FEATURES

www.iop.org/Journals/PhysEd

Introducing motion in a circle

John Roche
Linacre College, Oxford, UK
E-mail: john.roche@linacre.oxford.ac.uk

Abstract
Motion in a circle troubled Newton and his contemporaries and troubles
students today. This article attempts to give a clear presentation of certain
aspects, particularly centripetal acceleration and centrifugal force.

A difficult concept ideas almost seem like common sense? I believe

The founders of classical mechanics in the so.
seventeenth century had considerable difficulty I have chosen a very introductory—almost
with motion in a circle. The great Dutch physicist remedial—level of presentation, intended for
Christiaan Huygens (1629–1695) was the first those A-level and foundation students who may
to work out the correct dynamical laws for a not all be mathematically proficient. The emphasis
body moving uniformly in a circle. However, will be on interpretation and explanation because it
he thought that the force acting on the moving is here, I believe, that difficulties chiefly lie in this
body was directed away from the centre of topic, as in so many others in physics. Of course,
motion, and coined the term ‘centrifugal’ force there are many correct ways—and unlimited
to describe it [1]. Isaac Newton (1642–1727) numbers of unhelpful ways—of teaching motion
was unclear on this point until, in 1684, assisted in a circle or anything else. The structure, the pace,
by Robert Hooke (1635–1702), he coined the the use of computer animations and other visual
expression ‘centripetal’ force and begins to speak aids, worked examples in class and coursework
unambiguously of a centripetal force acting on will, of course, depend on the background of the
each planet and directed inwards towards the Sun students. Where I have found that an aspect of
[2]. Huygens’ mistaken version of the concept of this topic is generally well treated in textbooks, I
centrifugal force continues in use outside physics have barely touched upon it here. John Warren and
to this day, despite efforts to banish it by very Tom Duncan, for example, seem to me to provide
distinguished physicists, such as Alexis Clairaut very clear explanations of weightlessness [5], so I
(1713–1765) and Heinrich Hertz (1857–1894) [3]. do not deal with it. I have also benefited greatly
It may, indeed, seem counter-intuitive from recent pedagogical literature on mechanics
that a body moving uniformly in a circle [6]. The main focus of the article is, therefore, on
should somehow be accelerating towards the aspects of the interpretation of motion in a circle
centre. Richard Westfall, pre-eminent historian of which I judge to be poorly treated in the literature,
Newtonian physics, describes this insight as ‘the or entirely absent.
supreme act of imagination in the construction of As every physics teacher knows, to teach the
modern dynamics’ [4]. If centripetal acceleration kinematics and dynamics of motion in a circle
is so difficult to grasp, can we avoid teaching requires a knowledge on the part of the class
it—and the non-existence of Huygens’ centrifugal of the geometry of the circle, angular velocity
force—as brute facts that our students must and its relationship to tangential velocity, an
uncomprehendingly accept, or can we build up elementary knowledge of vector graphics [7], quite
their conceptual resources step by step so that these a sophisticated understanding of acceleration, and

0031-9120/01/050399+07$30.00 © 2001 IOP Publishing Ltd PHYSICS EDUCATION 399

J Roche

a knowledge of Newton’s laws of motion and his 1 radian will be swept out in 1/ω seconds, and
law of gravity. I have encountered difficulties of 2π radians will be swept out in 2π/ω seconds,
understanding in each of these areas—both on my which is just T , the period of revolution.
part and on the part of my students—and I shall
deal one by one with those that are relevant. Defining acceleration
I find that students often have a very unclear
The geometry of the circle understanding of the concept of acceleration in its
I introduce angular measure by asking students to full generality. I do not believe it is helpful to
explain each of the units found on their calculators, launch it immediately with a measuring definition
usually abbreviated to D R G. I ask how many and even less so with a mathematical formula, nor
‘radians’ there are in a right angle, approximately, is it particularly helpful at this point simply to say,
and often receive no response. I then say there are generally, that a body is accelerating if its velocity
about 1 21 radians in a right angle, or about 6 in a is changing in magnitude or direction. When I ask
circle, and sketch a radian. It is most important that students how acceleration is measured I usually
students can imagine the radian approximately, get the reply ‘metres per second squared’. When
and not simply know it as a concept associated I ask them what they mean by ‘second squared’
with a formula. (Again and again, I encounter the class usually falls silent. In fact, as I have
students who somehow confuse the symbol 2π attempted to show elsewhere, ‘second squared’
with 360.) Finally, I say that the ‘grade’ divides the has no meaning and this notation is the result of
right angle into 100 units, and that it was invented an incoherent merging of Fourier’s dimensional
during the French Revolution when all units were analysis with Gauss’s quantity calculus in the late
being decimalized [8]. nineteenth century [10].
What motivated the invention of the radian? I have found that the most successful
This requires more careful explanation. I say that explanatory definition of acceleration is the
6.283. . . (2π) of them are required to sweep out a ‘addition of velocity to a body’ since this applies
complete circle. I ask why such an inconveniently to all kinds of acceleration and is rigorous. (This
large unit is used, worse: one that does not even is the definition used by Galileo: see [11].) I
fit into a circle a whole number of times. I then also state that the direction of acceleration is the
turn to the circle and show that, if the degree is direction in which velocity is being added. The
used as an angular unit, the equation relating arc class will already be familiar with the straight
(s), radius (r) of a circular arc and angle (θ) has line acceleration of a car. This works them into
the clumsy form s = (2π/360)rθ. If the radian a concrete mode of thinking. I then ask how
[9] is used I show that this relationship becomes the acceleration of a car is measured. Primed
the far more convenient expression s = rθ. It is a by the earlier definition I may get the answer
considerable problem for some students to know ‘velocity added per second’. I then apply this to
when to use degrees and when to use radians. In free fall under gravity and persuade them to tell
the theory of motion in a circle angles are always me that a velocity of about 10 metres per second
measured in radians. is added every second to a falling body. After
I usually introduce angular velocity in terms 1 second this gives 10 metres per second, it adds
of the orbit of the Moon around the Earth. I up to 20 metres per second after 2 seconds, and
define it as the angle swept out per second by so on. Clearly, the measuring definition of the
the imaginary line running from the Earth to acceleration of gravity is 10 metres per second
the Moon1 . I find students generally have little added every second, or 10 metres per second, per
difficulty with the derivation of the formula v = second, for short. I discourage them from writing
ωr once they know s = rθ . It is also important to 10 metres per second squared and, instead, ask
derive for them the period formulas, T = 2π r/v them to write 10 m/s/s [12]. Once this concept
and T = 2π/ω. The latter they find rather abstract is clearly expressed more accurate values can be
but it can be made more intuitive by pointing out introduced.
that if ω radians are swept out in one second, The example of a projectile helps the class to
understand that velocity can be added to a body
1 About 2.5 microradian per second. in a direction that is perpendicular to the actual

400 PHYSICS EDUCATION

 Introducing motion in a circle

velocity of the body, or indeed in any direction velocity is so directed that it reduces the velocity
whatever. This means, of course, that acceleration a little in the tangential direction, but increases
can be in any direction, irrespective of the present it in a perpendicular direction, by just the right
direction of velocity. A vector graph or animation amounts so that the resultant velocity has the same
showing velocities added during the motion of the magnitude as before—and is turned through the
projectile is essential here. At this stage I find appropriate angle.
that students appear able to grasp the directional Since velocity is being added to the body, it
independence of velocity and acceleration without is accelerating. Acceleration in a technical sense
too much difficulty. may not, therefore, mean an actual increase in
speed. It is best to point out to the class here
Explaining centripetal acceleration that some of them may experience a conflict with
their pre-scientific understanding of acceleration,
Consider the Moon M (figure 1(a)), travelling for
which might lead them to assume that whenever
a moment parallel to the tangent to its orbit. If
a body accelerates its speed must increase.
velocity were not being added to the Moon in an
However, all that is required scientifically for
inwards direction, it would continue moving off
acceleration is that velocity is being added—
inertially along its tangent. If just the right total
whether or not it results in an increase, decrease
amount of velocity, in the right direction, is added
or no change in speed is unimportant. In this case
in each short period of time, the Moon will end
velocity is being added in a direction that always
up after each such period having the same velocity
points towards the centre of the circle. This means
magnitude as before, but it will be travelling along
that the acceleration is centripetal—towards the
the tangent to the orbit in its new position. The
centre.
angle in the diagram is greatly exaggerated to help
Bodies do not accelerate spontaneously. If a
visualize the explanation.
body is accelerating there must be another body or
It may be difficult for beginners to grasp that
agency exerting a force upon it. For a stone whirled
adding a velocity can mean that the body ends
around by a string the body directly acting on the
up with the same velocity magnitude as before,
stone is the string, and the force is the tension of
and a velocity vector construction (known as a
the string pulling the stone inwards. For a planet
hodograph) helps to make this clear (figure 1(b)).
orbiting the Sun it is the gravitational field of the
The figure can be used to show that the added
Sun, acting directly on the planet, that causes it
to accelerate centripetally towards the Sun. The
Sun does not make the planet move, but it does
impress a circular form on the existing motion
of the planet2 . Without the attraction of the Sun
the planet would fly off at a tangent, following its
natural inertial motion.
I find it very important not to tell the students
at this point that a centripetal acceleration is caused
by a ‘centripetal force’ because they will surely
come to believe that centripetal force is a new
kind of force along with gravity, contact forces
and electromagnetic forces. Furthermore, they
may even think of centripetal force as something
abstract, and not link in their minds to any agency
which exerts it. There is, of course, no such
special category of force as ‘centripetal force’. I
state that any force which truly causes centripetal
Figure 1. (a) Adding velocity to the Moon. The
direction in which velocity is being added, δ v , is the 2 Although the Moon and many planets have nearly circular

mean direction of acceleration. The angle θ is greatly orbits, all planetary orbits are, of course, more precisely
exaggerated for convenience. (b) The hodograph described as elliptical. In such orbits the Sun does modify
(velocity vector diagram) of (a). the actual speed of the planet, in a periodic manner.

PHYSICS EDUCATION 401

J Roche

acceleration in a moving body must be able to

follow the latter around (as a string does), or
otherwise act centripetally upon it at every point of
its path. Almost any familiar dynamic agency can
exert a force in such a centripetal manner. I also try
to talk about this force in a concrete way, such as
‘a tension acting centripetally’ or ‘a gravitational
force acting centripetally’. Some of these points
are well made in several, but not in all, textbooks Figure 2. The hypothetical lateral displacement of the
[13]. Moon from the Earth. Gravity, in the absence of lunar
tangential motion, would cause the Moon to fall
Students may not yet be entirely satisfied. towards the Earth about 17.5 km in one hour.
If the Moon is being pulled by the Earth’s Tangential inertial motion, in the absence of gravity,
gravitational field, if it is continually accelerating would cause the Moon to move laterally away from the
towards the Earth, why does it not actually strike Earth 17.5 km in one hour. The combined effects of
the Earth? This is an excellent question to put gravity and inertial motion force the path of the Moon
into a circular arc—following the curvature of the
to the class. If they answer vaguely this is Earth’s surface—and it keeps the same radial distance
because a rather new concept is involved. If the from the Earth.
Moon were instantaneously at rest with respect
to the Earth then it would, indeed, fall towards principle, which states that the negative of mass
the Earth a distance of about 17.5 km in one times acceleration is equal to the fictitious force
hour. Conversely, if there were no gravity, the required to bring the system to equilibrium [16].
tangential inertial motion of the Moon would carry D’Alembert’s principle can be very perplexing,
it away laterally (not radially: see figure 2) from and is no longer found in undergraduate textbooks.
the Earth a distance of about 17.5 km in one hour. A recent engineering mathematics textbook
The combined effects of tangential inertial motion states that ‘The centripetal force. . . [and]. . . the
and gravity force the path of the Moon into the centrifugal force. . . are in equilibrium at each
arc of a circle—which follows the curvature of instant of the motion’ [17]. We might be forgiven
the Earth’s surface—and maintains it at the same for thinking that this is what theologians call the
radial distance from the Earth as it moves3 . invincible blindness that can only be rectified by
prayer. Unfortunately, this view is widespread.
Centrifugal force For example, many students are likely to have
Teachers may groan to see yet another discussion absorbed uncritically the statement that the Earth’s
of this troubled subject, but I believe it is a attraction on the Moon is balanced by a centrifugal
touchstone of conceptual difficulty in mechanics. force. The standard physics response to this is to
I have identified at least three interpretations of point out that if the force of gravity on the Moon
centrifugal force in the literature: a valid meaning were balanced, then according to Newton’s second
in physics, an entirely different but equally valid law there would be no lunar acceleration, since
meaning in engineering, and a cluster of false there would be no resultant force, and the Moon
meanings. I will consider some of the false would fly off at a tangent. So there must be an
meanings first. One writer on a respectable unbalanced centripetal force acting on the Moon
website states that ‘When [centrifugal force] does to maintain the circular form of its orbit. This
exist, it is due to the acceleration of the mass explanation tacitly implies an inertial perspective,
of an object’ [15]. In reality, of course, force but students appear to adopt that naturally. Indeed,
causes acceleration (with mass as the control) it can confuse them here to talk about inertial
and not the other way around. Here, perhaps, frameworks.
we have a confused application of D’Alembert’s I add the following point. If there is a
centrifugal force on the Moon then it must be
3 In fact, since Newton, it has been often assumed in physics
caused by some body acting outwards on the
that gravitational acceleration is a rapid series of sudden, small,
discrete additions of velocity. Between each of these sudden
Moon, since all forces are caused by bodies: forces
additions the Moon will continue its tangential inertial motion are not abstract entities. Of course, no body or
[14]. other agency can be identified to cause the claimed

402 PHYSICS EDUCATION

 Introducing motion in a circle

lunar ‘centrifugal force’ (either in an inertial or in

a revolving framework) [18].
Well, ‘what about the centrifuge’, some may
object? I point out that the heavier particles in
the centrifuge fly off to the sides more or less
tangentially, rather than centrifugally, because the
differential liquid pressure cannot provide enough
centripetal force to maintain them in a circular path
in the centrifuge. The motion of these particles is
largely a tangential motion, therefore, and not a
Figure 3. Centripetal force on a body M, and
radial outwards motion, and there is no ‘centrifugal centrifugal force on a string S, in a gravity-free
force’ pushing them out. environment. FS is the elastic centripetal force exerted
There is, however, a valid concept of by the string S on the mass M, and FM is the
centrifugal force in physics. If the observer centrifugal inertial reaction of the mass M on the string
in a frame of reference rotating with the Earth S. Although equal and opposite, these forces do not
balance, because they act on different bodies.
pretends for mathematical convenience that it
is an inertial frame, then it becomes necessary
to postulate a fictitious outwards force on a Interestingly, the above considerations do not
geostationary satellite to explain why it does not apply to many situations that concern physicists.
plunge to Earth. This is the centrifugal force When the gravitational field of the Earth applies a
of physics, an entirely fictional force [19]. It centripetal force to the Moon, for example, the
has now virtually disappeared from school and Moon does not react centrifugally on the field.
undergraduate physics textbooks because it can There is no direct reaction on a field, whether it is
be highly confusing. Indeed, it is not uncommon gravitational, electric or magnetic [22]. The Moon
even for physics authors to confuse the language does, of course, act gravitationally on the Earth.
of inertial and rotating frameworks. True centrifugal force exists only as a reaction to
But we must leave the final word to macroscopic contact or binding forces.
the engineers. The stresses that develop in A physics teacher might agree with this but
rapidly rotating turbine blades are thought of by disagree that there should be any encouragement
mechanical engineers as being due to centrifugal whatever for centrifugal force in a physics class.
forces [20]. To take a simple example, an It is difficult enough to put across the concept of
object whirled on an elastic string pulls the string centripetal force, and banish the misuse of the
outwards, creating the tension in the string. Both
concept of centrifugal force. Furthermore, the
the inertial centrifugal force acting on the string
centrifugal force of physics is confusing, and the
and the elastic centripetal force acting on the
centrifugal force of engineering applies only in
moving body are reaction forces—they call each
special situations. These arguments would only
other into existence. Centrifugal and centripetal
allow ‘bad’ centrifugal force in again by the back
force are equal and opposite here but do not
balance because they act on different bodies door, and result in total incomprehension.
(figure 3). I agree fully that the fictional centrifugal
In a rotating turbine, for example, each outer force of physics should not even be mentioned.
section of the blade exerts an outwards pull on the However, if the bulk of the class intend to go
portion between it and the shaft, while at the same into engineering surely it would help them to
time the latter exerts an elastic inwards pull on the have the engineering concept of centrifugal force
former. It is the stresses in the blades and their explained clearly in their physics class beforehand.
causes that mainly interest engineers, rather than Indeed, there does seem to be some kind of
the centripetal forces. It follows that both elastic unhelpful conflict, with respect to this topic,
centripetal forces and inertial centrifugal forces between physicists and engineers. It must be
act in a rotating solid body [21]. What makes admitted, nevertheless, that this subject is subtle
centrifugal force rather difficult to grasp here is and the least confusing strategy for most physics
that it is an inertial reaction, and such forces are groups may be to teach them centripetal force only,
not easy to visualize. and leave centrifugal force to the engineers.

PHYSICS EDUCATION 403

J Roche

Calculating the centripetal acceleration really make sense to say that this equation ‘is’
I will conclude with the derivation of an expression the centripetal acceleration. Like all equations
for centripetal acceleration. My goal here is to in physics it represents a relationship between
provide the simplest and most intuitive proof. The distinct physical quantities, for example between
following version is standard, but not perhaps the lunar centripetal acceleration and the quantities
nuances, so I will give it in full. upon which it depends—tangential velocity and
In figure 1(b) δ v is the small velocity added angular velocity.
Since v = ωr, there are two other versions of
during the short time δt. The mean acceleration in
this equation,
magnitude, therefore, is given by
a = v 2 /r and a = ω2 r.
a = δv/δt
The version to choose depends, of course, on the
where δv is the scalar magnitude of the vector problem in hand, and it should be selected to
δ v . A scalar derivation is easier for beginners. simplify the calculations as much as possible.
(Later, this can be generalized to a = dv /dt.)
Received 29 January 2001, in final form 14 March 2001
The direction of acceleration is, of course, that of PII: S0031-9120(01)21457-9
δ v , the added velocity.
Since all three velocity scales are equal on
References
the diagram, the numbers measuring velocity in
[1] Huygens C 1888–1967 Oeuvres complètes vol 16
each case can equally well be interpreted as (The Hague: Martinus Nijhoff) pp 28–30
numbers measuring the lengths of the sides of [2] Dugas R 1958 Mechanics in the Seventeenth
the isosceles triangle MAB. For the second part Century (Neuchâtel: Editions du Griffon)
of the proof, therefore, we will think of the pp 294–6, 353–73
numbers representing v, v and δv simply as ad Westfall R 1971 Force in Newton’s Physics
(London: Macdonald) pp 167–77, 352, 426,
hoc numerical measures of the lengths of the sides. 432
This sudden transition to a purely geometrical [3] Dugas R 1958 Mechanics in the Seventeenth
interpretation here of δv and the two v’s needs Century (Neuchâtel: Editions du Griffon)
to be pointed out, otherwise the explanatory jump p 296
may perplex many students. [4] Westfall R 1971 Force in Newton’s Physics
(London: Macdonald) p 433
If AB is very small it can be regarded to a good [5] Warren J W 1979 Understanding Force (London:
approximation as representing the magnitude of John Murray) pp 20–21
the circular arc AB linking the radii MA and MB. Duncan T 1994 Advanced Physics (London:
It follows that John Murray) pp 170–1
[6] Warren J W 1979 Understanding Force (London:
AB = MA δθ John Murray) p 9
Goldring H and Osborne J 1994 Student’s
difficulties with energy and related concepts
or, in terms of their ad hoc measures, Phys. Educ. 29 26–32
Roche J 1995 Why is physics so difficult? Phys.
δv = vδθ. World 8 (6) 17–18
Slade I 1997 The dilemma of physics teaching
δθ in figure 1(b) should be measured in radians. Eur. J. Phys. 18 68–74
Now we switch back to a kinematic interpretation Ölme A 2000 Views on the physics curriculum
beyond 2000 Phys. Educ. 35 195–8
of δv and v. From its measuring definition, the
[7] Roche J 1997 Introducing vectors Phys. Educ. 32
acceleration 339–45
[8] Roche J 1998 The Mathematics of Measurement:
a = δv/δt = vδθ/δt = vω. a critical history (London: Athlone, Springer)
pp 146–7
This is perfectly accurate only when δθ is [9] Roche J 1998 The Mathematics of Measurement:
idealized to an infinitesimal size4 . It does not a critical history (London: Athlone, Springer)
p 134
4 Infinitesimals are still commonly used in physics as interim [10] Roche J 1998 The Mathematics of Measurement:
idealizations or useful fictions which immediately lead on to a critical history (London: Athlone, Springer)
the formal sense of derivatives [23]. pp 202–7

404 PHYSICS EDUCATION

 Introducing motion in a circle

[11] Galileo G 1952 Dialogues Concerning the Two [21] This analysis has a long history behind it, in both
New Sciences transl. H Crew and A de Silvio science and engineering:
Great Books of the Western World vol 28 Keill J 1745 An Introduction to Natural
(Chicago: Encyclopaedia Britannica; Philosophy (London) pp 285–6
1st Italian edn, 1638) p 200 Rankine W J M 1877 A Manual of Applied
[12] Roche J 1998 The Mathematics of Measurement: Mechanics (London: Griffin) pp 491–2
a critical history (London: Athlone, Springer) Jamieson A 1903 A Text-book of Applied
pp 227, 239, 281 Mechanics and Mechanical Engineering
[13] For an excellent explanation see Young H D and (London: Griffin) pp 88–96
Freedman R A 2000 University Physics Bevan T 1958 The Theory of Machines (London:
(San Francisco: Addison-Wesley) pp 140–1 Longmans) pp 226–7
[14] Westfall R 1971 Force in Newton’s Physics Mabie H M and Reinholtz C F 1987 Mechanisms
(London: Macdonald) p 429 and Dynamics of Machinery (New York: John
[15] Website http://observe.ivv.nasa.gov/nasa/space/ Wiley) pp 396–9
centrifugal/centrifugal1.html [22] Lorentz H A 1935–9 Collected Papers vol 5
[16] Goldstein H 1980 Classical Mechanics (Reading, (The Hague: Martinus Nijhoff) p 28
MA: Addison-Wesley) pp 16–18 [23] Roche J 1998 The Mathematics of Measurement:
[17] Kreyszig E 1999 Advanced Engineering a critical history (London: Athlone, Springer)
Mathematics (New York: John Wiley) p 436 pp 84–5
[18] For an excellent analysis of several of these
issues see Warren J W 1979 Understanding John Roche teaches courses in the
Force (London: John Murray) 17–19 history of physics at Linacre College,
Oxford, and courses in physics and
[19] Goldstein H 1980 Classical Mechanics (Reading, applied mathematics at Oxford Brookes
MA: Addison-Wesley) pp 178–9 University. His chief research interest
[20] Mabie H M and Reinholtz C F 1987 Mechanisms lies in the use of history and forensic
and Dynamics of Machinery (New York: John analysis to try to clarify some of the more
Wiley) pp 396–9 refractory concepts of classical physics.

PHYSICS EDUCATION 405