the order of 1011 N/m2 (E = 2×1011 N/m2 for steels,
E = 1011 N/m2 for Cu, etc.). The elastic limits are around 108 N/m2 (L = 3×108
N/m2 for steel, L = 108 N/m2 for Cu). A third quantity is the fracture strength σ f
, which is the stress under which the bar breaks. For the metals the values are
two or three times larger than the elastic limits. Once the plastic regime is
entered, the fracture is nearing. The issue of the resistance under stress is very
important for engineering and the definition of safety limits is a much more
complex issue than the definition we have given.
Typical values of the three quantities for some substances are given in
Table 3.1.

Table 3.1 Elastic characteristics of some materials in N/m2

Young module Elastic limit Fracture strength
10−10 × E 10−7 × L 10−7 × σ f
Iron 20 20 35
Steel 22 30 50–200
Copper 10 10 20–40
Lead 1.5 1 1
Glass 6 2.5 3–9
Rubber 10−4 10−4 3 × 10−4

Going back to the orders of magnitude, consider a steel wire S = 1 mm2 in

section and h = 1 m in length. We fix it at one extreme and pull the other with
the force F a . Its stretch is

The maximum force in the elastic regime is

(For example if the wire supports a 30 kg weight). The corresponding elongation
is and the stress is quite small, around 1.5 per
mille. A stress a few times larger would break the wire.
Much larger stress without fracture can be obtained with other materials, like
rubbers. Typical values are around E = 106 N/m2, L = 106 N/m2 and fracture
strength of 3 × 106 N/m2. Consider a rubber wire of the same geometry of the
steel one we considered above, namely with S = 1 mm2 and h = 1 m. Under the
action of the force F a , the elongation is . The
maximum force in the elastic regime is , which is quite
small, as expected for rubber.
The corresponding elongation is . Hence, a rubber
band can be stretched to twice its original length without reaching the elastic
limit.
The reason for such different behavior of the metals and the rubbers is
explained by the molecular structure of the materials. The metals are made of
microscopic crystals. In each of these microcrystals the atoms are arranged at the
nodes of a regular lattice. The distances between the atoms are such that the
intermolecular forces are in equilibrium. When we try to deform a crystal we are
attempting to change those distances by acting against intermolecular forces that
are quite strong. Consequently the system is stiff, difficult to deform. On the
other hand, rubber is composed of very long, spaghetti-like, molecules. These
molecular spaghetti form a sort of tangled skein. The molecules interact amongst
one another like a pasta that has been cooked too much and became sticky.
When we pull the rubber we make the molecules straighter, but we do not
change their length. Consequently, the process is much softer than for a crystal
lattice and is reversible within much wider limits. We mention that when heated
a metal expands, a piece of rubber contracts. With increasing temperature the
equilibrium intermolecular distances in a crystal increase, while in the rubber the
increased rate of collisions between molecules increases their entanglement.
In summary, the metal wires can be loaded with rather large stress and their
strain is small. The rubber bands can have large strains, but do not bear large
loads. If we need to work both with rather intense forces and relevant
elongations we can use a steel helical spring, as in Fig. 3.4. When we pull the
spring, its turns flex but the wire does not change appreciably its length. The
elastic force is proportional to the tilt angle, hence to the elongation (or
contraction) and the Hook’s law holds.
Fig. 3.4 Helical steel spring

3.2 Harmonic Motion

We have already discussed the motion, i.e. the oscillations , of the simple
pendulum. This type of motion is important in every branch of physics and in
physics based technologies. We shall now study the motion in its details.
To be concrete, let us consider the system shown in Fig. 3.5. The block, of
mass m, lies on a horizontal plane, which we assume to be frictionless (we can
use the trick explained in Sect. 2.4). In these conditions the resultant of the
vertical forces is zero, the weight being equilibrated by the normal force of the
constraint. A spring is connected to the block at one end and to a fixed point on
the other. We take a co-ordinate axis, x, horizontally in the direction of the
spring elongations and with the origin at the point in which the spring is attached
to the block when it is at rest. In this way, x will measure the deformation of the
spring. We assume we are in the range of validity of the Hook law . The force
acting on the block is then

Fig. 3.5 A mechanical oscillator

(3.5)
which is a restoring force proportional to the displacement.
The equation of motion is
which we write in the canonical form

(3.6)
We now introduce the positive quantity
(3.7)
This has a very important dynamical meaning. is the restoring force per
unit displacement and per unit mass. It depends on the characteristics of the
system. We can then write Eq. (3.6) as

(3.8)
We have already met it (with a different expression, Eq. (2.29) for ω 0) when
discussing the pendulum . This very important differential equation describes the
motion of many systems, including pendulums, near their stable equilibrium
position, when subjected to a return force proportional to the displacement. The
general solution, as learned by calculus, is
(3.9)
where the constants a and b must be determined from the initial conditions of
the motion. They are two in number because the differential equation is of the
second order.
The general solution can also be expressed in the, often more convenient,
form
(3.10)
where now the constants to be determined from the initial conditions are A
and ϕ.
To find the relations between two pairs of constants, we start from

Hence
(3.11)
and reciprocally

(3.12)
We now introduce the terms used when dealing with this type of motion. To
do that in a general way, consider the expression (with a generic ω)
(3.13)
The motion is not only periodic, but its time dependence is given by a
circular function. Such motions are said to be harmonic . A is called the
oscillation amplitude , the argument of the cosine, , is called the phase (or
instantaneous phase in case of ambiguity) and the constant ϕ is called the initial
phase (indeed, it is the value of the phase at t = 0). The quantity ω, which has the
physical dimensions of the inverse of time, is called angular frequency and also
pulsation. Its kinematic physical meaning is to be the rate of the variation of the
phase with time and, notice, is independent of the initial conditions of the
motion. In the specific case we have considered above, the harmonic motion is
the spontaneous motion of the system (in Sects. 3.8 and 3.9 we shall study
motions under the action of external forces) and the angular frequency, ω o, as in
Eq. (3.10), is called proper angular frequency .
The motion is periodic with period
(3.14)
The number of oscillations per unit time is called the frequency , ν.
Obviously it is linked to the period and to the angular frequency by

(3.15)
The period is measured in seconds, the frequency in hertz (1 Hz = 1 s−1), the
angular frequency in rad s−1 or simply in s−1. The unit is named after Heinrich
Rudolf Hertz (1857–1899).
The harmonic motion can be viewed from another point of view. Consider a
circular disc and a small ball attached to a point of its rim. The disc can rotate in
a horizontal plane around a vertical axis in its center. Suppose the disc is rotating
with a constant angular velocity ω. If we look at the ball from above, in the
direction of the axis, we see a circular motion, but if we look horizontally, with
our eye in the plane of the rotation, we see the ball oscillating back and forth
periodically. Indeed, the motion is not only periodic, it is harmonic, as we now
show.
Figure 3.6 shows the material point P moving on a circumference of radius A
with constant angular velocity ω. We call ϕ the angle between the position
vector at t = 0 and the x-axis. The co-ordinates of P at the generic time t are
Fig. 3.6 A point P moving of circular uniform motion

The projection of the motion on the axes, in particular on x, is harmonic.

The conclusion leads to the simple graphical representation of the harmonic
phenomena shown in Fig. 3.7. To represent an harmonic motion of amplitude A,
angular frequency ω and initial phase ϕ, we take a fixed reference axis x and a
vector A, of magnitude A, rotating around its origin in the plane of the figure at
the constant angular frequency ω and forming with the x axis the angle ϕ at t = 0.
The projection of A on the reference x axis is our harmonic motion.

Fig. 3.7 Vector diagram for the harmonic motion

We can use this representation also for velocity and acceleration of the
harmonic motion. The derivative of Eq. (3.13) gives

(3.16)
As written in the last side, the velocity is seen to vary in a harmonic way too,
with a phase that is forward of π/2 radians to the displacement. This is shown in
Fig. 3.8a.
Fig. 3.8 Vector diagram for harmonic motion a velocity, b acceleration

Differentiating once more we have the acceleration

(3.17)
The acceleration is proportional to the displacement with the negative
proportionality constant –ω 2, or, as seen in the last side, its phase is at π radians
to the displacement, or in phase opposition with it.
We now go back to the oscillator in the linear regime and consider its
potential, kinetic and total energies. The former one is the potential energy of the
elastic force, which we have already expressed in Eq. (3.2). We can use now
Eq. (3.7) and write directly for the total energy

where x(t) is given by Eq. (3.10), and we obtain

(3.18)
We see that neither the kinetic nor the potential energy are constant in time,
rather, they vary as and respectively, but their sum, the
total energy is, as we expected, constant. Notice also that kinetic, potential and
total energies are all proportional to the square of the amplitude and to the square
of the angular frequency.
The mean value of a quantity in a given time interval is the integral of that
quantity on that interval, divided by the interval. It is immediate to calculate that
the mean values of both functions cos2 and sin2 over a period are equal to ½ (the
period of the square of a circular function is half the period of that function).
Consequently the mean values of both potential and kinetic energy over a period
are one half of the total energy.
 (3.19)

3.3 Intermolecular Forces

All bodies are composed of very small particles that we call “molecules”. These
molecules combine to form gases, liquids and solids. Molecules are composed of
atoms, a different type for each chemical element. Atoms are also composite
objects. Each one has a positively charged central nucleus composed of protons
and neutrons, while electrons form what may be thought of as a cloud
surrounding the nucleus. Electrons and protons are equal in number, the atomic
number , so that each atom is globally neutral. Different elements have different
atomic numbers. Quantum mechanics, not classical mechanics, correctly
describes the molecular and atomic phenomena. We can however give here a
few semi-quantitative pieces of information, with a classical language, that are
consistent with the prediction of quantum mechanics.
An electron inside an atom cannot be thought of as moving on a well-defined
trajectory similar to the orbit of a planet (as was assumed in the early stages of
the development of the theory, i.e., the Bohr model). We must instead consider
the probability of finding an electron at a particular location around the nucleus.
This probability is a known function of position different for each different
atom. It, in particular, vanishes at a certain distance from the nucleus. It is the
“cloud” we have mentioned above. In order of magnitude, the radiuses are tenths
of nanometers (or 10−10 m). Nuclei are much smaller, 1–10 fm (10−15−10−14 m).
If we magnified a nucleus to the size of the dot on an “i” of this page, the
diameter of the atom would be of the order of meters.
The elementary constituents of chemical substances are molecules. For
example, a water molecule is made of two hydrogen and one oxygen atoms,
nitrogen one of two nitrogen atoms, etc. Atoms are bound in a molecule by
electric forces that are described by quantum mechanics. As a matter of fact,
electric forces are very strong inside molecules, but, the molecules being
globally neutral, are almost null outside the “cloud” of their electrons. Not
completely however, as two molecules when they are close enough do interact
with a force, much weaker than those inside the cloud, called van der Waals
force , after Johannes Diderik van der Waals (1837–1923)
The van der Waals force between two molecules as a function of the distance
between their centers r is shown in Fig. 3.9. It is repulsive at small distances,
attractive at larger ones. In the diagram we adopted the convention of having
positive repulsive forces.
Fig. 3.9 The force between two molecules as a function of the distance between their centers

When the centers of the molecules are at the distance r 0, at which the van
der Waals force is zero, they are in equilibrium. At smaller distances the force is
repulsive and becomes quickly enormous. In a very rough approximation we can
consider them as rigid spheres of radius r 0. The dotted line in Fig. 3.9 is for an
idealized rigid body, which would be non-deformable. The force would be
repulsive and infinite when trying to squeeze it and null at distances larger than r
0 where it is not touched.

3.4 Contact Forces. Constraint Forces

If we put a heavy body, for example a brick, on a horizontal plane, it does not
accelerate, it is in equilibrium. This implies that the plane, in general the
constraint, has developed a force, called normal because it is perpendicular to the
plane, which is exactly equal and opposite to the force that the body, the brick,
exerts on the plane. The latter may be the weight, as in the example, or not. If we
push with our hand on the brick, the magnitude of the normal force is equal to
the sum of the weight and our push. Similarly, if we push a wall with a hand, it
does not move. The normal force made by the wall is equal and opposite to the
push.
The normal force is a contact force . If we raise the brick or take back our
hand from the wall, even at very small distances, the force disappears. Contact
force is the resultant of the forces between the molecules of the constraint and
the molecules of the body. When the two are in contact, molecules on the two
surfaces are at distances between their centers equal to molecular diameters. The
applied force tends to bring the molecules of the body and of the constraint
nearer to each other, namely to reduce their radii. This is opposed by the van der
Waals force, which, as we have seen, quickly becomes enormous. For this
reason two solid bodies cannot penetrate into each other. We have also seen that
the intermolecular force goes quickly to zero at distances larger than the
equilibrium position. This explains why the force disappears if we separate the
surfaces even by very small distances. Already at a few molecular diameters the
surfaces no longer interact.
Contact forces are used in practice when we want to constrain a body to
move on a certain trajectory. For example, we have repeatedly used a horizontal
plane to force a block to move in that plane; the rails force the train to move on a
certain path, etc. The physical systems used for this purpose, the support plane,
the rails, etc., are called mechanical constraints, because they constrain the
motion. The constraints may inhibit motion on one side only or both, being
named respectively unilateral and bilateral . A support plane is unilateral
because it does not inhibit a body from rising above its surface. The rails of a
train are unilateral but those of a roller coaster are bilateral, the coaster cannot
detach from the rail.
Usually, forces produced by the mechanical constraints are not known a
priori. They depend on the motion of the body, hence on other forces acting on
it. For example the force exerted by the rail on the wheel of a train in a given
curve depends on the curvature, but also on the speed of the train and on the
mass of the wagon. Indeed, the rail develops a force that is exactly the
centripetal force needed to have the wagon moving at that speed on that
curvature with its mass. The forces exerted by the constraints are said to be
passive, the other ones, which are usually under control, are called active.
We can however, calculate the passive forces if we know the motion of the
body and all the active forces acting on it. Let us look at two examples.

Example E 4.1
We have already studied the pendulum in Sect. 2.9. We recall that the simple
pendulum is a material body, of mass m, constrained to move on a circular arc of
radius l. The easiest way to implement a mechanical constraint is like in
Fig. 3.10a, with an inextensible wire fixed in Ω that exerts the tension T on the
material point. Clearly, the constraint is unilateral, because the wire can fold. We
could make it bilateral by using a light bar instead of the wire. In Fig. 3.10b the
constraint is implemented with a wooden or plastic guide shaped as an arc of a
circle of radius l, in which the body can slide. Assuming friction to be negligible,
the guide will develop a normal force. We represent it with the same symbol as
the tension of the wire, namely T.
Fig. 3.10 Two different mechanical constraints for the same motion. a simple pendulum, b solid guide

In both cases, the second law gives: . We already know the

motion and are interested in the constraint force T. We observe that in both cases
T is directed always towards the center Ω. The radial component of the resultant
of the forces must be the centripetal one, corresponding to the velocity υ of the
body, namely , where υ is the velocity at the considered instant and
the minus sign means that the force is towards the center. The radial component
of the resultant is and we have

(3.20)
Clearly, T is not a constant, rather it depends on the position of the
pendulum, which is defined by the angle θ. We could do that using the equation
of motion we have found in Sect. 2.9. However, it is easier to employ energy
conservation. The reason is the term mυ 2 in the last expression, which is twice
the kinetic energy. If the pendulum is abandoned from the initial position θ 0,
corresponding to the height y 0, the energy conservation equation is
.
Hence . But, and , hence
, and we can write and finally,
substituting in Eq. (3.20), .

Example E 4.2
Consider, in a vertical plane, an inclined guide connected at its lower extreme
with a circular guide, as shown in Fig. 3.11. We want to study the motion of a
material point, a small rigid ball for example, on the circular rail, which is
unilateral, of radius r. We use the incline to launch the ball with a certain initial
velocity on that rail. More precisely, we want to find the minimum initial
velocity in order that the ball would travel through the entire circle without
detaching from the rail.

Fig. 3.11 a The forces on a ball moving on a vertical circular rail, b motion of the ball in case of
detachment

Two forces act on the ball, its weight m g and the force of the constraint,
which we suppose to be normal, N. The latter is directed as the radius, towards
the center. The normal force cannot be directed outwards.
Again, the radial component of the resultant of the forces must be the
centripetal force requested by the motion. This component is the sum of N and
of the radial component of the weight. The latter is a maximum at the highest
point of the guide.
To be sure that the ball does not detach, it is then sufficient to verify that in
this point. Here, the weight and the constraint normal force are both directed
vertically downwards. The condition of non-detachment is then .
Solving for the unknown N we have .
The condition of non-detachment is N > 0, hence the term υ 2 > gr. If the
velocity is smaller, the ball detaches following a trajectory as in Fig. 3.11b,
which gives a sequence of images of the ball in its motion. We can think that in
this situation the weight is providing a centripetal force too large for the radius
of curvature of the guide, at that velocity. The motion must follow a trajectory
with a smaller radius, and the ball detaches.

3.5 Friction
We have already seen several times that a physical rigid plane, when pushed by
a body in contact with it, reacts with a normal force which is equal and opposite
to the active force. In the example drawn in Fig. 3.12 the plane is horizontal and
the active force, which is vertical, is simply the weight F w of the block lying on
the plane. The normal reaction N is vertical upwards.

Fig. 3.12 a Active and constraint forces on a block, b friction force versus applied tangential force

We now apply to the block a force F parallel to the contact surface

(horizontal in this particular case), by attaching a wire to the block and pulling.
Suppose that we gradually increase the tangential force starting from zero. We
observe that initially, when F is not very strong, the block does not move, it is
still in equilibrium. This implies that the resultant of the forces must still be zero,
not only in the direction normal to the plane, where nothing is changed, but also
in the tangent one, where now there is a force. The constraint must have
developed also a force parallel to the contact surface, F t equal and opposite to F,
namely

The force developed by the constraint parallel to the contact surface, when
there is no motion, is called static friction .
If we continue to increase the tangential force on the block F, the tangential
force by the constraint increases too, as long as the block does not move. This
happens at a certain value of the active force, meaning that the friction force
cannot be larger than a maximum value that we call F t,max.
This behavior is followed in all cases in which two dry surfaces are in
contact. In these conditions, it is experimentally found that the maximum value
of the static friction is proportional to the normal force, namely that
(3.21)
The proportionality constant µ s is called the coefficient of static friction ,
which is clearly a dimensionless quantity.
We now study the motion of the block when the tangential applied force is
larger than F t,max. By measuring its acceleration, we infer that a tangential
contact force F t is present, which is in general somewhat smaller than F t,max as
shown in Fig. 3.12b. Also in the case of relative movements of the two contact
surfaces, it is experimentally found that the tangential force by the constraint is
proportional to the normal one. Its direction is always parallel and opposed to the
velocity, namely
(3.22)
where is the unit vector of the velocity. The dimensionless constant µ d is
called coefficient of kinetic friction .
Figure 3.12b shows schematically the tangential force of the constraint
versus the applied tangential force. We see that F t grows to be equal to the
applied force up to F t,max. Then, when the motion is started, it diminishes
somewhat, as we have already noticed, and then remains approximately, but not
exactly, constant. Notice that in the majority of the cases µ d < µ s but there are
also opposite cases.
As a matter of fact, the static and dynamic friction forces are due to the
interactions between the molecules on the surfaces of the two bodies.
Consequently, Eqs. (3.21) and (3.22) are a macroscopic description of a complex
microscopic situation. We observe that friction coefficients depend critically on
the status of the surfaces in contact, on how they have been machined, on their
cleanliness, etc. Notice carefully that the molecules on the surface of a body
made of a certain substance, for example copper or steel, are not only of that
substance. Water is almost always present, oxidation too. One can find
mentioned values of the friction coefficients between, say, copper and copper,
copper and steel, etc. But, there is no single copper on copper, etc. friction
coefficient, for the just mentioned reasons.
As a matter of fact, for example in the case of a piece of copper, it is possible
to obtain surfaces populated by copper molecules only. The piece must be
processed with ad hoc procedures under a vacuum, because in the presence of
air, copper will oxidize and water molecules will be deposited on the surface
immediately. Now suppose we have produced two such blocks in a vacuum and
put their surface in contact. They immediately stick one onto the other and you
will not be able to separate them. They became a unique copper bock. How are
molecules supposed to know to which block they belong?
The first astronauts to land on the Moon observed this phenomenon. Putting
two stones gathered from the soil in touch, they found them sticking together and
difficult to separate, even if their surfaces were obviously irregular.
There is no universal mechanism at the origin of the friction between two
contact surfaces. Consider the important case of two metal surfaces. Metallic
surfaces can be worked to be extremely smooth. Even in these conditions,
surfaces are not smooth if looked at nanometer scales. Figure 3.13 tries to show
the surfaces as seen at a large magnification. The irregular patterns have a
typical scale of 10 = 100 nm.

Fig. 3.13 Pictorial view of the contact surfaces between two metals, at nanometer scale

When two surfaces are, we think, in contact, the contact is indeed only
between the “crests” on the two sides. Consequently the surface really in contact,
say S c is much smaller than the nominal surface S (typical values of S c /S are
between 10−4 and 10−5). However, the larger is the normal force N pushing the
two surfaces one against the other, the larger is the number of crests touching
each other. We can then understand why the friction force is proportional to N.
We can also understand why it is independent of the area of contact. Suppose we
keep N constant and double the contact macroscopic surface S. The action of the
normal force will distribute on a doubled area and its effect on the crests per unit
area will halve. The number of contacts per unit surface will halve too, but they
will cover a twice as large area. The total number of contact has not varied. In
conclusion, S c is proportional to N and independent of S.
In the contact points the molecules of the two bodies interact strongly
attracting each other and becoming, so to say, welded. To have one surface
sliding on the other, these micro welding points must be broken. Again the
necessary force is proportional to S c and consequently to N and independent of
S.
What we have just described is relative to dry surfaces between solid bodies
and has nothing to do with the friction between lubricated surfaces. In this case,
a film of liquid is present between solid surfaces, the molecules of which are far
enough away from each other to have an interaction. In this case the friction is
due to the viscosity of the lubricant (see Sect. 3.6).
The rolling resistance or rolling friction is the force resisting the motion
developed by the constraint, for example the support surface, when a cylindrical
or spherical body, such as a reel or a ball, rolls on the surface. Figure 3.14
represents in cross section such a cylinder, say a reel, of radius r. We apply a
force F to the axis of the reel parallel to the support plane and normal to the axis.
We assume that the reel does not slide on the plane due to the static friction
force. This type of motion is called pure rolling . When the reel rolls, it does that
about an instantaneous axis that is the contact generator in the considered
instant. The moment of the applied force about the instantaneous rotation axis is
τ = rF. The moment τ necessary to have the rolling at a constant angular velocity
is experimentally found to be proportional to the magnitude of the normal force
N, namely

Fig. 3.14 Schematics of the rolling resistance

(3.23)
where γ is the rolling resistance coefficient . Its physical dimension is a
length, and is measured in meters. The applied moment is equal and opposite to
the moment developed by the constraint.
The rolling resistance force is generally smaller than the dynamic friction. As
a matter of fact it is due to quite complicated phenomena in the region of contact
between the reel and the support plane. In Fig. 3.14 this region is shown as a flat
area of longitudinal with δ. This is an idealization, because actually both the
cylinder and the plane deform into shapes that are not forward-backwards
symmetrical. We are here simplifying a lot. We can say that on the contact area a
number of the above considered “crests” of both bodies are in contact. The
difference is that now, to have movement, the microwelds are broken acting in a
direction normal, rather than parallel, to the surface. This requires, caeteris
paribus, a smaller force.

Example E 4.3
Consider Fig. 3.15. A brick lies on an inclined surface, the inclination of which,
α, can be varied. Given the coefficient of static friction µ s , what is the
maximum value of α at which the brick remains still?

Fig. 3.15 A brick on a slide and the forces acting on it

The forces on the brick are its weight m g and the force exerted by the
constraint. The latter can be decomposed in a normal, N, and a tangential, F t ,
component, which is the friction. For equilibrium the components of the
resultant must be zero. Namely, and . Hence
. But, the static friction force cannot be larger than µ s N, and the no-
slide condition is .
The maximum angle, say is called the friction angle . For
example, the slopes of the piles of sand or of the screes in the mountains
naturally settle on the corresponding friction angle.
We have seen in Sect. 2.11 that friction forces are dissipative, and that their
work is negative when their application point moves, because they are always in
a direction opposite to the motion, see Eq. (2.41). Indeed, the friction forces are
always such as to oppose the relative motion of the two bodies. This does not
imply that the friction acting on a body would always act to slow it down, on the
contrary it can also accelerate it.
As an example, let us consider our brick, of mass m, ling on the horizontal
platform of a cart. The latter moves straight forward with constant acceleration a
(see Fig. 3.16) in the direction of its velocity v. If the acceleration of the cart is
not too large, the block remains still relative to the platform; its motion is
accelerated with the same acceleration a as the cart. It must be acted upon by a
force equal to m a. But the only horizontal force acting on it is the friction F t .
Hence, F t = m a. The friction accelerates the brick. We know that F t can be at
most equal to µ s N = µ s mg. Consequently the maximum acceleration of the cart
at which the brick does not slide is µ s g.

Fig. 3.16 A brick on an accelerating platform and the forces acting on it

Notice that in this case the friction has the direction of the velocity, namely
of the displacement. Consequently its work is positive. In the same way, when
we start running we are accelerated by the friction force between our shoe soles
and the ground, when a car accelerates the accelerating force is the friction
between its reel and the road. Notice however, that in these cases the work of the
friction force is zero, because the point of application does not move.

3.6 Viscous Drag

A solid moving relative to a fluid, a liquid or a gas, is subject to a force, different
from friction, but as friction opposing the relative motion of the body and the
medium. It is called viscous drag or viscous resistance . Differently from
friction, there is no drag when the relative velocity is zero, and an increasing
function of the relative velocity. The direction of the drag force is always equal
and opposite to the relative velocity.
The magnitude of a force depends on the magnitude of the relative velocity,
on the shape of the body and on the fluid. Moving relative to the fluid, the body
induces a number of effects that may perturb substantially its flow. Think for
examples of vortices. Consequently, the dependence of the drag force on
velocity is complicated. We shall study it in the second volume of this course,
together with fluid dynamics. Here we anticipate only a few elements that are
needed in our study of the motions of bodies.
The force depends on the shape of the body, for example it is different for a
cylinder or a sphere, on its orientation, for example the case of a disc is different
for its orientation parallel or perpendicular to the flow, and, for a given
geometrical shape, on its size. We shall limit the discussion here to a spherical
body, of radius a.
The force depends on two characteristics of the fluid, its density ρ (mass per
unit volume) and the viscosity η. The latter will be discussed in the second
volume. It suffices to know here that it characterizes the difficulty with which
the fluid flows, so, for example, oil has larger viscosity, is more viscous, than
water, but is less viscous than honey. For a given fluid, the viscosity depends on
the temperature.
The physical units of viscosity are

(3.24)
where, in the third member we have taken into account that the dimensions
of the force are [F] = [MLT −2]. Pressure has the dimensions of a force per unit
surface (FL −2) and its unit is the pascal (Pa), from Blaise Pascal (1623–1662).
The unit for viscosity is then the pascal second (Pa s). For example, for some
everyday fluids at ambient temperature, their viscosities are for oils η ≈ 0.5–
1.5 Pa s, for water η ≈ 10−3 Pa s, and for air η ≈ 1.8 × 10−5 Pa s.
The Reynolds number is a parameter that gives relevant information on the
regime of the motion, named after Osborne Reynolds (1842–1912). It is
dimensionless, namely a pure number. The four quantities of the problem have
the physical dimensions , , and .
They can be arranged in a dimensionless quantity as
(3.25)
which is the Reynolds number for a sphere. Its expressions for other shapes
are similar.
Figure 3.17 shows schematically how the drag force on a body can be
measured. The body is fixed to a thin bar and to the pointer of a dynamometer
fixed on a support and is immersed in the fluid under study, which is moving at a
known velocity υ, that we can vary in a known manner. Experiments of this type
show that at small velocities the drag force can be written as the sum of a term
proportional to the velocity and one proportional to its square
Fig. 3.17 Measuring the drag force

(3.26)
where the coefficients A and B depend on the body and the fluid but, for not
too large velocities, are independent of velocity. As the ratio between the second
and the first term is proportional to the velocity, the first term dominates at small
velocities, the second at larger ones. We define as critical velocity υ c the
velocity at which the two terms are equal. It corresponds to a quite small value
of the Reynolds number
(3.27)
Consider now the sphere moving in air, as pendulums or free falling bodies,
at normal temperature and pressure conditions. The air density in these
conditions is ρ = 1.2 kg/m3. With the value for viscosity already given, the
Reynolds number is
(3.28)
and the critical velocity, in a round number
(3.29)
If for example a = 1 cm, the critical velocity is υ c = 4 cm/s. The time taken
to reach it by a body freely (in a vacuum) falling from rest is t = υ/g = 4 ms,
which is very short indeed. In this time it would travel in vacuum d = gt
2/2 = 80 µm. For larger dimensions bodies moving in the air the critical

velocities are even smaller.

We conclude that only for very small velocities, smaller than υ c , is the
viscous drag proportional to the velocity. However, it becomes proportional to
the square velocity very gradually, reaching that regime only at Reynolds
numbers two orders of magnitude larger than in Eq. (3.27), corresponding to
velocities of a few meters per second for a sphere of 1 cm radius.
As a second example consider the same sphere moving in water. With
ρ = 103 kg/m3 and the viscosity given above, η/ρ = 10−6 kg/m3, which is a value,
notice, smaller than for air. The Reynolds number at velocity υ for a = 1 cm is
Re = 104 υ. The critical velocity is only υ c ~ 2.5 mm/s.
In the elementary study of free fall, of the motion on an incline and of the
pendulum, the viscous drag of air is usually neglected. Is this a good
approximation? Let us control on a few typical cases. Consider a bronze (density
ρ = 8 × 103 kg/m3) ball of a = 2 cm radius and three cases: free fall from a
h = 20 m tall tower, descent of an incline of elevation h = 1 m and oscillation of
a pendulum abandoned at the height from the position at rest h = 0.5 m. The
weight of the ball is F p = 2.7 N. Neglecting the presence of the air, and the
energy of the rotation in the second case, the velocities at the end of the fall
would be in any case , hence υ 1 = 20 m/s, υ 2 = 4.5 m/s, υ 3 = 2 m/s in
the three cases respectively. In presence of air all velocities would be somewhat
smaller, but larger than the critical velocity. The drag force is approximately
proportional to the square velocity, but is not very large. For the just mentioned
velocities its values are approximately R 1 = 2.4 × 10−2 N, R 2 = 1.2 × 10−2 N, R
−3
1 = 2.4 × 10 N, which are in any case small compared to the weight.
Neglecting the drag in these cases is not a bad approximation. However, the
effect will be noticeable on much longer times.
Finally notice that, whatever its expression, the viscous drag is a dissipative
force. As it is always directed opposite to velocity, its work is negative for any
displacement of the application point.

3.7 Air Drag and Independence of Motions

In the study of the motion of a body under the action of one (or more) force, it is
often convenient to decompose the motion in its components on the Cartesian
axes. The component of the motion on an axis is due to the component of the
force (or forces) on that axis. The component motions are independent of each
other. This law of the independence of motions was discovered by Galilei , and
assumed by Newton as a corollary of the second law. We have already quoted in
Sect. 1.16 the following example from Galilei. Suppose we shoot a ball with a
gun at the top of a tower, aiming horizontally. Simultaneously we drop a ball
with zero velocity. The ball leaves the rifle barrel with a very high horizontal
speed and, under the action of its weight, describes a parabola finally touching
ground at a horizontal distance far away. The second ball falls vertically. Galilei
established that both balls touch ground at the same instant, provided that the
action of air is neglected, as he specifies.
We shall now analyze the motion in presence of the air and we shall see that
the law of independence of motions is not always valid.
We refer the motion of the ball, of mass m, to a frame having the y-axis
vertical and x horizontal in the plane of the motion as in Fig. 3.18. Two forces
are acting on the ball, its weight F p = m g vertical downwards and the viscous
drag

Fig. 3.18 The forces on a ball moving in air

(3.30)
where υ is the velocity and u υ is its unitary vector. The Newton law gives

(3.31)
The components on the axes of the equation, if θ is the angle of v with the
horizontal, are

(3.32)
This is a system of two non-linear differential equations, which cannot be
easily solved. However, we are only interested here in knowing if and when the
two motions are independent. To be so, only x and y components should appear
in the first and second equation respectively. This is indeed the case for low
velocities, when the term B can be neglected. In these conditions, considering
that and , Eq. (3.32) becomes
 (3.33)
The two motions are independent. However, if, as it is often the case, the
drag is proportional to the square velocity, Eq. (3.32) become

(3.34)
The motions are not independent. This is an obvious consequence of the
proportionality of the drag force to the square of the velocity, which depends on
both components. In the example of Galilei, the air resistance is larger for the
gun ball than for the vertically falling one, because the velocity of the former is
larger. The gun ball touches ground later than the ball falling from the tower if
the effects of the air are not neglected.

3.8 Damped Oscillator

In Sect. 3.2 we discussed the motion of the harmonic oscillator . We then
neglected the dissipative forces, which however, are always present. As we
know these are basically of two types, friction and viscous drag of the air. We
shall now include the viscous drag of the air, which we shall assume to be
proportional to the velocity.

Fig. 3.19 A mechanical damped oscillator

To be concrete, consider the system in Fig. 3.19, which is similar to that in

Fig. 3.5, with the addition of an element providing the viscous force . We can
think in terms of an absorber, like a piston moving in a fluid, but the element is
meant to represent all the viscous forces, including that due to the air. The
viscous drag is proportional to velocity in magnitude and opposite to it in
direction, namely

(3.35)
where β is a constant. We shall neglect the friction between the support plane
and the block. The force (3.35) tends to slow down or damp the motion. Hence
the oscillator is said to be damped. The second law gives

(3.36)
which we write, dividing by m and taking all the terms to the first member,
in the “canonical” form

(3.37)
In this form, the equation is valid for all harmonic damped oscillators. The
two parameters depend on how the oscillator is built, the strength of the spring,
the viscosity, etc. We have already met the first one while discussing the
harmonic oscillator. It is the restoring force per unit displacement and per unit
mass
(3.38)
The second, see Eq. (3.35) is the resistance force per unit velocity and unit
mass
(3.39)
Notice that both constants have the dimension of the inverse of time. We
already know that ω 0 is the angular frequency of the oscillator in absence of
dissipative forces. The inverse of the second
(3.40)
is the time that characterizes the damping, as we shall now see.
The solution of the differential Eq. (3.37) is given by calculus. The rule to
find it is as follows. First we write the algebraic equation obtained by
substituting in the differential equation powers of the variable equal to the
degree of the derivative. In our case it is
(3.41)
Then we solve it. The two roots are

(3.42)

The general solution of the differential equation is

(3.43)
where C 1 and C 2 are integration constants that must be determined from the
initial conditions.
Let us discuss the motion we have found. We observe that the effect of the
dissipative force, which is to damp the motion, is larger for larger values of γ.
Considering the two roots r 1 and r 2 of the algebraic Eq. (3.41), three cases
should be distinguished called respectively: under-damping if γ/2 < ω 0, the two
roots are real and different, over-damping if γ/2 > ω 0, the two roots are complex
conjugate, and critical damping if γ/2 = ω 0, the two roots are real and
coincident. Let us analyze the three cases.
Over-damping. The two solutions, which are real, are both negative. The
motion is the sum of two exponentials decreasing in time
(3.44)
The damping is so large that the system is not able to perform even a single
oscillation. The displacement from the equilibrium position decreases
monotonically. Mathematically speaking, Eq. (3.44) says that the time to reach
that is infinite. In practice, after some time both addenda are so small, and so is
the velocity, that other resistive forces that are always present, as the friction,
stop the motion in the equilibrium position (x = 0). This happens in a time
interval of a few times 1/|r 2| (which is larger than 1/|r 1|).
Critical damping. The two roots coincide, r = –γ/2 = –2/τ. In this particular
case, Eq. (3.42) is not the solution. This is
(3.45)
In this case too the system does not oscillate. The displacement reduces to
zero, in practice, in a time interval of a few times 2τ. It can be shown that in the
critical damping the time to reach equilibrium is minimum.
Under-damping. We can write the equation of motion in the form

where

(3.46)
We can now choose two different integration constants as a = C 1 + C 2 and
b = i(C 1 – C 2) and have the solution in the form

(3.47)
For damping tending to zero ( ) the equation of motion becomes
Eq. (3.9), as we expect since the oscillator is un-damped in these conditions. The
solution can be written in a form analogous to Eq. (3.10)
(3.48)
where now the integration constants are A and ϕ. The motion is an oscillation
similar to the harmonic motion with an amplitude, , which is not constant
but decreases exponentially in time with a decay time 2τ. The oscillations are
damped. A weakly damped, namely with γ ω1, motion is shown in Fig. 3.20.
The oscillation amplitudes diminish gradually in a time long compared to the
period. As a matter of fact, rigorously speaking, the motion is not periodic,
because the displacement after every oscillation is somewhat smaller than before
it. However, if the damping is small, γ ω1, we can still identify a period

Fig. 3.20 Weakly damped oscillations

(3.49)
The weak damping condition γ ω1 can be written as τ T, in words, the
decay time is much longer than the period.
Notice that the proper angular frequency ω 1 is smaller than the proper
angular frequency of the free oscillator ω 0, but that for γ ω1 the difference
becomes infinitesimal of the second order compared to γ/ω 0.
We have seen in Sect. 3.2 that the total, kinetic plus potential, mechanical
energy of the harmonic oscillator is constant in time. The difference now, even
in the case of weak damping, is that a dissipative force is present. We expect that
energy decreases. Without losing generality, we can assume the initial phase to
be zero. The initial amplitude of oscillation is A. At every oscillation, the
displacement reaches its maximum at, say, time t. The displacement is then
(3.50)
In that instant the velocity is zero and the total energy is equal to the
potential energy, which is proportional to the square of the amplitude
(3.51)
The total energy decreases exponentially in time, reducing to a value 1/e of
the initial value in a time τ, which is one half of the time in which the amplitude
reduces of the same factor. τ is called decay time of the oscillator.
An observation on the exponential function. The amplitude of a damped
oscillation in the Eq. (3.48) and the energy of the damped oscillator, Eq. (3.50)
are examples of physical quantities decreasing exponentially in time. This
behavior is often met in physics. We make here a simple but important
observation. Consider the function

and the ratio between its two values in two different instants t 1 and t 2 (t 1 < t
2). We immediately see that this ratio depends only on the interval t 2 − t 1 and
not separately on the two times or the constant (the initial value) f 0. Indeed

(3.52)
In particular, the function diminishes by a factor 1/e in every time interval
and not only in the initial one.
In particular, we can reformulate the above statement in: “τ is the time
interval in which energy reduces of a factor 1/e”.

3.9 Forced Oscillator. Resonance

Consider again the damped oscillator of the previous section and apply to the
body a force in the direction of the x-axis that oscillates as a circular function of
time with angular frequency ω and amplitude F 0. The component of the force
on the x axis (its magnitude or its opposite depending on the direction relative to
x) is given by
(3.53)
where we have chosen the origin of times and the instant in which the force is
zero; its initial phase is then null. The second law is

(3.54)
which we write in the form
 (3.55)
The left-hand side of this equation is that of the equation of the damped
oscillation (3.37). But the right-hand side, which is zero for the latter, is now
proportional to the external force. Equation (3.55) is a non-homogeneous
differential equation and Eq. (3.37) is its associated homogeneous differential
equation. A mathematical theorem states that the general solution of the former
is the sum of the general solution of the associated homogeneous equation and of
any particular solution of the non-homogeneous one.
We shall limit our discussion to the case of weak damping, as in Fig. 3.20.
We can guess that a possible motion might be a harmonic oscillation at the
angular frequency of the force; namely a particular solution might be
(3.56)
with some amplitude B and initial phase –δ to be determined. Let us check if
our guess is correct. The easiest way to do so is to consider an equation exactly
similar to (3.55) of the complex variable z(t) = x(t) + i y(t). The imaginary part
y(t) is some function that is irrelevant in our arguments. We then search for a
solution of the differential equation, of which (3.55) is the real part

(3.57)
Considering that the equations are linear, the real parts of the solutions of
Eq. (3.57) are solutions of (3.55). The function corresponding to our guessed
solution is
(3.58)
Let us try it in (3.57)

which must be satisfied in every instant of time. And so it is, because all the
terms depend on time by the same factor. Hence, Eq. (3.55) is a solution
provided that

(3.59)
which is an algebraic equation. The unknown, the parameter we must find to
have the solution, is the complex quantity z 0. This is immediately found to be

(3.60)
 We see that the solution is completely determined by the characteristics of
the oscillator, ω 0 and γ and of the applied force, F 0 and ω. It does not depend
on the initial conditions.
The particular solution of Eq. (3.57) is then

(3.61)
To have a particular solution of Eq. (3.55) we must now take the real part of
this expression. To do that it is convenient to write z 0 in terms of its modulus B
and its argument –δ (we shall soon see the reason for the negative sign)
(3.62)
Equation (3.60) gives z 0 as a ratio. The modulus of a ratio is the ratio of the
modulus of the nominator and the modulus of the denominator

(3.63)
The argument of the ratio is the difference between the argument of the
nominator, which is null, and the argument of the denominator, and its opposite
is

(3.64)
The particular solution of Eq. (3.57) is then
(3.65)
and, taking the real part, the particular solution of Eq. (3.55) is
(3.66)
Finally, the general solution of Eq. (3.55) is
(3.67)
Let us now discuss the motion we have found. It is the sum of two terms.
The first one represents a damped oscillation at the angular frequency ω 1 that is
proper for the oscillator. The constants A and ϕ, depending on the conditions
from which the motion started, appear in the first term. The second term depends
on the applied force. The motion is under these conditions quite complicated.
However, the amplitude of the first term decreases in time the faster the greater
is γ. It diminishes by a factor of e in every time interval 2/γ. After a few of such
intervals, the first term has practically disappeared. Once this transient regime
has gone, the regime of the motion is stationary. The stationary oscillation or
forced oscillation is described by our particular solution Eq. (3.66), which is
called a stationary solution . We write it as
(3.68)
We repeat that the stationary motion is a harmonic oscillation at the angular
frequency of the force, not at the proper frequency of the oscillator. However,
both the amplitude B and the quantity δ, which is not the initial phase but the
phase delay of the displacement x relative to the instantaneous phase of the
force, do depend on the characteristics of both the oscillator and the force as in
Eqs. (3.63) and (3.64). An important phenomenon, the resonance , happens
when the angular frequency of the force is near or equal to the proper angular
frequency of the oscillator: the amplitude is very large and the phase delay varies
very rapidly.
Figure 3.21 represents the amplitude of the forced oscillation as a function of
the angular frequency B of the force. It has a maximum at the resonance
frequency

Fig. 3.21 Dependence on the applied force angular frequency of forced oscillations a amplitude, b phase
delay, γ increases from continuous to dashed curve

(3.69)
as one obtains with the usual methods finding the derivative of Eq. (3.63).
Notice that ω R is close but not exactly equal both to the angular frequency of the
damped oscillations ω 1 and the proper angular frequency of the free oscillator ω
0. However for small damping, namely for γ/ω 0 1, all of them become almost
equal.
A simple way to observe the resonance phenomenon, and to understand the
reason for the noun, is using two tuning forks. The tuning fork is an acoustic
harmonic oscillator that vibrates at a specific frequency when set vibrating by
striking it. It is made, like a two-pronged fork, with U-shaped prongs, called
tines, and a stem of a metal, usually steel. The instrument is used to have a
definite pitch, typically an A at 440 Hz, to tune the music instruments.
 We strike a tine of one of the forks to have it vibrating, and we hear the
sound, with the other one a few meters far. We then bring the latter nearby and
stop the first fork by touching its tines. And we still hear the pitch. The second
fork, that has the same proper frequency, resonated. The first fork had excited
sound waves in the air, namely pressure oscillations at the frequency of its
vibrations (the sound we hear). These pressure oscillations act as a periodic force
on the second fork at its resonant frequency. We can double check that this is
true as follows. We fix, with a locking screw near the top of one of the tines of
the second fork, a small weight and repeat the experiment. This time we do not
hear the second fork sound. Its proper frequency is now different and it is no
longer in resonance with the first one.
Going back to Fig. 3.21a, we observe that calculations show that the full
width of the resonance curve at half maximum (FWHM) is equal to γ and that
the maximum is inversely proportional to γ. As a matter of fact, Eq. (3.63)
immediately shows that the amplitude is infinite in the ideal case of γ = 0.
We discuss now the behavior of δ, the phase delay of the displacement
relative to the force, given by Eq. (3.64) and shown in Fig. 3.21b. When the
frequency of the force is small relative to the proper one, ω ω 0, then δ ≈ 0,
namely force and displacement are in phase. On the contrary, if the frequency of
the force is much larger than the proper frequency, ω ω 0, then δ ≈ π. We can
easily understand the physical reasons for that, considering the relative
importance of the different terms in Eq. (3.53). At low frequencies the
accelerations are quite small and the applied force acts mainly against the elastic
force –kx and is consequently in phase with x. At high frequencies, as we have
just seen, force and displacement are in phase opposition; when the mass is on
the right, the force pushes to the left and vice versa. Accelerations are now very
large and the dominant term is , namely the inertia. Force acts
mainly against acceleration and is in phase with it, which we know to be in
phase opposition with displacement.
We also notice that our calculation shows the transition to be between the
two just described regimes and takes place in a angular frequency interval of the
order of γ. The less is the damping the more sudden is the transition. In
resonance, as immediately seen in Eq. (3.64) δ = π/2, namely the displacement is
in quadrature with the force, hence it is in phase with the velocity. The power
exerted by the force that is the product of the force and the velocity is a
maximum.
The resonance phenomenon is very common in nature and in technology, not
only in mechanics but also in electromagnetism, optics, atomic physics, nuclear
and particle physics. In fact, all the systems oscillate harmonically when
displaced close to a stable equilibrium configuration. We shall discuss this in
Sect. 3.11. These oscillations take place at definite frequencies characteristic of
the system. Engines, for example, have always a rotating part. Irregularities in
their structures, even if small, may produce periodic stresses of an axis and of
the support structures at the frequency of engine rotation. When this is varied
and reaches one of the resonance frequencies of the system (there may be more
than one) the amplitude of the vibration may become very large and, if the
damping is small, even destroy the engine, if it is not properly designed.
We now consider the energy stored in the oscillator when in its stationary
motion, Eq. (3.68). It is the sum of the kinetic and potential energies

(3.70)
The expression is similar to what we found for the free oscillator. However,
the two terms are now proportional one to and one to ω 2 while for the free
oscillator both were proportional to and the energy was constant in time.
Now the total energy varies periodically. This is because the power delivered by
the force is not equal at a single instant to the power dissipated by the viscous
force, while their averages on a period are equal. The instantaneous balance
exists, however, in resonance, when and the total energy

(3.71)
is constant.

3.10 Energy Diagrams in One Dimension

In our previous discussion, the role of the force has been the principal one, while
that of the potential energy was somewhat secondary. However, when in a more
advanced study of mechanics and in other fields of physics, the potential energy
has a central role. We have studied the problem: given a conservative force, find
its potential energy. We now consider the inverse problem: knowing the
potential energy, find an expression of the force.
For simplicity we consider the motion in one dimension only. The point P
moves on a line, which we take as the x-axis. Suppose we have only one force
acting on the point and call F x its x component. Suppose the point moves from x
1 to x 2. In general, the knowledge of x 1 and x 2 is not sufficient to know the
work done, but we also need to know the path taken. The point P might have
gone directly from x 1 to x 2, or have moved in the opposite direction and after a
while have come back, etc. For example, if the force is friction, its work is
proportional to the total length of the path. However, if the force is conservative,
as we shall assume, its work depends only on x 1 and x 2 by definition. For
example the force might be an elastic force, or the weight of the point. In this
case we have

(3.72)

We can fix the arbitrary additive constant by choosing a position x 0 in which

the potential energy is null by definition and write

(3.73)

We now want to invert Eq. (3.73). To do that we take the derivative of both
its members, immediately obtaining

(3.74)
In one dimension, the force is the opposite of the derivative of its potential
energy with respect to the position. For example, the potential energy of the
weight (x is vertical upwards) is and the corresponding force, by
derivation, is the one we know , the elastic potential energy is
and, by derivation, the force is .
Equation (3.74) can be written as
(3.75)
which shows that the elementary work of a conservative force is the
differential of a function, the opposite of the potential energy.
Suppose now that the potential energy U p (x) of the force F x (x) acting on
our point P (in one dimension) to be the function shown in Fig. 3.22. The study
of this type of diagram, called energy diagrams , is often useful to understand,
even if in a semi-quantitative way, the possible types of motion of the system.
Fig. 3.22 The energy diagram example discussed in the text

We start from the equilibrium positions of point P. A position, more

generally a state, is said to be of equilibrium when, if the system was abandoned
in that position with null velocity, it remains there indefinitely. This means that
in these positions the force is zero. We can recognize immediately these
positions on the diagram as those in which the derivative, i.e. the slope of the
curve, is zero, namely where the curve has a maximum, a minimum or a flex, x 4,
x 7, x 9, x 11 in the figure. However, in practice we can never position the body
exactly in a position and if we try to do that in a maximum or in a flex, the body
will run away. As a matter of fact, there are three types of equilibrium states. To
be general (for the material point) we define them in three dimensions.

1. Stable equilibrium . A position of a material point is of stable equilibrium if,

when it is removed in whatever direction by an infinitesimal distance, the
resultant of the forces tends to bring it back towards the equilibrium position
(restoring force).

2. Unstable equilibrium . A position of a material point is of unstable

equilibrium if at least a direction exists such that, when the point is moved in
that direction by an infinitesimal distance, the resultant of the forces tends to
bring it further away from the equilibrium position.

3. Indifferent or neutral equilibrium . If the point is removed by an infinitesimal

distance in any direction, the point remains there. In other words the
equilibrium position is surrounded by other equilibrium positions.
 Going back to Fig. 3.10, in one dimension, the position x 4, where the
potential energy has a relative minimum, is of stable equilibrium. Indeed, if we
move a small distance on the left, the force –dU p /dx is positive, hence in the
direction towards x 4. On the contrary, if we move to the right the force, –dU p
/dx, is negative, hence directed to the left.
The position at a maximum, like x 7 in the figure, is of unstable equilibrium.
If we move a bit on the left the force is to the left (–dU p /dx < 0), if we move to
the right, the force is to the right. In both cases the force tends to pull the point
farther from equilibrium.
In fact, one of these conditions is enough to make the equilibrium unstable.
This happen in the positions of the flexes, like x 9 in the figure. Moving the point
to the left, the force is restoring, but moving it to the right the force is of
removal.
Consider finally the position x 11. It is on a non-null segment on which dU p
/dx = 0, namely it is surrounded by other equilibrium positions and the
equilibrium is neutral.
We warn the reader that the curve representing U p suggests a ball moving on
hills and valleys, namely to think of the ordinate axis as the height. This is
indeed the case for weight, but not for other forces. Intuition should be
controlled.
Three other pieces of information can be extracted from the diagram. For
every value of potential energy, the material point may have different values of
kinetic energy. The sum of the two
(3.76)
is constant during its motion. In Fig. 3.22 we have drawn, as examples, four
different values of the total energy. In any case, the kinetic energy is the
difference between total energy and potential energy, the distance from the line
and the curve.
We now consider that, while total and potential energies may be positive or
negative (or zero), the kinetic energy cannot be negative. Consequently, if the
total energy is too low, as is U tot1 in the figure, for which in every point the
kinetic energy would be negative, it is not possible for our system. The total
energy cannot be less than the absolute minimum (the deepest in the figure) of
the potential energy.
 If the total energy is somewhat larger, as is U tot2 in the figure, the motion of
the point can happen only in two regions, between x 3 and x 5 in and between x 10
and x 12. In this example, the two regions are separated by a non-reachable
interval. If a point starts moving with total energy U tot2 in one of the two
regions, it cannot leave it. One might think that the point might jump from one
allowed interval to another, because total energy would remain the same. But
this cannot happen because between the two regions there is a forbidden one, in
which the total energy would be different or kinetic energy would be negative.
However, this type of phenomenon happens in quantum mechanics, in atomic
nuclei for example, which is called “tunnelling ” even if no tunnel exists,
because it looks as though the system would cross under the barrier in a tunnel.
We can learn something more from the diagram in Fig. 3.22. Suppose our
material point with total energy U tot2 to be at a certain instant in x 3. In this point
the line of U tot2 intersects the curve of the potential energy. All the energy of the
point is potential, its kinetic energy is zero. The point has zero velocity.
However, the point does not remain still, because it is not in an equilibrium
position. The force –dU p /dx is positive and accelerates the point in the direction
of increasing x, with increasing kinetic energy (the distance on the diagram from
the line to the curve increases). The force, and consequently the acceleration,
slows down as the slope of the potential energy curve diminishes. They become
zero when the point reaches the minimum in x 4. The motion does not stop there,
it continues now decelerated (the slope is positive, hence the force is negative,
opposite to x). We read from the diagram that the kinetic energy is now
diminishing. It does so up to zero when the point is in x 5. This point is reached
with zero velocity.
What does happen afterwards? The force –dU p /dx acting on the point is
now negative, namely in the direction opposite to x. The point restarts its motion
going back. In conclusion, the motion is an oscillation back and forth between x
3 and x 5. The motion is periodic, but generally not harmonic. We shall see in the
next section under what conditions it is so. Clearly also the motion with total
energy U tot2 between x 10 and x 12 is periodic.
Consider now a larger value of the total energy, U tot3 in the figure. There are
two possible motions. The first one is bounded, a periodic oscillation between x 2
and x 6 similar to that we have just discussed. The second motion is, for
example, the motion of a point approaching from infinite distance on the right.
Initially it accelerates, then, once x 11 is passed, decelerated up to stop in x 8.
Here it bunches back going through in inverted order all the phases moving
farther and never to come back again. The motion is unbounded.
At still higher values of the total energy, as U tot4, no periodic motion is
possible, but only unbound motions. A particle coming from far away, once
reached x 1 stops and bunches back to infinity.

3.11 Energy Diagrams for Relevant Forces

In this section we shall use the methods described in the previous section to
relevant types of motion: the oscillation under the action of a real elastic force ,
the oscillation of a pendulum and the oscillations of a diatomic molecule.
Let us start with a perfectly elastic spring on the x-axis, which has its origin
in the rest position of the spring. We know that the force it exerts on a point in
the generic position x is F x = –kx and the potential energy is U p = –kx 2/2. In
practice, as we saw in Sect. 3.1, no spring is perfectly elastic. For large
deformations the dependence of the force on the deformation is no longer linear
(see Fig. 3.2) or, in other words, the curve of the potential energy is not a
parabola, but as shown in Fig. 3.23.

Fig. 3.23 The potential energy versus deformation for an ideal (dashed curve) and real (continuous curve)
spring

The equilibrium position x = 0 corresponds to the minimum of the potential

energy. If the material point is abandoned outside the equilibrium position, its
oscillations are periodic. They are also harmonic if the displacement is not too
big, say within the vertical dashed lines that mark the region in which the
potential energy curve is at a good approximation a parabola.
Consider now a simple pendulum . We attach a small sphere of mass m to a
wire of length l and negligible mass with the other extreme fixed. If the sphere is
abandoned without velocity from out of equilibrium it will move on the arc of a
circle of radius l. The position can be measured with one variable, the angle θ
between the wire and the vertical. The potential energy is
(3.77)
Figure 3.24 shows this function. The variable θ can take any value from −∞
to +∞. However, the function is periodic and all the possible physical positions
are already described by the values of θ between –π and π.

Fig. 3.24 The energy diagram for a pendulum

In the figure we have taken the minimum potential energy as the zero of the
energy scale. We can see that the motion can be unbounded (in angle), if the
total energy is larger than 2mgl, which is the maximum potential energy, as U
tot2 in the figure. The angle θ grows indefinitely in time, the pendulum rotates on
the circle of radius l (in practice the wire would tangle around the nail). The
velocity varies from a minimum when the ball is in its highest position (θ = π,
3π, 5π,..), to a maximum when it passes through the equilibrium position (θ = 0,
2π, 3π,..).
If U tot < 2mgl, as for example U tot1 in the figure, the motion is limited. The
ball oscillates between the angles –θ 0 and +θ 0. In general however, the motion
is not harmonic, because the potential energy curve is not a parabola. If the
oscillations are small, however, the curve is approximately parabolic, as shown
in Fig. 3.25, and the motion is harmonic . The same thing can be seen
analytically. If we develop Eq. (3.77) in series and stop at the first term we
obtain
Fig. 3.25 The potential energy of a pendulum and its parabolic approximation

Notice that the approximation is quite good because the next term in the
development, the term in θ 3 is null, hence the first neglected term is the one in θ
4.
The last example is the diatomic molecule . To be concrete we consider HCl.
With a good approximation we can consider the two components as point-like.
The atomic clouds of the two atoms keep the two nuclei at the stable equilibrium
distance r 0. If the distance is different, a force appears, which tends to bring
back the equilibrium. These forces, which are responsible for chemical bonds,
are electromagnetic and of quantum nature. They are different from the van der
Waals forces we considered in Sect. 3.3. Figure 3.26 shows the potential energy
as a function of the distance between the nuclei of H and Cl. It is known as
Morse potential . The curve has a minimum, corresponding to the equilibrium
distance between the nuclei. The distances are of the order of the nanometers.
The energy is given in electronvolt (eV), which is a practical unit for atomic
energies. An electronvolt is the kinetic energy gained by an electron falling
under the electric potential difference of one volt. Its value is in round figures

Fig. 3.26 The potential energy of the HCl molecule

(3.78)
Suppose now we communicate a certain energy to the system, for example
by striking with another molecule. Also in this case there are two types of
motion. If the energy given to the molecule is large enough, as U tot2 in the
figure, the motion is unbounded. The two ions separate and the molecule
dissociates. If the energy is smaller, like U tot1, the molecule remains bound and
performs a periodic oscillation. As seen in Fig. 3.26, the potential energy curve
is not symmetric about its minimum. However, if the total energy is small
enough and the curve can be approximated with a parabola, the oscillation is
almost harmonic.
We also observe that the potential energy curve grows more rapidly at
energies smaller than the minimum than at higher ones. In other words, the
restoring force is larger than it would be if elastic for compression, smaller for
expansion. Macroscopically this translates in the asymmetry of the deviations
from the behavior described by Fig. 3.2.
The resonance phenomenon is present also in the molecular oscillators, at
quite high frequencies, of the order of 1013 Hz (10 THz). These are the
frequencies of the electromagnetic waves in the infrared. Imagine doing the
following experiment. We radiate a container with transparent walls containing a
HCl gas with an infrared radiation, of which we can vary the frequency and we
measure the intensity of the radiation transmitted by the gas in correspondence.
Taking the ratio between the transmitted and the incident intensities we have the
quantity of radiation absorbed by the gas as a function of frequency. We obtain
Fig. 3.27. It is a resonant curve, because in resonance much more energy is
transferred from the radiation to the molecular oscillators than for other
frequencies. However, two peaks, not one, are observed. The reason is that
Chlorine has two isotopes, 35Cl and 37Cl of atomic masses 35 and 37
respectively. The two proper frequencies squared are different, as the forces
are equal, the masses different, in the two cases. To be complete, in the spectrum
several doublets like the one in Fig. 3.27 are present. This is because quantum
oscillators have several, rather than a single one, proper oscillation frequencies.

Fig. 3.27 Absorption probability for HCl molecules versus frequency

 From the examples in this section we can draw an important conclusion. The
physical systems are found naturally in their (or one of their) stable equilibrium
state(s) corresponding to the minimum (or one of them) of the potential energy.
A small perturbation can take them out of equilibrium. The potential energy
curve is not in general a parabola. However, if the displacement from
equilibrium is small enough it can be well approximated by a parabola. In these
conditions the system oscillates harmonically. Consequently, the largest fraction
of natural oscillations are indeed harmonic .

3.12 Problems
3.1. Consider the oscillator of Fig. 3.5 with m = 0.3 kg and k = 30 N/m.
Calculate the proper angular frequency, the period and the frequency of its
oscillation. Write the equation of motion for an initial displacement, with
zero velocity, of 4 cm.

3.2. Show that the amplitude of a damped oscillator is halved in a time of 1.39/γ.
How much is the energy variation in this time?

3.3. A damped oscillator has the proper angular frequency ω 0 = 300 rad/s and ω
0/γ = 50. Calculate the angular frequency of the free oscillations ω 1 and the
resonance frequency ω R. Compare the values.

3.4. We build a mechanical oscillator as in Fig. 3.5. We can use a body with a
certain mass and two identical springs. We separately attach to the mass: (a)
one spring, (b) two springs in series, (c) two springs in parallel. What are
the ratios of the proper angular frequencies in cases (b) and (c) to case (a)?

3.5. A perfectly elastic spring stretches 10 cm when it hangs a mass of 10 kg. (a)
what is the value of the spring constant? (b) Lay the spring and the mass on
a horizontal plane without friction. Move the mass so as to stretch the
 spring 5 cm and let it go at t = 0. Write the equation of motion if (a) the
initial velocity is zero, (b) the initial velocity is 1 m/s in the direction of
increasing x.

3.6. Consider a forced oscillator vibrating at the angular frequency ω in its

stationary regime. Show that its energy is mainly potential when ω ω 0,
mainly when ω ω 0, exactly half and half when ω = ω 0.

3.7. We know the oscillation amplitudes of the displacement and the velocity of
a harmonic oscillator. How can we know the angular frequency?

3.8. A force with sinusoidal dependence on time acting on an oscillator makes it

oscillate, in a stationary regime, with amplitude A 1 = 20 mm. A second
force, acting alone on the same oscillator, makes it oscillate in the
stationary motion with amplitude A 2 = 40 mm. If both forces act together,
the amplitude in the stationary motion is A = 30 mm. What is the phase
difference between the forces?

3.9. A car of mass m = 1000 kg travels horizontally at 100 km/h. Suddenly an

obstacle appears at 100 m. The driver brakes immediately (neglecting the
reaction time, which is 1–2 s) and stops 10 m before the obstacle. Assuming
the force to have been constant how much was the magnitude of the force?
If the road were downhill with a slope of 15 % at which speed the car
would have hit the obstacle?

3.10. A block of mass m = 1 kg lies on a horizontal plane attached to a rope, the

other extreme of which is fixed to the point O of the plane. The block
under these constraints is moving on a circle of center O and radius
l = 1 m and velocity at the considered instant υ = 2 m/s. The coefficient of
kinetic friction between block and plane is µ d = 0.4. What is the
 magnitude of the resultant of the forces in that instant? What is the
direction relative to velocity?

3.11. A sphere of radius a moving with velocity υ acts in air with a drag force R.
The latter depends on the radius as with
and . Consider a raindrop falling
starting from null velocity. The drop moves under the action of its weight
and the resistance. When the velocity is small, the weight is larger than the
resistance and the drop accelerates. However, at a certain velocity the two
forces become equal and opposite and the velocity becomes constant. It is
called limit velocity. Calculate the limit velocities for a drop of radius
a = 0.1 mm and for one of radius a = 1 mm. In both cases assume the
second term in the above expression can be neglected. Verify a posteriori
if the assumption is reasonable. For a drop of radius a = 1 mm, now
assume that the first term is negligible and again verify a posteriori if the
hypothesis was reasonable.

3.12. A body of mass m is attached to an extreme of a rope of length R. The

other extreme is fixed. The body rotates in a vertical plane. a) Find the
expression of the tension T of the rope when the body passes, with
velocity υ, in the highest point of the trajectory. What is the agent of the
centripetal force in this point? Study the meaning of the found expression
for decreasing values of υ. What does T > 0, T = 0 and T < 0 mean? What
does happen when the velocity is such that T = 0? Repeat for the lowest
point.

3.13. A small body starts sliding, with negligible initial velocity, on a

frictionless wheel starting from its highest point, as in Fig. 3.28. The
radius of the wheel is R. (a) at what height h, measured from the center of
the wheel does the body detach and fall freely? (b) how would the result
change on the moon?
Fig. 3.28 Body sliding on a wheel, problem 3.13
© Springer International Publishing Switzerland 2016
Alessandro Bettini, A Course in Classical Physics 1—Mechanics, Undergraduate Lecture Notes in Physics,
DOI 10.1007/978-3-319-29257-1_4

4. Gravitation
Alessandro Bettini1
(1) Dipartimento di Fisica e Astronomia, Università di Padova, Padova, Italy

Alessandro Bettini
Email: alessandro.bettini@pd.infn.it

The first two books of Newton’s Principia establish the mechanics laws for
phenomena on the surface of earth. The third book, titled “The system of the
word”, applies the same laws to interpret the motions of extra-terrestrial bodies.
The grand unification of terrestrial and heavenly physics, started by G. Galilei
and J. Kepler, was completed. In the introduction to the volume, I. Newton wrote
It was the ancient opinion of not a few, in the earliest ages of philosophy,
that the fixed stars stood immovable in the highest parts of the world; that
under the fixed stars the planets were carried about the sun; that the earth,
as one of the planets, described an annual course about the sun, while by a
diurnal motion it was in the meantime revolved about its own axis; and the
sun, as the common fire which served to warm the whole, was fixed at the
centre of the universe.
This was the philosophy taught of old by Phylolaus, Aristarchus of
Samos, Plato in his riper years, and the whole sect of the Pythagoreans; and
this was the judgment of Anaximander, more ancient still …

A few lines below, after having mentioned the contributions of the Romans
and of the Egyptians, he added

It is not to be denied that Anaxagoras, Democritus, and others, did now and
then start up, who would have it that the earth possessed the centre of the
world, and that the stars of all sorts were revolved towards the west about
 the earth quiescent in the centre, some at a swifter, others at a slower rate.
However, it was agreed on both sides that motions of the celestial
bodies were performed in spaces altogether free and void of resistance. The
whim of solid orbs was of a later date, introduced by Eudoxus, Calippus
and Aristotle; when the ancient philosophy began to decline, and to give the
place to the new prevailing fictions of the Greeks

Observation of the night sky, with its moon and countless stars has, since
ancient times, never failed to astonish humanity throughout the world. Along
with astonishment, a deep curiosity aroused about the nature of these heavenly
bodies and the reasons of their existence. Along with the myth, truly scientific
activities developed in time in several cultures. Since the second millennium
B.C. mankind accurately and systematically registered the positions of the stars
in the sky. However, the mystical charm of the starry sky contributed to the
suggestion, in several periods, that the motion of the heavenly bodies should
have obeyed symmetry rules of a higher, often divine, order. This is the case of
the solid orbits of Aristotle, mentioned by Newton, and of the uniform circular
motions of Ptolemy and Copernicus. Gradually, beginning in the Renaissance,
there developed an inquiry leading to establishment of the physics laws that rules
the motions in the cosmos.
In this chapter we shall study universal gravitation, the physical law that
describes motions of the planets and their satellites, of the solar system and of
the galaxies and their clusters as well as the motions of all bodies up to the
boundaries of the Universe. We might start from the Newton law of gravitation
and analyze its consequences. We prefer to reach it following, albeit briefly, the
historical process that led to discovery of the law. Indeed, the path leading to
these discoveries has never been straight, but rather tortuous, through lateral,
sometimes wrong, paths, with successes and failures, laborious in any case.
Universal gravitation is one of the grand theories built by several scientists.
Knowledge, even if in a summary, of the historical roots of the process adds to
the depth of the physics laws. As a matter of fact, physics can be understood
even without knowing its history. The historical part of the chapter should be
considered as a, hopefully interesting, reading adventure. The parts to remember
are the laws and their experimental proofs.
Figure 4.1 shows the lifetime spans of the great authors of the development
of mechanics and astrophysics from the XVI to the XVIII century, the period of
the construction of a vast theoretical edifice.
Fig. 4.1 Life spans of the principal contributors

In Sect. 4.1 we shall briefly describe the geocentric and heliocentric models.
In Sect. 4.2 we shall see how the periods and diameters of the orbits of the
planets were measured from Greek civilization to the Renaissance. We shall then
see the fundamental contribution of Tycho Brahe with his systematic
measurements, with precision increased by an order of magnitude, of the
positions of the planets and how Johannes Kepler, based on those measurements,
discovered that the orbits of the planets are ellipses, rather than circles, and
established his three laws. The Kepler laws are very important but still
phenomenological. The dynamical theory was later established by Newton, as
discussed from Sects. 4.4 to 4.6.
The Newton law is a simple and symmetric mathematical expression. In the
fundamental physical laws the harmony of the world takes on an abstract
character, appearing as the simplicity of the mathematical expression that is able
to describe an enormous quantity of different phenomena, which, when that law
was not known, appeared to be uncorrelated.
The Newton law contains a universal constant, which is the same on earth
and in the Cosmos. In Sect. 4.7 we see how it was measured in the laboratory.
The gravitational force acts between bodies that are not in contact, rather
they may be very far from each other. The force acts through a vacuum. This is
also the nature of all the other fundamental forces, in particular of the
electromagnetic one. For all of them the concept of field of force is important.
The source of the force, for example the sun, creates a field of force in all the
space around it. The field then acts on every massive object as a force. We shall
see that in Sect. 4.8.
In Sect. 4.9 we shall go back to history and show how G. Galilei discovered
the satellites of Jupiter, discussing some of his data.
In Sect. 4.10 we shall see how the Newton law describes the motions of
cosmic objects of the most different sizes and distances and how it shows that
the nature of the largest fraction of matter is still unknown. It is called dark
matter.
In the first part of the chapter we assumed for simplicity the orbits of the
planets to be circular. In the final three sections, we relax this assumption and
discuss fully the problem of elliptic orbits. This is known as the “direct Kepler
problem”: knowing that the orbit is an ellipse with the centre of force in one of
the foci, find the force. We shall do that first using modern calculus formalism
(Sect. 4.10), then, in (Sect. 4.11) we shall read and explain, line by line, the
original demonstration of Newton, as a beautiful example of his thought. In the
last section, we shall consider the energy of a body in the gravitational field of a
central body.

4.1 The Orbits of the Planets

Observational astronomy is a very ancient science, dating back to the most
ancient civilizations to the third millennium B.C. The varying celestial co-
ordinates of the stars, of the moon and of the planets were accurately and
systematically measured and registered. The problem has always been to
understand what the data meant. Particularly complicated are the motions of the
planets , which owe their name to the Greek word for tramp.
The Heavenly bodies, including the planets, are so far away that their
distances could not be measured in ancient times, with the exception of the
moon. What we measure, for each body, is the direction at which it appears as a
function of the time of the observation. The directions are given by two angular
co-ordinates. However, it was natural to think of the stars as points on a sphere
of very large, but arbitrary, radius, which was called a celestial sphere . Its
center is on the earth and its axis coincides with the rotation axis of the earth.
The circle cut on the celestial sphere by the plane through the earth equator is
called the celestial equator .
The annual motion of the Heavenly bodies appears to an observer on earth as
a rotation around a common center, the earth. The stars, as different from the
planets, do move on the celestial sphere, but keeping all the distances between
them invariable. For this reason they have been called fixed. We know now that
the stars are not fixed at all and that they are at very different distances. They
appear to be fixed because the distances are enormous. The most striking
(apparent) motion is the diurnal one due to the rotation of the earth on its axis. A
further apparent motion of the fixed stars, due to the revolution of the earth
around the sun, is a rotation with a period called the sidereal year . The sidereal
year is also the time taken by the sun, in its apparent motion, to return to the
same position relative to the fixed stars. As such it is almost, but not exactly,
equal to our common year. As we shall see, the moon and the planets have more
complicated apparent movements (which are combinations of their own and of
earth).
As for distances, Aristarchus of Samos (310–230 BC) developed a brilliant
method to extract the distances from earth to the moon and to the sun by angular
measurement. He found correctly that the distance of the moon is about 60 times
the radius of the earth. However, due to an insufficient resolution in the
measurement of the angles, he concluded that the sun is 20 times farther away
than the moon, rather than about 400 times as it is. This was enough to conclude
that, considering the moon and the sun to have the same apparent size, the real
size of the sun had to be enormous. Aristarchus concluded that his finding
confirmed that the sun must be the center of the system. He then found the
correct order of the distances of the five planets around the sun that are visible
with the naked eye around the sun, which was standing still at the center of the
system. However, at least to our knowledge, he did not fully develop a
quantitative model of the planetary system.
A powerful quantitative model was developed three centuries later by
Claudius Ptolemy (90–168 AD), who lived at Alexandria in Egypt. By that time
the idea that the heavenly bodies had to move with constant (in magnitude)
velocity on circles, or combinations of circles, being brought around on a system
of solid spheres, had become dominant, as Newton recalls with the words quoted
in the introduction to this chapter. His book, originally written in Greek and
titled “Mathematical syntaxes” came to us through its Arabic translation and is
universally known as Almagest .
The “planets” were seven: sun, moon, Mercury, Venus, Mars, Jupiter and
Saturn. Figure 4.2a shows the basis of the model. Earth is at rest at the center of
the system. The sun describes a circle around the earth. The path of the sun on
the celestial sphere, through the fixed stars, is the ecliptic . The motion of each
planet, like P in the figure, is more complicated. In a first approximation it is
described by a circular uniform motion around earth performed by the point C
and by a second circular uniform motion of the planet itself around C. The
former circle is the deferent the latter the epicycle . The two motions are
(approximately) in the same plane and their combination is a curve, called an
epicycloid , shown in Fig. 4.2a. Clearly, both the deferent and the epicycle are
different for different planets. The observed trajectory of the planet is the
projection of its epicycloid on the celestial sphere, taking into account the angle
between its orbit and the plane of the celestial equator, which is also somewhat
different for different planets. Notice that for the largest fraction of its period the
planet moves forward, from East to West. However, in correspondence with the
smaller loops of the epicycloid it moves, for some time, backwards. This is in
accord with observations.

Fig. 4.2 Motion of an external planet relative to a the earth, b the sun. Figures are approximate

Ptolemy calculated, on the basis of the available measurements, results of

centuries of observations, the radii of the primary and secondary circles (with the
solar orbit radius as unit) and the corresponding periods. He found however, that
this relatively simple model did not work, namely did not explain all the data. To
make it work he added two features.
1. The primary circle (deferent) of each planet is not centered exactly on
earth but in a point not very far from her and different from planet to planet. It is
called equant, because it makes the motion on the deferent uniform. We now
know that the equant is the empty focus of the elliptical orbit of the planet. We
shall understand in Sect. 4.3 how it works.
2. A number of tertiary and quaternary circles, all called epicycles.
The model of Ptolemy, though even not particularly simple, was able to
reasonably explain all the observational facts and would remain such till the
accuracy in the measurements of the planets positions will be improved by an
order of magnitude by Tycho Brahe (1546–1601).
We can notice that the period of the deferent in two cases (Mercury and
Venus) and of the (first) epicycle in the other three cases (Mars, Jupiter and
Saturn) are all equal to a sidereal year. We know that the orbit of the first two
planets is smaller, the orbit of the other three is larger than the orbit of the earth
(Fig. 4.2). Ptolemy did not notice this feature. In his model all the circles are
independent. This important discovery is due to Nicolaus Copernicus (1473–
1543).
 Another feature that is not explained by the model is why both Mercury and
Venus never depart much from the sun. The maximum angle between Mercury
and the sun is θ m = 22.5° and Venus and the sun θ m = 46°.
Let us now go back for simplicity to the model with only one primary and
one secondary circle. Let us change our reference frame by choosing the sun at
rest at its center. We assume that the earth moves uniformly on a circle around
the sun with the radius of the epicycle of the planet and that the planet P moves
uniformly on the circle centered on the sun and radius equal to the deferent
radius, as in Fig. 4.2b. The relative positions of earth and planet are exactly the
same as before, but the description is now logically simpler. In addition, the
reason for which the period on the epicycle is the sidereal year becomes obvious.
Figure 4.2 represents an external planet. The reader can easily verify that an
analogue explanation works for the internal planets by simply exchanging the
roles of primary and secondary circles.
In the heliocentric frame the reason why Mars and Venus , which have orbits
smaller than earth around the sun, cannot be very far from the sun when viewed
by earth is also clear, as shown in Fig. 4.3. This argument was already known to
the Greeks, in particular to Aristarchus .

Fig. 4.3 The orbit of an internal planet, a viewed from earth, b viewed by the sun

The heliocentric description we have sketched, in its modern form, is due to

Nicolaus Copernicus. He gave a preliminary version of his model in the “
Commnetariolus ” distributed privately to his friends in 1514, and the final one
in “ De Revolutionibus Orbium Caelestium” published in 1543, the year of his
death. Differently from Aristarchus, Copernicus developed a full mathematical
model able to explain the observational facts.
The Copernicus model, as we have presented it so far, looks much simpler
than the Ptolemy model. One can then ask why it took so long to be accepted.
The reason is that, as for Ptolemy before him, such a simple model does not
work. The main reason was that Copernicus still believed that the orbits had to
be circles or combinations of circles and the motions on them uniform. The
reason of the belief was dogmatic, rather than scientific: the heavenly bodies
being the creation of God, their motion must be perfect. The bodies must be on a
rotating sphere, because, in his words, the sphere in its rotation moves

on itself through the same points, it expresses its form in the simplest body,
in which it is impossible to find either a beginning or an end or distinguish
the points from each other.

The consequence was that, to agree with the data, Copernicus, as long before
him Ptolemy, had to introduce both the equant and a rather large number of
epicycles. Indeed, the Copernicus model, in the form he presented it, was not
less arbitrary than the Ptolemy model.

4.2 The Periods of the Planets and the Radii of Their

Orbits
As we have already mentioned, two planets, Mercury and Venus , in their
motion as seen from earth never go far from the sun. The maximum angle
between Mercury and the sun is θ m = 22.5° and Venus and the sun θ m = 46°.
The model of Copernicus allows us to calculate the radii of the orbits of these
planets. Here the model shows its superiority to Ptolemy.
From Fig. 4.4, which is drawn for Venus, we have

Fig. 4.4 The earth and the Jupiter orbits

(4.1)
 Notice that the condition is for the ratio of the radius of the planet with the
radius of the earth orbit. Indeed the latter is the natural unit in astronomical
measurements and is called astronomical unit (au). To be precise the
astronomical unit is the mean distance of the earth from the sun. We shall not
discuss the different methods to measure r E . We simply mention that the
problem of the scales of the distances is a central one in astronomy.
The value of the astronomical unit was not known even to Kepler. He was
able to determine a lower limit (on the basis of the parallax of Mars) as 1 au > 15
Gm. The first measurements were made at the beginning of the XVII century by
Giovanni Domenico Cassini (1625–1712) and by Edmund Halley (1656–1742),
who found values between 140 and 150 Gm.
The value known today is
(4.2)
From the above values of θ m we have for Mercury r M ≈ 0.34 au and for
Venus r V ≈ 0.72 au.
For the external planets, the three known to Copernicus, the argument is
similar, but now the radius of the orbit of the planet is larger than that of the
earth. Figure 4.4 gives the geometry. The Copernicus interpretation is that the
larger circle, the deferent, is the orbit of the planet, and the smaller one, the
epicycle, is the orbit of earth. Consequently the angular diameter under which
the latter is seen from earth is 2θ m. From the figure we see that
(4.3)
Already Ptolemy knew the angles for the three planets, θ m = 41° for Mars , θ
m = 11° for Jupiter and θ m = 6° for Saturn . Equation (4.3) gives for the radii of
their orbits r Ma ≈ 1.5 au for Mars, r J ≈ 5.2 au for Jupiter and r J ≈ 9.5 au for
Saturn.
Let us see how to extract the periods from the observational data. For that we
must take into account that the observations are done from a frame moving in the
solar system. This problem is solved a little differently for the internal and for
the external planet, as in the case of the radii of the orbits. For the sake of brevity
we shall consider only one external planet, for example Jupiter.
Consider the two situations represented in Fig. 4.5. In both of them the
relative positions of earth, sun and Jupiter is the same. It is also such, being the
three bodies on the same line, to be easily and precisely recognized. This is
done, for a given observer, by taking the date at which Jupiter crosses the
celestial meridian at midnight. The celestial meridian is the projection of the
local meridian on the celestial sphere .
Fig. 4.5 Two consecutive transitions of Jupiter on the celestial meridian

The intervals between two consecutive recurrences of the phenomenon are

all equal and called the synodic period. Consequently, we can average on several
measurements and increase the precision. The synodic period of Jupiter is
τ = 399 d. In this period Jupiter travels through the angle θ (Fig. 4.5), the earth
travels that plus a revolution, namely 360° + θ. The number of revolutions of
Jupiter per unit time n J = 1/T J , where T J is its period. Similarly for earth, n E
= 1/T E . We can then write that is and also

Question Q4.1. Find the equivalent expression for an internal planet.

Table 4.1 gives the values of the orbit radii in astronomic units and of the
periods of the first six planets as known to Copernicus and as it is today.
Table 4.1 Orbit radii and periods of the first six planets
Planet Orbit radius (au) Orbit radius (au) Period Period
Copernicus Modern Copernicus Modern
Mercury 0.376 0.387 87.97 day 87.97 day
Venus 0.719 0.723 224.70 day 224.70 day
Earth 1.000 1.000 365.26 day 365.26 day
Mars 1.520 1.524 1.882 year 1.881 year
Jupiter 5.219 5.203 11.87 year 11.862 year
Saturn 9.174 9.539 29.44 year 29.457 year

We see that the values known to Copernicus, in particular for the periods,
were already close to the modern ones. We add that the values that can be
extracted from the data of Ptolemy are quite similar too. A millennium of
observations before 150 AD did allow great precision.
4.3 The Kepler Laws
As we have seen the (almost) heliocentric Copernicus system was not much
simpler than the Ptolemy (almost) geocentric one. Both systems make use of the
equant. To be precise, the center of the Copernicus system is not the sun, but the
equant of the earth (what we now know to be the empty focus of her elliptical
orbit). In both cases, beyond a primary circle, several secondary and tertiary
ones were necessary to fit the data. Since his youth, Tycho Brahe (1546–1601)
started his study of the astronomical texts and his observations of the night sky.
He soon found out that neither the tables of Ptolemy nor those of Copernicus
were very accurate. Both of them were in contradiction with the facts. When he
was 17 year old he had the opportunity to observe a not very frequent
phenomenon, the conjunction of Jupiter and Saturn (the two planets appear very
close to each other). Brahe calculated the conjunction time predicted by the
Ptolemy tables finding it to be off by about one month (which is not really so
much considering it is based on observations 1400 old) and that predicted by the
Copernicus tables finding it off by several days (being an extrapolation over a
few decennia, the relative error of Copernicus is much larger than that of
Ptolemy). Brahe was now sure that a correct model of the Cosmos (then the solar
system) could be found by planning and performing a systematic series of
measurements as accurate as possible, rather than interpreting the classic texts.
The observations still had to be done with the naked eye because the
telescope, as a scientific instrument, will not be invented by Galilei until 1609.
One of his first instruments is shown in Fig. 4.6. The star under consideration
must be seen through two small holes (D and E in the figure) fixed at the
extremes of a bar that can rotate over the arc of a circle. The angle of the bar
relative to the vertical, defined by the plumb line AH, is measured with a
goniometer on a scale giving the arc minute. To increase the sensitivity the
instrument had to be large. The graduated circle was almost seven meters in
diameter. The instrument had to be robust and accurately built to reduce
systematic errors. The instrument was built of timber and was so heavy that
twenty men were needed to install it in a garden.
Fig. 4.6 Instrument of Brahe to measure the position of the stars

Somewhat later Brahe succeeded to be funded by king Frederic II of

Denmark and Norway for the construction of a big astronomic observatory on
the island of Hveen near Copenhagen, the Uraniburg observatory . The castle in
which the observatory was built had a rich library, bedrooms, kitchens and
dining rooms. Brahe designed, built and installed a dozen different instruments,
apt at various types of observation. For the next 20 years, at Uraniburg and later
in Prague, Brahe continued his systematic observations. Before Brahe the
angular resolution had not improved from Grecian times, being about 10′. He
was determined to improve down to 1′ or better. He gathered the data in a series
of tables, which became the database that allowed Kepler and Newton to solve
the problem of the heavenly bodies’ motions.
Johannes Kepler (1571–1630) started his studies in the school of Tycho
Brahe in 1600. He began by searching through a large amount of available data
to determine if he could find any simple relation. Tables such as Table 4.1
pointed to existence of a relation between orbit radii and periods. The larger the
radius the larger is the period. But, is there really a mathematically simple
relation? Kepler finally found it and published it in the book “Harmonice mundi
” in 1618. He writes with confidence:

initially I thought I was dreaming…but it is absolutely certain and exact

that the ratio existing between the periodic times of any pair of planets is
exactly the ratio of the mean distances [from sun] to the power 2/3.

We can do the calculations ourselves. Starting from Table 4.1 we obtain the
data in Table 4.2. We can easily understand Kepler’s pride and satisfaction when
he found such a simple relation. We know it as the 3rd Kepler law , because it
came 10 years later than the discovery of the first two. The first two laws regard
the orbits of a single planet, the third gives a relation between different planets.
Table 4.2 Ratios of the cubes of the orbit radii and the squares of the period for the first six planets
Planet r 3/T 2 (au3 d–2)
Mercury 7.64 × 10–6
Venus 7.52 × 10–6
Earth 7.50 × 10–6
Mars 7.50 × 10–6
Jupiter 7.49 × 10–6
Saturn 7.43 × 10–6

Let us now briefly see how Kepler established that the orbits of the planets
are not complicated combinations of circles, but, simply, ellipses. Its great
discovery was based on the study of a single planet, Mars. The choice fell on
Mars because its deviations from the predictions of both models based on circles
where larger than for the other planets. Its strange behavior was the object of
study of several astronomers, but its anomalies remained unexplained. Brahe had
taken Kepler as his assistant in 1600 and charged him with a solution to this
problem. Kepler worked on the problem for 6 years, in which partial successes
alternated to partial failures, wrong paths were followed and retraced back,
before reaching the solution that we know.
Kepler fully accepted from the start a heliocentric view with the guiding idea
that the orbits should be a simple curve around the sun, but not necessarily a
circle. The problem to find the curve was made difficult by the fact that the
positions of the planet, Mars in his analysis, were measured in a frame fixed to
the earth, which moves in a non-uniform and unknown motion around the sun. It
took several years to solve this first problem, to find accurately enough, the
motion of earth. We shall not describe here the various mathematical methods he
employed, some of which are really elegant. We simply state that he found that
the earth orbit is indistinguishable from a circle. However, its center is not the
sun and its angular velocity about the sun is not uniform. The dogma that had
resisted from Aristotle to Copernicus included was broken.
With reference to Fig. 4.7, d is the distance from the center of the sun to the
center of the circle and R its radius. From the data of Brahe, Kepler found that
d/R = 0.018. The angular diameter of the sun, as seen from earth, varies
periodically during the year between a minimum and a maximum. Kepler had
Brahe’s measurements for that. With the above value of d/R, Kepler calculated
the variations of earth sun distance during the year and the consequent variations
of the apparent sun diameter. He found his results in agreement with the data. He
gained confidence that he was on the correct path.

Fig. 4.7 Scheme of the earth’s orbit. First approximation by Kepler. Continuous line is a circle, dotted line
an ellipse; the difference between them is exaggerated

In retrospect we know now, and Kepler himself was to learn that in a while,
that this model of the earth orbit is not correct, because the orbit is an ellipse.
However, the eccentricity of the earth orbit is so small that the maximum
difference between the preliminary Kepler model and the true orbit was smaller
than the experimental uncertainty. To fix the orders of magnitude, the distance
NN′ is about one half of a per cent of R. In conclusion the error introduced in the
analysis by the preliminary model is irrelevant.
Having defined the geometry of the orbit, Kepler had to find the motion. He
did that using a trick invented by Ptolemy, and that we have already quoted, the
equant . This is the point Q in the figure, lying on the line joining the center of
the sun and the center C of the circle, at the same distance d as the sun but on the
other side. Then the angular velocity of the position vector from Q to the earth is
constant. It is called equant for this reason. We shall see soon why it works.
Kepler now knew the motion of earth in a reference frame in which the sun
stood still. He could then calculate the positions of Mars at all times. It was an
enormous amount of calculations (by hand obviously). Once more, he assumed
the orbit of the planet to be an eccentric circle and a uniform angular velocity
around an equant (different from that of earth). He calculated 40 points on the
Mars orbit and compared it with the Brahe data. The maximum disagreement
was only 8′, a very small one, but larger than the uncertainties in the Brahe
measurements. Kepler knew he could trust Brahe. The model had to be wrong.
Kepler had to find another curve. Finally, his enormous computing effort
showed the light. Suddenly, everything became clear: the curve is the ellipse.
The first two Kepler laws were found. Kepler continued his work finding the
parameters of the ellipse of the orbits of the other planets, including earth,
calculating their positions and finding them in agreement with the rich and
precise Brahe data.
We notice now that the reason why an eccentric circle had worked for the
earth and not for Mars is the relatively large eccentricity of its orbit, which is
0.09, which is five time larger than that of the earth.
He published his results in 1609 in his book Astronomia nova .
The three Kepler laws are:

1. The orbits of the planets are ellipses, the sun occupying one of their foci.

2. The position vector from the sun to the planet sweeps out equal areas in equal
times

3. The ratio of the squares of the periods of any two planets is equal to the ratio
of the cubes of their average distances from the sun.

We can now show the reasons that make the equant work in a first
approximation. Indeed, the reason is in the second Kepler law. Consider Fig. 4.8
where an ellipse, in fact much more different from a circle than the real cases, is
shown. The equant, which is the center of a circle that tries to represent the
ellipse, is just the empty focus of the ellipse. In Fig. 4.8 the areas SCD and SAB
are travelled in the same time by the planet and are equal for the second Kepler
law. Consequently the arc CD is longer then AB proportionally at its distance
from the sun. However, there is a second effect. A given path length on the orbit
appears from the sun to be smaller, in its angular span, when it is closer than
when it is farther, once more proportionally to the distance. The two effects, one
due to the law of the areas and the geometrical one are identical. Consequently,
if we look to the planet from the other focus, the former effect remains while the
second inverts and the two cancel each other.

Fig. 4.8 Geometry plus area law explain the equant

The contribution of Brahe had been a systematic and accurate experimental

work, the work of Kepler an ingenious and superb analysis of the data. Both
were needed to discover three simple laws, which were able to interpret all the
available data. The work was not yet complete however. The marvelous Kepler
laws were still purely phenomenological. A fundamental step was missing: their
dynamical interpretation, which was going to lead to universal gravitation, one
of the highest creations of human genius, the genius of Isaac Newton (1642–
1727).

4.4 The Newton Law

We begin by showing that a consequence of the Kepler laws is that the angular
momentum , L, of any planet P about the position of the sun is constant. With
reference to Fig. 4.9, let r be the position vector, v the velocity and m the mass
of the planet. Its angular momentum is then
Fig. 4.9 The elementary area swept by the radius vector in dt

(4.4)
L is always perpendicular to both r and v, hence to the plane of the orbit that
is constant for the first Kepler law. Hence the direction of L is constant.
In addition L is constant also in magnitude for the second law. Indeed,
consider the area dA swept by the position vector in the time dt, which is the area
of the triangle in Fig. 4.9. Two of its sides are v dt and r. Remembering the
geometric meaning of the vector product we have

(4.5)
or

(4.6)
The quantity dA/dt is the area swept by the position vector in the unit of time
and is called areal velocity . It is constant for the second Kepler law . We
immediately recognize that the second member is proportional to the magnitude
of the angular momentum, namely

(4.7)
The areal velocity being constant, the magnitude of the angular momentum is
constant too. In conclusion the angular momentum vector about the sun is
constant. On the other hand, the planet is certainly subject to a force, because it
accelerates, but this force does not vary the angular momentum about a point
fixed in an inertial frame. Consequently, its moment about that pole must be
zero, namely its direction must be parallel to the position vector from the sun to
the planet. It must be towards the sun because in a curved motion the force is
always directed on the side of the curvature center.
In conclusion, the force on every planet must be directed towards the sun.
The conclusion suggests, better forces, us to think the sun to be the source of the
forces acting on all the planets.
We now consider the magnitude of the force. The symmetry of the problem
suggests choosing a reference frame with origin in the sun and polar co-ordinates
with an arbitrary polar axis. Let r be the magnitude and θ the azimuth of the
position vector of the planet r. Data show that the motion of the planets does not
slow down through the centuries, hence the force should be conservative.
Having just shown that it is also central, for the theorem we demonstrated in
Sect. 2.15, its magnitude cannot depend on θ, but depends only on the distance
from the center of the sun r (Fig. 4.10).

Fig. 4.10 The reference frame to study the motion of the planet

To make the demonstration as simple as possible we shall assume the orbits

to be circumferences rather than ellipses. In Sects. 4.11 and 4.13 the problem of
the ellipse will be treated exactly.
If the motion is circular, the area law implies that the angular velocity ω is
constant. The force should be the centripetal force of such a motion

(4.8)
where m is the mass of the planet and T is its period. The third Kepler law
states that
(4.9)
where K S is the proportionality constant, the same for all the planets of the
solar system (but not necessarily for other systems) and that, substituted in
Eq. (4.8), gives

(4.10)
We have found two fundamental properties of the force: 1. It is inversely
proportional to the distance from the sun, which is its source, 2. is proportional
to the mass of the planet. We now show the third property: the force is
proportional to the mass of its source. To find it, observational data on systems
similar to the solar one, but with a different central body, are needed. Newton,
had already compared the force exerted by earth on bodies on its surface, namely
the weight, and on the moon, as we shall see in Sect. 4.5. He had established
that, taking into account the difference in the distances from the center, the force
is the same. The characteristics of the gravitational force are universal.
Two small “solar systems” were known, Jupiter with its four principal
satellites (Io, Europa, Ganymede, Callisto), which had been discovered by
Galileo Galilei (1564–1642) (we shall tell of the discovery in Sect. 4.9), and
Saturn with its two larger satellites, which had been observed by Christiaan
Huygens and by Giovanni Domenico Cassini . These observations had
established the validity of the 3rd Kepler law for the systems (in both cases
more, smaller, satellites were discovered in recent times with the space
missions).
Gravity, Newton concluded, is of all the planets and satellites, and continued:

And since all attraction (by Law III) is mutual, Jupiter will therefore
gravitate towards all his own satellites, Saturn towards his, the earth
towards the moon, and the sun towards the primary planets.

To be concrete, consider one of the Jupiter satellites , Callisto. Jupiter is

attracted by the sun with a force proportional to its mass and attracts Callisto
with a force proportional to the mass of Callisto. For the 3rd Newton law
Callisto attracts Jupiter with a force equal and opposite. But, this latter force has
the same characteristics as the force that Jupiter receives from the sun, including
being proportional to the mass of Jupiter. We can conclude that the force that
Jupiter exerts on Callisto is proportional to the mass of Jupiter (beyond that of
Callisto). The property is general, namely the gravitational force between any
two (point-like) objects of masses m and M is proportional to the product of the
masses. We write

(4.11)
where G N is a universal constant, the Newton constant , that we shall soon
determine. This equation gives the magnitude of both the forces of mass M on m
and of m on M. Their directions are equal and opposite. If r is the position vector
from M to of m and u r is its unitary vector, the force exerted by M on m is

(4.12)
This is the Newton law of universal gravitation .
We first observe that, as written, the law is valid for point-like objects. In the
cases of the solar system and in the systems of Jupiter and Saturn, all the bodies,
sun included, can be considered as points because their distances are always very
much larger than their diameters. However, also two extended objects, for
example two bricks one close to the other, attract gravitationally one another. To
find the force we must ideally divide each body in infinitesimal parts. Every pair
of infinitesimal elements attracts each other with the force of Eq. (4.12) where r
is the position vector of one element relative to the other and the masses are
those of the two elements. The total force is obtained by taking the vector sum
(integrating) of all the pairs. There is certainly a case in which such an
integration is needed, namely the weight. Indeed, we state that the weight of an
object on the surface of earth is the gravitational force of the earth considered as
a point in its center. Why is this possible? The answer is in Sect. 4.6.
A second observation is on the masses in the Newton law Eq. (4.11). They
are clearly gravitational masses . However, in our demonstration we have started
from Eq. (4.8) where the mass is the inertial one. As we have seen in Sect. 2.9,
the equality of inertial and gravitational masses had been established by the
experiments of Galilei, which Newton had repeated. However, the experiments
had been done on terrestrial bodies and the question arises: does the same
relation hold for celestial bodies? Newton showed this to be true considering the
system of Jupiter and its four Galileian satellites. The system is a small replica of
the solar system, but is part of the solar system too. Observations had shown that
the satellites perform “exceedingly regular motions”. The radiuses of the orbits
about Jupiter and the periods had been measured. The periods turned out to be
proportional to the 3/2rd power of the orbits radiuses. Consequently, the force
exerted by Jupiter is inversely proportional to the distance. Suppose now the
ratio between gravitational and inertial mass of Jupiter and any of its satellites,
Callisto for example, to be different, say as

where ε is a positive small number. Then, Newton argues, the forces of the
sun on Jupiter and on Callisto, at equal distances from the sun, will differ by ± ε
also, and this would have an effect on the orbit of Callisto about Jupiter. The
calculation of the effect needs to solve a three-body problem, Jupiter, Callisto
and the sun, which cannot be done analytically. But Newton was able to find
that, if the forces of the sun on Jupiter and Callisto would differ in a certain
proportion, then the distances of the center of the orbit of Callisto (call it r CS )
about the sun and the center of Jupiter (r J ) from the sun would differ “nearly”
as the square root of the same proportion “as by some computations I have
found”, namely,
 He writes

Therefore if, at equal distances from the sun, the accelerative gravity (he
means the gravitational force) of any satellite towards the sun were greater
or less than the accelerative gravity Jupiter towards the sun but by one
1/1000 part of the whole gravity, the distance of the centre of the satellite’s
orbit from the sun would be greater or less than the distance of Jupiter from
the sun by one 1/2000 part of the whole distance; that is the fifth part of the
utmost satellite (Callisto) from the centre of Jupiter; an eccentricity of the
orbit which would be very sensible. But the orbits of the satellites are
concentric to Jupiter, and therefore the accelerative gravities of Jupiter, and
of all its satellites towards the sun, are equal among themselves.

Newton adds that if the ratios of gravitational to inertial mass of the earth,
, and of the moon, , would be different, the above-described
effect should be present and a deformation of the moon orbit should be
observable. Today, the moon-earth distance is measured with extreme precision
with LASER ranging techniques. In 1969 the Apollo 11 astronauts and later
other lunar missions deployed on the surface of the moon systems of mirrors
able to reflect back a LASER pulse sent from earth. The measurement of the
round-trip time of the pulse gives the moon distance with a few millimeter
precision as a function of time. The extremely sensitive technique did not detect
any effect, providing the very low upper limit

We now come back to the universality of the Newton law. If it is so, the
constant G N must be the same in any circumstance and is one of the
fundamental constants of physics, called the gravitational Newton constant . At
laboratory scale, between everyday life size objects, the Newton law is very
small and difficult to measure. This was first done by Henry Cavendish (1731–
1810) (see Sect. 4.7) leading him to a laboratory measurement of G N (which is
also called a Cavendish constant ).
The universality of the Newton law needs to be verified experimentally. This
has been done at all the length scales in many different conditions, finding it
valid. We shall discuss a few examples further in the chapter. However, a limit
of validity exists, as we shall see.
Equation (4.12) is mathematically very simple and symmetric in its elements.
It interprets a huge amount of phenomena, from the motion of planets to the free
fall of objects on earth, from the motion of the satellites, to that of the stars and
the galaxies. The expression shows us how Nature can be described in its most
fundamental aspects in simple and elegant mathematical form. The harmony of
the world that up to the Middle Age, and to Copernicus, was believed to be
substantiated in the existence of a mechanism of solid spheres, symmetric
objects, that rotate uniformly (simple motion), comes back, in an abstract form,
in the harmony, so to speak, of the physical law.
We finally come back on the constant K S in Eq. (4.9). From Eq. (4.12) we
can write, for the solar system

(4.13)
We see that the constant depends on the mass of the sun, namely the mass of
the central body. It is not universal. For example for the Jupiter system it is the
mass of Jupiter, for the earth-moon systems it is the mass of the earth, etc.

4.5 The Moon and the Apple

If Eq. (4.12) is universal, the force that earth exerts on the moon, the centripetal
force corresponding to her motion, must be the same as the force she exerts on a
body on her surface, for example an apple, which is its weight. In particular the
constant G N should be the same. As Newton himself recalls, in 1665 he started
to ask himself this question. He developed the following argument. Indeed, in
her circular motion the moon continuously falls accelerating towards earth. This
is simply another way to look at centripetal acceleration.
Suppose that the moon is at the point A of her orbit, as in Fig. 4.11, at a
certain instant. In the figure we have taken a reference frame with the origin O in
the center of earth and y-axis directed towards the moon in the considered
instant. After a certain short time, say after one second, if no force were present,
the moon would have moved to point B. On the other hand, if the moon would
be abandoned still in B, she would fall in a second, under the action of gravity,
from B to P. Point P is at the same distance r from the center of earth as A. Let
us calculate the drop h, taking into account that the angle θ is very small. The
Pythagorean theorem for the triangle ONP gives
Fig. 4.11 How the moon falls

If θ is infinitesimal, h 2 is an infinitesimal of second order and can be

neglected. We can also consider x equal to s and write

(4.14)
To evaluate the displacement of the moon in one second we can use the
proportion s:2πr = 1:T, where T is the period of the moon revolution, T = 27.3
d = 2.4 × 106 s and r = 3.8 × 108 m. We have s = 2πr/T ≈ 1000 m and
In a second the moon falls a little more than a millimeter.
We now compare this with the drop length of an object on earth, the famous
apple for example, which is

(4.15)
The ratio of the two drops in one second is equal to the ratio of their
accelerations. The latter, if the Newton law is valid, should be in the inverse
ratio of the squares of their distances. The ratio of the drops is
. Newton knew that the ratio of the distance of the
moon is about 60 times the radius of the earth and what we have just found is
about 602.
However, Newton had still the problem that we already mentioned. While
moon and earth can be considered as points, considering their large distance, for
what reason we should consider the apple, on a visually flat ground, should be
attracted towards a point 6380 km under the ground as if all the mass of earth
would be concentrated there?
 This is a “miracle” true only for forces inversely proportional to the distance
square. In the next section we shall prove the following theorem: the force
exerted by a homogeneous spherical mass in any point outside its surface is
equal to the force that would be exerted if all the mass were in a point at its
center.
Newton did not publish any result until he had made everything clear,
complete and perfect, in the Principia published in 1687.

4.6 The Gravitational Force of the Homogeneous Sphere

We shall calculate the force of a sphere of mass M on a point-like particle of
mass m outside the sphere at a distance r from the center. We assume that the
density of the sphere, if variable, depends only on the distance from the center
(spherical symmetry). We shall prove that the force is equal to that which would
exert all the mass M concentrated in the center.
We start by observing that is enough to prove the thesis for a spherical shell
of infinitesimal thickness. Indeed if it is true for one shell it is also true for the
sphere, which can be considered as composed of shells with the same center.
Consider the spherical shell, of radius R and center O, shown in Fig. 4.12,
having radius R and on it the ring AA′ limited by cones with their vertex in O and
semi-vertex angles θ and θ + dθ. Let ϕ be the semi-vertex angle of the cone with
vertex in P and the ring AA′ as base.

Fig. 4.12 Elements for calculation of the force of a spherical shell on an external point

All the elements of the ring AA′ are at the same distance from P and
consequently they exert on P forces, call them d 2 F, equal in magnitude, but not
in direction. The symmetry of the problem tells us that the resultant of these
forces, d F, is directed as OP. The contributions normal to it cancel each other.
The component in the direction OP of the force is proportional to the mass of the
element, to cosϕ and inversely to the square of the distance s 2. The resultant of
the forces on m in P due to the ring being

where dM is the mass of the ring. Now the mass of the ring is to the mass of
the shell as the area of the ring is to the area of the shell:

which gives us dM = (M/2)sinθ dθ. The force of the ring on the mass m is
then

(4.16)
The force of the shell is the integral of this expression for θ varying from 0 to
π, namely

(4.17)

Both s and ϕ are functions of the integration variable θ. It is convenient

however to express everything as functions of s. The Carnot theorem applied to
the triangle OAP gives

(4.18)
We differentiate the first equation, remembering that r and R are constant,
obtaining

We substitute this expression and the second Eq. (4.18) in the integral of
Eq. (4.17) and take into account that now the variable is s and the limits must be
changed in accord, obtaining

The integral does not present difficulties. The indefinite integral gives

which, evaluated in its limits, gives 4R. In conclusion the force of the shell
on a point P of mass m is
 (4.19)
which is, in particular, independent of the radius R of the shell. This proves
the theorem.
Consider now a point P of mass m inside the shell. What is the force on P
exerted by the shell? The reasoning remains exactly the same, but for the limits
on the integration on s. Now the angle θ varies between 0 and 2π and
correspondently s between R + r and R − r. The definite integral is zero. The
gravitational force exerted by a spherical shell on a point inside it is zero. This is
another property of the inverse square law forces.
Newton gave another proof of the last property using a simple geometric
argument. Consider point P inside the shell as shown in Fig. 4.13 and the cone
with vertex in P of very small vertex angle. The two napes intercept on the
shell’s two surfaces ∆S 1 and ∆S 2. As the density is constant, the masses of the
two surfaces are proportional to their areas. The latter are proportional to the
squares of their distances from P, say and . But the forces they exert in P
are proportional directly to the masses and inversely to the square distances. The
two forces are equal in magnitude. As their directions are opposite, their
resultant is null. As the shell can be divided in pairs giving null contribution, the
resultant is zero.

Fig. 4.13 The geometry to calculate the gravitational force of a spherical shell on an internal point

4.7 Measuring the Newton Constant

The Newton gravitational force Eq. (4.12) exerts between every pair of point-
like, or spherical, masses. It is important to control experimentally its validity
not only at the astronomical scales, but also at the laboratory scale. The
laboratory experiments are difficult because the force is, at these scales, very
small. Any disturbance such as small air currents, spurious electric forces, the
movement of the experimenter itself, is a possible cause of errors and must be
eliminated.
However, if we want to know the Newton constant , we must measure the
force between two known masses at a known distance. In the case of the
heavenly bodies in fact we do not know a priory the masses, but we infer them
from the Newton law.
The gravitational force was first measured by Henri Cavendish (1731–1810)
in 1798. His experiment is shown schematically in Fig. 4.14. A rigid metal bar
suspended on a very thin metal wire, carries to equal lead spheres at equal
distance from the wire. The system is in equilibrium and free to rotate about the
wire. This type of arrangement is called torsion balance and will be further
discussed in Sect. 8.9.

Fig. 4.14 The Cavendish experiment

Two more larger and heavier equal spheres, of mass M, are arranged
symmetrically, each at the same distance from one of the small ones.
Consequently each of the large spheres attracts the small one nearby with an
(equal) gravitational force. The arm of the couple is the distance between the
centers of the small spheres and can be accurately measured. The moment of the
couple induces a rotation to the bar. The wire reacts with an elastic torsion
moment, which is proportional to its rotation angle. The equilibrium is at an
angle at which the torsion moment and the moment of the gravitational couple
are equal. Hence, the measurement of this angle gives the moment of the couple
and, the arm being known, the forces.
The rotation angle is measured with the technique of the optical lever . A
narrow light beam is sent to a very light mirror, fixed to the wire. The mirror
reflects the beam on a scale located at a certain distance. The device is very
sensitive. Even a very small change in the orientation of the mirror causes a
sizeable movement of the light spot on the scale. Indeed, the moments are very
small. The wire must have a very small elastic constant and consequently be
very thin, but still capable of holding the weight of the small spheres and bar. All
the apparatus must be closed in a container to avoid air currents. The presence of
electrostatic charges must be avoided, etc.
The value of the gravitational constant obtained by Cavendish was
(4.20)
The present value is
(4.21)
To have a quantitative idea, consider that the large spheres of Cavendish had
a mass M = 158 kg, the small ones m = 0.73 kg and that the distance between
one small and one large was r = 0.225 m. The two forces to be measured are
about 10–7 N. This is about the weight of a hair.

4.8 The Gravitational Field

We interrupt in this section our discussion of experimental proofs of the Newton
gravitational law, to discuss an important property of the gravitational, and of the
other fundamental forces. Namely they are action at a distance. Other examples
are the electric force, which operates between electrically charged bodies, and
the magnetic force, for example between a magnet and a piece of iron.
In all these cases an extremely useful concept is the field of force or simply
field (gravitational, electric, magnetic field).
Consider the gravitational force exerted by the earth, on different objects. It
depends on the mass of the object (is proportional to it) and on the position of
the object. If we consider two particles of different masses in the same position
and divide the forces acting on each of them by its mass, we find the same result.
This vector function of the position is the gravitational field .
The gravitational field generated by a distribution of masses is a vector
function of the position. It is equal to the force acting on the unit mass in that
position. The masses giving origin to the field are called the sources of the field .
In particular to describe the field of the earth we can take a reference frame
with origin in the center of the earth. Consider a point P at the position vector r
with the unit vector u r . If we put a mass m in P, it feels the force

(4.22)
where M is the mass of the earth. The gravitational field is the vector
function of the position
 (4.23)
This expression is valid for points outside the earth in the approximation of
earth being spherical and with a spherically symmetrical distribution of masses.
The physical dimensions of the gravitational field are a force divided by a mass,
hence the dimensions of the acceleration. As a matter of fact, it is just the gravity
acceleration g.
The concept of field eliminates from our reasoning the idea of action at a
distance. We can think as follows. The earth, or any distribution of masses,
creates in all the space around it a physical entity, the gravitational field, which
extends, even if with decreasing intensity, to infinity. The field exists
independently of being perceived as a force. But if we place in a point of the
field a test body of mass m, it will feel like a force equal to the product of m
times the gravitational field in that point. By means of the field the gravitational
action becomes local.
We can now consider the potential energy of our test mass in the field of the
earth. Defining the potential energy to be zero at infinite distance, we have

(4.24)
The physical meaning is: the potential energy of the mass m in the point P is
the work to be done against the forces of the field to move the mass m from
infinity to P.
Obviously the potential energy, as the force, is proportional to m. If we
divide it by m we find a function of the position, independent of the body

(4.25)
This function is the gravitational potential . The relationship between
potential and field is the same as between potential energy and force. The
gravitational potential in a point is the work to be done against the forces of the
field to carry from infinity to that point a unitary mass. The physical dimensions
of the gravitational potential are a velocity squared. It is measured in m2/s2.
Consider now our mass m moving on a circular orbit of radius r with velocity
υ. It might be for example our moon. There is a simple relation between kinetic
and potential energy. Recalling that υ = 2πr/T, where T is the period, the kinetic
energy is

and for the 3rd Kepler law Eq. (4.13)

 (4.26)
This result, valid for circular orbits, is that the kinetic energy is one half of
the potential energy in absolute value. Consequently, the magnitude of the
gravitational potential in the points of the orbit is equal to the square of the
velocity of the body on that orbit
(4.27)
This expression will be useful in Sect. 7.13.
To appreciate the orders of magnitude, consider the motion of the earth
around the sun. The velocity is For Eq. (4.27), the potential of
the field of the sun in the points of the earth orbit is ϕ ≈ 109 m2 s–2.
We have already mentioned that limits of the validity of the Newton law
exist, when it must give place to general relativity. More precisely, the effects
that are in contradiction with the Newton law, and that are explained by general
relativity , are of the order of the gravitational potential compared to the square
of the speed of light, namely ϕ/c 2. Considering that c 2 ≈ 9×1016 m2 s–2, these
effects are usually very small (of the order of 10–8 on the earth orbit), but can be
detected with high precision observations, as in the case of the anomalous
precession of the Mercury perihelion (see Sect. 4.11). The effects become large
at very high gravitational potentials, near massive and compact objects, like
black holes.
The graphic representation of the gravitational field is very useful to have a
visual idea of its main features. It is done with the lines of force and with the
equipotential surfaces.
A line of force is drawn as shown in Fig. 4.15. We start from a point, 1 in the
figure, where we evaluate the vector of the field. Then we make a small step δ s
in the direction of the field, reaching point 2. We calculate the field in this point
and proceed another step as above, etc. In this way we obtain a broken line. It
becomes a field line for δs tending to zero. It is a continuous curve, in all the
points of which the field is tangent. Obviously the field lines are infinite in
number. However, there is only one line through any given point. If they were,
say, two, the field should have had two directions contemporarily. Graphically,
we draw a number of lines, which is enough to see the features of the field.
Fig. 4.15 Construction of a line of force

The equipotential surfaces are the loci of the points that satisfy the equations
ϕ(x,y,z) = constant, one for each value of the constant. These are infinite in
number too. It is convenient to draw a set of surfaces at constant steps of the
potential. An analogy are the geographic maps in which the level curves are
drawn every, say, one hundred meters of elevation. In the regions where the
level curves are denser, the elevation varies more rapidly and the slope of the
surface is steeper. The situation is analogous for equipotential surfaces.
Figure 4.16a shows some lines of force and equipotential surfaces for a
spherical mass M. The lines of force are radial and point to the mass, because the
force is attractive. The equipotentials are spherical and become denser getting
closer to the mass, which is the source of the field.

Fig. 4.16 Equipotentials and field lines for a a spherical mass M, b two masses one twice the other

Figure 4.16b represents the field originated by two spherical masses, one
double the mass of the other. In every point the field is the vector sum of the
fields of the two masses taken separately, the potential is simply the sum of the
potentials. Notice the “saddle” point on the line joining the two centers. Here
there is a minimum moving in that direction, a maximum moving
perpendicularly to it.
One sees that the lines of force are always perpendicular to the
equipotentials. This is a general property. Indeed, suppose we are moving with
the infinitesimal displacement d s. The potential difference between the two
points is If the displacement is on the equipotential, dϕ = 0 by
definition, hence G must be perpendicular to d s. The lines of force that have the
direction of G are perpendicular to the equipotential.
If we call G s the projection of G on the direction of the displacement we can
write
(4.28)
which can be also written as

(4.29)
We read this expression as: the component of the field in a given direction is
the directional derivative of the potential in that direction. Directional derivative
is just the name of the derivative in Eq. (4.29), it is the rate of change of the
function in that direction. As we have just seen the directional derivative is null
for directions on the equipotentials.
Consider infinitesimal displacements as those in Fig. 4.17, which are in
different directions but all leading from the equipotential ϕ to ϕ + dϕ. The
directional derivative is different for each of them because dϕ is the same and ds
is different. The derivative is a maximum when the direction is normal to the
surfaces because ds is there a minimum. The vector having the magnitude of the
maximum directional derivative and the direction of the normal to the
equipotential towards increasing potential is called the gradient of the potential.
Its symbol is grad ϕ. In conclusion we have

Fig. 4.17 Different steps between the same equipotentials

(4.30)
If we think of the level curves of a geographic map, the gradient is directed
as the line of maximal slope of the ground; its magnitude is greater the greater is
the slope.
On earth, the equipotential surfaces are materialized by the surfaces of the
lakes and of the seas (neglecting the waves).
We now see how to calculate the gradient starting from the potential. We
start from Eq. (4.28) and use the total differential theorem
 (4.31)
where dx, dy and dz are the Cartesian components of δ s. It immediately
follows that the Cartesian components of the gradient are the partial derivatives
of the potential

(4.32)
Obviously, similar relations exist between gravitational potential and
gravitational force of a mass m. It is just a matter of multiplying by m,
(4.33)
and

(4.34)

4.9 Galilei and the Jovian System

G. Galilei (1564–1642) was the first human to explore scientifically the sky
using the telescope , which he had developed. As a matter of fact, combinations
of two lenses put one after the other at a certain distance had existed for at least
30 years. The first written mention is in 1589, by Giovanni Battista Della Porta
(1535–1615). At the beginning of the XVII century telescopes were built in the
Netherlands by eyeglasses manufacturers. They were toys sold in exhibitions at
low prices. Galilei new of the Dutch telescope in 1609. He quickly envisioned a
way to transform the device into a scientific instrument and immediately started
his experimental work, without a solid theoretical basis. What is known today as
geometrical optics was developed only in 1611 by Johannes Kepler (motivated
by the desire to explain how the telescope works). Lenses had already been
produced since the XIII century, but their quality was not adequate for a
scientific instrument.
An important property of the telescope is angular magnification , which is
the ratio between the angle under which an object is seen through the telescope
and the angle under which it is seen with the naked eye. The second property is
the resolving power , namely the ability of the telescope to resolve, to see
separated, two point images very close one to the other. To increase both
properties the diameter of the objective lens (the one farther from the eye) must
be increased. However, the larger the lens, the more difficult is its production
without any defect. With a series of improvements, and the help of the Venetian
glass makers, Galilei developed the technique to the point that he could build a
telescope with magnification 10 and, some time later, one with magnification 30,
with lenses of perfect optical quality. With this magnification the light reaching
the eye in 302 = 900 times as with the naked eye.
Galilei published his first observation in the booklet “Sidereus nuncius ”
(astronomical notice) in 1610. In addition, the logbooks of his observations have
come down to us. One of his great discoveries was that around Jupiter’s four
satellites orbit, making a small replica of the solar system. A view of the system
with a modern telescope is shown in Fig. 4.18. Let us see how he describes his
discovery in the Sidereus Nuncius.

Fig. 4.18 Jupiter and his satellites. Image © NASA

On the night of the 7th of January 1610, looking to Jupiter , Galilei observed
three small “starlets”. They attracted his attention because they were perfectly
aligned between them and with Jupiter and on the ecliptic . He did not correlate
the starlets with Jupiter, thinking they were fixed stars in the background. He
took note of their positions in the logbook, as we try to reproduce in Fig. 4.19a.

Fig. 4.19 Sketches of the Galilei observations in January 1610 in the nights of a 7th, b 8th, c 10th, d 13th

The following night he repeated the observations and noticed that the relative
positions had changed, as in Fig. 4.19b. He thought the change to be due to the
movement of Jupiter relative to the stars, that he believed to be fixed, with some
doubts, because the motion did not match the calculations. He anxiously waited,
as he writes, the following night, but his hope was frustrated, because all the sky
was cloudy. The night of the 10th the stars were only two and had again changed
position, but still on a line, as in Fig. 4.19c. The third one, he thought, should be
hidden by Jupiter. Galilei had no more doubts. He writes (translated by the
author):

my perplexity changed to astonishment and I became sure that the apparent

movement was not of Jupiter, but in the stars I observed; hence I decided to
continue my investigation with increased attention and scrupulosity.

The 13th he saw for the first time the fourth satellite, which had entered the
field of view of the telescope, as in Fig. 4.19d.
After several more nights of observations he published the discovery,
together with other important ones on the moon and the Milky way in the above
quoted book in March 1610.
The next task was the measurement of the periods. The measurement was
extremely difficult, as much that Kepler had declared it impossible, because the
images of the four starlets were indistinguishable. Galilei understood that the
precision on his measurements of the angular distances from the center of Jupiter
had to be improved. He had measured them “by eye” with a precision of better
than one arc minute (1/60°). It was not enough. He developed the micrometer ,
with which he was to measure the positions with a precision “better than very
few arc seconds” (one arc sec = 1/3600°).
Galilei continued his systematic measurements for several years, but already
in 1611 he had been able to identify each of the satellites and to calculate their
period and the apparent diameters of the orbits. In Fig. 4.20 we report a subset of
his measurements made in spring 1611, as taken from his hand notes. For
simplicity they are for the two more external ones, Callisto and Ganymede. The
planes of the orbits are almost on the line of view from earth. Consequently, if
the orbit is an ellipse (or, in particular a circle) the motion appears as sinusoidal
functions of time. With a computer it is today easy to find the sinusoid that best
interpolates the data, the ones shown in the figure. Clearly, the data are in
agreement with the hypothesis. The procedure also gives us a value for the
amplitude and the period. Galilei had no computer and made his calculations by
hand.
Fig. 4.20 The distances from Jupiter of his two farther satellites as measured by Galilei in spring 1611. The
sinusoids are from my calculations

Table 4.3 reports the periods as measured by Galilei and how they are known
today. One sees that his measurements were quite good.
Table 4.3 Periods of the Jupiter satellites (in days)
Io Europa Ganymede Callisto
Galilei 1.76 3.53 7.16 16.3
modern 1.77 3.55 7.17 16.75

Accurate measurements of the apparent amplitudes are more difficult. Notice

that these quantities are measured relative to the apparent diameter of Jupiter,
namely they are, say, n = r/r J . Table 4.4 reports the values of n as measured by
Galilei in subsequent years, showing how the precision is increasing,
approaching the presently known values.
Table 4.4 Angular radii of the orbits relative to the radius of Jupiter
Io Europa Ganymede Callisto
1610? 3.5 5.7 8.8 15.3
1611 3.8 6.2 8.4 15
1611? 4 7 10 15
1612 5.7 8.6 14 “almost 25”
modern 5.58 8.88 14.16 24.90

The Jovian system is a small solar system. Is the third Kepler law verified?
Galilei did not check that, but Newton did. From the data in the two tables, we
can do it ourselves obtaining the following table (Table 4.5).
Table 4.5 The 3rd Kepler law in the Jovian system
Galilei Modern
T (d) n = r/r J n 3/T 2 T (d) n = r/r J n 3/T 2
Io 1.76 5.7 59.8 1.77 5.91 65.8
Europa 3.55 8.6 50.5 3.55 9.40 65.9
Ganymede 7.16 14.0 53.5 7.16 14.97 65.8
Callisto 16.3 24.9 58.1 16.69 26.33 65.5

The 3rd Kepler law is satisfied, better obviously by the modern data, for
which the experimental uncertainties are smaller.
We can finally check the universality, namely if the gravitational constant
has the same value in the Jovian and in the solar systems. We check if
Eq. (4.13), namely, is valid with the same G N , where
now M is the mass of Jupiter, r and T are orbit radium and period of any of the
satellites. For that we need absolute values. We now know the distance of Jupiter
and then the radii r. The Jupiter mass has been evaluated from his perturbing
effects on the other planets. With these values we find that, indeed, the
gravitational constant is the same.

4.10 Galaxies, Clusters and Something Else

In this section we shall give two examples of structures of larger scales than the
solar systems. The Newton law is valid also at the largest scales. However, we
shall also see that the same law gives us evidence that the mass of the Universe
is made for its largest fraction of components that are not visible, because they
do not emit or absorb light. This is the so-called dark matter , whose nature we
do not know.
A first example is shown in Fig. 4.21. It is a globular cluster , a system
containing millions of stars, which are very old, having an age comparable with
the Universe itself. The effect of the gravitational force keeping those stars
together is spectacular.
Fig. 4.21 The global cluster NGC 2808. Image © ESA

Figure 4.22. shows the image of a spiral galaxy , a system of hundreds and
millions of stars kept together by the gravitational attraction. All this enormous
system is rotating, as evident by the image. The angular momentum of the huge
gas cloud from which the galaxy originated billions of years ago remained
constant.

Fig. 4.22 The galaxy M74 from the Hubble Space Telescope. Image © NASA
 Let us more closely to the rotation. Let us start by considering how the
orbital velocity υ(r) of a body of mass m orbiting around a central body of mass
M (like a planet around the sun) varies with the distance from the center r.
Assume for simplicity a circular orbit. We state that the centripetal force must be
equal to the gravitational attraction

(4.35)
or
(4.36)
The velocity is inversely proportional to the square root of the distance from
the center. The validity of the law can be tested on the planets of the solar
system.
Five planets are visible with the naked eye and have been known since
ancient times. In order of distance from the sun, including earth, they are:
Mercury, Venus, Earth, Mars, Jupiter and Saturn. In 1781, William Herschel
(1738–1822) discovered a “star”, the image of which in the telescope had a non-
zero diameter. It was the seventh planet, Uranus. The object had been already
observed by Galilei and by more astronomers in the following years. They had
not recognized it as a planet, due to the limitations of their telescopes, but had
measured its coordinates. On the basis of these measurements, Herschel could
reconstruct the parameters of the orbit of Uranus. The motion of Uranus showed
some anomalies, when compared to the Newton law predictions. These were
interpreted in 1846, independently by Urbain LeVerier (1811–1877) and by
Johan Couche Adams (1819–1891), as possibly due to an eighth planet. When
his calculations were complete, LeVerrier sent a letter, with the calculated
coordinates, to the astronomer Johanne Grottfried Galle (1812–1910) in Berlin,
asking him to verify. The following night, Galle found Neptune within 1° of the
predicted position. Similarly, in 1930 Pluto was discovered, having its existence
predicted from the anomalies of the Neptune motion.
Figure 4.23 shows the orbital velocity of the planets as a function of their
distance from the sun. Equation (4.36) is fully satisfied.
Fig. 4.23 Inverse square root dependence of orbital velocities of the planets

Consider now the galaxy, a typical one, shown schematically in Fig. 4.24.
The image shows that its luminosity decreases for increasing distance r from the
center, till it disappears. This means that the star density decreases departing
from the center. We indicate with M(r) the total mass contained in a sphere of
radius r. We would guess it having the same behavior as the luminosity. But it is
not so. Let υ(r) be the (average) velocity of the points of the galaxy at the
distance r from the rotation axis. We can consider with a reasonable
approximation the mass distribution as spherically symmetrical. Then, the
gravitational force acting on a body, a star or a gas particle, at the distance r is
the same as the force of all the mass inside r, concentrated in the center, exactly
as for the weight of an apple. Differently from the apple, there is now a lot of
mass outside r, but, as we have proven in Sect. 4.6, its gravitational force inside
a spherical shell is zero.
Fig. 4.24 A spherical mass distribution. M(r) is the mass in a sphere of radius r

The force at the distance r is

(4.37)
and the rotation velocity at the distance r is

(4.38)
The image of the galaxy shows that the luminosity ends at a certain distance.
The visible part of the galaxy has a radius that we call r vis. Typical values vary
from 10 kpc to 100 kpc (1 pc, parsec,1 is 3 × 1016 m = 3.3 light years) from the
center. We then expect the function M(r) to increase with r and to become
constant at about r vis, because there is no more mass after that, as represented in
Fig. 4.24. Consequently, the function υ(r) for values of r larger than the radius of
the galaxy r vis should decrease as 1/√r.
How can we measure the rotation velocities of the galaxies at different
distances from the axis? The motion of the single stars is not observable from
earth. However, each of the elements in nature emits light having a well-defined
spectrum, which is characteristic of the element. If the source is moving, the
spectrum is shifted in a known way dependent on the relative velocity between
source and observer (it is called the Doppler effect).
Consequently, we measure the velocities of the different elements of a
galaxy by measuring the spectra of the light they emit. In practice the light
emitted by the huge clouds of gases, such as hydrogen and helium that extend
farther than the stars from the axis, but do not contribute substantially to the
mass.
Figure 4.25 shows the velocities relative to us of the galaxy NGC 2998 as
functions of the apparent distance from its center. We can deduce that the galaxy
has an average velocity (the velocity of its center) of about 4700 km/s. However,
on the left the velocities are systematically smaller, higher on the right. This is
because we are observing the rotation of the galactic disk at an angle different
from 90°. Consequently the disk is approaching on one side, withdrawing on the
other. To have the rotation curve of the galaxy, namely the orbital velocities at
different distances from its center, we subtract the average velocity. The distance
of the galaxy being known, we can convert the apparent distances from axes in
absolute distances. We obtain the diagram in Fig. 4.26.

Fig. 4.25 Rotation curve for the galaxy NGC 2998

Fig. 4.26 The rotation curve of the galaxy NGC 2998, the orbital velocity versus distance from center

We would expect the orbital velocity to decrease as 1/√r at distances larger

than the visible radius, which is in this case about 8 kpc. It is not so; the velocity
remains practically constant up to the maximum distance explored, much beyond
the distance at which no more stars are present.
The behavior of NGC 2998 is not an exception, rather is the norm. The same
phenomenon was found in all the spiral galaxies. We need to conclude that either
the Newton law is no longer valid in these circumstances, or that there is much
more matter in the galaxies than the visible one, which extends much beyond the
visible one. It has been called dark matter (but invisible matter would be a better
name). We now know that the right alternative is the latter. The conclusion
comes from a large number of observations, at different length scales, for
phenomena ruled by different physics, at different eras of the Universe. All point
consistently to the conclusion that dark matter is about five times more abundant
than the matter we know. The search for dark matter is one of the frontiers of
today’s physics.

4.11 Elliptic Orbits

In Sect. 4.4 we have seen the solution of the so-called direct Kepler problem ,
namely how to find the force from knowledge of the orbit. We have done that
however, in the particular case of circular orbits. It is instructive to solve the
problem in general, for elliptic orbits . We shall do that in this section using the
modern calculus. In the next section we shall show the same, following the
Newton demonstration.
We start finding the expressions of velocity and acceleration of a generic
material point P, moving an arbitrary plane curve, in polar co-ordinates. We
introduce a polar co-ordinate frame with origin O and polar axis x (see
Fig. 4.27). We call θ the azimuth of the position vector r, and u θ and u r the
unitary vectors respectively. The time derivatives of the latter is given by the
Poisson formula (1.59)

Fig. 4.27 The unit vectors of the polar co-ordinates

(4.39)
We now find the velocity, which is the time derivative of the position vector
r=rur

which, for the first of Eq. (4.39) is

 We now derive once more to have the acceleration

and finally

(4.40)
We have now the kinematic expressions we need. Pay attention to the fact
that υ r and a r are the components of the vectors on the position vector r from
the focus, not from the center of the ellipse.
We now consider the motion of the planet. The 1st Kepler law states that the
orbit is an ellipse with the sun in one of the foci.
We start by recalling the main properties of the ellipse (one of the conic
sections, together with the hyperboles and the parabola). We choose the polar
co-ordinate frame shown in Fig. 4.28 with the origin in the focus where the sun
is and the major axis as polar axis. (Notice that there are also polar co-ordinates
with the origin in the center O). The angle θ is called anomaly (to be precise, it is
sometimes called true anomaly, to distinguish it from the case in which the
origin is in the center), a and b the semi-major and semi-minor axes.

Fig. 4.28 The geometry of the ellipse and its main parameters

The equation of the ellipse , in its “canonical” form, is

(4.41)
where e is the eccentricity and p is the semi-latus rectum which is the
position vector for θ = 90°. The relation between eccentricity and semi-axes is
(4.42)
The circle can be considered a degenerate ellipse with e = 0. The smaller the
eccentricity the smaller is the difference of the ellipse from the circle. As a
matter of fact, the eccentricities of the planets are in any case quite small, much
smaller than in Fig. 4.28.
Consider now the force. First we observe that, being the force directed to the
sun, the F θ component is zero. This statement is equivalent to the 2nd Kepler
law and to the conservation of angular momentum. Indeed, from Fig. 4.29 we
see that the infinitesimal area swept by the position vector is
Hence the areal velocity is

Fig. 4.29 Elementary area swept by the position vector of the planet

(4.43)
and

(4.44)
In addition, calling L the angular momentum and recalling Eq. (4.7) we can
write

(4.45)
This expression will be useful in the following.
We are now ready to go to the acceleration a r and the force F r towards the
sun. We already found, Eq. (4.40), that

(4.46)
The polar co-ordinates r and θ are not independent, but linked by the ellipse
Eq. (4.41). Taking the time derivative of this equation, rearranging the terms and
using Eq. (4.45), we have
 (4.47)
We derive this again, because Eq. (4.46) contains the second derivative, and
use again Eq. (4.45), obtaining

We now substitute this in Eq. (4.46), use once more Eq. (4.45) and get

Looking back to the equation of the ellipse we recognize that the expression
in parenthesis in the last member is just –1/p. Finally we have

(4.48)
where the minus sign tells us that the force is opposite to r. We see that the
acceleration is inversely proportional to the square of the distance from the sun.
The same is true obviously for the force

(4.49)
This completes the proof. We have proven that if the orbit is an ellipse with
the sun in one of the foci, the force is inversely proportional to the square of the
distance. The remaining part of the argument to reach the Newton law is the
same we already did for circular orbits, with the conclusion

(4.50)
We did not need the 3rd Kepler law to reach this conclusion, as it had been
the case in the particular case of circular orbits. Indeed, in that case Eq. (4.41)
reduces to r = p = constant and not all of the arguments of this section any longer
hold.
Before concluding we stress once more that there is a unique dependence on
r of a central force F r (r) that produces elliptic orbits with the sun in a focus,
. As Newton showed, even the smallest difference in the exponent,
would produce an orbit of the type shown in Fig. 4.30, which is, so
to say, a slowly rotating ellipse, called a rosette . We shall not reproduce the
argument here, but only give a hint. In a motion on an ellipse or on a rosette,
both polar co-ordinates, r and θ, vary in time periodically. The period of the
latter is in any case the time to increase θ by 2π. The period of r depends on the
force. Only if the force is inversely proportional to r 2 is it equal to the period of
θ and the trajectory is closed. If the exponent of 1/r is not exactly 2, the two
periods are different, the orbit does not close and we have a situation like
Fig. 4.30. This effect cannot be seen if the orbit is circular, because a circle
rotating on itself is not different from a circle.

Fig. 4.30 A “rosette” orbit, showing a “snapshot” every 15° of precession

Astronomers have observed for centuries the apparent trajectories of the

planets in the sky with high accuracy. The absolute trajectories are obtained
subtracting the now well-known motion of earth. The aphelia and the perihelia ,
in particular, can be accurately identified. If the force is proportional to the
inverse square distance from the sun, these points should remain fixed. Indeed,
this is almost the case, but not quite. Very slow movements, called precession s,
of the perihelia are observed. They are the effects of the forces of the other
planets, the larger ones in particular, that act on them in addition to the sun.
Observations and calculations agree, with an exception, which was found by Le
Verrier in 1849. He calculated the precession of the perihelion of Mercury , the
nearest to sun, in 10 arc minutes per century. The largest fraction of that is
explained by the just mentioned effects of other planets. But not completely; 43
arc sec per century remained unexplained. A number of hypothesis were
advanced, but all of them failed. This was the first historical example of the
limits of the Newton law. The explanation of the anomalous precession of the
Mercury perihelion by Albert Einstein (1879–1955) in 1915 marked the success
of general relativity.
The 3rd Kepler law is a consequence of the 1st one. Let us prove that. We
start with the consideration that the period T is the area of the orbit A divided by
the areal velocity and expressing the latter in terms of the angular momentum L
using Eq. (4.45).

We now write the acceleration a r Eq. (4.48) using this equation and writing
the parameter p in terms of the axes, Eq. (4.42)

The force on the planet is the Newton force, and we can write

and finally

(4.51)
That is the 3rd Kepler law: the squares of the periodic times are proportional
to the cubes of the ellipse semi-major axis, for all the bodies orbiting the same
central body (of mass M).

4.12 The Newton Solution

In this section we shall not introduce any new concept, rather we shall show how
Isaac Newton demonstrates some of those we discussed in the previous sections.
Reading pages of the giants is, in fact, very instructive, even if, as is the case
with Newton , it is not always easy. After having given the necessary
preliminary information, we shall read one page of the Principia, explaining
their meaning line by line. As we shall see, the Newton arguments are mainly
geometrical. The novelty, with respect to what was already known to the Greeks,
is the final passage to the limit for the length of the considered orbit arc going to
zero.
After having stated the laws of motion in Sect. 1 of the Principia, Newton
dedicates Sect. 2 to “The determination of the centripetal forces ”. Here he
considers orbits of various geometrical shapes under the action of a force
directed to a immovable center (i.e. centripetal). The case we shall take is the
ellipse with the center of force in a focus. In the first two “Propositions” of
Sect. 2, he shows that, in any case, if the areal velocity is constant the force is
directed to the center and vice versa.
 Subsequently, in Proposition VI, Newton lays down the basic scheme he
shall use to solve the above-mentioned problems. The scheme is shown in
Fig. 4.31.

Fig. 4.31 The scheme of Proposition VI

A body moves on the arc PQ of its orbit in the short time interval ∆t. If there
were no gravitational force from the sun, the planet would move of rectilinear
uniform motion on the displacement PR. On the other hand, if abandoned still in
P the planet would drop in the time ∆t, under the gravitational attraction, by the
displacement PX. If the force is constant, the motion is uniformly accelerated
and PX is proportional to ∆t 2. If both conditions are present, the displacement is
the diagonal PQ. We now draw the segment QR parallel to PX. QR touches the
trajectory in Q.
What we just stated would be true if the force were constant during ∆t, which
is not true. However, the smaller is ∆t the smaller is the variation of the force in
that interval. This means going to the limit of The limit geometrically
corresponds to approximate the segment of the trajectory with a segment of
parabola. The motion is then equal to what was found by Galilei for the
projectiles on earth.
On the other hand, QR is also proportional to the acceleration and to the
force F we are looking for, namely , or .
For the constancy of the areal velocity, the time interval is proportional to the
area swept by the position vector in that interval, which is the area of the triangle
SQP. The latter, in turn, is proportional to the product of its base SP and its
height QT, and we have

(4.52)
This expression is valid for any curve. We shall see how it simplifies in the
case of the elliptic orbit, with the center of force in a focus. To do that, we shall
need to know some definitions and four properties of the ellipse. We give them
here without proof.
A diameter is a chord going through the center of the ellipse. Consider the
tangent to the ellipse in any given point P on it (see Fig. 4.32). Let be PP′ the
diameter passing in P and DK the diameter parallel to the tangent in P. The
diameters PP′ and DK are called conjugate diameters .

Fig. 4.32 PP′ and DK are conjugate diameters

Notice that the conjugate diameters bisect each other but, in general, do not
have equal lengths, neither t0 they cross at right angles.
Property 1. The sums of the distances of any point of the ellipse from the
two foci are equal and are equal to the major axis, 2a.
Property 2. (Fig. 4.33). All the parallelograms having conjugate diameters
as sides have the same area. It is equal to the area of the parallelogram
having, in particular, the axes as sides, namely 4ab.

Fig. 4.33 Property 2. Parallelograms with conjugate diameters sides

Property 3 (Fig. 4.34). The two focal lines that join any point P of the
ellipse form equal angles with the tangent in that point.
 Fig. 4.34 Property 3. Two focal lines and their angles with the tangent

Property 4 (Fig. 4.35). Every diameter bisects all the conjugate chords. For
any given diameter the ratio between the areas of the rectangles made by
the two segments of the diameter and the square of the corresponding semi-
chord are equal. Namely

Fig. 4.35 Property 4

(4.53)

We have now the properties of the ellipse we shall need and we can read
Proposition XI.
Proposition XI states:

if a body revolves in an ellipse; it is requested to find the law of the

centripetal force directed to the focus of the ellipse.

The proof shows that the ratio QR/QT 2 in Eq. (4.52), in the particular case of
the ellipse with the center of force in a focus, is equal to the latus rectum , which
we called 2p and he calls L. We shall use his symbol in this section (no risk of
confusion with the angular momentum).
Figure 4.36 reproduces the diagram on which the theorem is developed. The
first lines of the Proposition are:
Fig. 4.36 The Newton diagram for Proposition XI

Let S be the focus of the ellipse. Draw SP cutting the diameter DK of the
ellipse in E, and the ordinate QV in X; and complete the parallelogram
QXPR

The sun (the center of force) is in the focus S; H is the other focus, C is the
center, CA = a and CB = b are the semi-major and the semi-minor axes
respectively. At a certain instant the planet is in P, SP = r is the position vector
from the sun. We draw the tangent RPZ to the ellipse in P and the line QV
parallel to it. Be X and V the points were it cuts SP and PC respectively. We also
draw the lines of QRPT as in Fig. 4.31. To complete the diagram we draw the
perpendicular from P to the diameter DK and call F the point in which they
meet.
The Newton language is extremely synthetic. What is evident for him is not
always evident for us. We shall explain his lines immediately.

It is evident that EP is equal to the greater semi axis AC: for drawing HI
from the other focus H of the ellipse parallel to EC, because CS and CH are
equal, ES and EI will be also equal; and hence EP is half the sum of PS and
PI, that is (because of the parallels HI and PR, and the equal angles IPR,
HPZ) of PS and PH, which taken together are equal to the whole axis 2AC.

The geometric elements of Fig. 4.36 that are relevant for this step are
redrawn in Fig. 4.37. We start from the equation (Property 1)
Fig. 4.37 EP is equal to the major axis

(4.54)
The triangle IPH is isosceles with vertex in P. This is because:
the angles RPI and PIH are equal, as alternate interior angles of the two
parallel lines JL and RZ
the angles HPZ and IHP are equal as alternate interior angles of the same
lines
the angles PIR and HPZ are equal for the Property 3 of the ellipse
Consequently the angles PIH and IHP are equal, which proves the statement.
Hence PH = PI and we can simplify Eq. (4.54) as
(4.55)
The triangles ISH and ESC are similar because they have the same angle in
the vertex S and the sides opposite to it (EC and IH respectively) are parallel. In
addition, SH is twice SC and consequently SI = 2 ES, that is also ES = IE.
Substituting in Eq. (4.55) we obtain
(4.56)
and finally
(4.57)
Now Newton works on QR:

Draw QT perpendicular to SP [we did that already], and putting L for the
principal latus rectum of the ellipse (or for 2BC2/AC [see our Eq. (4.41)])
we shall have

Is not so simple to follow the Newton language. He uses proportions, which

we shall write as fractions to make them more readable. In addition, when he
takes a route he does not tell us the reasons, which are understood only at the
end. Let us trust him and follow. He starts from the ratio QR/PV with numerator
and denominator multiplied by L, because at the end this will be useful. The
relevant geometrical elements are drawn in Fig. 4.38.

Fig. 4.38 Working on QR

The first step is trivial

But QR = PX by construction. Let us find PX.

The triangles PXV and PEC are similar because they have a common vertex
in P and the two opposite sides, XV and EC, are parallel. Consequently
PE/PC = PX/PV and also

Using Eq. (4.57) namely PE = AC we have

and in conclusion

(4.58)
The next step is working on PV. The single line of Newton is:

also and

Once more Newton works on a ratio, L/GV, and multiplies numerator and
denominator by the same quantity, which is PV, the quantity we are now looking
for. The relevant geometrical elements are shown in Fig. 4.39.

Fig. 4.39 Work on PV

The first step is again trivial,

(4.59)
We use the Property 4 of the ellipse applied to the diameter PG and to the
semi-chords QV and DC conjugated to it, getting

(4.60)
Newton continues, finding a fourth proportion. Finally he will put the four
together. We take a breath, abandon him for a moment and put immediately
together the three Eqs. (4.58), (4.59) and (4.60) we found. We multiply them
member by member and obtain

Simplifying, but keeping L that will be useful, we have

(4.61)
We need another proportion, the last one.

By Cor. II, Lem. VII [is the rule for going to the limit], when points P and
Q coincide, QV2 = QX2 and QX2 or QV2: QT2 = EP2:PF2 = CA2:PF2, and
(by Lem. XII) = CD2:CB2.

We now need to express 1/QT 2. As usual, Newton works with proportions,

and we shall do the same with ratios. This time it is QX 2/QT 2. At the end we
shall take the limit for the length of the arc PQ going to zero, namely to have the
points P and Q coincident. In this limit, points X and V coincide too and it is then
convenient to consider QV 2/QT 2 in place of QX 2/QT 2. The relevant
geometrical elements are drawn in Fig. 4.40.

Fig. 4.40 Work on 1/QT 2

The triangles EPF and XQT are similar, because

the angles in F and T are equal because are both right,
the sides of the angles respectively in P and Q (in evidence in the figure)
are mutually perpendicular, hence they are equal.
Hence

and, as EP = CA,

(4.62)
To find PF we use Property 2. Figure 4.41a shows that PF is one half of the
height to DK of the drawn parallelogram on conjugate diameters.

Fig. 4.41 Using Property 2

 Property 2 gives: or

which, substituting in Eq. (4.62) and squaring, gives

(4.63)
The next step is to multiply the four proportions. Newton writes:

Multiplying together corresponding terms of the four proportions, and

simplifying, we shall have

since

We have already multiplied the first three ratios obtaining Eq. (4.61). Hence
we multiply now its members with those of Eq. (4.63)

and simplify

Recalling the relation between latus rectum and axes and

simplifying we finally have

(4.64)
Remember that the factor in Eq. (4.52) we want to express is QR/QT 2. We
have it now in Eq. (4.64). The final step is taking the limit for Remember
that the concept of limit was not known before Newton. He writes:

But the points Q and P coinciding, 2PC and GV are equal. And therefore
the quantities and QT2, proportional to these, will be also equal. Let
those equals be multiplied by and will become equal to

In the limit in which the arc PQ becomes infinitely small, point V coincides
with P. Consequently, GV becomes equal to 2PC, and the second member of
Eq. (4.64) goes to one, becoming

Now multiply both members by SP 2/QR and get

Finally, Newton concludes:

And therefore (by Cor. I and v, Prop. VI) the centripetal force is inversely
as , that is, inversely as the square of the distance SP.
Q.E.D.

Namely:

(4.65)
and, given that L, our 2p, is a constant for a given ellipse,

(4.66)
The force is inversely proportional to the square of the distance from the
center. That is what we had to show.

4.13 The Constants of Motion

We now go back to the main stream and consider the potential and the kinetic
energy of a body of mass m in the gravitational field of a body of mass M,
moving on an ellipse .
We start with its angular momentum, which we call L and go back to our
formalism calling p the semi-latus rectum. From Eq. (4.50) we can write

(4.67)
which, using Eq. (4.42) is

(4.68)
In words, the square of the angular momentum is proportional to the major
axis. For a given major axis, the angular momentum is the largest for e = 0,
which is the circle. It decreases for increasing e, i.e. for the ellipse becoming
more and more squeezed.
Consider now the potential energy , and make use for r of the ellipse
equation

(4.69)
For the kinetic energy , remember Eq. (4.39)

(4.70)
Using the expression of dr/dt given by Eq. (4.47) and using Eq. (4.52), we
have

We use now the Eq. (4.41) of the ellipse to express 1/r 2 and, taking into
account that obtain

(4.71)
Both potential energy, Eq. (4.69), and kinetic energy Eq. (4.56) depend on
the position of the planet and consequently on time. Not so the total energy,
which is their sum

(4.72)
which we can also write, in equivalent manner

(4.73)
In conclusion, the total energy of the planet depends only on the semi-major
axis . Different orbits, such as those in Fig. 4.42, which have the same semi-
major but different semi-minor axes have the same total energy. However, as we
have seen above, the angular momentum grows for decreasing eccentricity.
Fig. 4.42 Orbits of the same energy and different angular momenta

Pay also attention to the fact that the total energy is negative. However, this
is not the only possibility for a body moving about the sun, or any other source
of gravitational force. As a matter of fact, in our demonstration we have used
only Eq. (4.41). This is not only the equation of the ellipse, but, more generally
of all the conics, ellipse if e < 1, parabola if e = 1, hyperboles if e > 1. The three
cases correspond, from the physical point of view, to total energy (4.73)
negative, null or positive respectively. The potential energy is always negative,
tending to zero at infinite distance from the center. The kinetic energy can be
positive or zero. Consequently, at infinite distance the total energy is positive or,
as a minimum, zero. If the total energy of a body is negative, it must remain at
finite distances. The orbit is said to bound . The ellipse (including the circle as a
particular case) is the only conic that does not reach infinity. If on the contrary,
the total energy of a body is positive, it will be able to go farther and farther; at
infinite distances, or more realistically at distances large enough to have
negligible potential energy, all its energy is kinetic, positive in fact. The
intermediate case is when the body reaches infinity with zero kinetic (and total)
energy. The trajectory is a parabola.

4.14 Problems
4.1. A pendulum having 1 s period on the surface of earth is brought on the
surface of a planet having the same radius of earth and mass four times
larger. What is the period of the pendulum?

4.2. The gravitational potential difference between two points on the earth
surface (at the same latitude) is 1000 m2 s–2. What is the difference between
the heights were they are located?

4.3. We abandon a body at the distance from earth of the moon orbit with no
velocity. Will it fall with constant velocity? With constant acceleration?

4.4. We move a body from the sea level to the top of a mountain 5000 m high
 (same latitude). How does its mass vary? How does its weight vary?

4.5. Does the velocity at which a satellite moves in a circular orbit around the
earth depend on the mass of the earth? on the mass of the satellite? on the
radius of the orbit?

4.6. The apparent diameter of the sun as seen from earth is approximately
α = 0.55°. What would be the period of a hypothetical planet orbiting just
out of the sun?

4.7. We want to put an artificial satellite in orbit around the earth having a
period of 2 h. Knowing the gravity acceleration g on the surface of earth an
its radius R E , find the height of the requested orbit above the surface.

4.8. Consider a spring (of a ballpoint pen) with rest length 3 cm and elastic
constant k = 50 N/m. We fix to its two extremes two equal Pb spheres
(density ρ = 11 × 103 kg/m3), of mass m = 104 kg each. Assume,
unrealistically, that all frictions can be neglected. How much will the spring
shrink under the action of the gravitational attraction of the two spheres?

4.9. Knowing the values of g, of G N and of the radius of earth (R E = 6.4 × 106
m), make an estimate of the mass and of the mean density of earth.

4.10. Knowing the values of G N and of the radius of earth orbit (r E

= 1.5 × 1011 m) and of its period, make an estimate of the mass of the sun.
Knowing that its apparent diameter from earth is 0.55°, estimate its mean
density.
4.11. The sun moves on an orbit that we can consider circular about the center
of the Galaxy. The radius of the sun orbit is R S = 25
000 l year = 2.5 × 1020 m, his velocity is υ S = 250 km/s. Compare these
data with those relative to the motion of earth about the sun (r E
= 1.5 × 1011 m, υ E = 30 km/s). Make an estimate of the total mass M tot
around which the sun orbits; give it as a multiple of the solar mass M S .

4.12. Io, one of the Jupiter satellites, has the orbital period T I = 1.77 d and the
orbit radius r I = 4.22 × 108 m. Compare these data with those of the
motion of the earth about the sun (r E = 1.5 × 1011 m, υ E = 30 km/s).
Determine the mass of Jupiter in solar masses.

4.13. Find a procedure to determine the mass of earth.

4.14. Knowing that the earth moves around the sun with the velocity of υ E
= 30 km/s, find the gravitational potential of the sun ϕ S (E) in the points
of earth orbit. The gravitational potential in a point of the earth is the sum
of the just considered ϕ S (E) due to the sun and of the gravitational fields
of the earth itself, say ϕ E (E), and of all the Galaxy, say ϕ G (E). Calculate
the values of the latter two relative to ϕ E (E), knowing that the masses in
the three cases are approximately M S = 2×1030 kg, M E = 6×1024 kg, M G
= 2×1041 kg and taking as distances, from earth to sun r ES = 1.5 × 1011
m, radius of earth r E = 6.4 × 106 m, distance from sun to the center of
Galaxy r SG = 2.5 × 1020 m

Footnotes
1 A parsec is the distance at which the diameter of the earth orbit is seen under the angle of a second.
© Springer International Publishing Switzerland 2016
Alessandro Bettini, A Course in Classical Physics 1—Mechanics, Undergraduate Lecture Notes in Physics,
DOI 10.1007/978-3-319-29257-1_5

5. Relative Motions
Alessandro Bettini1
(1) Dipartimento di Fisica e Astronomia, Università di Padova, Padova, Italy

Alessandro Bettini
Email: alessandro.bettini@pd.infn.it

In our study of the kinematics of the material point, we have already seen that
the equations of motion depend on the reference frame. The law of motions, and
more generally all the laws of Physics, transform, as we say, from one frame to
another. This chapter is dedicated to the study of these transformations.
Two reference frames may differ in different ways.
The two frames have no relative motion, their co-ordinate homologous axes
are parallel, but have different origins; the frames differ for a rigid translation.
The two frames have no relative motion and coincident origins, but the
directions of the axes are different; the frames differ for a rigid rotation.
One frame can translate relative to the other in time with uniform or varying
velocity, or it can rotate, again with constant or varying angular velocity, or it
can translate and rotate contemporarily.
In Sect. 5.1, we shall consider two stationary frames relative to one another,
with a relative translation or rotation. We shall see that the laws of Physics have
the same form, namely the same mathematical expressions, in both frames. As
we say, the laws are covariant under translations and rotations. The meaning of
the term will be explained.
We shall then consider frames in relative motion and learn that, when the
relative motion is a translation with constant speed, the laws of mechanics are
also covariant. This is the relativity principle, a fundamental principle of physics,
established by Galilei. For example, experiments done inside a closed room in a
ship cannot establish whether the ship is moving in uniform motion or is
standing still. One of the consequences is that once we have found an inertial
frame, any other frame moving in a uniform translation motion relative to it is
also inertial.
In Sect. 5.3, we shall deal with the relative translatory accelerated motion. As
already anticipated, in any reference that accelerates relative to an inertial frame,
the Newton laws are not valid. For example, a body at rest can start moving
without any force acting on it. The motion can be described introducing
fictitious forces, which are known by several equivalent names, apparent forces
of the relative motions, pseudo-forces and inertial forces. We feel such “force,”
for example, when we brake suddenly in a car. In Sect. 5.4, we shall deal with
the general case (translation and rotation) and we shall see the relations between
velocities and between accelerations in two frames of any relative motion. In
Sect. 5.5, we shall discuss several examples of motion in frames rotating relative
to an inertial frame.
Any frame at rest in a laboratory on earth does, in fact, move with earth. In
initial, and quite good, approximation, these frames can be considered to be
inertial. Not completely, however, because earth rotates on its axis and moves
along its orbit around the sun, and even the sun moves along its orbit in the
galaxy. In Sect. 5.7, we shall study a few effects of the inertial forces in frames
at rest relative to earth: the variation with latitude of the magnitude of the
weight, the rotation of the oscillation plane of pendulums, the deviation from the
vertical of free fall and the circulation of winds.
The inertial forces acting on a body are proportional to its inertial mass,
while the gravitational attraction of earth is proportional to its gravitational mass.
This observation allows for the realization of very delicate experiments to check
whether the two masses are different or equal. We shall describe such an
experiment in Sect. 5.8.

5.1 Covariance of the Physical Laws Under Rotations

and Translations
Consider two Cartesian reference frames, which are stationary relative to each
other, S (coordinates x, y, z, origin O) and S′ (coordinates x′, y′, z′, origin O′).
A physical law is a mathematical equation between physical quantities. The
relation between the two frames can be a rigid rotation or a rigid translation . Let
us start with rotations.
We choose the origins of the two frames as coincident. For simplicity, we
consider their z-axes also to be coincident. The frames differ for a rotation, by an
angle θ, around this axis. The rotation is in the common plane xy, as shown in
Fig. 5.1.

Fig. 5.1 Two reference frames different for a rigid rotation

Suppose now that an observer in S makes a very simple experiment. He

measures, using a balance, the masses of two objects, finding the values m 1 and
m 2. He finds that the second mass is three times the first. He writes the relation
(let us call it the “law”)
(5.1)
Another observer in S′ performs the same experiment. We indicate with a
prime the homologous quantities he finds. In this very simple case, considering
that the procedure of measuring the mass with a balance does not depend on the
direction of the axes, we can conclude that he will find the same result, namely
(5.2)
The second observer also states that
(5.3)
Equation (5.3) has the same form as Eq. (5.1). Namely, the two observers
describe the same phenomenon with laws of the same form. Indeed, mass is a
scalar quantity, which is invariant under rotations of the axes. In general, a
relation between scalar quantities, if valid in a frame, is also valid in any frame
rotated relative to the first one, because both sides of the equation do not vary
going from one frame to the other.
But it is not always so. Suppose that the observer in S measures two
components of the velocity of a point, finding the values υ x and υ y . He finds,
again, the second quantity three times larger than the first and writes the
equation
 (5.4)
What would the observer in S′ find? We know the answer because we know
the relations between velocities in the two frames:
(5.5)
We calculate the ratio between and , also employing Eq. (5.5):

In conclusion, in S′, we have

(5.6)
The form of the “law” is different this time in the two frames, being (5.4) in
S and (5.6) in S′. This is an obvious consequence of the fact that the components
of a vector transform differently one from another.
But wait a moment, a law may be valid in both frames, even if its sides are
not invariant, as in the case of the masses; rather, it is sufficient that, if they vary,
in the same way. Let us see what happens for a law linking vector quantities.
The observer in S′, which we assume, for the sake of this example, to be
inertial, studies the motion of a material point. He measures the acceleration a
(namely its three components), the force acting on the point F (again, the three
components) and the mass m. He finds the relation
(5.7)
More explicitly, this vector relation corresponds to three equations:
(5.8)
We know how the components of the vectors, such as F and a are, transform
from one frame to the other, namely

(5.9)
and we can write

and, for Eq. (5.9),

(5.10)
which has the same form as Eq. (5.8). Both sides of the equations are
different, varying from one frame to the other. However, they vary in the same
way, because both sides are vectors. Thus, we say that the equation is covariant .
In conclusion, the laws of Physics keep the same form under rotations of the
axes, or, in other words, are covariant under rotations. And yet, from another
perspective, it is impossible experimentally to establish any privileged directions
of the reference axes. Space should be considered isotropic, without preferential
directions.
The case of the translations is very simple. Scalar quantities obviously have
the same values in two frames differing for a translation. This is also valid for
vectors, which are simply translated; hence, they are the same vector.

5.2 Uniform Relative Translation. Relativity Principle

Consider now two reference frames, S and S’, which are in relative motion. We
arbitrarily call one of them S (origin O and coordinates x, y, z) fixed and the
other one S’ (origin O’ and coordinates x’, y’, z’) mobile. We consider the case
of a uniform translation of S’. All the points of S’ move with the same velocity
relative to S, which is constant in magnitude and direction. The frames, for
example, might consist of one fixed on the ground, the other on a carriage
moving on straight rails, or a frame fixed at the shore and one on a ship moving
straight, in both cases with uniform motion.
The axes of the two frames do not change the relative directions and we can
take them as being parallel. Fig. 5.2 shows the two frames at a certain instant. At
a later time, the mobile frame will be in a different position, more on the right,
but its axes will still be parallel to the axes of S. We choose t = 0 as the time at
which the axes of the two frames overlap.

Fig. 5.2 Two reference frames in relative uniform translation motion

Figure 5.2 shows the material point P and its trajectory. The position vectors
r and r’ of P in the two frames have the well-known relation
(5.11)
where rO’ is the position vector of the origin O’ of the mobile frame S’ in
the fixed frame O, namely OO’.
A fixed and a mobile observer see the point P moving with different
velocities, v and v’. To find their relation, we take the time derivatives of Eq.
(5.11), obtaining
(5.12)
where v O’ is the velocity of the origin O’ of the mobile frame, and also of
all its points (because the motion is a translation) as seen by S. The velocity of
an insect flying in the ship in the above example relative to the shore is the
vector sum of the velocity of the insect relative to the ship and the velocity of the
ship relative to the shore.
A further time derivation gives the relation between accelerations
(5.13)
where aO’ is the velocity of the origin O’ of the mobile frame, and also of all
its points.
We now consider the important particular case in which the translation of S’
relative to S is uniform, namely the velocity of its origin, and of all its points,
seen by S is constant in time
(5.14)
Then, obviously,
(5.15)
and Eq. (5.13) becomes
(5.16)
The accelerations in the two frames are equal. The implications of this
simple conclusion are extremely important considering inertial frames.
If S is an inertial frame, any material point P not subject to forces moves at
constant velocity v (or remains at rest). In other words, its acceleration is zero, a
= 0. In the mobile frame, its acceleration a’, which is equal to a, is also zero.
Consequently, S’ is inertial too.
We conclude that, given an inertial reference frame, any other frame moving
relative to it by a uniform translation is also inertial.
What about the second Newton law? It is valid in the frame S, which is
inertial by assumption. Is it also valid in S’? In S, we have
(5.17)
 The observer in S’ measures the same mass (m’=m) and the same force (if,
e.g., he uses a dynamometer, the spring stretches by the same amount), F = F’.
The acceleration a’ that he measures is also equal to a, but only in the case we
are considering of relative translation at constant velocity. Then, in S’, the
relation between force, mass and acceleration is
(5.18)
In other words: the laws of mechanics are covariant under the
transformations that link two reference frames in relative uniform translation
motion.
As an example, consider a reference S’ fixed on a sailing ship moving on the
sea at constant velocity and S a frame fixed to the shore. As above, we choose
the axes of the two frames mutually parallel and with coincident origins at t = 0.
An experimenter climbs on top of the mast and drops a stone. Fig. 5.3 shows the
trajectories of the stone as seen by an observer on the shore, a), and on the ship,
b).

Fig. 5.3 Trajectory of a stone dropped from the top of the mast of a ship, as seen from the ship and the
shore

For the observer in S, the stone falls under the action of its weight, a constant
force (F = m g), directed downwards, opposite to the z-axis (that we have taken
to be vertical upwards). The initial velocity of the stone is the velocity of the
ship, and we have taken the x-axis in that direction. Hence, the motion of the
stone in the z direction is uniformly accelerated, while in the x direction, it is
uniform (neglecting the air resistance). The trajectory is a parabola. In the figure,
we marked the positions of the stone in time instants separated by the same time
interval.
In S’, the forces are the same, but the initial conditions are different; the
initial velocity of the stone is zero. Hence, it falls vertically along the z’-axis
with a uniformly accelerated motion.
Summarizing, in the two frames, the trajectories are different. The reason for
the difference is in the different initial conditions of the motion. On the contrary,
both observers describe the motion with the same law, F = m a. The two frames
are perfectly equivalent for every dynamic experiment. Each of them can be
considered as fixed or movable.
This conclusion is important and is known as the relativity principle. The
principle does not deal directly with the phenomena but rather with the laws that
describe the phenomena. It states that: the laws of Physics are covariant, namely
have the same form, in any reference frame moving of translational uniform
relative motion.
In our discussion, we have seen that the relativity principle is valid for the
laws of mechanics, which is the physics chapter we are studying. However, its
validity is completely general, including, in particular, all fundamental
interactions, gravitational, electromagnetic, nuclear strong and weak
interactions. In other words, it is impossible experimentally to establish the
relative motion, provided it is as uniform translation. Historically, the principle
was established by G. Galilei. He did not use that name, which was given to it by
Henri Poincaré (1854–1912) in 1904, but Galilei established it in complete
generality, describing, in a beautiful page, a series of experiments, some of
which were of an electromagnetic nature, below the deck of a large sailing ship.
The page of the Dialogue (transalted from Italian into English by the author) is:

Shut yourself with a few friends in the largest room below decks of some
large vessel, and have with you flies, butterflies and similar small flying
animals. Let a large bowl of water with several small fish in it be the cabin
too. Hang also, at a certain height, a bucket pouring out water drop by drop
into another vase with a narrow mouth beneath it. When the ship stands
still, carefully observe how those flying small animals fly with equal speed
towards all sides of the cabin; you will see the fish swim indifferently in all
directions; all the drops will fall into the vessel beneath; and you, when
throwing something to a friend, will not need throw it more strongly in one
direction than another, when the distances are equal; and jumping up feet
together, you will pass equal spaces in all directions.
Once you have observed all these things carefully, though there is no
doubt that when the vessel is standing they must happen like that, let the
vessel move with speed as high as you like. Then (provided the motion is
uniform and not unevenly fluctuating) you will not discover the slightest
change in any of the named effects, nor you will be able to understand from
any of them whether the ship is moving or standing still. In jumping you
will pass on the planking the same spaces as before, nor you will make
longer jumps toward the stern than toward the prow, as a consequence of
 the fast motion of the vessel, despite the fact that during the time you are in
the air the planking under you is running in a direction opposite to your
jump. In throwing something to your companion, no more force will be
needed to reach him whether he is on the side of the prow and you of the
stern or your positions are inverted. The drops will fall as before in the
lower bowl, without a single one dropping towards the stern, although,
while the drop is in the air, the vessel runs many palms. The fish in their
water will swim toward the forward part of their vase with no more effort
than toward the backward part, and will come with equal ease to food
placed anywhere on the rim of the vase. And finally the butterflies and the
flies will continue their flights indifferently towards every side, nor will
ever happen to find them concentrated close to the wall on the side of the
stern, as if tired from keeping up with the course of the ship, from which
they, remaining in the air, will have been separated for a long time. And if
some smoke will be made burning a bit of incense, it will be seen ascending
upward and, similar to a little cloud, remaining still and indifferently
moving no more toward one side than the other. The cause of all these
correspondences of effects is that the motion of the ship is common to all
things contained in it, and to the air also.

We notice here that the development of electromagnetism in the last part of

the XIX century led to doubts concerning the general validity of the principle.
The process of in depth analysis of the physical laws that followed, leading to
the relativity theory , showed that the Galilei relativity principle was, as we have
stated, valid in general. However, it was found that the transformations of the co-
ordinates, and of the time, between reference frames, valid at a small velocity
relative to the speed of light, do not hold at high speeds. We shall discuss that in
Chap. 5. Here, we simply anticipate the root of the issue. Consider the
transformation equations that link the co-ordinates in S′ and in S
(5.19)
where we have included the relation between times t and t′ measured by the
two observers. Indeed, the measurement of a time interval should be, we think,
the same on the shore as on the ship (to continue the example). However, this
conclusion, coming from our everyday experience and from experiments at the
usual velocities, is wrong at velocities not too small compared to the speed of
light. Two observers in two frames moving at those speeds measure different
time intervals between the same two events; in other words, t and t′ are not
equal. A consequence is that the relations between co-ordinates are different
from those of Eq. (5.19). The transformation Eq. (5.19), called Galilei
transformations , fail at high velocities and must be generalized into the Lorentz
transformations , as we shall see in Chap. 5. But the relativity principle remains
completely valid.

5.3 Non-uniform Translation. Pseudo Forces

We now consider the case in which the motion of the reference S′ relative to S is
still a translation, but with variable velocity. Consider, for example, S′ to be
fixed on a trolley moving on straight rails with an accelerated (or decelerated)
motion relative to S fixed on the ground. We still consider the motion of the
point P in Fig. 5.2 as seen by two observers in the two frames. The relation
between the accelerations is
(5.20)
As in the previous section, a O ′ is the acceleration in S of the origin of S′ and
also of all the points fixed in it (its motion being a translation). Suppose now that
S is an inertial frame. If no force acts on P, its acceleration is zero, a = 0. In S′,
however, a = – a O ′ ≠ 0. Namely, in S′, a body not subject to forces may
accelerate. The law of inertia does not hold. S′ is not inertial. Consider the trolley
in the above example initially moving at constant speed. If we put a ball on a
horizontal plane, it will not move. If the trolley now suddenly slows down, we
shall see the ball accelerating forward, without any force acting. This is the
interpretation of the observer in S′. The inertial observer in S thinks that there is
no force acting on the ball (suppose friction to be negligible), and that it is just
continuing its uniform motion (Fig. 5.3).
If a force F acts on the point P of mass m, the inertial observer in S finds the
relation
(5.21)
The observer in S′ measures the same value of the mass, m′ = m, the same
force, F′ = F, but a different acceleration a′, and finds ,
or
(5.22)
We also see that the second Newton law, not only the first one, does not hold
for the non-inertial observer S′.
However, the observer in S′ can play a trick. Indeed, he is accustomed to
thinking that any acceleration will be due to a force and will imagine that a force
has suddenly started to act on the ball on the table. Formally, the trick is by Jean
Baptiste d’Alembert (1717–1783); we can re-write Eq. (5.22) moving m′a O′ to
the left-hand side, as
(5.23)
and call −m′a O′ a force, or, more accurately, a fictitious force , or inertial
force
(5.24)
and Eq. (5.23) becomes
(5.25)
Namely, if we add to the “real” forces the fictitious, or inertial, ones, we re-
establish the validity of the 1st and 2nd Newton laws. However, these forces are,
as we said, fictitious, not real, because they are not produced by any physical
agent. Consequently, there is no corresponding reaction. The 3rd Newton law,
the action-reaction law, does not hold for the inertial forces.
Let us go back to the example of a sphere on a table on the trolley. The
resultant of the true forces, weight and normal reaction of the plane, is zero.
When the velocity of the trolley is constant, the fictitious force F in is also zero
because so is a O ′. But when the trolley slows down, the observer in the S′ sees
the effect of a fictitious force as in Eq. (5.24). It is directed forward, opposite to
a O ′. He can measure the fictitious force attaching the sphere to a spring and
measuring its stretch. In this way, he verifies that Eq. (5.24) is correct.

5.4 Rotation and Translation. Pseudo Forces

Consider now a stationary frame S (origin O and coordinates x, y, z) and a
mobile frame S′ (origin O′ and coordinates x′, y′, z′), the motion of which is
completely general. It may be a translation, with constant or variable velocity, a
rotation, again with constant or varying angular velocity, or both of them
together. Figure 5.4 represents the two frames at a certain time. At another time,
for example, a bit later, both the position of O′ and the direction of the axes of S′
will be, in general, different.
Fig. 5.4 Reference frame S′ moves in an arbitrary motion relative to S

We begin by finding a formula that will be useful in the following. Consider

a vector A, which does not vary with time relative to S′. Examples are the
position vector in S′ and the velocity of a point moving in rectilinear uniform
motion relative to S′. The vector A is not constant in S. We now find its time
derivative. We notice that A varies relative to S only in direction, not in
magnitude. More precisely, A rotates with the same angular velocity at which, in
that instant, the mobile frame S′ rotates relative to S. We indicate it with ω.
Notice that ω can vary in time, which is why we specify “in that instant”. Under
these conditions, the time derivative of A is given by the Poisson formula and
we have

where the subscript S specifies that it is the rate of change in the reference S.
If the vector A also varies in S′, we have to sum the rate of change in S′, and
finally we have

(5.26)
which is the formula we were looking for. Notice that in the preceding
sections, we did not take care to specify in which frame we were taking the
derivatives. This was allowed because, being the considered transformations
translations, the Cartesian components of the vectors were not modified. This
can be immediately verified in Eq. (5.26) in which, if ω = 0, the derivatives in
the two frames are equal.
We shall now find the relations between the kinematic quantities in S and in
S′. We shall call the former absolute and the latter relative, but we notice that the
definition is arbitrary; we could have started calling S′ stationary and S mobile.
 See that the relation between the position vectors is always
(5.27)
To obtain the relation between relative (in S′) and absolute (in S) velocities,
we need the time derivatives. To do that, we need to have on each side of the
equation only vectors in one frame. Hence, we re-write Eq. (5.27) as
(5.28)
and derive the vector r – r O ′ using the rule (5.26), obtaining

(5.29)
The meaning of the left-hand side of this equation is clear: it is the difference
between the absolute velocities of the point P, say v, and of the point O′, say v O
′. We substitute Eq. (5.28) on the right-hand side, obtaining

(5.30)
Now, we see that the first term on the right-hand side is the rate of change in
S′ of the position vector in S′, namely the velocity of P in S′, which we call
relative and indicate with v′. We then write
(5.31)
In other words, the velocity v of the point P in S is the sum of its velocity v′
in S′ and of two more terms that we have grouped in v t . The meaning of the
latter is understood considering the case in which the point does not move in S′,
namely if v′ = 0. Then, v t is the absolute velocity of the point. We can then state
that v t is the velocity of the point fixed in the frame S′, and call it Q, through
which the moving point P passes at the considered time. We can think of v t as
the velocity of the moving space. It is called the velocity of transportation . It
contains two terms,
(5.32)
which we discuss looking at Fig. 5.5. The first one is the velocity of the
origin of S′ and corresponds to the translational component of its motion relative
to S. The second term is due to the rotation of S′. We can think of this as taking
place about an instantaneous rotation axis passing through O′ with angular
velocity, in the considered instant, ω. Indeed, the velocity of the point Q
stationary in S′ where P is passing is just .
Fig. 5.5 The relative velocity in the rotating frame

We shall now find the accelerations, by a further derivative. We shall meet

more terms. We start from Eq. (5.31) in the form
(5.33)
and derive the left-hand side using Eq. (5.26), obtaining

(5.34)
Similarly to above, the left-hand side is the difference between the absolute
accelerations of P, say a, and O′, say a O ′. Still analogously, we use Eq. (5.33) to
substitute v – v O ′ on the right-hand side, obtaining

(5.35)

The last term looks a bit complicated, but its terms have well-defined
physical meanings. Let us examine them. The first term is the acceleration of P
in S, namely the relative acceleration, say a′. In the second term, the angular
acceleration of the motion of S′ relative to S appears. We shall name it

(5.36)
The next two terms are equal. We put them together and also group some
other terms, writing
 (5.37)
which expresses the Coriolis theorem, after Gustave de Coriolis (1792–
1843). We now define
(5.38)
which is called the acceleration of transportation and
(5.39)
which is called the Coriolis . Finally, we write Eq. (5.37) as
(5.40)
The meaning of the acceleration of transportation a t is analogous to that of
the velocity of transportation v t . Indeed, if both velocity and acceleration of P
in S′ are zero, then its absolute acceleration is a t , as the other two terms on the
right-hand side of Eq. (5.40) are then zero. The term a t is the absolute
acceleration of the point stationary in S′ through which the point P (call it Q
again) is passing at the considered instant. It is the sum of three terms. The first
is the acceleration relative to S of the origin of the mobile frame S′. The second
term is the absolute acceleration of Q due to the rotation of S′ relative to S. The
situation is shown in Fig. 5.6. Indeed, the velocity of Q (of position vector r′)
due to the rotation is . In turn, this velocity varies in time, and its rate of
change is, by the same formula . This is simply the centripetal
acceleration of the point Q. Indeed, as we understand looking at Fig. 5.6, we
have

Fig. 5.6 Geometry in a rotating frame

where d is the curvature radius (the radius of the osculating circle) of the
curve Q is describing. And further

which is the centripetal acceleration of Q.

Consider now the third term in Eq. (5.36). If the angular velocity ω is
constant, the absolute velocity of Q varies only in direction, and this term is
zero. If ω is not constant, the magnitude of the absolute velocity of Q also varies.
This acceleration is given by the third term, .
As for the Coriolis acceleration a Co , we see in Eq. (5.39) that it is zero in
three cases: when, in the considered instant, the point P does not move in S′
(v′ = 0), when the mobile frame does not rotate (ω = 0) and when the velocity of
the point P is parallel to the angular velocity. The Coriolis acceleration does not
depend on the position of P but does depend on its relative velocity and becomes
larger for larger relative velocities. It is always directed perpendicularly to the
motion and consequently is a cause of change in its direction, rather than of its
magnitude. We shall see examples in the next section.
We shall now assume that S is an inertial frame. As we have seen in the
previous section, if S′ accelerates relative to S, it is consequently not inertial. In
other words, the Newton laws in S do not hold. Let us look at the details.
In the inertial frame S, the law of motion of the mass m under the action of
the force F is F = m a. This can be written, using Eq. (5.40), as

The observer in S′ measures the acceleration a′ and wants to have that on the
right-hand side. We move the other terms to the left-hand side, obtaining
(5.41)
We get the Newton law back formally by defining two fictitious forces
(5.42)
and
(5.43)
which is called the Coriolis force , and we subsequently get
(5.44)
We can then state that, in a frame mobile with an arbitrary motion relative to
an inertial frame, the product of the mass times the acceleration is equal to the
resultant of both true and fictitious forces. However, as already stated, the
fictitious forces are not real and are not due to any physical agent. Consequently,
the action-reaction law is not satisfied.

5.5 Motion in a Rotating Frame

Consider now the simple case in which the reference frame S′ rotates relative to
the inertial frame S with angular velocity ω constant in magnitude and direction.
For example, S′ may be fixed on a rotating platform, for example, a merry-go-
round, and S stationary on earth. As we shall see in the next section, such a
frame is not exactly inertial due to the rotation of earth on its axis and its
revolution around the sun, but the effects of the difference are quite small and
we shall disregard them here.
We choose the origin of both frames in the center of the platform, their z and
z′ axes vertical upwards and, consequently, x, y and x′, y′ in the horizontal plane
of the platform, as shown in Fig. 5.7. The axes x and y are stationary relative to
the ground, while x′ and y′ rotate. With our choice of co-ordinates, the position
vectors in the two frames coincide, r = r′.

Fig. 5.7 The S reference frame is stationary to the ground, S′ rotates with constant angular velocity

In the particular case we are considering, the relevant expressions for

velocities and accelerations simplify in
(5.45)
(5.46)
(5.47)
(5.48)
 (5.49)
Let us consider the velocities. In general, the point P is not necessarily on the
platform. In Fig. 5.8, we have drawn it somewhat higher up. In general, the
vectors ω and r are not parallel. Recalling that v t is the velocity of the point Q
stationary in S′ in the instant position of P, we see that it is tangent to the circle
thorough P normal to the rotation axis and with its center on the axis. This circle
is the trajectory of Q. The magnitude of the velocity of transportation is then

Fig. 5.8 The velocity v t of the point Q

(5.50)
where d is the radius of the circle, namely the distance from the rotation axis.
v t is then simply the velocity of Q in its circular motion.
We now consider the accelerations. We immediately see that the a t term is
simply the centripetal acceleration of the point Q as seen in the inertial frame S.
Let us now consider a point P of mass m to be standing still, relative to S′, on
the platform at the distance r from the axis. Suppose that the friction is
negligible and that P is kept in position by a rubber band attached to a small ring
around the axis.
The inertial observer in S sees P moving in uniform circular motion with
velocity ωr. He knows that the motion has an acceleration towards the center,
the centripetal acceleration, of magnitude ω 2 r (this is the absolute acceleration
in this case). The (centripetal) force causing the acceleration is due to the rubber
band. The observer can check that measuring the stretch of the rubber band.
The non-inertial observer in S′, on the platform, also sees that the rubber
band is stretched, determining that a centripetal force is acting on P. He
measures it and finds the same result as the inertial observer. The mobile
observer now insists on having the first Newton law be valid and concludes that
a second force, equal and opposite to that of the rubber band, must exist. This is
the inertial force, due to the acceleration of transportation, –m a t , the direction
of which is opposite to the centripetal force. In this case, the force is centrifugal .
In this case, and always, the centrifugal forces are not real forces, but pseudo
forces of the relative motion. They appear only when we pretend to describe the
motion in a non-inertial, rotating frame as if it were inertial. However, the
centrifugal force is felt as a real force, such as, for example, in a fast rotating
merry-go-round.
We now discuss the Coriolis acceleration (Eq. 5.49) and the effects of the
corresponding fictitious Coriolis force
(5.51)
Consider again the point P lying on the rotating platform. If P does not move
relative to the platform, the Coriolis acceleration is null, as in the case just
discussed. Let v′ be this velocity, which we assume, for simplicity, to be parallel
to the platform. As we have already noticed, the Coriolis acceleration, and
consequently the Coriolis force , does not depend on the position of P on the
platform and is in any case perpendicular to the relative velocity. Consider
Fig. 5.9. If the angular velocity ω is directed out of the plane of the figure, as in
Fig. 5.9a, we see the platform turning counter-clockwise. In this case, the
Coriolis acceleration is directed towards the left of the motion, and the Coriolis
force to the right. Suppose you are the point P waking or running on the
platform. You will feel a push to the right of your speed, in whatever direction
you move. Contrastingly, if ω is directed inside the drawing, as in Fig. 5.9b, and
the rotation is clockwise, the Coriolis force pushes to the left of the speed.

Fig. 5.9 Coriolis acceleration and (pseudo)force on a platform rotating. a Counter-clockwise, b Clockwise

If we were to look at the earth from some distance from its surface on the
axis, we would see the northern hemisphere rotating counter-clockwise if we
were above the North pole, and the southern one clockwise if we were above the
South pole. The Coriolis forces are the dominant causes of the circulation of
winds in the atmosphere and cyclonic and anticyclonic phenomena. We shall
discuss that in the next section.
Consider now another example, namely a material point P, standing in
equilibrium above the platform in a fixed position relative to S, i.e., to the
ground. We might think about a fly located just above the platform. The
observer in S sees P at rest. Knowing that it is subject to its weight, he
understands that another force, equal and opposite to the weight, should exist.
The force is exerted by the beating of the fly’s wings.
For the observer in S′, the description is more complicated. He sees P
moving in a circular uniform motion on a circle of radius r with velocity ωr. The
motion is accelerated with a centripetal acceleration ω 2 r. He deduces that a
force mω 2 r should act on the fly. However, he also knows, as the result of
experiments he has done in the past, such as the one we just discussed, that a
centrifugal force exists on the platform, namely a force of magnitude mω 2 r
directed outwards. Considering that the point moves on a circle, he concludes
that the centripetal force on the fly must be twice as large, namely 2 mω 2 r.
From where is this force is coming? It is the Coriolis force . In this case, ω and
v′ are mutually perpendicular; Eq. (5.49) says that the magnitude of this force is
just 2 mω 2 r and that its direction is radial, towards the center. Physics is
difficult in non-inertial frames, but the factor two is needed!
As a final example, let us go back to the first one, in which the point P is
kept still on the platform by a rubber band attached to the axis. The motion seen
by S is circular uniform. At a certain instant when we cut the band, S will see P
sliding on the platform of a straight uniform motion at the velocity it had at the
moment of the cut, directed as the tangent to the circle in that moment. Indeed,
there is no net force acting on P.
How does the observer in S′ describe the motion? To be concrete, assume the
rotation to be counter-clockwise. When the rubber band is cut, the force that is
needed in the rotating system to keep the objects standing disappears, and we
might expect to see the point P moving outside along the radius of the platform.
But this is not what we observe; rather, the point moves outside describing a
curve. The reason is the Coriolis force. Before the rubber band was cut, P did not
move on the platform, and the Coriolis force was null, but it is not so any longer
since P has started moving. The Coriolis force acts, pushing P to the right all
along its trajectory. Observing from outside, we can better understand what is
going on. When the rubber band is cut, P moves with the same velocity as the
point of the platform on which it is seated. While moving outwards, P reaches
points of the platform having higher speeds, because they are farther from the
axis, and consequently is left behind by them.

5.6 The Inertial Frame

As we have already stated, a reference frame is defined as inertial, if in that
frame the first Newton law is valid. We have also seen that if a reference frame
is inertial, any other one moving in uniform translation motion relative to it is
also inertial. Indeed, the relativity principle we saw in Sect. 5.2 states that no
experiment can distinguish between them. In other words, there is no absolute
reference frame. Finally, we have seen that the Newton laws are covariant under
the Galilei transformations.
However, nature does not necessarily behave according to our definition, and
inertial reference frames might just not exist. The answer must come, as always,
from the experiment. Basically, we need to check if we can find one reference
frame in which material points not subject to forces, or, better yet, subject to
forces of null resultant, always move in a rectilinear uniform motion. As a matter
of fact, as is often the case in physics, we proceed through successive
approximations. We can find reference frames that can be considered inertial,
within a certain approximation, namely for experiments of a certain sensitivity or
precision. For more precise experiments, we must search for frames that are
closer to the inertial one, and we can find them.
Indeed, the largest fraction of the experiments takes place on earth, and is
described in a stationary frame relative to the walls of our laboratory. These
frames can be considered inertial within a quite good approximation, although
not perfect. Indeed, earth rotates on its axis, making a turn (2π angle) in a day
(84 600 s). The corresponding angular velocity, directed from the South to the
North pole, is ω rot = 7.3 × 10–5 s–1. Figure 5.10a, for example, shows a
stationary reference frame on earth at a certain latitude λ. In this frame, the
transportation and Coriolis acceleration are present.
Fig. 5.10 Three reference frames with, acceleration towards a the rotation axis of earth a 1 = 2.4 × 10–2
ms–2, b the sun a 2 = 5.9 × 10–3 ms–2, c the center of the Galaxy a 3 = 10–10 ms–2

Let us analyze the first one, to which the centrifugal (pseudo)force

corresponds. The magnitude of this force on a point P of mass m is the product
of the mass, the square of the angular velocity (equal everywhere on earth) and
the radius of the circle on which P moves. The latter is the distance from the
axis, Rcosλ, where R is the earth radius and λ is the latitude of P. Calling a 1 the
acceleration, the magnitude of the force is
(5.52)
Let us look at the numbers. Recalling that R = 6.4 × 106
m and taking, for
example, λ = 45˚, the acceleration, which is also the force per unit mass, is
(5.53)
which is quite small, less than a per mille of the gravitational acceleration.
However, for precise measurements, it can be relevant. The Coriolis force is
usually smaller. However, it is important for large-scale phenomena, as we shall
see in the next section.
However, a stationary reference frame on earth differs from an inertial frame
for a second reason, to even smaller effect. Indeed, earth moves along its orbit,
turning around in a year, with an angular velocity of ω riv = 2×10–7 s–1 on an
orbit of radius R orb = 1.49 × 1011 m (Fig. 5.10b). The centripetal acceleration is

(5.54)
which is an order of magnitude smaller than a 1. The effects of the
corresponding pseudo force are negligible, if not for the most precise
measurements. Usually, the Coriolis force is even smaller.
Even these small effects, however, can be eliminated by choosing a reference
frame with its origin in the sun and directions of the axes stationary to the fixed
stars. This frame is inertial to an extremely good approximation, although not
perfect. Indeed, the sun is located at the periphery of our spiral galaxy (1011 stars
in order of magnitude). The sun turns around the center of the galaxy in an orbit
of radius R S ≈ 2.4 × 1020 m over a period of about 150 million years,
corresponding to the angular velocity of ω S = 7.9 × 10–16 s–1. The
corresponding centripetal acceleration is
(5.55)
This is very small indeed. However, experiments exist that are so sensitive,
they are able to detect deviations from the state of inertia even at these extremely
small levels. As a matter of fact, our galaxy moves too, in a non-uniform motion.
However, when needed, we know how to eliminate the effects.
In conclusion, inertial reference frames exist in nature at every level of
approximation we need.

5.7 Earth, as a Non-inertial Frame

As we just saw, the rotation of earth on its axis, with the angular velocity, ω
–5 –1
rot = 7.3 × 10 s . This implies that in reference frame stationary on earth
dynamical effects of the transportation and Coriolis fictitious forces exist. We
shall discuss the principal ones in this section.
We take a reference system S with the origin in the center of earth and
stationary with it. We shall use the symbols v and a for velocities and
accelerations in this frame, omitting the prime we used in the previous sections.
The acceleration of transportation is

where a O is the acceleration of the earth’s center. This is the centripetal

acceleration of its motion around the sun in a very good approximation, and r E
is the radius of the orbit, as shown in Fig. 5.11a. The Coriolis acceleration on a
point moving with velocity v in S is

Fig. 5.11 a Forces and pseudo forces on matter point P; b Displacement to east in the free fall
(exaggerated)

In S, the equation of motion of a point with mass m subject to the real force F
true is then
 (5.56)
We can distinguish the following contributions to the true force F true: the
gravitational attraction of earth F E , the gravitational attraction of all the other
heavenly bodies F O , and of any other force that might be present (air resistance,
tension of a wire, etc.), with resultant F. We re-write Eq. (5.56), grouping the
terms according to their causes,
(5.57)
The gravitational force F O is due to all the heavenly bodies different from
earth, but is largely dominated by the sun. As the diameter of earth is much
smaller than the distance from the sun, in a first approximation, we can consider
F O equal in all the points of the earth. However, the small differences that are
present are one of the causes of the tides, as we shall see in Sect. 6.4. The
acceleration produced by F O on every body is proportional to the mass of the
body. Consequently, it is the same on the surface of the earth and in its center. In
other words, it is the acceleration a O , of the earth herself. Hence, F O –m a O
= 0.
We have reached an important conclusion, which is true as long as F O can
be considered not to vary on the points of the earth, that the gravitational forces
of the sun, the moon end of the other heavenly bodies do not appear in the
equations of motion in reference frames stationary on earth. These forces are
exactly balanced by the inertial forces resulting from the acceleration that those
agents impart to the earth.
We can simplify Eq. (5.57) as

(5.58)
Now, we are ready to consider several important examples.
The first case is of a body at rest, and F is simply its weight. This is the force
we measure with a balance and that we have written as
(5.59)
where g is a vector quantity, which is equal for all the bodies in a given
position. Up to now, we have talked of it as gravitational acceleration, but we are
now ready to see that it is only approximately so. Equation (5.58) indicates that
the force pushing a body downwards that does not move (v = 0, a = 0) is
. We can say that the gravitational force of the earth on the body is

(5.60)
and write
 (5.61)
where G is the gravitational field of earth, and
(5.62)
The acceleration is the same for all the bodies in the same location.
Equation (5.58) shows that a body dropped in absence of any force other than its
weight, from a position of rest, v = 0, moves with an acceleration a = g. We can
say that g is the acceleration of the free-fall of any body, provided its velocity is
null in the considered instant. If v ≠ 0, the Coriolis acceleration is, in general,
present too.
In any case, Eq. (5.61) tells us that the weight is the sum of two
contributions: the gravitational attraction m G of the earth, which largely
dominates, and the centrifugal force due to the rotation of earth, which is much
smaller and varies with the position. We will now discuss the observable
consequences of that.
The local value of g. Suppose we take a plumb and fix it at a support. In the
equilibrium position, its weight F w , given by Eq. (5.59), and the tension of the
wire are equal and opposite. The direction is given by the wire. The distance
from the rotation axis of a point P on the surface at the latitude λ is r E = Rcosλ,
where R is the earth radius (Fig. 5.11a). The weight F w can be decomposed in a
component, let us call it F w,r , directed to the center of earth, and a component,
F w,θ, in the direction of the meridian, to the North in the northern hemisphere
and to the South in the southern one. The two components are

(5.63)
The centrifugal term, the first one, is zero at the poles and maximum at the
Equator. The tangential component is zero both at the poles and at the Equator.
In these locations, but not elsewhere, the weight is precisely directed to the
center of earth. As for the magnitude, the measured values are g = 9.832 ms–2 at
the poles and g = 9.780 ms–2 at the Equator. If we approximate the shape of the
earth surface with a sphere, all its points are at the same distance from the center,
and if the mass distribution inside the earth is spherically symmetric, the
gravitational term G is equal everywhere. It should be equal to g at the poles,
G = 9.780 ms–2. Let us check by giving an estimate, starting from g at the
Equator.
 This value is close, but still a bit smaller than what we found from g at the
poles. The main reason for that is that earth is not really spherical but somewhat
squeezed at the pole, an effect of the centrifugal forces. Consequently, the poles
are a bit closer to the center than the Equator.
Notice however, that small differences on the value of g in the different
points of the surface are present, due to the local geology.
Absence of weight. If we measure the weight of an object with a balance on
the space station, we find it to be zero. Such is also the weight of all the objects
in the station, and in every artificial satellite. The arguments we just made are
still valid, if we put the station in the place of earth, and consider the earth as an
external body, as the sun, the moon and the other planets are. The spaceship is
small enough for the gravitational force of those bodies to be considered equal at
all the points of the ship. This force is exactly balanced by the inertial force to
the acceleration of the spaceship. If its engines are shot, the ship freely falls
under the action of gravitation. In this case, the equivalent of the weight on
earth, namely the gravitational attraction of the ship on the body inside it, is
completely negligible, F w = 0. The centrifugal term to the weight in the space
ship is also negligible because the ship does not rotate appreciably. The weight
in the ship is zero.
Eastwards shift in the free-fall . If a material point P of mass m is dropped
with null initial velocity at a height h from the ground, it initially falls under the
action of the weight, F w . However, as soon as its velocity, v, is appreciably
different from zero, a second inertial force, the Coriolis force , enters into action.
It is
(5.64)
The velocity v relative to earth is in the plane containing the earth’s axis and
point P, namely the plane PON in Fig. 5.11a. Consequently, the Coriolis force is
perpendicular to this plane. Considering that the direction of the angular velocity
is from South to North, and that v is downwards, we see that the Coriolis force is
toward East in both hemispheres. The situation is shown in Fig. 5.11b, where AB
is the direction of the plumb, i.e., the direction of F w (no Coriolis force on the
plumb that does not move) and C is the point in which the body reaches the
ground, falling from the height h. The shift from the vertical BC is very small,
and exaggerated in the figure. Let us calculate it.
We take a reference with the z-axis vertical, i.e., in the local direction of the
plumb, and the x-axis horizontal towards the East. Within a good approximation,
we can take the magnitude of the velocity to be υ = gt, as in the vertical fall. Its
direction is opposite to the z-axis. The equation of the component of the motion
on the x-axis is

We solve the equation by integrating twice on time and imposing the initial
conditions x(t) = 0, (dx/dt(0)) = 0, obtaining

(5.65)
The time of the fall is, with good approximation, , and we have

(5.66)
For example, at the latitude of 45˚ and a fall from h = 50 m, the eastward
shift is x ~ 5 mm, which is quite small, but has been measured, carefully
eliminating perturbing effects.
Horizontal wind circulation . As is well known, the earth’s atmosphere in a
certain instant contains zones of high pressure and zones of low pressure.
Naively, one would expect winds to blow from the former to the latter in the
direction of the pressure gradient. However, the direction of the winds is
substantially perpendicular to that, moving along the isobars, as you can see
watching weather forecasts on TV. The effect is due to the Coriolis force .
Figure 5.12 summarizes the situation. H is the pressure maximum, L a
pressure minimum, in the Northern hemisphere. Hence, the earth’s angular
velocity direction is out of the paper and the Coriolis force is directed,
perpendicular to the velocity, to the right. Consider, for simplicity, a horizontal
wind at constant velocity (in magnitude). Suppose we insulate a small mass of
air within an ideal film and follow its motion. Two vertical and two horizontal
forces act on our mass. The vertical ones are the weight and the Archimedes
force. As the motion is horizontal, they are equal and opposite. The horizontal
forces are the pressure (true) force and the Coriolis (pseudo) force. The pressure
force acts on the surfaces of our gas mass. The pressure on its left-hand face
pushes to the right, while the pressure on the right face pushes to the left. If the
pressure were equal on the two sides, the neat force would be null. However, if
there is a pressure maximum on the right of the gas mass we are following, as in
Fig. 5.12a, there is a neat pressure force F (P) pushing to the left. The Coriolis
force has an equal and opposite direction. Consequently, the two forces may
balance each other, or result in the right value being the centripetal force for the
curvature of the wind trajectory. This can happen only if the wind circulates in a
counter-clockwise direction around a pressure maximum (anticyclone).
Contrastingly, it must circulate clockwise around a minimum (cyclone), as in
Fig. 5.12b. The two situations are inverted in the southern hemisphere.

Fig. 5.12 Isobars around pressure maximum (left) and minimum (right-hand), in the Northern hemisphere
and the forces on a mass of air

Let us look at the orders of magnitudes. The magnitude of the Coriolis force
on an air mass m moving with horizontal speed υ at the latitude λ is
(5.67)
For example, the force on a kilogram of air, which is about 1 m3, moving at
10 m/s at 45˚ is about 10–3 N. This should be compared to the pressure forces on
the same volume. To be of the same order of magnitude, the pressure forces on
two opposite sides of our cubic meter volume should be different by 10–3 N.
This corresponds to a pressure difference of 10–3 Pa, being the surface unitary.
Hence, the pressure gradient should be of 10–3 Pa/m, corresponding, say, to a
distance of 100 km between two isobars of 100 Pa difference. This is reasonable
(have a look at the weather maps).
The Foucault pendulum . A simple pendulum abandoned in a non-
equilibrium position with null velocity oscillates in a vertical plane. However, if
we watch carefully for a long enough time, along the order of one hour, we can
see that the oscillation plane rotates relative to the laboratory, i.e., relative to a
reference fixed on earth. The reason for the rotation is, once more, that the frame
is not exactly inertial. As a matter of fact, the oscillation plane is fixed in an
inertial frame, relative to which the earth rotates, as in Fig. 5.13.
Fig. 5.13 The Foucault pendulum

While the effect has been known since its first observation by Vincenzo
Viviani (1622–1703) in 1661, the main experiment and its correct interpretation
were done by Léon Foucault (1819–1868) in 1851 in the Pantheon of Paris. His
pendulum was 67 m long and had a 28 kg mass.
A similar situation, shown in Fig. 5.14, helps in our understanding. There,
we have a pendulum, supported on a turning platform. If we put the pendulum in
oscillation and the platform in rotation, we observe the oscillation plane
remaining fixed, as expected, and the platform rotating under the pendulum. We
can easily imagine what an observer on the platform would see, namely the
plane of oscillation rotating in the opposite direction.

Fig. 5.14 Pendulum on a rotating platform

In this way, we easily understand what happens on earth if we are on a pole.

Here, the angular velocity vector ω rot is normal to the “platform”, the earth
surface, exactly as in that experiment. From the point of view of an inertial
observer, the oscillation plane is constant, and he sees the earth turning relative
to it. He understands why an observer at the pole sees the oscillation plane
rotating and making a complete turn in 24 h. In the reference fixed to earth, the
equation of motion is, once more, (5.58), with F the tension of the wire
(neglecting air resistance). The Coriolis force is normal to the
oscillation plane, and causes its rotation.
If the experiment is done not at the pole but at a latitude λ, we must pay
attention to the vector characteristic of ω rot. We decompose it in a horizontal
component, namely parallel to the ground in our position, ω h, and a vertical one
ω v: ω rot = ω h + ω v. We further decompose the horizontal component, which is
still a vector, in its components parallel, ω p, and normal, ω n, to the oscillation
plane and write Eq. (5.56) as
(5.68)
The term has the direction of the wire. Its effect is to change the
tension a bit, but it has no effect on the oscillation plane. The term is
perpendicular to the oscillation plane and causes its rotation. The third term
is also perpendicular to the rotation plane, but is very small. Indeed,
as we can see in Fig. 5.13b, it is proportional to sinθ, where θ is the angle
between the wire and the vertical and is small, for small oscillations. We can
then simplify Eq. (5.68) by writing
(5.69)
In concussion, the motion is similar to that at the pole with the sole
difference being that in place of the angular velocity ω, we must consider its
component along the local vertical, of magnitude
(5.70)
The oscillation plane makes a complete turn in the period

(5.71)
At 45˚ latitude, in one hour, the plane rotates by 10.6˚.
Figure 5.13c shows the projection on the horizontal plane of the trajectory of
the Foucault pendulum. The vector ω v is normal to the drawing towards the
observer. The Foucault force is always directed normally to the velocity to the
right of the direction of motion. The force bends the trajectory, as shown with
exaggeration in Fig. 5.13c. Suppose that the pendulum is initially in A and
abandoned with null velocity. Initially, when the Coriolis force is very small, the
pendulum heads to A′. But as soon as the velocity becomes appreciable, the
Coriolis force pushes to the right, bending the trajectory. The pendulum reaches
point B, where it stops. When the velocity has again sufficiently increased, but in
the opposite direction, the Coriolis force pushes in the opposite direction too,
although still to the right of the motion. The pendulum reaches C, etc.
In the Foucault experiment, the length of the pendulum was large, l = 67 m,
corresponding to a period T = 16.4 s. With such a long period, the lateral shift
can already be observed in a single oscillation. The oscillation amplitude was
A = 3 m. At the Paris latitude, sinλ = 0.753 and the rotation period is T
rot = 3.8 h = 14 480 s = 31,8 h = 14480 s. In an oscillation period T, the plane
rotates at the angle 2πT/T rot. Hence, the shift of the oscillation extreme in one
period is s = 2πAT/T rot = 2.7 mm.
Moreover, the length is important for another reason to which we can only
hint. In practice, it happens that the stress forces always present in the wire and
in the hook supporting the pendulum result in a spurious rotation of the
oscillation plane. The effect is slow, but important for observations of several
hours. It can be shown, however, that it is smaller for longer lengths.

5.8 The Eötvös Experiment

In Sect. 2.5, we have seen how Galilei and then Newton experimentally
established the identity between inertial and gravitational mass. This is a very
fundamental issue, and experiments have been done, and are still being done, to
increase the precision within which the equality is verified. We discuss here the
beautiful experiments conducted by the Hungarian physicist Loránd Eötvös
(1848–1919) in the last years of the XIX century.
In this chapter we gave a number of examples of the effects of the inertial
forces, the pseudo forces that appear in non-inertial frames. The inertial forces
acting on a mass are proportional to the mass, just like the gravitational force.
There is an important difference, however, as inertial forces are proportional to
the inertial mass m i , and the gravitational force is proportional to the
gravitational mass m g . Suppose the ratio between the two types of mass to be
different for different substances. We could then hang spheres made of the two
substances to two wires and look for any small difference in the directions of the
wires. In this section, we shall use different symbols, m i and m g , for the two
types of mass.
Consider a body hanging from a wire fixed in Ω, as in Fig. 5.15, at a point at
the latitude λ. The distance from the axis is r E Rcosλ, where R is the earth’s
radius. The centrifugal force has a direction perpendicular to the axis outwards
and magnitude

Fig. 5.15 The basis of the Eötvös experiment

(5.72)
and the gravitational force

(5.73)
If for two substances, m i and m g are different, the angle between the two
forces is also different, and so is the direction of the wire. As we saw in Sect.
5.6, the centrifugal acceleration on the earth’s surface is of the order of the per
mille of the gravity acceleration. Correspondingly, the sought-after effects can
be very small.
The Eötvös experiment directly compares the angles of wires to which
spheres of different substances are attached. The two wires are attached to the
extremes of a rigid bar. The bar is suspended by a metal wire that acts as a
torsion balance, as shown in Fig. 5.16, similar to what we described in Sect. 4.7.
Fig. 5.16 The scheme of the Eötvös experiment

Figure 5.16a shows the system in perspective, with Fig. 5.16b looking at it
parallel to the bar. If the ratio m i /m g is different for the two spheres, the
directions α and β of the two tensions are a bit different. This produces a
moment on the bar, due to the horizontal components of the two tensions, that
rotates it about the wire from which it hangs. Under rotation, the wire develops
an elastic moment, which increases with the angle. At the equilibrium angle, the
two moments are equal and opposite. Measuring the angle, the torsion balance
gives the moment.
The result of the very sensitive Eötvös experiment was null, allowing him to
give the upper limit , namely that the difference, if any, is less
than 5 parts per billion. An experiment of the same type by Robert Henry Dicke
(1916–1997) in the 1960s established the even smaller limit of
.

5.9 Problems
5.1. A kid sits in a carriage moving on straight rails. (a) If the speed of the
carriage is constant, in which direction should he launch a ball to take it
back in his hand without moving? In which direction if the carriage
accelerates forwards?

5.2. A train travels on straight horizontal rails at the velocity υ 0 = 30 m/s.

 Reaching a station, its stops, with constant acceleration, in s = 150 m. A
suitcase of mass m = 10 kg lies on the floor, with dynamical friction
coefficient µ d = 0.20. During the braking, it slides along the corridor. (a)
How much is its acceleration relative to the ground during this time? (b)
Which is the velocity of the suitcase when the train stops? (c) After the train
stops, the suitcase continues to slide for a while and then itself stops. Which
was the total displacement of the suitcase on the floor?

5.3. A man measures his weight in a lift, which is at rest, using a spring and
balance, and finds it to be 700 N. With the lift moving, he repeats the
measurement and finds it to be 500 N. What can he determine about the lift
acceleration? And about its velocity?

5.4. A person sits in a chair standing on the platform of a merry-go-round,

which is turning. The person holds a plumb. Draw separate force diagrams
for the plumb, the wire, the person, the chair and the platform. Describe
each of the forces in words. Identify the action reaction pairs, both for a
frame stationary on earth and for one stationary on the platform. In the
latter case, specify which of the forces are inertial.

5.5. A kid sits on a merry-go-round that turns at angular velocity ω, while his
friend is on the ground. The resultant of the forces on the latter is zero. (a)
What is the motion of the second kid seen by the first? (b) What is his
acceleration? (c) What are the forces causing it?

5.6. An old vinyl disk rotates at 33 turns per minute. Its radius is r = 15 cm. An
insect walks from the center towards the border. Will it be able to reach it if
the static friction coefficient is µ s = 0.1?

5.7. A tennis player at 45˚ latitude is imparting to the ball a speed of 100 km/s,
which we assume to be initially horizontal. Willing to hit ground at a
distance of 50 m, should he take into account the Coriolis force?
© Springer International Publishing Switzerland 2016
Alessandro Bettini, A Course in Classical Physics 1—Mechanics, Undergraduate Lecture Notes in Physics,
DOI 10.1007/978-3-319-29257-1_6

6. Relativity
Alessandro Bettini1
(1) Dipartimento di Fisica e Astronomia, Università di Padova, Padova, Italy

Alessandro Bettini
Email: alessandro.bettini@pd.infn.it

In the preceding chapters we have seen how the Galilei-Newton mechanics is

able to describe with simple laws an enormous number of phenomena both at the
everyday scale and at cosmic level. Newtonian mechanics is one of the major
conceptual constructions of human genius. However, the validity of the theory is
limited on two sides, on the side of high velocities and on the side of small
dimensions.
Newtonian mechanics is no longer valid for not very small velocities
compared to speed of light. The latter is very large, about 3 × 108 m/s. The
velocities of all the objects we have to deal with on earth, of planets and of the
majority of heavenly bodies are very small in comparison. The velocity of the
earth and the planets around the sun, for example, are of the order of one in ten
thousand of the speed of light. Consequently, Newtonian mechanics gives an
extremely good approximation for these phenomena. In this chapter we shall see
how the theory needs to be extended at speeds comparable with the speed of
light, in relativistic mechanics.
Classical mechanics, as are called both the Newtonian and relativistic
mechanics, cease to apply for objects at molecular or smaller scales, of the order,
say, of nanometers. These are one thousand times smaller than microbes. The
correct theory, valid at all orders of magnitude is quantum physics. Classical
mechanics is the limit of quantum physics for sufficiently large dimensions. In
this book we shall only warn the reader of the limits of classical mechanics when
needed.
As we have just stated, in this chapter we shall study the fundamental
principles of relativistic mechanics and of the high velocity phenomena it
describes. In the last part of the XIX century Maxwell formulated a set of
equations that completely describe with great accuracy all the electromagnetic
phenomena. However, these equations seemed to be in contrast with the
relativity principle. If it was so it would have been possible to experimentally
find an absolute reference frame. Astronomical observations and accurate
experiments, culminated in the experiment by Michelson and Morley in 1887
(described in Sect. 6.2), showed that was impossible. An in depth criticism of the
fundamental concepts, in particular on the measurement of time, followed,
leading to development of special relativity mainly by A. Lorentz, H. Poincaré
and A. Einstein, which was substantially completed in 1905.
The relativity principle was found to be universally valid, but new
transformations of coordinates and time, valid between two inertial frames, the
Lorentz transformations were established. The Lorentz covariance, first
established for electromagnetism, must be valid for all physics laws. We shall
emphasize this point after having followed the historical path.
In the sections from Sects. 6.4 to 6.6 we shall study how the concepts of
simultaneity, time interval and distance must be revised. In Sect. 6.7 we shall
find the law of addition of velocities, which shows, in particular, that the speed
of light is the largest possible one. Space and time become completely correlated
concepts and should be considered as a single four-dimensional manifold, space-
time, which we shall study in Sect. 6.8.
In Sects. 6.9 and 6.10, we shall discuss relativistic dynamics and see how the
fundamental concepts of mass, linear momentum and energy change. In
Sect. 6.11 we shall find the relativistic form of the second Newton law.
As already mentioned, all the physics laws should be relativistically
covariant. This in fact is the case. We shall give a few hints on that in Sect. 6.12.
Finally, in Sect. 6.13 we shall give a summary of the differences and
equalities between Newtonian and relativistic mechanics.

6.1 Does an Absolute Reference Frame Exist?

In Chap. 5 we studied the covariance of physical laws under transformation
between two reference frames. We have seen that inertial frames exist, which
are by definition the frames in which the inertia law is valid. In such frames also
the second and third Newton laws hold. In this chapter we shall consider only
inertial frames. Formally, we can state that the physics laws have the same form
in two inertial frames in the following cases. The two frames have no relative
motion but differ for a translation or a rotation of the axes, or the two frames are
in a relative uniform translation motion. In other words we have no means for
experimentally observing differences between one location or another
(invariance under translations, the space is homogeneous), or between one
direction or another (invariance under rotations, the space is isotropic), or, to
establish whether a frame is moving in a uniform translation motion or not
relative to another inertial system. The latter property is the relativity principle
established by Galilei .
We observe now that the set of all translations of the reference frames , all
their rotation, and all the transformations between two inertial frames in relative
uniform translation motion have the important mathematical characteristics of
being a group . Let us define what this means.
Consider a set of transformations A, B, C… for which a product
operation, ⊗ , is defined. The set is a group if the following conditions are
satisfied:

1. For any pair of transformations A and B of the set, the product C = A ⊗ B is

also a transformation of the set. The product is associative, namely
A ⊗ (B ⊗ C) = (A ⊗ B) ⊗ C.

2. The set includes the identity transformation E, such as A ⊗ E = A.

3. For every transformation A of the set, the inverse transformation, called A −1,
exists, such as A ⊗ A −1 = E

With product B ⊗ A we mean that we first apply transformation A and then,

on the result of that, transformation B.
To better understand that, consider the example of the static translation s or
displacements, which we take in two dimensions for simplicity. Suppose that the
transformation A is the displacement a in the x direction from the coordinates S
(x, y) to the coordinates Sʹ (xʹ, yʹ), as shown in Fig. 6.1.
Fig. 6.1 Two translations and their product

The transformation A is

(6.1)
Let the static transformation B be the displacement b in the yʹ direction of the
result of A, which is Sʹ(xʹ, yʹ), to Sʺ (xʺ, yʺ), namely

(6.2)
The product of the two is the transformation from S(x, y) to Sʺ (xʺ, yʺ). Is it a
translation? These relations are

(6.3)
which is the expression of a translation too. It is also easy to see that the
associative property holds. Property 2 for being a group is satisfied.
The other two properties are also satisfied. The identity is the translation of
null displacement (do nothing). Given a translation by a certain displacement,
the static translation of the opposite one is also such a translation. Doing one
after the other leads to the identity. In conclusion, static translations form a
group.
Particularly important are the rotation s. We recall that the covariance of the
laws under rotations of the axes correspond to the fact that the quantities
appearing in the equations that express the laws (position vector, velocity,
acceleration, force, energy, etc.) must have well-defined transformation
properties under rotations. They should be scalar , pseudoscalar , vector s or
pseudovector s, and both sides of the equation must share the property.
 Consider now the time. In Newtonian mechanics time is the same in all
reference systems. We need to look at that more carefully. The time interval, as
all the physical quantities, must be operationally defined. It is not obvious that
the operations to measure the time interval between two events is the same for an
observer at rest relative to the events and one moving relative to them. As we
shall see, this is not true at high enough velocities, in the domain of relativistic
physics.
We state immediately that the covariance properties of physics laws relative
to translations and rotations remain equal to those we know, in relativistic
physics. The changes are in the covariance properties between two frames in
relative uniform translation motion . Let us consider two (inertial) reference
frames. The first one has the coordinates x, y, z and time t. We call it S (x, y, z, t).
The second frame, Sʹ(xʹ, yʹ, zʹ, tʹ), has axes parallel to the first one. The relative
velocity is along the, overlapping, x and xʹ axes. The constant velocity of Sʹ, or
of its origin, is v Oʹ , is in the positive direction of x. We choose the origins of the
times in both frames in the instant in which Oʹ and O coincide. Figure 6.2 shows
the situation.

Fig. 6.2 Two reference frames in relative motion

The covariance of the laws under transformations between two inertial

frames is defined once the transformation equations are defined, namely the
relations between coordinate and time in Sʹ and in S. The transformation
equations we know, including the relation between the times, are

(6.4)

More generally, for a generic direction of the velocity v O , of the points of Sʹ

the transformations are
 (6.5)
These are called Galilei transformations . An important property of the
Galilei transformation is that they form a group .
The time intervals, in particular, are equal in the two frames. In other words,
time is absolute, independent of the motion of the observer. This implies that it is
possible to synchronize the clocks in S with the ones in Sʹ independently of the
relative velocity of the two frames. As is time, simultaneity is absolute. If two
events are simultaneous in S they are simultaneous in Sʹ too and in any other
(inertial) frame, whatever its velocity.
The Galilei and Newton laws of mechanics that we have studied were
established when only one of the fundamental interactions, gravitation, was
known. Three more fundamental interactions were discovered in the following
centuries. The first one was electromagnetism , including electric and magnetic
phenomena and will be treated in the third volume of this course. The other two
are the strong interaction between quarks in the nucleons and the weak
interaction responsible, in particular, for beta decay. Both of them act at the
nuclear and subnuclear scales and are quantum phenomena. Do they obey the
invariance principles we have discussed, in particular the relativity principle?
Let us see.
The study of electromagnetic phenomena was developed in the second half
of the XVII and, mainly, in the XIX Century. In 1820 Hans Christian Ørsted
(1777–1851) discovered that electric currents generate magnetic fields, linking
for the first time electricity and magnetism. Between 1820 and 1826 André
Marie Ampère (1775–1836) completely clarified the relation between electric
currents and magnetism with a series of experiments. In 1831 Michael Faraday
(1791–1867) discovered the electromagnetic phenomena: magnetic fields
varying in time give origin to electric fields. The progress became rapid and in
1865 James Clerk Maxwell (1831–1879) developed the complete theory of
electromagnetism. All the electric and magnetic phenomena are described by a
set of differential equations, called the Maxwell equations . In addition, the
theory predicted a new phenomenon. Electric charges in acceleration produce
electromagnetic waves , which propagate with a well-defined velocity. This
velocity can be expressed in terms of two quantities that measure the strength of
the force between two electric charges at rest and between two stationary electric
currents respectively. Maxwell himself accurately measured these quantities and
found the resulting value of the velocity to be, in round figures,
(6.6)
This is just equal to the velocity of light . And Maxwell noticed that
 the only use made of light in the experiment was to see the instruments.

He concluded

that light is an electromagnetic disturbance propagated through the field

according to electromagnetic laws.

The direct experimental confirmation of the existence and of the foreseen

properties of the electromagnetic waves was very difficult. Heinrich Rudolf
Hertz finally succeeded in that in 1887.
The Maxwell equations led to the unification of electric, magnetic and
optical phenomena. However, soon they showed an unexpected behavior. Their
form changes between two inertial frames when coordinates and time are
transformed according to the Galilei transformations Eq. (6.5). It looked like
Maxwell equations did violate relativity principle . If it were so, it should have
been possible to design and perform electromagnetic and optic experiments able
to establish an absolute reference frame .
Suppose for example we have a light source emitting a light pulse in the
positive x direction in Fig. 6.2. Let c be its velocity in S. Notice that light is a
wave phenomenon, similar to sound or sea waves. Consequently, its propagation
velocity is independent of the velocity of the source relative to the observer.
However, differently from the other mentioned cases, light propagates in a
vacuum too. Indeed it comes to us from very distant stars. Consequently there is
no substance perturbed by the wave and supporting its motion. This fact, which
is clear to everybody now, was not so at the end of the XIX Century, when the
existence of a substance pervading all space was assumed, the luminifer (light
supporting) ether . The ether hypothesis has been a serious problem in the
development of electromagnetic physics.
Anyway, as the speed of light is independent of the motion of the source, it
should transform as any other velocity, as we have by derivation of Eq. (6.4) in
the case of our example
(6.7)
As any other velocity, the speed of light should be different for the observer
in S and in Sʹ. If we then measure the speed of light in Sʹ and find it different
from c we would establish that S is the absolute reference frame, namely the
only one, amongst all the inertial frames in which the velocity of light has the
value of Eq. (6.6).
More specifically, the non-covariance of Maxwell’s equations under the
Galilei transformations requires us to establish which of the following
alternatives is the right one.

(1) The relativity principle is valid for the Newton laws of mechanics but not for
the Maxwell laws of electromagnetism. The Galilei transformations are
correct. This implies the existence of an absolute reference frame, which
should be experimentally found.

(2) The relativity principle is valid for both the Newton laws and
electromagnetism. The Galilei transformations are correct, but the Maxwell
equations are wrong. In this case we should find modifications to the
Maxwell equations that are necessary to have them covariant under Galilei
transformations and then experimentally control whether the predictions of
these modifications exist or not.

(3) The relativity principle is valid for mechanics and electromagnetism. The
Maxwell equations are correct, but the transformation equations between
reference frames are not the Galilei transformations. In this case we must
find new transformations, different from the Galilei ones and such as to
insure the covariance of the Maxwell equations. In addition, the Newton
laws would no longer be any more covariant under the new transformations.
We should find the modifications needed to guarantee the covariance also of
mechanical laws and experimentally verify whether the consequences of the
modifications we made are correct. The historical process leading to the
clarification of the problem was not straight, but rather along winding paths.
After the important contributions of Hendrik Antoon Lorentz (1853–1928),
in 1905 two fundamental articles were separately published, the first by
Henry Poincaré (1854–1912), the second a few weeks later by Albert
Einstein , that laid down the complete theory. It became known as special
relativity.

The crucial experiment to choose between the above stated alternatives is the
measure of the speed of light in inertial frames in relative motion, allowing us to
verify whether it is the same or not. The expected effects however, are extremely
small and very difficult to detect. The experiment was done by Albert Abraham
Michelson (1852–1931) in 1881 and, in a much more sensitive version, together
with Edward William Morley (1838–1923) in 1887. We shall describe the 1887
experiment in the next section. We shall see how it showed that the speed of
light is the same in all reference frames, so excluding alternatives (1) and (2).

6.2 The Michelson and Morley Experiment

We start with a bit of history. In 1879 Maxwell studied the possibility of
establishing the absolute motion of earth relative to the reference in which the
speed of light is c, namely the absolute reference frame , on the basis of
astronomical data. The absolute reference is the frame in which the ether, which
was thought to exist, is still. If this frame exists, it should be at rest relative to the
fixed stars, according to astronomical observations. We do not know the velocity
of the earth relative to this hypothetical frame, but we know that it should be at
least the velocity of the earth in its orbital motion around the sun. This is about υ
E ≈ 30 km/s in magnitude and varies in direction throughout the year. Let us
assume this velocity to be, in order of magnitude, what we have to detect. Its
ratio to the speed of light is

(6.8)
which is a very small value. Maxwell established that only in astronomical
phenomena could one expect effects of the first order in β E . In laboratory
experiments, in which the light leaves from a point, moves to a certain distance
and comes back to the starting point, or close to it, only effects of the second
order were expected, namely of the order of 10−8. This is really a very small
number. Maxwell’s argument is the following.
Suppose that in our laboratory, namely in a reference in which the earth
moves with speed υ E, we place a bar of length l in the direction of the motion. At
one end of the bar we have a source emitting flashes of light and a detector of
light nearby. At the other end there is a mirror sending the light pulses back to
the detector. The light pulse travels the distance l from the source to the mirror at
velocity c + υ E and when going back from the mirror to the detector at velocity
c – υ E . The total time is then

(6.9)
Now, 2 l/c would be the round-trip time if the bar were not moving. This is a
very short time. But the time to measure is of it. Maxwell concluded
that such an experiment was impossible.
The young, 25 years old, officer of the USA navy Albert Abraham
Michelson , who had already performed an accurate measurement of the speed
of light, did not accept as obvious the impossibility of a laboratory experiment
sensitive to the second order. Rather he worked on the problem and in 2 years
found a solution. In 1881, he had already a first result. The sensitivity of this
experiment was enough to detect the effect down to one half of the prediction.
The result was null. However, the conclusion was so important that a
confirmation was needed. Michelson, now with Morley, designed and performed
in 1887 a second experiment sensitive to effects 40 times smaller than the
predictions. Again the result was null.
The Michelson-Morley experiment is based on the employment of the
interferometer shown in Fig. 6.3, which had been developed by Michelson
himself ( Michelson interferometer ).

Fig. 6.3 The Michelson interferometer

The source L emits a monochromatic line. This means that the wave is a
sinusoid. The distance between two consecutive maxima is the wavelength
(λ = 0.6 µm). Each point on the wave moves up and down periodically with a
period T. In an equivalent manner we can say that if we looked at the wave
passing on a fixed point, the time interval between the passage of two maxima
would be T.
 Consequently the ratio between wavelength and period is the speed of the
wave. If this is c we have
(6.10)
In the Michelson interferometer, the light beam is divided in two by a
semitransparent mirror M at 45° with the incident beam direction. One of the two
beams after this mirror reaches the totally reflecting mirror M 1, is reflected
back, reaches again M, and is reflected towards the telescope C. The other beam
on the arm 2 is reflected back by M 2 and, after M, which partially transmits it,
rejoins with the first beam. The lengths of the two arms are made as equal as
possible. The two light waves are in phase when they leave M for the first time
and are also in phase when they recombine, namely in the telescope, provided
that the times, call them t 1 and t 2, are identical, or differ exactly by an integer
number of periods. This is the situation drafted in Fig. 6.4a. In this situation, the
signal they originate when they recombine is a maximum (constructive
interference).

Fig. 6.4 a waves in phase, b waves in phase opposition

If the travelling times t 1 and t 2 differ by half a period, or an odd number of

half periods, as in Fig. 6.4b, the two waves are in phase opposition and cancel
each other giving a zero signal (destructive interference). In the intermediate
cases, the intensity is intermediate too. If these were the conditions of the field
seen by the observer through the telescope, it would be clear in constructive,
dark in destructive interference. In practice however, the planes of the two
mirrors are never exactly at 90°. Consequently, the conditions of constructive
and destructive interference alternate through the width of the beam in the visual
field. The observer sees a series of clear and dark bands, called interference
fringes . One could change the planes of the mirrors by adjusting screws in order
to have the fringes horizontal, as in Fig. 6.5. A reference wire in the eyepiece
was used to measure the position of the fringes.
Fig. 6.5 The interference fringes before and after the rotation of 90° of the apparatus

We now evaluate the difference between the times t 1 and t 2. It is due to two
causes. The first one is instrumental and due to the fact that the lengths, say l 1
and l 2, of two arms are never exactly equal. Notice that here exactly means to be
so within a small fraction of the wavelength, namely a few dozens of
nanometers. The other cause is what we want to measure, namely a difference in
the light speed, relative to the instrument between the two arms due to the
motion of the earth.
Suppose we have aligned the arm 1 parallel to its transportation velocity and
evaluate t 1. In the path from M to M 1 the speed of light is c + υ E and in the path
back from M 1 to M is c – υ E . We have already calculated the round-trip time,
Eq. (6.9). We can write

(6.11)
We now calculate the time t 2. If earth moves with velocity υ E relative to the
absolute frame, in the time t 2 is displaced by υ E t 2 as shown in Fig. 6.6.
Looking at the figure we write

Fig. 6.6 Path of the light in arm 2

and hence
 (6.12)
Notice that we have just calculated t 1 in the frame fixed to earth and t 2 in
the supposed absolute frame. This was allowed because we have assumed the
Galilei transformations to be valid, in particular the time to be absolute. Notice
also that, as anticipated, the effect is of the second order, namely as .
The difference between the two times is then

(6.13)
As we anticipated, the two times differ by the searched for effect, i.e. the
term in , and for the difference between the arm lengths, 2(l 2 − l 1)/c. To get
rid of the second effect, Michelson employed a measurement method by
comparison. The comparison was between a measurement in the just described
conditions and one after rotating the whole apparatus by 90°. The time
difference, say ∆tʹ, is Eq. (6.9) with inverted l 1 and l 2, namely

(6.14)
We take the difference between the two differences and obtain

(6.15)
If the difference between the differences is zero, the position of the fringes
seen by the observer remains fixed relative to the reference wire when we rotate
the apparatus. If it is equal to one period the fringe pattern moves by one fringe.
In general, the number ∆n (not integer in general) of fringes crossing the
reference wire during the rotation, is given by

(6.16)
where, in the last member we have used Eq. (6.15) and introduced the mean
value l of the lengths of the two arms.
In the 1881 experiment the length of the arms was l = 1.2 m, corresponding
to an expected shift of Δn = 0.04 fringes. Michelson was able to appreciate a
shift of 0.02 fringes. He did not observe any and concluded that:

The consequence of a stationary ether results therefore contradicted by the

facts and it must be concluded that the hypothesis of the ether is false.
 The second experiment is shown in Fig. 6.7. The optical path, namely the
path of the beams, is increased to l = 11 m, having the beam going back and
forth on its arm eight times with a set of mirrors (Fig. 6.7b). The 90° rotation
was an extremely delicate operation. Any vibration even by a small fraction of a
wavelength had to be avoided. Everything had to be stable at this level.
Michelson and Morley mounted the interferometer on a massive granite bench,
which was floating on a mercury bath. The shift expected in the ether hypothesis
was now Δn = 0.40 fringes. Figure 6.4 reproduces the fringes before and after the
rotation. No shift can be seen. The sensitivity was one hundredth of a fringe,
corresponding to a distance of 6 nm. Figure 6.8 shows the result of the
measurements, which are the full lines. The dotted curves are 1/8th (to enhance
the visibility of a possible difference) of the expectations assuming the Galilei
transformations . No effect was detected repeating the experiment in day-time
and during the night, to check for any effect of the rotation velocity of the earth.
The conclusion was definitive: the experiment cannot establish the movement of
earth. This is an example of how a null result can give extremely important
information.

Fig. 6.7 The Michelson Morley experimentMichelson Morley experiment

Fig. 6.8 Observed shifts (continuous lines) and 1/8 of the expectations (dotted curves—to keep them inside
the diagram) for Galilei transformations

A first attempt to explain the result was done in 1889 by George FitzGerald
(1851–1901) and independently in 1992 by H.A. Lorentz . They advanced the
hypothesis that the objects, when in motion, contract, only in the direction of the
motion and not in the perpendicular ones. The contraction was able to cancel the
effect expected in the ether hypothesis. It was an ad hoc, and wrong, hypothesis
but an important step towards relativity theory.
In the following years the Michelson experiment was repeated with
increasing precision, always with a null result. Other experiments sensitive to the
absolute velocity were done, again with null result. In 1904 H. Poincaré , after a
careful analysis of the experimental evidence, drew the conclusion that the
relativity principle (so he named it for the first time) holds for all physical laws.
His words, similar to those of Galilei three centuries before him, are:

According the Relativity Principle the laws of the physical phenomena must
be the same, whether an observer is fixed, or for an observer moving in an
uniform translation motion: so that we have no means, and could not have
any, of discovering if are or are not carried along in such a motion.

His second conclusion was that the speed of light is the same in all inertial
reference frames, i.e., the speed of light is invariant .
From our side, we concede that only the third alternative of those considered
in the previous section can be valid. We must now, first of all, find new
transformation laws, in place of the Galilei transformations.

6.3 The Lorentz Transformations

We need now to find new transformation laws of coordinates and of time
between two inertial frames in relative uniform translation motion. They must be
such, in order to guarantee the relativity principle , that the Maxwell equations
are covariant, namely maintain their form, under such transformations. The
invariance of the speed of light is an immediate consequence of that. These are
the Lorentz transformations . After having recalled the important historical
elements, we shall give the result without demonstration and shall discuss it.
Finally we shall state which are the basic assumptions under which the Galilei
and Lorentz transformations are valid.
The Lorentz transformations were, found as the result of a difficult
theoretical effort in several subsequent steps of improving precision, by Hendrik
Lorentz between 1895 and 1904 and, with a further small correction, in final
form, by Henri Poincaré in 1905, who published the result on the 5th of June
1905. Albert Einstein reached the same result on the 30th June, when his
fundamental article was sent for publication.
Consider once more the two inertial reference frames S and Sʹ represented in
Fig. 6.2. We have in both frames rulers along the Cartesian axes, to measure the
coordinates, and, in every point of the space, we have identical clocks to
measure the time. All the clocks in each frame are synchronized with one
another. We shall discuss in the next section how this crucial operation can be
done. We choose the origins of the times in both frames, t = 0 and tʹ = 0, as the
time at which the two frames overlap.
We shall call something happening in a definite position an event, as defined
by the three Cartesian coordinates measured by the rulers in the considered
frame, at a definite instant of time, as measured by the clock in that position in
the considered frame. There are two relevant parameters, which always are
present in relativistic formulas. They are pure numbers and are functions of the
velocity υ O ʹ of Sʹ relative to S. The first one is the ratio of this velocity and the
light speed
(6.17)
and the second is

(6.18)
The Lorentz transformations are

(6.19)

We immediately see that they are a generalization of the Galilei

transformations , tending to them for , namely for velocities much
smaller than the light speed, .
The inverse transformations, to go from Sʹ to S, can be found by inverting the
system of equations or, in a simpler way, by changing the sign of the velocity.
Hence

(6.20)

The Lorentz transformations show very strange looking aspects. They mix,
so to say, space and time. We shall see the consequences in the next sections.
Here we shall look at them from a geometrical point of view. Indeed, Eq. (6.20)
are similar to the transformations between the coordinates in two frames
differing for a rotation of the axes. If the rotation is, for example, around the
common z axis, that we can call the height, the transformations are

(6.21)
Also in this case, the quantities in the second frame are mixtures, better
linear combinations, of the quantities in the first. If we look at an object we refer
to one of its dimensions as width, another as thickness. If we now rotate our
point of view by an angle around a vertical axis, the new width, namely the
angle under which we see the object in the horizontal plane, contains a part of
what we called depth before the rotation, and vice versa. It follows that depth
and width are not absolute properties, rather they depend on the point of view,
namely they are relative to the reference frame. The Lorentz transformations are
analogous. They tell us that the length measurements made by a person contain
some of the time measured by another person moving relative to the first one.
When speeds are high, close to the speed of light, the objects are mixtures of
space and time, as usually they are of width and depth. When we turn around an
object and we see it from different angles, our brain automatically recalculates
depth and width, because it developed under these conditions. If we were living
at high speed we might have a brain able to calculate the new mixture of space
and time every time we change speed. We do not have this automatic habit and
must understand the situation by carefully reasoning.
As we well know, the norm of a vector in our three dimensional space is the
sum of the squares of its Cartesian components. In particular the norm of the
position vector is
 (6.22)
If we consider for simplicity a plane, we have , which is the
Pythagorean theorem. Notice that the same is not true, for example, on a
spherical, rather than plane, surface. The Pythagorean theorem is valid if the two
dimensional space is flat. The same is true in three dimensions. A space in which
the squares of the distances are given by Eq. (6.22) is said to be an Euclidean
space .
We also know that a property of the rotation of the axes is to leave the norm
of the vectors invariant. We can see the reason for that writing Eq. (6.21) as a
product of matrices

(6.23)

We see that the square matrix in the transformation is orthogonal . Do the

Lorentz transformations have the same property?
Consider the following two events in S. The first one is the start of a light
pulse from its origin O at the instant t = 0, the second is the arrival of the pulse in
the point (x, y, z) at time t. We express the fact that the speed of light is c by
writing
(6.24)
In Sʹ too the light propagates with the same velocity c and we can write
(6.25)
The quantities in the left-hand side are very similar to the norm of a vector in
four dimensions. They are called interval s. The difference is the minus sign in
front of the temporal term. Technically, the four dimensional space—time is said
to be a pseudo-Euclidean space . Another way to cope with the issue is to define
an imaginary time coordinate, ict. To simplify the expressions we shall use the
symbols
(6.26)
An event is a point in space-time. Analogous to the position vector in three
dimensions is the four dimensional vector of coordinates given by Eq. (6.26).
We shall call these vectors, four-vectors , to distinguish them from the vectors in
three dimensions (three-vectors). The Lorentz transformations written as
products of matrices are
 (6.27)

As first established by Poincaré in 1905, the Lorentz transformations joined

with the space rotations, forming a group which he named the Lorentz group .
The matrix corresponding to the product of two transformations is the product of
their corresponding matrices.
Equations (6.24) and (6.25) tells us two things. Firstly, two events connected
by a light signal are separated by a null interval. This does not mean that they
coincide but that the norm of the interval between them is zero. This possibility
is a consequence of the minus sign in the temporal term. The norm of a four-
vector can be positive, zero, or negative in space-time. Secondly, if an interval is
null in S it is null in Sʹ too. This is a formal way to state that the speed of light is
invariant. As a matter of fact we can state more. The square matrix in Eq. (6.27)
is orthogonal. The consequence is that all the intervals, even more the norms of
all the four-vectors, are invariant under Lorentz transformations.
In three-dimensional space we dealt with vectors in three dimensions, which
we now call three-vectors. As the reader remembers, a three-vector is an ordered
triplet of real numbers that transforms under rotation of the axes as the position
vector.
In a similar manner in relativistic physics we deal with four-vectors. A four-
vector is a quadruple of numbers, real the first three, imaginary the fourth, which
transform from an inertial reference to another, in relative uniform translation
motion, as the coordinates do. The norms of all the four-vectors are consequently
invariant under Lorentz transformations, in other words they are four-scalars. As
such, they play very important roles in relativistic physics. We shall see
examples later.
In the next sections we shall discuss the deep consequences of the Lorentz
transformations on the basic concepts of space and time. Here we notice the
following.
Historically, the Lorentz transformations were found, as mentioned, by the
three main authors, in temporal order, Lorentz , then Poincaré and then Einstein .
Each of them started from somewhat different hypotheses and followed a
different logical path. The path we have followed here is to start from the
experimental discovery of the invariance of the speed of light. This was indeed a
revolutionary discovery. This was also one of the axioms, together with the
relativity principle, assumed by Einstein. From the logical point of view,
however, this second axiom is not necessary. Indeed, the relativity principle
imposes the covariance of the Maxwell equations. Once this is established, with
the Lorentz transformations, the invariance of the speed of light is an immediate
consequence.
However, the historical approach we have followed, as the vast majority of
textbooks do, tends to hide the logical structure of special relativity and to
overemphasize the role of electromagnetism in the foundations of the theory.
After more than one century from its creation we know that all the fundamental
interactions, not only the electromagnetic one, but also the gravitational, the
strong and the weak interactions obey the relativity principle. All the laws that
govern them are covariant under Lorentz transformations. The fields of the
fundamental interactions, which are analogous to the gravitational field we
studied in Chap. 4, in quantum mechanics, are mediated by “quanta”. The
quantum of light is the photon. Its velocity is the velocity of light. As we shall
see in Sect. 6.10 this implies that the mass of the photon is zero. However, it
would be logically possible that the photon would be massive. In this case, the
Lorentz transformations would not change, but the parameter c appearing in the
equations would not be the speed of light and the latter would not be invariant.
Indeed, this is the case of the weak interaction, the quanta of which, called Z 0
and W ± bosons, have mass and do not move at the speed of light. If that was the
case, the demonstration based on the invariance of the speed of light would not
hold. But the final result would still be valid.
From the logical point of view, we must ask ourselves the following
questions. Can we establish the relativity theory independently on
electromagnetism? What are the assumptions needed for that? The answer is yes;
only a few hypotheses on the basic structure of the space-time are needed. These
are the following:

1. Space-time is isotropic and homogeneous.

2. A class of inertial reference frames exist, namely frames in which the inertia
law holds.

3. The relativity principle is valid, namely there is no privileged reference

frame.
4. The transformations form a group.

5. A class of events exists for which the causality principle holds. In this class
the sign of the time differences between events, that is the nature of a possible
causal relation, is the same in all the inertial frames.

It can be demonstrated that only two transformations exist under these

hypotheses, the Galilei and the Lorentz transformations .1 The quantity c in the
latter has the dimension of a velocity and enjoys the properties of being invariant
and being the maximum possible velocity. Galilei transformations are the limit
of the Lorentz ones for . There is no need to rely on electromagnetism.
The electromagnetism enters the game only to give to c the physical meaning of
speed of light.

6.4 Criticism of Simultaneity

The most important difference between Galilei and Lorentz transformation is on
time measurements. In the former the result of the measurement of a time
interval is the same in S and in Sʹ. Time is absolute, independent of the reference
frame, in the Galilei and Newton physics. On the contrary, the last Eq. (6.19)
states, in particular, that the instant at which an event happens in Sʹ depends not
only on the time in which it happens in S (as expected) but also on its position in
S, as not expected. Hence, two events happening in two different points that
appear to be simultaneous to an observer in S do not appear to be so to an
observer in Sʹ. The simultaneity of two events is not an absolute concept, but
rather it is frame dependent. The in-depth criticism of the simultaneity concept
and of the time intervals measurement was made by H. Poincaré in 1898. We
shall explain the argument considering the ideal experiment represented in
Fig. 6.9.
Fig. 6.9 A light flasher and two detectors at equal distances

We suppose to have fixed in the Sʹ frame a rigid bar parallel to the xʹ axis. In
the middle point of the bar we have installed a light source, which emits a light
flash at a certain instant. The flash propagates in all directions, in particular
towards two detectors R 1 and R 2 at the two extremes of the bar. The observer in
Sʹ considers the two events of arrival of the flash at the two detectors as
simultaneous. Notice that this conclusion can be reached only assuming that
light propagates with the same velocity in both directions, namely that space is
isotropic. Notice that the assumption is different from the invariance of the speed
of light.
For the observer in S the two events are not simultaneous. Suppose that the
velocity v of the bar in S has the direction from R 1 to R 2. One flash travels
towards R 1 that is approaching, the other towards R 2 that is receding. The
former will then take a shorter time than the latter to reach its detector. The two
events are not simultaneous.
The fact that the simultaneity of two events happening in two different points
is not absolute is a consequence of the existence of a maximum velocity for the
propagation of the signals. This in turn has deep consequences on the
measurement of time. We have defined an event as the set of the three spatial
coordinates and the temporal one that characterize a phenomenon happening at a
certain time in a certain point. To give a physical meaning to this definition, we
need to define the sets of operations to be done to measure the space and time
coordinates. In particular, to measure the time of the events we need to have
identical clocks in all the points of the reference frame. All the clocks must be
synchronized. This means that the arms of all the clocks must reach the same
position simultaneously. As simultaneity is frame dependent, an observer
moving relative to a frame, the clocks of which have been synchronized by the
observer at rest in that frame, sees those clocks as not synchronized. The
consequence of the frame dependence of simultaneity is the frame dependence
of the time measurements. Let us see that in the details.
6.5 Dilation of Time Intervals
Consider two events happening in the same point x 1 of the frame S in two
different instants t 1 and t 2. In these conditions we can measure the time with a
single clock in x 1. In other words, we have no need to synchronize clocks in
different positions. The two events have the space and time coordinates (x 1, 0, 0,
t 1) and (x 1, 0, 0, t 2). They are separated by the time interval

where the subscript 0 is to recall that the time interval is measured in the frame in
which the object is at rest. Such intervals are said to be of proper time . The
observer in Sʹ obviously does not see the two events in the same point of his
frame, but, say, in x 1ʹ and x 2ʹ. If he wants to measure the times t 1ʹ and t 2ʹ, in
which the events happen he needs two clocks, one in x 1ʹ and one in x 2ʹ, which
must be synchronized. Equation (6.19) tell us that

The time interval in Sʹ is then

or
(6.28)
Consider for example a clock producing periodic ticks. The period, namely
the time interval between two consecutive ticks, in the frame in which the clock
is at rest, is, say ∆t 0. An observer moving with velocity υ Oʹ the clock appears
emitting ticks with the period

(6.29)

which is longer than the proper time ∆t 0.

It is useful to show this result also with a physics argument. Suppose that the
observer in S and in Sʹ have two identical clocks, built as in Fig. 6.10a. The light
source L emits a flash at a certain instant, which reaches the mirror R at the
distance l and is reflected back to the detector R. When the light pulse reaches R
a tick is emitted and the source L emits anther flash, and so on. Let us see now
how the two observers see the period of the clock. For both observers the period
of their clocks is the time to go twice through the distance l, as in Fig. 6.10a,
namely the proper time .

Fig. 6.10 A clock in, a seen in its rest frame, b seen from a moving observer

Also, to both observers the clock of the other one appears to move with
velocity υ Oʹ = υ O . Suppose that both clocks are oriented perpendicularly to the
relative motion. In these conditions, the path of light that the observer Sʹ sees in
the clock in S is as represented in Fig. 6.10b, and reciprocally. Light takes half a
period Δtʹ/2 to go from L to M, and the other half a period to go from M to L.
The distance travelled by the flash in half a period is then (Δtʹ/2)c. In the same
time interval the clock has moved a distance of (Δtʹ/2)υ Oʹ . Hence (see figure)

from which

which is Eq. (6.29).

In the just made argument we implicitly assumed that the length l of the
clock is independent of its motion, namely the same for both observers. As we
shall prove in the next section, this is true because it is perpendicular to the
relative motion. We can reach the same conclusion by observing that it is a
consequence of the relativity principle . Indeed, both observers may agree to cut
two notches in the positions of the extremes of the moving clock, respectively on
the y and yʹ axis, when this passes by. Such notches must result in the same
values of y and yʹ, otherwise the results would be able to distinguish which is
moving and which is still. Notice also that the hypothesis of the independence of
light speed of the direction is once more necessary.
The phenomenon of time dilation is observed every day in elementary
particles laboratories. Protons and electrons are accelerated in accelerators to
speed very close to the speed of light. If time dilation were not taken into
account, these machines would not work.
Consider as another example a natural phenomenon, the cosmic rays. These
are particles, mainly protons and atomic nuclei, accelerated in the galaxy, and
above it, to speed close to that of light and constantly entering the earth’s
atmosphere. In the atmosphere, sooner or later, one of these particles hits a
nucleus of the air producing a shower of secondary particles. Some of them are
unstable. Among them are the µ leptons, or muons, which are very similar to
electrons, if not for the mass that is about 200 times larger. Their lifetime is
2.2 µs. In the decay, a muon produces an electron, a neutrino and an
antineutrino.
The following experiment was done with didactic aims. Charged particles
can be detected using a block of transparent plastic material, doped with
substances that emit a flash of light when a charged particle goes through. The
small flash of light is converted into an electric current pulse, which is sent to an
electronic circuit. When a cosmic ray enters the block a pulse is observed. If it is
a muon and if it stops in the block, after a time of the order of the lifetime it dies
and the newborn electron gives a second pulse. This signature allows us to
discriminate the stopping muons from other events induced by cosmic radiation.
The apparatus was first used on Mount Washington (New Hampshire) at 1800 m
height. In 1 h 568 stopping muons were counted.
How much time is needed for the muons to travel from 1800 m height to sea
level? Obviously that depends on their speed. However, a lower limit is given by
assuming they move with the speed of light. This lower limit is 6.3 µs. The
experimenters then counted how many muons had lived more than 6.3 µs of the
568 detected on the Mount Washington. They found 27 of them. They then
moved their detector to sea level. In absence of time dilation, about 27 events
had to be detected. They found 412. This number agrees with Eq. (6.29) if the
average muon speed is β = 0.99.
Now consider an observer sitting on a muon. In this frame the lifetime is not
dilated and the muon survives only a few microseconds. How can so many reach
sea level? The reason is that, as we shall see in the next section, the distance
from the top of Mount Washington to sea level does not appear to the muon to
be 1800 m, rather it is contracted by the same factor, the Lorentz γ parameter, as
the time dilation. For β = 0.99, we have γ = 6.1 and the distance to travel is only
257 m.

6.6 Contraction of Distances

A second consequence of Lorentz transformations is the contraction of lengths .
We start by observing that the operational definitions of the length of an object
at rest and of an object in motion are not the same. The operations to be done in
the two cases are indeed completely different. Consequently, there is really no a
priori reason for which the two lengths should be equal. It is just every-day
experience with objects moving at relatively low velocities that makes us believe
in this equality. The lengths are equal, as is easily seen, for the Galilei
transformations , not, as we shall now see, for the Lorentz transformations .
According to the latter, when a body moves with velocity v relative to the
observer, its dimension parallel to v appears contracted by a factor 1/γ relative to
its value measured at rest. The transverse directions are equal for both observers.
To demonstrate these statements we imagine a rule fixed to the frame S lying
on its x-axis. The observer in the moving frame Sʹ determines the length of the
ruler by measuring the coordinates of its extremes x 1ʹ and x 2ʹ at the same instant
tʹ. The two corresponding events have the space-time coordinates (x 1ʹ, 0, 0, tʹ)
and (x 2ʹ, 0, 0, tʹ). The length found by the observer is lʹ = x 2ʹ– x 1ʹ. The
coordinates in S of the two events are

and their distance in S is

In conclusion, the relation between the length parallel to the relative velocity
of an object at rest and moving with velocity υ Oʹ is

(6.30)
where the subscript 0 recalls that this is the length at rest. This is called the
proper length . In any other moving frame the length appears contracted by the
factor 1/γ.
As for the dimension of the ruler, or any object, along y and z, perpendicular
to the motion, the fact that they do not vary follows immediately from the
second and third Eq. (6.19).
In this case too, let us demonstrate the result also with a physics argument.
This will show that the contraction of the length is a logical consequence of the
time dilation.
We still consider the ruler fixed along the x-axis of S. The observer in S
measures the length l, and establishes that the observer in Sʹ, which is travelling
at speed υ Oʹ , crosses the distance l in the time interval Δt = l/υ Oʹ . This time is
not a proper time, because it is between two events happening in different
locations, the passage of the mobile observer at one extreme and at the other. As
such it is measured with two different clocks. On the other hand, for the observer
in Sʹ the two events happen in the same point and he can measure the time
interval, ∆tʹ, with the same clock. ∆tʹ is a proper time interval and, for what we
saw in the last section, , and, as , it is . The
mobile observer sees the rule moving at the speed υ Oʹ and consequently
evaluates its length to be , which is the result that had to be
demonstrated.

6.7 Addition of Velocities

In this section we shall find the rule of addition of velocities in relativistic
physics. We just recall that for the Galilei transformations , if, for example, a
ship moves relative to shore with velocity u and on the ship a passenger moves
with velocity vʹ, relative to the ship, the velocity of the passenger relative to
shore is v = u + vʹ. This is the Galilean composition rules of velocities. We shall
now find the corresponding rule for Lorentz transformation , still in the
particular case in which the two frames S and Sʹ are those in Fig. 6.11.

Fig. 6.11 Two frames in relative motion

The velocity vʹ of a point in Sʹ is and the corresponding in S

. Notice that each derivative in each frame is with respect to the

time in that frame. We shall now use the Lorentz transformations Eq. (6.19) with
and . We have

By dividing the first three equations by the fourth we have

We then write the conclusion

(6.31)
Notice that not only the components parallel to the relative motion, but also
the normal ones, are different in the two frames. The complicated behavior of
the velocity stems from the fact that its components are not the three components
of a four-vector. This is because, while (dx, dy, dz) are such components, dt is
not a four-scalar.
It is easy to verify that the Eq. (6.31) tend to the Galilean one for .

Example E 6.1
Consider a particle moving with velocity υʹ x = c/2 relative to Sʹ, in the positive
direction of xʹ. The reference Sʹ moves relative to S at the speed u = c/2 in the
same direction. Notice that if the transformation were the Galilean ones the
velocity of the particle relative to S would have been equal to c. With the
Lorentz transformation we have

Example E 6.2
Consider Sʹ to be a (very fast) ship and shooting a ball vertically upwards with
velocity υ zʹ. Which velocity of the ball is seen from shore? With υ xʹ = υ yʹ = 0
Eq. (6.31) give

Consider now the important case of a light signal propagating along the xʹ axis of
Sʹ. Its velocity relative to S is

(6.32)
Namely, it has the same value in Sʹ and in S, whatever their relative velocity
can be. This result was expected considering that the speed of light is invariant
under the Lorentz transformations.
A corollary is that combing to velocities smaller than c the resulting velocity
is always smaller than c. The speed of light is the maximum possible velocity.

6.8 Space-Time
We have seen in Sect. 6.3 that the Lorentz are, from the geometric point of view,
rigid rotations in the space-time , of coordinates (x, y, z, ict).
We cannot represent the four dimensions of the space-time on the two
dimensions of a page of a book. However, we can learn a lot considering a
particle moving in just one dimension, x. The space-time diagram has then two
axes, the space coordinate x and the time, or, better to have the same physical
dimensions ct, as shown in Fig. 6.12.

Fig. 6.12 The space-time diagram

A point on this diagram represents an event happening in the space point x at

time t. A particle at rest in the frame is, in space-time, a sequence of events at
different times that have all the same coordinates. This is a line parallel to the ct
axis, as line 1 in the figure. Such a line, in general, is called the lifeline of the
particle.
If the particle moves with a constant speed υ its lifeline is a straight line, like
2 in the figure, having a slope relative to the ct axis equal to υ/c. Notice that the
scales of the axes are such that the lifelines of any particle moving at the usual
velocities are very near to being vertical (υ/c < <1). On the other hand, the
lifelines of the light signals are straight lines at +45° or −45° (dotted in Fig. 6.12)
depending on the direction of propagation being the same or opposite to the x-
axis. Line 3 is the life line of a particle that is at rest at a positive value of x at
time 0, and that later on moves in the positive x direction of an accelerated
motion, soon reaching speeds close to c. Notice that no lifeline can have a slope
relative to the ct axis larger than one, namely a velocity larger than c.
Consider now the event O in the origin of the space time reference frame,
namely the instant t = 0 in the point x = 0. Suppose this event being the start of
light signals in all directions (the two of the x axis in our case). The lifelines of
the signals are the bisectors of the axes as shown in Fig. 6.13. In the four
dimensional space time, these lines draw a hypercone with vertex in the origin
and half vertex angle equal 45°. It is called the light cone . The part of the light
cone on ct < 0 corresponds to a light signal reaching the point x = 0 at t = 0.

Fig. 6.13 The light cone

It is not difficult to see that a Lorentz transformation transforms the axes as

shown in the figures for xʹ, ctʹ. The rotation of the axes is different from rotations
in space because here the metric is pseudo-Euclidean rather than Euclidean. In
the space time the x and ct axes rotate in opposite directions, by the same angle,
approaching the light cone. The rotation angle is larger for higher relative
velocity and tends to 45° when that tends to c. Obviously, the light cones of the
two frames coincide, because the light velocity is the same in both.
Consider now events inside the light cone. The intervals between the origin
O and each of them, like A and B, are negative. Such intervals are said to be
time-like , because the purely time intervals are negative. The events outside the
light cone are separated from the origin by positive intervals, called space-like .
The events on the light cone are separated by null intervals and are called light-
like . The intervals being invariant, these properties are independent of the
reference frame.
Two events separated by a time-like interval can be joined by a signal
travelling at a speed smaller than light, if the interval is light-like interval, they
can be joined by a light signal, but if is space-like, they cannot be joined by any
signal. Such a signal should travel faster than light. Consequently, no cause and
effect relation can exist between two events at a space-like interval. This
conclusion is connected with the fact that the relation past-future is not an
absolute one for events outside the light cone, such as the event C in the figure.
This event is future relative to O in (it has t > 0), while it is past relative to O, the
same event, in Sʹ (tʹ < 0) as is clear from the figure.
The events separated from the origin by time-like intervals are, as we have
seen, inside the light cone. We can distinguish two parts of the cone. In the upper
half cone, with t > 0, we have the events future relative to O. In the lower half
cone, with t < 0, we have the events past relative to O (as B). Consider now, for
example, the event A. It is separated from O by a negative interval. As the
intervals are invariant there is no reference frame in which A is contemporary to
O, because in this case A would be separated from O by a positive, or null
interval. We can conclude that there is also no frame in which A is past relative
to O, because in this case, for continuity reasons, a frame would exist in which A
and O are simultaneous. In conclusion, all the events in the upper half light cone
are future to O in any reference frame (absolute future of O), those in the lower
half cone are past of O in any frame (absolute past of O).

6.9 Momentum, Energy and Mass

As we have seen, the Lorentz transformations between inertial frames are such
as to guarantee the validity of the relativity principle for the Maxwell equations,
the equations that govern electromagnetism. The principle requires however that
all the physical laws should be covariant under these transformations.
Consequently, we must also find the Lorentz covariant expression that
generalizes the second law of Newton. Once we have found the new law, its
predictions should be checked against experiments.
 We start with the observation that the new law should admit the Newton one
as a limit for small velocities. We also notice that, if an equation has to be
covariant, all its terms must transform in the same way under Lorentz
transformations . All of them must be four-scalar or four-vectors.
We already know a four-vector, the one that identifies the event in space-
time, having components (x 1, x 2, x 3, x 4). We have obtained “promoting” the
space three-vector r = (x 1, x 2, x 3) with the addition of the fourth time
component. Such a “promotion” is not always possible with every three-vector.
As we have already seen, for example, the three components of the velocity
three-vector v = (dx/dt, dy/dt, dz/dt) are not the three space components of a four
vector, because such are (dx, dy, dz), but dt is not a four-scalar.
The first step towards using relativistic dynamics is finding the correct
expression of linear momentum . As we well know, the linear momentum of a
particle of mass m and velocity small relative to c is

We can solve the problem of the non-invariance of dt by taking the

derivative relative to the proper time t 0, the time in the reference frame moving
with the particle, rather than relative to t. Recalling that dt 0 = dt/γ we have

(6.33)
We can immediately check that this expression tends to the Newtonian one
for small velocities, namely for . As a matter of fact, γ does not differ
much from 1 even at quite large velocities. For example, even at υ = 0.25c,
γ = 1.03, it has increased by only 3 %. However, when the velocity approaches c,
the increase of γ becomes very rapid, for example, for υ = 0.5c, γ = 1.15, for
υ = 0.75c, γ = 1.51, for υ = 0.99c, γ = 7.09, to diverge for υ → c. If we try to
accelerate a particle, when its velocity approaches the speed of light the work
necessary to increase the velocity further becomes larger and larger. The work
uses a larger and larger fraction of force to increase the γ factor and less and less
to increase the velocity. The work to reach c would be infinite.
We have now found the space vector Eq. (6.33) that can be promoted to four-
vector, which is called four-momentum . What is its fourth component? Taking
into account that dt/dt 0 = γ it is clearly
 (6.34)
This very important quantity is, as a part of a constant, the energy of a free
particle, as will become clear soon after having found the law of motion.
Before doing that we express the norm of the four-momentum .
As all the norms of the four-vectors, this is a Lorentz invariant quantity, a four-
scalar. Its expression is particularly simple in the rest frame of the particle, in
which p = 0, and we have
(6.35)
The norm of the four-momentum is proportional to the mass squared of the
particle.
We now state without demonstration that, once the expression of the
momentum is changed according to Eq. (6.33), the expression of the Newton law
does not need any further change. However, there are now two time dependent
factors in the derivative, the velocity and γ. We have

(6.36)
Notice that neither the force nor the time derivative of the momentum are the
space components of a four-vector. However, such are F dt and d p, and
consequently Eq. (6.36) is Lorentz covariant. Historically, the equation was
found for the first time in June 1905 by H. Poincaré , who demonstrated its
covariance and, in addition, that it is the unique expression enjoying such a
property.
We are now ready to see the physical meaning of the fourth components of
the four-momentum and of F dt, namely of . We shall proceed in a way
quite similar to what we did for the kinetic energy theorem. Let F(r) be the
resultant force acting on the particle at the position vector r. We calculate its
work when the particle moves from A to B on a certain trajectory, as shown in
Fig. 6.14.

Fig. 6.14 The trajectory of a particle and the force acting on it

The elementary displacement d s in the time interval dt is d s = v dt. The

work done by F is . To evaluate the last dot
product we differentiate Eq. (6.35), obtaining . Substituting
p = m γ v, and simplifying, we have
(6.37)
and the elementary work is then
(6.38)
The work done by the force when the particle moves from A to B is

(6.39)
Exactly as in Newtonian physics, the work done by the resultant of the forces
on the particle is the difference between the values of a function of the velocity
only at the end and at the beginning of the considered trajectory. In the following
we shall consider only free particles, namely in absence of potential energy. In
these conditions, we can say that the energy of the particle is
(6.40)
We see that the fourth component of the four-momentum is just the energy
of the particle, divided by c. For this reason, the four-momentum is also called
an energy-momentum vector . Its components are . Its norm, or better
the opposite of its norm is
(6.41)
The relativistic energy of a free particle, Eq. (6.40), is not only kinetic
energy. Indeed, the particle has energy also when it is at rest. It is called rest
energy and we shall indicate it with
(6.42)
We can say that the relativistic kinetic energy of a free particle is its total
energy less its rest energy, namely
(6.43)
2,
We see immediately, by developing in series of β that the relativistic
kinetic energy tends to the non-relativistic one at low velocities:

On the other hand, at very high velocities, Eq. (6.40) shows that the energy
of the particle grows without limits when its velocity approaches the speed of
light. As we have seen for the momentum, this is due to divergence of the γ
factor. The particle “accelerators” of the laboratories studying the elementary
particles work usually with protons or electrons “accelerated” at a speed very
close to c. Accelerators act to increase the energy of the particles, while their
velocity may change only by very small amounts. They should be more properly
called “energizers”. Indeed, particles of non-zero mass can never reach the speed
of light. Their energy and momentum would be infinite. We shall come back to
massless particles soon.
The fundamental mechanical quantities of a free particle are its mass , its
momentum and its energy. These quantities are linked by two fundamental
equations, Eq. (6.41) that we shall now write in a bit different form (multiplying
by c 2) and a somewhat different expression of Eq. (6.33). They are

(6.45)

(6.46)
We now observe that in nature elementary massless particles exist. Such are
the photons, the quanta of light, and also the quanta of the strong interaction
binding the quarks in a proton and in a nucleon, which are called gluons. When
m = 0, the expression Eq. (6.33) has no meaning, because it contains the ratio
between a null and an infinite quantity. The most general expression of the
relativistic momentum is Eq. (6.46) that is valid both for massive and for
massless particles.
Let us have a better look at Eq. (6.45) with the help of the “cartoon” of
Fig. 6.15. In the general case, Fig. 6.15a, the energy is like the hypotenuse of a
right triangle having mc 2 and pc as sides. It is given by the quadratic sum of the
two quantities, namely it is the square root of the sum of their squares. One of
them, mc 2, is the mass energy , the other one, pc, is the energy of its motion .

Fig. 6.15 Relation between energy, momentum and mass. a Generic, b particle at rest, c massless particle
 If the particle is at rest, its energy is only mass energy, or rest energy
(6.47)
Here we must warn the reader that this equation is often written in the press,
but also in the scientific literature, as E = mc 2, which is not true, because, as we
saw in general it is E = mγc 2, Eq. (6.40). The confusion is increased by writing
mγ “relativistic mass” and talking of mass varying with velocity. These are
archaic concepts that were introduced when relativity theory was being
developed, but should be avoided. Indeed the mass is an invariant quantity and
does not vary with velocity. The term mγ is apart from a factor c 2 not else
than the energy, which is the fourth component of a four-vector.
Equation (6.46) tells us that the mass energy is enormous, due to the c 2
factor. However matter and energy are not equivalent. Indeed, matter has existed
since the origin of the universe and does not convert into energy. The reason is
that the matter particles have charges, the electric, the weak and the strong ones.
These charges are conserved. We cannot destroy, for example, an electron and
get energy from its mass. We can however, annihilate an electron with its
antiparticle, the positron that has opposite charge. However, the quantity of
antimatter in the universe is very small. We shall come back to the mass and to
energy transformations in the next section.
Figure 6.15c shows the case of a massless particle, say a photon. For
Eq. (6.45), being massless means that
(6.48)
and from Eq. (6.46), for photons
(6.49)
a free massless particle can move at only one speed, the speed of light.

6.10 Mass, Momentum and Energy for a System of

Particles
We now consider a system of free particles, namely there are no forces, external
or internal, acting on them. As in non-relativistic physics, the total momentum
and the total energy of the system are the sum of the homologues quantities of
the single particles, namely

(6.50)
The situation is more complex if the particles interact with internal forces. In
particular, Eq. (6.50) are not valid. We do not have the time to discuss the issue
here, but only mention that, in addition to the mechanical ones of the particles,
there are both energy and momentum distributed in the fields of forces.
Coming back to the system of relativistic non-interacting particles, we shall
now look at its total mass . As for the single particle, the total momentum and
the total energy of a system are (taking into account the c factors) the four
components of a four-vector, of which Mc 2 is the norm.

(6.51)

We see here a fundamental difference from the non-relativistic case: the

mass of the system is not the sum of the masses of its constituents.
Consider now several examples.

Example E 6.3
Find the expressions for the mass of the system of two photons of the same
energy E, if they move in equal or opposite directions.
For the photon that has zero mass, pc = E. Consequently the total energy E tot
= 2E.
If the photons have the same direction, then the total momentum is p tot
= 2E/c and therefore the mass is m = 0.
If the velocities of the photons are opposite, it is still E tot = 2E, but p tot = 0,
and hence m = 2E/c 2 .
In general, if θ is the angle between the velocities,

and hence

Example E 6.4
Consider two particles with the same mass m moving with the same initial
velocity υ of opposite direction. The two particles collide and stick together. The
final kinetic energy is zero. Macroscopically we call the collision completely
inelastic. However, the total energy did not vary, because the rest energy has
increased by the same amount. In relativistic mechanics the inelastic collisions
do not exist. Energy is always conserved
 In other words, the mass of the final body is not M = 2 m, but
, which is larger than 2 m. The mass increase is extremely
small at low velocities. As an example, suppose that υ = 300 m/s, which is quite
large for everyday life, but very small compared to c, being that β = υ/c = 10−6.
Developing the above expression in series we have

which differs from m by, in order of magnitude, 10−12. This is so small that it
cannot be measured. In other words, the rest energy is so large that its increase
corresponding to the decrease in kinetic energy is undetectable. The decrease of
kinetic energy between initial and final state is on the contrary evident. It looks
like energy is not conserved. But, what appears to have been lost is rather hidden
in the mass energy.

Example E 6.5
The most massive nuclei, as some of the Uranium isotopes, are often unstable.
They can break up in fragments spontaneously, or make them absorb a neutron.
Suppose the fragments to be two and m 1 and m 2 their masses, while M is the
mass of the mother nucleus. We state that m 1 + m 2 < M. Indeed, the energy
conservation requires that

The final kinetic energy E K1 + E K2 is the energy produced for example in a

power station. The remaining energy difference may correspond
to a small mass difference, but the corresponding energy can be large due to the
factor c 2. Let us see a numerical example.
We profit from an example to introduce a measurement unit of mass that is
widely used in atomic and subatomic physics. As we have seen, energy can be
measured in electronvolt , Eq. (3.78). As the mass is equal to the rest energy
divided by c 2, we shall measure it in eV/c 2.
The simplest nucleus, the hydrogen one, is simply a proton, the mass of
which is . The mass of the neutron is a bit larger,
. The mass of the electron is about 2000 times smaller,
 . The most massive nuclei have masses of hundreds of GeV/c 2.
In a heavy nuclear fission, namely a break up, the released energy is of several
MeV. In other words, the mass difference between the initial and the final state
is, in relative value, of a few parts in hundred thousandths. These values are
small, but can be measured, and the predictions of the theory can be checked.
In the lightest nuclei the opposite process can happen. That process is fusion.
For example, two neutrons and two protons can join together to produce a He
nucleus. This is because the mass of the latter, , is smaller
than the sum of the initial masses. Let us calculate the mass defect, namely

The mass defect corresponds to the binding energy, namely to separate the
four components of a He nucleus we must give it an energy of 28.3 MeV.

Example E 6.6
Consider now the hydrogen atom, which is made of a proton and an electron. Its
binding energy, namely the energy to separate the electron from the proton is
∆E = 13.6 eV. The mass difference in relative values is

which is a very small fraction. The atomic energy scale is much smaller than the
nuclear one.

Example E 6.7
When energy is measured in eV, the momenta are measured in eV/c. Let us see,
for example, the value in SI of a 1 meV/c momentum. It is

6.11 Force and acceleration

As we have seen, the relativistic law of motion of a particle of mass m under the
action of the force F states that the force is equal to the rate of change of
momentum . This is p = mγ(υ)v. It contains the product of two functions of time.
Consequently, the derivative is the sum of two terms

(6.52)
 Taking the derivative of γ(υ), we obtain

We substitute this expression in Eq. (6.52) taking into account that dυ/dt is
the component of the acceleration in the direction of the velocity, namely that
, where u υ is the unit vector of velocity, obtaining

(6.52)
where β is the vector v/c.
We see that the force is the sum of two terms, one parallel to the acceleration
and one parallel to the velocity. Therefore, we cannot define any ‘mass’ as the
ratio between force and acceleration. At high speeds, the mass is not the inertia
to motion.
To solve for the acceleration we take the scalar product of the two sides of
Eq. (6.52) with . We obtain

Hence

(6.53)
and, by substitution into (6.52)
(6.54)
The acceleration is the sum of two terms, one parallel to the force, and one
parallel to the speed.
Equation (6.52) and its equivalent Eq. (6.54) have been the object of a large
number of experimental controls with high energy charged particles like protons,
nuclei and electrons under electric and magnetic forces in different
configurations. The engineers designing the accelerators at relativistic energies
use these formulas in their everyday work.
We notice that force and acceleration have the same direction in two cases
only: 1. force and velocity are parallel: F = mγ 3 a; 2. force and velocity are
perpendicular: F = mγ a. The proportionality constants are different. Consider
for example a particle moving with 95 % of light speed, that is β = 0.95 and
γ = 3.2. If the particle travels on a circle, the centripetal force should be 3.2 times
larger than what was foreseen by Newtonian mechanics. However, if it is in a
rectilinear accelerated motion the force necessary to give it the same acceleration
is γ 3 = 32.8 times larger than in Newtonian mechanics. We see that, even in
these special cases, we cannot consider mass as the inertia to motion.
6.12 Lorentz Covariance of the Physics Laws
We have seen how the relativity principle , originally established by G. Galilei in
the XVII century, was found to hold for electromagnetic interactions, provided
that the transformations of coordinates and time between two inertial reference
frames are Lorentz transformations . This led to special relativity. The theory,
however, can work only if all the physics laws turn out to be Lorentz covariant .
Indeed, we have already discussed that for the second Newton law.
We have already firmly stated that the Lorentz transformations, while they
historically discovered a guarantee for the relativity principle of a specific
interaction, can be demonstrated independently of electromagnetism, on the
basis of very general assumptions as we saw at the end of Sect. 6.4.
It remains to be seen, however, whether the other forces, or better
interactions, of nature satisfy the relativity principle , namely if the equations
that rule them behave in a Lorentz covariant form. The answer is yes, but we can
give here only a few hints.
The Newton law of the gravitational force,

(6.55)
is clearly not Lorentz invariant. Indeed, this expression implies instantaneous
propagation of the effects over any distance. If, for example, our sun would
suddenly disappear, the gravitational force on earth would go to zero
immediately. But Lorentz invariance requires that all the fundamental
interactions propagate with a speed not larger than c, which is the parameter in a
Lorentz transformation. Consequently we would be safe still for 8 min, the time
taken by the gravitational wave resulting from the explosion to reach us. The
relativistic theory of gravity is called general relativity , as we have already
mentioned. The equations were sent for publication at the end of 1915
independently by David Hilbert (1862–1943) and A. Einstein . We have now an
enormous quantity of experimental proofs of its validity. We only mention, as an
example, that the data of the global position system, the GPS, which is based on
a constellation of artificial satellites, would give wrong information on our
position if not elaborated with general relativity.
All the other forces we studied in Chap. 3, the elastic force, the forces of the
constraints, the force between molecules, etc. are, at a fundamental level, due to
electromagnetic interaction. As such, the laws by which they are governed are
Lorentz invariant.
The other two fundamental interactions, the weak interaction and strong
interaction, were discovered after the establishment of special relativity and their
equations, which are quantum theories, were written in a Lorentz covariant form
since the start. Their validity has been proven with a myriad of very high
precision experiments on high energy particles both from natural sources, like
the radioactive decays and cosmic rays, and, mainly, in the accelerator
laboratories.

6.13 What Is Equal and What Is Different

We summarize here the concepts that are different in relativistic mechanics
(r.m.) from Newtonian mechanics (n.m) and those that remain unaltered.
1. The relativity principle is valid both in n.m. and in r.m.

2. The coordinate transformations are different, Galilei in n.m., Lorentz in r. m.

3. Time and simultaneity are absolute in n.m., relative in r.m.

4. The law for summing velocity is different.

5. In n.m., velocities can have any value; in r.m. they cannot be larger than c.

6. The expressions of momentum are different.

7. The forces have the same expressions.

8. The force is equal to the time derivative of the momentum in both.

9. The total momentum (and the total angular momentum) of an isolated system
are conserved in both cases.

10. The energy has different expressions. The kinetic energy is directly
proportional to the square of velocity in n.m., not in r.m. The rest energy
does not exist in n.m.

11. The energy of an isolated system is conserved only if all the forces are
conservative in n.m., always in r.m.

12. The total momentum of a system of non-interacting particles is the sum of

the momenta of the single particles. The same is true for energies. This both
in n.m. and in r.m. In r.m. the same is not true for the systems of interacting
particles. We can only hint at the reason for that here. It lays in the fact that
the field of the interaction force contains both energy and momentum.

13. The mass of a composite body is the sum of the masses of its components in
n.m. it is not in r.m.

14. In n.m., force and acceleration are parallel; they are not so, in general, in
r.m.

15. In n.m. the proportionality constant between force and acceleration is the
mass, which acts as inertia to the motion. In r.m. acceleration is not
proportional to the force, there is no “inertial” mass.
16. The mass is invariant both under the Galilei and the Lorentz
transformations.

6.14 Problems
6.1. Consider two reference frames, S, which we call fixed, and Sʹ, which we
call mobile as in Fig. 6.2. In the two frames there are clocks as those in
Fig. 6.5. Develop the argument analogous to that of Sect. 6.5 if the arms of
the clocks are in the direction of the x axis, namely of the relative velocity.

6.2. A muon is produced by cosmic rays in the atmosphere. It travels at

υ = 0.99c for 4 km and then decays. (a) How long does it live in our
reference? (b) and in its frame? (c) How much is the thickness of the
atmosphere it crossed in its reference?

6.3. A particle of mass m moves in a straight motion along the x axis with
. Find its limit velocity for t → ∞. Find the expression of
the force acting on the point.

6.4. A particle of mass m moving with the speed υ = (4/5)c, hits a particle at rest
with the same mass. After the collision the two particles form a unique
body of mass M. Find M and the velocity of this body.

6.5. The cosmic rays contain protons with 1010 GeV energy. Find the time in the
reference frame of such a proton to cross the Galaxy.

6.6. Find its momentum (in MeV/c) of an electron of 1 meV kinetic energy.
6.7. Find the momentum, in MeV/c of an electron travelling at c/2.

6.8. Find the energy of an electron travelling at 80 % of the speed of light.

6.9. A particle called ρ having mass 770 meV/c 2 decays at rest in two particles
called π, which have mass m = 140 meV/c 2. Find their velocity.

6.10. In the LEP accelerator at CERN, electrons were accelerated up to an

energy of 50 GeV. Find the relative difference between the velocity of the
electrons and light.

6.11. A particle called tau has a lifetime of 0.3 ps. Find the velocity it should
have to travel 1 mm in a lifetime.

6.12. A Z° (mass 91.2 GeV/c 2) particle decays at rest in an electron and a

positron (they have equal masses). Find the energy and the momentum of
the electrons. How much does, in relative terms, the velocity of the
electron differ from c?

Footnotes
1 For an elementary proof of this result, see J-M. Lévy-Leblond “One more derivation of the Lorentz
transformation” American Journal of Physics 44 (1976) 271 and A. Pelissetto and M. Testa “Getting
Lorentz transformations without requiring an invariant speed” American Journal of Physics 83 (2015)
338.
© Springer International Publishing Switzerland 2016
Alessandro Bettini, A Course in Classical Physics 1—Mechanics, Undergraduate Lecture Notes in Physics,
DOI 10.1007/978-3-319-29257-1_7

7. Extended Systems
Alessandro Bettini1
(1) Dipartimento di Fisica e Astronomia, Università di Padova, Padova, Italy

Alessandro Bettini
Email: alessandro.bettini@pd.infn.it

In this chapter we shall discuss the mechanics of extended systems, namely of

mechanical systems composed of more than one particle or by bodies of finite
extension. As a matter of fact, even in the simplest case of a point-like body
under the action of a force, at least another body, giving origin to the force, must
exist. Every action is always accompanied by a reaction. In other words, the
simplest mechanical system consists of two interacting particles. We have
considered, for example, the motion of earth or of a planet around the sun. We
had ignored the sun. We could do that without much error because its mass is
enormously bigger. These however are particular cases.
In the first three sections we shall study two-body systems. We shall see
(Sect. 7.1) how the potential energy, corresponding to the force that one body
exerts on the other is, in fact, relative to the pair. In other words, it is an
interaction. We shall then introduce in Sect. 7.2 the concepts of center of mass
and of reduced mass. In Sects. 7.3 and 7.4 we shall discuss two examples of a
two-body system, the double stars and the tides, a phenomenon in another two-
body system, earth and moon.
The experimental study of collisions between two bodies had, and still has,
an enormous importance in the development of physics. In the sections from
Sects. 7.5 to 7.7 we shall see the collision experiments between two pendulums
that led Newton to establish the principle of conservation of linear momentum.
This is one of the fundamental principles in physics, strictly connected with the
action-reaction law.
We shall then move to systems of many particles, introducing the concepts
of total linear momentum (or quantity of motion) and total angular momentum of
a system. We shall find the fundamental laws giving their rate of change, and
study the properties of a privileged point, the center of mass of the system.
In the last two sections we shall come back to the study of collisions between
extended bodies.

7.1 Interaction Energy

In our discussions on potential energy in the preceding chapters, we have
analyzed the problems as if only one body existed, on which given forces were
acting. For example, we said that the potential energy of the weight of a body of
mass m at the height h is mgh. This is a perfectly correct statement when the
mass of the body under consideration, an apple for example, is much smaller
than the body with which it interacts, the earth in the example. In this situation, a
reference frame united with the larger object can be considered at rest. As a
matter of fact, when the apple falls towards the earth, also the earth falls towards
the apple in an accelerated motion. In practice, both earth velocity and
acceleration are completely negligible. Rigorously speaking however, we are
dealing with a two-body system , the apple and the earth and mgh is the variation
of potential energy of the earth-apple system, when the distance between the
center of the apple and the center of the earth increase by h. In other words, the
potential energy is a property of the couple of objects together; it cannot be
associated to one or the other individually. Indeed, if the two interacting bodies
have comparable masses, both of them accelerate considerably under the action
of interaction forces. The kinetic energy of each of them will vary at the
variation of the interaction potential energy . Let us now study the issue.
We start with a simple example in Fig. 7.1. It is made of two small spheres
of mass m 1 and m 2 joined by a spring. In the upper part of the figure the system
is in its configuration at rest. We now move both spheres and call x 1 and x 2 the
two displacements, in value and sign, measured each from its equilibrium
position, as in the lower part of the figure. Both forces now act, F 21 on sphere 1
and F 12 on sphere 2. The forces, an action reaction pair, are equal and opposite.
These are elastic forces, which are proportional to the stretch that is ∆x = x 1 – x
2 (N.B. x 1 is positive, x 2 is negative).
Fig. 7.1 Two masses linked by a spring

The elastic potential energy, Eq. (3.2), is

(7.1)
where k is the spring constant. Notice that this energy does not belong to one
or the other sphere, but to the whole system, in other words is the interaction
(through the spring) energy between the spheres.
The potential energy of any system in a given state is always the work that
must be done against the forces that the system develops to change its state from
the (arbitrarily) defined zero energy state to the given state. In our case the zero
energy state is when the spring is not deformed. In the above statement, all the
work must go into a change of the potential energy, namely it must be done at
constant kinetic energy (zero in particular). Let us check with a direct calculation
that our statements are correct.
Suppose we start from the equilibrium position. We first move sphere 1,
keeping 2 at rest. Call x the displacement (with sign) of sphere 1 from its
equilibrium position. We are moving it from x = 0 to x = x 1. During the
displacement the stretch of the spring is just x. The x component of the force is
consequently F 21x = –kx and the work to be done is against it,

We now move sphere 2, keeping 1 steady. We now call x the displacement of

sphere 2 from its equilibrium position. We are moving it from x = 0 to x = x 2.
The stretch of the spring is now x 1 – x and the x component of the force F 12x
= k(x 1 – x). The work to be done against it is
 Finally the total work is

which is clearly Eq. (7.1).

Consider now a second example: the potential energy of the gravitational
force. Consider a point-like body of mass m on the surface of the earth (mass M),
at the distance R E from its center. As we know, the potential energy is

(7.2)
Recalling the arguments of Sect. 2.14 one easily sees that this is the work to
be done against the gravitational force to move the mass m, say an apple, at zero
kinetic energy, from infinite distance (the state we have defined to have zero
potential energy) to the surface of earth. The energy is negative because, from
outside of the system, we must work against an attractive force. In other words,
the work we are considering is the opposite of the work of the gravitational
force. We also see that the energy is not in the apple alone but in the earth and
apple system.
As the last example we consider the weight force. The potential energy of a
body of mass m at height h over the level we have decided for the potential
energy to be zero, say the ground, is
(7.3)
We know that this energy is just Eq. (7.2), apart from an additive constant.
Indeed, in the two cases we made a different choice of the zero potential energy
state. At first sight the two equations look quite different. However, consider that
Eq. (7.3) is an approximate expression, valid for small level differences relative
to the earth’s radius, h « R E . We then start from Eq. (7.2) expanding it in series
of h/R E stopping at the first order. We get

Now consider that is simply the gravity acceleration on the earth’s

surface g. Equation (7.3) is valid when taking the potential energy on the earth’s
surface equal to zero, And the last equation becomes Eq. (7.3). In
conclusion, the energy of the weight force mgh is not of the body but of the
system body and earth.
7.2 Centre of Mass and Reduced Mass
We now come back to the simple mechanical system of two spheres joined by a
spring (Fig. 7.2) and consider its motions. We are interested in the motion of one
of them, say sphere 1. The sphere is subject to the elastic force F 21. In Sect. 3.2
we have already discussed the motion of a material point under the action of an
elastic force and found it to be harmonic. In that case, however, the other end of
the spring was fixed to a wall and did not move. We can think of the wall as
analogous in that case to sphere 2 in this case. In both cases, the force F 12 acts
on the second body. But the mass of the wall is so large that its acceleration is
completely negligible. In the present case, on the contrary, sphere 2 will
accelerate.

Fig. 7.2 Two spheres connected by a spring

The problem we have now is that both points move. As we shall see in this
chapter however, for every material system a privileged point, called center of
mass of the system, exists. It is a geometrical point, not a physical one. In the
presence of only internal forces, as in the case under discussion, the acceleration
of its center of mass, in an inertial reference frame, is zero. We shall profit from
that and describe the motion in a reference frame moving with the center of mass
and with its origin in it, called the center of mass frame , for a brief CM frame.
The center of mass of a two point-like bodies system is the point on the segment
joining the two points that divide it in parts inversely proportional to the masses
at the corresponding extremes.
We shall call C the center of mass, ξ 1 and ξ 2, the distances of the two
masses from it and r the coordinate of point 1 measured from point 2. By
definition of center of mass
(7.4)
Considering that r = ξ 1 + ξ 2 is the coordinate of point 1, the motion of
which we want to study is
 (7.5)
The force F 21 acting on point 1 will give it the acceleration a 1 according to
the Newton law

We have so found, in the last side of this equation, an important quantity,

called the reduced mass of the system

(7.6)
We can the write the equation of motion of point 1 as

(7.7)
which is a very simple expression indeed. The equation of motion of point 1
is identical to its equation of motion valid when point 2 is fixed, provided that
we are in the CM frame and we substitute for the mass of point 1 the reduced
mass of the system.
Let us check if the arguments we made in Sect. 3.2 agree. First, we observe
that when m 2 becomes very large compared to m 1, the reduced mass tends to
the smaller of the two masses, m 1. To see that, just write Eq. (7.6) as
, from which immediately for Clearly,
what we said in Sect. 3.2 is the limit case of what we are discussing here.
We now come back to the problem of the motion of point 1. We call r 0 the
length at rest of the spring and s its stretch. Hence r = r 0 + s and F 21 = –ks. But
and Eq. (7.7) becomes

(7.8)
which we recognize as the harmonic oscillator equation. We already know its
solution
(7.9)
where A and ϕ depend on the initial condition and
(7.10)
In the CM frame the motion of point 1 is a harmonic oscillation. The
difference with the case when point 2 is at rest is that in place of the mass of the
oscillating body we have the reduced mass of the system. Clearly, point 2 moves
with a harmonic motion of the same frequency because the reduced mass is the
same in both cases.
In Sect. 3.11 we have considered, as an example of mechanical resonance , a
diatomic molecule, in particular HCl. The two nuclei are small enough to be
considered point-like particles in a very good approximation. Call r 0 their
equilibrium distance. When the distance r is different from r 0, the electron cloud
that in the molecule surrounds the nuclei exerts a force, which, in a first
approximation, is proportional to the displacement s = r – r 0. The force is then
elastic and the system is quite similar to the one we just discussed. As a matter
of fact, the internal motions of molecules are correctly described by quantum
mechanics. Our discussion should be considered a first approximation.
The potential energy of the interaction between the two nuclei, which we
have already considered in Sect. 3.11, is shown in Fig. 7.3. The dotted parabola
around the minimum is an approximation of the potential energy corresponding
to the elastic force. In this approximation the potential energy is

Fig. 7.3 The energy potential of a diatomic molecule

(7.11)
The equation of the parabola is written in Fig. 7.3 in eV units of energy and
nanometer units of length. Expressing them in joule and meters respectively we
obtain and the “spring constant” equivalent is
.
We calculate now the reduced mass. In atomic mass units
(u = 1.66 × 10−27 kg) the masses of hydrogen and chlorine are (approximately)
equal to 1 u and 35 u. In the same units , which is
close to the smaller hydrogen mass.
 Finally, the proper oscillation frequency is , which is
the value we used in Sect. 3.11.
As a second example, consider a molecule of carbon oxide (CO). The
potential energy is quite similar to HCl, also quantitatively. We then take the
same value of the “elastic constant”.
As for the reduced mass we must consider that the masses of 12C and 16O are
respectively 12 u and 16 u. The reduced mass is then
Notice that, this time, the two
masses are similar and the reduced mass is substantially different from, and
smaller than, each of them. The reduced mass of a system of two equal masses is
one half of each of them.
Concluding our calculation, we find the oscillation frequency ,
which is not too different, considering our approximations, from the measured
value

7.3 Double Stars

In this section we shall consider a two-body system moving in two dimensions
rather than one dimension as the diatomic molecules. It will also be a much
larger astrophysical system. In Chap. 4 we discussed the motion of a planet, of
mass m, about the sun, of mass M or of a satellite around its planet, assuming the
sun in the first case, the planet in the second, to be at rest. From the discussion of
the last sections one clearly understands that the assumption is not rigorously
true. Indeed, both bodies move around their center of mass. However, in those
cases the mass of the central body is much larger than the one of the orbiting
body, and the approximation is quite good. We shall now consider an
astronomical system in which the two masses, say m 1 and m 2 are similar.
We know today that a large fraction of the stars are in fact double , or, in
several cases, even multiple. To establish these facts, the image of the star
system needs to be resolved in those of its components. Telescopes of adequate
resolving power are needed.
The first double star system was discovered in 1780 by Sir William Herschel
(1738–1822) in the Ursa Major constellation. It is called Xi Ursae Majoris. More
double stars were discovered by Sir William and his son John (1792–1871) in
the following years. The study of double stars gives a further opportunity to
check the Newton theory.
Figure 7.4 shows the apparent positions, namely the angles under which the
objects are seen from earth of the Xi, as measured for more than a century. The
motion must be studied in the CM frame, as in Fig. 7.5. C is the center of mass
of the system, r 1 and r 2 the position vectors of the two stars, which are point-
like in a good approximation, and m 1 and m 2 their masses. Let r be the vector
from m 2 and m 1.

Fig. 7.4 The apparent positions of one star relative to the other (the dot inside the curve) for the Xi Ursae
Majoris double star

Fig. 7.5 a Diagram for the motion of a double star system; b case of circular orbits

From the definition of center of mass and

(7.12)
We know that the force, call it F(r), acting on m 1 is the attraction of m 2 and
consequently that it is directed as r. The acceleration is and
the Newton law
 (7.13)
In this case too, as in one dimension, we have found that the motion of a
body of mass m 1 around another body of mass m 2 when both are moving is the
same as when m 2 is at rest if, (a) we substitute for m 1 the reduced mass of the
system, (b) we work in the CM taking into account that the center of the forces is
the center of mass.
Figure 7.4 shows that the orbit shape is an ellipse. However, one of the stars
does not look to be in a focus of the ellipse. This is an optical effect due to the
fact that we are not looking normally at the orbit plane, but at a certain angle.
An interesting feature of binary systems is that their period depends only on
the sum of the masses and not on their ratio. This is true in general, but, for
simplicity, we restrict ourselves to the circular ones, as shown in Fig. 7.5b. The
two stars rotate around the center of mass with the common angular velocity ω.
The motion of one of them, m 1 for example, is given by the Newton equation
and hence and, for Eq. (7.12)

(7.14)
By measuring the period T and the distance r between the stars we can
determine the sum of their masses.

7.4 Tides
The level of water contained by the seas and oceans varies during the day. The
level grows (flux) till it reaches a maximum level ( high tide ) and then decreases
(reflux) to a minimum ( low tide ) and so on. The phenomenon is periodic with a
period (for example between consecutive high tides) of 12 h 25′, which is
exactly equal to one half the time taken by the moon to come back to the same
position relative to earth, namely its revolution period. Consequently, since
ancient times tides were thought to be due to the moon . The explanation of the
phenomenon however is not at all simple and had to wait for Newton .
Considering that we observe the phenomenon on earth, we shall describe it
in a reference frame fixed on her. The first idea coming to mind is that the moon
attracts the parts of the oceans nearest to it more strongly, causing their rise. But
it does not work, because after half a period, when the moon is in its farthest
position, we observe another rise rather than a lowering. The explanation must
be different.
 We cannot consider here the earth as point-like. We must take into account
that the gravitational field of the moon is different in different points of the
earth’s surface, that have different distances from the moon. We shall work in a
reference frame with the origin in the center of the earth. Notice that it cannot be
considered inertial in the present discussion. If the gravitational force was equal
in all the points of earth, it would be exactly balanced by the inertial force
(centrifugal) due to the accelerated motion of the center of the earth, as we have
seen in Sect. 5.7. Actually, the gravitational force is exactly balanced by the
centrifugal one only in the earth’s center. On the part of the surface nearer to the
moon, the moon gravitational force is larger than the centrifugal one. On the
opposite part the centrifugal force is larger than the gravitational one.
We underline that the inertia force we are considering is due to the
acceleration of the origin of the reference frame (the center of the earth) that is
rotating during the day around the center of mass of the earth-moon system. We
also observe that we are neglecting the action of the sun on earth, which is much
more intense than that of the moon. We can do that, in a first approximation,
because what matters here is not the gravitational field itself but its differences
in the different points of the earth. As a consequence of the much larger distance
(400 times) of the sun than the moon, its field, even if stronger, is much more
homogeneous. However, the sun does have an influence. We shall come back to
that at the end of the section.
To simplify the problem, we shall consider the earth as a solid sphere with a
layer of water of constant depth on the surface. We also assume the moon
moving in the plane of the Equator. Figure 7.6 shows a view in this plane. The
earth and the moon, a two-body system, rotate about their common center of
mass. The accelerations of both are directed towards the center. We can also
think that both are continuously falling towards the center of mass.
Fig. 7.6 a The geometry of the problem, b the tide force in different points of the earth surface

In the point A, in which the moon is at the zenith, its gravitational attraction
is larger than in O, because A is closer to it. As a consequence, the water
particles in A fall towards the center of mass, and towards the moon too, with a
larger acceleration than the earth’s center O. On the contrary, in the point B in
which the moon is on nadir, the gravitational attraction of the moon is smaller
than in O and the water particles there fall towards the center of mass, and the
moon, with an acceleration smaller than O.
We have followed the argument of Newton till here. However, at this point,
Newton made a mistake (followed by several authors). The error is to extend
what was established for the accelerations of water particles to their
displacements. If we could do so, we would say that the water particles in A
move towards the moon more than the center O and the sea rises, while those in
B move towards the moon less than O. The sea moves away from the moon, and
rises here too. The situation is shown in Fig. 7.7a. The ocean presents two
bumps, diametrically opposed, on the line joining the moon with the earth’s
center. The bumps move in phase with the moon. We then expect high tides to
take place just when the moon passes at the zenith and at the nadir, the low tides
in quadrature, i.e. at a quarter of the period relative to those positions.

Fig. 7.7 Schematic view of the earth and of the tides. a Phase as foreseen by Newton. b Phase as actually
observed (approximately)

The observations, however, do not confirm these predictions. Rather, high

tides happen when the moon is about in quadrature, the low tides when she is at
the zenith and at the nadir as shown in Fig. 7.7b. The presence of the continents
and other effects make the situation more complex. In any case, however, the
delay between the passage of the moon on the zenith and nadir and the high tide
is always of several hours. The disagreement is a consequence of the above-
mentioned mistake. The displacement of a water particle at a certain time is not
parallel to the acceleration in that instant.
The correct treatment of the tides can be divided in two parts. In the first part
we calculate the tide-generating force as a function of a point on the earth’s
surface. The second part is the calculation of the forced oscillation of the oceans
under the action of that force. This calculation is complicated by the presence of
continents. We shall show the basic points of the argument.
Let us start with the force. Both the gravitational and the inertia force acting
on a water particle are proportional to its mass. We can then consider the force
per unit mass. The weight per unit mass, g, has no influence on the tides,
because it, even if different from point to point, is constant in time in each point.
If the moon did not exist, the surfaces of the sea would be in any point
perpendicular to g. The tide-generating force per unit mass, that we shall call f,
is, as we already said, the resultant of the gravitational attraction of the moon
and of the centrifugal force due to the rotation of the earth’s center around the
center of mass of the moon-earth system.
We shall not perform the calculation of f (which is not difficult). We show
the result in Fig. 7.6b. We shall however, evaluate the order of magnitude of f,
calculating it in A, where it is particularly easy. In the center O the gravitational
attraction of the moon and the centrifugal force are equal. In A the centrifugal
force is the same and is larger than in O. In this point they have equal and
opposite directions. The magnitude of the sum of the gravitational attraction and
the inertia force in A is consequently the difference between the gravitational
attraction in A and the gravitational attraction in O (because the latter is equal to
the inertia force both in O and in A) . In conclusion, with r EM the earth moon
distance, R E the earth radius and M M the moon mass, we have

Considering that the radius of the earth is much smaller than the earth-moon
distance, R E /r EM ~ 1/60, we can expand this expression in series of this
quantity and stop at the first term. We have

(7.15)
This is the tide-generating force per unit mass in the point A, which has the
dimensions of an acceleration. Let us compare it with the weight per unit mass,
, where M E is the earth mass. We have, in the right-hand side, with
and ,
 (7.16)
First we observe that the tide-generating force is inversely proportional to the
cube of the earth-moon distance. In fact it depends on the differences between
the gravitational force in different points, namely the derivative of the
gravitational force. The latter varies inversely as the square, its derivative as the
cube.
We observe that the tide-generating force is very small, but still enough to be
a cause of such important phenomena. As a matter of fact, the height of the tide
is of the order of a few to several meters, corresponding to a fraction of 10−7 of
the earth diameter.
Calculations show that the magnitude of the tide-generating force is the same
everywhere, hence is equal to what we calculated. Its direction, as shown in
Fig. 7.6b varies as a function of the point.
To be precise, we notice that the moon’s orbit is elliptic. Its distance from
earth varies between 57 and 63.7 earth radii. Consequently f/g varies from
1.33 × 10−7 to 0.96 × 10−7.
We now pass to the second part of the theory. Let us look at the situation in a
point of the earth’s surface. As we have said, the magnitude of the tide
generating force is constant in time, but its direction varies. Its variation is a
rotation at constant angular velocity. In other words, the components of the
force, say the horizontal and vertical ones, vary periodically in time. When the
former is a maximum the latter is null and vice versa. The ocean, which we still
imagine to cover the entire surface, is subject to a periodic force, varying in time
as a circular function. Even if the system is much more complex than a
pendulum, it behaves as a forced oscillator .
Consider for example a drop of water in the air of a spaceship. Its natural
shape is spherical. If we deform it a bit and then we let it go, it will tend to go
back to its natural shape. But it cannot do that directly. Rather, like a pendulum,
it will oscillate between different shapes and alternate between oblate and
prolate. The oscillations have a proper period, which depends on the physical
characteristics of the drop, and, if dissipative forces are present, are damped. The
same would happen if, in absence of the moon, we would deform the surface of
the ocean around the earth and abandon it. The system would oscillate at its
proper oscillation frequency or, in other words, with the period, call it T 0, of the
free oscillations of the system. Calculating T 0 is extremely difficult due to the
complicated shape of the continents and of the sea bottom. Calculations on
simplified models lead however, to values of T 0 = 20–30 h.
 We can imagine the ocean as an oscillator, with proper oscillation period T 0.
The oscillator is forced by a periodic force of period T = 12 h 25′, which is much
smaller than T 0. In other words, it is an oscillator forced at a frequency
substantially larger than the resonance frequency. In these conditions, as we
know (see Fig. 3.21b), displacement and force are in phase opposition.
Consequently, the correct shape is that of Fig. 7.7b, not that of Fig. 7.7a, in
substantial agreement with observations.
We now come back to the action of the sun. The reasoning is exactly the
same as for the moon, and the result analogous to Eq. (7.16) is reached,
obviously with the mass and the distance of the sun in the place of those of the
moon. It is so found that the magnitude of the tide-generating force due to the
sun is about half than that due to the moon. The two forces must be obviously
summed as vectors. The two forces reinforce one another when the sun and the
moon are about on the same line (new and full moon). The tides are then
particularly ample (a condition called a syzygy ), about one and a half larger
than the value for the moon only. On the contrary, when the moon is at the first
or last quarter, at 90° with the sun, the two forces partially cancel each other and
the tides have small amplitude (quadrature tides ), about one half as for the moon
alone.
In practice, the height of the tides depends on several other factors, like the
shape of the shores of the continents and the islands, the shape of the sea bottom,
the oceanic currents, the winds, etc. Near the oceanic islands the height of the
tides is typically one meter and near the continental shores it is about twice
larger. However, in some sites the tides reach three meters and in a few even six
meters. Particularly great tides are observed in deep gulfs or fiords facing the
open sea. The greatest tides are in the Bay of Fundy, in Nova Scotia, Canada.
Their amplitude is 4 m at the bay entrance, to reach 14 m at its end and even
more at the syzygy.

7.5 Impulse and Momentum

Consider a material point of mass m in an inertial reference frame. Let F be the
resultant force acting on the point. The second Newton law can be written in the
form
(7.17)
In words, the effect of a force in the time interval dt is a variation of
momentum equal to the product of the force and of the time interval. The vector
quantity F dt is called elementary (meaning infinitesimal) impulse of the force in
dt. The impulse of a force in a non-infinitesimal time interval from t 1 to t 2 is
defined to be

(7.18)

Integrating in that time interval Eq. (7.17) we immediately have

(7.19)

This equation expresses the impulse-momentum theorem : the momentum

change of a material point under the action of the force F in the time interval
from t 1 to t 2 is equal to the corresponding impulse, whatever is the time
variation of the force and whatever is the length of the interval.
The impulse-momentum theorem is useful when the force acts for a short
time, like in the collisions, strokes, explosions, etc. In these cases the force is
initially null, then it quickly grows and as quickly goes back to zero. In these
cases we do not usually know the instantaneous values of the force, but only its
average value. The average value of a quantity in a given time interval is, by
definition, the integral of the quantity over that time interval divided by the
length of the interval. For the force, as shown in Fig. 7.8,

Fig. 7.8 An impulsive force and its average value

Example E 7.1
The hammer is an instrument used since ancient times to amplify the muscular
force. Initially, at time t 1, a hammer of mass m, is at rest. With our arm we apply
to it a force of average value till the instant t 2 in which the hammer strikes
the head of the nail. In accordance with the impulse-momentum theorem, in this
instant the momentum of the hammer is After that, the hammer
slows down and stops (its momentum becomes zero) at time t 3. For the same
theorem, the average force on the nail in the interval from t 2 to t 3 is
In conclusion, Clearly, t 3 – t 2 is
much smaller than t 2 – t 1 so that we obtain a large amplification of the force, by
factors than can well be three orders of magnitude.

7.6 The Action-Reaction Law

Consider again a two-body system, made of two material points, which we call 1
and 2, and later, to follow Newton, A and B, of masses m 1 and m 2. The two
points interact, point 1 acting on 2 with the force F 12 and point 2 acting on 1
with the force F 21. The two forces are an action and reaction pair. The third
Newton law states that they are equal and opposite
(7.20)
We shall assume that no external force exists or, if some do, their resultant is
zero. We deal with an isolated system .
F 21 being the only force acting on point 1, it is equal to the rate of change of
its linear momentum, or quantity of motion, p 1, and similarly F 12 is equal to the
rate of change of p 2. Equation (7.20) immediately gives

(7.21)
and also

(7.22)
where we have put P = p 1 + p 2. This is total linear momentum (or total
quantity of motion ) of the system. Equation (7.22) implies that
(7.23)
This equation expresses the principle of conservation of linear momentum in
the case of a two-particle system. The principle states that total momentum of an
isolated system is constant. We shall prove its general validity later in this
chapter. In this section we shall use it in an experimental proof of the third law,
as Newton himself did. Indeed, we have just seen that, for a two-body system ,
the principle is a consequence of the action-reaction law. It is also true that, if
the total linear momentum of an isolated system is constant, the internal forces
must be pairs of equal and opposite ones. Indeed, the most accurate verifications
of the action-reaction law are, in fact, verifications of conservation of the total
momentum. We observe, however, that in this way we verify the interaction
forces to be equal and opposite, not that they have the same application line. We
shall come back later on to this point.
Historically, the first experimental checks of the action-reaction law were
done by Newton and his contemporaries Christopher Wren (1632–1723),
Christiaan Huygens (1629–1695) and John Wallis (1616–1703). Their
experiments are very accurate, conceptually simple and elegant. The
experiments study the collisions between two spheres of different sizes, measure
the momenta before the collision, say p 1 and p 2, and after, say p 1′ and p 2′, as
accurately as possible and check if the relation is satisfied or not. The
experiments were done by attaching the two spheres to two wires of equal
lengths,
(7.24)
thus making two pendulums of the same period. When at rest the two spheres
touch each other as in Fig. 7.9a. We move the spheres from equilibrium, each at
a certain distance, which we measure. If we let both spheres go at the same
instant from rest, they will accelerate, collide with each other in their lowest
points, separate and move back together.

Fig. 7.9 The two-pendulum experiment to verify the momentum conservation. a Position at rest, b an
initial configuration

The experiment profits from two properties of the pendulum . The first
property is the isochronism of the (small) oscillations. Having the same lengths,
the periods of the two pendulums are equal, independently of the masses of the
spheres and of their initial positions (amplitude). Consequently, also the times
taken to reach the equilibrium position are equal (a quarter of a period) and they
will always collide there, if abandoned at the same time with null velocity. The
second property is: the velocity of the pendulum when reaching the equilibrium
position starting from a certain distance with null velocity is proportional to that
distance. Let us show this property.
Let m be the mass and l the length of the pendulum. Let us remove it from
the equilibrium position by x 0 as in Fig. 7.10 and let it go with null velocity. In
this position, the pendulum is at a certain height, say h, above the horizontal
through the equilibrium position. For small displacement angles we can use for h
the approximate expression, Eq. (4.14)

Fig. 7.10 Geometry of the starting configuration of the pendulum

(7.25)
If υ 0 is the velocity of the pendulum in an equilibrium position, the energy
conservation law states that Hence, using Eq. (7.25),

and also if is the period of the pendulum

(7.26)
We conclude that the velocity υ 0 at a collision will be known if we measure
the period once and the initial position x 0 for each and every experiment.
 We are now ready to read how Newton describes his experiments in the
Principia . He does that just after having stated the third law to prove
experimentally its validity. Newton built two pendulums each 10 ft (about
3.25 m) long, attaching two spheres A and B of the materials to test, and fixing
the two wires in C and D as in Fig. 7.9a. We call m 1 and m 2 the masses of A
and B respectively and x 1 and x 2 their displacement, measured for each
pendulum from its equilibrium position (the position of its center to be precise).
We remove both spheres to x 10 and x 20 respectively and accurately measure
these distances. Notice that x 10 and x 20 can be on opposite sides, both on one
side or both on the other of O. If we let them go at the very same instant with
null velocities, they will collide in O with velocities
(7.27)
Let υ 1′ and υ 2′ be the velocities immediately after the collision. We can
determine them by measuring the maximum distances, x 10′ and x 20′ reached
(contemporarily) in their swing back. Indeed, we have
(7.28)
The two particles interact only during the instant of the collision. The
external forces acting on them, the weight and the tension of the wire, have zero
resultant. However, the system is not exactly isolated because air resistance
exists and is an external force. This is small, but it must be taken into account in
precision measurements. Newton did that as follows. He started operating with
one pendulum only. He removed it from equilibrium at each of the distances that
he was going to use in the following experiments. He let it go with zero velocity
and observed the position reached after one period, which did not coincide
exactly with the original one. He measured the miss. A quarter of that is what is
lost in a quarter of a period due to the air resistance.
He made a number of experiments with spheres of different substances. For
each of them, he tried different pairs of starting positions x 10 and x 20, measured
those reached after the collision x 10′ and x 20′ and applied the just described
correction. Each of them corresponds to a value of the quantity of motion ; he
calls that simply “motion”, before the collision. The linear momentum
conservation law (which is equivalent to the third law) that we need to verify is
(7.29)
He writes (in parenthesis some explanations):

Thus trying the thing with pendulums of ten feet (3.25 m) in unequal as
 well as equal bodies, and making the bodies to concur after a descent
through large spaces, as of 8, 12, or 16 feet (2.6, 3.9, 5.2 m), I found
always, without an error of 3 inches (8 cm), that when the bodies concurred
together directly (in a straight line), equal changes towards the contrary
parts were produced in their (quantities of) motions, and, of consequence,
that the action and reaction were always equal.

He continues giving numerical examples of his results. The initial and final
momenta are given in “parts of motion”, namely in an arbitrary unit. The unit is
clearly irrelevant. For clarity, we shall write the values of the two sides of
Eq. (7.29) for each quoted result at the beginning of each experiment. For each
experiment, he mentions also the changes of the momentum of each body.
In the first experiment B is initially at rest (9 + 0 = 2 + 7).

if the body A impinged upon the body B at rest with 9 parts of motion, and
losing 7, proceeded after reflection with 2, the body B was carried
backwards with those 7 parts.

In the second experiment the initial velocities have opposite directions (12 –
6 = –14 + 8).

If the bodies concurred with contrary motions, A with twelve parts of

motion, and B with six, then if A receded (in its motion after the collision)
with 14, B receded with 8; namely, with a deduction of 14 parts of motion
on each side. For from the motion of A subtracting twelve parts, nothing
will remain; but subtracting 2 parts more, a motion will be generated of 2
parts towards the contrary way; and so, from the motion of the body B of 6
parts, subtracting 14 parts, a motion is generated of 8 parts towards the
contrary way.

In the third experiment the two initial displacements are in the same direction
(14 + 5 = 5 + 14).

But if the bodies were made both to move towards the same way, A, the
swifter, with 14 parts of motion, B, the slower, with 5, and after reflection
A went on with 5, B likewise went on with 14 parts; 9 parts being
transferred from A to B. And so in other cases.

Newton then discusses the causes of the errors in the measurements of the
distances and, as we have read above, evaluates them less than 3 in., 8 cm. The
distances being several meters; this is about 2–3 % error. The relative error on
the momenta was similar (masses and periods being known with a much better
accuracy).

It was not easy to let go the two pendulums so exactly together that the
bodies should impinge one upon the other in the lowermost place AB; nor
to mark the places s, and k, to which the bodies ascended after congress.
Nay, and some errors, too, might have happened from the unequal density
of the parts of the pendulous bodies themselves, and from the irregularity of
the texture proceeding from other causes.

He, and we with him, then observe that the total momentum is conserved
both for elastic and non-elastic collisions . A collision is called elastic if energy
is conserved. This is an idealization; in practice perfectly elastic collisions do not
exist. However, the collision between two steel spheres is close to being so,
between two wax ones is not. In an elastic collision the two forces F 12 and F 21
are conservative. Elastic collisions conserve mechanical energy, inelastic ones
do not, but in both cases the total momentum is conserved. Let us go back to
Newton.

But to prevent an objection that may perhaps be alleged against the rule (the
action and reaction law), for the proof of which this experiment was made,
as if this rule did suppose that the bodies were either absolutely hard, or at
least perfectly elastic (whereas no such bodies are to be found in Nature), I
must add that the experiments we have been describing, by no means
depending upon that quality of hardness, do succeed as well in soft as in
hard bodies.

Obviously, the relative velocity of the bodies after a collision is smaller for
the inelastic than for elastic collisions with the same initial conditions. It may
even be null; the two bodies remain attached. The total momentum however is
always equal to the initial one.

This I tried in balls of wool, made up tightly, and strongly compressed.

He compared the results obtained with balls of steel, glass and cork. The
Newton conclusion is that

And thus the third Law, so far as it regards percussions and reflections, is
 proved by a theory exactly agreeing with experience

In collision experiments the interaction forces act for a very short time,
during which they are very intense. We talk of impulsive forces . The just
described experiments establish that the total momentum is conserved in an
isolated system in which the internal forces are impulsive. And if the forces are
not impulsive? To answer this question Newton did the following experiment.
He fixed a magnet on a piece of wood and a piece of iron on another one. He
leaned both of them on the surface of the water in a container, carefully
controlling them to be perfectly at rest. He let the two bodies go. The two bodies
moved one towards the other, under the attraction of the magnet, attached
themselves to each other and remained still. The important observation is that
the final body, iron plus magnet, does not move on water, even if there is no
impediment to do so. The total final momentum is zero, as the initial one was. In
this experiment too the system is isolated. Indeed, the external forces, weight
and Archimedes force equilibrate each other.
The conservation of linear momentum in an isolated system is a fundamental
law of universal validity.

7.7 Action, Reaction and Linear Momentum

Conservation
The conclusions from experiments we have described and many other ones can
be summarized as follows. We build an isolated two-body system. The resultant
external force is zero. The internal forces are F 21 = –F 12. We start from an
initial state i and measure the two momenta p i1 and p i2. We let the system
spontaneously evolve under the action of the internal forces. When the system
has reached the state, which we call final, f, we measure again the momenta, p f1
and p f2. We always find out that
(7.30)
The linear momentum is conserved. We stated that this proves the action and
reaction law (equality of the application lines apart). Let us look at that more
closely. Equivalently, we can write
(7.31)
In words, the changes of the linear momentum of the two bodies are equal
and opposite. For the impulse-momentum theorem we have
 The experimental verification of Eq. (7.31) is then a verification of

(7.32)

In conclusion, these experiments verify that the time of the force body 1
exerts on body 2 is equal and opposite to the time integral of the force body 2
exerts on body 1. In absence of any contrary evidence, we assume the
instantaneous values of F 21 and F 12 to be equal and opposite too.
Figure 7.11 shows the time evolution of internal forces in the example of a
hypothetical collision. Rigorously speaking, we know from the experiment only
that the two areas are equal and assume that the curves have, in addition, mirror
shapes, namely that the forces are equal and opposite in any instant.

Fig. 7.11 The time evolution of the internal forces during a collision

This assumption is basically a postulate. Moreover, the postulate, namely the

action-reaction law, is not true in every circumstance. There is no problem when
the two bodies interact through contact forces, as in a collision. Problems arise
when the two bodies are separated by a distance. As we have seen in Chap. 6, no
effect can propagate over a distance instantaneously. Consequently, when the
propagation time is comparable with the time in which the change in motion
takes place, the concept of instantaneous equality of action and reaction loses
validity.
Figure 7.12 shows a simple mechanical model of a “delayed” interaction.
The two bodies are two trolleys moving with negligible friction on straight rails.
Trolley 1 carries a gun, trolley 2 a block of material at the height of the gun. At a
certain instant the gun shoots a bunch of projectiles. Suppose the projectiles to
be invisible, have masses much smaller than the trolleys but very high speeds.
Consequently, the bunch carries an appreciable momentum p. We observe the
system and see, in the instant of the shot, trolley 1 to recoil with a momentum –p
while trolley 2 remains still. The total momentum of the two trolleys is changed.

Fig. 7.12 Mechanical model of an action at a distance

If υ is the velocity of the bullets and L the distance between the trolleys, the
bullets will reach trolley 2 in a time L/υ and stick to the block. We observe
trolley 2 acquiring a momentum p. The total momentum of the two blocks is
now null, as it was initially. Momentum conservation is restored.
The example looks a bit stupid. The momentum seems not to be conserved
during the time the bullets are in flight, just because we did not include their
momentum in the total, assuming them to be “invisible”. If we include that, as
we should, the total momentum is conserved in every instant. However, things
are not very different in the cases of actions at a distance as the gravitational and
electromagnetic ones. Light, in particular, is an electromagnetic phenomenon.
Consider again two trolleys, now very light, again with negligible friction. The
first trolley carries a lamp that emits a light flash at a certain instant. Now, light
carries momentum, even if in a very small amount. Consequently, the first
trolley recoils with an opposite momentum (p), while the second is still at rest.
Suppose the second trolley carries a black screen, which absorbs the light pulse
completely, acquiring the momentum –p. The situation is quite similar to the
“stupid” mechanical example. However now during the time of flight of the light
the total mechanical momentum is not conserved. The missing momentum is,
during this time, in the electromagnetic field. We shall study this in the 3rd
volume of this course. We only notice here that this is basically the reason for
which Eq. (6.50) that we found in discussing relativity is not valid for non-
interacting particles. In quantum mechanics the analogy is even closer; light is
made of “invisible” particles, the photons. A quite similar situation exists for the
gravitational interaction. In this case also the gravitational field carries
momentum. This is described by general relativity.

7.8 Systems of Particles

We shall now start our study of systems of several, say N, of material points. The
relevant physical quantities are shown in Fig. 7.13, in an inertial reference frame.
Let r i be the position vector of the generic point P i in a generic instant. We call
the set of positions of its constituent material points a configuration of the
system .

Fig. 7.13 A system of material points

Let m i be the mass of P i , v i its velocity and p i = m i v I its momentum. The

forces acting on each point can be usefully divided in internal, due to the other
points of the system, and external, due to agents external to the system.
Consider for example the system of Jupiter and its satellites. The forces on
one of them, Ganymede for example, are the internal ones due to Jupiter and to
the other satellites, Io, Europa, Callisto, and the external ones due to the sun and
the other planets. Obviously, being an internal or external force depends on the
system under consideration. If the system is the solar system, all the mentioned
forces are internal.
We call the resultant external force and the resultant internal force
acting on P i . All the forces, both external and internal, according to the Newton
law, determine the motion of P i ,

(7.33)
The motion of a system of N points is described by N independent Eq. (7.33).
Their solution is in general quite difficult. Indeed, just think of the fact that the
force acting on a certain point at a certain time depends not only on its position,
but on those of all the other points too. The problem is so complicated that even
in the simplest case N = 3 cannot be in general solved analytically. Numerical
methods are today available to solve the problem with the help of powerful
computers.
We shall not analyze the motions of single points, but rather consider
quantities relative to the whole system. We indicate with Ω a geometric point
that we choose as the pole of the linear momenta and of the moments of the
forces. This point is not necessarily at rest in the reference frame, rather it moves
with a velocity that is a function of time, v Ω . The angular momentum of point P
i about Ω is

(7.34)
Let f 1,i , f 2,i ,…. be the forces acting on the point P i and F i = f 1,i + f 2,i
+ …. their resultant. All these forces are applied to the same point and,
consequently, their total moment is equal to the moment of their resultant. The
external moment acting on P i is then

and the internal moment

The global quantities of the system that we shall need are the following:

(1) The total linear momentum of the system, which is the vector sum of the
linear momenta of the constituent points

(7.35)

(2) the total angular momentum

(7.36)

(3) the total kinetic energy

 (7.37)

(4) the resultant force

(7.38)
where the vectors in the last side are the resultants of internal and external
forces acting on the system.
We now make a very important observation that will greatly simplify
several problems. The internal forces come in pairs; the force exerted on
point P i by another point P j is equal and opposite to the force that P j exerts
on P i and their sum is null. Consequently the resultant internal force is zero,
, and Eq. (7.38) becomes

(7.39)

(5) The total moment about the pole Ω is

(7.40)
where the vectors in the last side are the total moment of the internal and of
the external forces respectively.

Notice that we can calculate the total moment of the forces acting on a single
point P i or calculate first the moments of the different forces and then sum them,
or sum the forces and then calculate the moment of the resultant. On the
contrary, to calculate the total moment acting on the system we must first
calculate the moments of the forces on the single points and then sum those
moments. Indeed, in this case the forces are applied in different points.
A second important observation is the following. The internal forces come in
pairs that, for the action-reaction law, not only are couples, but also zero arm
couples. Consequently, the moment of each couple is null, whatever is the pole.
The total internal moment is zero, and we can write

(7.41)

7.9 The Center of Mass

We continue with the system of N material points. Figure 7.14 represents the
situation.

Fig. 7.14 A material system and its center of mass

We define as the center of mass of the system the geometric point (it is not a
material point) defined by the position vector

(7.42)

where M is the total mass of the system. The coordinates of the center of
mass are, clearly

(7.43)

It can be shown, but we shall not do so, that the position of the center of
mass is independent of the choice of the reference frame. However, obviously,
its coordinates depend on that. We already met the center of mass in the
particular case of a two-point system. In this case the center of mass is the point
of the segment joining the two points at distances from them inversely
proportional to the masses. It can be shown that the two definitions agree in this
particular case.
We now consider the motion of points of the system. We call v i the velocity
of P i (which is a function of time). By deriving Eq. (7.42) we find that the
velocity of the center of mass is

(7.44)
We observe that the sum in the right-hand side of this equation is just the
sum of the linear momenta of the points, namely is the total momentum of the
system

(7.45)
We can write Eq. (7.44) as
(7.46)
which is a very important equation. It states that the total momentum of the
system is equal to the momentum of the center of mass , if considered as a
material point in which all the mass of the system is concentrated.
Consider now how the total momentum varies in time. We work in an
inertial reference frame. Taking the derivative of Eq. (7.45) we have

(7.47)
but, as we are in an inertial frame, m i a i is equal to the resultant force, both
external and internal, acting on P i .

(7.48)
Substituting this in Eq. (7.47) we have

but, as we know, the resultant internal force is zero, a fact that enormously
simplifies the equation. It becomes
 (7.49)
This fundamental equation states that the rate of change of the total
momentum of a mechanical system is equal to the resultant external force acting
on the system. The fact that the internal forces do not contribute to the variation
of the total momentum simplifies many problems.
We now go back to Eq. (7.46) and immediately see that
(7.50)
which is called the theorem of the center of mass motion : the center of mass
moves as a material point in which all the mass of the system is concentrated
and acted upon by the resultant external force. Notice that while the motion of
the center of mass is determined by the external forces only, the motion of each
point of the system depends on both external and internal forces.
As an example, suppose we take in our hand the handle of a hammer, and we
launch it in the air. The motion of the hammer will be a complicated
combination of rotations and displacements. The motion of its center of mass, on
the contrary, will be simply a parabola, with the hammer rotating about it
(neglecting air resistance). For that the body does not need to be rigid. If we
launch a chain in the air, its center of mass will describe a parabola too. In a
similar way, consider the bullet shot by a cannon. It describes a parabola. If at a
certain moment the bullet explodes, its pieces will describe complicated
trajectories, but their center of mass will continue on the same parabola, as long
as the first piece hits the ground. When this happens a new external force, due to
the action of ground, starts acting on the system.
The center of mass, as we have seen, is not a material point but behaves as
such.

7.10 Linear Momentum Conservation

The law (or principle) of conservation of linear momentum states that: if, in an
inertial frame, resultant external force on a system is zero, the total linear
momentum is constant in time. The property is immediately obtained from
Eq. (7.50)
(7.51)
We can also say that, under the same hypotheses
(7.52)
If the resultant external force is zero in an inertial frame the center of mass
remains still if initially still or continues on its rectilinear uniform motion.
In Sects. 7.5 and 7.6, we have already used the center of mass properties and
the linear momentum conservation principle in the particular case of two-body
systems and discussed the relations with the action-reaction law.

7.11 Continuous Systems

The mechanical systems we have considered so far are discrete, namely
composed of a number of point-like particles. We shall now consider continuous
mechanical systems. Such are the solid bodies when their physical dimensions
cannot be neglected. Figure 7.15 represents a continuous body of mass M and
volume V.

Fig. 7.15 A continuous body and an infinitesimal volume element

We can divide the body into small volumes dV, which we take as cubes with
sides parallel to the coordinate axes. Let r be the position vector of the generic
dV and Δm its mass. We define the density ρ(r) of the body in the position r to
be the ratio between the mass and the volume of the element in the limit in
which the volume becomes very small, namely

(7.53)
The density can vary from point to point. Think for example of the
atmospheric density that decreases with altitude. A body is said to be
homogeneous if its density does not vary from point to point.
Here we need to specify that the limit should be understood as a
physical rather than mathematical limit. Indeed, when seen at a molecular scale,
matter is not continuous, but made of small particles, the molecules, separated
one from another. Consequently, the limit for volumes going mathematically to
zero is not defined. However, the granularity of matter is so small compared to
the macroscopic sizes and we can safely state that the limit is taken for volumes
very small compared to macroscopic dimensions but still large enough to contain
a great number of molecules. Indeed, we can say, for volumes physically tending
to zero.
The definition of center of mass for a continuous system is completely
analogous to that we gave in Sect. 7.9 for a discrete system. We divide the
system in N small volumes ∆V i , then use Eq. (7.42) to define the center of mass
and take the limit for the small volumes tending physically to zero. We obtain

The position vector of the center of mass is then

(7.54)
or, its coordinates are

(7.55)
In this chapter we shall continue the study of material systems. For the sake
of simplicity, we shall consider them discrete. The discussion of continuous
systems is completely similar, just changing sums with integrals. The limitation
to discrete systems does not subtract anything from the physics conclusions.
As examples, we shall now calculate the position of the center of mass in
two examples of homogeneous bodies of simple geometrical shapes.

Example E 7.2
Figure 7.16 represents a thin sheet in the form of an isosceles triangle of height h
and base b. It can be considered two-dimensional and the volume integral (7.54)
becomes a surface integral. It is evident, for symmetry reasons, that the center of
mass must be on the height of the triangle (the same quantity of mass must lay
on the right and on the left). We need only to find its y coordinate. It is
convenient to take as surface elements strips of height dy running from one side
to the other. Indeed all points of such a strip have the same y and equally
contribute to the integral. The length l(y) of the strip at height y can be found
considering the proportion l(y):b = y:h. Hence we have The area of
the strip is and, if σ is the surface density, namely the mass per
unit area, its mass is We then calculate the integral
(Fig. 7.17)

Fig. 7.16 Calculating the center of mass of a homogenous isosceles triangle

Fig. 7.17 Calculating the center of mass of a homogenous cone

The mass M of the body is σ times the area hb/2 and we have

Example E 7.3
Figure 7.17 represents a homogeneous cone of height h and base radius R. As
evident in this case too, the center of mass is on the axis. To calculate its height
y, we take as volume elements thin sheets parallel to the base. All the points of a
sheet have the same height y. The volume of the sheet at y is dV = πR 2 (y) dy.
But r(y) = Ry/h and, if ρ is the density
which we must divide by the mass, that is , obtaining .

7.12 Angular Momentum

The fundamental equation Eq. (7.49), describes the evolution in time of the total
linear momentum of a mechanical system. We shall now see how the total
angular momentum varies in time. Figure 7.18 shows a mechanical system in a
reference frame, which we choose to be inertial. We arbitrarily choose a
geometric Ω to be the pole of the moments and angular momenta. The pole is not
necessarily still, and we call v Ω its velocity.

Fig. 7.18 The material system

The total angular momentum about the pole we have chosen is

(7.56)
We take the time derivative and obtain

(7.57)
The vector is the difference between two vectors, , both of
which vary in time. Consequently its time derivative is .
 In the second term in the right-hand side we have the rates of change of the
linear momenta of single points. As we are in an inertial frame, the rate of
change of p i is the resultant force, both internal and external, acting on the point
P i . We can write

The first term in the right-hand side is zero, being the sum of cross products
of parallel vectors. The sum in the second term is the total linear momentum P of
the system. The third term is the total moment of the external forces M (e). The
last term is the total internal moment, which is zero. In conclusion Eq. (7.57)
becomes

(7.58)
The expression becomes still simpler with two different choices of the pole.
If the pole is fixed in the (inertial) reference frame, v Ω = 0 and

(7.59)
This fundamental equation reads: the rate of change of the total angular
momentum of a mechanical system about a pole fixed in an inertial frame is
equal to the moment of the external forces about the same pole.
If the pole coincides with the center of mass , which generally moves, the
second term in the right-hand side of Eq. (7.58) is again zero. It is the cross
product of two parallel vectors, the velocity of the center of mass and the total
linear momentum. We can write

(7.60)
In words: The rate of change of the angular momentum of a mechanical
system about its center of mass as a pole is equal to the total external moment
(about the same pole).

7.13 Angular Momentum Conservation

The principle of conservation of angular momentum states that in an isolated
system the total angular momentum, about any pole fixed in an inertial frame is
conserved.
Indeed, in an isolated system the resultant external force and the total
external moment are zero. If the pole stands still in an inertial frame Eq. (7.59)
holds, we can state that the time derivative of the total angular momentum is
zero.
Similarly for Eq. (7.60) we can state also that in an inertial frame the total
angular momentum of any isolated system about its center of mass is constant.
Even if in the case of non-isolated systems, namely in the presence of
external forces, it is sometimes possible to choose a fixed pole, such that the
total external moment about it is zero. Then, the total angular momentum about
that pole is conserved. We shall see some examples in the following.
Notice also that the total external moment may be zero and their resultant
different from zero, or vice versa. Consequently, the linear momentum and the
angular momentum conservations are in general independent issues. If the
system is isolated however, both quantities are conserved.
Finally we observe the following. As we have seen, conservation of the total
linear momentum is a consequence of one aspect of the action-reaction law:
action and reaction are equal and opposite. The total angular momentum
conservation is a consequence of the second aspect of the third law: action and
reaction have the same application line. All the experimental evidence, without
exceptions, is in favor of the angular momentum conservation. Consequently,
also this second aspect of the third law must be considered experimentally
proven.
The linear and angular momentum conservation laws are fundamental
principles of physics, not only of mechanics. In advanced treatments we can
show that they are consequences respectively of the homogeneity of the space
(there are no privileged points in space) and of its isotropy (there are no
privileged directions).

7.14 Energy of a Mechanical System

We continue to consider a material system of N material points P i in an r i the
position vector of P i , m i its mass and v i its velocity. The generic point P i has
the kinetic energy , and the total kinetic energy of the system is

(7.61)
During the motion of the system its kinetic energy will, in general, vary,
because the single kinetic energies of the points vary under the action of the
forces. Let and be the resultants of external and internal forces acting on
P i respectively. In the generic elementary time interval dt the displacement of
the point is d r i . The corresponding elementary work of the forces is

Consider the point P i moving on a certain trajectory from an initial position

A in r iA to a final position B in r iB . The variation of its kinetic energy is given
by the kinetic energy theorem

In words, the variation of the total kinetic energy of a system is equal to the
works of both the external and internal forces. Differently from the cases of the
total linear and angular momenta, the contribution of internal forces is not zero.
If all forces acting on the system are conservative, the work can also be
expressed as a difference of potential energy. Calling U P the total potential
energy, which is the sum of the potential energies of all points of the system, we
immediately find that
(7.62)
We define the total energy of the system U tot as the sum of its potential and
kinetic energy and we see that it has the same values in A and in B. Considering
that these points are arbitrary, we conclude that the total energy is constant
during movement of the system
(7.63)
If the system is isolated, there are no external forces and only the internal
ones make work. This does not imply that the total energy is conserved. For that
to be the case all of the internal forces must be conservative. As an example
consider a system made by a block and a trolley supporting it. The trolley can
move on rails without appreciable friction, but there is friction between the plane
of the trolley and the block. The block moves on that plane. The plane exerts a
friction force on the block and so does the block on the plane. The two forces are
equal and opposite with the same application line. During a motion, the total and
angular momentum are conserved, but not kinetic energy.

7.15 Center of Mass Reference Frame

It is often useful to consider the motion of a mechanical system in the center of
mass frame , CM for brevity, even if that frame is not usually inertial. We start
from an inertial frame and we define as center of mass frame, the frame with
origin in the center of mass of the system and with axes parallel to the axes of
the inertial frame.
Figure 7.19 shows the two just mentioned reference frames and a generic
point of the system P i . The CM frame does not rotate relative to the inertial
frame, it translates with the center of mass velocity. This may vary and, as a
consequence, the CM frame is not in general inertial. It is so if the resultant
external force is zero (even if the total external moment is not) because then the
velocity of the center of mass is constant.

Fig. 7.19 The inertial frame xyz and the center of mass frame x*y*z*

We shall indicate with an asterisk the quantities in the CM frame. The

relation between the position vectors in the two frames is
(7.64)
and the relation between the velocities is
(7.65)
and similarly for the accelerations
(7.66)
Obviously, the center of mass position vector and the velocity are null in its
reference
(7.67)
In Sect. 7.9 we have found that in every reference frame, both inertial and
not, the total linear momentum of a system is equal to the mass of the system
times the velocity of the center of mass. But the latter is null in the CM frame
and we have

(7.68)
The CM frame is also the frame in which the total linear momentum is zero.
It is sometimes called the center of momenta frame .
We obtain another interesting property by expressing Eq. (7.42) in the CM
frame. For the first of Eq. (7.67) this becomes

(7.69)
We now consider the total angular momentum , which has an important role
in mechanics. We might expect it to be different in the two frames of Fig. 7.19,
the inertial and the center of mass frames. As a matter of fact they are equal.
Indeed the total angular momentum in the inertial system is

The first term in the last side is the angular momentum about the center of
mass, as a pole, in the CM frame, while the second is zero for Eq. (7.69). Hence
(7.70)
We conclude that the total angular momentum of a material system about its
center of mass is an intrinsic characteristic of the system, independent of the
reference frame.

7.16 The König Theorems

The two König theorems that we shall discuss in this section give two relations,
one between the kinetic energy in the inertial and CM frames U K and
respectively, and one between the angular momenta L and L*. In both cases the
quantity in the inertial frame is equal to the sum of two terms, one of the system
as a whole, the other corresponding to its motion relative to the center of mass .
The theorems are named after Johann Samuel König (1712–1757)
König kinetic energy theorem.
The kinetic energy of the system in the inertial frame is
 Using Eq. (7.65) it becomes

The expression in parenthesis in the last term of the last side is the total
momentum in the CM frame , hence is null. And we obtain

(7.71)
We read this expression as: the kinetic energy in the inertial frame is the sum
of two terms. One term is the kinetic energy “of the center of mass”, if we think
of it as being a material point with all the mass of the system. The second term is
the kinetic energy in the center of mass system, namely relative to the motion of
the parts of the system about the center of mass.

Example E 7.4
A child is sitting on a wheelchair near to a wall with his feet resting on it with
folded legs. The child, in stretching his legs, pushes on the wall and accelerates
backward. After his feet detach from the wall he continues to move at constant
velocity (neglecting frictions). What forces have caused the acceleration? Which
force is the variation of kinetic energy?

Our system is the child and the chair. We cannot consider it as point-like,
because the stretching of the legs changes the shape of the system. The resultant
external force is the normal reaction of the wall, N. This is the force causing the
acceleration. If m is the mass and a CM the center of mass acceleration, we have
N = m a CM .
The work of the external force N is, on the other hand, zero, because its
application point does not move. Which is the cause of the kinetic energy
variation?
In the analysis of this type of problem, the following mistake is often made.
It consists in application of the kinetic energy theorem to the center of mass, in a
form valid for the material point. Indeed, the center of mass behaves as a
material point from several points of view, but not from this one. Let us look at
that. We can write Eq. (7.50), which is valid for the center of mass, as

which is formally identical to the law and is valid for the material point. We
try now to go ahead as we did in Sect. 2.10 to show the kinetic energy theorem
for a material point. We indicate with d s CM the elementary displacement of the
center of mass in dt, in order to have We take the dot product of
the above equation and d s CM obtaining

We indicate with Γ the trajectory of the center of mass and we consider two
positions, A and B on Γ. As we did for the material point we integrate the above
expression on Γ from A to B obtaining

(7.72)

which has the same form as (2.36). Its meaning is however fundamentally
different. While the right-hand side of Eq. (7.72) is indeed the difference of
center of mass kinetic energy, the left-hand side is not the work of the resultant
external force. This is because d s CM is the displacement of the center of mass,
not of the application point of the resultant. The latter may not even have been
defined. It is defined only if all the forces are applied to the same point.
Consequently, Eq. (7.72) is not very useful in practice.
We can conclude that the work of the resultant external force has nothing to
do with the variation of kinetic energy. The latter is due to an internal force, the
one due to the muscles of the legs of the child.
Similarly, when a car accelerates, the force producing acceleration is the
friction of the road on the tires. The work of this force is null. The kinetic energy
variation is equal to the work of the internal forces due to the engine.

Example E 7.5
Figure 7.20 shows two blocks of masses m 1 and m 2 supported by a horizontal
plane with negligible friction. A spring, in its natural length, is fixed to the left-
hand side of the block on the right. Its elastic constant is k. The two blocks move
with velocities v 1 and v 2 in the same direction and with υ 1 > υ 2. Block 1
reaches block 2 and hits it, compressing the spring.

Fig. 7.20 Two locks on a horizontal plane of negligible friction

We then calculate the maximum spring compression.
We shall solve the problem in two ways, using a trivial reasoning first, then
using the König kinetic energy theorem. Let υ 1′ and υ 2′ be the velocities after
the collision. The linear momentum (one dimension) and energy conservation
give us two equations

From the first equation we express υ 2′ as a function of υ 1′. Then, with the
second equation, we express the spring energy as a function of υ 1′. We denote
by x the compression of the spring

The maximum spring deformation corresponds to the maximum elastic

energy. We obtain the latter by taking the derivative of the last side, putting it
equal to zero and solving for υ 1′:

By substituting this in the expression of the energy just found, we see that
the velocity corresponding to the maximum compression is the center of mass
velocity. Considering the symmetry of the problem, we expect υ 2′ to be equal.
This is immediately found from the above equation, as the reader can verify.
The second approach to solve the problem is much quicker. Moreover, it
immediately shows the reason for both velocities being equal to the center of
mass velocity. We write the energy in the form given by the König theorem.

where υ CM is the center of mass velocity and and

are the velocities relative to the center of mass. We then have

The first two terms in the right-hand side do not vary due to the energy and
linear momentum conservation respectively. The elastic energy is then a
maximum when the two last terms, namely the kinetic energies, and
consequently the velocities, relative to the center of mass are zero.
The angular momentum König theorem.
With reference to the inertial frame of Fig. 7.19, we choose the pole in the
origin O. The angular momentum is

which, using Eqs. (7.64) and (7.65) becomes

The last side contains four terms. The first term is the total angular
momentum in the center of mass about the center of mass as a pole, say .
The second term is the total linear momentum in the CM frame and is null. The
third term is zero for Eq. (7.69). The fourth term is the cross product of the
position vector of the center of mass and the total linear momentum in the
inertial system, say . We can write
(7.73)
We can state that the total angular momentum in the inertial frame is equal to
the sum of two terms. One term is the angular momentum “of the center of
mass”, which is the angular momentum that the center of mass would have if it
were a material point with the total mass of the system. The second term is the
angular momentum relative to the center of mass.

7.17 Elastic Collisions

In this chapter we have already discussed collision experiments. From the
observation of the linear momentum conservation we have deduced the validity
of the action and reaction law. We shall now take the opposite point of view.
Assuming the mechanics laws to be valid, we shall discuss in some detail
collision phenomena. We shall limit the discussion to material points. This is an
idealization. However, we can consider the real bodies as points coinciding with
their centers of mass, in which all their mass is concentrated, as long as in the
collision the kinetic energies of each body do not vary. If we are dealing, for
example, with the collision of two rigid balls, their motions should be
translations, with no rotation.
We specify that when talking of collision of two bodies we do not
necessarily imply that the two bodies come into contact. Considering for
example the Newton experiments on the collisions between two pendulums we
might substitute the balls with two bar magnets, with their north poles facing
each other. If we take the pendulums out of equilibrium and let them go, the two
magnets will approach each other subject to the repulsive force between the
magnets. This force will slow them down till they stop and bunch back, without
touching each other. As another example, consider two ions of the same charge
moving one towards the other. When they are far from one another they feel
practically no force and move with constant velocities. But when they become
close enough, the repulsive electric force will cause both trajectories to deflect.
The two ions will move on curved trajectories, approaching to a minimum
distance and then separating again. When they are far enough apart, the ions will
again move practically with constant velocities. The final velocities are in
general different in magnitude and direction from the initial ones.
In a collision process we can distinguish three phases. In the initial phase the
two bodies are distant and do not interact, namely the force exerted by one on
the other is negligible. The second phase is the phase of the proper collision,
which has a limited duration, say Δt. During this time the two bodies interact.
The interaction forces are internal, an action and reaction pair. The internal
forces are much larger than the external forces, which can consequently be
neglected. As a matter of fact we talk of a collision when this condition is met.
In the third phase the bodies no longer interact but move away from each other.
Notice that during the collision the two bodies may change their shape, the
internal energy can decrease or increase, one or both of them can break into a
number of pieces, or they may join in a single body, etc. In other words, the
bodies in the final state may be different, also in number, from the two initial
ones.
As during the collision only internal forces are present, we can state in
complete generality that the total linear momenta before and after the collision
are equal. Assume for simplicity to have two bodies both in the initial and final
state. We indicate with the subscripts i and f the quantities in the initial and final
states respectively and write
(7.74)
The collision is said to be elastic if each of the bodies after the collision is
the same as before the collision, its internal energy included, and if the total
energies after and before the collision are equal. As the external forces are
negligible and as the internal forces are zero in the initial and final states, the
initial and final total kinetic energies are equal as well. If m 1 and m 2 are the
masses of the two bodies, we have

(7.75)
and Eq. (7.75) can be written as
(7.76)
Equation (7.75) is one relation, Eq. (7.76) are three relations, in total four,
between the initial and final states. We shall now consider a few important cases.
Often one of the particles is at rest. If it is not so, we can always change the
reference frame by choosing a frame moving with one particle (think of an
observer sitting on the particle). The frame in which one particle stands still is
called a laboratory frame . The particle that is still, say particle 2, is called the
target particle . In the laboratory frame Eqs. (7.75) and (7.76) become
(7.77)
(7.78)
The velocity of the target particle after the collision is given by Eq. (7.78),

(7.79)
Consider the case in which the mass of the target is very large, namely
. We see that the final velocity of the target particle is very small. Its
final kinetic energy, namely the energy gained in the collision, is also very small.
In the limit of infinite target mass, the final velocity and kinetic energy of the
target are zero. For example, a standing railcar hit by a ping-pong ball does not
move, neither does a billiard table when a ball hits one of its sides. As a
consequence, the kinetic energies of a light particle hitting a very massive target
particle before and after collision are equal.
We now consider the case of two equal mass particles, which are at rest in
the laboratory frame. The masses being equal, we can eliminate it from
Eqs. (7.77) and (7.78) and write
(7.80)
The first of these equations tells us that the three velocity vectors can be
thought of as the sides of a triangle, as shown in Fig. 7.21. For the second
equation we have a right triangle, the hypotenuse of which is v i1. The final
velocities of two particles of equal masses in the laboratory frame are always at
90° from one another. This can be observed, for example in a billiard game.

Fig. 7.21 Initial and final velocities in a collision of two equal mass particles in the CM frame

Consider now Fig. 7.22, which represents the initial state of the collision
between two spherical bodies. One is initially at rest. The distance between the
line on which the center of the moving body travels and the center of the target is
called impact parameter . It is b in the figure. Clearly, the final state depends on
b. Suppose, for example, that the two bodies are rigid spheres. When they touch,
they interact with a force in the direction of the normal to the contact surface,
which depends on b. This is the direction also of the variation of the momenta.

Fig. 7.22 The impact parameter

The simplest case is when the impact parameter is zero. The collision is then
said to be central . The incoming particle travels on a line passing through the
center of the target. When the particles collide, the action and reaction forces are
directed on that line, and so are consequently the final momenta. After the
collision both particles will travel on this line. The momentum conservation law
Eq. (7.78) becomes a simple relation between magnitudes

(7.81)
The energy conservation equation Eq. (7.77) becomes

(7.82)
We seek two final velocities as functions of the initial one υ 1i . As Eq. (7.82)
can be written as , we can usefully divide it by
Eq. (7.81) obtaining . And finally
 (7.83)
Let us discuss the first equation. If the mass of the incoming particle is
smaller than the mass of the target (m 1 < m 2) its final velocity is negative,
meaning that after the collision it bounces back. On the contrary, if its mass is
larger than the mass of the target (m 1 > m 2), after the collision it continues to
move forward, even if with a smaller velocity. An interesting case is when the
two masses are equal. After the collision the velocities are υ f1 = 0 and υ f2 = υ i1.
The two balls exchange their velocities. The phenomenon is easily seen hitting
two pendulums of equal mass.
Finally, if , then υ f1 = –υ i1 and υ f2 = 0. This is the case of an elastic
collision of a ball, for example a tennis one, against a wall, shown in Fig. 7.23.
Here we suppose the wall to be smooth. In this case the force of the wall on the
ball is normal to the surface. We decompose the quantity of motion of the ball in
components normal and parallel to the wall. The latter is not changed by the
collision. To the normal component we can apply the results we found for the
central collisions. Particle 1 is the ball, particle 2 is the wall, hence .
After the collision the wall is still at rest while the normal component of the ball
velocity has changed its sign.

Fig. 7.23 Elastic collision on a wall

We now analyze the general case of the elastic collision between two
particles. As during the collision the system is isolated, the center of mass
velocity is constant and the CM frame is inertial. Recalling that υ i2 = 0, the CM
velocity in the laboratory frame is

(7.84)
We obtain the velocities of the particles in the CM frame, which we indicate
with an asterisk, by subtracting the CM velocity from their velocities in the
laboratory frame
(7.85)
In the CM frame the total linear momentum is zero both before and after the
collision. This means that the momenta of the two particles are equal and
opposite before the collision and similarly after it. These quantities are called
center of mass momentum before and after the collision respectively. If is the
momentum of particle 1 before the collision, the momentum of particle two is –
. Similarly, after the collision the momenta are, say, and – . We write the
kinetic energy conservation as

and also
(7.86)
In words, in an elastic collision in the CM frame, the magnitude of the linear
momentum of each particle is equal after and before the collision. The only
effect of the collision is to change the common direction of the momenta by an
angle, say, θ, as shown in Fig. 7.24.

Fig. 7.24 Elastic collision in the CM frame

The angle θ is called a scattering angle . It cannot be found only on the basis
of the conservation laws. First of all, it depends on the impact parameter b,
which in the CM frame is the distance between the lines on which the center of
mass of the two bodies travel in the initial state.
The dependence of the scattering angle on the impact parameter, given by
the function θ(b), depends on the structure of the colliding bodies. Suppose, for
example, that one of them, the incoming one in the laboratory frame, is point-
like, while the target body has a structure. We can think of the first as an
electron, the second an atom. We imagine the atom as a spherical cloud of
negative electric charge with the positively charged nucleus at the center. This is
very small and hard. If the impact parameter is larger than the atomic radius, the
electron is not deflected in its motion, namely the scattering angle is θ = 0. If the
impact parameter is smaller than the atomic radius, the electron penetrates in the
charged cloud, is deflected by the electric force and exits in a direction different
from the incident one. The scattering angle is now θ ≠ 0, which is increasing
with a decreasing impact parameter. In practice however it is never very large.
When the impact parameter is smaller than the nuclear radius, the collision is
with the nucleus, and is violent. The scattering angle is large. It can even reach
180°, namely the direction of motion can invert if the collision is central, b = 0,
because the mass of the nucleus is much larger than that of the electron.
This example shows how the measurement of the function θ(b) in a
scattering experiment (a it is called) can be extremely useful to understand the
structure of the objects that, like atoms, are too small to be visible. As a matter
of fact, the example we have just made is quite similar to the experiment
performed in 1911 by Hans Wilhelm Geiger (1882–1945) and Ernest Marsden
(1889–1970) that led Lord Ernest Rutherford (1871–1937) to discover the
atomic nucleus. Geiger and Marsden used energetic α particles (rather than the
electron in the example) sending them on a thin gold sheet and measuring how
many of them were scattered at different angles. They found, in particular, that
sometimes they were deflected backwards. If the atoms were soft clouds of
charges, as in the current model, this could not happen. Rutherford concluded
that a small hard nucleus had to be present inside the atom. In the same way the
internal structure of the atomic nuclei was studied and, in 1967, the presence of
the quarks in protons and neutrons was discovered.

7.18 Inelastic Collisions

As we have stated, linear momentum is always conserved in a collision. This is
not the case for energy. When the final energy is different from the initial one
the collision is said to be inelastic. Rigorously speaking, in real collisions
between every day size objects, at least a small fraction of the mechanical energy
is lost. For example, if we drop a steel ball on a rigid floor, it will bunch back
but will not reach exactly the initial height. If we do the same experiment with a
wax ball we see that it sticks on the floor. The real collisions are never perfectly
elastic, but have a smaller or larger degree of inelasticity. We shall give a
quantitative definition of this concept. Before doing that, let us consider the case
of the completely inelastic collision (the case of the wax ball in the example).
Consider two spherical bodies of masses m 1 and m 2 and initial velocities v
i1and v i2. The collision is completely inelastic if the two bodies stick together,
namely if their velocities after the collision are equal v f1 = v f2. We can indicate
simply with v f the final velocity and write the momentum conservation as
(7.87)
The final velocity is then

(7.88)
which is the same as the center of mass velocity (that does not vary in the
collision) as expected, considering that in the final state there is only one body.
We write down the initial kinetic energy , using the König theorem

where is the kinetic energy in the CM reference. The final kinetic energy
is

We see that in the completely inelastic collision all the kinetic energy
relative to the center of mass is lost in the collision. If we want to look at the
collision in the CM frame we can take over all the conclusions of the last
section, with the exception of equality of the magnitudes of the initial and final
momenta. If the collision is inelastic, the final center of mass momentum is
smaller than the initial one, null if it is completely inelastic. Figure 7.25 shows
the situation. In the completely inelastic collision all the momentum in the CM
reference and all the kinetic energy relative to the center of mass are lost. In the
laboratory frame not all the kinetic energy gets lost, because the velocity of the
center of mass must be the same after and before the collision, due to the
momentum conservation. Consequently, the kinetic energy “of the center of
mass” cannot be lost. In the completely inelastic collision all the energy that can
be lost is lost, but this is not all the energy.

Fig. 7.25 Collisions in the CM frame. a Elastic, b inelastic, c completely inelastic

 In the above example of the wax ball falling on the floor, the ball loses all its
energy. However, in this case the mass of the target is enormous, practically
infinite and consequently the velocity of the center of mass is zero.
An application of a completely inelastic collision is the ballistic pendulum
used to measure the velocities of bullets. Figure 7.26 shows the device, which is
made of a sand bag of mass M suspended with a bar to the pivot O. The bullet P
of mass m and velocity υ to be measured hits the pendulum, penetrates the bag
and sticks. By measuring the resulting oscillation amplitude we determine the
velocity of the bag υ f after the collision. Equation (7.81), considering that v 21
= 0, and that M » m, becomes , or . This gives
the bullet velocity, the two masses being known.

Fig. 7.26 Ballistic pendulum

Looking at Fig. 7.25 we understand immediately that all the intermediate

cases between elastic and completely inelastic collisions are possible. The
parameter characterizing the degree of elasticity is called the coefficient of
restitution and is defined as the ratio between the center of mass momentum
after and before the collision
(7.89)
By definition, the coefficient is a non-negative number. It is equal to one in
the elastic collision. Notice that it can be larger than one. Suppose for example
that the target body contains a spring, which is compressed and blocked by a
nail. In the collision the nail is broken, the spring expands and gives energy to
the colliding bodies. The final center of mass momentum is larger than the initial
ones. As another example, energy can be gained in collisions between two
molecules. Such are the exothermic chemical reactions.
Let us write down the kinetic energy in the CM reference frame

where µ is the reduced mass. Similarly the final kinetic energy is

and finally, for Eq. (7.89)
(7.90)
namely the ratio of the final and initial kinetic energies relative to the center
of mass is equal to the square of the coefficient of restitution.

7.19 Problems
7.1. What is the total momentum P of a system of particles in the CM frame?

7.2. A system of interacting bodies moves in the neighborhood of the earth’s

surface. Neglect air resistance. How does the center of mass move?

7.3. Two railcars move one against the other on a rail. The first one has a mass
of 1000 kg and moves at the speed of 2 m/s. The second one has twice the
mass. After the collision the two cars are at rest. What was the initial
velocity of the second car? Did the kinetic energy change?

7.4. A railcar of 5 t mass and speed 10 m/s is stopped by bumpers in 0.5 s. Find
the impulse and the average value of the force.

7.5. Two pendulums collide elastically. Initially, one of the two, of mass m 2
stands still in the equilibrium position, the other one, of mass m 1 is
abandoned at a certain height above that. After the collision the two
velocities are equal and opposite. (a) What is the ratio of their masses? (b)
What is the ratio between the center of mass velocity and the velocity of
pendulum 1 before the collision?
7.6.
In Problem 7.5, knowing the kinetic energy U Ki (1) of pendulum 1 immediately
before the collision, find: (a) the total kinetic energy in the CM reference, (b) the
kinetic energy U Kf (1) of the first pendulum immediately after the collision.

7.7. In a first approximation, the moon revolves around the center of the earth.
More precisely, earth and moon revolve around their common center of
mass. Knowing that the mass of the earth is about 81 times that of the moon
and that the distance between the two centers is about 60 earth radii, R E ,
calculate the position of the center of mass (in R E units).

7.8. A planet of mass M has a satellite of mass m = M/10. The distance between
their centers is R. (a) Express the revolution period as a function of R and
M. (b) Find the ratio between the (revolution) kinetic energies of the two
bodies.

7.9. We have measured the period of T earth years of a binary system and the
distance between the two stars in R astronomic units. Find the sum of the
two masses in solar mass (M S ) units.

7.10. Two point-like bodies have a completely inelastic collision. The first body
has a mass m 1 = 2 kg and the velocity before collision v 1i = (3, 2, –1)
m/s. The second body has a mass m 2 = 3 kg and the velocity before
collision v 2i = (–2, 2, 4) m/s. (a) Find the velocity V of the composite
body after the collision. (b) Find the total energy and the energy relative to
the center of mass before the collision and compare with the kinetic
energy after the collision.

7.11. A material point of mass m, moving with velocity v 1i collides with a

 second point, of mass 2m, that is standing. We measure the velocity after
the collision of the particle of mass m finding its direction at 45° with the
incident one and its magnitude one half of the initial value. (a) Find the
magnitude and direction of the velocity of a particle of mass 2m. (b) Was
the collision elastic?

7.12. The force F = (3, 4, 0) N is applied on the point P having coordinates (8,
6, 0) m. Find (a) its moment about the origin, (b) the lever arm b of the
force, namely the distance of its application line from the pole. (b) the
component F n of the force perpendicular to the position vector r.

7.13. A ball falls on the floor from 5 m. What are the heights it reaches when
bouncing back the first, the second and the third times if the coefficient of
restitution is 0.8? What are the corresponding energies? Neglect air
resistance.

7.14. An air guide is a rail with a series of small holes through which
compressed air is blown. A sledge can run on the guide practically without
friction. We put two such sledges on the rail. The first one, of mass m
1 = 2 kg is still. On its right side lies a spring of elastic constant
k = 300 N/m and 1 m long, in its natural length. The second sledge, of
mass m 2 = 3 kg is launched towards the first with velocity 5 m/s. It hits
the first sledge putting it and the spring in motion. What is the maximum
deformation Δx of the spring?

7.15. A material system is made of a particle of mass m 1 = 0.1 kg in the point

of coordinates (1, 2, 3) m, a particle of mass m 2 = 0.2 kg at the
coordinates (2, 3, 1) m and a particle of mass m 13 = 0.3 kg at the
coordinates (3, 1, 2) m. Find the coordinates of the center of mass.
7.16.
A body of mass m = 2 kg is shot vertically upwards with initial velocity υ
0 = 10 m/s from a point with coordinates (0, 20, 0) m. The z-axis is vertical
upwards. Find the difference ∆L O of the angular momentum of the body about
the origin between the instant when it is back in the initial position and the initial
instant.

7.17. A particle of mass m is launched with initial velocity v 0 at an angle α with

the horizontal. In the reference frame of Fig. 7.27, neglecting air
resistance, find the time dependence of (a) the moment of the force about
the origin O, (b) the angular momentum L O of the body about the same
pole.

Fig. 7.27 The trajectory of problem 7.17

© Springer International Publishing Switzerland 2016
Alessandro Bettini, A Course in Classical Physics 1—Mechanics, Undergraduate Lecture Notes in Physics,
DOI 10.1007/978-3-319-29257-1_8

8. Rigid Bodies
Alessandro Bettini1
(1) Dipartimento di Fisica e Astronomia, Università di Padova, Padova, Italy

Alessandro Bettini
Email: alessandro.bettini@pd.infn.it

In this chapter, we shall discuss the mechanics of an important class of extended

systems, the rigid bodies. In a perfectly rigid body, the distance between any pair
of its points does not vary for any acting forces or any motion. Clearly, this is an
idealization, but, in practice, the solid objects are rigid in a good approximation.
The motion of the rigid bodies is governed by two differential equations. The
known members are the ns of the external forces and of the external moments.
The solution to these may involve advanced calculus. We shall limit the
discussion to the simplest situations.
In Sect. 8.1, we shall define the rigid body and its motions, and then do the
same for the properties of the systems of applied forces in Sect. 8.2. In Sect. 8.3,
we shall consider the equilibrium conditions.
We then consider the rotations about a fixed axis, which are the simplest
motions. We shall find the expressions of the kinetic energy and the angular
momentum. We shall introduce a new kinematic quantity, the moment of inertia,
and see how the dynamics of the rotations about a fixed axis have some
similarities with the dynamics of the material point. In Sects. 8.9 and 8.10, we
discuss two important examples, the torsion balance and the compound
pendulum.
We shall then move on to more complex motions, those of a rigid body about
a fixed point. We shall first find the expression of the angular momentum about
that point, and of the kinetic energy in Sects. 8.12 and 8.13. We shall see that, in
general, the directions of the angular momentum and the angular velocity are
different. One consequence of that is that the forces develop on the supporting,
as demonstrated in Sect. 8.14. In Sects. 8.15 and 8.16, we study the pure rolling
motion of cylindrical and spherical rigid bodies on a plane.
In Sect. 8.17, we consider the gyroscopes, which are rigid bodies moving
about a fixed point, a top being a good example.
Finally, in Sect. 8.18, we shall study the collisions between rigid bodies.

8.1 Rigid Bodies and Their Movements

The solid bodies are approximately rigid . Upon first approximation, their shape
does not change if we stretch, compress or torque them. Clearly, this is never
rigorously true, because small deformations always take place. However, several
dynamical properties of these bodies can be studied considering them as rigid.
We define them as rigid if the distance between any pair of their points does not
vary.
The space location of a rigid body is called its configuration, which is the set
of the positions of its points. To define the configuration of a generic system of
N points, we need 3 N coordinates, but only six for a rigid body. Let us see why.
We start by defining the position of one point, for example, A in Fig. 8.1,
namely its three coordinates in the reference frame we have chosen. In a certain
instant, it is in A 1 = (x A1, y A1, z A1). We then shall give the coordinates of a
second point, like B, which is in B 1 = (x B1, y B1, z B1). But, wait a moment: we
cannot do that arbitrarily. We can choose only two coordinates, because the
distance between A and B is fixed, independent of the configuration, namely

Fig. 8.1 Two configurations of a rigid body

 With this, we do not yet know the configuration. We need the position of a
third point, like C, which is in C 1 in the considered instant. Of its three
coordinates, we can arbitrarily choose only one, the other two being defined by
the two conditions that the distances C 1 A 1 and C 1 B 1 are fixed. Now, the
positions of all the points are defined, hence the configuration of the body too.
In total, the configuration of a rigid body is defined by six coordinates. We
say that the system has six degrees of freedom .
Consider now two configurations. The transport from one configuration to
another can always be obtained with a translation followed by a rotation around
an axis. This is geometrical and not necessarily fixed. In Fig. 8.1, the translation
brings point A from A 1 to A 2 and the body goes to the dotted configuration. To
bring the other points to their final positions, we need a rotation about an axis
through A 2 (this point is already in position, and should not move any more).
The axis should have the right direction and the rotation angle the right value.
In the above argument, the choice of the point A was arbitrary, but it
determines the translation. If we had chosen, for example, B, the translation
would have been different. Consequently, there are infinite translation-rotation
pairs that produce the given displacement. It can be demonstrated, however, that
once given the initial and final configurations, the rotation direction and the
angle are determined.
Obviously, the same result can be obtained by performing the rotation first
and then the translation, or a sequence of rotation and translation pairs. As a
matter of fact, the motion of the rigid body can be thought of as a continuous
series of infinitesimal roto-translations . The rotation axis, in general, varies
continuously during the motion and we subsequently talk of an instantaneous
rotation axis .
While the choice of the translating point is, as we said, arbitrary, it is in
practice convenient to choose the center of mass, considering its privileged role
in dynamics. We recall here the dynamical equations governing every
mechanical system in an inertial frame that we found in Chap. 7. We choose a
point Ω fixed in the inertial frame as the pole. Let M Ω be the total external
torque about Ω and L Ω the total angular momentum about the same pole, F the
external resultant force and P the total linear momentum. The two dynamical
equations are

(8.1)
 (8.2)
We also recall that the second equation is similarly valid when we choose a
particular point, even if it is moving in the inertial frame, namely the center of
mass of the system

(8.3)
The two vector equations give six independent conditions. For any
mechanical system, these are necessary conditions, but in general, they are not
sufficient. They are, however, sufficient for a rigid body, which has six degrees
of freedom, as many as the conditions. In other words, if we know the external
resultant force and the total external torque (or moment) and the initial
conditions, we can know the motion of the body solving the above differential
equations.
We notice that Eq. (8.1) rules the motion of the center of mass of the body.
Remembering that P = m v CM , where m is the mass of the body and v CM the
velocity of its center of mass, we can write Eq. (8.1) in the equivalent form
(8.4)
where a CM is the center of mass acceleration. The motion of the center of
mass is exactly in the same way as the motion of a material point.
Equation (8.3) allows us to find the motion of the body about its center of
mass. This is general around an axis through the center of mass but of varying
direction and with varying angular velocity. The solution is, in general, quite
complicated. We shall consider the simplest cases here.
We immediately notice an important property of the rigid motions : the work
of the internal forces is always zero. Indeed, the internal forces come in couples
acting on pairs of points in the direction of the line joining the points. The work
done by one of the two for a given displacement of the body is equal to the force
times the projection of the displacement of the point on which it acts on the
direction of the force. The latter is the line joining the two points. The work done
by the couple of forces is then equal to the magnitude of the force times the
difference between the projections of the two displacements on the joining line.
But this difference is the change in the distance between the two points, and this
is zero, if the body is rigid.

8.2 Applied Forces

Suppose that several external forces are acting on a rigid body. As we have just
seen, the motion of the body is determined by their resultant and total torque.
Clearly, there is an infinite number of systems of forces having the same
resultant and the same torque. All these force systems applied to the same rigid
body produce the same motion, when starting from the same initial conditions.
Consequently, from the observation of the motion, we can know the resultant
force (from the center of mass acceleration) and the total torque (from the
angular acceleration), but not the single acting external forces. We define as
equivalent any system of applied forces with the same resultant and torque.
Notice such force systems are equivalent for the motion of a rigid body, but they
do not have the same effects if applied to a non-rigid body. Consider the very
simple example of a couple on the same line. The resultant force and torque are
zero. Acting on a rigid body, they tend to approach or separate the two points,
namely to change their distance. This distance, the body being rigid, cannot vary.
But if the body is a rubber band, the distance varies and both forces do work.
We now show a few simple properties of the force systems that will be
useful in the following.

(1) A force system has resultant F and total torque about the fixed point Ω, M Ω
. We show that the torque about any other fixed pole Ω′ is
(8.5)
With reference to Fig. 8.2, we can easily see that the relation between the
torques about the two poles of the generic force F i is

Fig. 8.2 A force and two different poles

which, summed on all the forces, gives Eq. (8.5)

Corollary 1
The torque of a force system of zero resultant is independent of the pole, namely
M Ω = M Ω ′.

Corollary 2
If two force systems have the same resultant and the same torque about the same
pole, they have the same torque about any pole.

(2) Consider a generic force system of resultant F and torque about Ω, M Ω .

The system is equivalent to a force system of a force F applied to the pole Ω
plus a torque couple M Ω. The demonstration is immediate. The two systems
have the same resultant and the same torque, as the torque of F about the
pole is null.

(3) A system of mutually parallel forces F i applied to different points P i of

position vectors r i is equivalent to their resultant F applied to the point C,
having the position vector

(8.6)

The point C is called the center of the force system . The demonstration of the
theorem is easy. First of all, the two systems obviously have the same resultant.
As for the torque, let us take the origin O as the pole, as in Fig. 8.3. The forces
being parallel, we can call u their common unit vector and write F i = F i u. The
torque about O is
Fig. 8.3 A system of parallel forces

which, by definition of the center of the forces, becomes

which proves the theorem.

The weight forces are a relevant example of parallel forces. Consider a
system of n material points (the argument is also valid for a continuous system)
P i of position vectors r i and masses m i . The weights m i g are parallel forces
applied to the points P i . The position vector of the center of the forces is

(8.7)
We see that the center of the weight forces, called the barycenter , is simply
the center of mass of the system. The motion of a rigid body under the action of
the weights of all its parts can be described as if a single force was acting, its
total weight applied to the center of mass. This property, which we have already
used, substantially simplifies several problems.
Notice, to be precise, that the coincidence between center of mass and center
of the weight forces exists for bodies that are not too large, such that the weights
of all their parts can be considered to be parallel. This is almost always true in
practice.

8.3 Equilibrium of the Rigid Bodies

A configuration of a rigid body is said to be of equilibrium if, leaving the body
at rest in that configuration, it keeps it indefinitely. The necessary and sufficient
condition for the equilibrium, in an inertial frame, is that the external resultant
force and the external moment are zero. Indeed, if the body is in equilibrium, the
acceleration of its center of mass is zero; hence, the resultant force is zero. In
addition, the angular momentum that is initially zero must remain as such.
Hence, the total moment is zero. On the other hand, if the resultant is zero, the
center of mass does not change its velocity, which is initially zero, and, if the
moment is zero, the angular momentum is constant and remains zero, if it is so
initially.
Notice that the two conditions are independent from one another. For
example, a couple of forces have zero resultant and non-zero moment, while a
force applied to the pole has zero moment and nonzero resultant.

Example E 8.1
Consider a rigid body on a horizontal plane under the action of its weight. The
position is of an equilibrium position if the vertical through the center of mass of
the body intersects its support base. Indeed, the external forces are the weights of
its elements and the constraint forces. The former are equivalent to the total
weight applied to the center of mass, the latter are normal to the base and
consequently are a system of parallel forces too. Consequently, they are
equivalent, with their resultant N applied to their center of forces D, as shown in
Fig. 8.4a. The constraint automatically adjusts its reaction in such a way that the
magnitude of N and the center D guarantee the equilibrium, in other words, that
m g and N are a couple with the same line of application. This implies that
N = −m g and that D should be on the vertical from C. This is possible if the foot
of this vertical is between A and B, namely inside the base. The insert in the
figure shows a possible configuration of the constraint forces. They are applied
between A and B. Consequently, their center must be a point of AB.

Fig. 8.4 a An equilibrium position, b a non-equilibrium position

In the configuration of Fig. 8.4b, the equilibrium is not possible. Even if the
constraint normal reaction N is concentrated in the extreme point B of the basis,
this is not enough to produce a couple of zero moments. The body overturns.
During the fall, the normal reaction is less than the weight, because the center of
mass is accelerating downwards. The difference mg − N is equal to the
acceleration of the center of mass times the mass of the body.
The center of the constraint forces can, however, be brought outside the
segment AB, and the equilibrium is also guaranteed in the conditions of
Fig. 8.5b, if part of the constraint forces is directed upwards. We can, for
example, drive a nail in A, as in Fig. 8.5, or attach a hook. If R is the reaction of
the nail, or of the hook, and N the reaction of the plane, the equilibrium is when
the resultant force and moment are zero, namely .

Fig. 8.5 The constraint R guarantees equilibrium

Example E 8.2
The ladder shown in Fig. 8.6 of length l is supported by a vertical wall, at an
angle of α. Suppose the friction on the wall to be negligible, while the coefficient
of static friction on the horizontal plane is µ S . Let us discuss the equilibrium
conditions.

Fig. 8.6 The forces acting on a ladder

In Fig. 8.5, C is the center of mass, and A and B are the footholds. We take
the reference frame with the x-axis horizontal in the plane of the figure, the z-
axis horizontal directed out of the figure and the y-axis vertical upwards. The
external forces are: the weight m g, applied to the center of mass, the constraint
reaction applied in B, which we consider decomposed in a vertical component,
N, and a horizontal component, F t , and finally, the constraint reaction applied
in A, N A that is horizontal (no friction here). At equilibrium, their resultant is
zero:

This equation gives two independent relations, its x and y components, the z
component being identically zero. The two relations are which gives
the unknown N, and , which links the other two unknowns. We now
state that the external moment should be zero too, namely

We have written the signs in this equation taking into account that N A must
be in the positive x direction, because the wall can only push. Consequently, N A
tends to rotate the ladder clockwise and the z component of its moment is
negative. On the other hand, for the above written equation, for the equilibrium
of the horizontal forces, F t must be in the opposite x direction. The z component
of its moment is consequently positive. Solving the two equations for F t and N A
, we immediately have .
The friction force cannot be too large, namely On the other hand,
Consequently, to be in equilibrium, the leaning angle should
not be too large, namely For larger angles, the ladder slides down.
We have assumed the vertical wall to be smooth and its reaction to be
normal. If there is friction, as there always is in practice, there is a vertical
component to the wall reaction too. We would have one more unknown, with the
same number of equations. Under these conditions, the problem is undetermined.
Indeed, there is an infinite number of pairs of the two tangential reactions that
lead to equilibrium. Another example of an undetermined problem is the
problem of finding the constraint reactions on the four wheels of a car, or the
four legs of a table, on a plane. These problems have a solution if more
information is available, such as the nature of the elastic forces of the tires on the
car or the lengths of the legs of the table.

8.4 Rotation About a Fixed Axis

An important and relatively simple class of rigid movements is the class of
movements about a fixed axis . Consider a rigid body of arbitrary shape, as
represented in Fig. 8.7, which can move around the axis a, which is fixed in an
inertial frame. Let u a be the unit vector, arbitrarily chosen in one of the two
directions of a. The configuration of the body is defined by the rotation angle,
which we call ϕ, around the axis a, relative to a fixed plane, which we choose as
the origin of the angles.

Fig. 8.7 A rigid body with a fixed axis

We now choose a point Ω on the axis as the pole of the moments and call M
Ω the total external moment and L Ω the total angular momentum about Ω. The
dynamic equation is

(8.8)
We now take the dot product of the two members with the unitary vector of
the rotation axis u a . We have

(8.9)
In this equation, we have the projections on the a-axis of the external
moment and of the angular momentum, namely
(8.10)
These quantities are called the external moment or the torque about the axis
and the angular momentum about the axis . Both quantities are the components
of a pseudo-vector. They can have both signs. It can be easily shown that they
are independent of the choice of the pole Ω, provided it is on the rotation axis.
We can write Eq. (8.9) as

(8.11)
which expresses the theorem of the angular momentum about an axis . In
other words, the rate of change of the angular momentum about a fixed axis, in
an inertial frame, is equal to the external moment about the same axis.
Let us find the expression of the angular momentum. The angular velocity,
which we call ω, is parallel to the axis. Its magnitude and its sign relative to the
axis can vary in time, but not its direction. We start by considering, for
simplicity, the body consisting of particles of mass m i , in the positions r i
relative to Ω, distance from the axis r′ i and velocity v i , as shown in Fig. 8.7.
The trajectory of the generic particle is a circle normal to the axis of radius r′ i .
Its velocity is tangent to this circle and has the magnitude .
We profit by the fact that the angular momentum about the axis is
independent of the pole on the axis and take it, for each particle, in the center O i
of its orbit. The angular momentum of the particle about this pole is

which, as in figure, has the direction of the axis. What we need is its
component on the axis. Its sign is the same as the sign of the projection on the
axis of the angular velocity, ω a . We have . We now sum over all the
particles and obtain the total angular momentum about the axis

(8.12)
where we have introduced the quantity

(8.13)
which is the moment of inertia of the body about the axis a.
We now consider the body as a continuous distribution of masses. Instead of
point particles of mass m i , we consider infinitesimal volume e dV, in the
position r and having mass dm = ρ(r) dV, where ρ is the density (that can be
different from point to point). Following the same arguments as for the discrete
body, one finds the same result
(8.14)
but now with an integral in place of the sum, namely

(8.15)
In Sect. 8.7, we shall calculate the moments of inertia of several bodies of
simple geometry. We observe here that the moment of inertia depends on the
axis, not only on the body. What matters is how the masses are distributed about
the axis. The equation of motion Eq. (8.11) can be written in equivalent forms.

(8.16)
and also
(8.17)
where

(8.18)
is the angular acceleration.
The last expression looks very similar to the dynamical equation for a point
moving along a straight line. If x is its coordinate, m the mass and F x the
component of the acting force, the equation of motion is, as we know,

Equation (8.17) is the same differential equation with ϕ in place of x (and

consequently, angular velocity and acceleration in place of the linear ones), the
external moment about the axis in place of the force and the moment of inertia in
place of the mass. Consequently, the solutions to Eq. (8.17) are the same as those
for the linear motion of a point.
The simplest case is when the external moment about the axis M a is
constant. Then, the angular acceleration α = M a /I a is constant too and,
analogous to the uniformly accelerated rectilinear motion, we have

(8.19)
where ϕ 0 and ω 0 are the angle and the angular velocity, respectively, at
t = 0.
Example E 8.3
Figure 8.8 shows a rigid disk, say a pulley, that can rotate around a horizontal
axis a passing through it center of mass . A wire, to which a mass m is attached,
is wrapped around the pivot. The radius of the pivot is r. The external moment
about the axis is clearly constant, M a = mgr. Suppose the disk to be initially at
rest and choose the origin of the angles such that ϕ 0 = 0. The motion is then
. Namely, the angle through which the system has turned is
proportional to the square of the time.

Fig. 8.8 A pulley and a weight

8.5 Conservation Angular Momentum About an Axis

We still consider a rigid body with a fixed (in an inertial frame) rotation axis a.
If the external moment about the axis is zero, the angular momentum about the
axis is constant, namely
(8.20)
The external moment about the axis is zero, a part of the trivial case of
absence of forces, in two important cases: (1) the directions of all the external
forces are parallel to the axis, and (2) the application lines of all the external
forces meet the axis. In these cases, for Eq. (8.14), as the moment of inertia is
constant, the angular velocity is constant too
(8.21)
Notice that, for zero external moment about the axis, Eq. (8.20) is also valid
for non-rigid bodies, while Eq. (8.21) is not. A simple experiment follows. A
person sits on a turntable stool holding in his hands two heavy objects with arms
horizontally outstretched. A second person pushes the first in rotation. The first
brings hands and heavy objects near his chest. His angular velocity increases
substantially. The initial moment of inertia of the body was, say, I 1 and was
quite large because heavy masses were far from the axis, while the final one, I 2,
is much smaller because the masses are close to the axis. We can say that the
external moment about the axis is zero, if we neglect frictions, because the
external forces, the weights, are parallel to the axis. The angular momentum is
conserved, and, if ω 1 and ω 2 are the initial and final velocities, we have
and consequently, as I 2 I 1, ω 2 ω 1. This trick is used by
skaters in their figures.

Example E 8.4
As an example, consider the system in Fig. 8.9, which shows an electrical motor
fixed on a support that can rotate about a vertical axis, coinciding with the axis
of the motor. The motor has two parts: the external one (stator) is fixed to the
platform, while the internal one (rotor) is free to rotate and has a flywheel (V in
the figure). The two parts are coaxial rigid bodies with moments of inertia, I 1
being the internal and I 2 the external.

Fig. 8.9 An electrical motor

Suppose that, starting from rest, we switch on the motor for some time and
then switch it off. We neglect frictions. We observe that the two parts rotate at
angular velocities ω 1 and ω 2, respectively.
The initial angular momentum is zero. The final one is zero as well, because
during the action of the motor, the forces are only internal. Hence, again,
or . We can measure the initial and final
angular velocities, repeat the experience with different flywheels, and verify if
the prediction is correct.

8.6 Work and Kinetic Energy

We continue our study of the rigid body rotating about the fixed axis a,
represented in Fig. 8.7. Its generic particle of mass m i , as shown in Fig. 8.10,
moves in a circle. We call O i its center and r′ i the position vector of the particle
from it.

Fig. 8.10 The motion of a particle of a rigid body rotating about an axis

We now calculate the total moment about the axis of the external forces F i
acting on the particle. We start from the moment τ i about any pole on the axis.
Once more, we take the center O i of the trajectory of m i as the pole. The force F
i can be thought of as the sum of three components, one parallel to the axis, one
to r′ i, and one tangent to the trajectory. The contribution of the first is normal to
the axis and has no axial component. The contribution of the second is zero,
because it is parallel to the arm. The only contribution is the third.
We call u t the unit vector tangent to the trajectory with positive direction in
accordance with the direction of increasing angles (which is not necessarily the
direction of motion). Let F ti be the component of the external force on u t . The
component of τ i on the axis is then, in magnitude and sign, .
Consider now the infinitesimal rotation of the body along the angle dϕ, and
calculate the corresponding total work of the forces. As we know, the body
being rigid, the total work of the internal forces is zero. As for the work of the
external forces, we start with the work on one particle. The displacement of the
particle is ds i = r′ i dϕ and the elementary work . To
find the total work of the external forces, we have now only to add up all the
particles. Taking into account that dϕ is the same for all and calling ,
we have
(8.22)
This important relation tells us that the elementary work of the external
forces for an infinitesimal rotation is equal to the external moment about the axis
times the rotation angle. Again, we have found an analogy with the elementary
work of the force on a point F x dx.
The work for a finite rotation, say from ϕ 1 to ϕ 2, is obtained by integration

(8.23)

For the rotations about a fixed axis, the kinetic energy theorem has a simple
expression. Recalling Eq. (8.16), we write

For a finite rotation, the work is equal to the difference of the kinetic
energies

(8.24)

We see that the kinetic energy of a rigid body rotating about a fixed axis is
(once again similar to the material point)

(8.25)

8.7 Calculating Inertia Moments

In this section, we shall calculate the moments of inertia of a few geometrically
simple bodies. We shall consider all of them as homogeneous, namely having
density independent on point. Consequently, their geometric centers coincide
with their centers of mass.
Cylindrical bar. Figure 8.11a represents a bar of mass m and length L. We
assume it to be thin, namely of transverse dimensions much smaller than the
length. We assume the faces to be perpendicular to the geometrical axis. The
shape of the faces is arbitrary. They can be circles, squares, anything. We
calculate its moment of inertia about the axis c normal to the bar through its
center C.

Fig. 8.11 Calculating the moment of inertia of a thin bar about a central transverse axis

We take a coordinate x along the bar originating in its center. We cut the bar
into infinitesimal slices between x e x + dx of mass dm. As the diameter of the
slice is very small, we can consider all the points of the slice at the same distance
from the axis c. The mass of the slice is clearly dm = (m/L)dx. We notice that
there are two slices at the same distance from c, on its two sides. Their
contribution to the moment of inertia is We integrate it
on half of the bar, namely from 0 to L/2, and obtain

(8.26)

Ring. Figure 8.12 represents a thin ring of mass m and radius R. We assume
the diameter of the section to be small compared to R. All the points of a section
can be considered at the same distance R from the center.

Fig. 8.12 Calculating the moment of inertia of a thin ring about the central axis

We calculate the moment of inertia about the axis c normal to the plane of
the ring through its center C. As all the mass sits at the same distance, we
immediately have
(8.27)
Cylindrical surface. The moment of inertia of a cylindrical surface (namely
of negligible thickness) about the geometrical axis is given by Eq. (8.27) as well,
because all the masses in this case are also at the same distance R from the axis.
Homogenous disk. Figure 8.13 represents a disk of radius R and mass m. We
calculate the moment of inertia about the geometric axis c shown in the figure.
We divide the disk into infinitesimal rings of rays between r and r + dr. The area
of a ring is 2πr dr, to be compared with the area πR 2 of the entire disk. The mass
of the ring is then Its contribution to the
moment of inertia is Integrating, we obtain

Fig. 8.13 Calculating the moment of inertia of a disk

(8.28)

Homogenous cylinder. Figure 8.14 shows a homogenous cylinder. It can be

thought of as a pile of disks. Hence, the moment of inertia about the symmetry
axis is given by Eq. (8.28).
Fig. 8.14 Homogenous cylinder and its axis

Homogeneous rectangular parallelepiped. Figure 8.15 represent a

parallelepiped of uniform density ρ, mass m and side lengths a, b and c. We
calculate the moment of inertia about the axes through the center C parallel to
the sides. These we call x, y and z and take as reference axes.

Fig. 8.15 Calculating the moment of inertia of a parallelepiped about three central axes

As a matter of fact, it will be enough to calculate the moment of inertia about

one axis, say z, and this will be analogous for all the axes. We have

Analogous expressions holding for the other axes, we can conclude that

(8.29)
Homogeneous cube. The moment of inertia about one, of the three,
symmetry axes is a particular case of what we have just found. If L is the length
of the side, we have

(8.30)
Homogeneous sphere. We give only the result without developing the
calculation. The moment of inertia about an axis through the center is

(8.31)

8.8 Theorems on the Moments of Inertia

In this section, we shall show two theorems that will help in several cases of
computing moments of inertia. The first one, the Steiner theorem, after Jacob
Steiner (1796–1863), concerns rigid bodies of arbitrary shape, while the second
one is for thin bodies, namely of negligible thickness.
Theorem of the parallel axes , or Steiner theorem. The theorem of the
parallel axes states that the moment of inertia about an arbitrary axis is equal to
sum of the moment of inertia about the parallel axis through the center of mass
and the product of the mass of the system and the square of the distance between
the two axes.
Figure 8.16 represents a rigid body of arbitrary shape. The c axis is through
its center of mass C. The moment of inertia about c is I c , the one relative to the
parallel axis a, at distance h, is I a . Let h be the vector from a to c in a plane
normal to the axes. Consider an arbitrary element of the body, of mass dm and
the plane normal to the axes through the element. In this plane, let r c ′ and r a ′
be the vectors to dm from c and a, respectively. Clearly, .
Fig. 8.16 A rigid body, an axis through the center of mass and a parallel axis

The contribution dI a of the moment of inertia about a is

Taking into account that the last term is dI c and integrating on the body, we
have

The integral in the first term is the mass of the body, while the second term is
the component on the considered plane of the position vector of the center of
mass from the center of mass, and is zero. We have
(8.32)
2
which is the parallel axes theorem . Notice that mh is a positive definite
quantity. For all the axes of a given direction, the moment inertia is minimum for
the axis through the center of mass.

Example E 8.5
Consider the right cylinder in Fig. 8.17, of mass m and radius R, its central axis c
and its generator a.
Fig. 8.17 Moment of inertia of a cylinder about a generator

The moment of inertia relative to c is given by Eq. (8.28). Hence, for the
parallel axes theorem, .

Theorem of the perpendicular axes .

The moment of inertia of a thin body about an axis perpendicular to its plane
through the point O of this plane is equal to the sum of its moments of inertia
about two mutually perpendicular axes passing through O.
Consider the body represented in Fig. 8.18. O is an arbitrary point of the
body that we take as the origin of the coordinates axes, z normal to the plane,
and x and y in the plane. The moment of inertia about z is where r is
the distance of the element dm from z. As , we have,
namely
Fig. 8.18 A thin body and two perpendicular axes in its plane

(8.33)
which is the theorem of the perpendicular axes.

Example E 8.6
Calculate the moment of inertia of a rectangular plate of sides a and b about the
perpendicular axis through its center, as in Fig. 8.19.

Fig. 8.19 A rectangular plate and the considered axes

Equation (8.29), with c = 0, gives and and, for the

theorem of the perpendicular axes

(8.34)
which is the third of Eq. (8.29)

Example E 8.7
Calculate the moment of inertia of a circular plate of radius R about a diameter,
as in Fig. 8.20.

Fig. 8.20 Circular plate and axes

The moment of inertia about the central perpendicular axis I z is given by

Eq. (8.24). On the other hand, obviously, I x = I y and, for the theorem of the
perpendicular axes, thus giving us

(8.35)

Example E 8.8
Find the moment of inertia of a circular disk about an axis tangent to its rim, as
in Fig. 8.21.

Fig. 8.21 A circular disk and an axis tangent to its rim

We just have to apply the theorem of the parallel axes to the result we just
found to have

Moment of inertia of a cylinder about the normal axis through the center.
Consider the (homogeneous) cylinder of radius R and length L represented in
Fig. 8.22.

Fig. 8.22 The cylinder and its longitudinal and perpendicular central axes

We want the moment of inertia about the axis y in the figure. This is the
same situation as we discussed in the previous section, but here, we do not
assume the section of the cylinder to be negligible. We call λ the linear density,
namely the mass per unit length of the cylinder. Consider an infinitesimal slice
between x and x + dx. Its mass is dm = λ dx. We can use Eq. (8.35) to find the
moment of inertia of the slice about the axis through it parallel to y (dotted in the
figure). For the theorem of parallel axes, we have dI y by adding to it x 2 dm,
namely Integrating along the entire length, namely in x from
−L/2 to L/2, we have
(8.36)

8.9 Torsion Balance

The torsion balance is a very sensitive instrument used to measure small
moments and, consequently, small forces. We have already seen how it was used
by Cavendish in Sect. 4.7 and by Eötvös in Sect. 5.8. We shall discuss it in more
detail now. Figure 8.23 shows the scheme of the device. A rigid bar AB is
suspended from a vertical wire through its center. The equilibrium position of
the bar is determined by the configuration of the wire at rest.

Fig. 8.23 Schematics of the torsion balance

When we apply a moment τ, the bar rotates about its center. The rotation
gives origin to an elastic moment τ e in the wire in the opposite direction,
proportional to the rotation angle ϕ
(8.37)
where the minus sign indicates that the elastic moment tends to bring the bar
back into its original position. The elastic constant k depends on the length and
the section of the wire and on its material. We can choose this constant when we
design the balance, depending on the torques we have to measure. For example,
thin quartz wires can be used for sensitivities down to several femtonewton.
The new equilibrium is reached when the rotation angle is such that the
elastic moment is equal to the applied one, τ = τ e . Hence, we can measure τ by
measuring ϕ, and knowing k.
The most accurate measurement of k is done using a dynamical method. We
rotate the bar at an angle ϕ 0 and let it go. It is the motion of a rigid body about a
fixed axis under the action of the external torque τ e . If I is the moment of
inertia, the equation of motion, Eq. (8.16), is

(8.38)
or

(8.39)
with
(8.40)
We recognize the differential equation of the oscillator. Its solution is an
harmonic motion in the angular coordinate ϕ with period
(8.41)
The period can be measured with high accuracy, because we can measure it
over many oscillations and count them. Once we know the period and the
moment of inertia by construction, we know the elastic constant.

8.10 Composite Pendulum

The composite pendulum is a rigid body , of mass m, which can rotate around a
fixed horizontal axis, an axis which should not be through the center of mass. In
Fig. 8.24, O is the trace of the axis, C is the center of mass and ϕ is the angle to
the vertical, taken to be positive counter-clockwise. The distance of the center of
mass from the axis is h.
Fig. 8.24 The composite pendulum

We take the pole for the moments to be the fixed point O. Two forces act on
the pendulum, the weight, which we can think of as being applied to the center
of mass, and the constraint reaction, applied to the axis of rotation. This is a
cylinder of radius r, as shown in the insert of the figure. The constraint reaction
is applied to the point P of its lateral surface. In the presence of friction, the
force has a direction different from the direction of the segment OP and its
moment about O is different from zero. If, however, the friction is negligible, as
we shall assume, the direction of the force is OP and its moment is zero. The
external moment on the system is, under these conditions, the moment of the
weight, which, at the angle ϕ, is −mgh sin ϕ. The equation of motion is

(8.42)
where I is the moment of inertia about the axis. For small angles, we can
approximate the sine with the angle, obtaining

(8.43)
with
(8.44)
Equation (8.43) is equal to that of the simple pendulum. Hence, the motion
of the composite pendulum is a harmonic motion in ϕ. Its period is
 (8.45)
The device is used, in particular, to measure g, knowing from construction
the other quantities in Eq. (8.45).
The period of the composite pendulum is equal to the period of the simple
one of length

(8.46)
which is then called the reduced length of the composite pendulum

8.11 Dumbbell
We have discussed several examples of rotations of rigid bodies around a fixed
axis. However, the axis will move if we do not provide the proper supports to
keep it fixed. In general, the axis is supported by a massive body at rest, on
which the axis rotates through a number of ball bearings to reduce the frictions
as much as possible. The relevant kinematic quantities are the angular velocity
and the angular momentum. Both are vector quantities. The former is by
definition parallel to the axis, the latter not necessarily so. Up to now, we have
used only the component on the axis of the angular momentum. In general, there
are also components perpendicular to the axis, which, in addition, vary in time.
Consequently, an external moment must be present. This is the action of the
supports. We shall now turn our attention to this action.
We shall start from the particularly simple case of the dumbbell in Fig. 8.25.
It is made of two equal spheres of mass m at the extreme ends of a rigid bar of
length 2d of negligible mass.

Fig. 8.25 A dumbbell rotating about a symmetry central axis

 The axis, a in the figure, is vertical through the center of the system O
perpendicular to the bar, namely a symmetry axis of the body. The frictions in
the rotation are negligible. The system is very similar to those we have already
discussed.
We indicate with and the position vectors of the two masses from O
and with u a the unit vector of the axis. The angular velocity has the direction of
the axis ω = ω u a .
The angular momentum about the fixed point O, .
Considering that we can write

(8.47)
where I a is the moment of inertia about a. In this case, the angular
momentum is parallel to the rotation axis. The external moment is zero. Indeed,
the moments of the weights of the two masses are equal and opposite and we are
neglecting the frictions. Under these conditions, angular momentum and angular
velocity are constant in time. If initially the system rotates at a certain angular
velocity, it will continue to do so forever. The ball bearings that keep the axis
must support the total weight, but do not exert any moment.
We now suppose the fixed rotation axis to be still through the center, but not
perpendicular to the bar, at the angle, say π/2 − θ, with it, as in Fig. 8.26. The
angular velocity still has the direction of the axis, ω = ω u a . If r 1 and r 2 are
the position vectors of the two masses, the angular momentum about O is

Fig. 8.26 A dumbbell rotating about a central, non-symmetry, axis

 (8.48)
Looking at the figure, we see that the two terms are equal both in magnitude
and in direction. The latter is the direction perpendicular to the bar in the plane
of the bar and the axis. We call n its unit vector. The plane rotates with angular
velocity ω. The magnitude of the angular momentum is
(8.49)
The moment of inertia relative to the axis is now .
Multiplying Eq. (8.49) by cos θ, and indicating with L a the angular momentum
about the axis, we have
(8.50)
We already knew this result. The axial angular momentum is equal to the
moment of inertia times the angular velocity. However, the axial angular
momentum is only one of the components of the angular momentum vector.
Equation (8.49) gives its magnitude, while its direction is n. We have
(8.51)
Even if the angular velocity is constant, the angular momentum vector is not.
It rotates at constant velocity on the cone of semi-vertex angle θ around the fixed
axis. Consequently, and the external moment is not zero. It is due to
the supporting ball bearings.
Let us look more closely at the situation. We decompose the momentum of
Eq. (7.59) in its components parallel and perpendicular, transverse, to the axis

as shown in Fig. 8.27a for the angular momentum. Its component parallel to
the axis L P is constant, and consequently, M P = 0. L T is constant in magnitude
and rotates around the axis at a constant angular velocity. Its derivative is
Fig. 8.27 a The angular momentum and its components, b the external torque

As seen in Fig. 8.27b, this derivative is also a vector rotating at constant

angular velocity ω in a plane perpendicular to the axis. It is at 90° with L T . This
derivative is just the external moment, which is exerted by the ball bearings.
These act with two forces, F 1 and F 2 in the figure, of constant magnitude and
rotating direction.
The situation is similar, for example, when the rotation axis of the reel of a
car is not exactly the symmetry axis. The periodic stress on the ball bearings
would induce vibrations in the vehicle.

8.12 Angular Momentum About a Fixed Pole

Let us summarize what we have established up to this point on the motions of
rigid bodies. The simplest is the rotation about a fixed axis. In this case, the
configuration of the body is defined by a single (angular) coordinate. Its rate of
change is the angular velocity. The axial angular momentum is the component
on the axis of the angular momentum about any point of the axis (and is
independent of its choice). The axial angular momentum is equal to the angular
velocity times the moment of inertia about the axis. The rate of change of the
axial angular momentum is equal to the component of the moment of the
external forces on the axis. This is the differential equation ruling the dynamics
of the system. In the last section, we saw the consequences of the angular
momentum components perpendicular to the axis on the constraints that
guarantee the stability of the axis. We shall now further study the relation
between angular velocity and angular momentum and the motion of a rigid body
about a fixed point, which is at rest in an inertial reference frame. We call it the
pole , O.
The rigid motion about a fixed pole is still a rotation with an angular velocity
ω, which now may vary both in magnitude and in direction. In other words, in
every instant, the body rotates about an “instantaneous rotation axis ” that passes
through O and has the direction of ω, which continuously changes. We choose
the inertial reference frame as shown in Fig. 8.28, with origin O. We shall also
take O as the pole.
Fig. 8.28 Motion of a rigid body about a fixed point O

We consider a system of material points. A continuous system can be treated

through the same arguments with integrations in place of sums. Let m i be the
mass of the generic point and r i its position vector. Its velocity is
(8.52)
which is obviously the same for all the points. The angular momentum of the
point about O is
(8.53)
We now use Eq. (1.29) to express the double cross product in the last
member, obtaining
(8.54)
The sum of these quantities is the total angular momentum we want to find.
In doing that, we would find a set of quantities analogous to the moment of
inertia about an axis. These are the nine elements of a 3 × 3 matrix. We shall
work on the Cartesian components. We start with the x component of the just
found equation. After simplification, we have

We now add up all the points, obtaining

where, in the last member, we have introduced the quantities

 (8.55)
The first quantity is immediately recognized as the moment of inertia about
the x-axis, while the second and third ones are the products of inertia . The same
argument for the other two components of the angular momentum lead to
analogous expressions. The final result can be expressed in a compact form with
the matrix formalism as

(8.56)

The 3 × 3 matrix is mathematically a tensor and is called the tensor of inertia

. Its elements in the first line are given by Eq. (8.55), and those of the other two
by analogous expressions. We shall not need to know its mathematical
properties. We only notice that the matrix is symmetric, namely the elements in
symmetric positions about the diagonal are equal, I xy = I yx , etc.
The situation looks quite complicated, but we can make it simpler with an
appropriate choice of the directions of the coordinate axes. This is because the
matrix of inertia is square and symmetric. Indeed, mathematics shows that this
type of matrix can always be put in diagonal form by a rotation of the axes. We
still refer to x, y, z as such axes, pose, for simplicity, I x = I xx , I y = I yy , I z = I
zz , and write

(8.57)

Another form of Eq. (8.57) that we shall use is

(8.58)
We have found that the angular momentum of the rigid body about the pole
O is the sum of three vectors. Each of them is directed as one of the axes, with
the magnitude equal to the product of the component of the angular velocity on
that axis and the moment of inertia about that axis. This is true only for the
particular choice of axes that makes the matrix of inertia diagonal. These are
called the principal axes of inertia relative to O. Their position is fixed relative
to the body and they move with it. Consequently, the reference Oxyz is NOT
generally an inertial one. If the pole O is the center of mass, the principal axes
are called central axes of inertia .
We shall now state without demonstration a few important properties of the
principal axes of inertia.
Firstly, as intuition suggests, if the body has symmetry axes relative to O,
these are also the principal axes.
For example, the principal axes of a homogeneous rectangular parallelepiped
relative to its center O are the axes parallel to its sides through O. If two sides of
the parallelepiped are equal, two of its moments of inertia are equal, say I x = I y
. Consider now an axis through O in an arbitrary direction in the plane xy
(namely defined by the two equal moments of inertia). It can be shown that the
moment of inertia about it is I = I x = I y , even if that axis is not a symmetry
axis. We notice that the symmetry of the moments of inertia is larger than the
symmetry of the distribution of the masses.
Consider as a second example a right homogenous cylinder. Its geometric
axis is both a symmetry axis and a central axis of inertia. Any axis in the plane
perpendicular to it through the center is a central axis too. Hence, again, there
are infinite central axes.
There are also cases in which all three moments about the central axes are
equal. Consider, for example, the symmetry axes of a homogenous cube parallel
to its sides. These are clearly central axes of inertia, with equal moments.
However, any other axis through the center is also a central axis of inertia with
the same moment. Again, we see that the symmetry of the moments of inertia is
larger than the symmetry of the distribution of the masses. The former is for a
cube, a spherical symmetry. Obviously, all the axes through the center of a
homogeneous sphere are central axes of inertia.
Consider now again a homogeneous cylinder, with height h and base radius
R. We put the origin of the reference in its center and the z on its axis. The other
two (central) axes are on the normal section. We already know the expression of
the moments of inertia, Eqs. (8.28) and (8.36), which give

We notice that all of them are equal if In these particular cases, all
the axes through the center are central axes of inertia. All the moments of inertia
about them are equal. Again, the symmetry of the moments of inertia is larger
than the symmetry of the masses. In other words, if there are symmetry axes,
these are principal axes of inertia, but a principal axis of inertia may not be a
symmetry axis. Indeed, any rigid body of whatever shape, with no symmetry at
all, like an irregular stone, has three principal axes of inertia about any point at
rest with it, even outside the body.
We state without proof that the principal axes of inertia about a point O and
those about another point O′ are not parallel, in general.
We shall now discuss a few important aspects of Eq. (8.58). First, it tells us
that angular velocity and angular momentum are not, in general, parallel vectors.
However, they are so if the rotation is around a principal axis, namely ω is
parallel to a principal axis. Consequently, the principal axes are also called
permanent rotation axes or spontaneous rotation axes . Consider a rotation about
a fixed point in an inertial frame. Its generic motion is a rotation about an
instantaneous axis through the fixed point, whose direction varies continuously
in time. As a consequence, the angular momentum about the point varies too.
This implies the existence of a non-zero external moment.
Consider now a rigid body with a fixed point which is otherwise free. The
external moment is zero. Consequently, its angular momentum about the fixed
point is constant. If, at a certain instant, the body rotates about a principal axis
with angular velocity ω, it is simply L being constant, ω is constant too,
in magnitude and direction. If, on the contrary, the body rotates around a non-
principal axis, L is constant, but ω is not necessarily so.
The same arguments are valid for the motion of a rigid body without any
constraint, provided the center of mass is chosen as the pole, for Eq. (7.60).

8.13 Kinetic Energy

In this section, we shall discuss the kinetic energy of a rigid body moving about
a fixed point O, which is not necessarily in an inertial frame. Figure 8.29 shows
the situation at a certain instant. The vector ω is the instantaneous angular
velocity, which, in general, varies both in magnitude and direction.

Fig. 8.29 A rigid body moving about a fixed point

 As usual, we think of the body as being made of material points of mass m i .
The kinetic energy of the generic point is , where υ
i is the magnitude of the velocity of the point and r′ I is its distance from the
instantaneous rotation axis. We obtain the kinetic energy of the body adding up
all the points. If I ω is the moment of inertia about the instantaneous rotation
axis, we have

(8.59)
We had already found this expression, Eq. (8.25), in the case of rotation
about a fixed axis.
If the reference is an inertial one and if the body is not subject to external
forces, the kinetic energy is constant in time, but the direction of the angular
velocity relative to the body does, in general, vary. Also in general, both ω and I
ω vary, while the product of the square of the former and the latter are constant.
In practice, Eq. (8.59) is not very useful. Let us find a more useful expression
proceeding in a way similar to what we did in Sect. 8.12 for the angular
momentum. We work in the reference frame of Fig. 8.29, with origin in the fixed
point O. The velocity of the generic point P i at the position vector r i is
(8.60)
The kinetic energy of the point is

We should now add up all the points. In the above expression, we have, for
example, the term Adding up the points, this gives and is
analogous for the other axes. The sums of the terms with the products of two
coordinates give terms propositional to the products of inertia . It is then
convenient to choose the coordinates on the principal axes relative to O, because
the products of inertia are zero. With this choice, we have

(8.61)
which we can write, recalling Eq. (8.58), as

(8.62)
In this expression, the components on the axes no longer appear.
Consequently, it is valid independent of the reference frame. We also notice that,
in absence of external forces, both kinetic energy and angular momentum are
conserved. Consequently, the component of the angular velocity on L O is
constant too.

8.14 Rotation About a Fixed Axis. Forces on the

Supports
We often deal, in practice, with symmetric rigid bodies that rotate about a fixed
axis at high angular velocities. This is the case with the rotating parts of electric
and internal combustion engines, with the reels of cars and bikes, turbines,
helices, etc. There are two vector quantities in the game: the angular velocity ,
which is, by construction, parallel to the axis, and the angular momentum ,
which can have a direction. We have seen an example of this situation in Sect.
8.11. In Sect. 8.12, we have seen that angular velocity and angular momentum
are parallel only if the rotation axis is a principal axis of inertia. If this is not the
case, an external moment must be applied to maintain the rotation axis as fixed.
This is done through the mechanical structure that supports the axis, in general,
through a ball bearing to reduce frictions.
To be concrete, consider the example in Fig. 8.30. The rotation axis is
through the center of mass C of the body, but is not the symmetry axis. The axis
is kept in position by two ball bearings, represented in the figure. The central
axes of inertia are the symmetry axis of the disk, that we shall take as coordinate
z, and any two mutually perpendicular directions in the plane through C normal
to z, which we take as coordinates x and y. The figure is a shot of the movement
when the x axis goes through the plane of the figure.

Fig. 8.30 Rotation of a rigid body about a central non-principal axis

The total force exerted by the supports is just equal to the weight of the body,
both if it rotates and if it is at rest. It will not enter into our arguments.
We shall take as the pole of the moments of the forces and of the angular
momentum the center of mass C, which is also a fixed point in this case. The
symmetry axis of the body forms an angle α with the rotation axis.
Consequently, angular momentum and angular velocity are not parallel. We shall
soon find the direction of the former.
We observe that the angular momentum can be usefully decomposed in one
component parallel and one perpendicular to the axis. The direction of the latter
rotates around the axis with angular velocity ω.
The component of the angular momentum on the axis is, with obvious
meaning of the symbols,
(8.63)
To vary the magnitude of the angular velocity, we must apply a moment
parallel to the axis. This is what engines do, when they accelerate or decelerate.
As a matter of fact, the ball bearings are used to decrease the friction, which,
however, cannot be completely eliminated. The friction moment opposes the
motion. If we abandon the body in rotation, we observe its angular velocity
gradually decreasing due to the moment of the frictions.
We now study the rotation of the components normal to the axis of the
angular momentum and of the moment exerted by the support. We assume the
frictions to be negligible and the moment of the forces to be perpendicular to the
axis. Consequently, both the magnitude of the angular velocity and the axial
component of the angular momentum are constant.
Equation (8.58) becomes, in the case under consideration,
(8.64)
If θ is the angle between the angular momentum and the rotation axis, as
seen in Fig. 8.30, we have

(8.65)
Both the ratio I x /I z and the relation between α and θ depend on the shape of
the body. If the body is a disk, as we saw in Sect. 8.8, I x /I z = 1/2, and
Eq. (8.65) gives If, as is often the case, the angles are small
and we can approximate the tangent with its argument, it is Hence, the
angle between angular momentum and rotation axis is constant in time. In
addition, as we have already observed, the component of the angular momentum
on the axis is also constant and, as a consequence, the magnitude of the angular
momentum is constant. In conclusion, the normal component of the angular
momentum is constant in magnitude and rotates around the axis with angular
velocity ω. The dynamical equation is

(8.66)
where M C is the external moment exerted by the ball bearings. The couple
of forces is shown in the figure. In the considered instant, the plane of the couple
is the plane of the figure. The magnitude of the moment is . And
also, writing Eq. (8.63) as ,

(8.67)
In conclusion, the stress on the support is periodic, with period 2π/ω, and
proportional to the square of the angular velocity. If the latter increases, for
example, by a factor of ten, the moment increases by one hundred.
We now consider a rotation at constant angular velocity around a fixed axis,
which is principal of inertia, but not through the center of mass , as in Fig. 8.31.
In this case, the angular momentum is parallel to the axis and, consequently, is
constant in time. The moment exerted by the ball bearings is zero. The force they
exert, however, must be equal to the centripetal force that is necessary to
maintain the center of mass in its circular motion, namely

Fig. 8.31 Rotation about a principal non-central axis

(8.68)
where r C is the position vector of the center of mass relative to the point O
on the axis (see figure) and u C is its unit vector. The force is exerted by the ball
bearings. Its direction rotates at angular velocity ω, its magnitude is constant,
proportional to the square of the angular velocity.
In conclusion, the ball bearings during the rotation must develop forces that
periodically vary in direction, having resultant F C and total moment M C . The
former is zero if the center of mass is on the axis; the latter is zero if the rotation
axis is a principal axis of inertia. Both are zero if the axis is central of inertia.
Clearly, this is the configuration engineers try to realize, especially if the
velocities are high. Under such conditions the system is said to be dynamically
balanced . Dynamic balance is obtained, for example, for car wheels, by
inserting small lead counterweights where necessary along the tire rim.

8.15 Rolling Motion

The wheels of a bike or of a car moving down the street normally roll without
slipping. If the friction between wheel and street is lower due to rain or snow,
slipping can set in, a situation that should obviously be avoided. The wheel can
be considered a disk. The hub is a central axis of inertia.
Consider, for example, a bike wheel. If we lift the bike and begin rotating the
wheel, which does not touch the ground, it rotates around its axis. If we put it
down and ride it, the motion of the wheel is the sum of a translation, with the
velocity of its center, and of a rotation.
We shall consider rolling without slipping here. If this is true, in every
instant, the contact point of the wheel with the ground is still. Figure 8.32
represents the wheel at a certain instant in full color and in four near instants,
two before and two after that in pale color. As one can see, the extreme of the
spoke near the ground is almost at rest, while all the other points move, to the
degree that they are farther from the contact point.

Fig. 8.32 A bike wheel moving down the road

As a matter of fact, there are two equivalent ways to describe the rolling
motion, shown in Fig. 8.33.
Fig. 8.33 Two possible representations of rolling without slipping

1. a translation with the velocity of the center of mass with a superposed

rotation around the symmetry axis with angular velocity ω

2. a rotation, again with angular velocity ω, around the instantaneous rotation

axis, which is the axis parallel to the symmetry one in contact with the ground
in the considered instant.

The type of motion we are discussing, rolling without slipping, can take
place for cylindrical and spherical shapes. To be concrete, we shall continue
considering a cylinder, of radius R, rolling on a plane, with reference to
Fig. 8.34.

Fig. 8.34 Cylinder rolling on a plane

We take the x axis on the ground in the direction of the motion. If there is no
slipping, the magnitude υ C of the velocity v C of the center of mass and the
angular velocity ω are linked by the relation
 (8.69)
The direction of the angular velocity vector ω is normal to the plane drawn
towards the inside. If R is the position vector of the center C relative to the
contact point A, we can write
(8.70)
We now find the expression of the kinetic energy of the body in both of the
above-mentioned points of view and verify that the result is the same.
In the first point of view, the kinetic energy is the sum of the kinetic energy
“of the center of mass ”, where m is the mass of the cylinder, and that of
the motion relative to the center of mass, where I C is the moment of
inertia relative to the central axis

(8.71)
In the second point of view, the motion is a pure rotation, with the same
angular velocity. The moment of inertia is, for the theorem of parallel axes,
. Hence, the kinetic energy is given by the last member of Eq. (8.71).

8.16 Rolling on an Inclined Plane

An important example of rolling motion is the descent of a rigid sphere on an
inclined plane . Figure 8.35 represents the system. The plane forms the angle θ
with the horizontal, and the radius of the sphere is R. The forces acting on the
sphere are its weight m g, applied to the center of mass C, and the reaction of the
constraint applied to the contact point A. The latter can be decomposed in two
components, one normal, N, and one tangent, F t . Notice that, in the case of the
latter, the friction force must be present in order to prevent slipping. As a matter
of fact, the magnitude of F t cannot be larger than µ S N, where µ S is the
coefficient of static friction. After that, slipping occurs. We shall assume that the
condition of pure rolling is satisfied.
Fig. 8.35 A sphere rolling on an inclined plane without slipping

We shall deal with the problem through three different methods.

Method 1. We consider the moment of the forces, M A , about the
instantaneous rotation axis through the contact point A. If I A is the moment of
inertia relative to this axis, we can write the dynamical equation

(8.72)
The moment of the constraint reaction, which is applied in A, is zero. The
moment of the weight is, in magnitude, and we have .
The velocity of the center of mass is because the motion does not
include slipping, and its acceleration is Substituting in the above
equation, we find

but, for the theorem of parallel axes, , where I C is the moment

of inertia about the central symmetry axis. Consequently,

(8.73)
Method 2. We consider the moments about the horizontal central axis
(through C), M C , and use the equation

(8.74)
The moment of the weight is zero because it is applied to C. The moment of
the normal reaction N is also zero because the force is parallel to the arm. The
magnitude of the tangent reaction of the constraint is F t R. We can write

(8.75)
This equation contains two unknowns, the angular acceleration and F t . A
second equation is given by the theorem of the center of mass motion
(8.76)
Recalling that we find back for a C Eq. (8.73) and for F t
 (8.77)
Method 3. In the process, we are considering that the mechanical energy is
conserved. Indeed, even if a non-conservative force is present, such as the
friction, its work is zero, because the contact point A, where it is applied, does
not move. Suppose that the body starts from rest at the point O of the plane at the
height h (see Fig. 8.35). We call x a coordinate along the inclined plane directed
downwards with the origin in O. The velocity of the center of mass is
We take the zero of the potential energy at h = 0. Initially, the energy of the body
is only potential, and its value is mgh. When the body is at the generic
coordinate x, its potential energy is mg(h − x sin θ). Its kinetic energy is the sum
of the kinetic energies of the center of mass, and of the rotation about the
center of mass, The energy conservation equation is then

or

(8.78)
from which we obtain the center of mass velocity at the generic x

(8.79)
At the end of the inclined plane, the center of mass velocity is then

(8.80)

The ratio that appears in this expression has the physical dimensions of
a length squared. This length, k, is called the radius of gyration of the body
about the central axis, namely
(8.81)
Using this quantity, the final center of mass velocity is

(8.82)
Using energy conservation, we have directly found the center of mass
velocity. Taking its time derivative, we get back Eq. (8.73) written in terms of
the gyration radius.

(8.83)
In the denominators of the expressions, we have found we have the ratio of
two lengths, the gyration radius and the geometric radius of the body. This ratio
depends on the distribution of the masses, as we shall now see in some
examples. Notice that the acceleration and the final velocity from a given height
are smaller for larger values of k/R. Indeed, as we have seen, part of the initial
potential energy becomes kinetic energy of the translation, while part becomes
kinetic energy of the rotation. The ratio between these two energies is

(8.84)
For example, using the expressions for the moments of inertia we found in
Sect. 8.7, we find for an empty cylinder k 2 = R 2 and for a full
homogeneous cylinder k 2 = R 2/2 and and for a full homogenous
sphere k 2 = 2R 2/5 and . In general, the empty bodies descend
slowly, followed by the full ones. This is because, for the same total mass, the
former have larger moments of inertia, and consequently, the fraction of kinetic
energy associated with the rotation is larger. To enhance the effect, we can build
the device shown in Fig. 8.36a, which is a disk with a cylindrical axis. The
radius R of the latter is much smaller than that of the disk. The axis lays on two
parallel inclined rails. The ratio k/R can be made very small, obtaining a quite
slow downward acceleration. Contrastingly, in the configuration of Fig. 8.36b,
the instantaneous axis of rotation is close to the central axis and the larger
fraction of the kinetic energy the energy of the center of mass.
Fig. 8.36 The fraction of kinetic energy in rotation is a large, b small

We shall now analyze when the conditions of pure rolling are satisfied. We
have already found the expression Eq. (8.77) for the tangential force that the
constraint must provide. We now write it in the form

(8.85)
The maximum tangential force the constraint can provide is
The normal reaction should equilibrate the normal component of the weight,
because there is no acceleration in that direction, namely . Hence,
. The no-slipping condition is then

This is a condition on the slope angle θ, namely, simplifying,

(8.86)
Suppose we study the motion of a sphere rolling on an inclined plane and we
gradually increase its slope. When we reach slopes larger than the value of
Eq. (8.86), we observe the contact point slipping on the inclined plane.
Let us briefly go back to what we saw in Sect. 2.12, as to how Galilei
experimentally established that the velocity of a sphere at the end of an inclined
plane is independent on its slope, depending only on the drop h. He did not
know, that part of the kinetic energy is in the rotation motion. However, we can
now show that this conclusion was independent of that. In the configuration of
Fig. 8.35, the velocity of the sphere after a drop h is

(8.87)
to be compared to that of a material point

(8.88)
Consequently, the motion of the center of mass of the sphere is the same for
a material point with 5/7 g in place of g. We notice, in addition, that he very
likely was using a cross-section of the beam similar to Fig. 8.36b for which the
factor in front of g is closer to 1. However, this factor is irrelevant, because the
scope of his experiments was the study of the accelerated motion, not the
measurement of the gravity acceleration.

8.17 Gyroscopes
A gyroscope is a rigid disc with a fixed point. Often, but not always, the fixed
point is the center or mass or, at least, a point on the symmetry axis. The
construction is such that the rotation axis is free to assume any orientation. If the
fixed point is the center of mass, the external moment is zero, and consequently,
the angular momentum is conserved when the disk rotates. The direction of the
axis is unaffected by tilting or rotation of the mounting. For this property,
gyroscopes of this type are useful for measuring or maintaining orientation.
Another example of a gyroscope is the spinning top.
The gyroscope in Fig. 8.37 is the disk in the center. The mounting, called a
Cardan mounting , after Girolamo Cardano (1501–1576), guarantees a complete
freedom to rotate in any direction with the center of mass fixed. The support is
made of three “gimbals ” or rings. The outer gimbal is a half circular, or fully
circular, ring fixed on the support basis. The second gimbal is mounted on the
outer one. It is free to pivot about an axis in its own plane (a in the figure) that is
always perpendicular to the pivot axis of the outer gimbal. The third gimbal is
mounted on the second one and is free to pivot about an axis in its own plane
perpendicular to the first axis (b in the figure). Finally, the axis of the disk is
mounted on the third gimbal, free to pivot around an axis in its plane
perpendicular to the second axis (c in the figure). This is a central axis of the
disk and, as such, a permanent rotation axis.
Fig. 8.37 A gyroscope with Cardan mounting

All the pivots are joined through ball bearings to minimize the frictions.
Notice that in the figure, the three axes are not only mutually perpendicular, but
also that b is vertical, and a and c are horizontal. The latter condition is not
necessary, however. Indeed, if one takes the basis in one’s hand and rotates the
external support, b will not be vertical and a and c will not be horizontal, but
they remain mutually perpendicular.
If we take the disk in our hand, we feel how it can be rotated in any direction
without any effort. Indeed, the disk is in an indifferent equilibrium configuration
and, as we just said, the frictions are negligible. We can give a rapid spinning
motion to the disk by wrapping many turns of wire around its axis and then
drawing it quickly. The rotation can last a long time, because the frictions are
very small.
The Cardan mounting is not necessary for gyroscopes having the fixed point
on the symmetry axis, but not in the center of mass . The most well-known
example is the spinning top . The top, the motion of which we shall study soon,
is a body of approximately conic shape supported on a horizontal plane spinning
about its axis. If the friction between the tip of the top and the plane is enough,
the support point remains (approximately) at rest and the top is a gyroscope.
We anticipate that the motions of the gyroscopes, when we apply an external
action onto them, look quite strange. Gyroscopes do not behave as our intuition
would suggest to us. To understand them, we should fix our attention on the fact
that the characteristic kinematic quantities of a rigid body in rotation are the
angular velocity and the angular momentum. Both are vectors. Pay attention to
the fact that, to modify the angular momentum, we need to apply a torque, or a
couple of forces, rather than one force. The induced change of angular
momentum (another vector) has the direction of the applied torque, which is
perpendicular to the force. If we apply a torque parallel to the angular
momentum, we modify its magnitude and not its direction, whereas if we apply
the torque perpendicular to the angular momentum, we modify its direction and
not its magnitude.
Let us now discuss a few simple examples.
In the first case, represented in Fig. 8.37, the fixed point is the center of mass
and the axis is the symmetry axis, which is an axis of permanent rotation . The
angular momentum L C and the angular velocity ω are parallel and
(8.89)
If the external moment is zero, the angular momentum is constant and the
angular velocity as well:
(8.90)
We observe that, if we take the support in one hand and change its
orientation, the spinning direction relative to the ground, which is an inertial
frame, does not change. The support gimbals change direction about the
invariable direction of the rotation axis (c, in this case).
Torpedoes, for example, make use of this property. One mounts a gyroscope
inside the torpedo and guarantees a continuous spinning with a motor. If the
torpedo deviates from the straight trajectory, due to a submarine current or some
other factor, the direction of the spinning axis changes relative to the torpedo. A
servomechanism then enters into action to modify the route acing on the helm.
The second case is the same gyroscope in the presence of an applied torque.
We can, for example, suspend a mass m to a point A of the c axis at a certain
distance from the center, as in Fig. 8.38. The angular velocity and the angular
momentum are still parallel and Eq. (8.89) is still valid. But now, the angular
momentum varies, according to the equation
Fig. 8.38 A gyroscope with an external torque

(8.91)
Our intuition suggests that we would see the point A lower under the action
of the weight. But this is not what we observe. Point A does not lower, but, on
the contrary, slowly moves in a horizontal circle. This motion of the rotation axis
c is called precession .
To understand this, as we have already stated, we must think about the
direction of the applied moment, not of the force. Let us start considering the
instant in which the axis of the gyroscope is still at rest and we apply the weight.
The vertical weight force exerts on the gyroscope a moment, or torque, the
direction of which is horizontal and perpendicular to the c axis and,
consequently, perpendicular to the angular momentum. In the time interval dt,
the variation of angular momentum is, for Eq. (8.91),

which has a direction perpendicular to L C in the horizontal plane, as shown

in Fig. 8.39. Consequently, L C varies in direction and not in magnitude. The c
axis, which has the direction of the angular velocity and, consequently, of the
angular momentum, also rotates in accord. The torque, which is perpendicular to
the axis, rotates as well, because the weight is applied to the point A of the axes.
The torque always remains perpendicular to the angular momentum.
Consequently, the just described situation is such in every instant, not only in the
initial one.

Fig. 8.39 The precession of the gyroscope

The angular velocity of the precession motion, which we call Ω, is directed

vertically, in our case, upwards. To have it downwards, namely to have the
gyroscope preceding in the opposite direction, we just have to attach the weight
at the other extreme of the c axis, or have the disk spinning in the opposite
direction and the weight still in A.
The motion of the gyroscope is a rotation with angular velocity equal to the
(vector) sum of ω and Ω. Rigorously speaking then, the rotation axis is not
exactly the symmetry axis. However, in practice, the spinning velocity is always
much larger that of the precession. The rotation is in a good approximation about
the symmetry axis and we can consider Eq. (8.89) valid. With reference to
Fig. 8.39b, we can write . If l is the distance of point A from the
center C, the magnitude of the moment is M C = mgl and the precession velocity
is

The third case we will consider is the following. The suspension point is not
the center of mass, but is on the symmetry axis anyway. The external moment is
not zero; its direction is always perpendicular to the rotation axis. Figure 8.40a
shows how such conditions can be realized. The axis of the disk ends with a
small sphere. The sphere lays on a concave support on top of a column, allowing
the axis to spin and to change its direction freely. We now give to the gyroscope
a rapid spin about its axis, keeping it horizontal with our hand. When we
abandon the axis, it does not fall downwards, but rotates in a precession motion
in the horizontal plane. The analysis of the motion is identical to the preceding
case, with the only difference being that the weight is now the weight of the
gyroscope itself.

Fig. 8.40 a A gyroscope with suspension point on the symmetry axis, but not in the center. b precession
and nutation

Notice how the behavior of the system is completely different when the disk
is spinning from when it is not. In the latter case, if we take the extreme of the
axis in our hand and then abandon it, the axis falls, rotating in the vertical plane.
If we do the same with the disk spinning, the axis rotates in the horizontal plane.
The acting torque, the moment of the weight, is equal in both cases, and so is the
change of the angular momentum in any time interval dt. This, however, in the
case of the spinning disk, adds to a pre-existent angular momentum, modifying
its direction, while, contrastingly, in the case of no spinning, the change is solely
to the angular momentum, which consequently has the direction of the torque.
To be sure, the angular velocity is the sum of ω and Ω, and, consequently, is
not exactly parallel to a principal, permanent rotation axis. The just-made
description is valid only in a first approximation. Let us look more carefully into
the issue.
As a matter of fact, the gyroscope’s strange immunity to its own weight is
not completely true. If we set the gyroscope spinning with the point A in our
hand, when we abandon it, it initially falls down vertically a bit. However, as
soon as the precession starts, the extreme A rises again, reaching the horizontal
plane, as shown in Fig. 8.40b. This is not all, however. The axis does not remain
horizontal. The precession has slowed down somewhat due to the rise of the axis
and is no longer fast enough to neutralize the weight. The extreme falls again to
the height of the first descent, the precession velocity increases and the extreme
rises again, and so on; the motion continues with a series of up and down
oscillations, which are ideally all equal. The motion is similar to the motion of
the head of somebody that nods, and is called nutation , which means ‘nodding’
in Latin.
We shall not further analyze this motion, which is quite complex. Rather, we
shall make a few observations. When a gyroscope spins about its symmetry axis
with angular velocity ω and precedes at the same time with angular velocity Ω,
its total angular velocity is not parallel to the symmetry axis. Consequently,
angular velocity and angular momentum are not exactly parallel. The effects are
generally small because ω Ω, but it is the basis of the nutation phenomena.
Consider a gyroscope rotating about an axis a bit different from a symmetry
axis in absence of external torque. In that case, the angular momentum is
constant and the angular velocity rotates around it, describing a cone. In our
case, however, an external torque exists. It is the moment of the weight that is
directed horizontally, perpendicular to the axis. The vertical component of the
angular momentum is constant, because the external torque is horizontal. The
magnitude of the angular momentum is constant too, because the torque is
perpendicular to its direction. Consequently, the angular momentum vector
rotates uniformly in the horizontal plane. This is the precession. The angular
velocity contemporarily describes a cone around the angular momentum. The
extreme of the axis describe a cycloid curve, as shown in Fig. 8.40b. This is the
nutation.
We can look at the phenomenon from another slightly different point of
view. When we abandon the axis horizontal of the spinning gyroscope, the
precession starts. This adds to the angular momentum the vector quantity I Ω Ω
where I Ω is the moment of inertia about the vertical axis through O. The external
torque being horizontal, the vertical component of the angular momentum is
conserved. Consequently, the spinning axis must fall a bit, or even better, rotate
downwards, in such a way that I C ω has a vertical component equal and
opposite to I Ω Ω (see Fig. 8.41). An oscillation starts in which I Ω Ω increases
and decreases alternatively, and so does the angle with the horizontal of I C ω.
Fig. 8.41 The vectors playing roles in the nutation

As a last example of precession, we consider the top . The top is a rigid body
of approximately conical shape, ending with a tip. Initially, we give the top a
rapid spin about its symmetry axis with angular velocity ω. The tip O lays on a
horizontal floor, as in Fig. 8.42. We assume that the friction is enough to keep
point O at rest. We observe that, beyond spinning, the top also has a precession
motion, with angular velocity that we shall call Ω.

Fig. 8.42 A top and its precession

Let r be the position vector of the center of mass C relative to O and m g the
weight of the top, applied, as usual, to the center of mass. The constraint forces
are applied in O, which we choose as the pole of the moments. Consequently,
the constraint forces do not contribute to the external moment. We have

(8.92)
The moment M O is horizontal, perpendicular to the spin axis. Consequently,
the direction, but not the magnitude of the angular momentum, varies. More
precisely, the angular momentum rotates with angular velocity Ω. Hence, for the
Poisson formula

(8.93)
 Considering that Ω ω, we can assume that and, for the above
equations, that Now, both Ω and g are vertical, something we
can express as . Substituting in the last expression, we have

or

and finally, for the magnitudes

(8.94)
This corresponds to the period of the precession

(8.95)
Let us look at the orders of magnitude. We approximate the top with a
homogeneous cylinder of radius R = 2 cm. Let r = 3 cm be the distance from the
center of mass to the tip. Suppose that the spinning angular velocity is
ω = 120 s−1 (that is, about 20 turns per second). Let us calculate the precession
period. The moment of inertia is I = mR 2/2. We have

The corresponding precession angular velocity is which is

qiute small in comparison to the spinning velocity.

8.18 Collisions Between Material Systems

In Chap. 7, we studied the collision phenomena. In that discussion, we
considered, for each colliding body, only the motion of its center of mass. We
did not consider there the motion of each body relative to its center of mass. For
example, when a football player kicks the ball, the momentum of the ball varies.
However, the player may wish to give an angular momentum to the ball as well,
to make it follow a curved trajectory. In general, in a collision, both the linear
and the angular momentum of each body vary. We shall now discuss this aspect
of collisions. We shall limit the discussion to the cases in which one of the
bodies can be considered as point-like, while the other is extended and rigid.
There are two different possible situations: the target body may be free or
may be subject to constraints. In the first case, but not in the second, we can
neglect the external forces during the collision, as in the case of point-like
objects. Consequently, the total linear momentum and the angular momentum
are conserved. The latter does not have any role in collisions between point-like
objects. It adds nothing to the linear momentum conservation. The angular
momentum conservation has, contrastingly, observable consequences in
collisions between extended objects. Even better, we observe that the angular
momentum conservation in an isolated system is a consequence of a particular
aspect of the action-reaction law, namely that action and reaction have the same
application line. This aspect cannot be experimentally controlled with collisions
between point-like bodies. We need to look at the angular momentum
conservation in collisions between extended bodies.
If constraints are present, there are external forces acting during the collision.
Consider, for example, a ball hitting a bar pivoted on an axis initially at rest. The
motion after the collision is bound to be a rotation about the axis. The forces
exerted by the constraints must be such as to equilibrate some of the components
of the impulsive forces that develop during the collision. Consequently, their
intensity is great, and cannot be neglected. The collision with a constrained body
is not, consequently, a process in an isolated system. Linear and angular
momentum, in general, are not conserved.
We shall limit the discussion to two examples.

Example E 8.9
A homogeneous disk of mass M and radius R lays on a horizontal plane. It is
initially at rest. A bullet of mass m and velocity v i1 hits the disk on its rim
tangentially, as in Fig. 8.43, and sticks. Find the motion of the system after the
collision.

Fig. 8.43 A bullet hitting a disk

There are no constraints. The system can be considered as isolated. Linear

and angular momentum are conserved. As in Fig. 8.43, we take a reference
frame at rest on the support plane, with the x axis in the direction of the initial
velocity of the bullet and through the center A of the disk. The y is perpendicular
to x in the support plane and the z axis is perpendicular to both. The angular
momentum of the system, which we take to be about the center of mass C, is in
the z direction. After the collision, we have one lone body moving at the velocity
of the center of mass v C . The angular momentum conservation gives the two
equations

The second equation tells us that the y component of the velocity of the
center of mass is zero, which is obvious. The first equation gives the velocity of
the center of mass

(8.96)
We choose the center of mass as being the pole of the angular momentums.
Its y coordinate does not vary during the motion and is given, by definition, by

(8.97)
The initial angular momentum is that of the bullet, because the disk is not
moving. Its direction is opposite to the z axis and its magnitude is

(8.98)
In the final state, the system disk plus the bullet rotates with angular velocity
ω, which we must determine. Its angular momentum about the center of mass is
. The angular momentum conservation then gives

(8.99)
which gives us ω once we know the moment of inertia I C . This is the sum of
the moments of the inertia of the bullet, I b and of the disk, I d . The former is
, and the latter can be found with the theorem of

parallel axes, giving In conclusion,

and finally, from Eq. (8.99),

Example E 8.10
Suppose now the disk in the previous example is constrained by a vertical axis
through its center A, about which it can rotate freely. In this case, the system is
not isolated. Neither the linear nor the angular momentum are necessarily
conserved. In this case, the external forces are the constraint ones, which are
applied to the point A. Their moment about A is zero. Consequently, the angular
momentum about A is conserved, say L A,f = L A,i .
In the final state, the system is a rigid body, disk plus bullet, rotating with the
angular velocity ω, still to be found. We can write .
Here, I A is the moment of inertia of the system about A, namely
Hence, the final angular velocity is

Notice the difference from the previous example.

8.19 The Virtual Works Principle

In this section, we shall discuss a method that often turns out to be useful for
establishing the equilibrium conditions for mechanical systems, rigid or not. The
method is based on the so-called virtual works principle .
A virtual displacement of a mechanical system is defined as any
infinitesimal displacement compatible with the constraints to which the system is
subject. For example, for a rigid body pivoted on a fixed axis, a rotation of an
infinitesimal angle about the axis, or for a carriage on a rail, an infinitesimal
translation in the direction of the rail, etc.
The work dW i that would be done for that displacement by the ith force
acting on the system is called the virtual work of that force. The virtual works
principle states that a mechanical system is in equilibrium in a given
configuration if the sum of the virtual works done by the forces acting on the
system for any virtual displacement from that configuration is zero.

(8.100)
We shall discuss a few examples.
Example E 8.11
Figure 8.44 shows a rigid bar, pivoted in O. Two forces, F 1 and F 2, are applied
to its extremes A 1 and A 2 perpendicular to the bar. The distances of the
extremes from O are b 1 and b 2, respectively.

Fig. 8.44 a Finding the equilibrium condition for a lever, b the same with two weights

The virtual displacements d s 1 and d s 2 of the extremes are infinitesimal

arcs of the circles of the center in O of radiuses A 1 and A 2. Indeed, the only
degree of freedom is the rotation angle ϕ about the axis.
Let us start with F 1. Its virtual work for the displacement d s 1 is
, where τ 1,z is the moment of the force about the z
rotation axis, with positive direction pointing outside the page of the drawing.
Once d s 1 is chosen, the displacement d s 2 of A 2 is fixed. The virtual work of
the force F 2 is . According to the virtual work
principle, the configuration is of equilibrium if We
recognize the known result that to have equilibrium, the total moment about the
rotation axis must be zero.
As a matter of fact, the virtual works principle is a consequence of the
energy conservation law. To see that, let us look at the present example from a
slightly different point of view, as in Fig. 8.44b. The force F 1 is the weight of a
block placed in A 1. We want to establish the equilibrium by placing another
block of weight F 2 in A 2. Which is the value of F 2 requested for that? We
imagine that when we place the second block, A 2 moves d s 2 down and A 1
moves d s 1 up. The corresponding variation of potential energy is

Being the total energy conserved, the variation of potential energy might be
compensated by an opposite variation of kinetic energy. However, in the virtual
change we are considering, the system is at rest both before and after the
displacement and the kinetic energy is always zero. We conclude that the
potential energy cannot vary, This is what the virtual works
principle states.

Example E 8.12
Figure 8.45a shows two blocks of masses m 1 and m 2 resting on two inclined
planes tilted to the horizontal at the angles θ 1 and θ 2 and connected by a rope.
Frictions are negligible. We want to know which is the ratio of the two masses to
have equilibrium.

Fig. 8.45 a Two blocks in equilibrium on different slopes, b the basis of the Stevin argument

We think to move block 1 of ds upwards on the plane. The work done by the
weight is dW 1 = − m 1 g(sin θ 1)ds. At the same time, block 2 moves on its plane
of the same ds downwards, because we want the rope to remain invariant. The
work of its weight is dW 2 = +m 2 g(sin θ 2)ds. The constraint forces are normal
to the displacements and do no work. The virtual works principles then requires
for equilibrium that dW 1 + dW 2 = 0. The ratio of the masses must be

Historically, as we have already mentioned, Galilei established the law of the

free fall with experiments on inclined planes of different slopes. He then
extended the validity of the law to the vertical motion with the exact above
argument, with one of the angles being equal to 90°.
His contemporary Simon Stevin (1548–1620) demonstrated the rule as
drawn in Fig. 8.45b. The chain is in equilibrium. In the case of this particular
right triangle, five rings balance three rings. In those times, the trigonometric
functions were still not known.
Example E 8.13
Figure 8.46 shows a rigid bar of mass m pivoted at its lowest point and held by a
rope. The bar holds, in turn, a block of mass M. Find the tension T of the rope.
This problem also has one degree of freedom, the rotation angle θ. To respect the
constraint, we can only diminish it, say by −dθ. The acting forces are three: the
weight of the bock Mg, the weight of the bar mg and the tension of the rope T.

Fig. 8.46 Find the equilibrium configuration

The displacement corresponding to −dθ of the block is d(b cos θ) = b sin θ

dθ upwards. The work of its weight is dW 1 = −Mgb sin θ dθ. The weight of the
bar is applied to its center of mass. Its work is dW 2 = −mg(b/2) sin θ dθ. The
application point of the rope has the displacement d(a sin θ) = a cos θ dθ and the
work of the tension is dW 3 = a cos θ dθ.
Imposing dW 1 + dW 2 + dW 3 = 0, we have .

8.20 Problems
8.1. Fig. 8.47 represents a rigid bar b, and v 1 and v 2 are the velocities of its
extremes. Is it possible?

Fig. 8.47 Problem 8.1

8.2. A rigid bar of length L = 8 m and mass m = 100 kg lays on two supports at
distances L 1 = 2m and L 2 = 1 m from the two extremes. Find the forces F 1
and F 2 on the two supports.

8.3. On which of the following elements does the moment of inertia of a body
depend? The mass of the body, the shape of the body, the angular velocity
of the body, the position of the axis relative to the body, or the external
resultant force?

8.4. A rigid body rotates about a fixed axis. How much does its kinetic energy
vary if the angular velocity doubles?

8.5. Two material points of masses m 1 and m 2 are linked by a rigid bar of
length L and negligible mass. Find the moment of inertia about a
perpendicular axis through the center.

8.6. The density ρ(r) of a cylinder of length L and radius R varies linearly with
the distance r from the axis from the value ρ 1 on the axis to the value ρ
2 = 3ρ 1 on the lateral surface. Find the moment of inertia about the axis.

8.7. Figure 8.48 represents a thin annular sheet of radii R 1 and R 2. Find the
moment of inertia about the a axis.
 Fig. 8.48 Problem 8.7

8.8. A rigid cylinder rolls on an inclined plane without slipping. Its density is
not necessarily uniform. Can the kinetic energy relative to the center of
mass be larger than that of the center of mass?

8.9. Two material points of masses m 1 and m 2 are fixed to the extremes of a
rigid bar of length L and negligible mass. We want to bring the bar into
rotation with angular velocity ω about an axis perpendicular to the bar
through one of its points. How should we choose this point so as to have the
minimum kinetic energy for the given angular velocity?

8.10. Under which conditions do the angular velocity ω and the angular
momentum L of a rigid body have the same direction?

8.11. In which cases is equation valid for a rigid body?

8.12. In which cases is the kinetic rotation energy of a rigid body given by
?
8.13. A homogeneous sphere of radius R and mass m rotates about an axis
through its center C with angular velocity ω. Find the angular momentum
about C. Does the angular momentum depend on the pole?

8.14. A homogeneous sphere of radius R and mass m rotates without slipping on

a horizontal plane. Its axis advances with velocity v. The points X, Y, and
Z in Fig. 8.49 are in the vertical plane containing the center of the cylinder
shown in the figure. Their heights are R/2 lower than C, equal to C and
R/2 higher than C, respectively. Find the angular momentum of the
cylinder about each of these points.

Fig. 8.49 Problem 8.14

8.15. A bar of length l = 3 m of mass m = 50 kg is initially vertical at rest with

one extreme O laying on the ground. With O maintained at rest, the bar
falls to the ground. Find the angular momentum about O and the velocity
of the other extreme at the instant in which the bar hits the ground.

8.16. A rigid homogeneous sphere is set free on a plane inclined at 40° with the
horizontal. At which values of the friction coefficient will the sphere roll
without slipping?

8.17. A yo-yo (Fig. 8.50), which we consider to be a homogenous cylinder, of

mass m = 100 g, hangs from a wire wrapped around its axis. The axis is
horizontal. Assume the radius of the wrapping to be equal to the radius of
 the cylinder. The yo-yo is released at rest. (a) What is time t that it takes to
drop to h = 50 cm? (b) Which is the tension T of the wire during the
descent?

Fig. 8.50 The yo-yo of Problem 8.17

8.18. A homogeneous disk of radius R in a vertical plane can rotate about its
geometrical axis (Fig. 8.51). The friction on the axis is not negligible but
exerts a torque M a about the axis, independent of the angular velocity. A
particle of mass m sticks to the rim of the cylinder at the level of the axis.
The system is released at rest. (a) Which is the minimum value of m for
the cylinder to start rotation? (b) Which is the value of m at which it
rotates a quarter of a turn and stops?

Fig. 8.51 Problem 8.18

8.19. A homogeneous disk of radius R and mass m rotates about its geometric
axis with angular velocity ω. The frictions of the axis slow it down until it
 comes to rest. How much work have they done?

8.20. A block of mass M on an inclined plane at the angle θ is held in

equilibrium by a set of pulleys, as shown in Fig. 8.52. Using the virtual
works principle, find the value of the mass m of the counterweight needed
to insure the equilibrium.

Fig. 8.52 The system of Exercise 8.20

8.21. The system in Fig. 8.53 is made of two identical dumbbells. Each of them
consists of two small spheres, each of mass m = 0.3 kg, separated by a bar
of negligible mass of length l = 1 m. The dumbbells move on a horizontal
plane with negligible friction with equal and opposite velocities υ = 1 m/s.
Two spheres, as shown in the figure, collide elastically. (a) Describe the
motion after the collision. Find the angular velocities (magnitude and
direction). (b) How long does the rotation last? (c) Then what happens?

Fig. 8.53 The system of Problem 8.21

Solutions
1.1. (a) No. (b) Yes, if Δ V has the same direction and verse as V .

1.2. Δ V = −2 V , ΔV = 0 and | Δ V | = 2 V .

1.3. (a) Δ v = (4,0,3), (b) | Δ v | = 5, (c) Δυ = 3.9.

1.4. (a) < υ > = υ , (b) < v > = 0.

1.5. (a) , a ( t ) = 6 j m/s 2 ; (b) υ ( t = 2 s) = 12.2 m/s.

1.6. R = υ 2 / a .

1.7. (a) v ( t ) = − i Aω sin ωt + j Aω cos ωt ; a ( t ) = − i Aω 2 sin ωt − j Aω

2 cos ωt ; υ ( t ) = Aω , a ( t ) = Aω 2 ; (b) , velocity is
perpendicular to the position vector (c) , acceleration is
parallel and opposite to the position vector; (d) x ( t ) 2 + y ( t ) 2 = A 2
= constant. The trajectory is a circle with center in the origin and radius A.
The motion is uniform in anticlockwise direction; (e) direction changes to
clockwise.

1.8. (a) The first step in solving this type of problems is drawing the vectors
they contain, as in Fig. 1 . v 1 is the cyclist velocity, v 2 is the wind velocity
relative to ground, v 2 − v 1 is the wind velocity as felt by the cyclist.
 Vectors and angle drawn with continuous lines are known. With the sine
law we get and β = 139.5°. Consequently, the wind
blows from 40.5° from North to East. (b) The new apparent direction of the
wind (the velocity of the cyclist is − v 1 ) is and its
apparent direction is 35.6° from South to East.

Fig. 1 Velocities and relative velocity of problem 1.8

1.9. She will cross at 3 miles towards stern in 18′.

1.10. (a) The rotation axis in the plane xz is at 27° to the x -axis (b) 20 rad. (c)
the magnitude of ω grows proportionally to the square of time, its
direction is constant.

1.11. (a) α = 78.5°, (b) t = 11.5 s (the smaller solution must be chosen); (c) s =
1.15 km.

1.12. The motion is the sum of a translation at the velocity v and a rotation
about the wheel axis. Hence, υ A = ( υ , υ , 0); υ B = (2 υ , 0, 0); υ C = ( υ , −
υ , 0).
2.3.

2.4. 3.94 N

2.5. The equation of motion of the body of mass M is − T + Mg = Ma and for

the body of mass m is − T + mg = ma . Hence .

2.6. The kinetic energy of the hammer is (1/2) mυ 2 when it hits and 0 at the end.
The change of kinetic energy is equal to the work done on the nail, which in
turn is equal to the mean force times the displacement s . The mean force is
then .

2.7. F , for the action reaction law.

2.8. − F on each hand, independently of the accelerations of the spheres being

different, as a consequence of the action-reaction law

2.9. Statement 1 is, in general, false. Statement 2 is true for the rings on the
guides b and c , for energy conservation. For the same reason the statement
is false for the guide a , because that ring cannot reach B .

2.10.

2.11. The initial kinetic energy transforms into elastic energy of the pole and
 then in potential gravitational energy of the athlete. .
(NB In practice, the athlete raises even more doing work with his arms.)

2.12. The two ropes have equal tensions. They break at the same time.

2.13. The lighter sphere rises four times more (energy conservation).

2.14. (a) F ( t = 0) = 0.09 N; (b) F max = 10 N.

2.15. If the rotation plane is horizontal, the wire is on a cone at an angle, say θ ,
with the horizon, in order to balance the weight mg with the vertical
component of the tension, T sin θ . Hence the radius of the circle is l cos
θ . We have two equations .

Eliminating θ , we have
If the circle is vertical, its radius is l . The tension varies along the
circle, reaching its maximum in the lowest point. Draw the situation. In
this point , hence .

3.1. In the SI units, ω 0 = 10 s −1 , ν = 1.6 s −1 , T = 0.63 s.

3.3. Expand Eq. ( 3.46 ) in series of the small quantity γ / ω 0 = 0.02 as

. ω 1 is smaller than ω 0 of five
parts in one hundred thousand.
3.4. (b) Each spring expands x /2, the restoring force reduces to ½, the proper
angular frequency is √2 smaller. (c) The proper angular frequency is √2
larger than in (a).

3.5. (a) k = 1 kN/m; (b) with x in centimeters and t in seconds x ( t ) = 5cos 10 t ;

(c) in the same units x ( t ) = 5cos 10 t + 10cos 10 t .

3.6. Take the average on a period of Eq. ( 3.70 ) and compare the members.

3.8 In a vector diagram like in Figs. 3.7 and 3.8 the two forces are represented
by rotating vectors at the same angular velocity. The angle between them,
which is the difference between their phases, ϕ , is constant. The phase
difference between forces is the same as between displacements. From the
geometry we have and ϕ =133°.

3.9. Initial velocity is υ = 28 m/s and the kinetic energy U k = 390 kJ. This is the
work of the force in 90 m. The magnitude of the force is 4.3 kN (43 % of
the weight of the car). With 15% slope, in 100 m the car descends h = 15 m
and the potential energy decreases by mgh = 150 kJ. To stop the car the
work of the braking force should be 430 kJ. After 100 m the kinetic energy
is reduced to 110 kJ and the velocity is 15 m/s (53 km/h).

3.10. The vertical forces are equal and opposite. The horizontal forces are the
tension of the wire T , which is the centripetal force of magnitude mυ 2 / l
directed towards O and the friction of magnitude µ d mg directed opposite
to velocity. The magnitudes of both are equal to 4 N. The angle between
them is 90”. Hence the magnitude of the resultant is 5.7 N and its direction
is at 135” with velocity.
3.11. At the limit velocity υ lim the drag force is equal to the weight mg . If the
term proportional to the velocity dominates, R = C 1 aυ , υ lim = 1.3×10 8 a
2 m/s. For a = 1 mm υ
lim = 130 m/s, for a = 0.1 mm υ lim = 1.3 m/s. The
term proportional to velocity is dominant only in the second case. If the
term proportional to the square velocity of the drag dominates, υ lim = 217
√ a m/s. Hence, for a = 1 mm, υ lim = 6.9 m/s. Neglecting the term
proportional to velocity is justified.

3.12. (a) T = mυ 2 / R − mg . The centripetal force is the sum of the weight and
the tension, which in the considered point have the same direction, vertical
downwards. If the velocity is smaller than the critical one the motion is not
circular. (b) T = mυ 2 / R + mg .

3.13. (a) h = 2 R /3. (b) same on the moon, it does not depend on g .

4.1. 0.5 s

4.2. g ( h 2 − h 1 ) = ∆ ϕ . Hence 1000/9.8 or about 100 m.

4.3. Neither velocity nor acceleration are constant.

4.4. Mass does not vary, weight diminishes.

4.5. Answers are found putting the centripetal force equal to the gravitational
 attraction.

4.6. (almost 3 h).

4.7. If r is the radius of the orbit of the satellite, its velocity is .

Remember that . The period of the satellite is then
and r − R E = 1500 km.

4.8. The radius of the spheres is r = 0.60 m, the distance between their centers is
d = 1.23 m. The gravitational force is F = 4.4 × 10 −3 N and the shrinking of
the spring is 90 µm.

4.9. , (notice that this value is

much higher than the density of the crust, which is about 2000 kg/m 3
showing that the central part of the earth must be much denser than the
average; it is made of iron).

4.10. , , a little denser than

water.

4.11. .

4.12. The orbital velocity of Io is υ I = 18 km/s and consequently

4.14. . We see that the
contributions to the potential of the distant masses are important,
differently than for the force. Indeed, the potential decreases as 1/ r , the
force as 1/ r 2 .

5.1. (a). Vertically. (b) At the angle arc tang ( a / g ) to the vertical, forward.

5.2. (a) During the braking, the acceleration of the train is a t = −3 ms −2 . In the
reference frame of the train, the forces acting on the case are the inertial
force − m a t and the friction force − µ d mg . Its acceleration relative to the
train is a r = 1 ms −2 and the absolute one a a = −2 ms −2 . (b) During the
time t b of the braking, the case moves relative to the train with acceleration
a r starting from rest. Its speed is 10 m/s, both relative to the train and the
ground (train has stopped). (c) The case travels a first path s 1 = 50 m
during braking (accelerated relative motion) and a second one s 2 when the
train has stopped. In the second path, the acceleration is a ′ = −2 ms − 2 ,
taking 5 to stop. The time to stop is s 2 = 25 m.

5.3. The acceleration of the lift is 2.8 ms −2 upwards. Nothing can be said on
velocity.

5.6. The angular velocity is ω = 3.45 rad/s, the centrifugal force at the rim is m
1.8 N, where m is the mass of the insect. The force of static friction is m
0.98 N. It does not make it.
5.7. Not really, because the lateral shift of the point where the ground is hit is
about 5 mm.

6.1. Suppose that the direction from the lamp to the mirror is the same as the
velocity v O ′ . (the analysis of the apposite case is quite similar). The time
in S taken by the pulse for its round-trip is always Δt 0 = 2 l / c . The
observer in S sees the clock of S ′ moving in the direction of the length; this
is . In addition he sees that, while the pulse is travelling, the
mirror recedes with speed υ O ′ . Call the time to reach the mirror, we
have , hence . When the pulse

comes back, the observer in S ′ sees the detector approaching. If we call

the return time, we have , hence
. The period of the clock is the sum of the two.

6.2. (a) 13.5 µs. (b) 1.9 µs. (c) 560 m.

6.3. The limit for t → ∞ is c. constant in time.

6.4. υ f = 0.62 c , M = 2.1 m.

6.5. γ = 10 10 , about 5 min.

6.6. p = 1.42 meV/c.

6.7. p = 0.295 meV/c.

6.8. E = 0.852 meV.

6.9. γ = 2.75, β = 0.93.

6.10. γ = 10 5 , .

7.1. It is zero.

7.2. It moves with acceleration g .

7.3. 1 m/s. Yes, the collision was completely inelastic.

7.4. i = 5 × 10 4 Ns, F = 10 5 N.

7.5. m 2 / m 1 = 3, υ CM / υ i = 1/4.

7.6. (a) , (b) .

7.7. 0.74 R E .

7.8. (a) . (b) 10.

7.9. .

7.10. The velocity after the collision is V = (0, 2, 2) equal to the center of mass
velocity. (b) 50 J, 30 J, 20 J.

7.11. (a) θ = −29°, . (b) Not.

7.12. (a) (0,0,14) Nm, (b) If α is the angle between vectors r and F ,
, hence b = 2.8 m, (c) , hence F n = 1.4 N.

7.13. 5 m, 3.2 m, 2.05 m and 1.31 m. The initial energy is U = mg 5, the

following ones are 0.8 U , (0.8) 2 U , (0.8) 4 U , (0.8) 6 U .

7.14. The maximum energy transfer is 15 J, corresponding to Δx = 32 cm.

7.15. (14/6, 11/6, 11/6).

7.16. ∆ L O = (0, −800, 0) kg m 2 s −1 .

7.17. , .

8.1. No.

8.2. The external resultant force is zero. The total external moment about one of
the support points is zero. F 1 = 590 N, F 2 = 390 N.

8.4. Quadruple.

8.5. I = µL 2 , where µ is the reduced mass.

8.6. .

8.7. Use the parallel axes theorem. .

8.8. A positive answer would require I / R 2 > M , where M is the mass and R is
the radius of the cylinder. Clearly, this is impossible for any distribution of
the masses.

8.9. At the distance from m 1 of , which is the center of mass.

8.11. When both M O and ω have the direction of a principal axis of inertia and
when O is fixed or is the center of mass and the three principal axes about
it are equal.

8.13. , which does not depend on the pole, if it is still.

8.14. , , L = 0.

8.15. , .

8.17. There are two unknown, the tension of the wire and the acceleration of the
center of mass. Use the equations ( 7.49 ) and ( 7.59 ) and solve them.
There are two alternatives for the second equation, namely taking the pole
in the center of mass or in the point Ω where the wire detaches from the
yo-yo. In the latter case, take into account that the velocity of the pole is
parallel to the total linear momentum.
, .

8.18. (a) , (b)

8.19.
8.20.

8.21. (a) Both dumbbells rotate with counter-clockwise angular velocity, and
their centers are at rest (angular and linear momentum conservation). The
magnitude of the angular velocity ω = 2 υ / l = 2 rad/s. (b) They rotate half
a turn, then collide again. It takes t = π / ω = 1.57 s. (c) The second
collision, which is symmetric to the first, blocks the rotations and the two
dumbbells separate with translations of speed opposite to the initial ones.

Index
A
Absolute reference frame
Accelerated motion
Acceleration
Acceleration of transportation
Action and reaction
Action line
Action-reaction law
Adams, Johan
Addition of velocities
Almagest
Ampère, André Marie
Angular frequency
Angular magnification
Angular momentum
Angular momentum about an axis
Angular momentum about an axis theorem
Angular velocity
Anomaly
Aphelion
Application point
Archimedes
Areal velocity
Aristarchus
Aristarchus of Samos
Arm
Astronomia nova
Astronomical unit
Atomic mass unit
Atomic number
Average value
Axial vector
Axis of permanent rotation
Azimuth

B
Ballistic pendulum
Barycenter
Base units
Bilateral
Bound orbit
Bound vector
Brahe
Brahe, Tycho

C
Cardan mounting
Cartesian frame
Cassini
Cassini, Giovanni Domenico
Cavendish
Cavendish constant
Cavendish, Henry
Celestial equator
Celestial sphere
Center of force system
Center of mass
Center of mass frame
Center of mass momentum
Center of mass motion
Center of momenta frame
Center of the forces
Central axes of inertia
Central collision
Central field
Centrifugal force
Centripetal acceleration
Centripetal force
Circular uniform motion
CM frame
Coefficient of kinetic friction
Coefficient of restitution
Coefficient of static friction
Collision
Commnetariolus
Completely inelastic collision
Composite pendulum
Composition of forces
Configuration
Conjugate diameter
Conservation of angular momentum
Conservation of linear momentum
Conservative
Conservative force
Contact force
Contraction of the lengths
Co-ordinate
Co-ordinate axis
Copernicus
Copernicus, Nicolaus
Coriolis acceleration
Coriolis force
Coriolis, Gustave
Coriolis theorem
Couple
Couple arm
Covariance
Critical damping
Critical velocity
Cross product
Curvature
Curvature radius

D
D’Alembert, Jean Baptiste
Damped oscillation
Damped oscillator
Dark matter
De Revolutionibus
Decay time
Deferent
Degrees of freedom
Della Porta, Giovanni Battista
Density
Derivative of a vector
Derived units
Descartes, René
Determinant
Dialogue
Diameter
Dicke, Robert
Dimensional equation
Directional derivative
Dissipative
Dissipative force
Dot product
Double star
Dynamical equations
Dynamically balance
Dynamometer

E
Eccentricity
Ecliptic
Einstein
Einstein, Albert
Elastic collision
Elastic constant
Elastic deformation
Elastic energy
Elastic force
Elastic hysteresis
Elastic limit
Electromagnetic waves
Electromagnetism
Electronvolt
Ellipse
Elliptic orbits
Energy
Energy conservation
Energy diagrams
Energy-momentum vector
Energy of motion
Eötvös
Eötvös experiment
Eötvös, Loránd
Epicycle
Epicycloid
Equant
Equilibrium
Equipollent segment
Equipotential surfaces
Equivalence principle
Equivalent force system
Ether
Euclidean space
Event
Exponential
External moment about an axis

F
Faraday, Michael
Fictitious force
Field
Field of force
FitzGerald, George
Fixed axis
Force
Force centrifugal
Forced oscillator
Force field
Force moment
Foucault, Léon
Foucault pendulum
Four-momentum
Four-vectors
Fracture strength
Free fall
Free fall acceleration
Free fall to East
Frequency
Friction
Friction angle
Full width

G
Galilei
Galilei, Galileo
Galilei transformations
Galle
Galle, Johanne
Geiger, Hans
General relativity
Gimbal
Globular cluster
Goniometer
Gradient
Gravitational attraction
Gravitational constant
Gravitational field
Gravitational force
Gravitational mass
Gravitational potential
Gravity acceleration
Group
Gyroscope

H
Halley, Edmund
Handness
Harmonic
Harmonic motion
Harmonic oscillation
Harmonic oscillator
Harmonice mundi
Herschel, John
Herschel, William
Hertz
Hertz, Heinrich
Hertz, Heinrich Rudolf
High tide
Hilbert, David
Homogeneity principle
Homogeneous
Hook law
Hooke, Robert
Hooke law
Huygens, Christiaan
Hyperboles

I
Impact parameter
Impulse
Impulse-momentum theorem
Impulsive forces
Inclined plane
Independence of motions
Inelastic collision
Inertia
Inertia law
Inertial force
Inertial frame
Inertial mass
Inertial reference frame
Initial phase
Initial position
Instantaneous rotation axis
Interaction
Interaction potential energy
Interference fringes
Interval
Invariant
Isochronism
Isolated system

J
Joule, James
Jule
Jupiter
Jupiter satellites

K
Kepler
Kepler, Johannes
Kepler law
Kepler problem
Kilogram
Kinetic energy
König, Samuel

L
Laboratory frame
LASER ranging
Latus rectum
Law of inertia
Left-handed
Length of the pendulum
Lever rule
Le Verrier
LeVerier, Urbain
Lifeline
Light cone
Light-like
Linear momentum
Linear regime
Line integral
Lines of force
Lorentz
Lorentz factor
Lorentz group
Lorentz, Hendrik
Lorentz transformations
Low tide

M
Mars
Marsden, Ernest
Mass
Mass energy
Massless particles
Material point
Matrix
Matrix minor
Matrix order
Matrix product
Maxwell
Maxwell equations
Maxwell, James Clerk
Mechanical energy
Mechanical oscillator
Mercury
Metre
Metrology
Michelson
Michelson, Albert
Michelson interferometer
Michelson Morley Experiment
Micrometer
Molecule
Moment
Moment of a couple
Moment of inertia
Momentum
Moon
Morley, Edward
Morse potential
Motion
uniformly accelerated
Motion about a fixed pole
Motion periodic
Multiples of units
Muon

N
Natural length
Neutral equilibrium
Newton
Newton constant
Newton law
Newton, Isaac
Non-conservative
Norm
Normal reaction
Nutation

O
Objective lens
Operational definition
Opposite vector
Optical lever
Oriented segment
Ørsted, Hans Christian
Orthogonal matrix
Oscillation amplitude
Oscillations
Osculating circle
Over-damping

P
Parabola
Parallel axes theorem
Parallel forces system
Parallelogram rule
Parsec
Particle
Pascal
Pascal, Blaise
Pendulum
Perihelion
Perihelion of Mercury
Period
Periodic motion
Permanent deformation
Permanent rotation axes
Perpendicular axes theorem
Phase
Phase opposition
Plane motion
Planets
Plastic deformation
Plastic regime
Poincaré
Poincaré, Henry
Poisson formula
Poisson, Siméon-Denis
Polar co-ordinates
Pole
Position
Position vector
Potential energy
Power
Precession
Precession of perihelion
Principal axes of inertia
Principia
Product
Products of inertia
Proper angular frequency
Proper length
Proper time
Pseudo-Euclidean space
Pseudoscalar
Pseudovector
Pseudscalar
Ptolemy
Ptolemy, Claudius
Pure rolling

Q
Quadrature tides
Quantity of motion

R
Radian
Radius
Radius of gyration
Rectilinear uniform
Reduced length
Reduced mass
Reference frame
Relative velocity
Relativistic mechanics
Relativity principle
Relativity theory
Resolving power
Resonance
Resonance curve
Resonance frequency
Rest energy
Rest length
Restoring force
Resultant
Reynolds number
Reynolds, Osborne
Right-handed
Rigid body
Rigid motions
Rolling
Rolling friction
Rolling resistance
Rolling resistance coefficient
Rosette
Rotation
Rotation curve
Roto-translation
Rutherford, Ernest

S
Saturn
Scalar
Scalar product
Scalar triple product
Scale non-invariance
Scattering angle
Second
Second Newton law
Semi-latus rectum
Semi-major axis
Sidereal year
Sidereus nuncius
Simultaneity
Sistème International
Sources of the field
Space
Space inversion
Space-like
Space rotation
Space-time
Special relativity
Spherical symmetry
Spinning top
Spiral galaxy
Spontaneous rotation axes
Spring constant
Square matrix
Stable equilibrium
Static friction
Static translation
Stationary field
Stationary oscillation
Stationary solution
Steiner
Stevin
Strain
Stress
Strong interaction
Submultiples of units
Symmetry properties
Synchronize clocks
Synodic period
Syzygy

T
Target particle
Telescope
Tension
Tensor of inertia
Tide-generating force
Tides
Time
Time dilation
Time interval
Time-like
Time translations
Top
Torque
Torque about an axis
Torsion balance
Total angular momentum
Total energy
Total mechanical energy
Total moment
Total momentum
Total torque
Trajectory
Translation
Triple vector product
Tuning fork
Tunnelling
Two-body system
Two new sciences

U
Under-damping
Unification
Uniform circular
Uniform field
Uniform motion
Uniform translation motion
Unilateral
Unit vector
Universal gravitation
Unstable equilibrium
Uraniburg observatory

V
van der Waals force
van der Waals, Johannes
Variable speed motion
Varignon experiment
Varignon, Pierre
Vector
Vector components
Vector diagram
Vector direction
Vector magnitude
Vector moment
Vector norm
Vector product
Vector sum
Velocity
Velocity of light
Velocity of transportation
Venus
Virtual displacement
Virtual work
Virtual works principle
Viscosity
Viscous drag
Viscous force
Viscous resistance
Viviani, Vincenzo
von Mayer, Juilus

W
Wallis, John
Water chronometer
Watt
Watt, James
Weak interaction
Weight
Wind circulation
Work
Wren, Christofer

Y
Young modulus
Young, Thomas

Z
Zenith