Andrew Pressley

Department of Mathematics, King's College, The Strand, London WC2R 2LS, UK Preface
Cover iJluslrutic}T/ elements reproduced by killdpermission ofi "
Aplech S)'lIteffill, Inc" PubJisha"a of the GAUSS Mathemuical and Sta~tic.al System, 23110"l S.E. Kent-Kanglq Road, Maple Valley, WA 980)8,
USA. Tel: (206) 432·7855 PU(2C6l 432·7832 email: infothptech,com URL: www.aptech.com
Amtrican Statilitical Association: Chance Vol 8 No I, 1995 article by KS and KW Heiner 'Tr~ Rings ofthe Northern Shawangunks' page 32 fig 2
Springer-Verlag: Mathematlca in Education and Researcli Vol 4 155ue 31995 article by Roman E Maeder, BeaU'ke Amrhein and Oliver Gloor
'lUustC81ed Mathematics: Yisualizetion ofMathc:matical Ob~' page 9 fig II, originlllly published u a CD ROM 'Illwtrated Mathematica' by
TELOS: ISBN 0-387-14222-3, german edition byBirkhauser: ISBN 3-7643-S1()(}'4.
MatbematiCll in Education and Research Vol4 l.uue31995 article by Richard J Gaylord and Kuume Nishidate 'Traffic Engineering with Cellular
Automata' page 35 fig 2. Mathematica in EdueadOll and Rellearch Vo15 Issue 2 1996llrticle by Mich~d TrOll 'The Implicitization of a Trefoil
Knot' page 14.
Mathematica in EdUClition and Research Vol 5 Issue 2 1996lrticle by Lee deColi 'Coins, Trees, Ban lItId Bells; Simulation of the Binomial Proc-
es.s page 19 fig 3. Mathernatica in Education and Research Vo15 Issue 2 1996llrticle by Richard GAylord and Kazume Nishidate 'Contligiow
Spreading' page 33 fig 1. Mathematka in Education and Reelrch Vol 5 Is.me 2 1996 article by Joe BuMer and Sian Wagon 'Secrets of the
Maddung Constant' page 50 fig 1.

Springer Undergraduate Mathematics Series ISSN 1615-2085

ISBN 1-85233-152-6 Springer-Verlag London Berlin Heidelberg
British Library Cataloguing in Publication Data The Differential Geometry in the title of this book is the study of the geometry
A catalogue record for this book is available from the British Library of curves and surfaces in three-dimensional space using calculus techniques.
Library of Congress Cataloging-in- Publication Data This topic contains some of the most beautiful results in Mathematics, and
Pressley. Andrew yet most of them can be understood without extensive background knowledge.
Elementary differential geometry I Andrew Pressley.
p. em. - (Springer undergraduate mathematics series. ISSN 1615·2085) Thus, for virtually all of this book, the only pre-requisites are a good working
Includes index_ knowledge of Calculus (including partial differentiation), Vectors and Linear
ISBN 1-85233-152-6 (alk. pap«)
I. Geometry, Differential. I. Title. II. Series.
Algebra (including matrices and determinants).
QA641 .P68 2000
Many of the results about curves and surfaces that we shall discuss are pro-
516.3'6-dc21 totypes of more general results that apply in higher-dimensional situations. For
00-058345
example, the Gauss-Bonnet theorem, treated in Chapter 11, is the prototype of
Apart from any fair dealing for the purposes of research or private study. or criticism or review. as a large number of results that relate 'local' and 'global' properties of geometric
permitted under the Copyright. Designs and Patents Act 1988. this publication may only be reproduced. objects. The study of such relationships has formed one of the major themes
stored or transmitted. in any form or by any means. with the prior permission in writing of the publishers,
or in the case of reprographic reproduction in accordance with the terms of licences issued by the of 20th century Mathematics.
Copyright Licensing Agency. Enquiries concerning reproduction outside those terms should be sent to the We want to emphasise, however, that the methods used in this book are
publishers.
not necessarily those which generalise to higher-dimensional situations. (For
© Springer-Verlag London Limited 2001 readers in the know, there is, for example, no mention of 'connections' in the
Printed in Great Britain
remainder of this book.) Rather, we have tried at all times to use the simplest
The use of registered names, trademarks etc. in this publication does not imply. even in the absence of a approach that will yield the desired results. Not only does this keep the pre-
specific statement, that such names are exempt from the relevant laws and regulations and therefore free
for general use. requisites to an absolute minimum l it also enables us to avoid some of the
The publisher makes no representation, express or implied, with regard to the accuracy of the information conceptual difficulties often encountered in the study of Differential Geometry
contained in this book and cannot accept any legal responsibility or liability for any errors or omissions in higher dimensions. We hope that this approach will make this beautiful
that may be made.
subject accessible to a wider audience.
Typesetting: Camera ready by author It is a cliche, but true nevertheless, that Mathematics can be learned only
Printed and bound at the Athenaeum Press Ltd.. Gateshead. Tyne & Wear
12/3830-543210 Printed on acid-free paper SPIN 10711768 by doing it, and not just by reading about it. Accordingly, the book contains
V
VI Elementary Differential Geometry p

over 200 exercises. Readers should attempt as many of these as their st.amina
permits. Full solutions to all the exercises are given at the end of the book, but Contents
these should be consulted only after the reader has obtained his or her own
solution, or in case of desperation. We have tried to minimise the number of
instances of the latter by including hints to many of the less routine exercises.

Preface , _. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
1. Curves in the Plane and in Space 1
1.1 What is a Curve? 1
1.2 Arc-Length. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3 Reparametrization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.4 Level Curves vs. Parametrized Curves 16
2. How Much Does a Curve Curve? 23
2.1 Curvature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.2 Plane Curves 28
2.3 Space Curves 36
3. Global Properties of Curves 47
3.1 Simple Closed Curves 47
3.2 The Isoperimetric Inequality 51
3.3 The Four Vertex Theorem 55
4. Surfaces in Three Dimensions 59
4.1 What is a Surface? 59
4.2 Smooth Surfaces 66
4.3 Tangents, Normals and Orientability 74
4.4 Examples of Surfaces 78
4.5 Quadric Surfaces 84
4.6 Triply Orthogonal Systems 90
4.7 Applications of the Inverse Function Theorem 93
VII
VIII
Elementary btfferenbal Geometry

11.4 Singularities of Vector Fields -- - . .. 269

5. The First Fundamental Form _ . 97
97 11.5 Critical Points __ __ -- 275
5.1 Lengths of Curves on Surfaces --
5.2 Isometries of Surfaces _ - -. 101 Solutions __ _ -- - 281
5.3 Conformal Mappings of Surfaces . 106 Chapter 1 _ - . . . . .. 272
5.4 Surface Area . 112 Chapter 2 _ - - 275
5.5 Equiareal Maps and a Theorem of Archimedes . 116 Chapter 3 - 279
123
Chapter 4 - - . . . . . . . . .. 280
6. Curvature of Surfaces .
Chapter 5 . . . . . . . . . . . . . . . .. 286
6.1 The Second Fundamental Form - 123
Chapter 6 _ - - . . . . . . . . . .. 289
6.2 The Curvature of Curves on a Surface -- . 127
Chapter 7 294
6.3 The Normal and Principal Curvatures . 130
Chapter 8 - . . . . . . . . . . . . . . . . . . .. 299
6.4 Geometric Interpretation of Principal Curvatures . 141
Chapter 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 307
7. Gaussian Curvature and the Gauss Map . 147 Chapter 10 311
7.1 The Gaussian and Mean Curvatures . 147 Chapter 11 314
7.2 The Pseudosphere . 151
Index _ - - 329
7.3 Flat Surfaces . 155
7.4 Surfaces of Constant Mean Curvature - 161
7.5 Gaussian Curvature of Compact Surfaces . 164
7.6 The Gauss map . 165

8. Geodesics . 171
8.1 Definition and Basic Properties . 171
8.2 Geodesic Equations - . 175
8.3 Geodesics on Surfaces of Revolution _ - 181
8.4 Geodesics as Shortest Paths - . 190
8.5 Geodesic Coordinates - - . 197

9. Minimal Surfaces _ _ . 201

9.1 Plateau's Problem -' . 201
9.2 Examples of Minimal Surfaces - . 207
9.3 Gauss map of a Minimal Surface - . 217
9.4 Minimal Surfaces and Holomorphic Functions - 219

10. Gauss's Theorema Egregium . 229

10.1 Gauss's Remarkable Theorem .. - - . 229
10.2 Isometries of Surfaces . 238
f··
10.3 The Codazzi-Mainardi Equations . 240
lOA Compact Surfaces of Constant Gaussian Curvature . 244

11. The Gauss-Bonnet Theorem . 247

11.1 Gauss-Bonnet for Simple Closed Curves . 247
11.2 Gauss-Bonnet for Curvilinear Polygons .. _ . 252
11.3 Gauss-Bonnet for Compact Surfaces - . 258
 1
Curves in the Plane and in Space

In this chapter we discuss two mathematical formulations of the intuitive notion

of a curve. The precise relation between them turns out to be quite subtle, so
we shall begin by giving some examples of curves of each type and practical
ways of passing between them.

1.1. What is a Curve?

H asked to give an example of a curve, you might give a straight line, say
y _ 2x = 1 (even though this is not 'curved'!), or a circle, say x 2 + y2 ::::: 1, or
perhaps a parabola, say y - x 2 = O.

y - 2x = 1
An of these curves are described by means of their cartesian equation
I(x, y) = c,
where f is a function of x and y and c is a constant. From this point of view,
1
2 Ei@iiiEiiCSfY b'ii@#eiTtid' GEO:ri@lIY

a curve is a set of points, namely

1'1 (t) = t, '"Y2 (t) = t 2. To get
every point on the parabola we must allow t to
take every real number value (since the x-coordinate of 'Y(t) is just. t, and the
c= {(X,y) E R 2 I f(x,y) = c}. (1) x-coordinate of a point on the parabola can be any real number), so we must
These examples are all curves in th~"plane R 2 , but we can also consider curves take (a,f3) to be (-00,00). Thus, the desired parametrisation is
in R 3 - for example, the x-axis in R 3 is the straight line given by 'Y: (-00,00) ---t R 2,
2
'Y(t) = (t,t ).

{(x,y,Z)ER 3
I y=z=O}, But this is not the only parametrisation of the parabola. Another choice is
and more generally a curve in R 3
might be defined by a pair of equations 'Y(t)= (t 3 ,t6 ) (with (a,f3) = (-00,00». Yet another is (2t,4t 2 ), and of course
there are (infinitely many) others. So the parametrisation of a given level curve
h(x,y,z) = Cl, h(x,y,z) = C2'
is not unique.
Curves of this kind are called level curves, the idea being that the curve in Eq.
(1), for example, is the set of points (x, y) in the plane at which the quantity Example 1.2
I(x, y) reaches the 'level' c.
Now we try the circle x 2 + y2 = 1. It is tempting to take x = t as in the previous
But there is another way to think about curves which turns out to be more
example J so that y = ~ (we could have taken y = -Jt=t2). So we get
useful in many situations. For this, a curve is viewed as the path traced out by
a moving point. Thus, if 'Y(t) is the position vector of the point at time t, the the parametrisation
curve is described by a function 'Y of a scalar parameter t with vector values 'Y(t) = (t, Jl=t2).
(in R Z for a plane curve, in R 3 for a curve in space). We use this idea to give
But this is only a parametrisation of the upper half of the circle, because
our first formal definition of a curve in R n (we shall be interested only in the
JI=t2 is always ~ O. Similarly, if we had taken y = -~, we would only
cases n = 2 or 3, but it is convenient to treat both cases simultaneously):
have covered the lower half of the circle.
H we want a parametrisationof the whole circle, we must try again. We
Definition 1.1 need functions 1'1 (t) and 1'2(t) such that
A parametrised curve in R n is a map "y (a,{3) ---t R n , for some a,{3 with 1'1 (t)2 + '"Y2(t)2 = 1 (3)
-00 ~ 0 < {3 ::; 00. for all t E (oJ (3), and such that every point on the circle is equal to C1'1 (t) ,1'2 (t»
for some t E (a, ,8). There is an obvious solution to Eq. (3): 1'1 (t) = COS t
The symbol (a, (3) denotes the open interval and 'Yz(t) =sin t (since cos 2 t + sin 2 t =
1 for all values of t). We can take
(a,{3) = {t E R I 0',< t < ,8}. (a,{3) = (-00,00), alt.hough this is overkill: any open interval (a f3) whose J

length is greater than 21T will suffice.

A parametrised curve whose image is contained in a level curve C is called
a parametrisation of (part of) C. The following examples illustrate how to pass The next example shows how to pass from parametrised curves t.o level
from level curves to parametrised curves and back again in practice.
curves.

Example 1.1
Example 1.3
Let us find a parametrisation 'Y(t) of the parabola y = x 2. If "y( t) = (1'1 (t), 1'2 (t» J
Take the parametrised curve (called an astroiri)
the components '"Y1 and '"Y2 of 'Y must satisfy
'Y(t) = (cos3 t,sin 3 t).
'"Y2(t) = '"Y1(t)Z (2)
Since cos2 t + sin2 t == 1 for all t, the coordinates x = cos3 t, Y = sin 3 t of the
for all values of t in the interval (a, (3) where 'Y is defined (yet to be decided), point 'Y(t) satisfy
and ideally every point on the parabola should be equal to (')'1(t),1'2(t» for
some value of t E (a, (3). Of course, there is an obvious solution to Eq. (2): take
4 Elementary Differential Geometry 1. Curves in the Plane and in Space

This level curve coincides with the image of the map 'Y. The following result is intuitively clear:

In this book, we shall be studying curves (and later, surfaces) using meth- Proposition 1.1
ods of calculus. To differentiate a vector-valued function such as 'Y(t) (as in
Definition 1.1), we differentiate componentwise: if If the tangent vector of a parametrised curve is constant, the image of the curve
is (part of) a straight line.

then Proof 1.1

d'Y (d'Yl d'Y2 d'Yn) If ..y(t) = a for all t J where a is a constant vector, we have, integrating compo-
dt = (ft'(ft""'dt '
nentwise,
~'Y ~'Y2
dt2 = (d?'Y1 rP'Yn)
dt2' dt2 , ... , dt2 ' etc.
'Y(t) = J~~ = J
dt adt = ta + b J
To save space, we often denote d'Y/dt by ..y(t), rP'Y/dt2 by ;yet), etc.
where b is another constant vector. If a ::j:. 0, this is the parametric equation
We say that "'( is smooth if each of the components 1'1,1'2,···, 'Yn of 'Y is of the straight line parallel to a and passing through the point with position
smooth, i.e. if all the derivatives d'Yi/dt, d?'Yddt2, d3'Yddt3, ... exist, for i. =
vector b:
1,2, ... , n. From now on, all parametrised curves studied in this book wW":be
assumed to be smooth.

Definition 1.2
If 'Y(t) is a parametrised curve, its first derivative d'Y/dt is called the tangent
vector of 'Y at the point 'Y( t) .
a
To see the reason for this terminology, note that the vector
'Y(t + 1St) - 'Y(t) o
1St If a = 0, the image of'Y is a single point (namely, the point with position vector
is parallel to the chord joining the two points 'Y(t) and 'Y(t + 1St) of the image C ~. 0
ofT

EXERCISES

1.1 Is "'(t) = (t 2 , t 4 ) a parametrisation of the parabola y = x2?

1.2 Find parametrisations of the following level curves:
(i) y2 _ x 2 = 1;
(ii) z4 +
2
= 1.f
We expect that, as 1St tends to zero, the chord becomes parallel to the tan- 1.3 Find the cartesian equations of the following parametrised curves:
gent to C at 'Y(t). Hence, the tangent should be parallel to (i) 'Y(t) = (cos 2 t, sin 2 t)j
(ii) 'Y(t) = (e t , t 2 ).
lim 'Y(t + 1St) - 'Y(t) = d'Y.
1.4 Calculate the tangent vectors of the curves in Exercise 1.3.
6t-+O 1St dt
6 Elementary Differential Geometry 1. Curves in the Plane and in Space 7

1.5 Sketch the astroid in Example 1.3. Calculate its tangent vector at 1.10 For the logarithmic spiral 'Y(t) = (e t cost, et sin t), show that the
each point. At which points is the tangent vector zero? angle between 'Y(t) and the tangent vector at -y(t) is independent of
1.6 If P is any point on the circle C in the xy-plane of radius a > 0 and t. (There is a picture of the logarithmic spiral in Example 1.4.)
centre (0, a), let the straight line through the origin and P intersect
the line y = 2a at Q, and let the line through P parallel to the x-axis
intersect the line through Q parallel to the y-axis at R. As P moves
around C, R traces out a curve called the witch of Agnesi. For this
1.2. Arc-Length
curve, find
If v = (Vi, ... , v n ) is a vector in R n , its length is
(i) a parametrisation;
(ii) its cartesian equation.
Q II v II = Jv? + ... + v~.
If u is another vector in R n, II u - v II is the length of the straight line segment
joining the points in R n with position vectors u and v.
To find a formula for the length of any parametrised curve 'Y, note that,
if 8t is very small, the part of the image C of'Y between 'Y(t) and 'Y(t + M) is
:¥----------.R nearly a straight line, so its length is approximately

o II 'Y(t + M) - 'Y(t) II·

1.7 A cycloid is the plane curve traced out by a point on the circum- Again, since c5t is small, ("{(t + at) - -y(t))/M is nearly equal to i'(t), so the
ference of a circle as it rolls without slipping along a straight line. length is approximately
Show that, if the straight line is the x-axis and the circle has radius II i'(t) II 8t. (4)
a > 0, the cycloid can be parametrised as
If we want to calculate the length of (a not necessarily small) part of C, we can
'Y(t) = a(t - sin t, 1 - cos t). divide it up into segments, each of which corresponds to a small increment rSt
1.8 Generalise the previous exercise by finding parametrisations of an in t, calculate the length of each segment using (4), and add up the results.
epicycloid (resp. hypocycloid), the curve traced out by a point on Letting c5t tend to zero should then give the exact length.
the circumference of a circle as it rolls without slipping around the
outside (resp. inside) of a fixed circle.
1.9 Show that 'Y(t) = (cos 2 t - !'sin t cos t, sin t) is a parametrisation of
the curve of intersection of the circular cylinder of radius ~ and axis
the z-axis with the sphere of radius 1 and centre (-!' 0, 0). (This is
called Viviani's Curve).

J!!!if

This motivates the following definition:

Definition 1.3
The arc-length of a curve 'Y starting at the point "Y(to) is the function s(t) given
8 Elementary Differential Geometry 9

by Definition 1.4
If 'Y : (a, f1) ---t R n is a parametrised curve, its speed at the point 'Y( t) is II -y{ t) I\,
set) = t II ..y(u) 11 duo
ltD and 'Y is said to be a unit-speed curve if-y(t) is a unit vector for all t E (a,,B).

We shall see many examples of formulas and results relating to curves that
Thus, s(to) = 0 and set) is positive or negative according to whether t
take on a much simpler form when the curve is unit-speed. The reason for this
is larger or smaller than to. If we choose a different starting point 'Y(to) , the
simplification is given in the next proposition. Although this admittedly looks
resulting arc-length s differs from s by the constant Jt~O lI..y(u) II duo uninteresting at first sight, it will be extremely useful for what follows.

Example 1.4 Proposition 1.2

For a logarithmic spiral Let net) be a unit vector that is a smooth function of a parameter t. Then, the
dot product
n(t).n(t) = 0
for all t, i.e. net) is zero or perpendicular to net) for all t.
In particular, if'Y is a unit-speed curve, then;Y is zero or perpendicular to

Proof 1.2
We use the 'product formula' for differentiating dot products of vector-valued
functions a( t) and b (t ):
d da db
dt (a.b) = dt· b + a. dt .
we have
Using this to differentiate both sides of the equation n.n = 1 with respect to t
gives
..y = (e t ( cos t - sin t) , et (sin t + cos t)) ,
.. II t 11 2 = e 2t(cost - sint)2 + e2t (sint + cost)2 = 2e
2t
•
n.n + n.n = 0,

Hence, the arc-length of'Y starting at 'Y(O) = (1,0) (for example) is so 2n.n = O.
The last part follows by taking n = ..y. o
s = it V2e 2u du = h(e t
- 1).

EXERCISES
If s is the arc-length of a curve 'Y starting at 'Y(to), we have
1.11 Calculate the arc-length of the catenary 'Yet) = (t, cosh t) starting at
~: = :t ltDr II t(u) II du = 11..y(t) 11· (5) the point (0, 1).
1.12 Show that the following curves are unit-speed:
Thinking of 'Y(t) as the position of a moving point at time t, ds/dt is the speed
of the point (rate of change of distance along the curve). For this reason, we
(i) 'Y(t) = 0(1 + t)3/2, t(1 - t)3/2, ~);
make (ii) 'Y(t) = Ucos t, 1 - sin t, -~ cos t).
10 Elementary DifferentIa eometry 1. Curves in the Plane and in pace 11

1.13 Calculate the arc-length along the cycloid in Exercise 1.7 correspond- Definition 1.6
ing to one complete revolution of the circle. A point "((t) of a parametrised curve "( is called a regular point if ..y(t) f:. 0;
otherwise 'Y(t) is a singular point of 'Y. A curve is regular if all of its points are
regular.
1.3. Reparametrisation Before we show the relation between regularity and unit-speed reparametri-
sation, we note two simple properties of regular curves. Although these results
We saw in Examples 1.1 and 1.2 that a given level curve can have many
are not particularly appealing, they will be very important for what is to follow.
parametrisations, and it is important to understand the relation between them.

Proposition 1.3
Definition 1.5
Any reparametrisation of a regular curve is regular.
A parametrised curve;y: (a,[J) -t- R n is a reparametrisation of a parametrised
curve "( : (a, (3) ---t R n if there is a smooth bijective map ,p : (a, [J) ---t (a, (3)
(the reparametrisation map) such that the inverse map ,p-l : (a,;3) -t- (a, fJ) is Proof 1.3
also smooth and Suppose that "( and ;y are related as in Definition 1.5, let t = ,p(l), and let
;y(i) = "((,p(t)) for all i E (a,[J). 'IjJ = $-1 so that i = 'IjJ(t). Differentiating both sides of the equation ,p(t{;(t)) = t
with respect to t and using the chain rule gives
Note that, since ,p has a smooth inverse, "( is a reparametrisation of;y: d,pd'IjJ -1
di dt - .
;y(,p-l(t)) = "((,p(¢-1(t))) = "((t) for all t E (a,;3).
This shows that dl'/di is never zero. Since .y(i) = "((l'(i)), another application
Two curves that are reparametrisations of each other have the same image, of the chain rule gives
so they should have the same geometric properties. d.y d'Y d¢
di=dtdi'
Example 1.5 which shows that d;y/di is never zero if d'Y/dt is never zero. o
In Example 1.2, we gave the parametrisation "((t) = (cos t, sin t) for the circle
x 2 + y2 = 1. Another parametrisation is Proposition 1.4
.y(t) = (sint,cost) If 'Y(t) is a regular curve, its arc-length s (see Definition 1.3), starting at any
(since t+
sin 2 cos2
t = 1). To see that .y is a reparametrisation of "(, we have to point of 'Y, is a smooth junction of t.
find a reparametrisation map ,p such that
(cos ¢(t), sin l'(t)) = (sin t, cos t). Proof L4
We have already seen that (whether or not 'Y is regular) s is a differentiable
One solution is ,p(t) = 1r/2 - t.
function of t and
As we remarked in the previous section, the analysis of a c;urve is simplified
when it is known to be unit~speed. It is therefore important to know exactly
~: = II .y(t) II·
which curves have unit-speed reparametrisations. To simplify the notation, assume from now on that "( is a plane curve, say
"((t) = (u(t),v(t)),
 Elementary Differential Geometry
12

2 injective, that its image is an open interval (ii,13), and that the inverse map
where u and v are smooth functions of t. Define f : R --t R by
8- 1 : (6,,8) --t (a,,8) is smooth. (Readers unfamiliar with the inverse function
f(u,v) = Ju 2 + v 2 , theorem should accept these statements for now; the theorem will be discussed
informally in Section 1.4 and formally in Chapter 4.) We take 1> = S-l and let
so that
l' be the corresponding reparametrisation of 'Y, so that
ds
dt = f('-)
u,v. (6)
t(s) = 'Y(t).
2
The crucial point is that f is smooth on R \{(0,O)}, which means that all the Then,
partial derivatives of f of all orders exist and are continuous functions except
d1'ds d'Y
at the origin (0,0). For example, ds dt dt'
of u of v II d1' II ds II II = (by Eq. (5»,
au = vu 2 + v2 ' av = vu 2 + v2 ' ds dt
= d'Y
dt
ds
dt
are well defined and continuous except where u v = =
0, and similarly for II ~; II = 1. o
higher derivatives. Since 'Y is regular, u and v are never both zero, so the chain
rule and Eq. (6) shows that ds/dt is smooth. For example,
The proof of Proposition 1.5 shows that the arc-length is essentially the
tPs af.. af .. only unit-speed parameter on a regular curve:
dt 2 = au U + 8v v,
and similarly for the higher derivatives of s. 0
Corollary 1.1
The main result we want is Let 'Y be a regular curve and let l' be a unit-speed reparametrisation of 'Y.'
i'(u(t» = "(t) for all t,
Proposition 1.5
where u is a smooth junction of t. Then, if s is the arc-length of'Y (starting at
A parametrised curve has a unit-speed reparametrisation if and only if it is any point), we have
regular. (7)
u = ±s+cl
where c is a constant. Conversely, if'U is given by Eq. (7) for some value of c
Proof 1.5
n
and with either sign, then l' is a unit-speed reparametrisation of 'Y.
Suppose first that a parametrised curve 'Y : (0:,,8) --t R has a unit-speed
reparametrisation 1', with reparametrisation map <p. Letting t = <p(i), we have
Proof 1.1
1'(t) 'Y(t) ,
=
The calculation in the first part of the proof of Proposition 1.5 shows that u
d1' d'Y dt
dt = dt di'
gives a unit-speed reparametrisation of"( if and only if

11 d~ II = 11 d'Y 111 d~l. du

dt = ± II d'Y
dt II = ± ds
dt (by Eq. (5»).
dt dt dt
Since i' is unit-speed, II di'/dt 11= 1, so clearly d"(/dt cannot be zero. Hence, u = ±s + c for some constant c. 0
Conversely, suppose that the tangent vector d'Y/dt is never zero. By Eq.
(5), ds/dt > 0 for all t, where s is the arc-length of"( starting at any point of Although every regular curve has a unit-speed reparametrisation, this may
be very complicated, or even impossible to write down 'explicitly', as the fol-
the curve, and by Proposition 1.4 s is a smooth function of t. It follows from
the inverse function theorem of multivariable calculus that s : (ct,,8) --t R is lowing examples show.
 ane and in Space 15

Example 1.6
Our final example shows that a given level curve can have both regular and
For the logarithmic spiral non-regular parametrisations.

"Y(t) = (etcost,etsint),
Example 1.8
we found in Example 1.4 that
For the parametrisation
11111 2 = 2e 2t .
"Y(t) = (t, e)
This is never zero, so 'Y is regular. The arc-length of'Y starting at (1,0) was found
to be a = v'2 (e t -1). Hence, t = In ( ~ + 1), so a unit-speed reparametrisation of the parabola y = x 2 , 1(t) = (1,2t) is obviously never zero, s0'Y is regular.
of"( is given by the rather unwieldy formula But

t(t) = (t 3 , t 6 )
t(a) = ( (~ + 1) cos (In (~ + 1)),(~ + 1) sin (In (~ + 1))).
is also a parametrisation of the same parabola. This time, ~ = (3t 2 , 6t 5 ), and
this is zero when t = 0, so t is not regular.
Example 1.7
The twisted cubic is the space curve given by EXERCISES
"(t) = (t, t 2
, t 3
), -00 <t < 00.
1.14 Which of the following curves are regular?
(i) "Y(t) = (cos 2 t, sin 2 t) for -00 < t < 00;
(ii) the same curve as in (i), but with 0 < t < 1r /2;
(iii) 'Y(t) ='(t,cosht) for -00 < t < 00.
Find unit-speed reparametrisations of the regular curve(s).

We have
'Y(t) = (1, 2t, 3t 2 ),
II 'Y(t) II = VI + 4t 2 + 9t 4 .
This is never zero, s0'Y is regular. The arc-length starting at "(0) = 0 is
s= it VI + 4u 2 + 9u4 duo
1.15 The cissoid of Diocles (see above) is the curve whose equation in
terms of polar coordinates (r,O) is
r = sin (;I tan 0, -1r/2 < (;I < 1r/2.
This integral cannot be evaluated in terms of familiar functions like logarithms
and exponentials, trigonometric functions, etc. (It is an example of something Write down a parametrisation of the cissoid using (I as a parameter
called an elliptic integral.)
 Elementary Differential Geometry 1. Curves in the Plane and in Space 17
16

function theorem has been formally introduced and used in our discussion of
and show that
surfaces.
"Y(t) = (t b),
2
, -1 < t < 1, To understand the significance of the conditions on f in Theorem 1.1, sup-
pose that (xo+dx, yo+dy) is a point ofC near P, so that f(xo+.1x, yo+dy) ==
is a reparametrisation of it. O. By the two-variable form of Taylor's theorem,
1.16 Let "y be a curve in lin and let ;Y be a reparametrisation of "y with a
reparametrisation map ¢ (so that .y(t) = "}'(¢(l»). Let to be a fixed f(xo + dx,yo + dy) == !(XO,yo) + dx ! + dya!
ax ay'
value of i and let to = iP(io). Let sand s be the arc-lengths of"Y and
neglecting products of the small quantities dx and dy (the partial derivatives
.y starting at the point 'Y(to) = .y(lo). Prove that s = s if diP/di > 0 are evaluated at (xo, Yo). Hence,
for all i, and s == -s if diP/dt < 0 for all i.
af af
dx- +dy- =0. (8)
ax ay
Since dx and dy are small, the vector (dx, dy) is nearly tangent to C at P,
1.4. Level Curves vs. Parametrised Curves so Eq. (8) says that the vector n = (M,~) is perpendicular to C at P.
Y
We shall now try to clarify the precise relation between the two types of curve
we have considered in previous sections. n
Level curves in the generality we have defined them are not always the c
kind of objects we would want to call curves. For example, the level 'curve'
x2 + y2 = 0 is a single point. The correct conditions to impose on a function
f(x,y) in order that f(x,y) = c, where C is a constant, will be an acceptable
level curve in the plane are contained in the following theorem, which shows p
that such level curves can be parametrised. Note that we might as well assume
that c = 0 (since we can replace f by f - c).
x
The hypothesis in Theorem 1.1 tells us that the vector n is non-zero at every
Theorem 1.1 point of~. Suppose, for example, that U
:f: 0 at P. Then, n is not parallel to
Let f(x, y) be a smooth junction of two variables (which means that all the par- the X-axiS at P, so the tangent to C at P is not parallel to the y-axis.
tial derivatives of f, 0/ all orders, exist and are continuous junctions). Assume
that, at every point of the level curve y
,
,,, c
C = {(x,y) E R 2
I f(x,y) == O},
a f / ax and a/ / ay are not both zero. If P is a point of C, with coordinates ,,
,
(xo,Yo), say, there is a regular parametrised curve 'Y(t), defined on an open
interval containing 0, such that"}' passes through P when t = 0 and 'Y(t) is ,,
,,
contained in C for all t.
Yo
.,
.. _- ........ - - - r - - ...

The proof of this theorem makes use of the inverse function theorem (one
version of which has already been used in the proof of Proposition 1.5). For the x
moment, we shall only try to convince the reader of the truth of this theorem.
The proof will be given in a later exercise (Exercise 4.31) after the inverse This implies that vertical lines x = constant near x = Xo all intersect C in a
18 Elementary Differential Geometry 1. Curves in the Plane and in Space 19

unique point (x, y) near P. In other words, the equation

f(x,y) = 0 (9)
has a unique solution y near Yo for every x near Xo· Note that this may fail to
be the case if the tangent to C at P is parallel to the y-axis:
y

~+y'=1 ~-~=1
p
~ith t.hese assumptions on f, there is a regular parametrised curve 'Y whose
Image IS the whole of C. Moreover, if C does not 'close up' (like a straight line
or a parabola), "I can be taken to be injective; if C does close up (like a circle
or.an ellIpse), then "I maps some closed interval [a,f] onto C, 'Y(a) = 'Y(f) and
x "I IS InjectIve on the open interval (a, f).

In this example, lines x = constant just to the left of x = Xo do not meet C A similar argument can be used to pass from parametrised curves to level
near P, while those just to the right of x = Xo meet C in more than one point curves:
near P.
The italicised statement about f in the last paragraph means that there is Theorem 1.2
a function g(x), defined for x near xo, such that y = g(x) is the unique solution
of Eq. (9) near Yo. We can now define a parametrisation"l of the part of C near Let'Y be a regular parametrised plane curve, and let "I(to) = (xo, Yo) be a point
III the .mage of 'Y. Then, there is a smooth real-valued function f(x, y), defined
P by
for x an~ ~ zn ~pen intervals containing Xo and Yo, respectively, and satisfying
"I(t) = (t, g(t)). the cond.tzons III Theorem 1.1, such that "I(t) is contained in the level curve
If we accept that g is smooth (which follows from the inverse function theorem), f(x,y) = 0 for all values oft in some open interval containing to.
then "I is certainly regular since
The proof of Theorem 1.2 is similar to that of Theorem 1.1. Let
"t = (1, g)
"I(t) = (u(t),v(t)),
is obviously never zero. This 'proves' Theorem 1.1.
where u and v are smooth functions. Since "I is regular, at least one of u(to)
It is actually possible to prove slightly more than we have stated in Theorem
and veto) IS non-zero, say u(to ). This means that the graph of u as a function
1.1. Suppose that f(x, y) satisfies the conditions in the theorem, and assume of t is not parallel to the t-axis at to:
in addition that the level curve C given by f(x, y) = 0 is connected. For readers
unfamiliar with point set topology, this means roughly that C is in 'one piece'.
For example, the circle x' + y' = 1 is connected, but the hyperbola x' - y' = 1
is not:
20 Elementary Differential Geometry 1. Curves in tne Plane and in Space 21

u EXERCISES

1.17 State a generalisation of Theorem 1.1 for level curves in R 3 given by

f(x, Y, z) = g(x, y, z) = O. (To guess the analogue of the condition
on f in Theorem 1.1, argue that (*, U' M) is perpendicular to the
surface !(x,y,z) = 0, and then think about the condition that two
planes intersect in a line. See Exercise 4.16 for a rigorous statement.)
~ ------------------
1.18 Generalise Theorem 1.2 for Curves in R 3 (or even R n ). (This is easy.)
1.19 Sketch the level curve C given by f(x, y) = 0 when f(x, y) = y -lxI-
to
Note that f does not satisfy the conditions in Theorem 1.1 because
af! ax does not exist at the point (0,0) on the curve. Show never-
As in the proof of Theorem 1.1, this implies that any line parallel to the t-axis theless that there is a smooth parametrised curve 'Y whose' image is
close to u = Xo intersects the graph of u at a unique point u(t) with t close to the whole of C. (Make use of the smooth function 8(t) defined after
to- This gives a function h(x), defined for x in an open interval containing xo, the proof of Theorem 8.2.) Is there a regular parametrised curve with
such that t = h(x) is the unique solution of u(t) = x if x is near Xo and t is this property?
near to. The inverse function theorem tells us that h is smooth. The function
f(x,y) =y - v(h(x»
For the remainder of this book, we shall speak simply of {curves',
has the properties we want. unless there is serious danger of confusion as to which type
It is not in general possible to find a single function f(x, y) satisfying the (level or parametrised) is intended.
conditions in Theorem 1.1 such that the image of 'Y is contained in the level
curve f(x, y) = 0, for 'Y may have self-intersections like the lima~on
'Y(t) = ((1 + 2 cos t) cos t, (1 + 2 cos t) sin t).

It follows from the inverse function theorem that no single function f satisfying
the conditions in Theorem 1.1 can be found that describes a curve near such a
self-intersection point.
 2
How Much Does a Curve Curve?

In this chapter, we associate to any curve in R 3 two scalar functions, called its
curvature and torsion. The curvature measures the extent to which a curve is
not contained in a straight line (so that straight lines have zero curvature), and
the torsion measures the extent to which a curve is not contained in a plane
(so that plane curves have zero torsion). It turns out that the curvature and
torsion together determine the shape of a curve.

2.1. Curvature
We want to find a measure of how 'curved' a curve is. Since this lcurvature'
should depend only on the 'shape' of the curve:
(i) the curvature should be unchanged when the curve is reparametrised.
Further, the measure of curvature should agree with OUf intuition in simple
special cases, for eXamPle:
(ii) the curvature of a straight line should be zero, and large circles should have
smaller curvature than small circles.
Bearing (ii) in mind, we get a clue as to what the definition of curvature
should be from Proposition 1.1: this tells us that, if"Y is a plane curve with
i = 0 everywhere, the curve "Y is part of a straight line, and hence should
have zero curvature. So we might be tempted to define the curvature of "Y to
be II 'Y II (we take the norm because we want the curvature to be a scalar, not
23
24 Elementary Differential Geometry ¥ 2. HOW NidtA Does 5 CuN@ CUNe f 25
i

a vector). Unfortunately, however, this depends (in a fairly complicated way) But what if we are given a curve 'Y(t) that is not unit-speed? If 'Y is regular
on the parametrisation of 1'- So let us remove this freedom to reparametrise (see Definition 1.6), then by Proposition 1.5 'Y has a unit-speed reparamet.risa-
by insisting that 'Y is unit-speed, so that II ly II = 1 everywhere. (Actually, this tion .y. We define the curvature of'Y to be that of the unit-speed curve .y. But
does not quite rule out the possibility of reparametrising - see Corollary 1.1.) since it is not always possible to find the unit-speed reparametrisation explicitly
So we make (see Example 1.7), we really need a formula for the curvature in terms of'Y and
t only.
Definition 2.1
Proposition 2.1
If'Y is a unit-speed curve with parameter s, its curvature ~(s) at the point 'Y(s)
Let 'Y(t) be a regular curve in R 3 . Then, its curvature is
is defined to be Ili(s) II·
11 i X ')' II (1)
The first part of condition (ii) will certainly be satisfied. As to the second "" = II ly 11 3 '
part, consider the circle centred at (xo, Yo) and of radius R. This has a unit- where the x indicates the vector (or cross) product and the dot denotes dldt.
speed parametrisation
Of course, since a curve in R 2 can be viewed as a curve in R 3 whose last
'Y(s) = (xo + R cos ~, Yo + R sin ~) .
coordinate is zero, Eq. (1) can also be used to calculate the curvature of plane
We have curves.
')'(8) = ( - sin ~, cos~) 1

Proof 2.1
111'(8) II = j (- sin ~f + (cos ~f = 1, Let .y (with parameter s) be a unit-speed reparametrisation of 'Y, and let us
showing that 'Y is indeed unit-speed, and hence denote dlds by a dash. Then, by the chain rule,
_I d8 .
i(8) = ( -~ cos~, -~ sin i) , 'Y dt = 'Y,

II i( 8 ) II = (- ~ cos ~ r + (- ~ sin Ii
so the curvature of the circle is inversely proportional to its radius.
r 1
R'
so

Now,
As to condition (i), recall from Corollary 1.1 that, if 'Y( s) is a unit-speed
curve, the only unit-speed reparametrisations of'Y are of the form 'Y(U) , where (~;) 2 = II Ay 11 2 = Ay.Ay,
u = ±s + c, and differentiating with respect to t gives
and c is a constant. Then, by the chain rule, ds d 2 s ...
d'Y = d'Y du = ± d"'( dt dt2 = 'Y.'Y.
ds du ds du' Using this and Eq. (2), we get
~'Y
2
d 'Y d (d'Y) du d (d'Y)
ds 2 = du ds ds = ± du ± du = du 2 ' K- = II i (*)2 - ly~* II = II i(ly·ly) - ly('Y·i) 11
This shows that the curvature of the curve computed using the unit-speed
(dsldt)4 II Ay 11 4
parameter 8 is the same as that computed using the unit-speed parameter u. Using the vector triple product identity
a x (b x c) = (a.c)b - (a.b)c
 26 Elementary DifFerential Geometry ¥ 2. How Much Does a Curve Curve ( 27

(where a, h, c E R 3 ), we get
for some value of B, then x 2 +y2 = a 2 , showing that the helix lies on the cylinder
l' x (i x 1') = i(1'·1) - 1'(1'.i). with axis the z-axis and radius lal; the positive number lal is called the radius
of the helix. As (J increases by 211", the point (a cos 0, a sin 0, bO) rotates once
FUrther, t and l' x t are perpendicular vectors, so
round the z-axis and moves up the z-axis by 21l"b; the positive number 21l"lbl is
II t x ("r x t) II = II l' IIII l' x l' II. called the pitch of the helix (we take absolute values since we did not assume
Hence, that a or b is positive).
Let us compute the curvature of the helix using the formula in Proposition
II 1'h·'Y) - 1'(1'·i) II _ II l' x (i x 1') II 2.1. Denoting dIdO by a dot, we have
II l' 11 4 - II 'Y 11 4
1'(8) = (-asin8,acosO,b),
_ II l' JIll i x l' II
II t 11 4 ,', Ili'(B) 11= Ja 2
+ b2 •
_ II i x l' II This shows that 'Y(8) is never zero, so 'Y is regular (unless a = b = 0, in which
- II l' 11 3 . o case the image of the helix is a single point). Hence, the formula in Proposition
2,1 applies, and we have
If 'Y is a non-regular curve, its curvature is not defined in general. Note, l' = (-a cosO, -asinB, 0),
ho~ever, that the formula (1) shows that the curvature is defined at all regular l' x l' = (-ab sin B, ab cos 0, _a 2 ),
pomts of the curve (where l' is non-zero). 2 2 2 4
K = II (-absinB,abcosO,-a ) II = (a b +a )l/2 _ la l . (3)
II (-a sin 0, a cos 0, b) 11 3 (a + b )3/2
2 2 a2 + b2
Example 2.1
Thus, the curvature of the helix is constant.
A circular helix with axis the z-axis is a curve of the form Let us examine some limiting cases to see if this result agrees with what we
already know. First, suppose that b = 0 (but a i= 0). Then, the helix is simply a
'Y(B) = (acosB,asinB,bB), -00 < B < 00,
circle in the xy-plane of radius lal, so by the calculation following Definition 1.1
where a and b are constants. its curvature is 1/1al. On the other hand, the formula (3) gives the curvature
as
lal lal lal 1
a2 + 02 a2 Aa j;;j'
Next, suppose that a = 0 (but b i= 0). Then, the image of the helix is just the
z-axis, a straight line, so the curvature is zero. And (3) gives zero when a = 0
too.

EXERCISES

2.1 Compute the curvature of the following curves:

(i) 'Y(t) = (~(1 + t)3/2, ~(1 - t)3/2, -jz);
(ii) 'Y(t) = (~cos t, 1 - sin t~ - ~ cos t);
If (x, y, z) is a point on (the image of) the helix, so that (iii) 'Y(t)= (t,cosht)j
(iv) 'Y(t) = (COg 3 t, sin 3 t).
x = acosB, y = asinB, z = bB,
 28 Elementary Differential Geometry ¥ 2. How Much Does a Curve Curve? 29

For the astroid in (iv), show that the curvature tends to 00 as we

approach one of the points (±l, 0), (0, ±1). Compare with the sketch
found in Exercise 1.5. t

2.2 Show that, if the curvature K(t) of a regular curve -y(t) is > 0 every-
where, then K(t) is a smooth function of t. Give an example to show
that this may not be the case without the assumption that K > O.

n.
2.2. Plane Curves
The signed curvature has a simple geometric interpretation:
For plane curves, it is possible to refine the definition of curvature slightly and
give it an appealing geometric interpretation.
Suppose that 'Y(s) is a unit-speed curve in R 2 • Denoting d/ds by a dot, let
Proposition 2.2
Let 'Y( s) be a unit-speed plane curve, and let ip(s) be t~e angl~ throu?h which
t='Y a fixed unit vector must be rotated anti-clockwise to bnng It mto comcldence
be the tangent vector of "y; note that t is a unit vector. There are two unit with the unit tangent vector t of 'Y. Then,
vectors perpendicular to tj we make a choice by defining lls, the signed unit dtp
normal of'Y, to be the unit vector obtained by rotating t anti-clockwise by 7r/2. "'s = ds'
Note that, even though the angle tp is only determined up to adding an
t
integer multiple of 27r, the derivative dip/ds is well defined.
Thus, the signed curvature is the rate at which the tangent -:ector. ~f the
curve rotates. As the above diagrams show, the signed curvature IS posItIve or
negative according as t rotates anti-clockwise or clockwise as one moves along
the curve in the direction of increasing s.

Proof 2.2
Let a be the fixed unit vector and let b be the unit vector obtained by rotating
a anti-clockwise by 7r /2. Then,
By Proposition 1.2, i = l' is perpendicular to t, and hence parallel to fi
s- Thus,
there is a number "'s such that t = a cos ip + b sin ip,
. dip
t = (-asin<p + bcosip) ds'
The scalar K s is called the signed curvature of -y (it can be positive, negative or . . dtp
t.a = - SlO<p ds'
zero). Note that, since II lls II = 1, we have
• d<p .) (5)
(4) "'s(ns.a) = - SlOtp ds (since t = "'slls .

so the curvature of'Y is the absolute value of its signed curvature. The following But the angle between lls and a is <p + 7r /2, since t must b~ rotated anti-
diagrams show how the sign of the signed curvature is determined (in each case, clockwise by 7r /2 to bring it into coincidence with lls (see the diagram below).
the arrow on the curve indicates the direction of increasing s).
30 Elementary Differential Geometry
• 2. How Much Does a Curve Curve? 31

Hence, Proof fU

ns·a = cos (/p + i) = - sincp.

For the first part, fix So E (a, (3) and define, for any s E (a, (3),

Inserting this into Eq. (5) gives the required result. o !pes) = r k(u)du,
ls o
(cf. Proposition 2.2),

n. t
"((s) = (1: COSlp(t)dt, 1: Sinlp(t)dt) .

Then, the tangent vector of "( is

1'( 8) = (cos!p(s), sin <p(s)),
which is a unit vector making an angle lp(s) with the x-axis. Thus, "( is unit-
speed and, by Proposition 2.2, its signed curvature is

a dlp
-d
S
d
= -d
S
18

80
k(u)du = k(s).
For the second part, let 'iJ( s) be the angle between the x-axis and the unit
tangent vector ~ of,;y. Thus,
'Y ~(s) = (cos lji(s), sin 43(s)),
The next result shows that a unit-speed plane curve is essentially determined
once we know its signed curvature at each point of the curve. The meaning of
'essentially' here is 'up to a rigid motion of R 2 ,. Recall that a rigid motion of
,;yes) = (1: coslji(t)dt, 1: sin'iJ(t)dt) +,;y(so). (6)

By Proposition 2.2,
R 2 is a map M : R 2 -t R 2 of the form
d'iJ
M = TaoR(J, ds = k(s),
where R(J is an anti-clockwise rotation by an angle (J about the origin, 43(s) = r k(u)du + lji(so).
180
R(J(x, y) = (x cos (J - y sin 0, x sin e + y cos (J),
Inserting this into Eq. (6), and writing a for the constant vector ,;y(so) and e
and Ta is the translation by the vector a, for the constant scalar 'iJ(so), we get

for any vectors (x,y) and v E R

Ta(v) = v
2
.
+ a, ,;yes) = Ta. (1: coS(<p(t) +O)dt, 1: sin(!p(t) +O)dt)

= T. (coso r cos <p(t)dt - sinO JBnr sin <p(t)dt,

180
Theorem 2.1
sinO r cos <p(t)dt + cosO r Sincp(t)dt)
Let k : (or.,{3) -t R be any smooth function. Then, there is a unit-speed curve 1 80 180
"( : (a, 13) -t R 2 whose signed curvature is k.
Further, if,;y : (or.,{3) -t R 2 is any other unit-speed curve whose signed
= T.R9 (1: cos <p(t)dt, 1: Sinlp(t)dt)
curvature is k, there is a rigid motion M 0/ R 2 such that = T.R9("((S)). 0
,;yes) = M("((s)) for all s E (a,{3).
Example 2.2
Any regular plane curve whose curvature is a positive constant is part of a
 ¥
32 Elementary Differential Geometry 2. How Much Does a Curve Curve? 33

circle. To see this, let K be the (constant) curvature of the curve 1', and let Ks If te s < 0, it is easy to check that the curve
be its signed curvature. Then, by Eq. (4),
'"I(s) = (Reos i, -Rsin ~)
(which is just another parametrisation of the circle with centre the origin and
A priori, we could have K s = K at some points of the curve and K s = -K at
radius R) has signed curvature -1/ R. Thus, if R = -1/K s we again get a circle
others, but in fact this cannot happen since K s is a continuous function of 8 (d.
with signed curvature K 8 •
Exercise 2.4), so the Intermediate Value Theorem tells us that, if K a takes both
the value K and the value -K, it must take all values between. Thus, either
"-a = "- at all points of the curve, or "-a = -K at all points of the curve. In Example 2.3
particular, te s is constant. Theorem 2.1 shows that we can find a plane curve with any given smooth
The idea now is to show that, whatever the value of te s , we can find a function as its signed curvature. But simple curvatures can lead to complicated
parametrised circle whose signed curvature is "-a. The theorem then tells us curves. For example, let the signed curvature be Ks (8) = 8. Following the proof
that any curve whose signed curvature is tea can be obtained by applying a rigid of Theorem 2.1, and taking 80 = 0, we get
motion to this circle. Since rotations and translations obviously take circles to
circles, it follows that any curve whose signed curvature is constant is (part of) r.p(8) =
r
Jo udu =
S2
2'
a circle.

R is
A unit-speed parametrisation of the circle with centre the origin and radius
1'(8) = (l
a
cos c:) dt, l C:)
s
sin dt).

1'( 8)
8 R sm. R8) .
= (R cos R'
Its tangent vector

t = ""(8) = (-sin ~,cos ~)

is the unit vector making an angle 1r /2 + s/ R with the positive x-axis:

L..---'-----+---------7X These integrals cannot be evaluated in terms of 'elementary' functions.

(They arise in the theory of diffraction of light, where they are called Fres-
nel's integrals, and the curve l' is called Cornu '$ Spiml, although it was first
considered by Euler.) The picture of l' above is obtained by computing the
Hence, the signed curvature of l' is integrals numerically.

~ (% + i) = ~.
Thus, if "-8 > 0, the circle of radius 1/"-8 has signed curvature "-8'
 _._ ... _ .. _-'J _ ... _. _ .. _._ ........ .., •. , ........ )

y (i) if () :::: 0, then 'Y is part of a straight line (write 'Y :::: rt and
deduce that "'s = 0);
(ii) if () = 1r/2, then 'Y is a circle (write 'Y = rn.);
°
(iii) if < () < 7r /2, then 'Y is a logarithmic spiral (show that "-. =
-1/8 cot ()).
2.7 Let 'Y(t) be a regular plane curve and let A be a constant. The parallel
------:;;_---r--""""""'--~:c
curve 'Y A of 'Y is defined by
'YA(t) :::: 'Y(t) + Ans(t).
Show that, if IA"-s(t)1 < 1 for all values of t, then 'Y A is a regular
curve and that its signed curvature is ,,-./(1 - A,,-.).
2.8 Let 'Y be a unit-speed plane curve with nowhere zero curvature.
It is natural to ask whether Theorem 2.1 remains true if we replace 'signed Define the centre of curvature ~(s) of "Y at the point 'Y( 8) to be
curvature' by 'curvature'. The first part holds if (and only if) we assume that 1
k ~ 0, for then 'Y can be chosen to have signed curvature k and so will have ~(s) = 'Y(s) + -(-)n.(s).
"-. s
curvature k as well. The second part of Theorem 2.1, however, no longer holds. Prove that the circle with centre f(8) and radius 11/Ks (s)1 is tangent
For, we can take a (smooth) curve "Y that coincides with the x-axis for -1 ~ to 'Y at 'Y(8) and has the same curvature as "'( at that point. This
x ~ 1 (say), and is otherwise above the x-axis. (The reader who wishes to write (
circle is called the osculating circle to 'Y at the point 'Y( s). (Draw a
down such a curve explicitly will find the solution of Exercise 1.19 helpful.) We
°
nOw reflect the part of the curve with x ~ in the x-axis. The new curve has
the same curvature as 'Y (see Exercise 2.12), but obviously cannot be obtained
picture.)
2.9 With the notation in Exercise 2.8, we regard f(S) as the parametrisa-
tion of a new curve, called the evolute of "I (if "I is any regular plane
by applying a rigid motion to 'Y. See Exercise 2.13 for a version of Theorem 2.1
curve, its evolute is defined to be that of a unit-speed reparametrisa-
that is valid for curvature instead of signed curvature.
tion of "'(). Assume that ks(s) ¥- 0 for all values of s (a dot denoting
d/ds), say k s > 0 for all s (this can be achieve~ by repl~cing s by
-8 if necessary). Show that the arc-length of E 1S UQ - ~, where
EXERCISES
UQ is a constant, and calculate the signed curvature of E.

2.3 Show that, if'Y is a unit-speed plane curve, Show that the evolute of the cycloid
'Y(t) = aCt - sin t, 1 - cos t), 0 < t < 21r,

2.4 Show that the signed curvature of any regular plane curve 'Y(t) is a where a > °is a constant, is
smooth function of t. (Compare with Exercise 2.2.) E(t) ::::a(t+sint,-I+cost)
2.5 Let 'Y(t) = (e kt cos t, e kt sin t), where -00 < t < 00 and k is a non- (see Exercise 1.7) and that, after a suitable reparametrisation, f can
zero constant (a logarithmic spiral - see Example 1.4). Show that be obtained from 'Y by a translation of the plane.
there is a unique unit-speed parameter son 'Y such that 8 > 0 for all
2.10 A string of length £ is attached to the point 8 :::: 0 of a unit-speed
t and s -+ 0 as t -+ =f00 if ±k > 0, and express s as a function of t.
plane curve "1(8). Show that when the string is wound onto the curve
Show that the signed curvature of'Y is 1/ ks. Conversely, describe
while being kept taught, its endpoint traces out the curve
every curve whose signed curvature, as a function of arc-length 8, is
I/ks for some non-zero constant k. £(8) :::: 'Y(S) + (£ - shes),
2.6 A unit-speed plane curve 'Y has the property that its tangent vector where 0 < s < l and a dot denotes d/ds. The curve £ is called the
t(8) makes a fixed angle () with 'Y(s) for all 8. Show that involute of'Y (if "I is any regular plane curve, we define its involute
36 Elementary Differential Geometry * 2. How Much Doe~ a Curve Curve? 37

to be that of a unit-speed reparametrisation of 'Y). Suppose that the of R 3 , and is right-handed, i.e.
signed curvature "'8 of'Y is never zero, say 1\;8(8) > 0 for all s. Show b = t x n, n = b x t, t = n x b.
that the signed curvature of " is 1/(£ - 8).
2.11 Let 'Y be a regular plane curve. Show that b
(i) the involute of the evolute of'Y is a parallel curve of 'Yi
(ii) the evolute of the involute of'Y is 'Y.
(These statements might be compared to the fact that the integral n

of the derivative of a smooth function J is equal to J plus a constant,

while the derivative of the integral of J is J.)
2.12 Show that applying a reflection in a straight line to a plane curve
changes the sign of its signed curvature.
2.13 Show that, if two plane curves 'Y(t) and i(t) have the same non- Since b(s) is a unit vector for all 5, b is perpendicular to b. Now we use the
zero curvature for all values of t, then i can be obtained from 'Y by 'product rule' for differentiating the vector product of vector-valued functions
applying a rigid motion or a reflection in a straight line followed by u and v of a parameter s:
a rigid motion. d du dv
\ -(u x v) = -d x v + u x -d .
ds 5 5
Applying this to b =t x n gives

2.3. Space Curves b = i x n + t x 0 = t x 0, (9)

since by the definition (7) of n,
Our main interest in this book will be in curves (and surfaces) in R 3 , Le.
space curves. While a plane curve is essentially determined by its curvature i x n = II:n X n = O.
(see Theorem 2.1), this is no longer true for space curves. For example, a Equation (9) shows that b is perpendicular to t. Being perpendicular to both
circle of radius one in the xy-plane and a circular helix with a = =
b 1/2 t and b, b must be parallel to n, so
(see Example 2.1) both have curvature one everywhere, but it is obviously
impossible to change one curve into the other by any combination of rotations
b = -Tn, (10)
and translations. We shall define another type of curvature for space curves, for some scalar 1', which is called the torsion of 'Y (the minus sign is purely a
called the torsion, and we shall prove that the curvature and torsion of a curve matter of convention). Note that the torsion is only defined if the curvature is
together determine the curve up to a rigid motion (Theorem 2.3). non-zero.
Let "y( s) be a unit-speed curve in R 3 , and let t = l' be its unit tangent Of course, we define the torsion of an arbitrary regular curve 'Y to be the
vector. If the curoature 1\;(s) is non-zero, we define the principal normal of 'Y at torsion of a unit-speed reparametrisation of "y. As in the case of the curvature,
the point 'Y( 8) to be the vector to see that this makes sense, we have to show that if we make a change in the
1 . unit-speed parameter of "y of the form
n(8) = "'(s) t(8). (7)
u = ±s + c,
Since II i II = "', n is a unit vector. Further, by Proposition 1.2, t.i = 0, so t where c is a constant, then l' is unchanged. But this change of parameter has
and n are actually perpendicular unit vectors. It follows that the following effect on the vectors introduced above:
b=txn (8) t I-t ±t, i I-t i, n ~ n, b I-t ±b, b I-t b.
is a unit vector perpendicular to both t and n. The vector b(s) is called the It follows from Eq. (8) that 'T I-t 1', as required.
binormal vector of 'Y at the point "y( s). Thus, {t, n, b} is an orthonormal basis
38 Elementary Differential Geometry 2. How Much Does a Curve Curve? 39

Just as we did for the curvature in Proposition 2.1 l it is possible to give a Then,
formula for the torsion of a space curve "( in terms of"( alone, without requiring . ds,
a unit-speed reparametrisation: "(=dt"('
.. _(dS)2"
Proposition 2.3 "( - dt "( + rl?s_/
dt 2 " ,

Let "((t) be a regular curve in R' with nowhere vanishing curvature. Then, ... _ (dS)'", 3ds rl?s -" d's,
"( - dt "( + dt dt 2 ' + dt'''('
denoting d/dt by a dot, its torsion is given by
Hence,
(1' x i)·'f
Ih x il1 2 '
(~:)' "(' xi',
7= (11)
l' x i =

Note that this formula shows that 7(t) is defined at all points ,,((t) of the ... (."( x "(") = (dS)6 '" ('
curve at which its curvature I«t) is non-zero, since by Proposition 2.1 this is "(. dt "(. "( x "(") ,
the condition for the denominator on the right-hand side to be non-zero.
and so
';;.(1' x i) _ "("'.("(' xi')
o
Proof 2.3 II l' x i II' - II "(' x "(" 11 2 •
We could 'derive' Eq. (11) by imitating the proof of Proposition 2.1. But it
is easier and clearer to proceed as follows, even though this method has the Example 2.4
disadvantage that one must know the formula for 7 in Eq. (11) in advance.
We first treat the case in which "( is unit-speed. Using Eqs. (7) and (10), We compute the torsion of the circular helix

7 = -n.b = -n.(t x nr = -n.(t x n + t x ill = -n.(t x ill. "((8) = (acos8,asin8,bO)

Now, n = ~i = ~;Y, so studied in Example 2.1. We have

7 = 1.. (."( x
-K"(' d
dt (1..))
K"(
1'(8) = (-asin8,acos8,b),
i(8) = (-acos8,-asin8,0),
1.. (."( x
= --"(.
K,
(1... i< .. ))
-"( - -"(
K.,..,2
';;(8) = (a sin 8, -acos8,0).
Hence,
1 ... (. ..)
= 2"(' "( x "( ,
I<
l' x i = (absin8, -abcos8,a 2 ),
t II l' x i 11 2 = a2(a 2 + b2),
since i.('Y x i) = 0 and i·('Y x ';;) = -';;.(1' x i). This agrees with Eq. (11), for,
since "( is unit-speed, l' and i are perpendicular, so (1' x i).';; = a 2 b,
I
II l' x i II = II 1'1111 ill = II ill = 1<. I
I
and so the torsion
(1' x i).';; a2 b b
In the general case, let s be arc-length along "( and denote d/ds by a dash.
7 = II l' x i II' = a2 (a 2 + b2 ) = a2 + b2 '
Note that the torsion of the circular helix in Example 2.4 becomes zero
when b = 0, in which case the helix is just a circle in the xy-plane. This gives
I us a clue as to the geometrical interpretation of torsion, contained in
I
I
46 Elementary Differential Geometry o 2. How Much Does a Curve Curve 7 41

Proposition 2.4 Hence,

3
Let'Y be a regular curve in R with nowhere vanishing curvature (so that the iJ. ::::; b x t + b x i = -Tn x t + ",b x n::::; -d + Tb.
torsion T of'Y is defined). Then, the image of'Y is contained in a plane if and
only if T is zero at every point of the curve. Putting all this together, we get

Proof 2·4 Theorem 2.2

We can assume that 'Y is unit-speed (for this can be achieved by reparametrising Let'Y be a unit-speed ~urve in R 3 with nowhere vanishing curvature. Then,
'Y, and reparametrising changes neither the torsion nor the fact that 'Y is, or is i = ",n
not, contained in a plane). We denote the parameter of'Y by s and djds by a n = -/d +Tb (14)
dot as usual. b = -Tn.
Suppose first that the image of 'Y is contained in the plane r.a ::::; d, where
a is a constant vector and d is a constant scalar (r is the position vector of an
Equations (14) are called the Frenet-Serret equations (or sometimes the
arbitrary point of R 3 ). We can assume that a is a unit vector. Differentiating
Serret-Frenet equations). Notice that the matrix
'Y.a::::; d with respect to s, we get
t.a ::::; 0, (12) 0 ° 0) I(,

i.a ::::; 0 (since a::::; 0), (o-I(,

-'1'
T
0
Kn.a ::::; 0 (since i ::::; Kn),
which expresses 1, Ii. and b in terms of t, nand b is skew-symmetric, i.e. it is
n.a ::::; 0 (since", 1:- 0). (13)
equal to the negative of its transpose. This helps when trying to remember the
Equations (12) and (13) show that t and n are perpendicular to a. It follows equations. (The 'reason' for this skew-symmetry can be seen in Exercise 2.22.)
that b ::::; t x n is parallel to a. Since a and b are both unit vectors, and b(5) Here is a simple application of Frenet-Serret:
is a smooth (hence continuous) function of 5, we must have b(s) ::::; a for all 5
or b(s) ::::; -a for all s. In either case, b is a constant vector. But then b ::::; 0,
so T = O.
Proposition 2.5
Conversely, suppose that T ::::; 0 everywhere. By Eq. (10), b::::; 0, so b is a Let 'Y be a unit-speed curve in R 3 with constant curvature and zero torsion.
constant vector. The first part of the proof suggests that 'Y should be contained Then, 'Y is (part of) a circle.
in a plane r.b ::::; constant. We therefore consider

1 h·b) = "f.b = t.b = 0, Proof 2.5

so 'Y.b is a constant (scalar), say d. This means that 'Y is indeed contained in I This result is actually an immediate consequence of Example 2.2 and Propo-
the plane r.b = d. 0 sition 2.4, but the following proof is instructive and gives more information,
namely the centre and radius of the circle and the plane in which it lies.
There is a gap in our calculations which we would like to fill. Namely, we By the proof of Proposition 2.4, the binormal b is a constant vector and 'Y
know that, for a unit-speed curve, we have is contained in a plane perpendicular to b. Now consider
i = ",n and b = -Tn d (
ds
1) = + 1
'Y + ~n t ~i:I = 0,
(these were our definitions of nand T, respectively), but we have not computed
n. This is not difficult. Since {t, n, b} is a right-handed orthonormal basis of where we have used the fact that the curvature K is constant and the Frenet-
R 3, Serret equation
t x n = b, n x b = t, b x t = n. n= -",t + Tb = -K,t (since T = 0).
 4
42 Elementary Differential Geometry 2. How Much Does a Curve Curve 7 43

Hence, 'Y + ~ n is a constant vector I say R, and we have for ii.n and b.b. It follows that A(s) ::; 3, with equality holding if and only
1 if t = t, ii = nand b = b. Thus, if we can prove that A is constant, it will
'Y + -n = 8, (15) .y
follow in particular that t = t, i.e. that =..y, and hence that ..y(s) - 'Y(s) is a
I<
1 1 constant. But by Eqs. (16) again, this constant vector must be zero, so l' = 'Y-
II 'Y - aII = II--n
I<
II = -.
I< For the first part of the theorem, we are therefore reduced to proving that
This shows that 'Y lies on the sphere of centre a and radius 1/1<. Since the A is constant. But, using the Prenet-Serret equations,
intersection of a plane and a sphere is a circle, this completes the proof. (Note A= t.t + ii.n + b.b + H + ii.n + b.b
that the plane actually intersects the sphere in a great circle: this is because
n is parallel to the plane, so by Eq. (15) the centre a of the sphere lies in the
= I<ii.t + (-J<t + rb).n + (-rii).b + t.l<n + ii.( -I<t + rb) + b.( -rn),
plane.) 0 and this vanishes since the terms cancel in pairs.
For the second part of the theorem, we observe first that it follows from the
We conclude this chapter with the analogue of Theorem 2.1 for space curves. theory of ordinary differential equations that the equations
Recall that a rigid motion of R 3 is a rotation about the origin followed by a T=kN, (17)
translation.
N= -kT+tB, (18)
13= -tN (19)
Theorem 2.3
have a unique solution T(s), N(s), B(s) such that T(so), N(so), B(so) are the
Let 'Y( s) and 1'(s) be two unit-speed curves in R 3 with the same curvature
standard orthonormal vectors i = (1,0,0), j = (0,1,0), k = (0,0,1), respec-
1« s) > 0 and the same torsion r(s) for all s. Then, there is a rigid motion M
tively. Since the matrix
of R 3 such that
1'(s) = M('Y(s)) for all s.
Ok0 0)
-k t
Further, if k and t are smooth functions with k > 0 everywhere, there is a
(o -t 0
unit-speed curve in R 3 whose curoature is k and whose torsion is t. expressing T, Nand 13 in terms of T, N and B is skew-symmetric, it follows
that the vectors T, N and B are orthonormal for all values of s (see Exercise
Proof 2.3 2.22).
Now define
Let t, n and b be the tangent vector, principal normal and binormal of 'Y,
and let t, ii and b be those of 1'. Let So be a fixed value of the parameter 'Y(s) = /.' T(u)du.
s. Since {t(so), n(so), b(so)} and {t(so), ii(so), b(so)} are both right-handed '0

orthonormal bases of R 3 , there is a rotation about the origin of R 3 that takes Then, ..y = T, so since T is a unit vector, 'Y is unit-speed. Next, T = kN by Eq.
t(so), n(so) and b(so) to t(so), ii(so) and b(so), respectively. Further, there is (17), so since N is a unit vector, k is the curvature of'Y and N is its principal
a translation that takes 'Y(so) to 1'(so) (and this has no effect on t, n and b). normal. Next, since B is a unit vector perpendicular to T and N, B = AT x N
By applying the rotation followed by the translation, we can therefore assume where A is a smooth function of s that is equal to ±1 for all s. Since k = i x j,
that we have A(so) = 1, so it follows that A(s) = 1 for all s. Hence, B is the binormal
of'Y and by Eq. (19), t is its torsion. 0
'Y(so) = 1'(so), t(so) = t(so), n(so) = ii(so), b(so) = b(so), (16)
The trick now is to consider the expression
A(s) = t.t + ii.n + b.b.
In view of Eqs. (16), we have A(so) = 3. On the other hand, since t and tare
unit vectors, t.t :0; I, with equality holding if and only if t = t; and similarly
 Elementary Differential Geometry ¥ 2. How Much Does aCUNe CuNe i 43
44

a general helix. (Thus, Examples 2.1 and 2.4 show that a circular
EXERCISES helix is a general helix.)
2.14 Compute I'i-, 7, t, nand b for each of the following curves, and verify 2.21 Let -yet) be a unit-speed curve with ,,(t) > 0 and T(t) # 0 for all t.
that the Frenet-Serret equations are satisfied: Show that, if -y lies on the surface of a sphere, then
(i) -yet) = (~(1 + t)3/2, ~(1 - t)3/2, +,); T d(i<) (20)
;;: = ds T,,2 .
(ii) -y(t) = (~cost,l-sint,-~cost).
2.15 Show that the curve (If -y lies on the sphere of centre a and radius r, then (-y - a).( -y - a) =
2 r 2 ; now differentiate repeatedly.) Conversely, show that if Eq. (20)
1+t 1-t)
-y(t)= ( -t-,t+1'-t- holds, then

is planar. p2 + (pa)2 = r2
2.16 Show that the curve in Exercise 2.14(ii) is a circle, and find its centre, for some (positive) constant r, where p = 1/" and a = l/T, and
radius and the plane in which it lies. deduce that -y lies on a sphere of radius r. (Consider -y + pn + pab.)
2.17 Describe all curves in R 3 which have constant curvature" > 0 and Verify that Eq. (20) holds for Viviani's curve (Exercise 1.9).
constant torsion T. (Observe that it is enough to find one curve with 2.22 Let (aii) be a skew-symmetric 3 x 3 matrix (i.e. aU = -aii for
curvature K. and torsion T.) all i,j). Let v" V2 and v, be smooth functions of a parameter s
2.18 Show that the torsion of a regular curve -yet) is a smooth function satisfying the differential equations
of t whenever it is defined. ,
2.19 Let -yet) be a unit-speed curve in R 3 , and assume that its curvature Vi::::: L aijVj,
j=1
,,(t) is non-zero for all t. Define a new curve 6 by
for i = 1,2 and 3, and suppose that for some parameter value So
6(t) = d-y(t) . the vectors v,(so), V2(SO) and v,(so) are orthonormal. Show that
dt
the vectors v,(s), V2(S) and v,(s) are orthonormal for all values of
Show that 6 is regular and that, if s is an arc-length parameter for s. (Find a system of first order differential equations satisfied by the
6, then dot products vi,vi, and use the fact that such a system has a unique
ds solution with given initial conditions.)
dt = ".
Prove that the curvature of 6 is

( 1+
T
K,2
2
)! 1
For the remainder of this book,
all parametrised curves will be assumed to be regular.

and find a formula for the torsion of 6 in terms of ", T and their
derivatives with respect to t.
2.20 A regular curve -y in R 3 with curvature> 0 is called a general
helix if its tangent vector makes a fixed angle 0 with a fixed unit
vector a. Show that the torsion T and curvature" of -y are related
by T = ±" cot O. (Assume that -y is unit-speed and show that a =
tcosO + bsinO.)
Show conversely that, if the torsion and curvature of a regular curve
are related by T = A" where A is a constant, then the curve is
$

3
Global Properties of Curves

All the properties of curves that we have discussed so far are 'local': they
depend only on the behaviour of a curve near a given point, and not on the
'global' shape of the curve. In this chapter, we discuss some global results about
curves. The most famous, and perhaps the oldest, of these is the 'isoperimetric
inequality', which relates the length of certain 'closed' curves to the area they
contain.

3.1. Simple Closed Curves

Our first task is to describe the kind of curves that we shall be considering in
this chapter, namely 'simple closed curves'. Intuitively, these are curves that
'join up', but do not otherwise self-intersect. A precise definition is as follows:

Definition 3.1
Let a E R be a positive constant. A simple closed curve in R 2 with period a is
a (regular) curve'"( : R -+ R 2 such that
'"((t) = '"((t') if and only if t' - t = ka for some integer k.

Thus, the point '"((t) returns to its starting point when t increases by a, but
not before that.

47
 ¥
Elementary Differential Geometry 3. Global Properties of Curves 49
48

simple closed curve non-simple closed curves

It is a standard, but highly non-trivial, result of the topology of R 2 , called

the Jordan Curve Theorem, that any simple closed cur~e in thellane has an Since every point in the image of a simple closed curve "( of period a is
'interior' and an 'exterior': more precisely, the set of pomts of R that are not traced out as the parameter t of')' varies through any interval of length a, e.g.
on the curve 'Y is the disjoint union of two subsets of R 2 , denoted by int(-y) o :S t :S a, it is reasonable to define the length of"( to be
and ext(-y), with the following properties:
(i) int('"f') is bounded, i.e. it is contained inside a circle of sufficiently large £(')') = La II -yet) II dt, (1)
radius;
where a dot denotes the derivative ,with respect to the parameter of the curve 'Y.
(ii) ext('Y) is unbounded; Since "( is regular, it has a unit-speed reparametrisation .y with the arc-length
(iii) both of the regions int{')') and exth') are connected, i.e. they have the
property that any two points in the same region can be joined by a curve
contained entirely in the region (but any curve joining a point of int('"f') to
s = l t
111'(11.) II du

a point of ext('Y) must cross the curve 'Y). of "( as its parameter (so that .y( s) :::: 'Y( t)). Note that

s{t + a) = 1t+
a
r rt+ a
Ih(u) II du = J II 1'{u) Ildu+ 1 II 1'{u) Ildu:::: l{"() + set),
Example 3.1 0 o a
since, putting v = u - a and using "(11. - a) :::: 'Y(U), we get
The parametrised circle
rt+a 111'(11.) II r t

'Y(t) = (cos (2:t) (2:t))

,sin Ja du:::: 1 1I1'(v) II
0
dv = set).

Hence,
is a simple closed curve with period a. The interior and exterior of'Y are, of
course, given by {(x,y) E R21x2 +y2 < I} and {(x,y) E R 2 ]X 2 +y2 > I}, .y(s(t)) = -reset')) <=> "(t) = ')'(t') <=> t' - t = ka <=> set') - set) = kf("(),
respectively. where k is an integer. This shows that .y is a simple closed curve with period
f(-y). Note that, since .y is unit-speed, this is also the length of.y. In short, we
Not all examples of simple closed curves have such an obvious interior and
can always assume that a simple closed curve is unit-speed and that its period
exterior, however. Is the point P in the interior or the exterior of the simple is equal to its length.
closed curve shown at the top of the next page?
We shall usually assume that our simple closed curves 'Yare positively-
oriented. This means that the signed unit normal n s of 'Y (see Section 2.2)
points into int("() at every point of "(. This can always be achieved by replacing
the parameter t of"( by -t, if necessary.
50 Elementary Differential Geometry
• 3. Global Properties of Curves 51

Note that, although the formula in Eq. (3) involves the parameter t of
'Y, it is clear from the definition (2) of A(int('Y)) that it is unchanged if 'Y is
reparametrised.

EXERCISES
t 3.1 Show that the length t('Y) and the area A(int('Y)) are unchanged by
applying a rigid motion to 'Y (see Section 2.2).
3.2 Show that the ellipse
1(t) == (acost,bsint),
positively-oriented not positively-oriented where a and b are positive constants, is a simple closed curve and
compute the area of its interior.
In the above diagrams, the arrow indicates the direction of increasing parame-
3.3 Show that the lima<;on
ter. Is the simple closed curve shown at the top of the previous page positively-
oriented? 'Y(t) = «1 + 2cost) cost, (1 +2cost)sint)
In the next section, we shall be interested in the area contained by a simple is a (regular) curve such that 'Y(t + 211") == 'Y(t) for all values of t, but
closed curve 1, i.e. that "y is not a simple closed curve.
A(int('Y)) = j. r
Jint('Y)
dxdy. (2) 3.4 Show that, if 'Y(t) is a simple closed curve of period a, and t,
and
are its unit tangent vector, signed unit normal and signed curvature,
Ua "'a
This can be computed by using Green's Theorem, which says that, for all respectively, then
smooth functions f(x, y) and g(x, y) (i.e. functions with continuous partial t(t + a) == t(t), ns(t + a) :;::: ns(t), /\'s(t + a) :;::: "'a(t).
derivatives of all orders),
(Differentiate the equation 'Y{t + a) :;::: 'Y(t).)
r (0ox
j Jint('Y) 9 _ Of) dxdy
oy
= rf(x,y)dx + g(x,y)dy,
J'Y
if 1 is a positively-oriented simple closed curve.
3.2. The Isoperimetric Inequality
Proposition 3.1 The most important global result about plane curves is
2
If -y(t) = (x(t),y(t)) is a positively-oriented simple closed curve in R with
period a, then Theorem 3.1 (The Iso peri metric Inequality)
A(int('Y)) =! r (xi; -
2 Jo
yx)dt. (3) Let'Y be a simple closed curve, let f(-y) be its length and let A(int('Y)) be the
area of its interior. Then,
Proof 3.1
A(int('Y)) ::; 4~ f('Y)2,
Taking f = -h, 9 = ~x in Green's theorem, we get with equality holding if and only if'Y is a circle.
A(int('Y)) ==! ( xdy - ydx,
2 J'Y Of course, it is obvious that equality holds when 'Y is a circle, since in that
which gives Eq. (3) immediately. o case t('Y) :;::: 211" Rand A(int('Y)) :;::: 11" R 2 , where R is the radius of the circle.
52 Elementary Differential Geometry ¥ 3. Global Properties of Curves 53

To prove this theorem, we need a result from analysis called WirtingerJs since (dx/ds)2 + (dy/ds)2 = 1. Further, by Eq. (3), we have
Inequality: 1 r 1
A(int('Y)) :::: 2" 1 (xy - yx)dt = 21
r r 0dt.
.
2
(6)
0 0

Proposition 3.2 To prove Theorem 3.1, we have to show that

Let F: [O,rr] -+ R be a smooth function such that F(O) :::: F(1I') :::: O. Then, £~~2 _ A(int('Y)) ~ 0,
l (a::)
1f

2 dt ~ fo
1f

F(t)2dt, with equality holding if and only if'Y is a circle. By Eq. (5),

with equality holding if and only if F(t) :::: A sin t for all t E [0, rr], where A is
10
r (r2 + r 2iP)dt :::: £("()2 . 11'
a constant.
Hence, using Eq. (6),
Assuming this result for the moment, we show how to deduce the isoperi-
metric inequality from it.
f..~~2 _ A(int{'Y)) :::: ~ l 1f

{f 2 + r 2 (P)dt - ~ fo
1f
2
r iJdt = ~I,
where
1f
Proof 3.1
I = 10 (r 2 + r 2iJ2 - 2r 2iJ)dt. (7)
We start by making some assumptions about 'Y that will simplify the proof.
First, we can if we wish assume that 'Y is parametrised by arc-length s. However, Thus, to prove Theorem 3.1, we have to show that I ~ 0, and that I ;;;;; 0 if
because ofthe 11' that appears in Theorem 3.1, it turns out to be more convenient and only if 'Y is a circle.
to assume that the period of 'Y is 11'. If we change the parameter of 'Y from s to
t :::: 1I'sjf..('Y), (4)

the resulting curve is still simple closed, and has period 11' because when
s increases by f..(o"() , t increases by 11'. We shall therefore assume that 'Y is
parametrised using the parameter t in Eq. (4) from now on.
For the second simplification, we note that both f..('Y) and A('Y) are un-
changed if -y is subjected to a translation 'Y(t) f-t -yet) + h, where h is any
constant vector (see Exercise 3.1). Taking h :::: -'Y(O), we might as well assume
that 'Y(O) :::: 0 to begin with, Le. we assume that 'Y begins and ends at the
origin.
To prove Theorem 3.1, we shall calculate f('Y) and A(int("()) by using polar
coordinates
X :::: r cos 8, Y = r sin 8.
Using the chain rule, it is easy to show that
i:2 + ii = f2 + r 2 iP, xy - yx = r 2 iJ,
with d/dt denoted by a dot. Then, using Eq. (4),
By simple algebra,

f2 + r 2 iJ2:::: (~:Y + (~~y = ((~:r + (~~y) (~:Y e~r, (5) I = 10

1r

r 2(8 - 1)2dt + l 1f

(r 2 - r 2 )dt. (8)
 ¢
Elementary Differential Geometry 3, Global Properties of Curves 55
54

(where a and bare positive constants), prove that

The first integral on the right-hand side of Eq. (8) is obviously 2: 0, and the
second integral is ~ 0 by Wirtinger's inequality (we are taking F == r: note {2"
that r(O) == r(rr) == 0 since 1'(0) = 1'(rr) == 0). Hence, I ~ O. Further, since
1 J a 2 sin2 t + b2 cos 2 t dt 2: 2rr~,
0
both integrals on the right-hand side of Eq. (8) are ~ 0, their sum I is zero if with equality holding 'if and only if a == b (see Exercise 3.2).
and only if both of these integrals are zero. But the first integral is zero only if
iJ == 1 for all t, and the second is zero only if r == A sin t for some constant A (by
Wirtingeragain). SaO == t+o, where ois a constant, and hence r == Asin(O-a).
It is easy to see that this is the polar equation of a circle of diameter A, thus 3.3. The Four Vertex Theorem
completing the proof of Theorem 3.1 (see the diagram above). 0
We conclude this chapter with a famous result about convex curves in the
We now prove Wirtinger's inequality. plane. A simple closed curve l' is called convex if its interior int(1') is convex, in
Let G(t) == F{t)/ sin t. Then, denoting d/dt by a dot as usual, the usual sense that the straight line segment joining any two points of int('Y)
is contained entirely in int('Y).
l 1f

P2dt == 1'1r (0 sin t + G cos t)2dt

= 11'( 0 2 sin 2 tdt + 2 !art GGsintcostdt+ 11'( G 2 cos2 tdt.
Integrating by parts:

2 LGO
Jr

sin t cos t dt == G2 sin t cos tl~ - 11f

2 2
G (cos t - sin t)dt
2

== 1" G2 (sin 2 t - cos t)dt.

2

convex not convex

Hence,
lf

111: P2 dt== 1 L G2(sin 11'( G

1f
2 2 2 2
02 sin 2 tdt+ t-cos t)dt+ cos tdt Definition 3.2

== 1" (G 2 + 0 2 ) sin 2 t dt == 1 10
1r
2
p dt +
1f
2 2
sin t dt, A vertex of a curve 1'(t) in R 2 is a point where its signed curvature K.s has a
stationary point, Le. where dK.s/dt == O.
and so
= 1" 0
It is easy to see that this definition is independent of the parametrisation
11'( P dt - LI< P
2 2
dt
2 2
sin t dt. of l' (see Exercise 3.7).
The integral on the right-hand side is obviously 2: 0, and it is zero if and only
if (; == 0 for all t, i.e. if and only if G(t) is equal to a constant, say A, for all t. Example 3.2
Then, F{t) = A sin t, as required. 0 The ellipse "Y(t) == (a cos t, b sin t), where a and b are positive constants, is
a convex simple closed curve with period 271" (see Exercises 3.2 and 3.6). Its
signed curvature is easily found to be
EXERCISES ab
K.s(t) == 2 .
3.5 By applying the isoperimetric inequality to the ellipse (a 2 sin t + b2 cos 2 t)3/2
x2 y2
a 2 + b2 == 1
55 Elementary Differential Geometry p 3. Global Properties of Curves 57

Then,
dl';,s 3ab(b 2 - a 2 ) sin t cos t
dt - (a 2 sin 2 t + b2 cos 2 t)5/2
vanishes at exactly four points of the ellipse, namely the points with t
0, 7r/2, 1r and 37r/2, which are the ends of the two axes of the ellipse.

The following theorem says that this is the smallest number of vertices a
convex simple closed curve can have.

Theorem 3.2 (Four Vertex Theorem) Suppose that P and Q are the only vertices of "Y. Since "Y is convex, the
straight line joining P and Q divides "Y into two segments, and since there are
Every convex simple closed curve in R 2 has at least four vertices.
no other vertices, we must have K,s > 0 on one segment and k s < a on the
other. But then the integrand on the left-hand side of Eq. (10) is either always
It is actually the case that this theorem remains true without the assump-
> a or always < a (except at P and Q where it vanishes), so the integral is
tion of convexity, but the proof is then more difficult than the one we are about
"definitely> 0 or < 0, a contradiction.
to give.
Hence, there must be at least one more vertex, say R. If there are no other
vertices, the points P, Q and R divide "Y into three segments, on each of which
Proof 3.2 k s is either always> 0 or always < O. But then k s must have the same sign on
two adjacent segments. Hence, there is a straight line that divides "Y into two
We might as well assume that the curve "Y(t) is unit-speed, so that its period
segments, on one of which k s is always positive, and on the other K,s is always
is the length eof "Y. We consider the integral
< O. The argument in the preceding paragraph shows that this is impossible.
1 i
ks(th(t) dt,
So there must be a fourth vertex. 0

where a dot denotes d/ dt. (Recall from Exercise 2.4 that K s is a smooth function
oft.) Integrating by parts, and using the equation D. s = -Kat (see Exercise 2.3), EXERCISES
we get
3.6 Show that the ellipse in Exercise 3.2 is convex. (You may need to use
1 i
ks"Ydt = -1 i
"'s'Ydt = -1 i
Kst dt = 1 i
D. a dt = na(e) - ns(O) = O. (9) the inequality 2XIX2 :::; xi + x~.)
3.7 Show that the definition of a vertex of a plane curve is independent
Now, K s attains all of its values on the closed interval [O,f], so K s must of its parametrisation.
attain its maximum and minimum values at some points P and Q of "Y, say.
3.8 Show that the limacQn in Exercise 3.3 has only two vertices.
We can assume that P i- Q, since otherwise "'a
would be constant, "Y would be
a circle (by Example 2.2), and every point of"Y would be a vertex. Let a be
a unit vector parallel to the vector PQ, and let b be the vector obtained by
rotating a anti-clockwise by 1T /2. Taking the dot product of the integral in Eq.
(9) with the constant vector b gives

1
i
ksb·b) dt = O. (10)
F

4
Surfaces in Three Dimensions

In this chapter, we introduce several different ways to formulate mathematically

the notion of a surface. Although the simplest of these, that of a surface patch,
is all that is needed for most of the book, it does not describe adequately most
of the objects that we would want to call surfaces. For example, a sphere is not
a surface patch, but it can be described by gluing two surface patches together
suitably. The idea behind this gluing procedure is simple enough, but making
it precise turns out to be a little complicated. We have tried to minimise the
trauma by collecting the most demanding proofs in a section at the end of the
chapter; this section is not used anywhere else in the book and can safely be
omitted if desired. In fact, surfaces (as opposed to surface patches) will be used
in a serious wayan only a few occasions in this book.

4.1. What is a Surface?

A surface is a subset of R 3 that looks like a piece of R 2 in the vicinity of any
given point, just as the surface of the Earth, although actually nearly spherical,
appears to be a flat plane to an observer on the surface who sees only to the
horizon. To make the phrases 'looks like' and 'in the vicinity' precise, we must
first introduce some preliminary material. We describe this for R n for any
n 2: 1, although we shall need it only for n = 1,2 or 3.
First, a subset U of R n is called open if, whenever a is a point in U, there
is a positive number < such that every point u E R n within a distance < of a is
59
 p
60 Elementary Differential Geometry 4. Surfaces in Thrp.e Dimensions 61

also in U: Example 4·1

a E U and II u - a II < to ===} u E U. Every plane in R 3 is a surface with an atlas containing a single surface patch.
In fact, let a be a point on the plane, and let p and q be two unit vectors that
For example, the whole of R n is an open set, as is
are parallel to the plane and perpendicular to each other. Then, any vector
Vr(a)={uERn Illu-all<r}, parallel to the plane is a linear combination of p and q, say up + vq for some
the open ball with centre a and radius r > O. (If n = 1, an open ball is called scalars u and v. If r is the position vector of any point of the plane, r - a is
an open interval; if n ::= 2 it is called an open disc.) However, parallel to the plane, so
r - a= up+vq,
Vr(a)={uERn Illu-al1:~r}
r = a+up +vq,
is not open, because however small the positive number to is, there is a point
for some scalars u and v. Thus, the desired surface patch is
within a distance to of the point (al + r, a2, ... ,an) E 15" (say) that is not in
Vr(a) (e.g. the point (al + r + ~,a2,"" an))' u(u,v) = a + up + vq,
Next, if X and Y are subsets of R m and R n , respectively, a map f : X ---t Y and its inverse map is
is said to be continuous at a point a E X if points in X near a are mapped by
f onto points in Y near f(a). More precisely, f is continuous at a if, given any u- 1 (r) = ((r - a).p, (r - a).q).
number to > 0, there is a number 8 > 0 such that These formulas make it clear that u and u- 1 are continuous, and hence that u
is a homeomorphism. (We shall not verify this in detaiL)
u E X and II u - a II < 8 ===} II f(u) - f(a) II < E.
Then f is said to be continuous if it is continuous at every point of X. Com- The next example shows why we have to consider surfaces, and not just
posites of continuous maps are continuous. surface patches.
In view of the definition of an open set, this is equivalent to the following:
f is continuous if and only if, for any open set V of R n, there is an open set U
Example ..{ 2
of R m such that f maps un X into V n Y.
If f : X ---t Y is continuous and bijective, and ifits inverse map f- 1 : Y ---t X The unit sphere
is also continuous, then f is called a homeomorphism and X and Yare said to S2={(X,y,Z)ER3 1 x 2 +y2+ z 2=1}
be homeomorphic.
is a surface. The most obvious pararnetrisation is probably that given by lati-
We are now in a position to make our first attempt at defining the notion
tude () and longitude y>:
of a surface in R 3 .
u ((), y» = (cos () cos y>, cos () sin y>, sin ()).
Definition 4.1
A subset 5 of R 3 is a surface if, for every point P E 5, there is an open set U in
R 2 and an open set W in R 3 containing P such that 5 n W is homeomorphic
to U.

Thus, a surface comes equipped with a collection of homeomorphisms u ;

U ---t 5 n W, which we call surface patches or parametrisations. The collection
of all these surface patches is called the atlas of S. Every point of S lies in
the image of at least one surface patch in the atlas of 5. The reason for this
terminology will become clear from the following examples.
x
 52 Elementary Differential Geometry
p 4. Surfaces In Three DimenSions 53

Without some restriction on (e, ip), (J is not injective (and so is not a homeo-
morphism). To cover the whole sphere, it is clearly sufficient to take

- ~2 -< () -< ~2' 0 <_ '".,.. <_ 21r.

However, the set of points (B, 'P) satisfying these inequalities is not an open
2
subset of R , and so cannot be used as the domain of a surface patch. The
largest open set consistent with the above inequalities is

U = { «(}, 'P) I - ~ < (} < i, 0 < 'P < 21T} ,

but now the image of tT : U -t R 3 is not the whole of the sphere, but rather
the complement of the great semi-circle C consisting of the points of the sphere
of the form (x, 0, z) with x ~ O. Hence, (T : U -t R 3 covers only a 'patch' of the
It is intuitively obvious, although not quite trivial to prove, that the sphere
sphere. Again, we shall not verify in detail that 11 is a homeomorphism from U
cannot be covered by a single surface patch (see Exercise 4.5).
to the intersection of the sphere with the open set
W={(x,y,z)E R3 1 x<Ooryt-O}. Our last example (for the moment) is a subset of R 3 that is nearly, but not
quite, a surface.

Example 4.3

We consider the double cone

S={(X,y,Z)ER 3 1 x 2 +y2=Z2}.

To show that the sphere is a surface, we must therefore produce at least one
more surface patch covering the part ofthe sphere omitted by tT. For example,
let iT be the patch obtained by first rotating tr by 1r about the z-axis and then
by 1r /2 about the x-axis. Explicitly, iT : U ---) R 3 is given by
iT((}, ip) = (- cos () cos 'P, - sin (}, - cos Bsin 'P)
(the open set U is the same as for (T). The image of iT is the complement of the
great semi-circle C consisting of the points of the sphere of the form (x, y, 0)
with x ~ 0 (see the diagram at the top of the next page).
It is clear that C and Cdo not intersect, so the union of the images of 11 and
if is the whole sphere. Note that most points of the sphere are in the images of
both surface patches. To see that this is not a surface, suppose that (T : U -t S n W is a surface
patch containing the vertex (0,0,0) of the cone, and let a E U correspond to
64 Elementary Differential Geometry p 4. Surfaces In r hree DimenSions 65

the vertex. We can assume that U is an open ball with centre a, since any and V t; U, respectively. The composite homeomorphism tr- 1 0 jj : V -+ V is
open set U containing a must contain such an open ball. The open set W must called the transition map from tr to iI. If we denote this map by P, we have
obviously contain a point p in the lower half S_ of S where z < 0 and a point u(u, v) = a(<f>(u, v))
q in the upper half S+ where z > 0; let band c be the corresponding points
in U. It is clear that there is a curve 1l' in U passing through band c, but not for all (u,v) E V.
passing through a. This is mapped by tr into the curve "y = tr 0 1l' lying entirely
in S, passing through p and q, and not passing through the vertex. (It is true
that "y will in general only be continuous, and not smooth, but this does not EXERCISES
affect the argument.) This is clearly impossible. (The reader versed in point set
topology will be able to make this heuristic argument rigorous.) 4.1 Show that an open disc in the xy-plane is a surface.
4.2 Show that the circular cylinder
S = {(x, y, z) E R I x 2 + y2 = 1}
3

\
,,
, can be covered by a single surface patch, and so is a surface. (Take U
/ ',,(' C
I
, to be an annulus.)
I
,,
I • a
I
,, 4.3 Define surface patches O'i : U ~ R 3 for the unit sphere by solving
,, the equation x 2 + y2 + z2 = 1 for x in terms of y and z:
,,
\

\\ b ,
a±(u,v) = (±Jl - u 2 - v 2 , U, v),
defined on the open set U = {(u,v) E R 2 I u 2 + v 2 < 1}. Define tr~
and u± similarly (with the same U) by solving for y and z, respec-
If we remove the vertex, however, we do get a surface S_ uS+. It has an atlas tively. Show that these six patches give the sphere the structure of a
consisting of the two surface patches tr± : U ~ R 3 , where U = R 2 \{(o, On, surface.
given by the inverse of projection onto the xy plane:
tr±(u, v) = (u, v, ±Ju2 + v 2 ).

As the example of the sphere shows, a point a of a surface S will generally lie
in the image of more than one surface patch. Suppose then that tr : U ~ S n W

and u: U ~ SnW are two patches such that a E SnWnw. Since tr and u are
homeomorphisms, 0'-1 (S n WnW) and 1
u-
(S n WnW) are open sets V ~ U
 F 4. Surfaces In Three DimenSions 61
66 Elementary Differential Geometry

4.4 The hyperboloid of one sheet is computed componentwise. For example, if m = 2 and n = 3, and
S = {(x,y,z) E R3 I x 2 +y2 - Z2 = I}. f(u,v) = (h(u,v),h(u,v),13(u,v»,

(A picture of the hyperboloid can be found in Proposition 4.6.) Show then

that, for every (), the straight line 8f _ (8lI 8h 8h) 8f = (8II, 8lz, 813),
~- ~'~'~ , ~ & & &
(x - z) cos () = (1 - y) sin 0, (x + z) sin 0 = (1 + y) cos () and similarly for higher derivatives. We often use the following abbreviations:
is contained in S, and that every point of the hyperboloid lies on one 8f 8f
of these lines. Deduce that S can be covered by a single surface patch, 8u :::: f u , 8v = f v ,
and hence is a surface. (Compare the case of the cylinder in Exercise 82f 82f 82 f 82 f
4.2.) 8u 2 = f uu , 8u8v = 8v8u =
f uv , 8v2:::: f vv ,
f vu ,
Find a second family of straight lines on S, and show that no two and so on. Note that f uv = f vu , because all the partial derivatives of the com~
lines of the same family intersect, while every line of the first family ponents of f are continuous.
3
intersects every line of the second family with one exception. One It now makes sense to say that the surface patches u : U ---+ R in the atlas
says that the surface S is doubly ruled. of a surface S are smooth. But we shall require one further condition.

Definition 4.2
A surface patch u : U ---+ R 3 is called T'egular if it is smooth and the vectors (J u
and U v are linearly independent at all points (u, v) E U. Equivalently, u should
be smooth and the vector product Uu X (Iv should be non~zero at every point
ofU.

We can finally define the class of surfaces to be studied in this book.

Definition 4.3
A smooth surface is a surface u whose atlas consists of regular surface patches.
4.5 Show that the unit sphere cannot be covered by a single surface patch.
(This requires some point set topology.)
Example 4.4
The plane in Example 4.1 is a smooth surface. For
4.2. Smooth Surfaces u(u,v) =a+up+vq
is clearly smooth and u u = P and u v = q are linearly independent because p
In Differential Geometry we use calculus to analyse surfaces (and other geomet~ and q were chosen to be perpendicular unit vectors.
ric objects). We must be able to make sense of the statement that a function
on a surface is differentiable, for example. For this, we have to consider surfaces
Example 4.5
with some extra structure.
First, if U is an open subset of R m , we say that a map f : U ---+ R n is For the unit sphere 8 2 in Example 4.2, it is again clear that u and 17 are smooth.
smooth if each of the n components of f, which are functions U ---+ R, have As for regularity, we compute
continuous partial derivatives of all orders. The partial derivatives of f are then Uo = (-sinOcoslp,-sin(}sinlp,cosO), u'" = (-cos()sinlp,cos(}cOSlp,O),
----!!6B'!:==================:!E~Ie~m~e~nr~a~ryE'!b~lf~fe~r~en~I~la~i~G~e~o~m~e~lr~y---'?"--,!4:.:!3~d1~iJi!'C:l!@,!§~iI~i:'i~ii~'E~@!,:b~iii,!,:e~il~S~Ib,!!ii!S::::::::::::::::~~~~~~~~~~~~~--

which gives The scalar on the right-hand side of this equation is the determinant of the
jacobian matrix
17, X17 ~ = (- cos' 0 cos 'P, - cos' 0 sin '1', - sin 0 cos 0) au
and hence 1117, x u~ 11= IcosOI. But if (0,'1') E U, then -,,/2 < 0 < ,,/2 so J(P) = ( g~
au
cos 0 i- O. Similarly, one checks that if is regular.
of P. We recall from calculus that, if 1f! and oP ';e two maps between open sets
In Exercise 4.3 we gave another family of surface patches covering the unit in R2 ,
sphere 8', and it is easy to check that they are regular (see Exercise 4.7). J(oP 0 1f!) = J(oP)J(1f!).
Together with Example 4.5, this gives two atlases for 8 2 consisting of regular
(In fact, this is equivalent to the chain rule that expresses the first partial
surface patches, and an obvious question is: which atlas should we use to study
derivatives of oP 0 1f! in terms of those of oP and 1f!.) Taking 1f! = P and oP =
the sphere? The answer is that we can use either, or both. For the eight patches
in Exercise 4.3 and Example 4.5 together form a third atlas. In most situations
P-" we see that J(p- 1) = J(p)-l. In particular, J(p) is invertible, so its
determinant is non-zero and Eq. (1) shows that if is regular. 0
(although not in all - see Definition 4.5), one might as well use the maximal
atlas for a given surface 8 consisting of all the regular surface patches 17 : U --+
If regular surface patches 17 and if are related as in this proposition, we say
8 n W, with U and W being open subsets of R' and R 3 , respectively. Such
that iT is a reparametrisation of u, and that ~ is a reparametrisation map. Nate
surface patches are called allowable surface patches for 8. The maximal atlas
that tT is then a reparametrisation of i1, since t1 = is 0 p-l .
is independent of any arbitrary choices.
Note also that, if 17 : U --+ 8 n Wand if : U --+ 8 n Ware two allowable
Aithough not at first sight very interesting, the next two results are very
surface patches of a smooth surface 8, and if V ~ U and V ~ U are the open
important for what is to follow.
subsets such that 17(V) = if(V) = 8 n wnw, then P = 17- 1 0 if ; V --+ V is
bijective, smooth and has a smooth inverse by Proposition 4.1. Thus, if is a
Proposition 4.1 reparametrisation 0/17 where they are both defined.
The transition maps of a smooth surface are smooth. These observations give rise to a very important principle that we shall use
throughout the book. The principie is that we can define a property 0/ any
The proof of this will be given in Section 4.7. The next result is a kind of smooth sur/ace provided we can define it for any regular sur/ace patch in such
converse. a way that it is unchanged when the patch is reparametrised.
As an illustration of this principle, let us define what is meant by a smooth
map / ; 5, --+ 8 2 , where 8 , and 8 2 are smooth surfaces. By our general
Proposition 4.2
principle, we can assume that 5, and 8 2 are covered by single surface patches
Let U and U be_open subsets 0/ R 2 and let 17 : U --+ R 3 be a regular sur/ace 171 : U, --+ R 3 and 172 : U, --+ R 3 provided we verify that the definition we
patch. Let P : U --+ U be a bijective smooth map with smooth inverse map give is unaffected by a reparametrisation of 0"1 and 0"2. Since 0"1 and 0"2 are
p- 1 : U --+ U. Then, if = u 0 P : U --+ U is a regular sur/ace patch. bijective, any map f : 51 -+ 52 gives rise to the map 0'2"1 0 f 00'1 : V 1 -+ V 2 ,
and we say that / is smooth if this map is smooth (we already know what
it means for a map between open subsets of R 2 to be smooth). Now suppose
Proo/4·2
that if, ; U, --+ R 3 and if 2 : U2 --+ R 3 are reparametrisations of 171 and 172,
The patch if is smooth because any composite of smooth maps is smooth. As with reparametrisation maps P, : U, --+ U, and P2 : U2 --+ U2, respectively. We
for regularity, let (u,v) = P(ii,v). By the chain rule, have to show that the corresponding map if,' 0 / 0 if , : U, --+ U2 is smooth if
_au av _au av 0"2"1 0 fOO'l : U1 -+ U2 is smooth. But this is true, since
Uu = 8r/'u + 8fJ. UVl Uti = 8v t1u + 8fJtTVl if,' 0/0 if , = if,' 0 (172 017,') 0/0 (U1 0 U,') 0 if ,
so
--1
= ( (72 00'2 ) 00'2
(-1 0 / 00"1 ) 0 (-1
0'1 -)
00"1
_ _v (auav auav) 4','
au x a = 8ft8fJ - au au au x Uv' (1) = 0 (17,' 0 fOUl) 0 P"
70 Elementary Differential Geometry p 4. Surfaces in Three Dimensions 71

and tP] and P2"l are smooth maps (between open subsets of R 2 ). The reader We wrap the strip around t.he cylinder by wrapping the line z = v parallel to
should check that composites of smooth maps bet.ween surfaces arc smooth. the y-axis around the 'waist' of the cylinder at height v above the xy-plane.
We shall be especially int.erest.ed in smooth maps f ; 51 --+ 52 which are Since the widt.h of the strip is equal t.o the circumference of the cylinder, a
bijective and whose inverse map 1-1 : 52 --+ 51 is smooth. Such maps are point in the strip with y-coordinate u will go to a point on the cylinder with
called dijJeomorphisms, and SI and S2 are said to be diffeomorphic if there is polar angle u. Putting t.his together, we see t.hat the wrapping map takes the
a diffeomorphism between them. The following observation will be useful. point (O,u,v) of the strip to the point (cosu,sinu,v) of the cylinder, i.e. in the
notation above,
Proposition 4.3 f(O,u,v) = (cosu,sinu,v).
Let I : SI --+ 52 be a diffeomorphism. If tTl is an allowable surface patch on From this, it is clear that the expression for f in terms of parameters is simply
51, then f 0 tTl is an allowable surface patch on 52 .
F(u, v) = (u, v),
since
Proof 4.3

We can assume that 51 and 52 are covered by single allowable patches tTl ; Ul --+
R 3 and tT2 ; U2 --+ R 3 , respectively. Since I is a diffeomorphism, 1(0'1 (u,v)) =
tT2(F(u,v)), where F : U l --+ U2 is bijective, smooth and F-l is smooth. The We conclude this section with a general construction of smooth surfaces. In
result now follows from Proposition 4.2. 0 fact, it can be shown that every smooth surface arises in this way.
As we have seen in the above examples, surfaces are often given to us as
Example 4.6 level surfaces {(x,y,z) E R 3 I f(x,y,z) = c}, where f is a smooth function
and c is a constant (of course, we can always assume that c = 0 by replacing f
We consider the map which wraps the plane onto the circular cylinder of radius by f - c). For example, the unit sphere 8 2 is the level surface x 2+y2 + z2 = 1. In
1 and axis the z-axis, which we pararnetrise by 0'2 : U --+ R 3 , where these examples, we constructed atlases by fairly ad hoc methods. The following
tT2(U,V) = (cosu,sinu,v), U = {(u,v) E R 2 I 0 < u < 21T}. result gives conditions under which a level surface is a smooth surface. In fact,
it deals with a slightly more general situation in which different regions of a
If we wrap the whole plane onto the cylinder, we shall not get a bijective map, surface may be defined by different functions.
since the plane will wrap around infinitely many times. So we consider the
infinite strip in the yz-plane of width 27f pararnetrised by 0'1 : U --+ R 3 , where
Theorem 4.1
O'l(U,V) = (O,u,v).
Let 5 be a subset ofR3 with the following property: for each point PES, there
is an open subset W of R 3 containing P and a smooth function f : W --+ R
such that
(i) snw={(x,y,z)EW I f(x,y,z)=O};
(ii) the partial derivatives fx, f v and fz do not all vanish at P.
Then, 5 is a smooth surface.

We postpone the proof to Section 4.7.

Example 4.7

For the unit sphere 8 2 , we can take W = R 3 and use the single function
f(x, y, z) = x 2 + y2 + Z2 - 1. Then, (Ix ,Iv' fz) = (2x, 2y, 2z) so II (Ix, f v ' fz) II
72 Elementary Differential Geometry 4 Surfaces in Three Dimensions 73

= 2 at all points of S2. In particular, (Ix, jy, jz) is non-zero everywhere on S2. two analogous parametrisations for the part z < 0 of the hyperbolic
Hence, the theorem tells us that 8 2 is a smooth surface. paraboloid.
4.9 Show that the level surface
Example 4.8 x2 y2 Z2
2"
a + -b2 + 2"
c = 1,
For the double cone of Example 4.3, f(x, y, z) = x 2 + y2 - Z2. Hence,
(Ix, jy, fz) = (2x, 2y, -2z), and this vanishes only at the vertex (0,0,0). Hence, where a, band c are non-zero constants, is a smooth surface (called
removing the vertex gives a smooth surface, as we have already seen. an ellipsoid). (A picture of an ellipsoid can be found in Proposition
4.6.)
For the rest of this book, by a surface we shall mean 4.10 A torus is obtained by rotating a circle C in a plane II around a
a smooth surface, and by a surface patch we shall mean straight line £. in II that does not intersect C. Take II to be the
a regular smooth surface patch (or equivalently an allowable surface patch). xz-plane, £. to be the z-axis, a > 0 the distance of the centre of C
from £, and b < a the radius of C. Show that the torus is a smooth
Unless we indicate otherwise, we shall also assume that all surfaces we
surface
consider are connected, which means that any two points of S can be joined by
(i) by showing that it has an atlas consisting of surface patches
a curve lying entirely in S. This is not a serious restriction, for it is not difficult
to prove that any surface is a disjoint union of connected surfaces, and so can 0'((), tp) = «a + bcos 0) cos tp, (a + bcos 8) sin tp, b sin 0),
be studied by studying each of its connected parts separately. All the surfaces
with (0, tp) in suitable open subsets of R 2 ;
we have encountered so far are connected except the double cone of Example
(ii) by showing that it is the level surface given by
4.3, which breaks into the union of two disjoint half cones S± when the vertex
is removed, as it must be to have a surface. (x 2 + y2 + z2 + a2 _ b2)2 = 4a2(x2 + y2).

EXERCISES

4.6 Show that, if f(x, y) is a smooth function, its graph

{(x, y, z) E R 3 I Z = f(x, yn
is a smooth surface with atlas consisting of the single regular surface
patch
O'(u,v) = (u,v,j(u,v)).
4.11 If S is a smooth surface, define the notion of a smooth function
4.7 Verify that the six surface patches for the unit sphere in Exercise 4.3
S --+ R. Show that, if S is a smooth surface, each component of the
are regular. Calculate the transition maps between them and verify
inclusion map S --+ R 3 is a smooth function S --+ R.
that these maps are smooth.
4.12 Show that translations and invertible linear transformations of R 3
4.8 Show that
take smooth surfaces to smooth surfaces.
O'(r,()) = (rcoshO,rsinh9,r2 )
is a parametrisation of the part z > 0 of the hyperbolic paraboloid
z = x 2 - y2. (A picture of the hyperbolic paraboloid can be found in
Proposition 4.6.) Use Exercise 4.6 to find another parametrisation iJ
of the same part, and verify that u is a reparametrisation of 0'. Find
74 Elementary Differential Geometry 4. Surfaces in Three Dimensions 75

4.3. Tangents, Normals and Orientability Since (T" and (Tv are linearly independent, the tangent space is tWQ-
dimensional, and will be called the tangent plane from now on. Note that Def-
A natural way to study a surface S is via the (smooth) curves "'( that lie in S. inition 4.4 shows that the tangent plane is independent of the choice of patch
3
If "y : (0:, (3) --t R is contained in the image of a surface patch (I : U --t R 3 in containing P, even though this is not immmedi~tely obvious from Proposition
the atlas of 8, there is a map (0:, f3) --t U, say t t-+ (u(t), v(t)), such that 4.4 (see Exercise 4.15).
"Y(t) = (I(u(t), v(t)). (2) Since the tangent plane at PES passes through the origin of R 3 , it is
completely determined by giving a unit vector perpendicular to it, called a
The functions u and v are necessarily smooth (see Exercise 4.30). Conversely, unit normal to S at P. There are, of course, two such vectors, but Proposition
it is obvious that if t t-+ (u(t), v(t)) is smooth, then Eq. (2) defines a curve lying 4.4 shows that choosing a surface patch (I : U --t R 3 containing P leads to a
in 8. In general, if"Y is a curve in 8 and some point "'((to) of"Y lies in a surface definite choice, namely
patch (T of 8, then Eq. (2) will hold for all t in some open interval containing
to. Thus, we may restrict ourselves to curves of the form (2). (3)

(with the derivatives evaluated at the point of U corresponding to P), for this
Definition 4.4 !.'.
is clearly a unit vector perpendicular to every linear combination of (T" and (I v.
The tangent space at a point P of a surface S is the set of tangent vectors at This is called the standard unit normal of the surface patch (I at P. Unlike the
P of all curves in S passing through P. tangent plane, however, N(T is not quite independent of the choice of patch (T
containing P. In fact, if U : (j --t R 3 is another surface patch in the atlas of S
Proposition 4.4 containing P, we showed in the proof of Proposition 4.2 that

Let (T ; U --t R 3 be a patch of a surface S containing a point P of S, and let Uu x Uv = det(J(p)) (ltL x (Iv,
(u, v) be coordinates in U. The tangent space to 8 at P is the vector subspace where J(p) is the jacobian matrix of the transition map {jj from IT to U. So the
of R 3 spanned by the vectors (TtL and (Iv (the derivatives are evaluated at the standard unit normal of jj is
point (uo, vo) E U such that (I(uo, vo) = P). iTfL x iT (T" X (Tv
N iT = II iT u x iTvv II = ±
II (I" x (Tv II
= ±N(T,
Proof 4.4 where the sign is that of the determinant of J(p). This leads to the fallowing
Let "'( be a smooth curve in S, say definition.

"Y(t) = (I(u(t), v(t)).

Definition 4.5
Denoting d/dt by a dot, we have, by the chain rule,
An orientable surface is a surface with an atlas having the property that, if
it = (ltL U + (lv v. P is the transition map between any two surface patches in the atlas, then
Thus, it is a linear combination of (ltL and (Tv. det(J(P)) > 0 where P is defined.
Conversely, any vector in the vector subspace of R 3 spanned by (I tL and (I v
is of the form {(Tu + 1](lv for some scalars ~ and 1]. Define The preceding discussion gives
"'((t) = (T(uo + ~t, VO + Tlt).
Then, "'( is a smooth curve in S and at t = 0, i.e. at the point P in S, we have Proposition 4.5

it = {(Tu + 1](Tv' An orientable surface S has a canonical choice of unit normal at each point,
obtained by taking the standard unit normal of each surface patch in the atlas
This shows that every vector in the span of (I u and (Tv is the tangent vector at ~S. 0
P of some curve in S. 0
76 Elementary Differential Geometry
T--~4-. ~S~urf~a~c~e~s ~in-T~hrllllee~D~im-e-n-s~io-n-s -------------------7-7---
In fact, the converse is also true: if a surface S has a unit normal N defined We compute the standard unit normal NO" at points on the median circle
at each point PES and depending smoothly on P, then Sis orientable. To see (where t = 0). At such points, we have
this, start with the maximal atlas of S and retain a patch oo(u, v) if Uu xu v is a . 8 8 . (J • ()
positive multiple of N at all points in the image of 00, otherwise discard it. The
).

oot = ( -sJn2"cos ())

,-sJn2"sJn ,cos2" ' .uo = (-sin(),cos8,0),
patches that remain form an atlas satisfying the condition in Proposition 4.5. so

at x ao = (- cos () cos~, - sin () cos ~2' - ~)

We leave the details of this to the interested reader (the argument is similar to
sin
that used in the next example). 2 2'
Most of the surfaces we shall discuss are orientable (see Exercise 4.16). Here This is a unit vector, so it is equal to N a .
is one that is not. If the Mobius band were orientable, there would be a well defined unit
normal N defined at every point of S and varying smoothly over S. At a point
Example 4.9 a(O, 0) on the median circle, we would have

The Miibius band is the surface obtained by rotating a straight line segment £ N = A«())Nu ,
around its midpoint P at the same time as P moves around a circle C, in such where A : (0,271") ---t R is smooth and A«()) = ±1 for all O. It follows that either
a way that as P moves once around C, £ makes a half-turn about P. If we take >"(9) = +1 for all () E (0,271"), or >"(8) = -1 for all 0 E (0,271"). Replacing N by
C to be the circle x 2 + y2 = 1 in the xy-plane, and £ to be a segment of length =
-N if necessary, we can assume that A 1. At the point a(O,O) =
a(O, 211"),
1 that is initially parallel to the z-axis with its midpoint P at (1,0,0), then we must have (since N is smooth)
after P has rotated by an angle 0 around the z-axis, £ should have rotated by
N = limNa = (-1,0,0)
0/2 around P in the plane containing P and the z-axis. The point of £ initially 1/.).0

at (1,0,t) is then at the point and also

oo( t, 0) =( (1 - t sin ~) cos 0, (1 - t sin ~) sin 0 I t cos ~) . N = lim NO" = (1,0,0).

Ot2n
This contradiction shows that the Mobius band is not orientable.
We take the domain of definition of a to be
U = {(t,O) E R 2 I -1/2 < t < 1/2, °< () < 211"}.
We can define a second patch it by the same formula as a but with domain of EXERCISES
definition U = {(t,()) E R 2 I -1/2 < t < 1/2, -11" < () < 7I"}. It can be checked
that these two patches form an atlas for the Mobius band consisting of regular 4.13 Find the equation of the tangent plane of the following surface
patches at the indicated points:
surface patches, making the Mobius band into a smooth surface S.
(i) a(u,v) = (u,v,u 2 - v 2 ), (1,1,0);
(ii) u(r, B) = (rcosh(),rsinhO,r 2 ), (1,0,1).
4.14 A helicoid is a surface swept out by an aeroplane propeller, when
both the aeroplane and its propeller move at constant speed. (A
picture of the helicoid can be found in Example 9.3.) If the aeroplane
is flying along the z-axis, show that the helicoid can be parametrised
as
u(u,v) = (vcosu,vsinu,>"u),
where>" is a constant. Show that the cotangent of the angle that
the standard unit normal of u at a point P makes with the z-axis is
proportional to the distance of P from the axis.
78 Elementary Differential Geometry 4. Surfaces in Three Dimensions 79

4.15 If u(u, v) is a surface patch, show that the set of linear combinations for u to be a surface patch (and hence injective), no straight line parallel to
of Uti and U v is unchanged when u is reparametrised. a should meet "I in more than one point. Finally, Uu = "t, (Iv = a (with a dot
4.16 Consider the surface S defined by f(x, y, z) = 0, where f is a smooth denoting d/du), so iT is regular if and only if the tangent vector of'Y is never
function such that Ix, Iy and Iz do not all vanish at any point of 5. parallel to a.
Show that the vector

is perpendicular to the tangent plane at every point of 5, and deduce

that 5 is orientable. (Compare Exercise 1.17.)
4.17 Let S be a surface and let F : S -+ R be a smooth function (see
Exercise 4.11). Show that, at each point PES, there is a unique
vector V sF in the tangent plane at P such that ------------- ...

(VsF) ...,.(O) = :t!t=o F(')'(t))

for all curves "I in S with "1(0) = P. Deduce that VsF = 0 if F has
a local maximum or a local minimum at P. The parametrisation is simplest when "I lies in a plane perpendicular to a.
Show that, if S is the surface in Exercise 4.16, then V sF is the In fact, this can always be achieved by replacing "I by its projection onto such
perpendicular projection of V F onto the tangent plane to S, and a plane (see Exercise 4.22). The regularity condition is then clearly satisfied
deduce that, if F has a local maximum or a local minimum at P, provided 'Y is never zero, i.e. provided "I is regular. We might as well take the
then V F = AV I for some scalar A. (This is called Lagrange's Method plane to be the xy-plane and a = (0,0,1) to be parallel to the z-axis. Then,
of Undetermined Multipliers.) 'Y( u) = (J (u), 9 (u), 0) for some smooth functions f and g, and the parametri-
sation becomes
I1(U,V) = (J(u),g(u),v).
4.4. Examples of Surfaces As an example, starting with a circle, we get an ordinary (circular) cylinder.
Taking the circle to have centre the origin, radius 1 and to lie in the xy-plane,
We now describe some of the simplest classes of surfaces. Others will be intro- it can be parametrised by
duced later in the book.
'Y(u) = (cosu,sinu,O),
Example .{10 defined for 0 < u < 21T and -1T < u < 1T, say. This gives an atlas for the
cylinder consisting of two patches, both given by
A (generalised) cylinder is the surface S obtained by translating a curve. If the
curve is "I : (a,fJ) -t R 3 and a is a unit vector in the direction of translation, u(u,v) = (cosu,sinu,v),
the point obtained by translating the point 'Y(u) of "I by the vector va parallel and defined on the open sets
to a is
{(u,v) E R 2 I 0 < u < 21T}, {(u,v) E R 2 I -1T < u < 1T}.
u(u, v) = 'Y(u) + va.
Then, u: U -t R 3 , where U = {(u,v) E R 2 I a < u < ,6}, and clearly u is Example ..pt
smooth. Since
A (generalised) cone is the union of the straight lines passing through a fixed
point and the points of a curve. If p is the fixed point and "I : (a,,6) -+ R 3 is
 Elementary Differential Geometry 4. Surfaces in Thrf!e Dimensions 81
80

the curve, the most general point on the straight line passing through p and a point P of the surface lies on one of the given straight lines which intersects C
point ")'( u) of the curve is at Q, say. If -y is a parametrisation of C with -y(u) = Q, and if 6(u) is a non-zero
vector in the direction of the line passing through 'Y(u), P has position vector
cr(u, v) = (1 - v)p + v")'(u). of the form
Then, U is clearly smooth. Now,
IT(u, v) = 'Y(u) + v6(u),
u(u, v) = cr(u' , Vi) {::::::} v")'(u) - v''')'(u') + (Vi - v)p = 0,
for some scalar v.
which says that the points p, -y(u) and -y(u /) are collinear. So, for cr to be a
surface patch, no straight line passing through p should pass through more
than one point of -y (in particular, -y should not pass through p). Finally, we
have Uu = v'Y, lTv =-y-p (with d/du denoted by a dot), so cr is regular provided
v f. 0, Le. the vertex of the cone is omitted (d. Exercise 4.3), and none of the
straight lines forming the cone is tangent to -y.

\
i
\
\
\
\
\ We have, with d/du denoted by a dot,
\
\
,... ----- lT u =.y+v6, (1v =6.
------- ----'" \

Thus, (1 is regular if i'+v6 and 6 are linearly independent. This will be true, for
example, if 'Y and 6 are linearly independent and v is sufficiently small. Thus,
to get a surface, the curve C must never be tangent to the rulings.
The parametrisation is simplest when 'Y lies in a plane. If this plane contains
p, the cone is simply part of that plane. Otherwise, we can take p to be the
Example 4.13
=
origin and the plane to be z 1. Then, -y(u) (f(u),g(u), I) for some smooth =
functions f and g, and the parametrisation takes the form A surface of revolution is the surface obtained by rotating a plane curve, called
the profile curve, around a straight line in the plane. The circles obtained by
u(u, v) = v(f(u), g(u), 1). rotating a fixed point on the profile curve around the axis of rotation are called
the parallels of the surface, and the curves on the surface obtained by rotating
Examples 4.10 and 4.11 are both special cases of the next class of surfaces. the profile curve through a fixed angle are called its meridians. (This agrees
with the use of these terms in geography, if we think of the earth as the surface
Example 4·12 obtained by rotating a great circle passing through the poles about the polar
axis and we take u and v to be latitude and longitude, respectively.)
A ruled surface is a surface that is a union of straight lines, called the rulings of
the surface. Suppose that C is a curve in R 3 that meets each of these lines. Any
82 Elementary Differential Geometry 4. Surfaces in Three Dimensions 83

Let us take the axis of rotation to be the z-axis and the plane to be the EXERCISES
xz-plane. Any point P of the surface is obtained by rotating some point Q of
the profile curve through an angle v (say) around the z-axis. If 4.18 The surface obtained by rotating the curve x = cosh z in the xz-
plane around the z-axis is called a ca~enoid. Describe an atlas for
")'(u) = (f(u), 0, g(u» this surface. (A picture of the catenoid can be found in Section 9.2.)
is a parametrisation of the profile curve containing Q, P has position vector of 4.19 Show that
the form
0'( u, v) = (sech u cos v, sech u sin v, tanh u)
O'(u, v) = (f(u) cos v, f( u) sin v, g( u».
is a regular surface patch for the unit sphere (it is called Merca-
tor's projection). Show that meridians and parallels on the sphere
z
correspond under (I to peri>endicular straight lines in the plane.
4.20 A loxodrome is a curve on the unit sphere that intersects the meridi-
ans at a fixed angle, say a. Show that, in the Mercator surface patch
(1 (see Exercise 4.19), a unit-speed loxodrome satisfies

it = cos ct cosh u, iJ = ± sin a cosh u

(a dot denoting differentiation with respect to the parameter of the
loxodrome), Deduce that loxodromes correspond under u to straight
lines in the uv-plane.
y 4.21 A right conoid is a ruled surface whose rulings are parallel to a given
plane II and pass through a given straight line {, perpendicular to
II. If II is the xy-plane and {, is the z-axis, show that
u(u, v) = (v cos B(u), vsinB(u), u)
x
is a regular surface patch for the conoid, where B( u) is the angle that
To check regularity, we compute (with a dot denoting d/du): the ruling through (0,0, u) makes with the z-axis (B(u) is assumed
to be a smooth function of u). Taking 9( u) = u gives a helicoid
O'u = (jcosv,jsinv,g), (Iv = (-fsinv,fcosv,O), (Exercise 4.14).
O'u x 0'1,1 = (fgcosv, - fgsinv,fj),
x (11,1 11 = reP + Ii)·
2
II (Iv.

Thus, au x (11,1 will be non-vanishing if feu) is never zero, i.e. if'Y does not
intersect the z-axis, and if j and 9 are never zero simultaneously, i.e. if"Y is
regular. In this case, we might as well assume that j(u) > 0, so that feu) is
the distance of u(u,v) from the axis of rotation. Then, 0' is injective provided
that "Y does not self-intersect and the angle of rotation v is restricted to lie in
\ n
an open interval of length::; 27r. Under these conditions, surface patches of the
form (1 give the surface of revolution the structure of a smooth surface.
84 Elementary Differential Geometry 4. Surfaces in Three Dimensions 85

4.22 Show that, if u(u, v) is the (generalised) cylinder in Example 4.10: is one of the following:
(i) the curve i(u) = 1'('11,) - ('Y(u).a)a is contained in a plane per-
pendicular to a; (i) ellipsoid: ~
p + ~
q + ~
r = 1
= =
(ii) /1(u, v) i'(u) + va, where v v + 'Y(u).a;
(iii) &(u, v) = 'i'(u) + va is a reparametrisation of u(u, 11).

4.5. Quadric Surfaces

The simplest surfaces, namely planes, have cartesian equations that are linear
in x, y and z. From this point of view, the next simplest surfaces should be (ii) hyperboloid 0/ one sheet: 5 + ~ - 5-r =
p q
1
those whose cartesian equations are given by quadratic expressions in x, y and
z. This leads to the following definition.

Definition 4.6
A quadric is the subset of R 3 defined by an equation of the form
(rA).r + b.r + c = 0,
3
where r = (x, y, z), A is a constant symmetric 3 x 3 matrix, b E R is a constant
vector, and c is a constant scalar. (iii) hyperboloid of two sheets: 5- - ~p' - ~ = 1
r
y2

To see this more explicitly, let

A = (:: :: ::), b = (b1 , b2 , b3 ).

a6 a5 a3
Then, the equation of the quadric is
alx 2 + a2y2 + a3 z2 + 2a 4x y + 2a5YZ + 2a6xz + b1x + b2y + b3z + c = O. (4)
A quadric is not necessarily a surface. For example, the quadric with equa-
tion x 2 + y2 + Z2 = 0 is a single point, and that with equation x + y2 = 0
2
(iv) elliptic pamboloid: ~ +~
q =z
is a straight line. A more interesting example is the quadric xy = 0, which is p

the union of two intersecting planes, which is also not a surface. (Intuitively, it
has a 'corner' along the line of intersection of the planes.) The following theo-
rem shows that it is sufficient to consider quadrics whose equations take on a
particularly simple form.

Proposition 4.6 \
By applying a rigid motion o/R3 , every non-empty quadric (4) in which the co-
efficients are not all zero can be transformed into one whose cartesian equation
86 Elementary Differential Geometry 4, Surfaces in Three Dimensions 87

(v) hyperbolic paraboloid: ?- 2 2

~ = z (ix) parabolic cylinder: ~
p
= y

(vi) quadric cone:

2 2
}r + ~ -
2
~ = 0 (x) plane: x = 0
(xi) two parallel planes: x 2 = p2
(xii) two intersecting planes: ~ - ~ =0
(xiii) straight line: ~ + ~ =0
222
(xiv) single point: ?- + ~ + ~ = O.
In each case, p, q and r are non-zero constants.

Proof 4·6
The proof depends on the following fact. If A is a real symmetric matrix, there
(vii) elliptic cylinder: ~+~ = 1
is a matrix P with pt P = I and det(P) = 1 such that pt AP is a diagonal
matrix A' (pt denotes the transpose of P). The diagonal entries of A' are the
eigenvalues of A, and the rows of P are the corresponding eigenvectors.
With A as in Definition 4.6, we define r ' = (Xl, y/, z/), b ' = (b~ 1 b~, b3), where
(Xl y' z') = (xy z}P, (b~ b~ b;) = (b1 b2 b3 }P.
Writing the equation of the quadric as
(xyz}A(xyz}t + (b 1 b2 b3 )(xyz}t + c = 0

2 2
and noting that
(viii) hyperbolic cylinder: ?- - ~ =1
we get
+ (b~ b~ b;)(x' yl Z/}t + c :::: 0,
(Xl y' z/}A' (Xl y' z'}t
(r ' A').r' +b/.r' + c = 0,
,,2
a1 x
/ 12
+ a 2 y + a3z 1,2
+b'1 X I + b'2Y I + b'3 Z I + c = 0 ,
where a~ ,a2 and a3 are the diagonal entries of A', Le. the eigenvalues of A. Since
any 3 x 3 matrix P with pt P = I and det(P) = 1 represents a rotation of R 3 ,
this new quadric is obtained from the given one by a rigid motion. Hence, we
88 Elementary Differential Geometry 4. Surfaces in Three Dimensions 89

might as well consider the quadric in (4), but assume that a4 = as = a6 = 0, Finally, if ai = a2 = a3 = 0, (5) is the equation of a plane, so after applying
i.e. a rigid motion we are in case (x) again. 0
(5)

Suppose now that, in Eq. (5), ai -# O. If we define x' = x + bI/2ai, corre-

Example 4-14
sponding to a translation of R 3 , the equation becomes Consider the quadric
aix,2 + a2y2 + a3z2 + b2y + b3 z + c' = 0, x2 + 2y 2 + 6x - 4y + 3z = 7.

where c' is a constant. In other words, if ai -# 0, we can assume that bi = 0, Setting x' = x + 3, y' = y - 1 (a translation), we get
and similarly for a2 and a3, of course. X,2 + 2y,2 + 3z :::: 18.
If al, a2 and a3 in Eq. (5) are all non-zero, we may therefore reduce to the
Setting z' = z - 6 (another translation) gives
form
x,2 + 2y,2 + 3z' = O.
aix 2 + a2y2 + a3z2 + c = O.
Finally, setting x" :::: x', y" = -y', zIt = - Zl (a rotation by IT about the x-axis)
If c i= 0, we get cases (i), (ii) and (iii), depending on the signs of ai, a2, a3 and
gives
c, and if c = 0 we get cases (vi) and (xiv).
1 2
If exactly one of ai, a2 and a3 is zero, say a3 = 0, we are reduced to the -x
3 3
112
=z ,
+ -y 112 II

form
which is an elliptic paraboloid. It can be parametrised by setting x" :::: U, y" :=
(6) V, Zll = ~U2 + ~V2. This corresponds to x :::: U - 3, y :::: 1 - v, Z :::: 6 - ~u2 - ~V2,

If b3 i' 0, we may define z' :::: Z + c/b 3 • Thus, by a translation (and by dividing and shows that the given quadric is.a smooth surface with an atlas consisting
by b3 ), we are reduced to the case of the single surface patch

alx 2 + a2y2 + z = o. a(u, v) :::: (u - 3,1 - v,6 - ~U2 - ~V2) .

This gives cases (iv) and (v).
°
If b3 = in Eq. (6), we have
alx 2 + a2y2 + c = O. EXERCISES
If c = 0 we get cases (xii) and (xiii). If c i' 0, dividing through by it leads to 4.23 Write down parametrisations of each of the quadrics in parts (i)-(xi)
cases (vii) and (viii). of Proposition 4.6 (in case (vi) one must remove the origin).
Suppose now that a2 = a3 = 0, but al i- O. Then we have 4.24 Which quadric surfaces are
(7) (a) generalised cylinders;
If b2 and b3 are not both zero, by rotating the yz-plane so that the y-axis (b) generalised cones;
becomes parallel to the vector (b 2 , b3 ), we can arrive at the situation b2 i- (c) ruled surfaces;
0, b3 = 0, and then by a translation along the y-axis we can arrange that c = O.
(d) surfaces of revolution?
This leads to the equation 4.25 By setting
x y x y
alx
2
+Y = 0, u= - - -
p q'
v= -
p
+-,
q
which gives case (ix). If b2 = b3 :::: 0 in Eq. (7), then c = 0 gives case (x) and find a surface patch covering the hyperbolic paraboloid
c i- 0 gives case (xi). x2 y2
-
p2 - -q2-- z .
 90 Elementary Differential Geometry 4. Surfaces in Three Dimensions 91

Deduce that the hyperbolic paraboloid is doubly ruled. It follows from these properties and the Intermediate Value Theorem that
4.26 Show that, if a quadric contains three points on a straight line, it there is at least one value of t in each open interval (-00, p2), (p2, q2) and
contains the whole line. (Parametrise the line bY'Y(t) = a + bt, and (q2, r 2 ) such that Ft(a, b, c) :::::: 1. On the other hand, the equation Ft(a, b, c) :::::: 1
note that substituting into Eq. (4) gives a quadratic equation for t.) is equivalent to the cubic equation Gt(a, b, c) :::::: 0, where
Deduce that, if L 1 , L 2 and L 3 are non-intersecting straight lines in
3 Gt(a, b, c) :::::: a2(q2 - t)(r 2 - t) + b2(p2 - t)(r 2 - t) + C2(p2 - t)(q2 - t) (8)
R , there is a quadric containing all three lines. (Take three points _ (p2 _ t)(q2 _ t)(r 2 - t),
on each line and show that there is a quadric passing through all
nine points.) and so has at most three real roots. It follows that there are unique numbers
4.27 Use Exercise 4.26 to show that any doubly ruled surface is (part of) u E (-00, p2), V E (p2, q2) and W E (q2, r 2) (depending on (a, b, c), of course)
a quadric surface. (A surface is doubly ruled if it is the union of each such that
of two families of straight lines such that no two lines of the same Fu(a, b, c) :::::: 1, Fv(a, b, c) :::::: 1, Fw(a, b, c) :::::: 1. (9)
family intersect, but every line of the first family intersects every line
The three quadrics Fu(x,y,z) :::::: 1, Fv(x,y,z) :::::: 1 and Fw(x,y,z) :::::: 1 are
of the second family, with at most a finite number of exceptions.)
ellipsoids, hyperboloids of one sheet and hyperboloids of two sheets, respec-
Which quadric surfaces are doubly ruled?
tively, and we have shown that there is one of each passing through each point
(a, b, c) E R 3 that does not lie on any of the coordinate planes. We show that
they form a triply orthogonal system.
4.6. Triply Orthogonal Systems
We conclude this section by showing how quadric surfaces furnish some beau-
tiful examples of triply orthogonal systems. By such a system, we mean three
families of surfaces, each depending on a single parameter, with the property
that, if P is a point that is on one surface of each family, the tangent planes
of these surfaces at P are mutually perpendicular. The simplest example, of
course, consists of the planes parallel to one of the three coordinate planes.
Other examples, almost as obvious, are given in Exercise 4.28. But a more
interesting example can be constructed in which the three families consist of
ellipsoids, hyperboloids of one sheet and hyperboloids of two sheets, respec-
tively.
To see this, let p, q and r be constants, and assume that 0 < p2 < q2 < r 2.
For (x, y, z) E R 3 , t i- p2, q2 or r 2, let
x 2 y2 z2
Ft(x,y,z):::::: -2-- + -2-- + -2--'
p-t q-t r-t
3
Fix a point (a, b, c) E R with a, band c all non-zero. The following properties
are clear:
(i) Ft(a, b, c) is a continuous function of t in each of the open intervals
(-OO,p2), (p2,q2), (q2,r 2) and (r 2,oo);
(ii) Ft(a,b,c) ~ 0 as t ~ ±oo;
(iii) Ft(a, b, c) ~ 00 as t approaches p2, q2 or r 2 from the left, and Ft(a, b, c) ~
- 00 as t approaches p2 ,q2 or r 2 from the right.
 92 Elementary Differential Geometry 4. Surfaces in Thr~e Dimensions 93

Indeed, the vector

(r ~ t' q2 Y- t ' r 2 ~ t)
is perpendicular to the tangent plane of the surface Ft (x, y, z) = 1 at (x, y, z)
(see Exercise 4.16). Thus, to show that the first two surfaces in (9) are perpen-
dicular at (a,b,c), for example, we have to show that
~ ~ 2
-;-;::----::-;---;;------;-
(p2 - U)(p2 - v)
+ (q2 - U)(q2 - v)
+ (r 2 - u)(r 2- -
v) -
0
.
But the left-hand side of this equation is
Fu(a,b,c)-Fv(a,b,c) 1-1
--=-.o.....:.....:.......;'--------=-~....:.....-:..
u-v
=- - = O.
u-v 4.29 By considering the quadric surface Ft(x, y, z) = 0, where
We can also construct a simultaneous parametrisation of the three families. x2 y2
Note that the cubic Gt(a, b, c) is equal to (t-u)(t-v)(t-w), since it is divisible
Ft(x,y,z) = -P2--t- + -q2--t- - 2z + t,
by this product and the coefficients of t 3 agree. Putting t = p2, q2 and r 2 and construct a triply orthogonal system consisting of two families of
solving the resulting equations for a 2, b2 and c2, we find that elliptic paraboloids and one family of hyperbolic paraboloids (see
(p2 _ u)(p2 - v)(p2 - w) above). Find a parametrisation of these surfaces analogous to (1O).
a= ± (r2 _ p2)(q2 _ p2)

(q2 - u)(q2 - V)(q2 - w)

b= ± (p2 _ q2)(r2 _ q2) (10) 4.7. Applications of the Inverse Function
Theorem
(r 2 - u)(r 2
- v)(r 2 - w)
c= ± (p2 _ r2)(q2 _ r2)
In this section we give the proofs of Proposition 4.1 and Theorem 4.1.
Define O'(u,v,w) = (x,y,z), where x,y and z are the right-hand sides of the Suppose first that f : U -+ R n is a smooth map, where U is an open subset
three equations in (10), respectively, with any combination of signs. For fixed of R m • If we write (Ul' ... ,un) = f(Ul, . .. , urn), the jacobian matrix of f is
u (resp. fixed v, fixed w), this gives eight surface patches for the correspond-
fu fu Bu]m )
8u
ing ellipsoid F u (x, y, z) =
1 (resp. hyperboloid of one sheet F v (x, y, z) = 1,
BUl
~
BU2
~ BU2
hyperboloid of two sheets F 11J (x,y,z) = 1). J (f) = 8~l 8U2 8~m.
(
~ ~ BUn
lJu 1 lJU'2 8u m
EXERCISES This has already been used in the case m =n = 2 in Section 4.2, but now we
shall need it in other cases too.
4.28 Show that the following are triply orthogonal systems: The main tool that we use is
(i) the spheres with centre the origin, the planes containing the
z-axis, and the circular cones with axis the z-axis;
Theorem 4.2 (Inverse Function Theorem)
(ii) the planes parallel to the xy-plane, the planes containing the
z-axis and the circular cylinders with axis the z-axis. Let f : U -+ R n be a smooth map defined on an open subset U of R n (n 2: 1).
Assume that, at some point Xo E U I the jacobian matrix J(f) is invertible.
 elementary Ultterentlal Geometry 4. Surfaces in Three Dimensions

Then, there is an open subset V of R n and a smooth map 9 : V --+ R" such on TV where F::;: 7fOo-. Since p-l and F are smooth on WI so is the transition
that map ~-1 0 a. Since 0'-1 0 jj is smooth on an open set containing any point
(i) Yo=f(Xo)EV; (uo, va) where it is defined, it is smooth. 0
(ii) g(yo) = Xo; We now give the proof of Theorem 4.1. Let. P, Wand 1 be as in t.he
(iii) g(V) <; U; statement of the theorem, and suppose t.hat P = (xo,. yo, zo) and that. f, ,. 0 a~
(iv) g(V) is an open subset ofR n ; P. (The proof is similar in the other two cases.) Consider t.he map F: W --+ R
(v) f(g(y)) = y for all y E V. defined by
In particular, 9 : V --+ g(V) and 1 : g(V) --+ V are inverse bijections.
F(x, y, z) = (x, y, l(x, y, z)).
Thus, the inverse function theorem says that, if J(J) is invertible at some The jacobian matrix of F is
point, then 1 is bijective near that point and its inverse map is smooth. A proof
of this theorem can be found in books on multivariable calculus.
We use the inverse function theorem to give the proof of Proposition 4.1.
We want to show that, if a: U --+ R 3 and" : fj --+ R 3 are two regular patches
Ox J. JJ,
and is clearly invertible at P since fz ,. O. By the inverse function theorem,
in the atlas of a surface S, the transition map from a to " is smooth where it
there is an open subset V of R 3 containing F(xo, yo, zo) = (xo, ~o, 0) and a
is defined.
smooth map G : V --+ W such that W = G(V) is open and F : W --+ V and
Suppose that a point P lies in both patches, say a(uo, vol = "(Uo, vo) = P.
G : V --+ W are inverse bijections.
Write
Since V is open, there are open subsets U, of R' containing (xo, Yo) and
a(u, v) = (J(u, v),g(u, v), h(u, v)). U, of R containing 0 such that V contains the open set Ul x U, of all points
(x,y,w) with (x,y) E U1 and w EU,. Hence, we might as well assume that
Since au and av are linearly independent, the jacobian matrix
V = U, X U,. The fact that F and G are inverse bijections means that
lu
9u
Iv)
91) G(x,y,w) = (x,y,g(x,y,w))
(h h
u v for some smooth map 9 : U, x U, --+ R, and
of a has rank 2 everywhere. Hence, at least one of its three 2 x 2 submatrices
l(x,y,g(x,y,w)) = w
is invertible at each point. Suppose that the submatrix
for all (x,y) E U1 , W E U,.
lu
( 9u
Jv) Define a : U1 --+ R 3 by
91)
is invertible at P. (The proof is similar in the other two cases.) By the inverse a(x, y) = (x, y, g(x, y, 0)).
function theorem applied to the map F : U --+ R' given by Then a is a homeomorphism from U, to S n W (whose inverse is the restriction
to S n W of the projection ,,(x,y,z) = (x,y)). It is obvious that a is smooth,
F(u, v) = (J(u, v), g(u, v)),
and it is regular because
there is an open subset V of R' containing F(uo, vol and an open subset W of
U containing (uo, vol such that F : W --+ V is bijective with a smooth inverse ax x a y = (-g" -g., 1)
F-l : V --+ W. Since a : W --+ a(W) is bijective, the projection" : a(W) --+ V is nowhere zero. So iT is a regular surface patch on S containing the given point
given by ,,(x, y, z) = (x, y) is also bijective, since" = F 0 a-Ion a(W). It P. Since P was an arbitrary point of S we have constructed an atlas for S
follows that W = ,,-1 (a(W)) is an open subset of fj and that making it into a (smooth) surface.
I

0
0'-1 0 a= F- 1 0 F
96 Elementary Differential Geometry

EXERCISES 5
4.30 Show that, if'Y : (a:,;3) ~ R 3 is a curve whose image is contained in a The First Fundamental Form
surface patch (f : U ---t R 3 , then 'Y(t) = (f(u(t), vet)) for some smooth
map (0:,,8) ~ U, t t-+ (u(t),v(t)). (Imitate the proof of Proposition
4.1.)
3
4.31 Prove Theorem 1.1 and its generalization to level curves in R (Ex-
ercise 1.17).

Perhaps the first thing that a geometrically inclined bug living on a surface
might wish to do is to measure the distance between two points of the surface.
Of course, this will usually be different from the distance between these points
as measured by an inhabitant of the ambient three dimensional space, since
the straight line segment which furnishes the shortest path between the points
in R 3 will generally not be contained in the surface. The object which allows
one to compute lengths on a surface, and also angles and areas, is the first
fundamental form of the surface.

5.1. Lengths of Curves on Surfaces

If 'Y(t) = (f(u(t), vet)) is a curve in a surface patch fT, its arc-length starting at
a point 'Y(to) is given by

s = r II -y(u) II duo
t

lt~
By the chain rule, -y == (fuU + (fv V, so
II -y 11 2 == «(fuU + (fvV).«(fu U + (fv V)
=«(fu.(fu)u2 + «(fu.(fv)UV + «(fv.(fu)i1u + «(fv.(fv}r}
2
== «(fu.(fu)it 2 + 2«(fu.(fv)itv + «(fv.(fv)v
== Eit2 + 2FitiJ + GiJ 2 ,
97
 where (see Example 4.2), we have
0'0 = (-sin(Jcoscp,-sin8sincp,cos(J), 0'1{) = (-cos8sin<p,cos8costp,O),
So ... E = II O'(! 11 2 = 1, F = O'(!.O'", = 0, G = 110'", 1l 2 = cos2 0,
t
s = r(Eu
ito
2 + 2Fuv + GiJ2)1/2dt. (1)
and so the first fundamental form is
d(J2 + cos2 (J dcp2. .
If we bring the dt inside the square root and write (~~}2 (dt)2 = du 2, etc. (1),
we see that s is the integral of the square root of the expression
Example 5.3
Edu 2 + 2Fdudv + Gdv 2 • (2)
We consider a (generalised) cylinder
This is called the first fundamental form of iT. Since
O'(u, v) = 'Y(u) + va
s = J..(d;i, (!)
defined in Example 4.10. As we saw in that example we can assume that 'Y is
Eq. (1) is sometimes written as unit-speed, that a is a unit vector, and that 'Y is contained in a plane perpen-
dicular to a. Then, denoting djdu by a dot,
ds 2 = Edu 2 + 2Fdudv + Gdv 2 .
We shall not attempt to justify these apparently dubious manipulations. We
simply regard the expression in (2) as a way of keeping track of the functions so E = II O'u 11 2 = II,.. 11 2 = 1, F = O'u·O'v = "'.a = 0, G = II O'v 11
2
= II a 11 2 = 1,
E, F and G from which the length of any curve in the surface patch can be and the first fundamental form of 0' is
computed using Eq. (1).
du 2 + dv 2 .
If now 'Y is a curve on an arbitrary surface S, its length can be computed
by breaking 'Y into segments, each of which lies in a surface patch, and using Note that this is the same as the first fundamental form of the plane (see
Eq. (1) to compute the length of each segment. The first fundamental form will Example 5.1). The geometrical reason for this coincidence will be revealed in
change when the surface patch is changed, in the manner described in Exercise the next section.
5.4.
Example 5.4
Example 5.1
We consider a (generalised) cone
For the plane
O'(u, v) = (1 - v)p + v'Y(u)
0'( u, v) = a + up + vq (see Example 4.11). Before computing its first fundamental form, we make some
(see Example 4.1) with p and q being perpendicular unit vectors, we have simplifications to 0'.
O'u = p, O'v = q, so E = II O'u 11 2 = II p W= 1, F = O'u.O'v = p.q = 0, First, translating the surface by p (which does not change its first funda-
G = II O'v 11 2 = II q 11 2 = 1, and the first fundamental form is simply mental form by Exercise 5.3), we get the surface patch 0'1 = U - P = v('Y - p),
du 2 + dv 2 • so if we replace 'Y by 'Y1 = 'Y - P we get 0'1 = v'Y1' This means that we might as
well assume that p := 0 to begin with. Next, we saw in Example 4.11 that for
0' to be a regular surface patch, 'Y must not pass through the origin, so we can
Example 5.2 define a new curve .y by .y(u) = 'Y(u)j II 'Y(u) II· Setting u = u, V = vj II 'Y(u) II,
we get a reparametrisation u(u,v) = v-y(ii) of u with II -y II = 1. We can there-
For the sphere in latitude longitude coordinates
fore assume to begin with that q(u,v) = v'Y(u) with 11 'Y(U) II = 1 for all values
0'( (J, cp) = (cos (1 cos cp, cos (J sin <p, sin (J) of u (geometrically, this means that we can replace 'Y by the intersection of the
100 Elementary Differential Geometry 5. The First Funda,nental Form 101

cone with the unit sphere). Finally, reparametrising again, we can assume that 5.2. Isometries of Surfaces
"yis unit-speed, for we saw in Example 4.11 that for u to be regular, "y must be
regular. We observed in Example 5.3 that a plane and a (generalised) cylinder, when
With these assumptions, and with a dot denoting djdu, we have suitably parametrised, have the .~ame first fundamental form. The geometric
reason for this is not hard to see. A plane piece of paper can be 'wrapped' on
a cylinder in the obvious way without crumpling the paper (we take the case
so E = II vi' 11 2 = v 2 II i' 11 2 = v 2 , F = v-Y.'}' = 0 (since II')' II = 1), G = II'}' 11 2 = 1, of a circular cylinder for simplicity):
and the first fundamental form is
v 2 du 2 + dv 2 •
Note that, as for the generalised cylinder in Example 5.3, there is no trace of
the curve "y in the first fundamental form.

EXERCISES

5.1 Calculate the first fundamental forms of the following surfaces:

(i) 0'( u, v) = (sinh u sinh v, sinh u cosh v, sinh u);
(ii) O'(u, v) = (u - v, u + v, u 2 + v 2 );
(iii) 0'( u, v) = (cosh u, sinh u, v);
(iv) u(u,v) = (u,v,u 2 +v 2 ).
What kind of surfaces are these? If we draw a curve on the plane, then after wrapping it becomes a curve on the
5.2 Sketch the curve on the cone cylinder. Because there is no crumpling, the lengths of these two curves will be
the same. Since the lengths are computed as the integral of the (square root of
O'(u,v) = (ucosv,usinv,u)
the) first fundamental form, it is plausible that the first fundamental forms of
given by u = eAt, v = t, where A is a constant. Find the length of the the two surfaces should be the same. Experiment suggests, on the other hand,
part of the curve with 0 ~ t ~ 1r. that it is impossible to wrap a plane sheet of paper around a sphere without
5.3 Show that applying a rigid motion to a surface does not change its crumpling. Thus, we expect that a plane and a sphere do not have the same
first fundamental form. first fundamental form (whatever the choice of parametrisation).
The following definition makes precise what it means to wrap one surface
5A Let a surface u(u, v) be a reparametrisation of a surface O'(u, v), and
onto another without crumpling.
let

Definition 5.1
be their first fundamental forms. Let
au If 8 1 and 8 2 are surfaces, a diffeomorphism f : 8 1 ~ 8 2 is called an isometry
J- au lJU)
lJil if it takes curves in 8 1 to curves of the same length in 8 2 , If an isometry
- ( av lJv
Bu lJil f : 8 1 -t 8 2 exists, we say that 8 1 and 8 2 are isometric.
be the jacobian matrix of the reparametrisation map (it, v) r-+ (u, v),
and let Jt be the transpose of J. Prove that
Theorem 5.1
( ~F ~) = Jt
G
(EF GF) J. A diffeomorphism f : 8 1 ~ 8 2 is an isometry if and only if, fOT any surface
102 Elementary Differential Geometry 5.

patch 0'1 of 51, the patches 0'1 and f 00'1 of 51 and 5 z , respectively, have the Example 5.5
same first fundamental form. Let 51 be the infinite strip in the xy-plane given by 0 < x < 211", and let 52
be the circular cylinder x 2 + y2 == 1 with the ruling given by x == 1, Y == 0
Proof 5.1 removed. Then, 51 is covered by the single patch (11 (u, v) == (u , V, 0), and 52 by
the patch U2(U,V) == (cosu,sinu,v), with 0 < u < 211" in both cases. The map
Since the length of any curve can be computed as the sum of the lengths of 1 : 51 --t 52 that takes O'r( u, v) to (12 (u, v) is an isometry since, by Example
curves each lying in a single surface patch, we can assume that 51 and 5 z 5.3, 0'1 and 0'2 have the same first fundamental form.
are covered by single surface patches. Moreover, since f is a diffeomorphism,
Proposition 4.3 shows that we can assume that these patches are of the form A similar argument shows that a suitable subset of a circular cone is isomet-
0'1 : U --t R:i (for 51) and f 0 0'1 == (1z (for 52). We have to show that 1 is an ric to part of a plane (see Exercise 5.5), and this even extends to generalised
isometry if and only if 111 and O'z have the same first fundamental form. cylinders and cones (see Exercise 5.7). It turns out that there is another class of
Suppose first that 0'1 and (12 have the same first fundamental form. 1£ surfaces that are isometric to the plane, called tangent developables. (In older
t H- (u(t),v(t» is any curve in U, and ')'l(t) == O'I(U(t),v(t» and ')'2(t) == works, a 'development' of one surface on another was the term used for an
~z{u(t), v(t» are the corresponding curves in 51 and 52, then 1 takes 'Y1 to 'Y2' isometry.) A tangent developable is the union of the tangent lines to a curve in
smce R 3 _ the tangent line at a point -y(u) to a curve 'Y is the straight line passing
/(')'1 (t» == 1(0'1 (u(t), v(t))) == O'z(u{t), v(t» == ')'2(t). through 'Y(u) and parallel to the tangent vector 1'(u).

It is clear that 'Y1 and ')'z have the same length, since both lengths are found by
integrating the expression (Eu Z + 2Fit:iJ+Ov 2 ) 1/2 , where Edu 2 + 2Fdudv +Odv 2
is the (common) first fundamental form of 0'1 and O'z· "'I(u)
Conversely, suppose that 1 is an isometry. If t H- (u(t),v(t») is any curve
in U, defined for t E (0.,13), say, the curves 'Yl(t) == (1r(u(t),v(t) and 'Yz(t) ==
0'2(U(t), v(t) have the same length. Hence,

i~tl
(E 1UZ + 2F1uv + 0 1i})1/2dt == r
ho
h
(Ezu z + 2Fzuv + G ZV2)1/2dt,

for all to, t l E (0.,13), where E 1 , F 1 and 0 1 are the coefficients of the first
fundamental form of U1, and E z , Fz and G z those of 112. This implies that the
two integrands are the same, and hence that
We might as well assume that 'Y is unit-speed. The most general point on the
E1 u2 + 2F1 itv + G 1 v2 == E 2 u2 + 2F2 uiJ + G zv2 . (3) tangent line at ')'(u) is
Fix to E (0',13), and let Uo == u(to), Vo == v(t o)· We now apply Eq. (3) for the (1(u, v) == ')'(u) + v1'(u) ,
following three choices of the curve t l-t (u( t), v (t» in U: for some scalar v. Now
(i) u == Uo + t - to, v = Vo: this gives E 1 == E z; (1", Xliv == (i' + vi') xi' == vi' x 'Y.
(ii) u == Uo, v == Vo + t - to: this gives G 1 == G 2; For (1 to be regular, it is thus necessary that ;Y is never zero, or in other words,
(iii) U == UQ + t - to, v == Vo + t - to: this gives the curvature", == II i' II is > 0 at all points of ')'. Now, 'Y == t, the unit tangent
E 1 + 2F1 + G 1 == E 2 + 2F2 + G 2 , vector of ')', and i' == i == "'0, where n is the principal normal to ')', so

and hence (in view of (i) and (ii) F 1 == F 2 . o (1,. X 'O'v == ",vn x t == -K,vb,
where b is the binormal of ')'. Thus, (1 will be regular if K, > 0 everywhere and
v =I- O. The latter condition means that, for regularity, we must exclude the
104 Elementary Diff iaJ Geometry 5. The First Fundamental Form 105

curve "y itself from the surface. Typically, the regions v > 0 and v < 0 of the But since l' is a plane curve, its tangent lines obviously fill out part of the plane
tangent developable form two sheets which meet along a sharp edge formed by in which i lies. 0
the curve 'Y where v = 0, as the following illustration of the tangent developable
of a circular helix indicates: There is a converse to Proposition 5.1: any sufficiently small piece of a sur-
face isometric to (part of) a plane is (part of) a plane, a (generalised) cylinder,
a (generalised) cone, or a tangent developable. The proof of this will be given
in Section 7.3.

EXERCISES

5.5 The circular cone

u(u, v) = (ucosv,usinv,u), u > 0, 0 < v < 211",
can be 'uwrapped\ hence is isometric to (part of) the xy-plane (say).
Write down this isometry explicitly, and describe exactly which part
of the plane iT is isometric to. Verify that the map really is an isometry.
5.6 Is the map from the circular half-cone x 2 + y2 = z2, Z > 0, to the
Our interest in tangent developables stems from the following result. xy-plane given by (x, y, z) t-+ (x, y, 0) an isometry?
5.7 Show that every (generalised) cylinder and every (generalised) cone
is isometric to (part of) the plane. (See Examples 5.3 and 5.4 and
Proposition 5.1
Exercise. 5.5.)
Any tangent developable is isometric to (part of) a plane.

Proof 5.1

We use the above notation, assuming that 'Y is unit-speed and that K, > O. Now,
2 t=O t:::: 0.6
E = II UU 11 = (it + vi').(-), + vi')
= -)'.-)' + 2v.y.;Y + v 2 ;y.i' = 1 + v2 K,2,
F = Uu.U v = (-)' + v;Y).-)':::: .y..y + v.y.i' = 1,
G = II U 11 2 = .y.-)' = 1,
V

since -)'.-)' = 1, -)'.;Y = 0, ;Y.1' :::::: ",2. So the first fundamental form of the tangent
developable is t = 0.2 t = 0.8

(4)
We are going to show that (part of) the plane can be parametrised so that it
has the same first fundamental form. This will prove the proposition.
By Theorem 2.1, there is a plane unit-speed curve i' whose curvature is '"
(we can even assume that its signed curvature is K). By the above calculations, t:::: 0.4 t=l
the first fundamental form of the tangent developable of i' is also given by (4).
 ementary Differential Geometry 5. The First Fundamental Form 107

5.8 Consider the surface patches Using the dot product formula for the angle between vectors, we see that () is
u(ulv) = (coshucosv,coshusinv,uL 0<v < 271'", given by
u(u,v) = (u cos v, u sin v, v), 0 < v < 271", ".i'
representing the catenoid (Exercise 4.18) with one meridian removed cos () = -II-~'--Ill'--;1~-II
and the part of the helicoid (Example 4.14) between the planes z = 0 By the chain rule,
and z = 21T. Show that the map from the catenoid to the helicoid . . .
that takes u( u, v) to u(sinh Ul v) is an isometry. Which curves on the 'Y = O'uu + O'vv, i' = O'uu + O'vv,
helicoid correspond under this isometry to the parallels and meridians so
of the catenoid ?
'y.~ = (O'uu + O'vv)'(O'ut1 + (Tvb)
In fact, there is an isometric de/ormation of the catenoid into the
= (O'u.O'u)ut + (O'u.O'v)(ub + tv) + (uv.O'v)vt
helicoid. For 0 :5 t :5 1T /2, define a surface
== Eut + F(ufJ + iiiJ) + GiJt.
ut(u, v) = costO'(u, V + t) + sin tu(sinhu, v + t - 71'"/2),
so that O'O(u, v) = O'(u, v) and O'1r / 2(u, v) = o-(sinh u, v). Show that,
Replacing i' by "1 (resp. "1 by i) gives similar expressions for II" 11 2= ..y...y (resp.
II 4-11 2 ), which finally give the formula
for all values of t, the map u(u,v) t-t O't(u,v) is an isometry. The
surfaces O't are shown above for several values of t. _ But + F(uiJ + £tv) + Gvt
cos () - .2 .. .2 . (5)
(Eu + 2Fuv + GiJ 2)l/2(Eu + 2FUfJ + Gv )l/2
2

Example 5.6
5.3. Conformal Mappings of Surfaces
The parameter curves on a surface patch 0'( u l v) can be parametrised by
Now that we understand how to measure lengths of curves on surfaces, it is
natural to ask about angles. Suppose that two curves "1 and i' on a surface
'Y(t) = O'(a, t), i'(t) == u(t, b),
5 intersect at a point P that lies in a surface patch 0' of S. Then l 'Y(t) = respectively, where a is the constant value of u and b the constant value of v in
O'(u(t),v(t)) and i'(t) = O'(u(t),ii(t)) for some smooth functions ulv,u and the two cases. Thus, ,
ii, and for some parameter values to and to, we have O'(u(to), v(to)) = P = u(t) == a, v(t) = t, u(t) = t, v(t) = b,
O'(u(to) ,v(to)).
u = 0, iJ == 1, fi = II t = O.
These parameter curves intersect at the point O'(a, b) ofthe surface. By Eq. (5),
their angle of intersection () is given by
F
cos() = VEC'
In particular, the parameter curves are orthogonal if and only if F = O.
Corresponding to the Definition 5.1 of an isometry, we make
(u,u)

Definition 5.2
If S1 and S2 are surfaces, a diffeomorphism f : 51 -t 52 is said to be conformal
The angle 8 of intersection of "1 and i' at P is defined to be the angle between if, whenever / takes two intersecting curves "11 and i1 on 51 to curves "12 and'1-z
the tangent vectors l' and ~ (evaluated at t = to and t = to, respectively).
108 Elementary Differential Geometry 5. The First Fundamental Form 109

on S2, the angle of intersection of 'Yl and 1'1 is equal to the angle of intersection For the converse, we must show that if
of 'Y2 and i2'
2
In short l f is conformal if it preserves angles. (El u 2 + 2Fl itv+ G l V2)1/2(E1tl + 2Fl tlt + G l t/)1/2
As a special case, if tT : U -+ R 3 is a surface, then a may be viewed as a (6)
= E 2 utl + F 2 (ut + fiv) + G 2 vfJ
map from part of the plane (namely U), parametrised by (u,v) in the usual 2 2 2
(E2u + 2Fzuv + G2V2)l/2(E2ti + 2FzIit + G 2t )l/2
way, and the image 5 of a, and we say that a is a conformal parametrisation
or a conformal surface patch of S if this map between surfaces is conformal. for all pairs of intersecting curves
'Y(t) = 0'1 (u(t), v(t)) and i(t) = tTl (ii(t), vet))
Theorem 5.2 in 51, then the first fundamental forms of a1 and a2 are proportional. Fix
A diffeomorphism f : 51 -+ S2 is conformal if and only if, for any surface patch (a, b) E U and consider the curves
al on 51, the first fundamental forms of al and f 0 a1 are proportional. -y(t) =al(a+t,b), i(t) =O'](a+tcos¢,b+tsin¢),
where ¢ is a constant, for which
Proof 5.2
U = 1, v= 0, Ii = cos ¢, t = sin ¢.
As in the proof of Theorem 5.1 l we can assume that 51 and S2 are covered by
Substituting in Eq. (6) gives
the single surface patches al : U ---+ R 3 and (12 = f 0 O'I, respectively. Suppose
that their first fundamental forms E l cos ¢ + Fl sin ¢

E 1 du 2 + 2Fl dudv + G 1 dv 2 and E 2 du 2 + 2F2 dudv + G 2 dv 2 VE 1 (E1 cos .¢ + 2Fl sin¢ cos ¢
2
+ Gl sin 2 ¢)
(7)
are proportional, say = r===~E~2=C=OS=¢~+=P.~2;,s;;"in~¢=====
E 2 du 2 + 2F2 dudv + G 2 dv 2 = >'(E1 du 2 + 2Fl dudv + G l dv 2 ) V E 2 (E2 cos 2 ¢ + 2F2 sin ¢ cos ¢ + G 2 sin 2 ¢) .

for some smooth function >.(u, v), where (u, v) are coordinates on U. Note that Squaring both sides of Eq. (7) and writing
>.. > 0 everywhere, since (for example) E 1 and E 2 are both> O. If 'Y(t) = (El cos ¢ + F] sin ¢)2 = E 1 (E 1 cos 2 ¢ + 2Fl sin ¢ cos ¢ + Gl sin 2 ¢)
O'I(U(t),v(t)) and .y(t) = al(ii.(t),v(t)) are curves in SI, then f takes "y and i - (El G l - F't) sin 2 ¢,
to the curves a2(u(t), v(t)) and a2(ii.(t), vet)) in 52, respectively. Using Eq. (5),
we get
the angle 0 of intersection of the latter curves on 52 is given by
(EIGI-F12)E2(E2 cos 2 ¢ + 2F2 sin¢cos¢ + G 2 sin 2 ¢)
E2uti + F2 (ut + tiv) + G2 vt
cos () = .2 . . .2 = (E2 G 2 - F:j)E l (El cos 2 ¢ + 2F1 sin 41 cos ¢ + G 1 sin 2 41),
(16)
(E2u + 2F2uv + G2v2 )l12 (E2u + 2F2uv + G 2v )l12
2
>..El uft + >..F1(itt + tlv) + >'G l vt or, setting>.. = (E 2 G 2 - Fi)EtI(E1 G 1 - F{)E2 ,
2
>..Fd sin¢cos¢ + (G 2 - >"Gd sin z 41 = O.
2 2
= (>..E l u + 2>..Fl uv + >"G l v 2 )1/ 2(>..E 1 tl + 2>..F1 tlt + >..G l t )l/2
2 (E2 - >..Ed cos ¢ + 2(F2 -

E l uti + F1 (ut + tv) + G l vt Taking 41 = 0 and then ¢ = IT /2 gives E z = >..E1 , G 2 = >"G 1, and then
= (E1 u + 2F1 uv + G1 v 2 )1/2 (E l t/ + 2Fl fit + Gl t )l/2 '
2
2 substituting in the last equation gives F 2 = >.F1 . 0

since the >"'s cancel. But, using Eq. (5) again, we see that the right-hand side
is the cosine of the angle of intersection of the curves 'Y and .y on 51. Hence, f Example 5.7
is conformal.
We consider the unit sphere x 2 + y2 + Z2 = 1. If P = (u, v, 0) is any point in the
xy-plane, draw the straight line through P and the north pole N = (0,0,1).
110 Elementary Differential Geometry 111

This line intersects the sphere at a point Q, say. Every point Q of the sphere As to 0'1, we get
arises as such a point of int.ersection, with the sole exception of the north pole 2(V
2
- u2 + 1) -4uv 4U)
itself. (ud" = ( (u2 + v2 + 1)2 , (u 2 + v 2 + 1)2' (u2 + v2 + 1)2 '
0' _ ( -4uv 2(u 2 - v 2 + 1) 4v )
N ( d.. - (u2 + v2 + 1)2' (u:.! + v 2 + 1)2' (u 2 + v 2 + 1)2 .
This gives
4(v 2 - u2 + 1)2 + 16u 2 v 2 + 16u 2
E = (ud"·(O'd,, =
1 (u 2 + v 2 + 1)4

=
which simplifies to E 1 4/(u 2 +v 2 + 1)2. Similarly, PI = 0, G 1 = E l . Thus, the
first fundamental form of 0'2 is .\ times that of 0'1, where .\ = t(u 2 + v2 + 1)2.

EXERCISES

5.9 Show that every isometry is a conformal map. Give an example of a

conformal map that is not an isometry.
The vector NQ is parallel to the vector NP, so there is a scalar, say p, such
5.10 Show that the curve on the COne in Exercise 5.2 intersects all the
that the position vector q of Q is related to those of Nand P by
rulings of the cone at the same angle.
q - n = pep - 0), 5.11 Show that Mercator's parametrisation of the sphere
and hence D'(u, v) = (sech u cos v, sech u sin v, tanh u)
q= (O,O,I)+p«u,v,O) - (0,0,1)) = (pu,pv,l-p). is conformal.
Since Q lies on the sphere, 5.12 Let lex) be a smooth function and let
p2 U2 + p2 V2 + (1 _ p)2 = 1 O'(u,v) = (ucosv,usinv,f(u))
which gives p = 2/(u 2 + v 2 + 1) (the other root p = 0 corresponds to the other be the surface obtained by rotating the curve z = f(x) in the xz-
intersection point N between the line and the sphere). Hence, plane around the z-axis. Find all functions f for which 0' is confor-
2 2 mal.
2u 2v U +V -1)
q = ( u 2 + v2 + 1 ~ u 2 + v 2 + l' u 2 + v 2 + 1 . 5.13 Let 0' be the ruled surface
If we denote the right-hand side by O'l(U,V), then 0'1 is a parametrisation of O'(U, v) = "'(u) + v6(u),
the whole sphere minus the north pole. Parametrising the plane z = 0 by
where "'( is a unit-speed curve in R 3 and 6(u) is a unit vector for all
0'2(U,V) = (u,v,O), the map that takes Q to P takes O'l(U,V) to 0'2(U,V).
This map is called stereographic projection. We are going to prove that it is u. Prove that 0' is conformal if and only if 6(u) is independent of u
conformal. and 'Y lies in a plane perpendicular to 6. What kind of surface is 0'
in this case?
According to Theorem 5.5, we have to show that the first fundamental forms
of 0'1 and 0'2 are proportional. The first fundamental form of 0'2 is du 2 + dv 2 . 5.14 Show that the surface patch
O'(u,v) = (J(u,v),g(u,v),O),
112 Elementary Differential Geometry 5. The First Fundamental Form 113

where 1 and 9 are smooth functions on the uv-plane, is conformal if Recalling that the area of a parallelogram in the plane with sides a and b is
and only if either II a x b II, we see that the area of the parallelogram on the surface is approxi-
mately
111. = 911 and 111 = -911.,
or
II O'u Llu x O'v Llv [I = II au x O'v II Llu.;lv.

/11. = -gv and /v = guo This suggests the following definition.

The first pair of equations are called the Cauchy-Riemann equa·

tions; they are the condition for the map from the complex plane to Definition 5.3
itself given by u + iv f.4 I(u, v) + ig(u, v) to be holomorphic. The The area k(R) of the part O'(R) of surface patch a : U --+ R 3 corresponding
second pair of equations says that this map is anti-holomorphic, i.e. to a region R ~ U is
that its complex-conjugate is holomorphic. We shall say more about
holomorphic functions in relation to surfaces in Section 9.4. k(R) = JLII O'u x O'v II dudv.

Of course, this integral may be infinite - think of the area of a whole plane,
5.4. Surface Area for example. However, the integral will be finite if, say, R is contained in a
rectangle that is entirely contained, along with its boundary, in U.
Suppose that q : U -+ R 3 is a surface patch on a surface S. The image of The quantity II O'u X a v II that appears in the definition of area is easily
a is covered by the two families of parameter curves obtained by setting u = computed in term~ of the first fundamental form Edu 2 + 2Pdudv + Gdv 2 of 0':
constant and v = constant, respectively. Fix (uo,vo) E U, and let.1u and Llv
be very small. Since the change in 0'(u, v) corresponding to a small change Llu
in u is approximately O'uLlu and that corresponding to a small change Llv in v is
Proposition 5.2
approximately O'v Llv, the part of the surface contained by the parameter curves II00u x O'v II = (EG - P 2)1/2.
in the surface corresponding to u = Uo, U = Uo + Llu, v = Vo and v = Vo + Llv
is almost a parallelogram in the plane with sides given by the vectors O'u Llu
Proof 5.2
and O'vLlv (the derivatives being evaluated at (uQ,vo)):
We use a result from vector algebra: if a, b, c and d are vectors in R 3 , then
(a X b).(c x d) = (a.c)(b.d) - (a.d)(b.c).
Applying this to II au x O'v 11 2 = (O'u X (1v).(O'u X (1v), we get
II 0'11. X 0'11 11 =
2
((1U'O'U)(O'11'O'V) - (O'u.(1v)2 = EG ~ p 2. o

Note that, for a regular surface, EG - p2 > 0 everywhere, since for a regular
surface 0'11. x O'v is never zero.
Thus, our definition of area is

k(R) = JL (EG - p 2 )1/2dudv. (8)

We sometimes denote (EG - P2)1/2dudv by dAq. But we have still to check

that this definition is sensible, Le. that it is unchanged if 0' is reparametrised.
114 Elementary Differential Geometry 5. The First Fundamental Form 115

This is certainly not obvious, since E, F and G change under reparametrisation 5.16 A surface is obtained by rotating about the z-axis a unit-speed curve
(see Exercise 5.4). "( in the xz-plane that does not intersect the z-axis. Using the stan~
dard parametrisation of this surface, calculate its first fundamental
form, and deduce that its area is
Proposition 5.3
The area of a surface patch is unchanged by reparametrisation. 27f / p( u) du,

where p(u) is the distance of "(u) from the z-axis. Hence find the
Proof 5.3 area of
Let u : U -t R 3 be a surface patch and let iT : U -t R 3 be a. reparametrisation (i) the unit sphere;
of u, with reparametrisation map ~ : U -t U. Thus, if 4i( i/., v) = (u, v), we have (ii) the torus in Exercise 4.10.
5.17 Let "(8) be a unit-speed curve in R 3 with principal normal n and
a(ii.,v) = u(u,v).
binormal b. The tube of radius a > 0 around "( is the surface
Let R ~ (j be a region, and let R = ifJ(R) ~ U. We have to prove that parametrised by

!hi' Uu x UU II dudv = !kII iTf> x ail II dii.dv. u(8,8) = "(8) + a(n(s) cos 8 + b(s) sin 0).
Give a geometrical description of this surface. Prove that u is regular
We showed in the proof of Proposition 4.2 that
if the curvature J\, of"( is less than a-I everywhere. Assuming that
au x ail = det(J(tll) Uu x uu, this condition holds, prove that the area of the part of the surface
given by'so < 8 < Sl, 0 < e < 27f, where 80 and 81 are constants, is
where J(~) is the jacobian matrix of ifJ. Hence,
21ta(sl - 80).

!kII au x ail II di/.dv = !kIdet(J(~»lll Uu x UU II di/.dv.

By the change of variables formula for double integrals, the right-hand side of
this equation is exactly

/hi' Uu x Uv 11 dudv. o

This proposition implies that we can calculate the area of any surface S by
breaking S up into pieces, each of which are contained in a single surface patch,
calculating the area of each piece using Eq. (8), and adding up the results (cf.
Section 11.3, where an analogous procedure is carried out).

EXERCISES
The tube around a circular helix
5.15 Determine the area of the part of the paraboloid z = x 2 + y2 with
Z :S 1 and compare with the area of the hemisphere x + y2 + Z2 = 1,
2

z < o.
 Elementary Differential Geometry 5. The First Fundamental Form 117
116

z
5.5. Equiareal Maps and a Theorem of
Archimedes
We are going to use the formula (8) for the area of a surface to prove a theorem
due to Archimedes which, legend has it, was inscribed onto his tombstone by the
Roman general Marcellus who led the siege of Syracuse in which Archimedes
perished. Naturally, since calculus was not available to him, Archimedes's proof
of his theorem was quite different from ours. From his theorem, we shall deduce
a beautiful formula for the area of any triangle on a sphere whose sides are arcs
of great circles.
In modern language, the Theorem of Archimedes asserts that a certain map
between surfaces is equiareal, in the following sense:

Definition 5.4
Let 8 1 and 8 2 be two surfaces. A diffeomorphism f : 8 1 -t 8 z is said to be
equiareal if it takes any region in 8 1 to a region of the same area in 8 2 ,
To find a {orrI\ula for f, let (x, y, z) be the cartesian coordinates of P, and
We have the following analogue of Theorem 5.1. (X, Y, Z) those of Q. Since the line PQ is parallel to the xy-plane, we have Z = z
and (X, Y) = -X(x, y) for some scalar .\. Since (X, Y, Z) is on the cylinder,
Theorem 5.3 1 = X 2 + y Z = .\2(X Z +y2),
A diffeomorphism f : 8 1 -t 82 is equiareal if and only if, for any surface patch .\ = ±(x2 + yZ)-1/2.
0'( u, v) on 8 1 , the first fundamental forms Taking the + sign gives the point Q, so we get
z
E 1 du 2 + 2F1 dudv + G 1 dv 2 and E 2 du z + 2F2 dudv + G 2 dv
of the patches (1 on 8 1 and f 0 (1 on 8 2 satisfy
f(x, y, z) = Cx 2 +Xy2 )l/2' (x 2 +Yy2 )1/2 I z) .
We shall show in the proof of the next theorem that f is a diffeomorphism.
E 1G 1 - F~ =: E 2GZ - F:j. (9)

Theorem 5.4 (Archimedes's Theorem)

The proof is very similar to that of Theorem 5.1 and we leave it as Exercise
5.22. The map f is equiareal.
For Archimedes's theorem, we consider the unit sphere x Z + y2 + z2 = 1 and
the cylinder x 2 + yZ = 1. The sphere is contained inside the cylinder, and the Proof 5·4
two surfaces touch along the circle x Z + y2 = 1 in the xy-plane. For each point
P on the sphere other than the poles (0,0, ±1), there is a unique straight line We take the atlas for the surface Sl consisting of the sphere minus the north
parallel to the xy-plane and passing through the point P and the z-axis. This and south poles with two patches, both given by the formula
line intersects the cylinder in two points, one of which, say Q, is closest to P.
Let f be the map from the sphere (minus the two poles) to the cylinder that
°
a1(O,tp) = (cos cos tp, cos 9sintp,sinO),

takes P to Q. and defined on the open sets

{-7r!2< (J<7r!2, O<tp<27r} and {-7r!2<O<rr!2, -rr<tp<rr}.
118 Elementary Differential Geometry .+ 3. 'Ii@ r Ii §t , ahUlY'ii@hCJi , 61 iii ii9

The image of tJ1 (0, rp) under the map f is the point
(f2(O,({i) = (cos({i,sin({i,sinO) (10)
of the cylinder. It is easy to check that this gives an atlas for the surface 52,
consisting of the part of the cylinder between the planes z :::: 1 and z :::: -1,
with two patches, both given by Eq. (10) and defined on the same two open
sets as tJi. We have to show that Eq. (9) holds.
We computed the coefficients E 1 , Ft and G 1 of the first fundamental form
of tJ1 in Example 5.3:

For tJ2, we get If we now apply the isometry which unwraps the cylinder on the plane,
(tJ2)6:::: (O,O,cosO), (0'2)'1':::: (-sin<p, cosrp, 0), this curved rectangle on the cylinder will map to a genuine rectangle on the
2
plane, with width 0 and height 2. By Archimedes' theorem, the lune has the
E2 :::: cos 0, F2 :::: 0, G 2 :::: 1. same area as the curved rectangle on the cylinder, and since every isometry
It is now clear that Eq. (9) holds. is an equiareal map (see Exercise 5.18), this has the same area as the genuine
Note that, since / corresponds simply to the identity map (8, i.p) H- (0,1,0) in rectangle in the plane, namely 2{). Note that this gives the area of the whole
terms of the parametrisations 0'1 and (12 of the sphere and cylinder, respectively, sphere to be 4'1l'.
it follows that / is a diffeomorphism. 0
\
Theorem 5.5
Example 5.8 Let ABC be a triangle on a sphere of unit radius wh08e side8 are arcs of great
We use Archimedes's theorem to compute the area of a 'Iune', i.e. the area circles. Then, the area 0/ the triangle is
enclosed between two great circles:
LA + LB + LC - IT,

, ~-
where LA is the angle of the triangle at A, etc.

Proof 5.5
The three great circles, of which the sides of the triangle are arcs, divide the
sphere into 8 triangles, as shown in the following diagram (in which A' is the
antipodal point of A, etc.). Denoting the area of triangle ABC by A(ABC),
etc., we have, by Example 5.8,
., A(ABC) + A(A' BC) :::: 2LA,
~ .. ~
A(ABC) + A(AB'C) :::: 2LB,
We can assume that the great circles intersect at the poles, since this can be
A(ABC) + A(ABC' ) :::: 2LC.
achieved by applying a rotation of the sphere, and this does not change areas
(see Exercise 5.3). If () is the angle between them, the image of the lune under
the map / is a curved rectangle on the cylinder of width 8 and height 2:
---~.~::::::::::::::::::::::::::::~~~::::::~E~;~@'~ii@~,!!a~a'!9:!b~;~"@~i~@j~It~;a!i~G!@6~I!!ii@~E~'Yr---T S. i he Plrst Pundamental Form 121

EXERCISES

5.18 Show that every isometry is an equiareal map. Give an example of

an equiareal map that is not an isometry.
5.19 Show that a map between surfaces that is both conformal and
equiareal is an isometry.
5.20 A sailor circumnavigates Australia by a route consisting of a triangle
whose sides are arcs of great circles. Prove that at least one interior
angle of the triangle is 2:: j + 116°9 radians. (Take the Earth to be a
sphere of radius 650Qkm and assume that the area of Australia is
7.5 million square km.)
5.21 The unit sphere in R 3 is covered by triangles whose sides are arcs of
great circles, and such that the intersection of any two triangles is
either empty or a common edge or vertex of each triangle. Suppose
that there are F triangles, E edges (a common edge of two triangles
being counted only once) and V vertices (a common vertex of several
triangles being counted only once). Show that 3F :::: 2E. Deduce
from Theorem 5.5 that 2V -F = 4. Hence show that V -E+F:::: 2.
(This result will be generalised in Chapter 11.)
Adding these equations, we get 5.22 Prove Theorem 5.3.
2A(ABC) + {A(ABC) + A(A' BC)+A(AB' C) + A(ABC')}
(11)
:::: 2LA + 2LB + 2LC.
Now, the triangles ABC, AB'C, AB'C' and ABC' together make a hemisphere,
so
A(ABC) + A(AB'C) + A{AB'C') + A(ABC' ) :::: 211". (12)
Finally, since the map which takes each point of the sphere to its antipodal
point is clearly an isometry, and hence equiareal (Exercise 5.18), we have
A(A' Be) :::: A(AB'C').
Inserting this into Eq. (12), we see that the term in { } on the right-hand side
of Eq. (11) is equal to 211". Rearranging now gives the result. 0

In Chapter 11, we shall obtain a far-reaching generalization of this result

in which the sphere is replaced by an arbitrary surface, and great circles by
arbitrary curves on the surface.
 6
Curvature of Surfaces

In this chapter, we introduce several ways to measure how 'curved' a surface is.
All of these rest ultimately on the second fundamental form of a surface patch.
It turns out (see Theorem lOA) that a surface patch is determined up to a rigid
motion of R 3 by its first and secorid fundamental forms, just as a unit-speed
,J plane curve is determined up to a rigid motion by its signed curvature.

6.1. The Second Fundamental Form

To see how we might define the curvature of a surface, we start by finding a
new interpretation of the curvature of a plane curve. Suppose then that "/ is a
unit-speed curve in R 2 • As the parameter t of "/ changes to t + .dt, the curve
moves away from its tangent line at ,,/(t) by a distance (,,/(t + .dt) - ,,/(t)).n,
where n is the principal normal to "/ at "/(t). By Taylor's theorem,

,,/(t + .dt) = ,,/(t) + -y(t).dt + ~1'(t)(.dt)2 + remainder,

where (remainder)/(.dt)2 tends to zero as .dt tends to zero. Now, n is perpen-

dicular to the unit tangent vector t = -y, and l' = i = "n, where" is the
curvature of 'Y. Hence, ;Y.n = K, and the deviation of'Y from its tangent line is

(-y(t)Llt + ~1'(t)(Llt)2 + .. ·).n = ~"(Llt)2 + remainder. (1)

123
 6. CuMtdPe Of SdPf5EE5 liS
124 Elementary Differential Geometry .¥
is the analogue for the surface of the curvature term K,(6.t)2 in the case of a
curve. One calls the expression
Ldu 2 + 2Mdudv + N dv 2 (4)
the second fundamental form of 0'. As in the case of the first fundamental form,
we regard the expression (4) simply as a convenient way of keeping track of
the three functions L, M and N. We shall soon see that a knowledge of these
functions (together with that of the first fundamental form) will enable us to
compute the curvature of any curve on the surface 0'.

")'(') Example 6.1

Now let tI be a surface patch in R 3 with standard unit normal N. As the
Consider the plane
parameters (u,v) of tI change to (u+,1u,v+Llv), the surface moves away from
its tangent plane at tI(u, v) by a distance 0'(1.£, v) = a + up + vq
(tI(u + ,1u, V + ,1v) - O'(u, v)).N. (see Example 4.1). Since tlu =
P and O'v =
q are constant vectors, we have
O'uu = U uv = tl vv = O. Hence, the second fundamental form of a plane is zero.

N
Example 6.2
"'(,. + <lu,v + <lv)
",---- ......
, Consider a patch tI on a surface of revolution:
,,
,, O'{ U , v) = (f(1.£) cos v, f( u) sin v, g(u)).
, ,,
, ,, Recall from Example 4.12 that we can assume that f(u) > 0 for all values of u
,, and that the profile curve u t-+ (J(u),O,g(1.£)) is unit-speed, i.e. j2 + Ii = 1 (a
"'(u, v)
dot denoting d/du). Then:
Uu = (j cos v, j sin v,iJ), O'v = (- f sin v, f cosv, 0),
By the two variable form of Taylor's theorem, tI(u + ,1u,v + .dv) - tI(1.£,v) is E = II 2
0'" 11 =j2 + il = I, F = tlu.O'v = 0, G = II O'v 11 = f2,
2

equal to 0' u X 0' v = (- f iJ cos v, - f iJ sin v, f j),

1 x U v II = f (since j2 + Ii = I),
II tlu
tl u ,1u + O'v Llv + '2 (tl uu (.d1.£)2 + 2t1 uv ,1w1v + O'vv(LlV)2) + remainder,
tlu X a v (. .. f')
where (remainder)/«Ll1.£)2 + (LlV)2) tends to zero as (.d1.£)2 + (LlV)2 tends to N = II II = -gcosv,-gsmv, ,
tlu X O'v
zero. Now tlu and tl v are tangent to the surface, hence perpendicular to N, so
tl uu = (j'cosv,lsinv,ii),
the deviation of tI from its tangent plane is
tl uv = (-jsinv,jcosv,O),
'12 (L(Llu)2 + 2M Llu,1v + N(.dV)2) + remamder,
.
(2) a vv = (- f cosv, - f sin v, 0),
where L = tluu.N = j§ - jiJ, M = O'uv· N = 0, N = tluv·N = f9,
L = tluu.N, M = O'uv.N, N = O'vv.N. (3) so the second fundament~l form is
Comparing Eq. (2) with Eq. (1), we see that the expression
L(Ll1.£)2 + 2M ,1u.dv + N(Llv)2
126 Elementary Differential Geometry
6. Curvature of Surfaces 127

If the surface is the unit sphere, we can take feu) = casu, g(u) = sinu,
6.2. The Curvature of Curves on a Surface
with -11"/2 < u < 11"/2, giving
£=1, M=O, N=cos 2 u. Another natural way to investigate how much a surface curves is to look at
the curvature of various curves on the surface. If 'Y(t) ::::; O'(u(t) ,vet)) is a unit-
(Note that the conditions f > 0 and j2 + fP = 1 are satisfied.) Replacing u
speed curve in a surface patch 0', then l' is a unit vector and is, by definition, a
and V by the more usual () and <p, we get the second fundamental form of the
tangent vector to 0'. Hence, i' is perpendicular to the standard unit normal N
unit sphere:
of 0', so 1', N and N x l' are mutually perpendicular unit vectors. Again since
d(J2 + cos 2 () d<p2. 'Y is unit-speed, l' is perpendicular to 1', and hence is a linear combination of
If the surface is a circular cylinder of unit radius, we can take feu) ::::; 1,
Nand N x 1'=
g(u) ::::; u (again, the conditions f > 0 and j2 + iP = 1 are satisfied). This gives (5)
L=M=O, N= 1,
so the second fundamental form of the cylinder is dv 2 •

EXERCISES

6.1 Compute the second fundamental form of the elliptic paraboloid Nx1'
O'(u, v) = (u, V, u 2
+v 2
).
I
I
I
I
6.2 The second fundamental form of a surface patch 0' is zero everywhere. I
I
Prove that 0' is part of a plane. (By computing expressions such as
(O'u.N)u, prove that N u and Nt! lUe perpendicular to O'u and 0'11, and
deduce that the unit normal N of 0' is a constant vector.) This is the
analogue for surfaces of the theorem that a curve with zero curvature
everywhere is part of a straight line.
'Y
6.3 Let a surface patch iT(u, v) be a reparametrisation of a surface patch The scalars K n and Kg are called the normal curvature and the geodesic curva-
=
O'(u, v) with reparametrisation map (u, v) !l'(u, v). Prove that ture of 'Y, respectively. Since Nand N x l' are perpendicular unit vectors, Eq.
(5) implies that
(i %) = (t ±J
t
~) J, I\:n = i·N, I\:g = i·(N x 1')
where J is the jacobian matrix of !l' and we take the plus sign if and
det(J) > 0 and the minus sign if det(J) < O.
II i 11 =
2
K; + 1\:;.
6.4 Show that the second fundamental form of a surface patch is un-
changed by applying a rigid motion to the patch. Hence, the curvature K = II i 11 of'Y is given by
..2
",
= /\,2
n
+ /\,2g' (6)
 6. Curvature of Surfaces 129
128 Elementary Differential Geometry

Moreover, if n is the principal normal of "Y, so that l' = Kn, we have EXERCISES
I\,n = Kn.N = KCOStP, (7) 6.5 Compute the normal curvature of the circle 'Y(t) = (cos t, sin t, 1) on
where'lj; is the angle between nand N. Then, from Eq. (6), the elliptic paraboloid 0'(1.£, v) = (u, V, u 2 + v 2 ) (see Exercise 6.1).
6.6 Show that if a curve on a surface has zero normal and geodesic
Kg = ±Ksin'lj;. (8)
curvature everywhere, it is part of a straight line.
It is clear from their definition that K n and Kg either stay the same or both 6.7 Show that the normal curvature of any curve on a sphere of radius
change sign when (f is reparametrised (since this is the case for N). r is ±1/r.
If "Y is regular, but not necessarily unit-speed, we define the geodesic and
6.8 Compute the geodesic curvature of any circle on a sphere (not nec-
normal curvatures of"Y to be those of a unit-speed reparametrisation of 'Y. When
essarily a great circle).
a unit-speed parameter t is changed to another such parameter ±t + c, where
c is a constant, it is dear that K n t-+ K n and Kg t-t ±Kg , so K n is well defined 6.9 Consider the surface of revolution
for any regular curve, while Kg is well defined up to sign. Equations (7) and (8) 0'(1.£, v) = (f(u) cos v, I(u) sin v, g(u)),
continue to hold in this more general situation.
where u t-+ (f(u), 0, g(u) is a unit-speed curve in R 3 . Compute the
An important special case is that where'Y is a normal section of the surface,
Le. 'Y is the intersection of the surface with a plane II that is perpendicular to geodesic curvature of
the tangent plane of the surface at every point of "Y. (i) a meridian v = constant;
(ii) a parallel u = constant.
6.10 A unit-speed curve -y with curvature K > 0 and principal normal n
forms the intersection 9f two surfaces Sl and S2 with unit normals
N 1 and N z . Show that, if K1 and K.2 are the normal curvatures of 'Y
when viewed as a curve in Sl and S2, respectively, then
I\,lN2 - K2Nl = K.(N 1 x N 2) x n.
Deduce that, if a is the angle between the two surfaces,
K.
2
sin 2 a = K.I + K~ - 2K.1K.Z cos a.

6.11 Let 'Y be a unit-speed curve on a surface patch (f with curvature

I\, > O. Let 'If; be the angle between l' and N, and let B = t x N (in

the usual notation). Show that

N = n cos 'lj; + b sin 'If;, B = b cos 1/J - n sin 'If;.
Since 'Y lies in II, the principal normal n is parallel to II, and since II is
perpendicular to the tangent plane, N is also parallel to II. Since n and N are Deduce that
both perpendicular to "y, and since "y is parallel to II, n and N must be parallel t = KnN - KgB, N = -I\:nt + TgB, B = I\:gt - TgN,
to each other, i.e. 'lj; = 0 or 1r. From Eqs. (7) and (8), we deduce that
where T g = T + ¢. (Tg is called the geodesic torsion of -y; cr. Exercise
K. n = ±K., K. g =0 8.4.)
for a normal section. 6.12 A curve 'Y on a surface S is called asymptotic if its normal curvature
We shall now study the normal curvature K. n in more detail. The study of is everywhere zero. Show that any straight line on a surface is an
Kg will be taken up in Chapter 8. asymptotic curve. Show also that a curve 'Y with positive curvature
 y=
130 Elementary Differential Geometry I 6. Curvature of Surfaces 131
I

is asymptotic if and only if its binormal b is parallel to the unit using the definition (3) of L, M and N, and the fact that N is perpendicular
normal of S at all points of "I. to (lu and U v ' 0
6.13 Prove that the asymptotic curves on the surface
This proposition implies the following classical result, which takes longer to
u(u, v) = (u cos v, usinv, in u) state than to prove.
are given by
In u = ±(v + c), Proposition 6.2 (Meusnier's Theorem)
where c is an arbitrary constant. Let P be a point on a surface S and let v be a unit tangent vector to S at P.
Let IIe be the plane containing the line through P pamllel to v and making an
6.14 Show that an asymptotic curve with positive curvature has torsion
angle (J with the tangent plane to S at P. Suppose that IIe intersects S in a
equal to its geodesic torsion (see Exercise 6.11). (Show that B is
curve with curvature Kf). Then, K.e sin (J is independent of ().
parallel to n.)

IIe
6.3. The Normal and Principal Curvatures v

The most important single fact about the normal curvature K. n of a curve "I on
a surface iT is contained in

Proposition 6.1
If'Y(t) = u(u(t),v(t)) is a unit-speed curve on a surface patch (1, its normal
curvature is given by
K.n = Lu 2 + 2Muv + NiP,
where Ldu 2 + 2Mdudv + Ndv 2 is the second fundamental form of (1,

This result means that two unit-speed curves passing through a point P
on a surface and with the same tangent vector at P have the same normal
curvature at P, since both K n and the tangent vector 'Y = u u it + (1 v iJ depend
only on u, v, u and v (and not on any higher derivatives of u and v).
Proof 6.2

Proof 6.1 Assume that "If) is a unit-speed parametrisation of the curve of intersection of
IIf) and S. Then, at P, 'Ye = ±v, so 1'e is perpendicular to v and is parallel to
We have, with N denoting the standard unit normal of u, IIf). Thus, in the notation of Section 6.1, 't/J = rr/2 - () and so Eq. (7) gives

Kn
··
= N ."1 =
Nd(.) d( . .
. dt "I = N. dt uuu + uvv) Kf) sin () =Kn·
= N.(uuu + (1vV + ((1uuU + uuvv)u + (uuvu + (1vvv)v) But Kn depends only on P and v, and not on (J. o
= Lu2 + 2Muv + Nv 2 ,
 132 Elementary Differential Geometry 6. Curvature of Surfaces 133

To analyse "'n further, it is useful to use matrix notation. If the principal curvatures are always real. To motivate this result, note that. if :.Fr
happens to be the identity matrix (as it is for the standard parametrisation of
Edu 2 + 2Fdudv + Gdv 2 and Ldu 2 + 2Mdudv + Ndv 2
the plane, for example), Eq. (11) would become the equation for the eigenvalues
are the first and second fundamental forms of a surface tT, we introduce the of FII. But a standard result from linear algebra states that the eigenvalues
following symmetric 2 x 2 matrices: of any real symmetric matrix, such as :Fr I, are real numbers. In general, F I is
not the identity matrix, but it is always invertible (see the remark following
Proposition 5.2), so Eq. (11) is equivalent to
Let det(FI(Fi 1 FII - K)) = 0,

t1 = 60'u + 1]l tT v, t2 = 60'u + 1]20'11 det(Fddet (:Fi 1 FII - 11:) = 0,

1
.. det(Fi :Fl I - K) = 0,
be two tangent vectors at some point of 0'. Then,
and hence the principal curvatures are the eigenvalues of :Fi 1 fII· However,
t 1 .t 2 = (60'u+ 1]10'v).(60'u + 7720'11) :Fi 1 F1[ is not usually symmetric, so the above result from linear algebra does
= E66 + F(6772 + 6771) + G1}I'1/2 not immediately imply the reality of the principal curvatures in the general

= (6 '1/d (; ~) (~~) . case.

If II: is one of the principal curvatures, Eq. (11) says that FIl - "'fI is not
Thus, writing invertible, so, assuming that", is real, there is a non-zero 2 x 1 column matrix
T with real number entries such that
(13)
we get
Definition 6.2
(9)
If T = (;) satisfies Eq. (13), the corresponding tangent vector t = ~O' u + '1/0'1)
On the. other han~,. the tangent vector l' = uO'u + VO'VI and if T = (~),
to the surface O'(u, v) is called a principal vector corresponding to the principal
then by usmg ProposltlOn 6.1 we see by a similar matrix calculation that
curvature 11:.
Kn = T t :FlIT. (10)
The justification for the next two definitions will appear in Proposition 6.3 Proposition 6.3
and Corollary 6.1. Let 11:1 and "'2 be the principal curvatures at a point P of a surface patch 0'.
Then,
Definition 6.1 (i) Kl and"'2 are real numbers;
(ii) if "'1 = 11:2 = 11:, say, then F II = K:FI and (hence) every tangent vector to
The principal curvatures of a surface patch are the roots of the equation
0' at P is a principal vector;
det(FII - "':F[) = 0, (11) (iii) if"'l -:j:. /£2, then any two (non-zero) principal vectors t 1 and t2 corre-
i.e. sponding to "'1 and K2, respectively, are perpendicular.

L - ",E M - KF 1_ (12) In case (ii), P is called an umbilic.

IM - ",F N - KG - O.

Proof 6.3
Since (12) is a quadratic equation for ,.., there are two roots. A priori, these
may be complex numbers. However, we shall prove in the next proposition that For (i), let t 1 and t2 be any two perpendicular unit tangent vectors to the
134 Elementary Differential Geometry 9' 6. Curvature of Surfaces 135

surface at P (not yet known to be principal vectors). Define ~i, 1]i and T i , for For (ii), suppose that the principal curvatures are equal, to I'C, say. Then
i = 1,2 as in the discussion preceding Definition 6.1, and let At = A2 = /\, and Eqs.(14) and (15) give

A = (6
1]1
~2).
Tf2
CtFIC == I, CtFIIC = KI,
Ct(FIl - KFI)C = 0,
By multiplying out the matrices, it is easy to check that FII -I'LFI = 0,
AtFIA = (T{FITI T~FIT2) because C and ct are invertible. Obviously then, if T is any 2 x 1 column
T 2FIT1 T 2FIT2 matrix,
= (t 1.t 1 t 1.t 2 ) (byEq.(9))
t2.t1 t2.t 2

=(~ ~),
It follows that every tangent vector to iT at P is a principal vector.
Finally, for (iii) let
since t 1 and t2 are perpendicular unit vectors. Let T. -
t -
({i)
1]i '
gIl = At FIlA.
for i = 1,2. Then, by Eq. (9),
Then, gIl is still (real and) symmetric because
OjI = AfFJI(At)f = AfFIlA = gIl.
and by Eq. (13),
By the theorem from linear algebra referred to above, there is an orthogonal
matrix B (so that B t B = I), say, such that (16)

= (~l ~2)'
Hence,
BtOIlB
TiFIITl = "'1 (t1.t 2), T;FIlT2 = 1'L2(t1.t2). (17)
for some real numbers >'1 and >'2. Let C = AB. Then,
But since Tf FIlT2 is a 1 x 1 matrix, it is equal to its transpose:
CtFIC = Bt(AtFIA)B = BtB = (~ ~), (14) TfFIIT2 = (TfFIlTd = TiFj I T 1 = TiFIlT1 ,
because B is orthogonal, and where the last equality uses the fact that FIl is symmetric. Hence, Eq. (17)
gives
CtFIlC = Bt(AtFIIA)B = BtgIlB = (~1 ~2)' (15)

Now C is invertible (being the product of two invertible matrices), so o

det(FIi - ",FI) = 0 if and only if det( C t (FIl - ",FI )C) = 0,

... det(FIl - /\,FI) == 0 if and only if det ( (>'~ ~2) - I'L ( ~ ~)) == o. Example 6.3
It is intuitively clear that a sphere curves the same amount in every direction,
Hence, the principal curvatures are the roots of and at every point of the sphere. Thus, we expect that the principal curvatures
>'1 - '" 0 I = 0, of a sphere are equal to each other at every point, and are constant over the
I o A2 - K sphere. To confirm this by calculation, we use the latitude longitude parametri-
sation as usual. We found in Example 5.2 that
E = 1, F = 0, G = cos 2 8,
 ay 137
136 Elementary Differential Geometry 6. Curvature of Surfaces

and in Example 6.2 that To find the principal vectors t 1 and t2, recall that ti = £iO'u + r/iO'v, where
L::; 1, M::; 0, N::; cos 2 e. Ti = (~:) satisfies
So the principal curvatures are the roots of

[I ~ K cos 2 e _0" cos2 eI::; 0, i.e.

i.e. K. ::; 1 (repeated root), as we expected. Any tangent vector is a principal For 1';,1 (= 1), we get
vector.

Example 6.4 6 = o.
We consider the circular cylinder of radius one and axis the z-axis, parametrised
in the usual way: So T 1 is a multiple of (~), and hence t 1 is a multiple of 00' u + 10' v = 0' v =
O'(u, v) = (cos v, sin v, u). (- sin v, cos v, 0). Similarly, for "2 (= 0), one finds that T 2 is a multiple of ( ~) ,
We found in Example 5.3 that and hence t 2 is a multiple of Uu = (0,0,1).
E ::; 1, F = 0, G = 1,
As we mentioned above, one reason for introducing the principal curvatures
and in Example 6.2 that and principal vectors is contained in the following result, which shows that, if
L = 0, M = 0, N = 1. we know the principal curvatures and principal vectors of a surface, it is easy
to calculate the normal curvature of any curve on the surface:
So the principal curvatures are the roots of

°
[ O-K. ° 1=0
1-11: '
Corollary 6.1 (Euler's Theorem)

11:(1';, - 1) = 0, Let 'Y be a curve on a surface patch 0', and let 11:1 and K.2 be the principal
curvatures of 0', with non-zero principal vectors t 1 and t2. Then, the normal
K. = 0 or 1.
curvature of'Y is

where () is the angle between -j and t 1 ·

Proof 6.1
~ 1 - 7 tt
We can assume that 'Y is unit-speed. Let t be the tangent vector of -y, and let
t = ~O'u + 'f/U v , T = (~). Suppose first that 1';,1 = "2 = 11:, say. By Proposition
6.3(ii), the normal curvature of'Y is
1';,1'1 = T t FIlT = I';, T t FIT = II: t.t = 11:.

This agrees with the formula in the statement of the corollary, since
K.1 cos2 () + 1';,2 sin2 8 = l';,(cos2 (} + sin2 9) = 1';,.
 4
138 Elementary Differential Geometry 6. Curvature of Surfaces 139

Assume now that K1 =f. K2, so that t 1 and t 2 are perpendicular by Proposi- Proof 6.2
tion 6.3(iii). We might as well assume that t1 and t2 are unit vectors. Let
If the principal curvatures 1t1 and K2 are different, we might as well suppose
ti = {iD'u + rJiD'11' Ti = ( rJi~i ) '
that K1 > K2. Let K n be the normal curvature of a curve 'Y on the surface.
Then, since
for i = 1, 2. Now 2
K n = 1t1 cos 0 + K2 sin (} = 1t1 - (K1 - K2) sin 8,
2 2

"y = COS8t1 + sin8t 2 , it is clear that Itn :5 1t1 with equality if and only if (} = 0 or 1r, i.e. if and only
so if the tangent vector i' of'Y is parallel to the principal vector t1- Similarly, one
cos8(6D'u + 711D'11) + sin8(6D'u + fJ2D'...,) = ~au + 11D'v, shows that Kn ~ K2 with equality if and only if i' is parallel to t 2 .
If K1 = K2, the normal curvature of every curve is equal to K1 by Euler's
... 6 cos 8 + 6 sin 8 = e, 7]1 cos 8 + 7]2 sin8 = 11,
Theorem and every tangent vector to the surface is a principal vector by Propo-
sition 6.3(ii). 0

T = cos8Tl + sin8T2 • We conclude this section with the following computation which will be very
useful on several occasions later in the book.

Proposition 6.4
Let N be the standard unit normal of a surface patch a( u, v). Then,

(20)
where

(: ~) 1
= -:Fi :Fl/.
Hence, by Eq. (10), the normal curvature of'Y is
It n = T t FIlT The matrix :Fi 1:FII is called the Weingarten matrix of the surface patch D',
and is denoted by W.
= (cos8T; + sin OTi)Fl/(cos OT1 + sinOT2 )
= cos 2 0T;:FIIT1 + cosO sin o(T;:Fl/T2 + TjFl/Td + sin 2 (}TiFl/T2 . (18)
Proof 6.4
By Definition 6.2 and Eq. (9),
Since N is a unit vector, we know that N u and N..., are perpendicular to N,
Tl :FIlTj = KiTl:FITj = {OKi
if i = j, (19) hence are in the tangent plane to a, and hence are linear combinations of D' u
otherwise. and a...,. So scalars a, b, c and d satisfying Eq. (20) exist.
Substituting this into Eq. (18) gives the result. o To calculate them, note that N.au = 0 implies, on differentiating with
respect to ti, that
Corollary 6.2 Nu.D'u+N.D'uu = 0,
The principal curvatures at a point of a surface are the maximum and minimum Nu.D'u = -L.
values of the normal curvature of all curves on the surface that pass through Similarly,
the point. Moreover, the principal vectors are the tangent vectors of the curves
giving these maximum and minimum values.
 ¥
140 Elementary Differential Geometry 6. Curvature of Surfaces 141

Taking the dot product of each of the equations in (20) with tlu and tllJ thus 6.19 Show that a curve on a surface is a line of curvature if and only if
gives its geodesic torsion vanishes everywhere (see Exercise 6.11).
-L = aE + bF, - M ;::: cE + dF, 6.20 Two surfaces 8 1 and 8 2 intersect in a curve C that is a line of cur-
-M;:::aF+bG, -N=cF+dG. vature of Sl. Show that C is a line of curvature of S2 if and only if
the angle between the tangent planes of 8 1 and s,~ is constant along
These four scalar equations are equivalent to the single matrix equation
C.
-(t ~);::: (; ~) (: ~) 6.21 Let (f : W ~ R 3 be a smooth function defined on an open subset
W of R 3 such that, for each fixed value of u (resp. v, w), tI(u, v, w)
-FII;::: fl ( : ~) is a (regular) surface patch. Assume also that
(21)
(: ~) = -fIlfI/. o This means that the three families of surfaces formed by fixing the
values of u, v or w is a triply orthogonal system (see Section 4.6).
(i) Show that tlu.tl lJW ;::: tlv.tl uw ;::: (fw.(fuv = O. (Differentiate the
Eqs. (21).)
EXERCISES (ii) Show that, for each of the surfaces in the triply orthogonal
system, the matrices fl and fII are diagonal. (Note that the
6.15 Calculate the principal curvatures of the helicoid and the catenoid,
standard unit normal of the surface obtained by fixing u is
defined in Exercises 4.14 and 4.18, respectively. I.
parallel to tIu, etc.)
6.16 Let -y(t);::: CT(U(t),v(t)) be a regular, but not necessarily unit-speed, (iii) Deduce that the intersection of any surface from one family
curve on a surface CT, and denote d/dt by a dot. Prove that the of the triply orthogonal system with any surface from another
normal curvature of -y is family is a line of curvature on both surfaces. (This is called
Lu 2 + 2Muv + NiP Dupin's Theorem.)
It
n = Eit 2 + 2Fitv + Gv 2 . 6.22 The third fundamental form of a surface patch tI(u, v) is
2 2
6.17 By using the results of Exercises 5.4 and 6.3, show that the principal II N u 11 du + 2Nu ·N v dudv+ 2 2
II N v 11 dv ,
curvatures of a surface either stay the same or both change sign when
where N(u,v) is the standard unit normal at tI(u, v). Let :FIll be
the surface is reparametrised, according to whether the unit normal
the symmetric 2 x 2 matrix associated to the third fundamental form
stays the same or changes sign. Show also that the principal vectors
in the same way as fI and fII are associated to the first and second
are unchanged by reparametrisation.
fundamental forms (see Section 6.3). Show that FIll = fIIFil fH.
6.18 A curve C on a surface 5 is called a line of curvature if the tangent (Use Proposition 6.4.)
vector of C is a principal vector of 5 at all points of C. If 'Y is a
parametrisation of the part of a curve C lying in a surface patch tI
of 5, and if N is the standard unit normal of tI, show that C is a line
of curvature if and only if 6.4. Geometric Interpretation of Principal
N;::: -Ai', Curvatures
for some scalar A, and that in this case the corresponding principal The relative values of the principal curvatures at a point P of a surface patch
curvature is A. (This is called Rodrigues' Formula.) tell us much about the shape of the surface near P. To see this, note first that,
Show that the meridians and parallels of a surface of revolution are by applying a rigid motion of R 3 and a reparametrisation of (f (which does not
lines of curvature. change the shape of the surface), we can assume that
 ¥ 6. Curvature of Surfaces 143
142 Elementary Differential Geometry

if we neglect terms of order greater than two.

(i) P is the origin and u{O, 0) = Pj
We distinguish four cases:
(ii) the tangent plane to (1 at P is the xy-plane;
(iii) the vectors parallel to the x and y-axes are principal vectors at P, corre- (i) K,1 and K,2 are both> 0 or both < O. Then, (22) is the equation of an elliptic
sponding to principal curvatures K,1 and K,2, say. paraboloid (see Proposition 4.5) and one says that P is an elliptic point of the
The unit principal vectors can be expressed in terms of "11. and Uv by surface.
(ii) 1\;1 and K2 are of opposite sign (both non-zero). Then, (22) is the equation of
a hyperbolic paraboloid and one says that P is a hyperbolic point of the surface.
say. Then, if
(iii) One of K,1 and K,2 is zero, the other is non-zero. Then, (22) is the equation
of a parabolic cylinder and one says that P is a parabolic point of the surface.
(iv) Both principal curvatures are zero at P. Then, (22) is the equation of a
the point (x, y, 0) in the tangent plane at P is equal to
plane, and one says that P is a planar point of the surface. In this case, one
X(6CT u + 1}1(1v) + y(6uu + 1}2C1v) = (x6 + Y6)uu + (X1}1 + Y1}2)U v = sU11. + tuv, cannot determine the shape of the surface near P without examining derivatives
say. Thus, neglecting higher order terms, of order higher than the second (in the non-planar case, these terms are small
compared to K,IX2 + K2y2 when x and yare small). For example, the surfaces
1 2
u(s,t) = (1(0,0) + sUu + tu v + 2(s Uuu + 28tuuv + t 2Uvv ) below (the one on the right is called the monkey saddle) both have the origin
as a planar point, but they have quite different shapes.
1 2 a + 2sta + t 2 a ) ,
= (x, y, 0) + 2(8 uu uv vv
all derivatives being evaluated at the origin. Thus, neglecting higher order
terms, the coordinates of u(s,t) are (x,y,z), where
z = u{s, t).N

= !(Ls 2 + 2Mst + Nt 2 )
2

=~(s t)(~ ~)(:).

Now,

(8)t = (x6 ++ Y6)

X1}1 Y1}2
x({I) + (6) =
=
'11
Y
1}2
xTI + yT2 ,
so
1
= 2(xTI + yTz ) FIl(xTI + yT2
t )
z Note that this classification is independent of the surface patch CT, since
T{FIlTI + xy(T{FIlT2 + T~FIITI) + y 2 T!;.FIlT2 )
reparametrising either leaves the principal curvatures unchanged or changes
= !(x
2
2
the sign of both of them.
= 21 (K,lX
2 + K,2Y 2) ,
Example 6.5
using Eq. (19). We conclude that, near a point P of a surface at which the prin-
cipal curvatures are K,1 and K,2! the surface coincides with the quadric surface On the unit sphere, K,1 = K,2 = ±1 (the sign depending on the parametrisation)
so all points are elliptic (and umbilics). On a circular cylinder, K,1 = ±l, K,2 = 0,
z = ~(K,lX2 + K, 2y 2) (22)
 Elementary Differential Geometry ¥ 5. Curvature of Surfaces 145
144

so u(U) is part of the sphere with centre /\,-l a and radius /\,-1.
so every point is parabolic (and there are no umbilics). On a plane, K:l = K:2 = 0 We have now proved the proposition when S is covered by a single surface
so all points are planar (1) (and umbilics).
patch. For an arbitrary surface S, the preceding argument shows that each
patch in the atlas of S is contained in a plane or a sphere. But clearly two
We conclude this chapter with the analogue for surfaces of Example 2.2,
which tells us that a plane curve with constant curvature is part of a circle. overlapping patches must then be part of the same plahe or the same sphere.
It follows that the whole of S is contained in a plane or a sphere. 0

Proposition 6.5
Let S be a (connected) surface of which every point is an umbilic. Then, S is EXERCISES
either part of a plane or part of a sphere.
6.23 Find the elliptic, hyperbolic and parabolic points on the torus de-
scribed in Exercise 4.10.
Proof 6.5
6.24 Show that every point on the surface of revolution
Let U : U ---+ R 3 be a surface patch in the atlas of S with U a (connected) open
subset of R 2 • Let K: be the commmon value of the principal curvatures of (I. u(u, v) = (f(u) cos v, feu) sin v, g(u))
By Proposition 6.3(ii), is parabolic if and only if u is part of a circular cylinder or a circular
cone.
6.25 Show that, if p, q and r are distinct positive numbers, there are
so the Weingarten matrix
i ' exactly four umbilics on the ellipsoid
W;
1
fi FII = (~ ~). x2 y2 Z2
2+2"+2";l.
p q r
By Proposition 6.4,
(23) (Use Proposition 6.3(ii).)

Hence,

so

Since U is regular, (I,. and u" are linearly independent, so the last equation
implies that K:,. = K:" ; O. Thus, K: is constant.
There are now two cases to consider. If K = 0, Eq. (23) shows that N is
constant. Then,
(N.u),. = N.u,. = 0, (N.u),,; N.u" = 0,
so N.(I is a constant, say c. Then (I(U) is part of the plane r.N = c.
If K: f:. 0, Eq. (23) shows that
N; -K:u+a,

where a is a constant vector. Hence,

I 2
I I 2
II u- -a 11 ;11--N 11 = 2'
K: K: K:
 7
Gaussian Curvature and the Gauss Map

We shall now introduce two new measures of the curvature of a surface, called
its gaussian and mean curvatures. Although these together contain the same
information as the two principal curvatures, they turn out to have greater geo-
metrical significance. The gaussian curvature, in particular, has the remarkable
property, established in Chapter 10, that it is unchanged when the surface is
bent without stretching, a property that is not shared by the principal cur-
vatures. In the present chapter, we discuss some more elementary properties
of the gaussian and mean curvatures, and what a knowledge of them implies
about the geometry of the surface.

7.1. The Gaussian and Mean Curvatures

We start by defining two new measures of the curvature of a surface.

Definition 7.1
Let Kl and K2 be the principal curvatures of a surface patch. Then, the gaussian
curvature of the surface patch is

and its mean curvature is

1 .
H = '2(Kl + K2)'
147
148 Elementary DMerentlal Geometry I. UaUSSi5i1 Cdl05tdi@alld til@ daMs tvlap 149

Note Some authors omit the 1/2 in the definition of H, even though this con- Example 7.2
flicts with the usual meaning of 'mean'.
In Example 6.2 we considered the surface of revolution
Note from Exercise 6.17 that the gaussian curvature stays the same when
a surface patch is reparametrised, while the mean curvature either stays the /7(U, v) = (f(u) cosv,j(u) sin v,g(u)),
same or changes sign. It follows that the gaussian curvature is well defined for where we can assume that f > 0 and j' + g' = 1 everywhere (a dot denoting
any surface S. d/du). We found that
It is easy to get explicit formulas for H and K:
E=l, F=O, G=f',
Proposition 7.1 L=jg-jg, M=O, N=fg.
By Proposition 7.1(i), the gaussian curvature is
Let /7( u, v) be a surface patch with first and second fundamental forms
Edu' + 2Fdudv + Gdv' and Ldu' + 2Mdudv + Ndv', LN - M' (ig - Ig)fg
K= EG-F' = f' (1)
respectively. Then,
(,') K -- LN M'. We can simplify this formula by noting that j' + g' = 1 implies (by differenti-
EO F7.1
ating with respect to u) that
(ii) H -- LG2(EG:2)
'MFf,NE.
,
if+gg = 0,
(iii) the principal curvatures are H ± v' H' K.
(ig - jg)g = - PI - jg' = - j(i' + g') = - I,
.. ..
Proof 7.1
By Definition 6.1, the principal curvatures are the roots of
K=-~~ =-t.
L-I<E M-I<FI_o
Example 7.3
IM-I<F N-I<G - ,
.. (L - I<E)(N - I<G) - (M - I<F)' = 0, For a ruled surface, take a patch
(EG - F')I<' - (LG - 2MF + N E)I<+LN - M' = O. /7(u, v) = 'Y(u) + v6(u),
Now, recall that in a quadratic equation aI<' + bl< + c = 0, the sum of the roots
(see Example 4.12). Denoting d/du by a dot, we have
is -b/a and the product of the roots is c/a. So,
LN-M' au = .y + v6, tr1) = 6,
K = 1<,1<, = product of roots = EG _ F' ' trUf} = 6, tl vv = O.
1 1 lLG-2MF+NE
H = '2(1<1 + 1<,) = '2(sum of roots) = 2 EG _ F' . Hence, if N = (/7 u x /7 v )/ II /7u x /7 v II is the standard unit normal of /7, then
M = /7 uv .N = ,s.N and N = O. So
By the definition of H and K, 1<1 and 1<, are the roots of
1<' - 2HI< + K = 0, K = LN - M' = -(,s.N)' < 0
EG - f' EG - F' - ,
i.e. H ± v'H' - K. o
Le. the gaussian curvature of a ruled surface is negative or zero, a result that is
proved by a different method in Exercise 7.8. We shall return to this example
Example 7.1
in Section 7.3.
For the unit sphere, we found in Example 6.3 that 1<1 = 1<, = 1, so K = H = 1.
For a circular cylinder of radius one, we found in Example 6.4 that 1<1 = 1,
K2 = 0, so H = ~, K = O.
150 Elementary Differential Geometry 7. Gaussian Curvature and the Gauss Map 151

EXERCISES 7.10 Use Exercise 7.9 to show that, if 'Y(t) is a curve in a surface patch
0', then along '1,
7.1 Calculate the gaussian and mean curvatures of the surface
N.N + 2HN.-y + K-Y.-Y == o.
O'(u,v)::;; (u + v,u - v,uv)
at the point (2,0,1).
(Note that, if'Y(t) ::;; O'(u(t),v(t)), T = (~), then N.N::;; TtFII1T,
7.2 Calculate the gaussian curvature of the helicoid and catenoid (see etc.) Deduce that the torsion T of an asymptotic curve on a surface
Exercises 4.14 and 4.18). is related to the gaussian curvature K of the surface by T 2 ::;; - K.
(Use Exercise 6.12.)
7.3 Calculate the gaussian curvature of the surface z ::;; f(x, y), where f
is a smooth function.
7.4 In the notation of Example 7.3, show that
(i) if 6 is the principal normal n of'Y or its binormal b, then K ::;; 0 7.2. The Pseudosphere
if and only if 'Y is planar (use Proposition 2.4);
(ii) if'Y is a curve on a surface S and 6 is the unit normal of S, then We have seen in the examples in Section 7.1 some surfaces of zero and constant
K ::;; 0 if and only if 'Y is a line of curvature of S (use Exercise positive curvature. For an example of a surface with constant negative gaussian
6.18.) curvature, however, we have to construct a new surface. To this end, we examine
7.5 Let 0' be the parametrisation of the torus in Exercise 4.10, and let again the surface of revolution
K be its gaussian curvature. Show that
0'( u, v) ::;; (f(u) cos v, feu) sin v, g( u)).
We found in Example 7.2 that its gaussian curvature is
(The 'explanation' of this result will appear in Section 11.3.)
(2)
7.6 Show that the gaussian and mean curvatures are unchanged by ap-
plying a rigid motion, and that the dilation (x,y,z) M a(x,y,z), Suppose first that K 0 everywhere. Then, Eq. (2) gives i
= =
0, so feu) =
where a is a non-zero constant, multiplies them by a- 2 and a-I, au + b for some contants a and b. Since j2 + iJ2 = 1, we get iJ = ±v'l - a 2 (so
respectively. we must have lal :::; 1) and hence g(u) = ±v'1 - a 2 u + e, where c is another
7.7 Show that the gaussian and mean curvatures of a surface patch constant. By applying a translation along the z-axis we can assume that c = 0,
3
0' : U -t R are smooth functions on U. Show that the principal and by applying a rotation about the x-axis (say) we can assume that the sign
curvatures are smooth functions on any open subset of U on which is +. This gives the ruled surface
0' has no umbilics. (Use Proposition 7.1.)
O'(u, v) = (bcosv,bsinv,O) +u(acosv,asinv,~).
7.8 Show that K S 0 at every point of an asymptotic curve on a sur-
face (use Corollary 6.2). Hence give another proof that the gaussian If a=::O this is a circular cylinder; if Jar = 1 it is the xy-plane; and if 0 < lal < 1
curvature of a ruled surface is S 0 everywhere (use Exercise 6.12). it is part of a cone (to see this, put it = au + b).
Now suppose that K = 1 everywhere. (Any surface with constant positive
7.9 Show that, if FIll is the third fundamental form of a surface patch
gaussian curvature can be reduced to this case by applying a dilation of R 3 -
0' (see Exercise 6.22), then
see Exercise 7.6.) Then, Eq. (2) becomes
FIll - 2H:FIl + K:FI ::;; 0,
j' + f = 0,
where K and H are the gaussian and mean curvatures of 0'. (Use
which has the general solution
the fact that any 2 x 2 matrix A == (: ~) satisfies the equation
feu) = acos(u + b),
A2 - (a + d)A + (ad - be)! = 0.)
152 tlementary Differential GMM@£fy t. Gaussian CdMidPe arid tHE Gauss iOiap 153

where a and b are constants. We can assume that b = 0 by performing a Putting x = feu), z = g(u), we see that the profile curve in the xz-plane has
reparametrisation u = u + b, v = v. Then, up to a change of sign and adding a equation
constant,
z = Vi - x 2 - cosh- 1 (~) • (4)
g(u) = ! J1- a2 sin
2
udu.
Rotating this curve around the z-axis thus gives a surface called the pseudo-
This integral cannot be evaluated in terms of 'elementary' functions unless sphere, which has gaussian curvature -1 everywhere. Note that, since u S 0,
a = 0 or ± 1. The case a = 0 does not give a surface, so we consider the case x = e is restricted to the range 0 < x :S 1.
U

a = 1 (the case a = -1 can be reduced to this by rotating the surface by IT

around the z-axis). Then, feu) = casu, g(u) := sinu, and we have the unit
sphere.
Suppose finally that K = -1. The general solution of Eq. (2) is then
feu) = aeU + be- u ,
where a and b are arbitrary constants. For most values of a and b we cannot
express g in terms of elementary functions, so we consider only the case a = 1
and b = O. Then, feu) = eU and we c~ take

g(u) = ! J1 - e2u duo (3)

Note that we must have u ~ 0 for the integral in Eq. (3) to make sense, since
otherwise 1 - e2u would be negative.
The integral in (3) can be evaluated by putting v = eU • Then,
The curve defined by Eq. (4) is called the tractrix, and.it has an interesting
! J1-e 2U du= 17 dv
geometrical property. Consider the tangent line at a point P of its graph, and
suppose that it intersects the z-axis at the point Q. Let us compute the distance

=!(~-v)h PQ.

_ ~
-v1-v~+
1 ~.
dv
vv 1 - v2
z

(1,0)
x
Put w = V-I in the last integral. This gives

~_! p

! V1-e 2U du=

= ~ -cosh-
dw
";w
1
2

w
-1

= ~ -cosh- 1 (~) Q

= J1- e2u - cosh- 1 (e- U ) .

We have omitted the arbitrary constant, but we can take it to be zero by
suitably translating the surface in the direction of the z-axis. So
Suppose that P is the point (xo, zo). Either by a direct calculation or by
feu) =e U
, g(u) = \11 - e2u -
1
cosh- (e-
U
).
154 Elementary bdferentrai Geometry 7. Gaussian Curvature and the Gauss Map 155

inspecting the calculation of the integral (3), one finds that (called the upper half-plane moden;
(ii) setting
dz JI=X2
dx= x v2 + w2 - 1 -2v
U= V=~~'::':"--
Hence, the tangent line at P has equation
v2 + (w + 1)2' v2 + (w + 1)2
defines a reparametrisation a2(U, V) of the pseudosphere with
Z-Zo= ~(x-xo).
Xo
first fundamental form
dU' + dV 2
This meets the z-axis at the point (0, Zl), where (1 - U2 - V2)2'

Zl-ZO = v'1-X5(O-xo ) = ~
-y1-xij.
This is called the disc model, because the region w > 0 of the
Xo vw-plane corresponds to the disc U 2 + V 2 < 1 in the UV -plane.
Hence, the distance PQ is given by Which regions in the half-plane model and the disc model correspond
to the region u < 0, -1r < V < 1r where the parametrisation a of the
(PQ)2 = x5 + (Zl - zO)2 = x5 + 1 - x5 = 1, pseudosphere given in this section is defined?
i.e. the distance PQ is constant and equal to one. 7.13 Let S be a surface of revolution with axis the z-axis, and let its
This means that the tractrix has the following description. Let a donkey pull profile curve be a unit-speed curve ')'(u) in the xz-plane. Suppose
a box of stones by a rope of length one. Suppose that the donkey is initially that')' intersects the z-axis at right angles when u = ±1r /2, but does
at (0,0), the box is initially at (1,0), and let the donkey walk slowly along the not intersect the z-axis when -1r /2 < u < 1r /2. Prove that, if the
negative z-axis. Then, the box of stones moves along the tractrix. gaussian curvature K of S is constant, that constant is equal to one
The study of geometry on a pseudosphere is a subject in its own right called and S is the unit sphere.
non-euclidean geometry. Many of the results of plane euclidean geometry have
analogues for the pseudosphere, but there are several differences. For example,
the sum of the interior angles of a triangle on the pseudosphere with 'straight'
sides is always less than 1'- (We shall explain the meaning of 'straight' here 7.3. Flat Surfaces
in Chapter 8 - see especially Example 8.8.) This should be compared with
Theorem 5.5 which shows that the sum of the interio, angles of a triangle on In Section 7.2, we gave some examples of surfaces of constant gaussian curvature
the unit sphere whose sides are arcs of great circles is always greater than 1T. K, but this certainly falls well short of a complete classification of such surfaces.
It is possible, however, to give a fairly complete description of flat surfaces, i.e.
surfaces for which K = 0 everywhere. To do so, we shall make use of a special
EXERCISES parametrisation, valid for any surface, described in the following proposition.

7.11 For the pseudosphere: Proposition 7.2

(i) calculate the length of a parallel;
Let P be a point of a surface S, and suppose that P is not an umbilic. Then,
(ii) calculate its total area;
there is a surface patch a( u, v) of S containing P whose first and second fun-
(iii) calculate the principal curvatures;
damental forms are
(iv) show that all points are hyperbolic.
7.12 Show that Edu2 + Gdv 2 and Ldu 2 + Ndv 2,
(i) setting w = .-U
gives a reparametrisation al (v, w) of the pseu- respectively, for some smooth functions E, G, Land N.
dosphere with first fundamental form
dv 2 + dw 2 We recall from Section 6.3 that a point P of a surface S is an umbilic if the
w2 two principal curvatures of S at P are equal. From Section 6.3, we see that for
156 Elementary DifFerential Geometry 7. Gaussian Curvature and the Gauss Map 157

the patch (7 in the statement of the proposition, (7u and (7v are principal vectors Since
with corresponding principal curvatures LIE and N IG. We call a a principal au x a v = ~ x 6 + v.s x 6,
patch.
We assume Proposition 7.2 for the moment, and use it to give the proof of and 6.(6 x 6) = 0,
K = 0 if and only if 6.(.y x 6) = o. (6)
Proposition 7.3 Thus, K = 0 if and only if i', 6 and 6 are everywhere linearly dependent.
To proceed further, let us assume, as we may, that 6(u) is a unit vector for
Let P be a point of a fiat surface 5, and assume that P is not an umbilic. Then,
all values of u. Then, 6.6 = O. Suppose first that 6(u) = 0 for all values of u.
there is a patch of 5 containing P that is a ruled surface.
Then, 6 is a constant vector and a is a generalised cylinder.
Suppose now that 6 is never zero. Then, 6 and 6 are linearly independent
Proof 7.3 as they are non-zero and perpendicular, so if 1', 6 and 6 are linearly dependent,
We take a patch a : U ---t R 3 containing P as in Proposition 7.2, say P = then
a(uo,vo). By Proposition 7.1(ii), the gaussian curvature K = LNIEG. Since 1'(u) = f(u)6(u) + g(u)6(u)
the gaussian curvature is zero everywhere, either L = 0 or N =0 at each point for some smooth functions f and g. Assume first that f = 9 everywhere. Then,
of U, and since P is not an umbilic Land N are not both zero. Suppose that
i' = (g6)' and so 'Y = g6 + a, where a is a constant vector; hence,
L(uo, vo) ¥- O. Then, L(u, v) ¥- 0 for (u, v) in some open subset of U containing
(uo, vo). Hence, by shrinking U if necessary, we can assume that L ¥- 0 at every u(u, v) = a + (v + g(u))6(u).
point of U. Then, N = 0 everywhere, and the second fundamental form of (7 is Putting it = u, ii :::: v + g(u), we see that this is a reparametrisation of a
Ldu 2 • generalised cone.
We shall prove that the parameter curves u = constant are straight lines. Suppose finally that 6 and f - 9 are both nowhere zero. If we define
Such a curve can be parametrised by v f-t a(uo, v), where UQ is the constant _ v + g(u)
value of u. A unit tangent vector to this curve is t = tl v lGl/2, so by Proposition 't(u) = 'Y(u) - g(u)6(u) , v = f(u) _ g(u) ,
1.1 what we have to prove is that tv = O.
By Proposition 6.4, the derivatives of the unit normal are a short calculation gives

(5)
a(u, v) = .:y(u) + ii~(u),
so a is a reparametrisation of part of the tangent developable of .:y.
Hence, tv.a u = -EL-1tv.N u . Now, t.N u =
0 and N uv =
0 by Eq. (5), so
tv.N u = -t.N uv = O. Hence, tv.a u = O. Next, tv.t = 0 since t is a unit vector
by construction, so tv.a v = O. Finally, tv.N = -t.N v = 0 by Eq. (5) again.
Since the vectors au, a v and N form a basis of R 3 , we have proved that tv = O.
o
Our task, then, is to describe the structure of flat ruled surfaces. We
parametrise the ruled surface as in Example 7.3:
a(u, v) = 'Y(u) + v6(u).
We found there that au =.y + v6, U v = 6, the dot denoting dJdu, and that the
gaussian curvature of 0' is zero if and only if
6.(a u x a v ) = O. Of course, it could be that none of the conditions on 6, f and 9 considered
above are satisfied. In fact, we have only shown that the parts of the surface
 Elementary Differenti(ll Geometry 7. Gaussian Curvature and the Gauss Map 159
158

corresponding to certain open subsets of U are parts of generalised cylinders, FIl ::::
(
"IE
0 0)
"2
G
l SO the second fundamental form of tr .
IS Ldu 2 + N dv 2 ,
generalised cones or tangent developables. It is not true that the whole surface
must be one of these three types, since flat surfaces of different types can be where L = "1E and N = "2G.
We are thus left with the proof of Proposition 7.4. Let
joined together to make a smooth surface, as shown in the diagram above. In
general, a flat surface is a patchwork consisting of pieces of generalised cylinders, e :::: aiT u + biT v ,
generalised cones and tangent developables, joined together along segments of
where a and b are smooth functions of (ii, ii) E U. We first observe that we
straight lines. can find a curve "y in iT with'" :::: e and with any given point Q :::: &(0:,13) as
The remainder of this section is devoted to the proof of Proposition 7.2
starting point "Y(O). For finding such a curve "Y(t) :::: u(u(t),ii(t)) is equivalent
and can safely be omitted by readers uncomfortable with the use of the inverse
to solving the pair of ordinary differential equations
function theorem. In fact, we can prove a more general result with no additional
effort: ft :::: a(u, ii), fj:::: b(u, ii)
with initial conditions u(O) :::: 0:, v(O) :::: /3. It is proved in the theory of ordinary
Proposition 7.4 differential equations that this problem has a unique solution u(t), ii(t) defined
Let u- : U- -t R 3 be a surface patch, and suppose that for all (u, ii) E U- we are on some open interval containing t = O. Moreover, u and v are smooth functions
given tangent vectors e1 (u, v) and e2(u, ii), whose components depend smoothly of the three variables t, 0: and /3.
on (u,ii). Assume that, at some point (fio,vo) E U, the vectors e1(uo,vO) and
e2(uO, vo) are linearly independent. Then, there is an open subset V of U con-
taining (uo,iio) and a reparametrisation u(u,v) ofiT(u,ii), for (u,ii) E V, such
that tf" and tf" are parallel to e1 and e2, respectively.

Proposition 7.2 is a special case of Proposition 7.4. In fact, let iT be any

surface patch of S containing P, and let P :::: iT(fio, vo). Since the principal
curvatures Kl and K2 of iT are distinct at P, and are continuous functions by
Exercise 7" 7, they remain distinct for (u, ii) in some open set U containing
(uo, va) on which iT is defined. With the notation of Section 6.3, let (~:) be
non-zero eigenvectors of PIl - KiP! for i :::: 1,2, where p! and PIl are the
matrices associated to the first and second fundamental forms of ii" Then, We can thus find a curve "Yl (sd in ii with "Yl (0) :::: ii(uo, iio) and d"Yl/ dsl ::::
e1. Then, for each value of 81 close to 0, we can find a curve 82 r-+ >'(81,S2) in
el :::: 6iT,. + 771 iifJ, e2:::: 6iT,. + 772 iT" iT with 8>./8s 2 :::: e2 and >'(SI'O):::: "Yl(sd. Define (u,ii) as functions of (81,82)
are principal vectors corresponding to K1 and K,2, and they are perpendicular by
by Proposition 6.3(iii). We can assume that el and ez are unit vectors (by (7)
using etlll el 11 and e2/11 ezll instead). Let u(u,v) be a reparametrisation of
f1 as in Proposition 7.4. Then, Uu'U" :::: 0 because el and ez are perpendicular, Differentiating with respect to SI and 82 gives
so the first fundamental form of u is of the form Edu2 + Gdv z . Also, Uu and _ 8u _ Bii , _ au _ aii ,
Cf1,] are principal vectors corresponding to Kl and K2, so we have
tf"-a +u"-8 :::: "Sll
81 81
Cf u -
8 S2 +tr v-
a82 :::: "S2"

(FIl - "1FI) (~) :::: (FI1- K,2FI) (~) :::: (~), We have
(8)
where F 1 and FI1 are the matrices associated to the first and second fun-
damental forms of tf. Since FI :::: (~ ~), these equations imply that
160 Elementary Differential Geometry 7. Gaussian Curvature and the Gauss Map 161

so at the point i1(uo, vo), where 81 = 82 = 0, the jacobian matrix so IT,, and lTv are parallel to el and e2 everywhere.

g5~ g~) =
( .fuL.fuL (ab dc) . (9)
85 1 8'2 EXERCISES
Since el and e2 are linearly independent at (uo, vo), this matrix is invertible.
By the inverse function theorem 4.2, Eq. (7) can be solved for (81,82) as smooth 7.14 Let P be a hyperbolic point of a surface S (see Section 6.4). Show
functions of (u,v) when (u,v) is in some open set Waf (; containing (uo,vo). that there is a patch of S containing P whose parameter curves are
Thus, ~ is an allowable surface patch; by Eq. (8), it has the property that asymptotic curves (see Exercise 6.12).
~51 = el when 82 = 0, and .\52 :::::: e2 everywhere.
We now repeat the procedure, this time starting with a curve 'Y2(t2) with
d'Y2l dt 2 = e2 and 'Y2(O) = i1(uo,vo), and then taking a curve h t-+ P(h,t2)
with aplatl :::::: el and p(O, t2) :::::: 'Y2(t 2). This gives an allowable patch P(tl, t2) 7.4. Surfaces of Constant Mean Curvature
such that
We now consider surfaces whose mean curvature H is constant. As we shall see
p(tl> t2) = i1(u, v) in Chapter 9, such surfaces are encountered in real life as the shapes taken up
for (u, ii) in some open subset Z of (; containing (uo, vo). This patch has the by soap films.
property that Ph :::::: el everywhere and Pt2 :::::: e2 when tl = O. We shall discuss the surfaces for which H is everywhere zero in detail in
The parametrisation we want is D'(u,v), where O'(U, v) is the intersection of Chapter 9. In this section, we are going to describe a construction which gives
the curve 82 t-+ .\(U,82) with the curve h t-+ P(tl,V). Thus, we consider the a correspondence between surfaces of constant non-zero mean curvature and
equations surfaces of constant positive gaussian curvature.

Definition 7.2
iFrom Eq. (9),
Let a be a surface patch with standard unit normal N, and let A be a constant
au af! _ b
-=a, au - , scalar. The parallel surface 0'>" of 0' is
au
and similarly (7)'' = 0'+ AN.
au
av = c, af;
av = d.
Hence, the jacobian matrix
lJU)
!~ =
(a c)
b d .
As usual, the fact that this matrix is invertible means that (u , v) can be ex-
pressed as smooth functions of (u, ii), for (u, ii) in some open subset V of W n Z
containing (uo,iio), and we get a reparametrisation O'(u,v) of u(u,ii). Finally,
the equation O'(u, v) = p(tl, v) implies that
at l atl tT

0'" = au Ph = au el,
and similarly Roughly speaking, IT>' is obtained by translating the surface a a distance
A perpendicular to itself (but this is not a genuine translation since N will in
general vary over the surface).
 162 Elementary Differential Geometry 7. Gaussian Curvature and the Gauss Map 163

Proposition 7.5 The principal curvatures of a>' are the eigenvalues of the Weingarten matrix
Let 1\.1 and 1\.2 be the principal curvatures of a surface patch a : U --; R 3, W>. of a A• By Proposition 6.4, this is the negative of the matrix expressing N~
and suppose that there is a constant C such that 11\.11 and 11\.21 are both ~ C and N; in terms of a~ and a;. Equation (10) says that the matrix expressing
everywhere. Let -\ be a constant with 1-\[ < l/C, and let all. be the corresponding (I~ and (I~ in terms of (I u and 0'11 is I - AW, and the fact that N l\ = N implies
parallel surface of a. Then, that - W is the matrix expressing N ~ and N; in terms of (I u and iT 11' Combining
(i) a), is a (regular) surface patch; these two observations we get
(ii) the standard unit normal of a>' at a),(u,v) is the same as that of a at W A = (I - AW)-lW.
a(u, v), for all (u, v) E U;
If T is an eigenvector of W with eigenvalue 11:, then T is also an eigenvector of
(iii) the principal curvatures of (I), are I\.d(1 - AI\.t} and K2/(1 - AK2), and the
corresponding principal vectors are the same as those of a for the principal
W A with eigenvalue 11:/(1 - AII:). The assertions in part (iii) follows from this.
curvatures 11:1 and K2, respectively;
Part (iv) follows from part (iii) by straightforward algebra. 0
(iv) the gaussian and mean curvatures of a>' are
K H - >.K Corollary 7.1
and
1- 2AH + A2K 1- 2AH + A2K' If 0' is a surface patch with constant non-zero mean curvature H, then for
respectively. A = 1/2H, 0'). has constant gaussian curvature 4H 2 . Conversely, if 0' has
constant positive gaussian curvature K, then for A = ±l/VK, a>' has constant
Proof 7.5 mean curvature =f! vK.
By Proposition 6.4,
Proof 7.1
l\
(I u = au + AN = (1 + Aa) au + Ab a v,
tJ.

>. (5) This follows from part (iv) of the proposition by straightforward algebra. 0
av = a + ANv = AC(lu + (1 + Ad) (Iv,
11

where the Weingarten matrix

W=-(: ~). EXERCISES

Hence, 7.. 15 The first fundamental form of a surface patch O'(u, v) is of the form
(I~ x (I~ = (1 + A(a + d) + A2(ad - bc)) au x (Iv. E(du 2 + dv 2 ). Prove that (luu + O'vv is perpendicular to O'u and (111'
Deduce that the mean curvature H = 0 everywhere if and only if
Since 11:1 and 1\.2 are the eigenvalues of W (see Section 6.3), and since the sum the laplacian
and .product of the eigenvalues of a matrix are equal to the sum of the diagonal
entnes and the determinant of the matrix, respectively, O'uu + a 1111 = O.
1\.1 + K2 = -(a + d), 1\.11\.2 = ad - be. Show that the surface patch
3
Hence, (I(U, v) = ( U - 3u3 + uv 2 ,v - 3v + U 2 V,u 2 - v
2)

(I~ x a; = (1 - AKt}(l - A1I:2) au x (Iv. (6)

has H = 0 everywhere. (A picture of this surface can be found in
Since IAI < 11C and 111:11 and 11\.21 are ~ C, it follows that IAll:1I and 1-\1\.21 are Section 9.3.)
< 1, so (1 - AI\.I)(1 - AK2) > 0, and Eq. (6) shows that (I), is regular and that
7.16 Compute the mean curvature of the surface with cartesian equation
its standard unit normal is
>. ), z =f(x,y)
NA_ (lu X (111 _ (lu X (111
- II (I~ x (I; 11 - II (lu x lTv II = N.
164 Elementary Differential Geometry 7. Gaussian Curvature and the Gauss Map 165

where f is a smooth function of x and y. Prove that H = 0 for the and Q in X such that f(Q) $ f(R) $ f(P) for all points R in X, so that f
surface attains its maximum value on X at P and its minimum at Q.
z = In (cosY).
cos x
Proof 7.6
(A picture of this surface can also be found in Section 9.3.)
7.17 Let .,.(u,v) be a surface patch with first and second fundamental Define f : R 3 ---> R by f(v) = II V 11 2 Then, f is continuous so the fact that S
is compact implies that there is a point P in S where f attains its maximum
forms Edu 2 + Gdv 2 and Ldu 2 + N dv 2 , respectively (cf. Proposition
7.2). Define value. Let P have position vector p; then S is contained inside the closed ball
of radius II p II and centre the origin, and S intersects its boundary sphere at
E(u, v, w) = .,.(u, v) + wN(u, v), P. The idea is that S is at least as curved as the sphere at P, so its gaussian
where N is the standard unit normal of .,.. Show that the three curvature should be at least that of the sphere at P, i.e. at least 1/ II p W·
families of surfaces obtained by fixing the values of u, v or w in E To make this argument precise, let 'Y(t) be any unit-speed curve in S passing
form a triply orthogonal system (see Section 4.6 and Exercise 6.21). through P when t = O. Then, f('Y(t)) has a local maximum at t = 0, so
The surfaces w = constant are parallel surfaces of.,.. Show that the
surfaces u = constant and v = constant are flat ruled surfaces. :/('Y(t)) = 0, ::2 f('Y(t)) $ 0

at t = 0, Le.

'Y(0).-y(0) = 0, 'Y(O)·i(O) + 1 $ o. (12)

7.5. Gaussian Curvature of Compact Surfaces The equation in (12) shows that p = "1(0) is perpendicular to every unit tangent
vector to S at P, and hence is perpendicular to the tangent plane of Sat P.
We have seen in Section 6.4 how the relative signs of the principal curvatures Choose a surface patch .,. in the atlas of S containing P, and let N be its
at a point P of a surface S determine the shape of S near P. In fact, since standard unit normal. By the preceding remark,
the gaussian curvature K of S is the product of its principal curvatures, the p
discussion there shows that N = ±TIPli' (13)

(i) if K > 0 at P, then P is an elliptic point; The inequality in (12) implies that the normal curvature "n = i(O).N of "1 at P
(ii) if K < 0 at P, then P is a hyperbolic point; (computed in the patch.,.) is $ -1/ II p II or ::': 1/ II p II, according to whether
(iii) if K = 0 at P, then P is either a parabolic point or a planar point (and in the sign in Eq. (13) is + or -, respectively. By Corollary 6.2, the principal
the last case we cannot say much about the shape of the surface near P). curvatures of.,. at P are either both $ -1/ II p II or both 2': 1/ II p II· In each
In this section, we give a result which shows how the gaussian curvature case, K ::': 1/ II p II' > 0 at P. 0
influences the overall shape of a surface. We shall give another result of a similar
nature in Section 10.4.

Proposition 7.6
7.6. The Gauss Map
If S is a compact surface, there is a point P of S at which its gaussian curvature Proposition 2.2 shows that, if 'Y(s) is a unit-speed plane curve, its signed cur-
K is > O. vature ". = d<p/ds, where <p is the angle between its tangent vector -y and a
fixed direction, i.e. the (signed) curvature is the rate of change of direction of
We recall that a subset X of R 3 is called compact if it is closed (i.e. the set the tangent vector of "1 per unit length. We are going to look for an analogue of
of points in R 3 that are not in X is open) and bounded (i.e. X is contained this for surfaces.
in some open ball). In the proof, we shall make use of the following fact about The 'direction' of the tangent plane to a surface patch.,. : U ---> R 3 is
compact sets: if f : R 3 ---> R is a continuous function, then there are points P measured by its standard unit normal N, so we might expect that the curvature
 u
166 Elementary Differential Geometry 7. Gaussian Curvature and the Gauss Map 167

i
I

of CT is measured by the 'rate of change of N per unit area'. To make sense of I Proof 7.1
this, note that N is a point of the unit sphere
By Definition 5.3,
2
8 = {v E R
3
III v II = 1}.
AN(RJ) ffR II Nit x N v II dudv
--;---'- - $ (14)
The Gauss map is the map S --+ 8 2 , where S is the image of CT, which sends Au (Ro) ffR6 II t7 u x t7 v II dudv .
the point O'(u,v) of S to the point N(u,v) of 8 2 . We denote the Gauss map
By Proposition 6.4,
by 9. More generally, the Gauss map can be defined for any orientable surface
S (see Section 4.3), since such a surface has a well defined unit normal N at N u x N v = (at7 u x bt7 v ) x (CO'u x dtT1I )
every point. = (ad - bc)O'u x 0'11
= det(-Fi1F/l)O'u x 0'11
det(.'F/l )
= det(.'Fr) t1 u x CT v

, -------~~~~------
._\
~-------_Q_--------
=
ItI; ~II
~ tr u X tr v (by the definition of Fr and F/l)

LN-M2
= EG _ F2 t7u X 1711

= Ktr u X t1 v (by Proposition 7.1(i)). (15)

Sustituting in Eq. (14), we get
If R C;;;; U is a region, the amount by which N varies over the part u(R) of
S is measured by the area of the part N(R) of the sphere. Thus, the rate of AN(R
_ - , - - -D
,- _
) ffR IKIII O'u x O'v II dudv
6

change of N per unit area is approximately Au(Ro) - ffR$ II tr u X tr v II d11.dv .

area of N(R) AN(R) Let € be any positive number. Since K(u,v) is a continuous function of
area of t7(R) = Au(R) , (11., v) (see Exercise 7.7), we can choose & > 0 so small that
in the notation of Section 5.4. The following theorem shows that, as the region IK(11., v) - K(uo, vo)1 < €
R shrinks to a point, this ratio becomes the absolute value of the gaussian
curvature of 0' at the point. if (11., v) E R o. Since, for any real numbers a,b, la - bl ~ lIal-lbll, it follows
that IIK(11.,v)I-IK(11.o,vo)11 < € if (u,v) E R o, i.e.

Theorem 7.1 IK(11.o, vo)1 - to < IK(u, v)1 < IK(11.o, vo)1 +€
,
Let 17 : U --+ R 3 be a surface, let (uo, vo) E U, and let r5 > 0 be such that the if (11., v) E R o. Multiplying through by II tr... x tr v II and integrating over RO''fft
closed disc get

~ = {(u,v) E R 2 I (u - UO)2 + (v - VO)2 ~ r5 2 } (IK(11.o,vo)I-€)!!lltru Xtr v II dudv <!!IK(u,v)llICTu XCT v Ildudv

with centre (uo, vo) and radius r5 is contained in U (such a r5 exists because U
is open). Then, < (IK(11.o,vo)1 +€)!!II t7tJ. x t7 v lld11.dv,
lim AN(RD) = IKI, AN(Ro)
0-+0 Au(Ro) .. IK(uo, vo)1 - € < Aq(R ) < IK(uo, vo)1 + (. (using Eq. (14))
o
where K is the gaussian curvature 0117 att7(uo,vo).
168 Elementary Differential Geometry 7. Gaussian Curvature and the Gauss Map 169

EXERCISES

7.18 Let 0' : U -+ R 3 be a patch on a surface S. Show that the image

under the Gauss map of the part O'(R) of S corresponcljnl!" to a region
This proves the theorem. o R ~ U has area

Although this proposition only gives the absolute value of the gaussian
curvature K, the sign can be recovered from the Gauss map if we define the
signed area of N(R) to be ±AN(R), where the sign is + or - according to where K is the gaussian curvature of S. (Inspect the proof of The-
whether N u x N v points in the same or the opposite direction as N. By Eq. orem 7.1.)
(15), this sign is that of K, so K is the limit of the ratio 7.19 Let S be the torus in Exercise 4.10. Sketch the parts S+ and S-
signed area of N(R) of S where the gaussian curvature K of S is positive and negative,
area of O'(R) respectively. Show, without calculation, that
as the region R shrinks to a point.
As the following examples show, Theorem 7.1 sometimes allows one to find
fh+ K dA =- fL- K dA = 41r.
the gaussian curvature of a surface with no calculation. (The meaning of these integrals should he clear: if S+, say, is con-
tained in a single surface patch 0' of S, so that S+ = O'(R+), say,
Example 7.4 then the left-hand side means IIR+
K dAa; if S+ is not contained
in a single patch, it can be broken into pieces, each of which is con-
For a plane, the unit normal is constant. Thus, for any R, N(R) is a single tained in a single patch, and then the integral over S+ is the sum of
point, and thus has zero area. By the theorem, a plane has gaussian curvature the integrals over each piece of S+. Further details can be found in
zero everywhere. Section 11.3.)
For a (generalised) cylinder, the unit normal is clearly always perpendicular fI
Of course, it follows that s K dA = 0, a result that will be 'ex-
to the rulings of the cylinder, so the image of the Gauss map is contained in plained' in Section 1l.3.
the great circle on the unit sphere formed by intersecting the sphere with the
plane passing through the centre of the sphere and perpendicular to the rulings
of the cylinder. Any great circle obviously has zero area, so the cylinder has
zero gaussian curvature too.
Finally, for the unit sphere 8 2 itself, the unit normal at a point P is clearly
parallel to the radius vector from the centre of the sphere to P. In other words,
the Gauss map is the identity map or the antipodal map (depending on the
choice of parametrisation). Both of these maps are obviously equiareal, so the
(absolute value of the) gaussian curvature of 8 2 is one. In fact, this discussion
shows that, for any parametrisation 0' of 8 2 , we have N = ±O'. The sign depends
of the choice of parametrisation, but in any case N u x N l1 = AU X (1v so the
gaussian curvature is +1.
 8
Geodesics

Geodesics are the curves in a surface that a bug living in the surface would
perceive to be straight. For example, the shortest path between two points
in a surface is always a geodesic. We shall actually begin by giving a quite
different definition of geodesics, since this definition is easier to work with. We
give various methods of finding the geodesics on surfaces, before finally making
contact with the idea of shortest paths towards the end of the chapter.

8.1. Definition and Basic Properties

Recall from Proposition 1.1 that a curve 'Y is a straight line if it has zero
acceleration l' everywhere. A bug living in a surface and travelling along a curve
'Y, however, would perceive only the component of1' parallel to the tangent
plane, so would argue that 'Y is straight if this component of the acceleration
is everywhere zero. This suggests

Definition 8.1
A curve'Y on a surface S is called a geodesic if i(t) is zero or perpendicular to
the surface at the point 'Y(t), Le. parallel to its unit normal, for all values of
the parameter t.

Note that the unit normal is well defined up to sign for any surface S, so
this definition makes sense.
171
 Elementary Differential Geometry 8, Geodesics 173
172

There is an interesting mechanical interpretation of geodesics: a particle If i is parallel to N, it is obviously perpendicular to N x 'Y, so by Eq. (1),
moving on the surface, and subject to no forces except a force acting per- 1<9 = O.
pendicular to the surface that keeps the particle on the surface, would move Conversely, suppose that 1<9 = O. Then, i is perpendicular to N x 'Y. But
3
along a geodesic. This is because Newton's second law of motion says that the then, since 'Y, N and N x 'Y are perpendicular unit vectors in R , and since i
force on the particle is parallel to its acceleration i, which would therefore be is perpendicular to 'Y, it follows that i is parallel to N. 0
perpendicular to the surface.
We begin our study of geodesics by noting that there is essentially no choice The next result gives the simplest examples of geodesics.
in their parametrisation.
Proposition 8.3
Proposition 8.1 Any (part of a) straight line on a surface is a geodesic.
Any geodesic has constant speed.
By this, we mean that every straight line can be parametrised so that it is
a geodesic. A similar remark applies to other geodesics we shall consider whose
Proof 8.1 parametrisation is not specified.
Let 'Y(t) be a geodesic on a surface S. Then, denoting d/dt by a dot,
Proof 8.3
i!.-
dt
II 'Y 11 2 = dd (H) = 2.:y..y.
t
This is obvious, for a straight line may be parametrised by
Since 'Y is a geodesic, i is perpendicular to the tangent plane and is therefore
perpendicular to the tangent vector 'Y. So i.'Y = 0 and the last equation shows 'Y(t) = a + ht,
that II 'Y II is constant. 0 where a and h are constant vectors, and clearly i = O. o
It is clear that a unit-speed reparametrisation of a geodesic 'Y is still a
geodesic, since Proposition 8.1 shows that the effect of the reparametrisation is Example 8.1
to multiply i by a non-zero scalar. Thus, we can always restrict to unit-speed All stralght lines in the plane are geodesics, as are the rulings of any ruled
geodesics if we wish. surface, such as the generators of a (generalised) cylinder, or those of a (gen-
We observe next that there is an equivalent definition of a geodesic expressed eralised) cone, or the straight lines on a hyperboloid of one sheet.
in terms of the geodesic curvature 1<9 (see Section 6.2). Of course, this is why '.
1<9 is called the geodesic curvature! The next result is almost as simple:

Proposition 8.2 Proposition 8.4

A curve on a surface is a geodesic if and only if its geodesic curvature is zero Any normal section of a surface is a geodesic.
everywhere.
Proof 8.4
Proof 8.2
Recall from Section 6.2 that a normal section of a surface S is the intersection
It is sufficient to consider a unit-speed curve 'Y contained in a patch fr of the C of S with a plane II, such that II is perpendicular to the surface at each
surface. Let N be the standard unit normal of fr, so that point of C. We showed in Section 6.2 that 1<9 = 0 for a normal section, so the
(1) result follows from Proposition 8.2. 0
1<9 = i·(N x 'Y)'
I
 Elementary Differential Geometry 8. Geodesics 175
174

Example 8.2 EXERCISES

All great circles on a sphere are geodesics. 8.1 Describe four different geodesics on the hyperboloid of one sheet
x 2 + y2 _ z2 =1
passing through the point (1,0,0). (Use Propositions 8.3 and 8.4.)
8.2 Consider the tube of radius a > 0 around a unit-speed curve 'Y in R 3
defined in Exercise 5.17:
(1(S,O) = "(s) + a{cos 8 n{s) + sin 0 b(s)).
Show that the parameter curves on the tube obtained by fixing the
value of s are circular geodesics on S.
8.3 Let 'Y{t) be a geodesic on an ellipsoid S (see Proposition 4.6). Let
2R(t) be the length of the diameter of the ellipsoid parallel to i'(t),
and let B(t) be the distance from the centre of the ellipsoid to the
For a great circle is the intersection of the sphere with a plane II passing
tangent plane of S at 'Y{ t). Show that the curvature of "( is S (t) I R (t ) 2 ,
through the centre 0 of the sphere, so if P is a point of the great circle, the
and that the product R{t)S(t) is independent of t. (Take the ellipsoid
vector OP lies in II and is perpendicular to the tangent plane of the unit sphere 2 2 ~

at P. Hence, II is perpendicular to the tangent plane at P. to be ~ + ~ + ~ = 1 and note that 'Y(t) ::::;: (f(t),g{t),h(t)) is a
fr,
geodesic if and only if (I, §, h) ::::;: >"( ~, ~) for some scalar >..(t).)
8.4 Show that the torsion ofa geodesic with nowhere vanishing curvature
Example 8.3
is equal to its geodesic torsion (see Exercise 6.11). (Use the fact that
The intersection of a (generalised) cylinder with a plane II perpendicular to the principal normal of the geodesic is parallel to the unit normal of
the rulings of the cylinder is a geodesic. For it is clear that the unit normal N the surface.)
is perpendicular to the rulings. It follows that N is parallel to II, and hence 8.5 A geodesic "( on a surface S lies in a plane and has nowhere vanishing
that II is perpendicular to the tangent plane. curvature. Show that 'Y is a line of curvature of S. (If II is the plane,
show that the unit normal N of S is parallel to II and deduce that
N is parallel to 'Yi then use Exercise 6.18.)

N 8.2. Geodesic Equations

II
Unfortunately, Propositions 8.3 and 8.4 are not usually sufficient to determine
all the geodesics on a given surface. For that, we need the following result:

Theorem 8.1
A curve "( on a surface S is a geodesic if and only if, for any part 'Y(t) =
u(u(t), v(t)) of"! contained in a surface patch (1 ofS, the following two equations
176 Elementary Differential Geometry 8. Geodesics 177

are satisfied: Example 8.4

2 We determine the geodesics on the unit sphere 8 2 by solving the geodesic
:(EU + FiJ) = 4(E,.U 2 + 2F...uiJ + G,.il ),
(2) equations. For the usual parametrisation by latitude 8 and longitude 'P,
2
: (Fu + Gil) = 4(Ev u + 2Fv uv + G v v
2
), tI(8, 'P) = (cos 8 cos 'P, cos (J sin 'P, sin 8),
where Edu 2 + 2Fdudv + Gdv 2 is the first fundamental form of (1. we found in Example 5.2 that the first fundamental form is
d(J2 + cos2 () dlp2 •
The differential equations (2) are called the geodesic equations.
We might as well restrict ourselves to unit-speed curves -y(t) = u(O(t), 'P(t)), so
that
Proof 8.1
iP + "p2 cos2 e = 1,
Since {(1.. , u v} is a basis of the tangent plane of (1, "1 is a geodesic if and only
if;Y is perpendicular to (1,. and (11" Since l' = tit1... + iJu u, this is equivalent to and if -y is a geodesic the second equation in (2) gives

(:t (tit1,. + V(1u)) .(1,. =0 and (:t (tit1,. + V(1v)) .U v = O. (3)

~ (ep cos2 8) = 0,
so that
We show that these two equations are equivalent to the two geodesic equations.
The left-hand side of the first equation in (3) is equal to ep cos 2 (J = [),
d ((.U(1.. + Vtl
. v).tI..
) - (Uti..
. + Vtl
.) dtr,.
where n is a constant. If [} = 0, then ep = 0 so 'P is constant and -y is part of a
dt v .---;jt meridian. We assume that [} =I- 0 from now on.
The unit~speed condition gives
= ~ (Eu + Fv) - (tit1... + Vtlv).(Utl,... + V(1..v)
n2
= :t (Eu + Fil) - (u ((1,..(1..... ) + uv(tI...tI,.u
2 2
+ (1v· tI....) + v ((1u.(1,.u))' (4)
so along the geodesic we have
o'2 =1-~(),
cos

Now,
( dd'P8 ) 2 'Po:22 = cos 2 0(0- 21cos 2 0 - 1)'
so (1... (1..... = ~E.... Similarly, u v .(1,.u = ~G,.. Finally, and hence
d8

Substituting these values into (4) gives

(1.... tl UIJ + tllJ.U..,. = (tI,..uv ).. = F,.. ±('P - 'Po) =
I cos Ov' [}-2 cos 2 ()
where 'Po is a constant. The integral can be evaluated by making the substitu-
- 1
,

(:t (Uti,. + Vtlv)) ·(1u = :t (Eu + FiJ) - ~(EuU2 + 2Fu uv + Guv ). 2 tion U = tan 0, This gives
(u)
This shows that the first equation in (3) is equivalent to the first geodesic
equation in (2). Similarly for the other equations. 0
±
('P
- -
'Po) - I v'n-2 _ du
1- u2 -~ sin-
1
y'n- 2 - 1 '
and hence

The geodesic equations are non-linear differential equations, and are usually tan e = ±J[}-2 - 1 sin('P - 'Po).
difficult or impossible to solve explicitly. The following example is one case in This implies that the coordinates x = cos e cos 'P, y = cos (} sin 'P and z = sinO
which this can be done. Another is given in Exercise 8.9. of -y(t) satisfy the equation
z = ax + by,
 178 Elementary Differential Geometry 8. Geodesics 179

where a = 'fy'n '-lsin'Po, b = ±y'n '-lcos'Po. This shows that -y is Example 8.5
contained in the intersection of 8' with a plane passing through the centre of
the sphere. We already know that all straight lines in a plane are geodesics. Since there is
a straight line in the plane through any given point of the plane in any given
Hence, in all cases, -y is part of a great circle.
direction parallel to the plane, it follows from Corollary 8.1 that there are no
other geodesics.
The geodesics on a sphere can actually be determined much more simply
hy using the following consequence of Theorem 8.1.
Example 8.6
Corollary 8.1 Similarly, on a sphere, the great circles are the only geodesics, for there is
Let P be a point of a surface S, and let t be a unit tangent vector to S at P. clearly a great circle passing through any given point of the sphere in any given
Then, there exists a unique unit-speed geodesic 'Y on S which passes through P direction tangent to the sphere. (If p is the position vector of the point and
and has tangent vector t there. t the tangent direction, let II be the plane passing through the origin and
parallel to p and t (i.e. with normal p x t); then take the intersection of the
In short, there is a unique geodesic through any given point of a surface in sphere with II.)
any given direction.
The following consequence of Theorem 8.1 can also be used in some cases
to find geodesics without solving the geodesic equations.
Proof 8.1
The geodesic equations are of the form Corollary 8.2
u = f(u,v,u,il), ii = g(u,v,u,il), (5) An isometry between two surfaces takes the geodesics of one surface to the
where f and g are smooth functions of the four variables u, v, it and il. It geodesics of the other.
is proved in the theory of ordinary differential equations that, for any given
constants a, b, c, d, and any value to of t, there is a solution of Eq. (5) such that Proof 8.2
u(to) = a, v(to) = b, u(to) = c, il(to) = d, (6) Let S, and S, be the two surfaces, let f : S, --t S, be the isometry, and let 'Y
and such that u(t) and v(t) are defined and smooth for all t satisfying It-tol < be a geodesic in S,. Let O'(u, v) be a patch in S, and suppose that the part of
e, where e is some positive number. Moreover, any two solutions of Eq. (5) the geodesic lying in this patch is given by -y(t) = O'(u(t), v(t)). Then, u and v
satisfying (6) agree for all values of t such that It - tol < e', where e' is some satisfy the geodesic equations (2), with E, F and G being the coefficients of the
positive number < €. first fundamental form of 0'. By Theorem 5.1, f 0 0' is a patch of S, with the
We now apply these facts to the geodesic equations. Suppose that P lies in same first fundamental form as 0'. Hence, by Theorem 8.1, t>-+ f(O'(u(t), v(t)))
a patch O'(u,v) of S, say that P is the point O'(a, b), and that t = CO'u + dO'v, is a geodesic on S" in other words, f 0 -y is a geodesic. 0
where a, b, C, d are scalars and the derivatives are evaluated at u = a, v = b. A
unit-speed curve -y(t) = O'(u(t), v(t)) passes through P at t = to if and only if Example 8.7
u(to) = a, v(to) = b, and has tangent vector t there if and only if
On the circular cylinder x' + y' = 1, we know that the circles obtained by
CO'u + dO'. = t = "t(t o) = u(to)O'. + il(to)O'., intersecting the cylinder with planes parallel to the xy-plane are geodesics
i.e. u(to) = c, il(to) = d. Thus, finding a (unit-speed) geodesic-y passing through (since they are normal sections). We also know that the straight lines on the
P at t = to and having tangent vector t there is equivalent to solving the cylinder parallel to the z-axis are geodesics. However, these are certainly not
geodesic equations subject to the initial conditions (6). But we have said above the only geodesics, for there is only one geodesic of each of the two types passing
that this problem has a unique solution. 0
180 Elementary Differential Geometry B. Geodesics 181

through each point of the cylinder (whereas we know that there is a geodesic tangent vector makes a constant angle with the rulings of the cylin-
passing through each point in any given tangent direction). der.
8.9 Find the geodesics on a circular cylinder by solving the geodesic
equations.
,,
, 8.10 Let "I(t) be a unit-speed curve on the helicoid
,,
/7(u,v) = (ucosv,usinv,v).
> Show that
u· + (1 + u·),,· = 1
.~_.- '-"-"-"-"-"-. (a dot denotes dldt). Show also that, if "I is a geodesic on /7, then
. a
To find the missing geodesics, we recall that the cylinder is isometric to the v = 1 + u2'
plane (see Example 5.5). In fact, the isometry takes the point (u, v, 0) of the where a is a constant. Find the geodesics corresponding to a = 0
xy-plane to the point (cos u, sin u, v) of the cylinder. By Corollary 8.2, this map and a = l.
takes geodesics on the plane (i.e. straight lines) to geodesics on the cylinder,
8.11 Show that, if N is the standard unit normal of a surface patch /7( u, v)
and vice versa. So to find all the geodesics on the cylinder, we have only to
with first fundamental form Edu· + 2Fdudv + Gdv·, then
find the images under the isometry of all the straight lines in the plane. Any
line not parallel to the y-axis has equation y = mx + c, where m and c are N x au = Eu v - Fu u , N x a = FtT v - GCf u
y' EG F. v -y'r.E;i';G;;;==_=;F;ii".
constants. Parametrising this line by x = U , Y = mu + c, we see that its image
is the curve (Use Proposition 5.2.) Deduce that, if /7(u(t), v(t» is a unit-speed
curve on tT, its geodesic curvature
"I(u) = (cosu,sinu,mu+c)
1<9 = (iiu - "u)VEG - F. + Au' + Bu·" + Cu,,' + D,,',
on the cylinder. Comparing with Example 2.1, we see that this is a circular helix
of radius one and pitch 27flml (adding c to the z-coordinate just translates the where A, B, C and D can be expressed in terms of E, F, G and their
helix vertically). Note that if m = 0 we get the circular geodesics that we derivatives. (Use the method of proof of Theorem 8.1 to calculate
already know. Finally, any straight line in the xy-plane parallel to the y-axis is dot products such as /7"./7 vv.) This gives another proof of Corollary
mapped by the isometry to a straight line on the cylinder parallel to the z-axis, 8.2.
giving the other family of geodesics that we already know. 8.12 Show directly that the parameter of any curve satisfying the geodesic
equations (2) is proportional to arc-length.

EXERCISES

8.6 Show that, if P and Q are distinct points of a circular cylinder, there 8.3. Geodesics on Surfaces of Revolution
are either two or infinitely many geodesics on the cylinder joining P
and Q. Which pairs P, Q have the former property? It turns out that, although the geodesic equations for a surface of revolution
8.7 Use Corollary 8.2 to find all the geodesics on a circular cone. (Use cannot usually be solved explicitly, they can be used to get a good qualitative
Exercise 5.5.) understanding of the geodesics on such a surface.
8.8 Use Corollary 8.2 to show that the geodesics on a generalised cylin- We parametrise the surface of revolution in the usual way
der are exactly those constant-speed curves on the cylinder whose /7(u, v) = (f(u) cos v, flu) sin v, g(u»,
 • 8. Geodesics 183
182 Elementary Differential Geometry

where we assume that J > 0 and (-~f +

Proof 8.5
(*)2
= 1 (see Examples 4.13 and
6.2 - note that in these examples we used a dot to denote d/du, but now a dot On a meridian, we have v = constant so the second equation in (7) is obviously
is reserved for djdt, where t is the parameter along a geodesic). We found in satisfied. Equation (8) gives i.J. = ±1, so i.J. is constant and the first equation in
Example 6.2 that the first fundamental form of II is du2 + J(u)2dv 2. Referring (7) is also satisfied.
to Eq. (2), we see that the geodesic equations are For (ii), note that if U = Uo is constant, then by Eq. (8), i.J = ±l/f(uo)
is non-zero, so the first equation in (7) holds only if df /du = O. Conversely,
U = J{u) : i?, :t (f(U)2i.J) = O. (7) if df /du == 0 when u = uo, the first equation in (7) obviously holds, and the
second holds because i.J = ±1/ f{uo) and feu) = f(uo) are constant. 0
We might as well consider unit-speed geodesics, so that
i.J.2 + J(u)2i.J2 = 1. (8) Of course, this proposition only gives some of the geodesics on a surface of
revolution. The following result is very helpful in understanding the remaining
l.From this, we make the following easy deductions:
geodesics.

Proposition 8.5 Proposition 8.6 (C1airaut's Theorem)

On the surface of revolution Let 'Y be a geodesic on a surface of revolution $, let p be the distance of a
lI{U, v) = (f{u) cosv, J{u) sin v, g{u)), point of $ from the axis of rotation, and let 1/J be the angle between "t and the
(i) every meridian is a geodesic; meridians of $. Then, p sin 'l/J is constant along 'Y.
(ii) a parallel u == Uo (say) is a geodesic if and only if df /du == 0 when U == Uo, Conversely, if p sin 1/J is constant along some curve 'Y in the surface, and if
i. e. Uo is a stationary point of f· no part of'Y is part of some parallel of $, then'Y is a geodesic.

geodesics

In the second paragraph of the proposition, by a 'part' of'Y we mean 'Y(J),

where J is an open interval. The hypothesis there cannot be relaxed, for on a
 p
184 Elementary DifFerential Geometry 8. Geodesics 185

parallel ¢ = 11'/2, so p sin 'IjJ is certainly constant. But parallels are not geodesics u = Uo when It - tol < to, contrary to our assumption. Hence, the term in
in general, as Proposition 8.5(ii) shows. brackets must vanish everywhere on 1', i.e.
.. dp'2
u=p-v
Proof 8.6 du '
showing that the first equation in (7) is indeed satisfied. 0
Parametrising 5 as in Proposition 8.5, we have p = f(u). Note that u ti / II Uti II
= U u and CTvl II CT v II = p-1CT v are unit vectors tangent to the parallels and the
meridians, respectively, and that they are perpendicular since F = O. Assuming Clairaut's Theorem has a simple mechanical interpretation. Recall that the
that 1'(t) = u(u(t), v(t)) is unit-speed, we have geodesics on a surface 5 are the curves traced on 5 by a particle subject to no
forces except a force normal to 5 that constrains it to move on 5. When 5 is
l' = cos 1/J Uu + p-l sin 'l/J U v a surface of revolution, the force at a point P of 5 lies in the plane containing
(this equation actually serves to define the sign of 'I/J, which is left ambiguous the axis of revolution and P, and so has no moment about the axis. It follows
in the statement of Clairaut's Theorem). Hence, that the angular momentum {} of the particle about the axis is constant. But,
if the particle moves along a unit-speed geodesic, the component of its velocity
Uu x l' = p-l sin 1/J au x a v'
along the parallel through P is sin tP, so its angular momentum about the axis
Since "t = ita.. + vU v , this gives is proportional to p sin 1/J.
VU11. x CT v = p-l sin 1/J au x a v ,
piJ = sin t/J. Example 8.8
Hence, We use Clairaut's Theorem to determine the geodesics on the pseudosphere
(Section 7.2):
psin 1/J = p2 iJ.
But the second equation in (7) shows that this is a constant, say {}, along the u(u, v) = (e U cos v, e11. sin v, VI - e 2u + cosh- 1 (e- U
)).

geodesic. We found there that its first fundamental form is

For the converse, if p sin 'ljJ is a constant {} along a unit-speed curve 'Yin 5,
the above argument shows that the second equation in (7) is satisfied, and we du 2 + e2u dv 2 .
must show that the first equation in (7) is satisfied too. Since It is convenient to reparametrise by setting w = e- U • The reparametrised
. surface is
v = -sin'ljJ
P
{}
=-,
p2
(9)
u(v,w) = (.!.cOsv,.!.sinv,jI- 12 +COSh-1W)
Eq. (8) gives w w w

i;.2 =1_ {}2 and its first fundamental form is

2 . (10)
P dv 2 + dw 2
Differentiating both sides with respect to t gives w2
. .. 2{}2. 2{}2 dp . We must have w > 1 for u to be well defined and smooth.
2
uu= ----p3p= 7 du U ' If 1'(t) = u(w(t),v(t)) is a unit-speed geodesic, the unit-speed condition
gives
u. (..u-p-v
dp
du
'2) =a.
(11)
If the term in brackets does not vanish at some point of the curve, say at and Clairaut's Theorem gives
1'(to) = u(uo, vo), there will be a number f > a such that it does not vanish for
1 . •1,
It - tol < f. But then u = a for It - tol < f, so l' coincides with the parallel -
W
Sin 'f/ = W21.V = n J£ 1 (12)
 L..1~.II""rl"g.J LJ.I~\,..\,..I"II;rr ...... 1.,;...,.1.1'1;,.1.1' ~u,

where rl is a constant, since p = l/w. Thus, iJ = [}w 2 • If [} = 0 we get a Returning now to an arbitrary surface of revolution 5, we describe how
meridian v = constant. Assuming now that fl i- a and substituting in Eq. (ll) Clairaut's Theorem allows us to describe the qualitative behaviour of the
gives geodesics on S.
Note first that, in general, there are two geodesics passing through any
given point P of S with a given angular momentum fl, for v is determined by
Hence, along the geodesic, Eq. (9) and u up to sign by Eq. (10). In fact, one geodesic is obtained from
the other by reflecting in the plane through P containing the axis of rotation
dv = i!.. = ± flw ,
(which changes n to -n) followed by changing the parameter t of the geodesic
dw w ..;'1 - [}2 W 2
/ to -t (which changes the angular momentum back to n again).
(v - va) = =F ~ )1- fl2 W 2, The discussion in the preceding paragraph shows that we may as well assume
2 1 that n > 0, which we do from now on. Then, Eq. (10) shows that the geodesic
(v - Vo ) + w2 = [}2' (13) is confined to the part of S which is at a distance ~ n from the axis.
where Va is a constant. So the geodesics are the images under i1 of the parts If all of S is a distance > {} from the axis, the geodesic will cross every
of the circles in the vw-plane given by Eq. (13) and lying in the region w > 1. parallel of S. For otherwise, 'U would be bounded above or below on S, say the
Note that these circles all have centre on the v-axis, and so intersect the v-axis former. Let 'Ua be the least upper bound of'U on the geodesic, and let {} + 210,
perpendicularly. The meridians correspond to straight lines perpendicular to =
where f > 0, be the radius of the parallel 'U 'Ue. If 'U is sufficiently close to 'Uo,
the v-axis. the radius of the corresponding parallel will be ~ n + f, and on the part of the
w geodesic lying in this region we shall have

lui ~ V1 - (n~fr > 0

by Eq. (10). But this clearly implies that the geodesic will cross u = 'Uo, con-
1
tradicting our assumption.
Thus, the interesting case is that in which part of S is within a distance {}
v of the axis. The discussion of this case will be clearer if we consider a concrete
example whose geodesics nevertheless exhibit essentially all possible forms of
Since w > 1, any geodesic other than a meridian has a maximum value of w, behaviour.
which it attains, and a maximum and minimum value of v, which it approaches
arbitrarily closely but does not attain (see the diagram below). This shows that ,
I
the pseudosphere is 'incomplete', i.e. a geodesic on the pseudosphere cannot be Example 8.9
continued indefinitely (in one direction if it is a meridian, in both directions We consider the hyperboloid of one sheet obtained by rotating the hyperbola
otherwise) .
x2 - z2 = 1, x > 0,
in the xz-plane around the z-axis. Since all of the surface is at a distance 2:: 1
from the z-axis, we have seen above that, if 0 :5 {} < I, a geodesic with angular
momentum n crosses every parallel of the hyperboloid and so extends from
Z = -00 to z = 00.
 p
188 Elementary Differential Geometry 8. Geodesics 189

1 passing through a point P. If P is on the waist r of the hyperboloid (i.e.

the unit circle in the xy-plane), which is a geodesic by Proposition 8.5(ii), then
p = 1 at P so ¢ = 1f /2 and C is tangent to r at P. It must therefore coincide
with r.

0< n<1
Suppose now that n > 1. Then the geodesic is confined to one of the two
regions
{}>1
z 2: J n2 - 1, z ~ - J n2 - 1,
which are bounded by circles r+ and r- , respectively, of radius n. Let P be a Assume now that P is in the region below r. Then 0 < 'IjJ < tr/2 at P, so
point on r-, and consider the geodesic C that passes through P and is tangent as it leaves P in one direction, C approaches r. It must in fact get arbitrarily
=
to r- there. Then, ¢ ::::: 1f /2 and p n at P, so C has angular momentum n. close to r. For if it were to stay always below a parallel i' of radius 1 + E, say
Now C cannot be contained in r-, since r- is not a geodesic (by Proposition (with t;: > 0), then we would have
8.5(ii)), so C must head into the region below r- as it leaves P. Moreover, C
must be symmetric about P, since reflection in the plane through P containing
the z-axis takes C to another geodesic that also passes through P and is tangent
lui ?- J C~ J
1- 2

everywhere along C by Eq. (10), which clearly implies that C must cross every
to r- there, and so must coincide with C by the uniqueness part of Corollary
parallel, contradicting our assumption. So, if n = 1 the geodesic spirals arou~d
8.1. Since u :j; 0 in the region below r- by Eq. (10), the geodesic crosses every
parallel below r- and z -+ -00 as t -+ ±oo. the hyperboloid approaching, and getting arbitrarily dose to, r but never qmte
reaching it.
Suppose now that C is any geodesic with angular momentum n > 1 in the
region below r-. Then a suitable rotation around the z-axis will cause C to
intersect C, say at Q, and so to coincide with it (possibly after reflecting in
the plane through Q containing the z-axis and changing t to -t). We have
EXERCISES
therefore described the behaviour of every geodesic with angular momentum
n > 1 that is confined to the region below r-. Of course, the geodesics with 8.13 There is another way to see that all the meridians, and the paral-
angular momentum n > 1 in the region above r+ are obtained by reflecting lels corresponding to the stationary points of f, are geodesics on a
those below r- in the xy-plane. surface of revolution considered in this section. What is it?
Suppose finally that n = 1. Let C be a geodesic with angular momentum 8.14 A surface of revolution has the property that every parallel is a
geodesic. What kind of surface is it ?
 p
190 Elementary Differential Geometry 8. Geodesics 191

8.15 Show that a geodesic on the pseudosphere with non-zero angular the part of'Y between p' and q' by this shorter path, thus giving a shorter path
momentum {) intersects itself if and only if n < (1 + 11"2)-1/2. How from p to q in S.
many self-intersections are there in that case? (The condition for a We may therefore consider a path 'Y entirely contained in a surface patch u.
self-intersection is that, for some value of w > 1, the two values of v To test whether "'( has smaller length than any other path in (1 passing through
satisfying Eq. (13) should differ by an integer multiple of 211".) two fixed points p and q in u, we embed 'Y in a smooth family of curves on u
8.16 Let f: a(v,w) i-7 a(v,w) be an isometry ofthe pseudosphere, where passing through p and q. By such a family, we mean a curve 'YT on (1, for each
7 in an open interval (-0,0), such that
the parametrisation a is that in Example 8.8. /
/
(i) Show that f takes meridians to meridians, and deduce that u (i) there is an E > 0 such that 'YT(t) is defined for all t E (-f,f) and all
does not depend on w. (Use the fact that meridians are the only 7 E (-0,0);
geodesics on the pseudosphere that can be extended indefinitely
in one direction.)
(ii) for some a, b with -f < a < b < f, we have
(ii) Deduce that f takes parallels to parallels. (Parallels are the 'YT(a) =p and 'YT(b) =q for all 7 E (-0,8);
curves that are perpendicular to every meridian.)
(iii) Deduce from (ii) and Exercise 7.11 that w = w. (iii) the map from the rectangle (-8,0) x (-E, E) into R 3 given by
(iv) Show that f is a rotation about the axis of the pseudosphere
(7, t) H "'(T(t)
or a reflection in a plane containing the axis of rotation.
8.17 What do the geodesics on a pseudosphere correspond to in the disc is smooth;
model (Exercise 7.12(ii»? (Use the solution to Exercise 7.12 and the (iv) 'Yo = 'Y.
fact that the Mobius transformation z H ~~: is a conformal map
from the (complex) plane to itself that takes lines and circles to lines
and circles.)
8.18 Describe the geodesics on
(i) a spheroid, obtained by rotating an ellipse around one of its
axes;
(ii) a torus (Exercise 4.10).

,'\ The length of the part of 'YT between p and q is

8.4. Geodesics as Shortest Paths
Everyone knows that the straight line segment joining two points P and Q
£(7) = l II t
b
T
II dt,

in a plane is the shortest path between P and Q. It is almost as well known where a dot denotes d/dt.
that great circles are the shortest paths on a sphere. And we have seen that
the straight lines are the geodesics in a plane, and the great circles are the Theorem 8.2
geodesics on a sphere.
To see the connection between geodesics and shortest paths on an arbitrary With the above notation, the unit-speed curve 'Y is a geodesic if and only if
surface S, we consider a unit-speed curve "'( on S passing through two fixed d
d7£(T) = 0 when T = O.
points p and q on the surface. If "'( is a shortest path on the surface from p to
q, then the part of"'( contained in any surface patch u of S must be the shortest Note Although we assumed that 'Y = 'Yo is unit-speed, we do not assume that
path between any two of its points. For if p' and q' are any two points of 'Y in 'YT is unit-speed if T "I O.
(1, and if there were a shorter path in u from p' to q' than 'Y, we could replace
 p
192 Elementary DifFerential Geometry 8. Geodesics 193

Proof 8.2 using integration by parts. Now, since "Y1"(a) and "Y1"(b) are independent of T

(being equal to p and q, respectively), we have

We use the formula for 'differentiating under the integral sign': if f(T, t) is
fJ'yT
smooth, -ar = 0 when t :::: a or b.
Since

d
dT leT) = dT
d r b
1a II ",1' II dt
we see that
au :::: av :::: 0 when t = a or b.
= -d
dT
l(l
b
(Eu 2 + 2Fuil + Gil 2)1/2 dt
ar ar
Hence, the first term on the right-hand side of Eq. (15) is zero. Inserting the
ba remaining terms in Eq. (15) back into Eq. (14), we get
=
l
1
(l
-(geT, t)1/2) dt
ar
(b a d
drL(r) = 1(1r (au
b
av)
U ar + v ar dt, (16)
= "21(1 g(r, t)-1/2 a~ dt, (14)
where
where
U(r, t) :::: ~g-1/2(Euit2 + 2Fu uv + G u v 2) - ~ {g-1/2(Eu + FiJ) } ,
g(r, t) = Eu 2 + 2Fuil + Gil 2
and a dot denotes dldt. Now, VCr, t) = ~g-1/2(EtJit2 + 2Fvuil + Gvii) - :t {g-1/2(Fu + Gil)} .
(17)

ag 8E' 8F.. 8G' .8u (au ail) ail Now 'Yo = 'Y is unit-speed, so since II ",1" 11 2 = g(r, t), we have g(r, t) = 1 for all
ar = aT U 2 + 2 ar uv + aT v 2 + 2Eu ar + 2F ar iI + it ar + 2GiI ar t when r = O. Comparing Eqs. (17) with the geodesic equations in (2), we see
=
that, if"( is a geodesic, then U v = 0 when r :::: 0, and hence by Eq. (16),
= ( Eu ~~ + E ~~) it 2 + 2 (F ~~ + F ~~) itil + (Gu ~; + G ~~) iJ2
v u v v
d
drl(r) =0 when r:::: O.
. au 2
( au 2
av )
2
av
2

+ 2Eu arat + 2F aTat il + it aTat + 2Gil aTat For the converse, we have to show that, if

=(E '2+2F"
uU uUV +

au2
G u V.2)aU
-aT + (E' 2 2F"UV G .2)aV
vU +

av2
v + vV -aT l (U~~ + V:~)
b
dt = 0 when r = 0 (18)
+ 2(Eit + FiJ)a r a t +2(Fu+GiJ)--. aTat for all families of curves "(", then U :::: V = 0 when r = 0 (since this will
The contribution to the integral in Eq. (14) coming from the terms involving . prove that 'Y satisfies the geodesic equations). Assume, then, that condition
the second partial derivatives is (18) holds, and suppose, for example, that U i=- 0 when r = O. We will show
that this leads to a contradiction.
lb g-1/2 { (Eu + FiI) ::;t +(Fit + Gil) ::;t} dt Since U i=- 0 when T = 0, there is some to E (a, b) such that U(O, to) i=- 0,
say U(O, to) > O. Since U is a continuous function, there exists 11 > 0 such that

= g-1/2 { (Eu + FiI) ~~ + (Fu + Gil) ~~} I:~: U (0, t) > 0 if t E (to - 11, to + 11)·

-lb (:t {g-1/2(Eu + Fv)} ~~ + :t {g-1/2(Fu + Gil) } :~) dt, (15)

Let IjJ be a smooth function such that
¢(t) >0 if t E (to -11, to + 11) and ¢(t) = 0 if t ~ (to -11, to + 11)· (19)
 p
194 Elementary Differential Geometry 8. Geodesics 195

(The construction of such a function ¢ is outlined in Exercise 8.20.) Sup- Thirdly, in general, a shortest path joining two points on a surface may not
pose that 'Y(t) = 0'( U (t), v (t)), and consider the family of curves 'YT (t) exist. For example, consider the surface S consisting of the xy-plane with the
0'(U(7, t), V(7, t)), where origin removed. This is a perfectly good surface, but there is no shortest path
u( 7, t) = u(t) + n;f>(t), v(r, t) = v(t). on the surface from the point p =
(-1,0) to the point q =
(1,0). Of course,
the shortest path should be the straight line segment joining the two points,
Then, au/ar = ¢ and av/ar = 0 for all T and t, so Eq. (18) gives but this does not lie entirely on the surface, since it passes through the origin

0=
1a
Ua+V av) I dt=
(au
6
7 ar
o

T=O
/t
+'1 U(O,t)¢(t)dt.
to-1j
(20)
which is not part of the surface. For a 'real life' analogy, imagine trying to walk
from p to q but finding that there is a deep hole in the ground at the origin.
The solution might be to walk in a straight line as long as possible, and then
But U(O, t) and ¢(t) are both> 0 for all t E (to -1], to + TJ), so the integral on skirt around the hole at the last minute, say taking something like the following
the right-hand side of Eq. (20) is > O. This contradiction proves that we must route:
have U(O, t) = 0 for all t E (a, b). One proves similarly that V(O, t) = 0 for all
t E (a, b). Together, these results prove that "y satisfies the geodesic equations.
o
It is worth making several comments on Theorem 8.2 to be clear about what ----a.f---....:::>~--------.J(\L---~>1lo--••-----
it says, and also what it does not say. p < > q
Firstly, if'Y is a shortest path on 0' from p to q, then £(T) must have an
absolute minimum when T = O. This implies that d~£(T) = 0 when T = 0, and • I

hence by Theorem 8.2 that "y is a geodesic.

Secondly, if "y is a geodesic on 0' passing through p and q, then £( T) has
a stationary point (extremum) when T = 0, but this need not be an absolute This path consists of two straight line segments of length 1 - 10, together with
minimum, or even a local minimum, so 'Y need not be a shortest path from p a semicircle of radius 1:, so its total length is
2(1 - 1:) + 11"1: = 2 + (11" - 2)t:.
Of course, this is greater than the straight line distance 2, but it can be made
as close as we like to 2 by taking t: sufficiently small. In the language of real
analysis, the greatest lower bound of the lengths of curves on the surface joining
p and q is 2, but there is no curve from p to q in the surface whose length is
equal to this lower bound.
----------
..
- - - " '-----.... - -
p q
Finally, it can be proved that if a surface 5 is a closed subset of R 3 (i.e. if
the set of points of R 3 that are not in 5 is an open subset of R 3 ), and if there
is some path in 5 joining any two points of 5, then there is always a shortest
path joining any two points of S. For example, a plane is a closed subset of R 3 ,
so there is a shortest path joining any two points. This path must be a straight
line, for by the first remark above it is a geodesic, and we know that the only
geodesics on a plane are the straight lines. Similarly, a sphere is a closed subset
of R 3 , and it follows that the short great circle arc joining two points on the
to q. For example, if p and q are two nearby points on a sphere, the short sphere is the shortest path joining them. But the surface S considered above
great circle arc joining p and q is the shortest path from p to q (this is not is not a closed subset of R 3 , for (0,0) is a point not in 5, but any open ball
quite obvious - see below), but the long great circle arc joining p and q is also containing (0,0) must clearly contain points of 5, so the set of points not in S
a geodesic. is not open.
 ¥
196 Elementary Differential Geometry 8. Geodesics 197

Another property of surfaces that are closed subsets of R 3 (that we shall (iv) Show that the function

(t-to)
also not prove) is that geodesics on such surfaces can be extended indefinitely,
Le. they can be defined on the whole of R. This is clear for straight lines in the </>(t) = 1jJ -1)-

plane, for example, and for great circles on the sphere (although in the latter
has the properties we want.
case the geodesics 'close up' after an increment in the unit-speed parameter
equal to the circumference of the sphere). But, for the straight line 'Y(t) =
(t-l,O) on the surface S defined above, which passes through p when t = 0, the /
/
largest interval containing t = 0 on which it is defined as a curve in the surface
is (-00,1). We encountered a less artificial example of this 'incompleteness'
8.5. Geodesic Coordinates
in Example 8.8: the pseudosphere considered there fails to be a closed subset
The existence of geodesics on a surface S allows us to construct a very useful
of R 3 because the points of its boundary circle in the xy-plane are not in the
atlas for S.
surface.

a(u,v)
EXERCISES

8.19 The geodesics on a circular (half) cone were determined in Exercise

8.7. Interpreting 'line' as 'geodesic', which of the following (true)
statements in plane euclidean geometry are true for the cone? p ,,(V)
(i) There is a line passing through any two points.
(ii) There is a unique line pa..sing through any two points.
(iii) Any two distinct lines intersect in at most one point. For this, let P be a point of S and let 'Y, with parameter v say, be a unit-speed
(iv) There are lines that do not intersect each other. geodesic on S with 'Y(O) = P. For any value of v, there is a unique unit-
(v) Any line can be continued indefinitely. speed geodesic iV, with parameter u say, such that iV(O) = 'Y(v) and which is
(vi) A line defines the shortest distance hetween any two of its perpendicular to 'Y at 'Y(v). We define a(u, v) = i V (u).
points.
(vii) A line cannot intersect itself transversely (i.e. with two non-
parallel tangent vectors at the point of intersection).
Proposition 8.7
8.20 Construct a smooth function with the properties in (19) in the fol- With the above notation, there is an open subset U of R 2 containing (0,0)
lowing steps: such that a : U --> R 3 is an allowable surface patch for S. Moreover, the first
(i) Show that, for all integers n (positive and negative), tn.-,/t' fundamental form of a is
tends to 0 as t tends to o. (Use L'Hopital's rule.) du 2 + G(u, v)dv 2 ,
(ii) Deduce from (i) that the function
where G is a smooth function on U such that
Ott) = {.-I/t' if t ::': 0,
= 1, = 0,
o ift$O G(O, v) Gu(O, v)
is smooth everywhere. whenever (O,v) E U.
(iii) Show that the function
1jJ(t) = 0(1 + t)O(l - t) Proof 8.7
is smooth everywhere, that 1jJ(t) > 0 if -1 < t < 1, and that
°
1jJ(t) = otherwise.
The proof that a is (for a suitable open set U) an allowable surface patch makes
 p
198 Elementary Differential Geometry 8. Geodesics 199

use of the inverse function theorem, and is similar to the proof of Proposition A surface patch a constructed as above is called a geodesic patch, and u and
4.1 (see Section 4.6). v are called geodesic coordinates.
Note first that, for any value of v,

~ -y(v),
Example 8.10
Uu(O,v) = dd i'V(U)! ' Uv(O,v) = ddVi'V(O) =
U 1£=0 If P is a point on the equator of the unit sphere 8 2 , take -y to be the equator
and that these are perpendicular unit vectors by construction. If u(u, v) with parameter the longitude I.{J, and let 1'''' be the meridian parametrised by
(J (u, v), g(u, v), h(u, v)), it follows that the jacobian matrix latitude 8 and passing through the point on the equator with longitude I.{J. The
corresponding geodesic patch is the usual latitude longitude patch, for which
lugu gvIv) the first fundamental form is
(h u hv d82 + cos2 8 dI.{J2 ,
has rank 2 when u = v = O. Hence, at least one of its three 2 x 2 submatrices in accordance with Proposition 8.7.
is invertible at (0,0), say

(
11£
g1£
Iv).
gv
(21)
We shall give an application of geodesic coordinates in the proof of Theorem
10.4.

By the inverse function theorem 4.2, there is an open subset U of R 2 such that
the map given by EXERCISES
F(u, v) = (J(u, v), g(u, v))
8.21 Let P be a point of a surface S and let v be a unit tangent vector to
is a bijection from U to an open subset F(U) of R 2 , and such that its inverse S at P. Let ')'9 (r) be the unit-speed geodesic on S passing through
map F(U) ~ U is also smooth. The matrix (21) is then invertible for all P when r = 0 and with tangent vector v there. Let a(r,8) = ')'9 (r).
(u, v) E U, so U1£ and U v are linearly independent for (u, v) E U. It follows that It can be shown that a is smooth for - f < r < f and all values of 8,
u : U -l' R 3 is an allowable surface patch.
where f is some positive number, and that it is an allowable surface
As to the first fundamental form of a, note first that patch for S defined for 0 < r < f and for 8 in any open interval of
E =
2
II au 11 = II d~ .yV(u) 11 2 = 1
length::::; 211". This is called a geodesic polar patch on S. Show that,
if 0 < R < f,
because .yv is a unit-speed curve. Next, we apply the second of the geodesic
equations (2) to i'v. The unit-speed parameter is u and v is constant, so we get
Fu = O. But when u = 0, we have already seen that au and a v are perpendic-
l R
II :~ 11
2
dr = R.
By differentiating both sides with respect to 8, prove that
ular, so F = O. It follows that F = 0 everywhere. Hence, the first fundamental
form of u is a r .U9 = O.

We have

G(O,v) = II uv(O,v)
2
11 = II ~z 2
11 = 1

because'Y is unit-speed. Finally, from the first geodesic equation in (2) applied
to the geodesic -y, for which u = 0 and v is the unit-speed parameter, we get
G1£(O, v) = 0. 0
geodesics
 p
200 Elementary Differential Geometry

This is called Gaus.'i Js Lemma - geometrically, it means that the

parameter curve r = R, called the geodesic circle with centre P
and radius R, is perpendicular to each of its radii, i.e. the geodesics
/ 9
passing through P. MInimal Surfaces
Deduce that the first fundamental form of a is
dr 2 + G(r, B)dB 2 ,
for some smooth function G(r, B).

In Section 8.4 we considered the problem of finding the shortest paths between
two points on a surface. We now consider the analogous problem in one higher
dimension, that of finding a surface of minimal area with a fixed curve as its
boundary. This is called Plateau's Problem. The solutions to Plateau's problem
\ turn out to be surfaces whose mean curvature vanishes everywhere. The study
of these so-called minimal surfaces was initiated by Euler and Lagrange in the
mid-18th century, but new examples of minimal surfaces have been discovered
quite recently.

9.1. Plateau's Problem

In Section 8.4, we found the condition for a curve on a surface to mininiise
distance between its endpoints by embedding the given curve in a family of
curves passing through the same two points, and studying how the length of
the curve varies as the curve varies through the family. Accordingly, we shall
now study a family of surface patches aT : U --+ R 3 , where U is an open subset
2
of R independent of 7', and 7' lies in some open interval (-6,6), for some 6 > O.
Let a = aD. The family is required to be smooth, in the sense that the map
(u, v, 7') --+ aT(u, v) from the open subset {(u, v, 7') 1 (u, v) E U, 7' E (-6, 6)}
of R 3 to R 3 is smooth. The surface variation of the family is the function
'P: U --+ R 3 given by

f{J = .TI r=O'

iT
201
202 Elementary Differential Geometry 9. Minimal Surfaces 203

where here and elsewhere in this section, a dot denotes d/dr. and this energy is proportional to its area. A soap film spanning a wire in the
Let lI" be a simple closed curve that is contained, along with its interior shape of a curve C should therefore adopt the shape of a surface of least area
int(lI"), in U (see Section 3.1). Then lI" corresponds to a simple closed curve with boundary C. By Corollary 9.1, this will be a minimal surface.
...,T = aT oll" in the surface patch aT, and we define the area function A(T) to More generally, if the soap film separates two regions of different pressure,
be the area of the part of aT inside ...,T: the film will adopt the shape of a surface of constant mean curvature. This is

A(T) = I!.
int(1f)
dAa·.
/
the case for a soap bubble, for example, for which the air pressure inside the
bubble is greater than the pressure outside. To see this, we apply the principle
Note that, if we are considering a family of surfaces with a fixed boundary curve of 'virtual work'. This tells us that, if the soap film is in equilibrium, and we
=...,
..." then...,T =
for all T, and hence cpT(U,V) 0 when (u,v) is a point on the imagine a ('virtual') change in the surface, the change in the energy of the film
must be the same as the work done by the film against the air pressure. If p
curve 11'.
is the pressure difference, the force exerted by the air on a small piece of the
surface of area LlA is pLlA, so the work done when it moves a small distance
Theorem 9.1 a perpendicular to itself is apLlA. On the other hand, the formula in Theorem
With the above notation, assume that the surface variation cpT vanishes along 9.1 shows that the change in area of the surface is proportional to aHLlA (note
the boundary curve lI". Then, that a is the component of the variation cp perpendicular to the surface). So p
is proportional to H. Since the pressure difference must be the same across the
A(O) = -2 j' (
lint(1f)
H(EG - p2)'/2 a dudv, (1) whole surface, so must the mean curvature H. Surfaces of constant non-zero
mean curvature were discussed in Section 7.4.
where H is the mean curvature of a, E, P and G are the coefficients of its first
For the moment, we give only one eXll.mple of a minimal surface; others will
fundamental form, a = cp.N, and N is the standard unit normal of a.
be given in the next section. This example already shows, however, that the
converse of Corollary 9.1 is false.
We defer the proof of this theorem to the end of this section.
If a has the smallest area among all surfaces with the given boundary curve
..." then A must have an absolute minimum at T = 0, so A(O) = 0 for all smooth Example 9.1
families of surfaces as above. This means that the integral in (1) must vanish
for all smooth functions a : U -t R. As in the proof of Theorem 8.2, this can
happen only if the term that multiplies a in the integrand vanishes, in other
"
A catenoid is the surface obtained by revolving the curve x = ~ cosh az in
the xz-plane around the z-axis, where a is a non-zero constant (a picture of
the catenoid can be found in Example 9.2). We take a = 1 for simplicity. The
words only if H = O. This suggests the following definition. catenoid can be parametrised by

Definition 9.1 a(u,v) = (coshucosv,coshusinv,u).

Then,
A minimal surface is a surface whose mean curvature is zero everywhere.
(It/, = (sinhucosv,sinhusinv, 1), av = (-coshusinv,coshucosv,O),
Theorem 9.1 and the preceding discussion then give (Iv. x (Iv = (-coshucosv, - coshusinv,sinhucoshu),
N = (-sech u cosv, - sech usin v, tanhu),
Corollary 9.1 D'uu = (coshucosv,coshusinv,O),
If a surface S has least area among all surfaces with the same boundary curve, (luu = (- sinh usin v, sinh ucosv, 0),
then S is a minimal surface. a vu = (- cosh u cos v, - cosh usin v, 0).
This gives the coefficients of the first and second fundamental forms of a as
Minimal surfaces have an interesting physical interpretation as the shapes
taken up by soap films. A soap film has energy by virtue of surface tension, 2
E=G=cosh u, P=O, L=-I, M=O, N=l.
204 Elementary Differential Geometry 9. Minimal Surfaces 205

The first three of these equations show that the parametrisation u is conformal, The graphs of 1 + e- 2a and 2a as functions of a clearly intersect in exactly one
and Proposition 7.1(ii) gives point a = ao, say, and the inequality (2) holds if a > ao. If this condition is
H= LG-2MF+NE =0 satisfied, the catenoid is not area minimising.
2(EG - P2) , It can be shown that if a < no the catenoid does have least area among all
showing that the catenoid is a minimal surface. surfaces spanning the circles C+ and C-.

It is time to prove Theorem 9.1.

Proof 9.1

Let cpr = iT"T, so that cpo = cp, and let NT be the standard unit normal of u T.
There are smooth functions a"T ,j3"T and "'( of (u, v, T) such that
tp"T = aTN"T + j3"T O'~ + ,"T u~,
so that a = aO. To simplify the notation, we drop the superscript T for the rest
of the proof; at the end of the proof we put T = O.
We have
Fix a > 0, and let b = cosh a. The surface S consisting of the part of the
catenoid with Izl < a has boundary the two circles C± of radius b in the planes
A(T) = /1 int(1l')
110''1,1. x Uv II dudv = /1 int(1l')
N·(O'u x O"v) dudv,

z = ±a with centres on the z-axis. Another surface spanning the same two so
circles is, of course, the surface So consisting of the two discs x 2 + y2 ~ b2 in
the planes z = ±a. The area of S is, by Proposition 5.2, A= /1 : int(1f) T
(N·(O'u x O'v)) dudv. (3)

{21r
i o La
r (EG _ p 2 )1/2dudv = (21r
i o La
r cosh 2 ududv = 21f(a + sinh a cosh a). Now,

The area of So is, of course, 21fb2 = 21T cosh 2 a. So the minimal surface S
will
~ (N.(uu x O"v)) = N.(O'u x Uv ) + N.(uu x O"v) +-N.(u" x uv). (4)
not minimise the area among all surfaces with boundary the two circles C± if Since N is a unit vector,
cosh 2 a < a + sinh a cosh u, i.e. if
N.(O'u x O'v) = N.N II 0"'1,1. x O'v II =o.
1 + e- 2a < 2a. (2)
On the other hand,
.
N •(0"'1,1.
) _ x O'v).(u" X O'v)
(0''1,1.
X Uv ~ 11
uuxUv II
2a
_ (uu·U'l,l.)(Uv.O'v) - (u"'O"v)(uv'O''')
- Iluuxuvll
_ G(uu'u,,) - F(O'v.uu)
- (EG - F2)l/2 '
using Proposition 5.2. Similarly,
N ( x·) _ E(O"v.uv) - F(u".uv)
. 0''1,1. O"V - (EG _ F2)l/2
a
206 Elementary Differential Geometry f 9. Minimal Surfaces 207

Substituting these results into Eq. (4), we get since such a parallel variation causes the surface to slide along itself and will
8« )) _ E(/Tv.lTv ) - F(iTu./T v + /Tu.iT v ) + G(/Tu.iTu) not change the shape, and in particular the area, of the surface. Thus, the
8T N. /T u X /Tv - (EG _ F2)1/2 (5)
main point is to prove Theorem 9.1 for normal variations, i.e. those for which
Now (3= =
"I 0 everywhere on the\surface. Making this restriction simplifies the
above proof considerably.
iru ::: CPu ::: nuN + f3u tl u + lut1v + aN u + {3tT uu + 'Ytruv,
/Tu.iT u = EfJu + F"'{u + (o-u.Nu)a + (/Tu./Tuu)fJ + (/Tu·/Tuvh·
Since O'u.Nu ::: -(J'uu.N ::: -L, (111..(111.11. ::: ~Eu and Uu.t1uv ::: ~Ev, we get EXERCISES

/Tu.iT u = E(3u + F"'{u - La + ~Eu(3 + ~Ev"'{. 9.1 Show that any rigid motion of R 3 takes a minimal surface to another
minimal surface, as does any dilation (x, y, z) >4 a(x, y, z), where a is
Similarly, a non-zero constant.
/Tv.iTu = F(3u + G"'{u - Ma + (Fu - ~Ev)(3 + ~Gu"'{, 9.2 Show that z = l(x, y), where 1 is a smooth function of two variables,
is a minimal surface if and only if
+ F"'{v + ~Ev(3 + (Fv - ~Guh,
/Tu.iT v = E(3v - Ma
(1 + 1;)1.. - 21.1.1•• + (1 + mI•• = o.
/Tv.iTv = F(3v + G"'{v - Na + ~Gu(3 + ~Gv"'{. 9.3 Show that every umbilic on a minimal surface is a planar point (see
Substituting these last four equations into the right-hand side of Eq. (5), sim- Proposition 6.3).
plifying, and using the formula for H in Proposition 7.1(ii), we find that 9.4 Show that the gaussian curvature of a minimal surface is S 0 ev-
erywhere, and that it is zero everywhere if and only if the surface is
:T (N.(/T u x /Tv)) = ((3(EG - F 2)1/2)u + ("'{(EG - F 2)1/2) v part of a plane. (Use Proposition 6.5.) We shall obtain a much more
precise result in Corollary 9.2.
- 2aH(EG - F 2)1/2. (6)
9.5 Show that there is no compact minimal surface. (Use Proposition 7.6
Comparing with Eq. (3), and reinstating the superscripts, we see that we must and Exercise 9.4.)
prove that

If.
IOt(1r)
{((3°(EG - F 2)1/2)
11.
+ ("'{O(EG - F 2)1/2) } dudv = O.
v
(7)
9.2. Examples of Minimal Surfaces
But by Green's Theorem (see Section 3.1), this integral is equal to

L (EG - F 2)1/2(fJodv - "'{Odu) , The simplest minimal surface is, of course, the plane, for which both principal
curvatures are zero everywhere. Apart from this, the first minimal surfaces to
and this obviously vanishes because (30 = "'{o = 0 along the boundary curve 1I'. be discovered were those in the following two examples.
This completes the proof of Theorem 9.1. 0
Example 9.2
Note that we did not quite use the full force of the assumptions in Theorem
9.1, since they imply that aO (= a) vanishes along the boundary curve, and A catenoid is obtained by rotating a curve x = ~ cosh az in the xz-plane around
this was not used in the proof. So Eg. (1) holds provided the surface variation the z-ij,Xis, where a > 0 is a constant. We showed in Example 9.1 that this is a
minimal surface (we only dealt there with the case a = 1, but the general case
I{) is normal to the surface along the boundary curve.
follows from it by using Exercise 9.1).
Note also that Theorem 9.1 is intuitively obvious for variations I{) that are
parallel to the surface, i.e. those for which a = 0 everywhere on the surface,

\
208 Elementary Differential Geometry
y 9. Minimal Surfaces
I 209

We suppose now that, for some value of u, say u = uo, we have g(uo) f:. O.
We shall then have iJ(u) 1- 0 for u in some open interval containing uo. Let
(a,!3) be the lar~est such interval. Supposing now that u E (a,{3), the unit-
speed condition P + g2 == 1 gives-Cas in Example 7.2)
. .. !
/9 - fg = --;-,
g
so we get

H ==!
2
(! _1)f iJ •
Since tl = 1 - j2 , S is minimal if and only if

f!=l-j2. (8)
To solve the differential equation (8), put h == j, and note that
! = dh == dh df = h dh
dt df dt dj"
Hence, Eq. (8) becomes
The catenoid is a surface of revolution. In fact, apart from the plane it is dh 2
the only minimal surface of revolution: fh df =:: 1- h .

Note that, since iJ 'f; 0, we have h 2 f:. 1, so we can integrate this equation as
Proposition 9.1 follows:

Any minimal surlace 01 revolution is either part 01 a plane or can be obtained hdh =Jdf
by applying a rigid motion to part 01 a catenoid. J 1- h 2 f'
1
.~==af,
v1- h-
Prool9.1
_ v"a2"--:f"'""2---1
By applying a rigid motion, we can assume that the axis of the surface S is the h - af '
z-axis and the profile curve lies in the xz-plane. We parametrise S in the usual where a is a non-zero constant. (We have omitted a ±, but the sign can be
way (see Example 6.2): changed by replacing u by -u if necessary.) Writing h = df/ du and integrating
(T(u, v) = (f(u) cos v, feu) sinv, y(u», again,

where the profile curve u H (f(u) , 0, y(u)) is assumed to be unit-speed and

I > O. From Example 6.2, the first and second fundamental forms are
J-vra=:;::f=:'f~;;=;lf-=1 = J du,
1
du 2 + l(u)2dv 2 and (jg - !iJ)du 2 + fiJdv 2, f = (tVl + a2 (u + bp,
respectively, a dot denoting djdu. By Proposition 7.1(ii), the mean curvature where b is a constant. By a change of parameter U H u + b, we can assume that
is b == O. So
iJ)
H="21 ( /9-fiJ+7
. ..
.
--------------------------aaya
210 Elementilry Differentiill Geometry 9. Minimal Surfilces 211

To compute g, we have line starts along the x-axis, at time v the centre of the line is at (0,0, a.v) and
it has rotated by an angle wv. Hence, the point of the line initially at (u, 0, 0)
'2 _
g -
1_ /'2 _ 1_ h2 _ _ 1_
- - a 2 j2' is now at the point with position vector
dg _ ± 1 u(u,v) = (u'coswv, usinwv,O'v).
~ . du - VI + a~:Lu2 t
We leave it to Exercise 9.6 to check that this is a minimal surface.
g= ±~ sinh-1(au) + c (where c is a constant),
a
.. au = ± sinh(a(g - en,
1
f = - cosh(a(g - c)).
a
Thus, the profile curve of S is
1
x = - cosh(a(z - c)).
a
By a translation along the z-axis, we can assume that c = 0, so we have a
catenoid.
We are not quite finished, however. So far, we have only shown that the part
of S corresponding to u E (a, (3) is part of the catenoid, for in the proof we
used in an essential way that 9 ¥= O. This is why the proof has so far excluded
the possibility that S is a plane. To complete the proof, we argue as follows. We have the following analogue of Proposition 9.1.
Suppose that (3 < 00. Then, if the profile curve is defined for values of u ~ (3,
we must have 9«(3) = 0, for otherwise il would be non·zero on an open interval Proposition 9.2
containing (3, which would contradict our assumption that (a, {3) is the largest
Any ruled minimal surface is part of a plane or part of a helicoid.
open interval containing UQ on which il ¥= O. But the formulas above show that

·2 1 if U E (a,I3), Proof 9.2

9 = 1 + a2 u 2
We take the usual parametrisation
so, since il is a continuous function of u, 9(13) = ±(l + a 2/32)-1/2 ¥= O. This
contradiction shows that the profile curve is not defined for values of U ~ (3. Of t7(u, v) = "Y(u) + v6(u)
course, this also holds trivially if /3 = 00. A similar argument applies to a, and (see Example 4.12), where 'Y is a curve that meets each of the rulings and 6(u)
shows that (a., (3) is the entire domain of definition of the profile curve. Hence, is a vector parallel to the ruling through 'Y( u). We begin the proof by making
the whole of S is part of a catenoid. some simplifications to the parametrisation.
The only remaining case to consider is that in which g(u) = 0 for all values First, we can certainly assume that 116(u) II = 1 for all values of u. We assume
of U for which the profile curve is defined. But then g(u) is a constant, say d, also that 6 is never zero, where the dot denotes d/du. (We shall consider later
and S is part of the plane z = d. 0 what happens if 6(u) = 0 for some values of u.) We can then assume that 6 is
a unit-speed curve (we do not assume that 'Y is unit-speed). These assumptions
Example 9.3 imply that 6.6 = 6.6 = O. Now we consider the curve

A helicoid is a ruled surface swept out by a straight line that rotates at constant .y(u) = 'Y(u) - (;-.6)6(u).
speed about an axis perpendicular to the line while simultaneously moving at If ii = v + ;-.ts,
the surface can be reparametrised using" and the parameters
constant speed along the axis. We can take the axis to be the z-axis. Let w be u and v, namely
the angular velocity of the rotating line and a its speed along the z-axis. If the u(u, v) = ,,(u) + v6(u) ,
 212 Elementary Differential Geometry
i'" 9. Minimal Surfaces 213

but now we also have Equation (12) shows that the curvature of the curve 6 is 1, and that its principal
normal is -6. Hence, its binormal is 6 x (-6), and since
+6 = ('\' - :.. (,\,.6)6 - (,\,.6)6).6 = 0,
d . - ..
-d (6 x 6) = 6 x 6 + 6 x 6 = -6 x 6 = 0
since 6.6 = 0 and 6.6 = 1. This means that we could have assumed that '\'.6 =0 u '
at the beginning, and we make this assumption from now on. it follows that the torsion of 6 is zero. Hence, 6 parametrises a circle of radius
We have 1 (see Proposition 1.5). By applying a rigid motion, we can assume that 6 is
the circle with radius 1 and centre the origin in the xy-plane, so that
au ='Y + v6, (Iv = 6,
~ E = II'Y + v611 2 , F = ('\' + v6).6 = '\'.6, G = 1. 6(u) = (cosu,sinu,O).

Let A = (EG - F 2 )'(2. Then, i,From Eq. (12), we get 6.(..y x 6) = -6.(,\, x 6) = 0, so by Eq. (10),

N = A-'(,\, + v6) x 6. i.(6 x 6) = O.

Next, we have It follows that i is parallel to the xy-plane, and hence that

D'uu = i' + v6, 6,

tT uv = (1V11 = 0, -y(u) = (f(u), g(u), '!u + b),
L = A-'(i + v6).((,\, + v6) x 6), where f and 9 are smooth functions and a and b are constants. If a = 0, the
surface is part of the plane z = b. Otherwise, Eq. (9) gives
M = A-'6.((..y + v6) x 6) = A-'6.(..y x 6),
N=O. 9 cosu - i sin u = 2(j cosu + gsin u). (13)
Hence, the minimal surface condition We finally make use of the condition ..y.6 = 0, which gives
H = LG - 2MF+NE =0 isinu = gcos u. (14)
2A2
Differentiating this gives
gives
isinu + icosu = gcosu -; gsinu. (15)
(i + v6).((,\, + v6) x 6) = 2(6...y)(6.(..y x 6».
Equations (13) and (15) together give
This equation must hold for all values of (u, v). Equating coefficients of powers
of v gives
., fcosu + gsinu = O.
i.('\' x 6) = 2(6.'\')(6.('\' x 6», (9) and using Eq. (14) we get i
= g = O. Thus, f and 9 are constants. By a
i.(6 x 6) + 6.(,\, x 6) = 0, (10) translation of the surface, we can assume that the constants f, 9 and b are
zero, so that
6.(6 x 6) = o. (11)
-y(u) = (0,0, au)
Equation (11) shows that 6,6 and 6 are linearly dependent. Since 6 and 6
are perpendicular unit vectors, there are smooth functions a(u) and f3(u) such and
that
a(u,v) = (vcosu,vsinu,au),
6 = a6 + 136. which is a helicoid.
But, since 6 is unit-speed, 6.6 = O. Also, differentiating 6.6 = 0 gives 6.6 = We assumed at the beginning that 6 is never zero. If 6 is always zero, then
-6.6 = -1. Hence, a = -1 and 13 = 0, so 6 is a constant vector and· the surface is a generalised cylinder. But in fact a
generalised cylinder is a minimal surface only if the cylinder is part of a plane
6= -6. (12)
(Exercise 9.8). The proof is now completed by an argument similar to that used
 +
214 Elementary Differential Geometry 9. Minimal Surfaces 215

at the end of the proof of Proposition 9.1, which shows that the whole surface
is either part of a plane or part of a helicoid. 0

After the catenoid and helicoid, the next minimal surfaces to be discovered
were the following two.

Example 9·4
Eimeper's minimal surface is

D'(u, v) = (u - ~u3 + uv 2tV - ~V3 + vu2,u2 - 2

V) .

It was shown in Exercise 7.15 that this is a minimal surface.

The white squares have centres of the form (mll', mr), where m and n are
integers with m + n even. Since, for such m, n,
cos(y+ nrr) cosy
cos(x + mrr) = cosx'
it follows that the part of the surface over the square with centre (mn, nrr) is
obtained from the part over the square with centre (0,0) by the translation
(x, y, Z).H (x + mrr, y + nrr, z). So it suffices to eX}~~bit the part of the surface
Strictly speaking, this is n~t a surface patch in the sense used in this book as over a smgle square: C

it is not injective. The self-intersections are clearly visible in the picture above.
However, if we restrict (u t v) to lie in sufficiently small open sets, (1 will be
injective by the inverse function theorem.

Example 9.5
Scherk's minimal surface is the surface with cartesian equation
z- _ In (COSY)
-- .
cos x
It was shown in Exercise 7.16 that this is a minimal surface. Note that the
surface exists only when cos x and cos y are both > 0 or both < 0, in other
words in the interiors of the white squares of the following chess board pattern,
in which the squares have vertices at the points (rr /2 + mrr, rr /2 +mr), where m
and n are integers, no two squares with a common edge have the same colour,
and the square containing the origin is white:
216 Elementary Differential Geometry 9. Minimal Surfaces 217

EXERCISES (ii) the parameter curve u = IT is a parabola;

(iii) the parameter curve v = 0 is a cycloid (see Exercise 1.7).
9.6 Show that the helicoid is a minimal surface. Show also that each of these curves, when suitably parametrised, is
9.7 Show that the surfaces u t in the isometric deformation of the helicoid a geodesic on Catalan's surface.
into thr catenoid given in Exercise 5.8 are minimal surfaces.
9.8 Show that a generalised cylinder is a minimal surface only when the
cylinder is part of a plane. 9.3. Gauss Map of a Minimal Surface
9.9 A translation surface is a surface of the form
3
z = f(x) + g(y), Recall from Section 7.3 that the Gauss map of a surface patch u : U -+ R
associates to a point u(u, v) of the surface the standard unit normal N (u, v)
where f and 9 are smooth functions. (It is obtained by mov- regarded as a point of the unit sphere 52. By Eq. (15) in Chapter 7,
ing the curve u I-t (u, 0, f(u)) parallel to itself along the curve
v t-+ (O,v,g(v)).) Use Exercise 9.2 to show that this is a minimal
surface if and only if where K is the gaussian curvature of u, so N will be regular provided K is
d 2f/dx 2 ~g/dy2 nowhere zero, and we assume this for the remainder of this section.
=
1 + (df /dx)2 1 + (dg/dy)2'
Deduce that any minimal translation surface is part of a plane or Proposition 9.3
can be transformed into part of Scherk's surface in Example 9.5 by
Let O'(u, v) be a minimal surface patch with nowhere vanishing gaussian cur-
a translation and a dilation (x,y,z) I-t a(x,y,z) for some non-zero
vature. Then, the Gauss map is a conformal map from u to part of the unit
constant a.
sphere.
9.10 Verify that Catalan's surface
u (u, v) = (u - sin u cosh v, 1 - cos u cosh v, -4 sin ~ sinh ¥) We should really be a little more careful in the statement of this propo-
sition, since for us conformal maps are always diffeomorphisms (see Section
is a conformally parametrised minimal surface. (As in the case of 5.3). However, even if N u >< N" is never zero, it does not follow that the map
Enneper's surface, Catalan's surface has self-intersections, so it is u(u, v) I-t N(u, v) is injective (see Exercise 9.12(ii)). Nevertheless, the inverse
only a surface if we restrict (u, v) to sufficiently small open sets.) function theorem tells us that, if (uo, vo) E R 2 is a point where 0' (and hence
N) is defined, there is an open set U containing (uo,vo) on which u is defined
and on which N is injective. Then, N : U --+ 52 is an allowable surface patch
on the unit sphere 52, and the Gauss map is a diffeomorphism from O'(U) to
N(U).

Proof 9.3
By Theorem 5.2, we have to show that the first fundamental form
(N u .Nu )du 2 + 2(N u .Nv )dudv + (N".N,,)dv 2
of N is proportional to that of 0'. Form the symmetric 2 >< 2 matrix
Show that
(i) the parameter curve on the surface given by u = 0 is a straight
line;
 218 Elementary Differential Geometry 9. Minimal Surfaces 219

in the same way as we associated symmetric 2 x 2 matrices F j and Fj1 to the We saw in Exercise 5.14 that a conformal parametrisation of the plane is
first and second fundamental forms of (f in Section 6.3. Then, we have to show necessarily holomorphic or anti-holomorphic, so this proposition strongly sug-
that . I·
gests a connection between minimal surfaces and holomorphic functions. This
connection turns out to be very extensive, and we shall give an introduction to
(16)
I it in the next section.
for some scalar A. I

By Proposition 6.4 l

N".Nu = a2 tlu.(f... + 2aba....tTv + b2 tlv.tlv = a2 E + 2abF + b2G , EXERCISES

where (: ~) = -Wand W = Fi 1 Fll is the Weingarten matrix. Computing 9.11 Show that the scalar A appearing in the proof of Proposition 9.3 is
equal to -K, where K is the gaussian curvature of the surface.
N ....N II and NII.NII in the same way gives
2 2 9.12 Show that
:FIll = ( a E + 2abF + b G acE + (ad + bc)F + bdG) (i) the Gauss map of the catenoid is injective and its image is the
acE + (ad + be)F + bdG c2 E + 2edF + d2G
whole of the unit Jlphere except for the north and south poles;
= (~ :) (; ~) (~ ~) (ii) the image of the Gauss map of the helicoid is the same as that
of the catenoid, but that infinitely many points on the helicoid
= (-W)tF1(-W) are sent by the Gauss map to any given point in its image.
1
= (-:Fi :FII)t:Fr(-:Fi 1:F11)
1 1
= FII:Fi F1:Fi :Fll
= FlI Fi 1 :F11. 9.4. Minimal Surfaces and Holomorphic
Hence, Eq. (16) is equivalent to Functions
:Fi 1 :F11:F;1 F11 = AI,
In this section, we shall make use of certain elementary properties of holomor-
i.e. W2 = AI. phic functions. Readers without the necessary background in complex analysis
But, may safely omit this section, whose results are not used anywhere else in the

W2 = (a
b d
C)2 = ( a2 + bc
b(a + d)
c(a + d))
~ + be .
book.
We shall need to make use of special surface patches on a minimal surface.
Recall from Section 5.3 that a surface patch tI : U -t R 3 is called conformal if
Now, recall from Section 6.3 that the principal curvatures K:l and I\, are the its first fundamental form is equal to E(du 2 + dv 2 ) for some positive smooth
. I 2
elgenva ues of W. Since the sum of the eigenvalues of a matrix is equal to the
function E on U.
sum of its diagonal entries,

Kl + K:z = -(a + d). Proposition 9.4

If tI is minimal, the mean curvature H = ~(K:l + 1\,2) vanishes, so a + d = 0 and
Every surface has an atlas consisting of conformal surface patches.
hence
We shall accept this result without proof (the proof is non-trivial).
Let tI : tJ - t R 3 be a conformal surface patch. We introduce complex
as we want. o " coordinates in the plane in which U lies by setting
(= u+iv for (u,v) E U,
220 Elementary Differential Geometry
+ 9. Minimal Surfaces 221

and we define We compute

lp«) = tT u - itT v • (17) .:::ltT.tT u :::::: tTuu.tT u + tTvv·tTu
1
Thus, I{) = (If!l, 1f!2, 1f!3) has three components, each of which is a complex-valued = "2(tT u .tT u )u + (O'v.tTu)v - (O'v.O'uv)
function of (u, V),I i.e. of (. The basic result which establishes the connection 1
between minimal surfaces and holomorphic functions is = "2(O'u.tTu - 0'1I.tT1I)u + (0'1I. tTu)v.
But, since 0' is conformal, tTu.(fu :::::: O'v.tTv and O'u.tTv :::::: O. Hence, .:::lO'.O'u :::::: O.
Proposition 9.5 Similarly, .:1tT.tT 11 :::::: O. 0

~et tT : U : R be a con!ormal surface patch. Then tT is minimal if and only

3
The holomorphic function <P associated to a minimal surface tT is not arbi-
zf the functzon I{J defined m Eq. (17) is holomorphic on U.
trary, however:
Saying that lp is holomorphic means that each of its components /I' '" and
. h 1 h' .,...1,.,...2 Theorem 9.2
'P3 IS 0 omorp IC.
If 0' : U ---+ R 3 is a conformally parametrised minimal surface, the vector-valued
Proof 9.5 holomorphic function V' = ('P1, <P2, <P3) defined in Eq. (17) satisfies the following
conditions:
Let ~(u,~) be a complex-valued smooth function, and let a and (3 be its real
and lmagmary parts, so that 'P = a + ifJ. The Cauchy-Riemann equations
(i) If'I+ 4'~ + 4'~ :::::: 0;
(ii) V' is nowhere zero.
au = fJv and Q v = -f3u Conversely, if U is simply-connected, and if If!l, 4'2 and 1f'3 are holomorphic
functions on U satisfying conditions (i) and (ii) above, there is a conformally
ar~ the necessary and sufficient conditions for <P to be holomorphic. Applying
parametrised minimal surface tT : U ---+ R 3 such that V' = (<PI, 4'2, <P3) satis-
thiS to each of the components of I{), we see that I{) is holomorphic if and only if fies Eq. (17). Moreover, tT is uniquely determined by "Pl, l{)2 and 4'3 up to a
(tTu)u = (-tTv)v and (tTu)v = -(-tTv)u. translation.
The second equation imposes no condition on tT, and the first is equivalent to An open subset U of R 2 is said to be simply-connected if every simple closed
tT uu +tT vv = O. S~ we have to show that (f is minimal if and only if the laplacian
curve in U can be shrunk to a point staying inside U. Intuitively, this means
.::ltT = tT uu + tTvv IS zero.
that U has no 'holes'.
By Proposition 7.2(ii) and the fact that tT is conformal the mean curvature
of tT is given by ,

H:::::: L +N
2E '
so tT is minimal if and only if L +N :::::: 0, Le.

(tT uu + tTvv)·N :::::: O. (18)

Obviously, then, tT is minimal if .:::ltT = O. For the converse, we have to show
t~at .:::ltT = 0 if E~. (18) holds. It is enough to prove that .:::ltT.tT" = .:::ltT:tTv = 0,
smce {tTu,tTv,N} IS a basis ofR3 •
simply-connected not simply-connected
In the course of the following proof, and in the proof of Proposition 9.7
below, we shall need to recall that, if F is a holomorphic function of ( = u+iv,
 222 9. Minimal Surfaces 223
Elementary Differential Geometry

then Before giving some examples, we observe that, if a holomorphic funetio.n

Ip satisfies the conditions in Theorem 9.2, so does il/). If I() is the holomorp~lc
Fu =F I
, Fv = iF I
, (F)u = pt, (F) v = -iF', function corresponding to a minimal surface S, the minimal surface to whIch
where F
I
=
dF/d( is the complex derivative of F, and the bar denotes complex- il/) corresponds is called the conjugate of S. It is well defined by S up to a
conjugate. translation.

Proof 9.2 Example 9.6

Suppose first that iT = (0'1,0'2,0'3) is minimal, where a k : U --t R for k = 1,2,3. The parametrisation
We have to show that I/) = (i;?1,ip2,ip3) satisfies conditions (i) and (ii). Since tI(u,v) = (coshucosv,coshusinv,u)
ipk = a~ - ia~ for k = 1,2,3,

3 3 of the catenoid is conformal (see Example 9.1). The associated holomorphic

L ip% = L (0':)2 - (a~)2 - 2ia~a~) = II UU 11 2
- II tl v W-2iuu .tlv , (19) function is
1:=1 1:;;::1 I/)(() = tlu - iCT v
which vanishes since a is conformal. Finally, I{J = 0 if and only if tlu tl v = 0, = :::: (sinh u cos v + i cosh u sin v, sinh u sin v - i cosh u cos v, 1)
and this is impossible since u is regular. = (sinh(u + iv), -i cosh(u + iv), 1)
For the converse, take I/) satisfying conditions (i) and (ii). We must show
= (sinh(, -icosh(, 1).
that I{J arises from a minimal surface as above, and that this minimal surface
is unique up to a translation of R 3 . Fix (uo, vo) E U and define tI as the real Note that conditions (i) and (ii) in Theorem 9.2 are satisfied, since cp is clearly
part of a complex line integral: never zero and the sum of the squares of its components is

u(u,v) = 9te L 1p(I;)dl;,

sinh 2 ( - cosh 2 ( + 1 = O.
Let us determine the conjugate minimal surface it of the catenoid. From the
where'" is any curve in U from (uo, vo) to (u, v) E U. The fact that U is proof of Theorem 9.2,
simply-connected implies, by virtue of Cauchy's Theorem, that J1r 1p(1;)dl; is in-
dependent of the path", chosen, and hence so is tI(u, v). Now, .«() = J", Ip(!;)~
is a holomorphic function of ( = u + iv, and ./«() = Ip«(). Hence, by the facts
u(u, v) = 9le 1 (i sinh!;, cosh 1;, i) dl;

stated just before the beginning of the proof, = !ne(i cosh (, sinh (, i()

tlu = !ne(.u) = 9le(.') = 9le(Ip), = (- sinh u sin v, sinh u cos v, -v),

Uv = 9le(• .,) = 9le(i.') = -Jm(Ip), (20) up to a translation. If we reparametrise by defining 11, = sinh u, v = v + 7[" /2, we
get the surface
so Ip = tlu -
iuv .
To complete the proof, we have to show that u is a conformal surface patch. (fi,ii) H (11,cosii,usinii,-ii),
But, condition (ii) and Eq. (20) show that tlu and tl v are not both zero. By after translating by (0,0, -7["/2), which is obtained from the helicoid in Exercise
condition (i) and Eq. (19), II tlu II ~ II U V II and tlu.tI., = O. Since au and U v are 4.14 by reflecting in the z-axis.
not both zero, this proves that Uu and U v are both non-zero and perpendicular, Note that the parametrisation of the helicoid given in Example 4.14 is not
hence linearly independent, so that u is a regular surface patch; it also proves conformal, so the constructions in this section cannot be applied to it.
that tI is conformal.
If another conformal minimal surface 17 corresponds to the same holomor- It is actually possible to 'solve' the conditions on Ip in Theorem 9.2.
phic function tp as tT, then Uu = tlu and Uv = tI" everywhere on U, which
implies that iT - a is a constant, say a, so that 17 is obtained from tI by trans-
lating by the vector a. 0
 ---r------ep-------------------------
---n-4-------------E-1ementary Dif erential Geometry 9. Minimal Surfaces 225

Proposition 9.6 Simple algebra shows that Eqs. (22) and (23) imply Eq. (21). Finally, Eq, (23)
implies that f g2 is holomorphic, and the argument with Laurent expansions in
Let f (() be a holomorphic junction on an open set U in the complex plane, the first part of the proof now gives the condition on the zeros and poles of f
not identically zero, and let g(() be a meromorphic junction on U such that, if
andg. 0
(0 E U is a pole of g of order m > 1, say, then (0 is al30 a zero of f of order
I -
~ 2m. Then, I,
We give only one application of Weierstrass's representation.
I{) = Of(l- g2), ~f(l+ g2), f9) (21)
Proposition 9.7
satisfies conditions (i) and (ii) in Theorem 9.2, and conversely every holomor-
The gaussian curvature 0/ the minimal surface corresponding to the junctions
phic junction I{) satisfying these conditions arises in this way.
f and 9 in Weierstrass's representation is
The correspondence given by Theorem 9.2 and Proposition 9.6 between -161dg/ d(1 2
pairs of functions f and 9 and minimal surfaces is called Weierstrass's repre- K = If12(1 + IgI 2)4'
sentation. Proof 9.7
This is a straightforward, if tedious, computation, and we shall omit many of
Proof 9.6 the details. Define <p by taking complex-conjugates of each component of I{).
Suppose that f and 9 are as in tbe statement of the proposition. If 9 has a pole Then,
of order m 2:: 1 at (0 E U, and f has a zero of order n 2:: 2m at (0, then the I{)+<p I{)-I{)
tT u = -2-' (Ttl = ~.
Laurent expansions of f and 9 about (0 are of the form
b Hence, the coefficients of the first fundamental form are given by
f(() = a(( - (o)n + ... and g(() = (( _ (o)m + ... ,
E = G= ~ III{) + iO 11 2 = ~1{).iO, F = 0,
where a and b are non-zero complex numbers and the ... indicates terms in-
volving higber powers of ( - (0. Then, since I{).I{) = iO.iO = O. Substituting the formula for I{) into Eq. (14) and simpli-
fying, we find that the first fundamental form is
f(1 ± g2) = ±ab2(( - (0)n-2m + ... and fg = ab(( - (ot- m + ...
1 2
-lfI
4 (1 + IgI 2)2(du 2 + dv 2). (24)
involve only non-negative powers of (- (0, so I{) is holomorphic near (0. Since
it is clear that I{) is holomorphic wherever 9 is holomorphic, it follows that the Next,
function I{) defined by Eq. (21) is holomorphic everywhere on U. It is clear that 1 1
I{) is identically zero only if / is identically zero, and simple 'algebra shows that tT u X tT v = 4i (I{) + iO) x (iO -I{)) = 2il{) x <p,
I{) satisfies condition (i) in Theorem 9.2.
Conversely, suppose that I{) = ('PI, 'P2, 'P3) is a holomorphic function satis-
II tTu x tT v 11 2 = -~(I{) X <p).(1{) x iO),
fying conditions (i) and (ii) in Theorem 9.2. Define = -~((I{).I{))(<p.<p) - (l{)·iO)2),
f = 'PI - i'P2, 9 = 'P3. (22) -)2 ,
1 (I{).I{)
'PI - 1'P2 = 4
This definition makes sense, for if 'PI and 'P2 were both zero, then 'P3 would
be zero by condition (i), and this would violate condition (ii). Since I{) is holo- N=iiO
I{).I{)
X
.!.
morphic, f is holomorphic and 9 is meromorphic. Condition (i) implies that
In terms of / and g, this becomes
('PI + i'P2)( 'PI - i'P2) = -'P~, and hence that
1
'PI +i'P2 = _fg2. (23) N = 1 + Igl 2 (g + g, -i(g - g), Igl 2 - 1) . (25)
 LL/
elementary UlTrerentlal LJeometry 9. Minimal Surfaces

Using the remarks preceding the proof of Theorem 9.2 and the formulas 9.15 Calculate the functions f and 9 in Weierstrass's representation for
the catenoid and the helicoid.
L = -uu.Nu, M = -uu.NlJ' N = -uv .N lJ
9.16 Find the holomorphic function cp corresponding to Enneper's mini-
(which follow by differentiating uu.N = uv.N = 0), we find that the second mal surface given in Example 9.4. Show that its conjugate minimal
fundamental form is surface coincides with a reparametrisation of the same surface ro-
I

-~ ((Vg' + f gl)(du 2 + dv 2 ) + 2i(fg' - fgl)dudv) . (26) tated by 1r / 4 around the z-axis.

9.17 Find a parametrisation of Henneberg's surface, the minimal surface
Combining Eqs. (24), (25) and (26), and using the formula for the gaussian corresponding to the functions f«() = 1 - (-4, g(() == ~ in Weier-
curvature K in Proposition 7.1(i), we finally obtain the formula in the statement strass's representation. (Reparametrise by putting ( == e<.) The fol-
of the proposition. 0 lowing are a 'close up' view and a 'large scale' view of this surface.
':.

Corollary 9.2
Let S be a minimal surface that is not part of a plane. Then, the zeros of the
gaussian curvature of S are isolated.

This means that, if the gaussian curvature K vanishes at a point P of

S, then K does not vanish at any other point of S sufficiently near to P.
More precisely, if P lies in a surface patch u of S, say P = u(uo,vo), there
is a number f > 0 such that K does not vanish at the point u(u, v) of S if
0< (u - UO)2 + (v - VO)2 < (02.
Henneberg: close up
Proof 9.5

From the formula for K in Proposition 9.8, K vanishes exactly where the
meromorphic function g/ vanishes. If i is zero everywhere, so is K and S is
part of a plane (this was shown in Proposition 6.5, but follows immediately
from Eq. (18) which shows that N is constant if 9 is constant). But it is a
standard result of complex analysis that the zeros of a non-zero meromorphic '.
function are isolated, so if K is not identically zero its zeros must be isolated.
o

EXERCISES

9.13 Find the holomorphic function cp corresponding to the plane pass-

ing through the origin with unit normal a. What is its conjugate
surface?
9.14 If a minimal surface S corresponds to a pair of functions f and 9
in Weierstrass's representation, to which pair of functions does the
conjugate minimal surface of S correspond? Henneberg: large scale
228 Elementary Differential Geometry v-------------------------
9.18 Let 9 be the Gauss map of a minimal surface in Weierstrass's repre-
sentation. Let 7r : 8 2 \(0,0,1) -t R 2 be the stereographic projection
map defined in Example 5.7, and identify the point (u,v) E R 2 with
10
the complex number u + iv as usual. Show that 7r 0 9 is the function Gauss's Theorema Egregium
g.

One of Gauss's most important discoveries about surfaces is that the gaussian
curvature is unchanged when the surface is bent without stretching. Gauss
called this result 'egregium', and the Latin word for 'remarkable' has remained
attached to his theorem ever since.

10.1. Gauss's Remarkable Theorem

The purpose of this section is to prove the following important theorem and to
study its consequences.

Theorem 10.1
The gaussian curvature of a surface is preserved by isometries.

This means, more precisely, that if 8, and 8 2 are two surfaces and if f :
8, -t 8 2 is an isometry between them, then for any point P in 8, the gaussian
curvature of 8, at P is equal to that of 82 at f(P). The theorem is sometimes
expressed by saying that the gaussian curvature is an intrinsic property of a
surface, for it implies that the gaussian curvature could be measured by a bug
living in the surface.
In proving the theorem, it is enough by Theorem 5.1 to consider the case
of a surface patch a on 8" and to prove that, if a and f 0 a have the same
229
 230
¥
,
Elementary Differential Geometry 10. Gauss's Theorema Egregium 231

first fundamental forms, then they have the same gaussian curvature. This is Lemma 10.1
far from obvious, for the formula
With the above notation, we have
LN-M 2
'~= EG-F2 e~ .e~ - e~ .e~ = >..' JJ." - >..U JJ.' (2)
=at/-f3u (3)
that we obtained for the gaussian curvature K in Proposition 7.1(i) depends
on the coefficients L, Nand M of the second fundamental form, as well as the LN-M2
(4)
coefficients E, F and G of the first fundamental form. Hence, the theorem is = (EG - F2)1/2'
telling us that LN - M 2 can be expressed in terms of E, F and G (although
it is not saying that L, M and N individually can be so expressed). Proof 10.1
To prove the theorem, we shall make use of a smooth orthonormal basis Equation (2) follows immediately from Eq. (1), since e', en and N are perpen-
{e', e"} of the tangent plane at each point of the surface patch, where 'smooth' dicular unit vectors.
means that e' and e" are smooth functions of the surface parameters (n, v). Next, we compute
Then, {e', e", N} is an orthonormal basis of R 3 (N being the standard unit
normal of 0'), and we shall assume that it is right-handed, i.e. that N = e' x e". av - /3 u = a (e ''')
au a ('e .e")u
.ev - av (by Eq. (1))
This can always be achieved by interchanging e' and e" if necessary. Note that ,,, 'If ,,,,,,
= eu·et/ + e .eut/ - et/.e u - e .eut/
the dashes on e' and e" have nothing to do with derivatives.
'" ,,,
= eu·e v - ev·e u ·
We can express the partial derivatives of e' and e" with respect to u and
This proves Eq. (3).
v in terms of the orthonormal basis {e', e", N}. Since both partial derivatives
To prove Eq. (4), we use the formula
of e' are perpendicular to e', the e' components of e~ and e~ are zero (and J

similarly for e"). Thus, N u x N v = KtT u x (Tv

e~ = ae/l + A'N, (see Eq. (15) in Chapter 7). Combining this with the formulas
e~ = /3e" + J.t'N, N = au x a v , II00u X a v II = (EG _ F 2)'/2
e~ = -a'e' + A"N,
II au x av II
(see Proposition 5.2), we get
e~ = -f3' e' + /JON,
LN-M 2
for some scalars a, fi, a' ,[3', ..\' , Ii, A" ,J1." (which may depend on u and v). More- N u x N v = (EG _ F2)'/2 N,
over, by differentiating the equation e' .e lf = 0 with respect to u, we see that
e~.e" = -e/.e~. i.e. Q/ = a (and similarly /3' = fj). Thus,
and hence
LN-M 2
e~ = ae" + >"'N, (N x Nv).N = (EG _ F2)'/2'
u (5)
e~ = {lelf + JiN, Since N = e' x e", we get
(1)
e~ = -ae ' + ,\"N, (N u x Nv).N = (N u x Nv).(e' x e"),
e~ = -pe' + Il'N. = (Nu.e')(Nv.e") - (Nu.e")(Nv.e')
The following lemma is the crucial step in the proof of Theorem 10.1. = (N.e~)(N.e~) - (N.e~)(N.e~)
= A' J.t" - A" J.t' (by Eq. (1)), (6)
where in passing from the second line to the third we used the equations
N u·e' = - N .e u'' Nu·"
e = - N"
.e u '
N t/.e ' = - N
.ev'' N " = - N"
v·e .e v '
232 Elementary Differential Geometry 10. Gauss's Theorema Egregium 233

which follow by differentiating N.e ' = 0 = N.e" with respect to u and v. We now compute a and (3. First,
Putting Eqs. (5) and (6) together shows that the right-hand sides of Eqs. (2)
a: = e~.e'l (by Eq. (1))
and (4) are equal. Since Eq. (2) has already been established, this proves Eq.
(4)., 0 = (lOuUu + fUuu).('"'(U u + M v ) (by Eq. (9)
= ffu (fUu).(JUu + oO'v) + flO'uu·O'u + fOO'uu'O'v
Now we can prove Theorem 10.1.
= fUe/.e"
f
+ !fl(O'u.O'u)u + fO«O'u.O'v)u -
2
O'u.O'uv)
Proof 10.1 1
= -flEu + fO (
F u1 )
- -Ev
(.
SInce e.e
I II
= 0) , (10)
2 2
Combining Eqs. (3) and (4), we get
which does indeed depend only on E, F and G (because the same is true for
av - f3u (7) 1,0 and f). And finally,
K = (EG _ F2)1/2 '
(3 = e~.e"
so to prove the theorem it suffices to show that, for a suitable choice of {e / , e"},
the scalars a and f3 depend only on E, F and G. We shall construct {e ' , e"} by
= (fvO'u + fO'uv).(Juu + rSO'v)
applying the Gram-Schmidt process to the basis {Uu,C7 v } of the tangent plane, = lOv I "
-e.e + fIUUT)'UU + fVUuv'U
1>
v
lO
and will then show that they have the desired property. _1 1 1>G
So we first define
= "2fJEv + "2 fV u, (11)

I Uu which also depends only on E, F and G.

e = II U u II = fUu , This completes the proof of Gauss's Theorem. o
where f = E-l/2. Now we look for a vector e" = '"'(uu + r5u v, for some scalars
By substituting the actual values of '"'(, rS and f into these formulas for a and
1,15, such that e" is a unit vector perpendicular to e ' . These conditions give
f3, and then using Eq. (7), we get an explicit formula for K in terms of E, F
E- 1/ 2 ('"'(E + r5F) = 0, '"'(2 E + 21 r5F + 62 G = 1. and G. Here is the result:
The first equation gives I = -6F/ E, and substituting in the second equation
then gives Corollary 10.1
2
2 F F2 ) The gaussian curvature is given by
6 ( E - 2"E + G = 1,
El/2
-tEvv + Fuv - tG uu tEu Fu - tEv 0 tEv tau
Fv - !Gu E F - !Ev E F
15 = (EG _ F2)1/2 ' 1= - (EG _ F2)1/2' (8)
!Gv F G !Gu F G
K=
(We could change the sign of 15, and hence also that of I' but it would make no
difference in the end.) Thus,
We shall not go through the details of this calculation, partly because the
(9)
proof is very tedious, and partly because the following special cases are often
where I' r5 and f depend only on E, F and G. all that is needed:
 234 Elementary Differential Geometry 10. Gauss's Theorema Egregium 235

We are now in a position to give the application of geodesic coordinates

Corollary 10.2 that we promised in Section 8.5.

(i) If F = 0, we have Theorem 10.2

Any point 0/ a surface of constant gaussian curvature is contained in a patch
K= - 2Y~G {:u (y;a) + ~ (v~a)}' that is isometric to part of a plane, a sphere or a pseudosphere.

(ii) If E = 1 and F = 0, we have

Proof 10.2
K I_c)2.j(j
- .;a au2 • Let P be a point of a surface S with constant gaussian curvature K. By applying
a dilation of R 3 (see Exercise 7.6), we need only consider the cases K = 0,1
Proof 10.2
and-I.
If F = 0, Eq. (8) gives We take a geodesic patch u(u, v) with u(O,O) = P. Writing 9 = v'G, the
first fundamental form is
'Y = 0, & = a- 1 / 2 , £ = E- 1 / 2 •
Substituting in Eqs. (lO) and (11), we get
By Corollary lO.2(ii)~
a = _~(EG)-1/2 E v , /3 = ~(EG)-1/2Gu. ()2 g
Hence, 8u 2 + Kg = O. (13)

K - av
-
- /3u _ {a ( E
(EG)1/2 - - 2vEG au vEG
1 v ) a( au )}
+ au vEG ' (12)
Note that
g(O, v) = 1, gu(O, v) = 0, (14)
proving the formula in (i).
by Proposition 8.7.
If, in addition, E = 1, the first term on the right-hand side of (12) vanishes, If K = 0, the solution of Eq. (13) is g(u,v) = au + /3, where a and /3 are
so
smooth functions of v only. The boundary conditions (14) give a = 0, f3 = 1,
K _ 1 8(Gu) _ 1a .j(j
- - 2.;Gau .j(j - - .;G au
2

2 '
so 9 = 1 and the first fundamental form of (T is
du 2 + dv 2 .
proving the formula in (ii). o
This is the same as the first fundamental form of the usual parametrisation of
the plane (see Example 5.1), and Theorem 5.1 now shows that (T is isometric
Example 10.1 to part of the plane.
For the surface of revolution If K = 1, the general solution of Eq. (13) is 9 = acosu + f3sinu, where a
and /3 only depend on v. The boundary conditions (14) give 0' = 1, f3 = 0, and
u( u , v) = (f(u) cos v, feu) sin V, g( u)), the first fundamental form of (T is
where / > 0 and j2 + 92 = 1 (a dot denoting d/du) , we found in Example 6.2
that E = 1, F = 0, G = f(u)2. Hence, Corollary lO.2(ii) applies and gives
This is the first fundamental form of the unit sphere, with u and v being latitude
K _ 1 a2 .j(j _ j and longitude, respectively (see Example 5.2). Hence, (T is isometric to part of
- - .jG {)u 2 - --j'
the unit sphere.
in agreement with Eq. (2) of Chapter 7.
236 Elementary Differential Geometry 10. Gauss's Theorema Egregium 237

Finally, if K = -1, we find in the same way that the first fundamental form 10.3 With the notation of Exercises 8.21 and 10.2, show that
of (1 is
(i) the circumference of the geodesic circle with centre P and ra-
du 2
+ cosh 2
U dv 2
• dius R is
We have not encountered thi's'first fundamental form before. However, let us
reparametrise f7 by defining
OR == 2rrR (1 - K~P) R 2 + remainder) ,

where remainder/R2 tends to zero as R tends to zero;

v = e"tanh u, W = ettsech u. (ii) the area inside the geodesic circle in (i) is
We then find, using the formulas in Exercise 5.4, for example, that the first
fundamental form becomes AR = R (1 - Ki:) R
11"
2 2
+ remainder) ,
dV 2 + dW2
W2 where the remainder satisfies the same condition as in (i).
Comparing with Example 8.8, we see that this is the first fundamental form of Calculate OR and A R exactly when S is the unit sphere, and verify
the pseudosphere. 0 that the results agree with those in (i) and (ii) above.
10.4 Let ABC be a triangle on a surface f7 whose sides are arcs of
geodesics. Assume that ABC is contained in a geodesic patch f7 as in
EXERCISES Exercise 8.21 with P == A. Thus, with the notation in that exercise,
if we take v to be parallel to AB at A, then AB and AC are the
10.1 If a surface patch has first fundamental form e>'(du 2 + dv 2 ), where parameter curves 0 == 0 and 0 == LA, and BC can be parametrised
A is a smooth function of u and v, show that its gaussian curvature by 'Y(8) ;: f7(1(O) , 8) for some smooth function f and 0 ~ 8 ~ LA.
K satisfies
LlA + 2Ke>' ;: 0,
where Ll denotes the laplacian (P18u 2 + fj2I8v 2 •
10.2 With the notation of Exercise 8.21, define u = rcosO, v;: rsinO,
and let o-(u, v) be the corresponding reparametrisation of u. It can
be shown that 0- is an allowable surface patch for S defined on the
open set u 2 + v 2 < f2. (Note that this is not quite obvious because u
is not allowable when r ;: 0.) Show that the first fundamental form
of 0- is Edu 2 + 2Fdudv + Gdv 2 , where (i) Use the geodesic equations (Chapter 8, Eqs. (2» to show that
- u
E == 2'
r
2
Gv
+ -4-'
r
2
~
F = ( 1- G) uv
2"
r
-2'
r
_ v2
G;: - 2
r
Gu 2
+ -r 4-
1" l' AI
~).2-28r'
_ 1 8G

(use Exercise 5.4). Show that u 2 (E - 1) = v 2 (G - 1), and by con- where a dash denotes didO and A = II 'Y' II.
(ii) Show that, if t/J(O) is the angle between (1T and the tangent
sidering the Taylor expansions of E and (; about u ;: v == 0, deduce
that vector to BC at -y(O), then

G(r, 9) = r 2 + kr 4 + remainder 'IjJ'(8) ;: - 8-:; (1(8),8).

for some constant k, where remainder/r4 tends to zero as r tends to (iii) Show that, if K is the gaussian curvature of S,
zero. Show finally that k = -K(P)/3, where K(P) is the value of
the gaussian curvature of Sat P (use Corollary 1O.2(ii». j'lABc
f K dAq ;: LA + LB + LO - rr.
238 Elementary Differential Geometry 10. Gauss's Theorema Egregium 239

This result will be generalised in Chapter 11. Proposition 10.2

The only isometries of a helicoid

10.2. Isometries'of Surfaces O'(u,v) = (ucosv,usinv,v)

are SA' R x 0 SA' R y 0 SA and R z 0 SA for some value of A, where SA is the
The Theorema Egregium provides a necessary condition for the existence of an screwing motion O'(u, v) H O'(u, V + A), and R"" R II and R z are rotations by 1r
isometry between surfaces: if such an isometry exists, the gaussian curvature around the x, y and z-axes.
must be the same at corresponding points of the two surfaces. We give two
examples of this idea; others can be found in the exercises.
Proof 10.2
Our first result shows that it is impossible to draw a 'perfect' map of the
Earth (which is why cartography is an interesting subject). Suppose that an isometry of the helicoid takes 0'( u, v) to 0'( ii, v), where ii and
v are smooth functions of u and v. Since the gaussian curvature at 0'( U, v) is
-1/(1 + u 2 )2 (see Exercise 7.2), the Theorema Egregium tells us that
Proposition 10.1
-1 -1
Any map of any region of the earth's surface must distort distances. (1 + u 2 )2 = (1 + ii,2)2 '
so ii, = ±u. Applyi!1g a rotation R z by 11' around the z-axis changes u to -u
Proof 10.1 (and fixes v), so we assume that ii = u. Let v = f(u, v). By Theorem 5.1,
A map of a region of the earth's surface which did not distort distances would the patches O'(u l v) and u(u, v) = O'(u, f(u, v)) have the same first fundamental
be a diffeomorphism from this region of a sphere to a region in a plane (the form. That of 0' is du 2 + (1 + u 2)dv 2, and that of u is found to be
map) which multiplied all distances by the same constant factor, say C. We (1 + (1 + u2)f~)du2 + 2(1 + 11,'l)!U!11dudv + (1 + u 2)f;dv 2.
~:- :- .
might as well assume that the plane passes through the origin. Then, by com-
posing this map with the map from the plane to itself which takes a point
Equating these, we find that fu = 0 and f11 = ±1. Hence,
with position vector r to the point with position vector C-1r, we would get v = f(u,v) = ±v + A,
an isometry between this region of the sphere and some region of a plane. This
where A is a constant. A rotation R", by 1r around the x-axis changes v to -v
would imply, by the Theorema Egregium, that these regions of the sphere and
(and fixes u), so we take the + sign. This gives the isometry
the plane have the same gaussian curvature. But we know that a plane has
gaussian curvature zero everywhere, and a sphere has constant positive gaus- SA: O'(U, v) H O'(u,v + A).
sian curvature everywhere (if the sphere has radius R, the gaussian curvature
is 1/ R 2 ). So no such isometry can exist. 0
This proves the proposition (the isometry Ry 0 SA arises because R y =R x 0 R z .)
o
Note, on the other hand, that it is possible to draw a map of the Earth that
correctly represents angles, for we saw in Example 5.7 and Exercise 5.11 that
EXERCISES
the stereographic and Mercator projections are conformal, and Archimedes's
Theorem 5.4 shows that there is a map that correctly represents areas.
10.5 Show that there is no isometry between any region of a sphere and
Our next example shows how the Theorema Egregium can sometimes be
any region of a (generalised) cylinder or a (generalised) cone. (Use
used to determine all the isometries of a surface.
Proposition 10.1 and Exercise 5.7.)
10.6 Show that the gaussian curvature of the Mobius band in Example
4.9 is equal to -1/4 everywhere along its median circle. Deduce
that this Mobius band cannot be constructed by taking a strip of
 f-""':":""':--:-:~~---------
Elementary Differential Geometry
! 10. Gauss's Theorema Egregium 241
240

paper and joining the ends together with a half-twist. (The analytic
The six r coefficients in these formulas are called Christoffel symbols.
description of the 'cut and paste' Mobius band is more complicated
Proof 10.3
than the version in Example 4.9.)
10.7 Consider the surf~ce patches Since {(7 u, (1'11' N} is a basis of R 3 1 scalar functions 01, '.' . 11'3 satisfying

tr(u,v) = (ucosv,usinv,lnu), u(u,v) = (ucosv,usinv,v). O'uu = O'IO'u+ Ct20'v + (}:a N ,

Uuv = {3,uu + {32uv + {33N, (15)
Prove that the gaussian curvature of tr at tr(u, v) is the same as that
of iT at iT(u, v), but that the map from tr to U which takes tr(u, v) to tT vv = "rlau + 'Y2t1v + 1'a N ,
u(u, v) is not an isometry. Prove that, in fact, there is no isometry certainly exist. Taking the dot product of each equation with N gives
from is to a-. a3 = L, {33 = M, ,3 = N.
10.8 Show that the only isometries from the catenoid to itself are products Now we take the dot product of each equation in (15) with Uu and U v ' This gives
of rotations around its axis, reflections in planes containing the axis, six scalar equations from which we determine the remaining six coefficients. For
and reflection in the plane containing the waist of the catenoid. example, taking the dot product of the first equation in (15) with Uu and U v
gives the two equations
1
E0. 1 +Fa:2 = tl uu ·t1 u = 2'E Ul

10.3. The Codazzi-Mainardi Equations 1

Fa! + Ca2 = tluu.tT v = (tlu.tlv)u - tlu·tl uv =. F u - '2Ev.
Gauss's Theorema Egregium shows that the coefficients of the first and second
Solving these equations gives a, = rl" a2 = r;,,; similarly for the other four
fundamental forms of a surface cannot be arbitrary smooth functions, for it
coefficients in Eqs. (15). 0
shows that LN - M 2 can be expressed in terms of E, F and G. It is natural to
ask if there are any further relations between these coefficients. In this section,
The new relations between the coefficients of the first and second funda-
we find that there are indeed some additional relations, and we show that, in
mental forms of a surface patch are contained in the following result.
a sense we shall explain, there are no others.
We begin with a computation similar to that in Lemma 10.1.
Proposition 10.4 (Codazzi-Mainardi Equations)
Define the Christoffel symbols of a surface patch u(u, v) as above. Then,
Proposition 10.3 (Gauss Equations)
Lv - M u = Lrl2 + M(r;'2 - rA) - Nr;",
Let tr( u, v) be a surface patch. Then,
M v - N u = Lri2 + M(ri2 - rl2) - Nr;'2'
a uu = rila u + r{lt7 v + LN,
Proof 10·4
tl uv ri2l1u + r't2tTV + MN,
=
tT vv = ri2 D'u + r;2t1'v + NN,
We write down the equation (uuu)v = (uuv)u, using the Gauss equations for
tl uu and tT uv :
where (rl,tr u + r;',uv+LN)v = (rl2u u + r;'2trv + MN)u,
r' _ GEu 2FFu + FE v r 2 _
-
2EFu - EEv - FEu
11 - 2(EG _ F2) 11 - 2(EG _ F2) arl , _ arl2 ) u u + (ar;" _ aT;'2) u v + (L _ M )N
( av 8u 8v Bu v U
, GEv - FG u 2 EGu-FEv
r'2 = 2(EG _ F2)' r'2 = 2(EG _ F2)' = rl2u uu + (r;'2 - rl,)uuv - r;"uvv - LNv + MNu
r' _ 2GFv - GG u - FG u 2 _ r EG v - 2FFv + FG u = rl2(rl,uu+r;"uv + LN) + (r;'2 - rl,)(rl2u u + r;'2Uv + MN)
22 - 2(EG _ F2) , 22 - 2(EG _ F2)
- r;" (ri2Uu + ri2Uv + NN) - LN v + MNu , (16)
242 Elementary Differential Geometry +
,
10. Gauss's Theorema Egregium 243

using the Gauss equations again. Now, N u and N v are perpendicular to N, all the Christoffel symbols are zero and the Codazzi-Mainardi equations are
and so are linear combinations of IT. and lTv. Hence, equating N components obviously satisfied. The formula in Corollary 10.1 gives K = 0, so the only
on both sides of the last equation gives other condition to be checked is LN - M2 = 0, and this clearly holds since
M = N = O. Theorem 10.3 therefore tells us that a surface patch with the
Lv - M. =LFI2 + M(Fl2 - F,!,) - NFl" given first and second fundamental forms exists.
which is the first of the Codazzi-Mainardi equations. The other equation follows To find it, we note that the Gauss equations give
in a similar way from (D'uv)tI = (uvv)u' 0
tTuu = -N, (1uv = 0, 6 vtJ = o.
At first sight, it seems that we could get four other identities like those The last two equations tell us that lTv is a constant vector, say a, so
in Theorem 10.2 by equating the coefficients of IT. and lTv in Eq. (16) and in IT(u, v) = btu) + av, (17)
its analogue coming from (IT. v ). = (IT. v ).' It turns out, however, that these
identites are all equivalent to the formula in Corollary 10.1 (and so, in partic- where b is a function of 1.1 only. The first equation then gives N = - b" (a dash
ular, they give another proof of the Theorema Egregium). In fact, there are no denoting d/dv). We now need to use the expressions for N. and N v in terms
further identites to be discovered, as the following theorem shows. of IT. and lTv in Proposition 6.4. The Weingarten matrix is

Theorem 10.3
W = :F1 ':FIl = (~ 0)-' (-I 0) = (-1 0)
1 0 0 0 0 '

Let IT : U ....; R 3 and ir : U ....; R 3 be surface patches with the same first and so Proposition 6.4. gives
second fundamental forms. Then, there is a rigid motion M of R 3 such that Nu = tTU1 Nt) = O.
ir=MolT.
The second equation tells us nothing new, since we already knew that N = - b"
Moreover, let V be an open subset of R 3 and let E,F,G,L,M and N be
depends only on u. The first equation gives
smooth functions on V. Assume that E > 0, G > 0, EG - F 2 > 0 and that the
equations in Corollary 10.1 and Proposition 10.4 hold, with K = 11& 'I: and b'" + b' = O.
the Christoffel symbols defined as in Proposition 10.3. Then, if (uo,vo) E V,
Hence, b" + b is a constant vector, which we can take to be zero by applying
there is an open set U contained in V and containing (1.10, vo), and a surface
a translation to IT (see Eq. (17)). Then,
patch IT : U ....; R 3 , such that Edu 2 +2Fdudv+Gdv 2 and Ldu2 +2Mdudv+Ndv 2
are the first and second fundamental forms of IT, respectively. btu) = ccosu + dsinu,
where c and d are constant vectors, and N = -b" = b. This must be a unit
This theorem is the analogue for surfaces of Theorem 2.3, which shows that
vector for all values of u. It is easy to see that this is possible only if c and d
unit-speed plane curves are determined up to a rigid motion by their signed
are perpendicular unit vectors, in which case we can arrange that c = (1,0,0)
curvature. We shall not prove Theorem 10.3 here. The first part depends on
and d = (0,1,0) by applying a rigid motion, giving btu) = (cosu,sinu,O).
uniqueness theorems for the solution of systems of ordinary differential equa-
Finally, IT u x lTv = AN for some non-zero scalar A, so b' x a = Ab. This forces
tions, and is not particularly difficult. The second part is more sophisticated
a = (0,0, A), and the patch is given by
and depends on existence theorems for the solution of certain partial differential
equations. The following example will illustrate what is involved. IT(u,v) = (cosu,sinu,Av),
a parametrisation of a circular cylinder of radius 1 (which the reader had proh-
Example 10.2 ably guessed some time ago).
Consider the first and second fundamental forms du 2 + dv 2 and -du 2 , respec-
tively. Let us first see whether a surface patch with these first and second
fundamental forms exists. Since all the coefficients of these forms are constant,
244 Elementary Differential Geometry + 10. Gauss's Theorema Egregium 245

EXERCISES Lemma 10.2

Let u : U -7R 3 be a surface patch containing a point P = U(uo, vol that is
10.9 A surface patch has first and second fundamental forms not an umbilic. Let"'1 ;::: "'2 be the principal curvatures of a and suppose that
z
cos 2 v diP- + dv 2 and - cos2 v du - dv 2 , Itl has a local maximum at P and 1t2 has a local minimum there. Then, the
gaussian curvature of U at P is :0; O.
respectively. Show that the surface is part of a sphere of radius one.
(Compute the Weingarten matrix.) Write down a parametrisation
of the unit sphere with these first and second fundamental forms. Proof 10.2
10.10 Show that there is no surface patch whose first and second funda- Since P is not an umbilic, It! > K2 at P 1 SO by shrinking U if necessary, we
mental forms are may assume that 1\.1 > "-2 everywhere.
By Proposition 7.2, we can assume that the first and second fundamental
du 2 + cos2 U dv 2 and cos2 u du 2 + dv 2 ,
forms of u are
respectively.
Edu 2 + Gdv 2 and Ldu2 + N dv 2 ,
10.11 Suppose that the first and second fundamental forms of a surface
patch are Edu 2 + Gdv 2 and Ldu2 + N dv 2 , respectively (cf. Propo- respectively. By Exercise 10.8,
sition 7.2). Show that the Codazzi-Mainardi equations reduce to 2E 2G
Ev = (ltl)v, G.. = (1t2) .. ,

21Ev (LE + N) 21 G.. (LE + N)

K1 - K2 K1 - "-2
Lv = G ' N.. = G . and by Corollary 1O.2(ii), the gaussian curvature
Deduce that the principal curvatures Itl = LjE and
satisfy the equations
1t2 = NjG
K = - 2V~G (:u (vc;G) + :v (v~G)) .
= :~(1t2 -Itl), = ~G(ltl -
Since P is a stationary point of 1<1 and 1t2, we have (1t1)V = (1t2) .. = 0, and
(ltl)v (1t2) .. 1t 2).
hence E v = G.. = 0, at P. Hence, at P,
1 1 ( 2G 2E)
K = -2EG(G.... + Evv ) = -2EG "'1 - K,2
(1t2) .... -
Itl - K2
(ltdvv

10.4. Compact Surfaces of Constant Gaussian (again dropping terms involving E v , G. and the first derivatives of 1<1 and 1t2).
Curvature Since Itl has a local maximum at P, (I<dvv :0; 0 there, and since 1t2 has a local
minimum at P, (1t2) •• 2': 0 there. Hence, the last equation shows that K :0; 0
We conclude this chapter with a beautiful theorem that is the analogue for ~P. 0
surfaces of the characterisation given in Example 2.2 of circles as the plane
curves with constant curvature. Proof 10.4
a
The proof of Theorem 10.4 will use little point set topology. We consider the
Theorem 10.4 continuous function on the surface S given by J = (It 1 - 1t2)2, where 1<1 and 1<2
are the principal curvatures. Note that this function is well defined even though
Every compact surface whose gaussian curvature is constant is a sphere.
K,1 and "'2 are not, partly because we do not know which principal curvature
is to be called 1<1 and which 1t2, and partly because the sign of the principal
Note that, by Proposition 7.6, the value of the constant gaussian curvature
curvatures depends on the choice of parametrisation of S. We shall prove that
in this theorem must be > o.
this function is identically zero on S, so that every point of S is an umbilic.
The proof of this theorem depends on the following lemma.
Since the gaussian curvature K > 0, it follows from Proposition 6.5 that S is
part of a sphere, say S. In fact, S must be the whole of S. For, any point P of
246 Elementary Differential Geometry +
S is contained in a patch a : U ---> R 3 of S, and a(U) = S n W, where W is an
open subset of R 3 ; it follows that S is an open subset of S. On the other hand,
since S is compact, it is necessarily a closed subset of R 3 I and hence a closed
11
subset of S. But since S is c,?nnected, the only non-empty subset of S that is The Gauss-Bonnet Theorem
both open and closed is S itseif.
Suppose then, to get a contradiction, that J is not identically zero on S.
Since S is compact, J must attain its maximum value at some point P of S, and
this maximum value is > O. Choose a patch t1 : U ---> R 3 of S containing P, and
let It,
and 1t2 be its principal curvatures. Since 1t,lt2 > 0, by reparametrising
if necessary, we can assume that It,and 1t2 are both> 0 (see Exercise 6.17).
Suppose that It,> 1t2 at Pi then by shrinking U if necessary, we can assume
that It,
> 1t2 everywhere on U. Since K is a constant> 0, the function (x _ ~) 2
increases with x provided that x > K / x > O. Since It, It,
> K / = 1t2 > 0, this
function is increasing at x = It" so It,
must have a local maximum at P, and
then 1t2 = K / It, must have a local minimum there. By Lemma 10.2, K :::; 0 at
P. This contradicts the assumption that K > O. 0

The Gauss-Bonnet theorem is the most beautiful and profound result in the
EXERCISES theory of surfaces. Its most important version relates the average over a surface
of its gaussian curvature to a property of the surface called its 'Euler number'
10.12 Show that a compact surface with gaussian curvature> 0 every- which is 'topological', i.e. it is unchanged by any continuous deformation of
where and constant mean curvature is a sphere. (As in the proof the surface. Such deformations will in general change the value of the gaussian
of Theorem 10.4, ifIt, has a local maximum at a point P of the curvature, but the theorem says that its average over the surface does not
surface, then 1t2 = 2H - It,
has a local minimum there.) change. The real importance of the Gauss-Bonnet theorem is as a prototype
of analogous results which apply in higher dimensional situations, and which
relate geometrical properties to topological ones. The study of such relations is
one of the most important themes of 20th century Mathematics.
J

11.1. Gauss-Bonnet for Simple Closed Curves

The simplest version of the Gauss-Bonnet Theorem involves simple closed
curves on a surface. In the special case when the surface is a plane, these
curves have been discussed in Section 3.1. For a general surface, we make

Definition 11.1
A curve ,),(t) = a(u(t), vet)) on a surface patch t1 : U ---> R 3 is called a simple
closed curve with period a if 1I'(t) = (u(t), vet)) is a simple closed curve in
R 2 with period a such that the region int(1I') of R 2 enclosed by 11' is entirely
247
248 Elementary Differential Geometry +
I 11. The Gauss-Bonnet Theorem 249

contained in U (see the diagrams below). The curve 'Y is said to be positively- integral:
oriented if 1r is positively-oriented. Finally, the image of int(1r) under the map (Wf)
q is defined to be the interior inth) of "Y. I= J e/,e"ds
o
{Wy}
= 1 e'.(e~u + e~v)ds

L
0

= (e/.e~)du + (e',e~)dv,
By Green's theorem (see Section 3.1), this can be rewritten as a double integral:

i I = I·{
Jint(1r)
{(e/.e~)u - (e/,e~)1J}dudv

= f· r
Jint('1r)
{(e~.e~) - (e~.e~)}dudv

=!{ Jint('1r}
LN - M2/2dudV
(EG - F2)1
(by Lemma 10.1)

= f· r
Jirlt('1r)
LN - M: (EG - F 2 )1/2dudv
EG - F
allowed not allowed
We can now state the first version of the Gauss-Bonnet Theorem.
= I·(
Jint('1r)
KdAq. (1)

Now let O( s) be the angle between the unit tangent vector 1 of'Y at ,,( 8) and
Theorem 11.1 the unit vector e l at the same point. More precisely, (J is the angle, uniquely
determined up to a multiple of 211", such that
Let 'Y(8) be a unit-speed simple closed curve on a surface (1 of length l('Y), and
assume that 'Y is positively-oriented. Then, 1 = cos Be' + sin (Je" . (2)

1 o
£("Y) Kgds = 211" - 11
int('Y)
KdAq,
Then,
N x 'Y = - sin (Je ' + cos (Je'l. (3)
where Kg is the geodesic curvature of "Y, K is the gaussian curvature of tr and
dAq = (EG - F2)1/2dudv is the area element on q (see Section 5.4). N

We use s to denote the parameter of'Y to emphasize that 'Y is unit-speed.

Proof 11.1
As in the proof of Theorem 10.1, choose a smooth orthonormal basis {e', e"} of
the tangent plane of tr at each point such that {e', e", N} is a right-handed or-
thonormal basis of R 3 , where N is the unit normal to tr. Consider the following
 Elementary Differential Geometry 11. The Gauss-Bonnet Theorem 251

Now, by Eq. (2), Note, however, that it is crucial that the interior of 11" is entirely contained in
U, otherwise such a family will not exist, in general:
i' = cos (}e
l
+ sin (}i/' + 8( - sin Be' + cos Be"), (4)
so by Eqs. (3) and (4) the geodesic curvature of"Y is
/'\,9 = (N x 1').'5' (see S~ction 6.2)
= 8(- sin Be' + cos Be"). (- sin eel + cos (je")
+ (- sin 8e' + cos (Jell). (cos Oe' + sin (jell)
=iJ + cos 2
8( e .e") - sin 2 B( e".e' )
l

+ sinO cosO(elJ.e" - e/.e /) (by Eqs. (2) and (3)).

l
Since e and e" are perpendicular unit vectors,

e'.e' = e".e" = 0, e'.e" = -e'.e ll

.

Hence,

/'\,g
I'll ,
= O· - e.e
and by the definition of I, Observe next that the integral I:('Y~) 0 ds should depend continuously on T.
Further, since 'YT and e l return to their original values as one goes once round
rl('*Y)
'}'T, the integral is always an integer multiple of 211". These two facts imply
I =1 (8 - /'\,g)ds.
0 that the integral must be independent of T - for by the Intermediate Value
Thus, to complete the proof of Theorem 11.1, we must show that Theorem a continuous variable cannot change from one integer to a different
ri("Y) integer without passing through some non-integer value. To compute Iol('Y) iJ ds,
10 Ods=2rr. (5) we can therefore replace "Y by any other simple closed curve i in the interior
of '}', since this will not change the value of the integral. We take"" to be the
Equation (5) is called 'Hopf's Umlaufsatz' - literally 'rotation theorem' in image under t1 of a small circle in the interior of 11". It is 'intutively clear' that
German. We cannot give a fully satisfactory proof of it here because the proof
l (1')
would take us too far into the realm of topology. Instead, we shall justify Eq.
(5) by means of the following heuristic argument.
The main observation is that, if it is any other simple closed curve contained because
lo Ods = 2rr,

in the interior of 'Y, there is a smooth family of simple closed curves "Y T , defined (i) e/ is essentially constant at all points of"" (because the circle is very small),
for 0 :$ T :$ 1, say, with '}'O = 'Y and 'Y 1 = i (see Section 8.4 for the notion of and
a smooth family of curves). The existence of such a family is supposed to be
'intuitively ohvious'. (ii) the tangent vector to l' rotates by 27f on going once round 'Y because the
interior of l' can be considered to be essentially part of a plane, and it is
'intutively clear' that the tangent vector of a simple closed curve in the
plane rotates by 2rr on going once round the curve.
This completes the 'proof' of HopPs Umlaufsatz, and hence that of Theorem
11.1.
252 Elementary Differential Geometry
+ 11. The Gauss-Bonnet Theorem 253

EXERCISES

11.1 A surface patch t1 has gaussian curvature :$ 0 everywhere. Prove that

there are no simpl~closed geodesics in iT. How do you reconcile this
with the fact that the parallels of a circular cylinder are geodesics? "r-
11.2 Let ')'(8) be a simple closed curve in R 2, parametrised by arc-length
and of total length i(,),). Deduce from Hopf's Umlaufsatz that, if

( ')')
11: 8 (8) is the signed curvature of ')', then

1o It.(s)ds:::: 27f.

(Use Proposition 2.2.)

Now let iT : U -t R 3 be a surface patch and let 1r : R -t U be a curvilinear

11.2. Gauss-Bonnet for Curvilinear Polygons
polygon in U, as in Definition 11.2. Then, 'Y = u 0 1r is called a curvilinear
polygon on the ,.Surface patch tT, int('Y) is the image under iT of int(n"), the
For the next version of Gauss-Bonnet, we shall have to generalise our notion
vertices of'Y are the points "Y(ti) for i = 1, ... , n, and the edges of iT are the
of a curve by allowing the possibility of 'corners'. More precisely, we make the
segments of it corresponding to the open intervals (ti-1, ti)' Since u is allowable,
following definition.
the one-sided derivatives

Definition 11.2
A curvilinear polygon in R 2 is a continuous map 1r : R -t R 2 such that, for exist and are not parallel.
some real number a and some points 0 = to < t1 < ... < t n = a, Let Or be the angles between -y±(ti) and e/, defined as in Eq. (2), let
(i) 1r(t):::: 1r(t/) if and only if t' - t is an integer multiple of a; 8i = 8t - 0i be the external angle at the vertex ')'(ti), and let ai = 7f - 8i be
(ii) 1r is smooth on each of the open intervals (to, t1), (h, t2), ... , (tn-I, tn); the internal angle. Since the tangent vectors -y+ (ti) and -y- (ti) are not parallel,
(iii) the one-sided derivatives the angle r5i is not a multiple of 7f. Note that all of these angles are well defined
only up to multiples of 27f. We assume from now on that 0 < ai < 27f for
?i'-(td :::: lim 1r(t) -1r(t i ), ?i'+(ti):::: lim 1r(t) -1r(ti) (6)
ttt, t - ti tJ,t, t - ti i = 1, ... ,n.
A curvilinear polygon 'Y is said to be unit-speed if II -y II = 1 whenever -y is
exist for i = 1, ... ,n and are non-zero and not parallel.
defined, i.e. for all t such that 'Y(t) is not a vertex of 'Y. We denote the parameter
The points 'Y( ti) for i :::: 1, ... , n are called the vertices of the curvilinear polygon
of 'Y by s if 'Y is unit-speed. The period of 'Y is then equal to its length l(-y),
'Jf, and the segments of it corresponding to the open intervals (ti-1 , t i ) are called
which is the sum of the lengths of the edges of 'Y.
its edges.

It makes sense to say that a curvilinear polygon 'JI' is positively-oriented: Theorem 11.2
for all t such that 'Jf(t) is not a vertex, the vector n. obtained by rotating ?i'
Let 'Y be a positively-oriented unit-speed curvilinear polygon with n edges on a
anticlockwise by 7f/2 should point into int('Jf). (The region int('Jf) enclosed by 'Jf
surface u, and let aI, a2, . .. , an be the interior angles at its vertices. Then,
makes sense because the Jordan Curve Theorem applies to curvilinear polygons
("Y) !r
1
in the plane.) n
II: g dS=L a i-(n-2)7f- Kdk-
o i=l Jint('Y)
254 Elementary DifFerential Geometry
..,..----------------
11, The Gauss-Bonnet Theorem 255

Proof 11.2

Exactly the same argument as in the proof of Theorem 11.1 shows that
i ('Y)
~gds~'
," l i ('Y)
J[ , !

l
o 0
Ods -
Jint('Y)
KdAo-.

We shall prove that

(7) ---_ ... -

Assuming this, we get

(('Y) n
Ino Kgds = 2'11" - LIS,
,=1
r KdAa
- /'Jint('Y)
Le. 'Y and i agree except when s belongs to a small interval (s~,s~'), say, con-

= 2'11" - t (-11' - ct,) - /' ~ K dAo-

taining Si, so the contribution from the ith vertex is

1~~1 r~i Ods - 1~~' Ods.

,=1 Jint('Y)
n
Ods -
= L Q , - (n - 2)'11" - /' ~ KdAo-.
S, Jsi. 81.

,=1 Jint('Y) The first integr81 is the angle between ~(sn and ~(si), which as si and si'
To establish Eq. (7), we imagine 'smoothing' each vertex of'Y as shown in tend to s, becomes the angle between -Y+(Si) and -y-(Si), Le. h On the other
the following diagram. hand, since 'Y(s) is smooth on each of the intervals (s~, Si) and (8i, the last sn,
two integrals go to zero as si and si' tend to 8i. Thus, the contribution to the
expression (9) from the ith vertex tends to c5i as si and sil tend to Si . Summing
over all the vertices, we get
i ('Y) . l i ("1)
L lSi.
n

l o
Ods -
0
{Jds =
i=1
o
Equation (7) now follows from this and Eq. (8).
, '

Corollary 11.1
If'Y is a curvilinear polygon with n edges each of which is an arc of a geodesic,
then the internal angles aI, a2, ... , an of the polygon satisfy the equation
If the 'smoothed' curve ;Y is smooth (l), then, in an obvious notation,
(('1) . t a i = (n - 2)'11" +
i=1
/1 mt('Y)
KdAo-.
J Ods = 211". (8)
o
Since 'Y and ;Y are the same except near the vertices of 'Y, the difference Proof 11.1
£('1) ;., 1£('Y) . This is immediate from Theorem 11.2, since ~g = 0 along a geodesic. 0
Ods - Ods (9)
1o 0
As a special case of Corollary 11.1, consider an n-gon in the plane with
is a sum of n contributions, one from near each vertex. Near 'Y(s,), the picture
is
256 Elementary Differential Geometry t 11. The Gauss-Bonnet Theorem 257

straight edges. Since K = 0 for the plane, Corollary 11.1 gives Finally, for a geodesic n-gon on the pseudosphere (see Section 7.2), for which
n K = -1, we see that E ai is less than (n - 2) 11" by the area of the polygon. In
L>l!i = (n - 2)11", particular, for a geodesic triangle ABC on the pseudosphere,
~=l
A{ABC) = 11' - LA - LB - LC.
a well known result of element~ry geometry.

EXERCISES

11.3 Suppose that the gaussian curvature K of a surface patch (f satisfies

. For a curvilinear n-gon on the unit sphere whose sides are arcs of great
K :::; -1 everywhere and that 'Y is a curvilinear n-gon on (f whose
cIrcles, we have K = 1 so E ai exceeds the plane value (n - 2)11' by the area
sides are geodesics. Show that n ;::: 3, and that, if n = 3, the area
II dAo- of the polygon. Taking n == 3, we get for a spherical triangle ABC enclosed by 'Y cannot exceed 11'.
whose edges are arcs of great circles,
11.4 Consider the surface of revolution
A(ABC) = LA + LB + LC - 1r.
(f(U, v) = (f(u) cos v, f(u) sin v, g(u)) ,
This is just Theorem 5.5, which is therefore a special case of Gauss-Bonnet.
where 'Y(u) = (f(u),O,g(u)) is a unit-speed curve in the xz-plane.
Let Ul < U2 be constants, let 'Yl and 'Y2 be the two parallels u = Ul
and u = U2 on (f, and let R be the region of the uv-plane given by

Compute
("'(1)
o Itgds,
1(('Y2) Itgds and
J{J KdAo-,
1 0
R
and explain your result on the basis of the Gauss-Bonnet theorem.
 11. The Gauss-Bonnet Theorem
258 Elementary Differential Geometry

The version of the Gauss-Bonnet theorem we are aiming for is obtained by

11.3. Gauss-Bonnet for Compact Surfaces covering a compact surface S with curvilinear polygons that fit together nicely,
applying Theorem 11.2 to each one, and adding up the results. We begin to
The most important version of the Gauss-Bonnet Theorem applies to a compact
surface S. It is a surprising' result that there are very few compact surfaces in make this precise with
R 3 up to diffeomorphism, and they can all be described explicitly. The simplest
example is, of course, the sphere. The next simplest is the torus, which can be Definition 11.3
obtained by rotating around the z-axis a circle in the xz-plane which does not 3
Let S be a surface, with atlas consisting of the patches l1i : Ui ---+ R • A
intersect the z-axis: triangulation of S is a collection of curvilinear polygons, (the interior of) each
of which is contained in one of the (1i(Ui) , such that
(i) every point of S is in at least one of the curvilinear polygons;
(ii) two curvilinear polygons are either disjoint, or their intersection is a com-
mon edge or a common vertex;
(iii) each edge is an edge of exactly two polygons.

Thus, situations like

One can also join such tori together:

are not allowed. A triangulation of the unit sphere with eight polygons is ob-
tained by intersecting the sphere with the three coordinate planes:
z

,\-
This surface is denoted by T g , where 9 is the number of holes, called the genus
of the surface (we take 9 = 0 for the sphere). We accept the following theorem
without proof:

Theorem 11.3
For any integer 9 ~ 0, T g can be given an atlas making it a smooth surface.
Moreover, every compact surface is diffeomorphic to one of the T g • We state without proof:
260 Elementary Differential Geometry f--~i~i:. ~r~h!e~G~a~u~ss~-~B~8n~n~e~t !i~H~@6~F~@M!!:====::::::::::==~~~~~~~~--

Theorem 11.4 where X is the Euler number of the triangulation.

Every compact surface has a triangulation with finitely many polygons. We need to explain what is meant by the left-hand side of the equation
in Theorem 11.5. Fix a triangulation of S with polygons Pi, say. Each Pi is
We introduce the following number associated to any triangulation: 3
\
contained in the image of some patch (f i : Ui -t R in the atlas of S, say
Pi = (fi(R i ), where R i ~ Ui. Then, by definition,
Definition 11.4
The Euler number X of a triangulation of a compact surface S is
X=V-E+F, where K is the gaussian curvature of (fi. Unfortunately, we have to show that
where this is a good definition, Le. that it does not depend on our choice of patches
v= the total number of vertices of the triangulation, (fi (since, even for a given triangulation, Pi may be contained in more than one
patch), nor does it depend on the triangulation itself.
E = the total number of edges of the triangulation, To see this, we note first that, if Ui : Ui -t R 3 is a reparametrisation of (fi
F = the total number of polygons of the triangulation. and if Pi = O'i(R i ), where Ri ~ Ui, then

For the triangulation of the sphere given above, V = 6, E = 12 and F = 8,

so X = 6 - 12 + 8 = 2. The importance of the Euler number is that, although
because both the area element dAtr and the gaussian curvature K are un-
different triangulations of a given surface will in general have different numbers
changed by reparametrisation (see Proposition 5.3 and Exercise 6.17).
of vertices, edges and polygons, X is actually independent of the triangulation
Next, if {Pi} and {PH are two triangulations of S, it is intuitively clear
and depends only on the surface. For example, we can get another triangulation
that we can find a third triangulation {Pk} of S such that each Pi is the union
of the sphere by 'inflating' a regular tetrahedron:
of some of the Pk', as is each Pi- For example, if some ~ and some Pj overlap
as follows,

p~
This time, V = 4,' E = 6 and F = 4, so X = 4 - 6 + 4 = 2, the same as J

before. This property of X is a consequence of the following theorem, which is

the desired extension of Theorem 11.2 to compact surfaces:

Theorem 11.5
Let S be a compact surface. Then, for any triangulation of S, then by inserting additional vertices we can create the appropriate polygons
f1KdA = 2rrx, P"·
k •
 262 Elementary Differential Geometry ¥
I 11. The Gauss-Bonnet Theorem 263
I
I

where ni is the number of vertices of Pi, 'Yi is the curvilinear polygon that
forms the boundary of Pi, l('Yi) is its length, and L i is the sum of its interior
angles. We must therefore sum the contributions of each of the three terms on
the right-hand side of Eq. (10) over all the polygons Pi in the triangulation.
First, Ei L i is the sum of all the internal angles of all the polygons. At
each vertex, several polygons meet, but the sum of the angles at the vertex is
obviously 21T, so
(11)

where V is the total number of vertices.

It is then clear that (in an obvious notation)

since both sides are equal to

L/LIIKdAa~.
k Jo

This is just because the integral of K over the union of finitely many polygons, Next,
all contained in a single surface patch and any two of which are either disjoint
or intersect only in a common edge or vertex, is simply the sum of the integrals
of K over each polygon.
~(ni - 2)11':::: (~ni) 11' - 211'F:::: 211'E - 211'F, (12)

Together with the fact that lIs KdA is independent of the triangulation, where F is the total number of polygons and E the total number of edges, since
Theorem 11.5 thus implies in the sum Li ni each edge is counted twice (as each edge is an edge of exactly
two polygons).
Corollary 11.2
The Euler number X of a triangulation of a compact surface S depends only on
S and not on the choice of triangulation.

We now give the proof of Theorem 11.5.

Proof 11.5

As above, we fix a triangulation of S with polygons Pi, say, each of which is

contained in the image of some patch t1 i : Ui -t R 3 in the atlas of S, say
Pi :::: t1i(Ri), where R i ~ Ui . By Theorem 11.2,

J JrRi KdAa. :::: L i - (ni - 2)1T +

l
0
l
('Y;)
"'9 ds, (10)
Finally, we claim that
l ('Yi)
Ll i 0
K.gds = O. (13)
264 Elementary Differential Geometry f I
11. The Gauss-Bonnet Theorem 265

Indeed, note that in the sum in Eq. (13), we integrate twice along each edge, lIs KdA = 41T. (More generally, this discussion shows that the Euler number
once in each direction. By Eq. (5) in Chapter 6, "'9 changes sign when we of any compact surface is unchanged when the surface is deformed without
traverse a given curve in the opposite direction, so these two integrals cancel tearing.)
out. The various contributiolls to the sum in Eq. (13) therefore cancel out in
pairs, thus proving Eq. (13). We complete the picture by determining the Euler numbers of all .the com-
Putting Eqs. (10), (11), (12) and (13) together gives pact surfaces.

JIs KdA
. JL.
= l;.
'
KdAq,

[[("til
Theorem 11.6
The Euler number of the compact surface T g of genus g is 2 - 2g.
= ~Li - ~(ni -2)rr+ ~ 1 "'gds
•• • 0
= 2rrV - (211B - 211P) +0 Proof 11.6
= 2rrx, The formula is correct when 9 = 0, since we know that X ::;:; 2 for a sphere.
proving Theorem 11.5. o We now prove it for the torus T 1 • To find a triangulation of the torus, we use
the fact that it can be obtained from a square in the plane by gluing opposite
To see why Theorem 11.5 is so remarkable, let us apply it to the unit sphere edges:
52. Then, X = 2 so we get

JIs~ KdA = 41T. (14)

Of course, this result is not remarkable at all because K = 1 so the left-hand

side of Eq. (14) is just the area of the sphere. But now suppose that we deform
the sphere, i.e. we think of the sphere as being a rubber sheet and we pull and /1\ 1[\
stretch it in any way we like, but without tearing:

~~

We subdivide the square into triangles as shown:

For such a deformed sphere S, K will not be constant and the direct computa-
tion of the integral IIsK dA will be difficult. But if we start with a triangulation
of the undeformed sphere, then after deformation we shall have a triangulation
of the deformed sphere with the same number of vertices, edges and polygons
as the original triangulation. It follows that the Euler number of the deformed
sphere is the same as that of the undeformed sphere, i.e. 2, so by Theorem 11.5,
Lt>t> t.lementary UlTlerentlal l.Jeometry 1 11, The Gauss-Bonnet Theorem 267

This leads to a triangulation of T 1 with V = 9, E = 27 and F::;:; 18. One must Suppose we carry out the gluing by removing a curvilinear n-gon from T g and
count carefully: for example, the four circled vertices of the square correspond T1 and gluing corresponding edges (having fixed suitable triangulations of Tg
to a single vertex on the torus. Note also that not just any subdivision of the and T 1 ). If V', E ' and pi are the numbers of vertices, edges and polygons in
square into triangles is acceptable. For ~xample, the subdivision the triangulation of T g , and V", E" and pI! those for T 1 , the numbers V, E
and F for T g + 1 are given by
V ::;:; Vi - n + VII - n + n = Vi + V" - n,
E = E' - n + EI! - n + n = E' + E" - n,
F ::;:; pi - 1 + F 'I - 1 ::;:; F ' + pI! - 2.
Indeed, V is the number V' of vertices in T g plus the number V" in T 1 , except
that the n vertices of the polygon along which T1 and T g are glued have been
counted twice, so V = V' + V" - n; a similar argument applies to the edges;
and F is as stated because the polygon along which T 1 and T g are glued is not
is not acceptable, since after gluing, the two shaded triangles intersect in two part of the triangulation of Tg+l' Hence,
vertices, which is not allowed:
X(Tg+l} =V - E +F
=. (Vi + V" - n) - (E' + E" - n) + (F' + F" - 2)
= Vi - E' + pi + VII - Ell +F II
- 2
= X(Tg) + X(Tt} - 2
= 2 - 2g + 0 - 2 (by the induction hypothesis)
::;:; 2 - 2(g + I},
But the finer subdivision above does work, and gives proving the result for genus 9 + 1. o
x ::;:; 9 - 27 + 18 "" 0 ::;:; 2 - 2 xl,
proving the theorem when 9 ::;:; 1. Corollary 11.3
We now complete the proof by induction on g, using the fact that Tg +l can We have
be obtained from T g by gluing on a copy of Tl:
JlTr g
KdA::;:; 41T(1- g).

Proof 11.3
Just combine Theorems n.5 and 11.6. o

EXERCISES

11.5 Show that, if a triangulation of a compact surface with Euler num-

T. ber X by curvilinear triangles has V vertices, E edges and F trian-
gles, then

3F:=2E, E=3(V-X), V~~(7+J49-24X},

268 Elementary Differential Geometry 11. The Gauss-Bonnet Theorem 269

11.6 A triangulation of the sphere has n curvilinear triangles, and r (ii) x 2 + y2 + Z4 = l.

triangles meet at each vertex. Explain why there are 3n/r vertices Which is which, and why? Sketch the compact surface and write
altogether in the triangulation, and write down the total number down its Euler number.
of edges. Show that
6 4
---=1.
r n
Deduce that r ::; 5, and sketch triangulations of the sphere corre-
11.4. Singularities of Vector Fields
=
sponding to r 3,4 and 5.
Suppose that S is a surface and that V is a smooth tangent vector field on S.
11.7 Show that, given 5 points on a sphere, it is impossible to connect This means that, if (l : U ~ R 3 is a patch of Sand (u,v) are coordinates on
each pair by curves on the sphere that intersect only at the given U, then
points. (Such a collection of curves would give a triangulation of
the sphere for which 2E ;::: 3F, since each face would have at least V = a(u, v)O'u + f3(u, v)O'v,
3 edges.) Deduce that the same result holds if 'sphere' is replaced where a and f3 are smooth functions on U. It is easy to see that this smoothness
by 'plane'. (Use stereographic projection.) condition is independent of the choice of patch 0' (see Exercise 11.14).
11.8 Let Pl,P2,P3 and Ql,Q2,Q3 be points on a sphere. Show that it
is impossible to join each Pi to each Qj by 9 curves on the sphere
Definition 11.5
that intersect only at the given points. (This is sometimes called
the 'Utilities Problem', thinking of PI, P2 and P3 as the gas, water If V is a smooth tangent vector field on a surface S, a point P of S at which
and electricity supplies to three homes Ql, Q2 and Q3') V = 0 is called a stationary point of V.
11.9 Show that, if a compact surface S is diffeomorphic to the torus T 1 ,
The reason for this terminology is as follows. We saw in the proof of Propo-
then
sition 7.4 that, if P is any point of 5, there is a unique curve 'Y(t) on S such
JIs KdA=O. that "I = V and '1(0) = P; '1 is called an integral curve of V. We can think of '1
as the path followed by a particle of some fluid that is flowing over the surface.
Can such a surface have K = 0 everywhere?
If V = 0 at P, the velocity "I of the flow is zero at P, so the fluid is stationary
11.10 Show that, if S is the ellipsoid there.
x2 + y2 Z2 We are going to prove a theorem which says that the number of stationary
a2 + b2 = 1, points of any smooth tangent vector field on a compact surface S, counted
where a and b are positive constants, with the appropriate multiplicity, is equal to the Euler number of S. To define
this multiplicity, let P be a stationary point of V contained in a surface patth
JIs K dA = 47r. 3
e
0' : U ---t R of S, say, with 0'( Uo, vo) = P. Let be a nowhere vanishing smooth
e
tangent vector field on O'(U) (for example, we may choose = O'u or O'v), and
By computing the above integral directly, deduce that let 1/J be the angle between V and e.
"/2 ab2 cos 9
j --."--::--,,",,,:,,"",:::-:-;:: d()
-,,/2 (a 2 sin 2 9 + b2 cos2 9)3/2
= 2.
Definition 11.6
11.11 Suppose that S is a compact surface whose gaussian curvature K With the above notation, the multiplicity of the stationary point P of the
is> 0 everywhere. Show that S is diffeomorphic to a sphere. Is the tangent vector field V is
converse of this statement true?
1 rl('Y) d'IjJ
11.12 One of the following surfaces is compact and one is not: JJ(P) = 211" Jo ds ds,
(i) x 2 - y2 + Z4 = 1;
270 Elementary Differential Geometry ¥ n. The Gauss Bonnet Theorem 271

(iv) V(x, y) :::: (x, -y); IJ.:::: -1

where '1(s) is any positively-oriented unit-speed simple closed curve of length
l('Y) in O'(U) with P in its interior.

It is clear that IJ.(P) is an integer, a~d an argument similar to our heuristic

proof of Hopf's Umlaufsatz in Section 1i.l shows that I-'(P) does not depend on
the choice of simple closed curve 'Y. It is also easy to see that it is independent
e
of the choice of 'reference' vector field (see Exercise 11.15).

Example 11.1 The stationary point in examples (i), (ii), (iii) and (iv) is called a source, sink,
The following smooth tangent vector fields in the plane have stationary points vortex and bifurcation, respectively.
of the indicated multiplicity at the origin (we have shown the integral curves Let us verify the multiplicity in case (iv), for example. Take the 'reference'
of the vector fields for the sake of clarity);
tangent vector field to be the constant vector field (1,0). Then, the angle e:: :
'l/J is given by
(i) V(x,y):::: (x,Y)j J.I.:::: +1.

(cos 1/', sin 'l/J) :::: ~

IIVII
== (
.jx2 +y2
X I - Y
.jX2 +y2
)
.

Taking 'Y(s) :::: (cos s, sin s) to be the unit circle, at 'Y( s) the angle 1/' satisfies
(cos 'l/J, sin 1/') :::: (cos s, - sin s),
so 1/' :::: 21T - s. Hence,

(ii) V(x,y):::: (-x,-y); J.I.:::: +1 (l(0,0) = 21T

1 1 0
2
11" d
ds (271" - s) ds :::: -1.

Theorem 11.7
Let V be a smooth tangent vector field on a compact surface S which has only
finitely many stationary points, say P1 ,P2 , ••• ,Pn . Then,
n

L{l(Pi ):::: X,
i=l
(iii) V(x, y) :::: (y, -x)j J.I. :::: +1 the Euler number of S.

Proof 11.7
Let 'Yi be a positively-oriented unit-speed simple closed curve contained in a
patch O'i of S with Pi in the interior of 'Yi' Assume that the 'Yi are chosen so
small that their interiors are disjoint. Choose a triangulation of the part Sf of S
r
outside 'Yu 'Y2,'" ,'Yn by curvilinear polygons j . Note that the edges of some
of these curvilinear polygons will be segments of the curves 'Yi:
272 Elementary Differential Geometry
+ 11. The Gauss-Bonnet Theorem 273

Combining Eqs. (15), (17) and (18), we get

Ln l i (";.)
' (f'.f" - e ' ,e") ds = 27rX· (19)
i;;:l 0

But, from the proof of Theorem 11.2,

,./1
e.e n
=11-/\'g, f l' f '=tp-K,g,
'''

where ,,"g is the geodesic curvature of 'Yi and () and tp are the angles between 'Yi
and e ' and f', respectively, Then, 'I./J = tp - () is the angle between e ' and fl , Le.
.',"
between V and the 'reference' tangent vector field fl on (Ii. So the left-hand
side of Eq, (19) is
Note also that, when these polygons are positively-oriented, the induced orien-
tation of the 'Yi is opposite to their positive orientation (see the diagram above,
in which the arrows indicate the sense of positive orientation).
We can regard the curvilinear polygons in $I, together with the simple
as we want. o
closed curves 'Yi and their interiors, as a triangulation of 8, so by Theorem
11,3,

[ K dA +
15'
t [
i;;:l llnt('YJ
K dA = 27rX, (15) We now give some simple examples of vector fields on surfaces (we show
their integral curves for clarity).
where X is the Euler number of 8. On 8', we choose an orthonormal basis
{e', e"} of the tangent plane of 8 at each point so that e' is parallel to the Example 11.2
tangent vector field V, Arguing as in the proof of Theorem ILl, we see that
A vector field on the sphere with 1 source and 1 sink: X = 2
l(rj)

1 5'
K dA = L
j l 0
e'.e" ds, (16)

where s is arc-length on r j and f(r}) is its length. Any common edge of

r
two of the curvilinear polygons j is traversed once in each direction so their
contributions to the sum in Eq. (16) cancel out. What remains is the integral
along the segments of the curves 'Yi that are part of the polygons j ' r In view
of the remark about orientations above, we get
n l l ('Y)

1 5'
KdA=-L
i=l 0
. e'.e"ds, (17)

where s is arc-length along 'Yi and f('Yi) is its length.

Now choose an orthonormal basis {fl, f"} of the tangent plane of 8 on each
patch C1i. By the proof of Theorem 11.1,
l('Y;) •

1 int{'Y,)
K dA =
l 0
f',f" ds. (18)
274 Elementary Differential Geometry T
I
11. The Gauss-Bonnet Theorem 275

(fof S, the three components of V at the point u(u, v) are smooth

Example 11.3 functions of (u, v).
I 11.15 Show that the Definition 11.6 of the multiplicity of a stationary
A vector field on the torus with no sta~ionary points: X = 0
point of a tangent vector field V is independent of the 'reference'
e
vector field (. (If is another reference tangent vector field, and ()
........... , ..... e,
is the angle between ( and then d(}jds =
-(1- p2)-1/2 p, where
p = cos(). Now use Green's theorem to show that J~("()(d()jds)ds =
........... ~ .
0.)

11.5. Critical Points

If I (ti, v) is a smooth function defined on an open subset U of R 2 , we say that a
point (uo, vo) is a critical point of I if al jau and allav both vanish at (uo, vo).
Example 11.4 Equivalently, the gradient of I,
A vector field on the double torus T2 with 2 bifurcations: X = -2
VI = (a l al)
au' av '
should vanish at (uo, va).
If now F : S --+ R is a smooth function on a surface S (see Exercise 4.11),
and if q : U --+ R 3 is a surface patch of 5 then I = F 0 q is a smooth function
I

on the open subset U of R 2 . This suggests

Definition 11.7
Let 5 be a surface and F : S --+ R a smooth function on S. A point P of Sis
a critical point of F if there is a surface patch q of S, with P = q(uo, vo), say,
EXERCISES such that I = F 0 (T has a critical point at (uo, vo).

11.13 Let k be a non-zero integer and let V (x, y) = (a, f3) be the vector It is easy to check directly that the definition of a critical point is indepen-
field on the plane given by dent of the choice of patch IT (see Exercise 11.18), but this will follow imme-
diately from another characterisation of critical points that we shall now give
'f3_{(X+i y )k ifk>O,
and which is independent of any arbitrary choices.
a +t - (x . )-k
- ty 'f k <.
1 0
If F : S -t R is a smooth function, we want to define an analogue of the
Show that the origin is a stationary point of V of multiplicity k. gradient of I that is a tangent vector field on S. This is accomplished in
(Take,,( in Definition 11.6 to be the unit circle and use de Moivre's
theorem.)
Proposition 11.1
11.14 Show that the definition of a smooth tangent vector field is inde-
pendent of the choice of surface patch. Show also that a tangent If F is a smooth junction on a surface 5, there is a unique smooth tangent
vector field V on S is smooth if and only if, for any surface patch vector field V sF on S such that, if P is a point of Sand "(t) is a curve in S
276 Elementary Differential Geometry T 11, The Gauss-Bonnet Theorem 277

which passes through P when t = 0, we have matrix

~n
~ 1,=0 F(')'(t)). 1i = (~¥
(VsF) ..y(O) = (20)
}juJv a
"a",av )

is invertihle, where f = F 0 u and the derivatives are evaluated at (uo, vol.

Moreover, P is a critical point of F if and only if V sF = 0 at P.
In this c""e, the point P is called a local maximum, a saddle point or a local
minimum if 1i h"" 2, 1 or 0 negative eigenvalues, respectively.
Proof 11.1
The tangent vector field V sF is obviously unique, if it exists. Indeed, every It is not difficult to show that this definition is sensible, i.e. independent of
tangent vector to 5 at P is of the form .y(0) for some curve')' on 5 with the choice of patch u (see Exercise 11.16). Note that the matrix 1i is real and
')'(0) = P, so any two choices of V sF at P would differ by a vector perpendicular symmetric, so it always h"" two real eigenvalues (not necessarily distinct).
to every tangent vector to 5 at P, which must be zero.
To see that V sF exists, choose a surface patch u(u, v) for 5 with u(uo, vol = Proposition 11.2
P, say, and let f = F 0 u. Let {e', e"} be the b""is of the tangent plane of 5
at P such that Let P be a critical point of a smooth /unction F on a surface 5. Then, the
/
multiplicity of P as a stationary point of V sF is
e .t1 u = efl.tlt) = I, e'.t1 v = e".flu = O. (21)
J-t(P) = {' 1 if P is a local maximum or a local minimum,
Explicitly, -1 if P is a saddle point.
, Guu - Fuv " Eu v - Fu u (22) Example 11,5
e = EG-F2 e = EG-F2 '
where Edu 2 + 2Fdudv + Gdv 2 is the first fundamental form of u. We take The function on the plane given by F(u, v) = _u 2 _ v 2 (resp. u 2 _ v 2, u 2 + v')
h"" a local maximum (resp. saddle point, local minimum) at the origin.
VsF = fue' + fve", (23)
where the derivatives are evaluated at (uo, vol. If')' is "" in the statement of We shall not give a complete proof of Proposition 11.2 here. But the fol-
the proposition, say ')'(t) = u(u(t), v(t)), then (with dldt denoted by a dot) lowing argument should convince the reader of its truth. Let us ""sume that
(uo, vol = (0,0) for simplicity, and write the matrix in Definition 11.8 at
(V sF) ..y(O) = (fue' + fve").(uuu + uvv)
u=v=Oas
= fuu + fvv (by Eqs. (21))
= j, 1i=C ~).
the derivatives with respect to t being evaluated at t = O. It is clear from Eqs. Then, Taylor's theorem tells us that
(22) and (23) that V sF is smooth and that P is a critical point of F if and 1
only ifVsF=O at P. 0 flu, v) = 2"(AU' + 2J1.uv + vv') + r(u, v),
where r(u, v)/(u' + v') tends to zero as u and v tend to zero. It is plausible,
Since V sF is a smooth tangent vector field on 5, we can apply Theorem then, that the behaviour of V sF near P is the same"" that of V sF, where
11.7 to it. To do so, we must compute the multiplicity of the stationary points
- 1
of V sF. For this, we shall make an additional ""sumption about F, contained F(u(u, v)) = 2"(AU2 + 2J-tuv + vv').
in
In particular, F and F should have the same type of critical point at P.
But the multiplicity of P "" a critical point of F is e""y to compute. To do
Definition 11.8 so, note first that there is a rotation matrix P such that
A critical point P of a smooth function F on a surface 5 is said to be non-
p'(A J-t)p=(f 1 0)
degenerate if, whenever u(u, v) is a patch of 5 with P = u(uo, vol, say, the J1. v 0 f,
 Elementary Differential Geometry
+ 11, The Gauss-Bonnet Theorem 279
278

is a diagonal matrix. This means that, by applying a rotation in the uv-plane Example 11.6
(i.e. a reparametrisation of 0'), we can assume that If we take the surface T g of genus 9 described at the beginning of Section 11.2
- ,1 2 2 and stand it upright on the xy-plane, the height above the plane is a smooth
F(O'(u, v)) = 1'2(€1 u + €2V ).
function F on T g • The critical points of F are as shown in the following diagram,
We can also assume that the patch (f is conformal, so that angles measured in and they are all non-degenerate (cf. Exercise 11.18).
the uv-plane are the same as those measured on the surface (see Section 9.4).
Taking our 'reference' tangent vector field to be tT u , the angle TjJ between
R
VsF and tT u is given by
(€lU,€2 V )
(cos TjJ, sin TjJ) = ---j.~;=~""2
J€~U2 + ' 4v
We take the simple closed curve in the uv-plane given by the ellipse
€~U2 + €~V2 =r 2 ,
where r is a small positive number, which can be parametrised by
r r .
u = i4T cost, v = ~ smt.
Hence,

cos TjJ = i4T cos t,

fl . .1.
smOf' = ~smt,
f2.

and so
t if fl > 0 and f2 > 0,
TjJ = 211" - t if fl > 0 and f2 < 0,
{ 11" - t if fl < 0 and f2 > 0,
11" + t if fl < 0 and f2 < O.
This gives the multiplicity of P as a stationary point of V sF as
r
~ 21r dTjJ dt = { 1 if €1 and f2 have the same sign,
211" i o dt -1 otherwise,
in accordance with Proposition 11.2.
If we accept this heuristic argument, we can combine Theorem 11.7 and
Proposition 11.2 to give There is a unique local minimum P, 2g saddle points Ql, Q2, ... , Q2g 1 and a
unique local maximum R. Hence, Theorem 11.8 gives the Euler number of Tg
as
Theorem 11.8
1 - 2g + 1 = 2 - 2g,
Let F : S ~ R be a .~mooth function on a compact surface S with only finitely
many critical points, all non-degenerate. Then, in accordance with Theorem 11.6.
number of local) _ (numb~r of saddle) + (nu,":b~r of local) = X,
( maxima of F po1.nts of F mmzma of F
the Euler number of S.
280 Elementary Differential Geometry
+

EXERCISES
Solutions
11.16 Show directly that the de~nition of a critical point, and whether
it is non-degenerate, is ind'ependent of the choice of surface patch
in Definitions 11. 7 and 11.8. Show that the classification of non-
degenerate critical points into local maxima, local minima and
saddle points is also independent of this choice. (Show that the
critical point is a local maximum (resp. local minimum) if and only
if v t llv < 0 (resp. > 0) for all non-zero 2 x 1 matrices v.)
11.17 For which of the following functions on the plane is the origin a non-
degenerate critical point? In the non-degenerate case(s), classify
the origin as a local maximum, local minimum or saddle point.
(i) x 2 - 2xy + 4y 2;
(ii) x 2 + 4xYi
(iii) x 3 - 3xy 2.
11.18 Let S be the torus obtained by rotating the circle (x-2)2 +Z2 = 1 in
the xz-plane around the z-axis, and let F : S -+ R he the distance
from the plane x = -3. Show that F has four critical points, all Chapter 1
non-degenerate, and classify them as local maxima, saddle points,
or local minima. (See Exercise 4.10 for a parametrisation of S.) 1.1 It is a parametrisation of the part of the parabola with x ~ O.
1.2 (i) 'Y(t) = (sect, tan t) with -1r/2 < t < 1r/2 and 1r/2 < t < 31r/2.
(ii) 'Y(t) = (2cost,3sint).
=
1.3 (i) x + y 1.
(ii) Y = (lnx)2.
1.4 (i) .y(t) = sin 2t( -1,1).
(ii) .y(t) = (e t ,2t).
= =
1.5 .y(t) = 3 sin t cos t(- cos t, sin t) vanishes where sin t 0 or cos t 0, Le.
t = n1r /2 where n is any integer. These points correspond to the four
cusps of the astroid.

281
282 Elementary Differential Geometry + solutions 283

1.6 (i) Let OP make an angle e with the positive x-axis. Then R has coordi- 1.9 A point (x,y,z) lies on the cylinder if x 2 + y2 = 1/4 and on the sphere
nates "I(e) = (2acote,a(1 - cos2e)). if (x + ~)2 + y2 + Z2 = 1. From the second equation, -1 ::; z ::; 1 so
let z = sin t. Subtracting the two equations gives x + ~ + sin t :;; ~' so
2
(ii) From x = 2acote, y = a(l- cos 28), we get sin 2 8 = y!2a, cos 2 8 =
e
cot 2 sin 2 0 :;; x 2 y!8a 3 , so the cwtesian equation is y!2a + x 2 y!8a 3 = 1. x = ~ - sin 2 t :;:: cos 2 t - ~. From either equation we then get y :;:: sin t cos t
1.7 When the circle has rotated through an angle t, its centre has moved to (or y :;:: - sin t cos t, but the two solutions are interchanged by t I-t 71" - t).
(at, a), so the point on the circle initially at the origin is now at the point 1.10 1'(t) :;:: (et(cost - sint),et(sint + cost» so the angle () between 'Y(t) and
(a(t - sint), a(l- cost». i'(t) is given by
'Y'Y e2 t(cos 2 t-sintcost+sin 2 t+sintcost) 1
cosB:;:: II 'Y 11'11 i' II :;:: e2 t«cost-sint)2+ (sint+cost)2) :;:: 2'
and so 8 :;; 1f /3.
1.11 t(t) :;:: (1, sinh t) so II l' II:;:: cosh t ~d the arc-length is s :;:: f~ cosh u du :;;
sinh t.
1.12 (i) 111' 11 2 :;:: t(l + t) + t(l - t) + 1. t :;::
(ii) II i' 11 2 :;:: ~~ sin 2 t + cos 2 t + 295 sin 2 t :;:: cos 2 t + sin t :;:: 1.
2

.t 1.13 The cycloid is parametrised bY'Y(t) :;:: aCt - sint, 1 - cost), where t is
the angle through which the circle has rotated. So i' :;:: a(l - cos t, sin t),
1.8 Let the fixed circle have radius a, and the moving circle radius b (so that
b < a in the case of the hypocycloid), and let the point P of the moving
II t 11 2 :;:: a 2 (2 - 2 cos t) :;:: 4a2 sin 2 ~' and the arc-length is
circle be initially in contact with the fixed circle at (a, 0). When the moving t I
t t ;2".
1
2".
circle has rotated through an angle c.p, the line joining the origin to the 2asin 2' dt:;:: -4a cos 2' :;:: Ba.
o t=o
point of contact of the circles makes an angle () with the positive x-axis,
1.14 (i) l' :;:: sin 2t( -1, 1) vanishes when t is an integer multiple of 1f /2, so 'Y is
where a() :;; bc.p. The point P is then at the point
not regular.
'Y(B) = «a + b) cosO - bcos(B + c.p), (a + b) sinO - bsin(B + c.p» (ii) Now 'Y is regular since i' =f:. 0 for 0 < t < 11'/2.
= «a + b) cosO - bcos«a + b)B/b), (a + b) sinO - bsin«a + b)8/b» (iii) i' :;:: (1, sinh t) is obviously never zero, s0'Y is regular.
in the case of the epicycloid, 1.15 x :;:: r cos e :;::sin 2 B, y :;:: r sin 8 :;:: sin 2 0 tan 8, so the parametrisation in
e
terms of () is () I-t (sin2 8, sin2 8 tan 0). Since 8 I-t sin is a bijective smooth
map (-1f /2, 1f /2) ~ (-1,1), with smooth inverse t I-t sin- 1 t, t :;:: sin 0 is
a reparametrisation map. Since sin 2 8 = t 2 , sin 2 () tan 0:;:: t 3 /~, the
reparametrised curve is as stated.
1.16 We have s = ft~ II d'Y/du II du, s = flo II diif/dilll dil. By the chain rule,
d"l/du = (diif/dil)(dil/du), so s :;:: ± hto II diif/dil II (dil/du) du :;:: ±s, the
sign being that of du/du.
1.17 If 'Y(t) :;:: (x(t), yet), z(t» is a curve in the surface f(x, y, z) :;:: 0, differen-
tiating f(x(t), Yet), z(t» = 0 with respect to t gives xfx + ilf1l + if" = 0,
so t is perpendicular to (fx, Ill' f,,)· Since this holds for every curve in the
surface, (fx, /11' f,,) is perpendicular to the surface. The surfaces / :;:: 0 and
and g :;:: 0 should intersect in a curve if the vectors (fx, /11' J,,) and (gx, gy, gz)
'Y(8) = «a - b) cos() + bcos(c.p - 8), (a - b) sinO - bsin(c.p - 8» are not parallel.
= «a - b) cos() + bcos«a - b)B/b), (a - b) sinO - bsin«a - b)8/b» 1.18 Let 'Y(t) = (u(t), vet), wet»~ be a regular curve in R 3 . At least one of u, V, tV
is non-zero at each value of t. Suppose that u(to) =f:. 0 and Xo :;:: u(to)· As
in the case of the hypocycloid.
 sy
284 Elementary Differential Geometry Solutions 285

in the 'proof' of Theorem 1.2, there is a smooth function hex) defined (iii) Using Proposition 2.1,
for x near Xo such that t = hex) is the unique solution of x = u(t) for
II (1, sinh t, 0) X(0, cosh t, 0) II cosh t 2
each t near to. Then, for t near to, 'Y(t) is contained in the level curve '" = = -- 3 = sech t.
II (1, sinh t, 0) 11 3 cosh t .
f(x, y, z) = g(x, y, z) = 0, where j,(x,
I
y, z) = y - v(h(x» and g(x, y, z) =
Z - w(h(x)). The functions! and g satisfy the conditions in the previous (iv) (-3 cos 2 t sin t, 3 sin 2 t cos t, 0) X (-3 cos3 t +6 cos t sin2 t, 6 sin t cos 2 t-
exercise, since (fx, fy, fz) = (-iJh J , 1, 0), (gx, gy, gz) = (-tiJh J , 0,1), a dash 3 sin3 t, 0) = (0,0, -9 sin2 t cos 2 t), so
denoting d/dx. II (0,0, -9 sin2 t cos 2 t) II 1
1.19 Define B(t) = tan 1rO(t)/2, where 8 is the function defined in Exercise ~- -.,....,...-------,.
- II (-3 cos 2 t sin t, 3 sin 2 t cos t, 0) 11 3 - 31 sin t cos tI'
8.20. Then e is smooth, B(t) = 0 if t ~ 0, and e ; (0,00) ---t (0,00) is a
bijection. The curve This becomes infinite when t is an integer multiple of 1r /2, i.e. at the four
cusps (±1, 0) and (0, ±1) of the astroid.
'Y(t) = { (e(t), e(t» if t ~ 0, 2.2 The proof of Proposition 1.4 shows that, if vet) is a smooth (vector)
(-e( -t), B( -t)) if t :::; 0,
function of t, then II vet) II is a smooth (scalar) function of t provided vet)
is a smooth parametrisation of y = Ixl. is never zero. The result now follows from the formula in Proposition 2.1.
The curvature of the regular curve 'Y(t) = (t, t 3 ) is K:(t) = 6Itl/(1 +9t4 )3/2,
which is not differentiable at t = O.
2.3 Differentiate t.D s = 0 and use i = ~sns.
2.4 If 'Y is smooth, t = "I is smooth and hence so is n s (since it is obtained by
applying a rotation to t) and 1. So ~s = t.n s is smooth.
2.5 The arc-length s = J II "I II dt = J II e kt (k cos t - sin t, k sin t + cos t) II dt
= J Jk 2 + 1e kt dt = ~ekt + c, where c is a constant. Taking c = 0
makes s -+ 0 as t ---t TOO if ±k > O.
Since t = 'Y/ 11"1 11= v'k~+1(kcost - sint,ksint + cost), we have
There is no regular parametrisation of y = Ixl. For if there were, there n s = v'k~+1 (-ksin t - cost, kcost - sin t). So dt/ds = (dt/dt)/(ds/dt) =
ke
would be a unit-speed parametrisation 'Y(t) , say, and we can assume that e-
k'+1 (
- k sint-cost, k cost-sint ) = Jk2+1nS
-kt
so"'s=1 / ks.ByTheorem
'Y(O) = (0,0). The unit tangent vector ~ would have to be either ~(1, 1) or 2.1, any other curve with the same signed curvature is obtained from the
- ~(1,1) when x > 0, so by continuity we would have .y(0) = ± ~(1, 1). logarithmic spiral by applying a rigid motion.
2.6 (i) Differentiating 'Y = rt gives t = ft + ~BrnS. Since t and n s are per-
But, by considering the part x < 0 in the same way, we see that .y(0) =
pendicular unit vectors, it follows that ~s = 0 and 'Y is part of a straight
±~(1, -1). These statements are contradictory.
line.
(ii) Differentiating 'Y = rn s gives t = rD s + rn s = fn. - ",.rt (Exercise
2.3). Hence, f = 0, so r is constant, and "'s = -l/r, hence "'s is constant.
Chapter 2 So 'Y is part of a circle.
(iii) Write'Y = ret cos O+ns sin 0). Differentiating and equating coefficients
2.1 (i) 'Y is unit-speed (Exercise 1.12(i)) so of t and D s gives f cos 0 - ~sr sin 0 = 1, f sin (J + ~sr cos 0 = 0, from which
f = cosO and ~sr = - sinO. From the first equation, r = scosO (we can
~ = II i II = II (-41 (1 + t) -1/2, -41 (1 - t) -1/2,0) 11= 1 assume the arbitrary constant is zero by adding a suitable constant to
J8(1- t 2 ) s) so ~s = -l/scotO. By Exercise 2.5, 'Y is obtained by applying a rigid
(ii) 'Y is unit-speed (Exercise 1.12(ii)) so motion to the logarithmic spiral defined there with k = - cot (J.
2.7 We can assume that 'Y is unit-speed. Then "I>' = (1 - AK:s)t, and this is
K: = II i II = II (- ~ cos t, sin t, ~ cos t) II = 1. non-zero since 1 - AK: s > O. The unit tangent vector of 'Y>' is t, and the
286 Elementary Differential Geometry 4 Solutions 287

arc-length s of 'Y'\ satisfies ds/dt = 1 - AK,. Hence, the curvature of 'Y'\ is 2.14 (i) t = (~(1 + t)'/2, -~(1 - t)'/2, Jz) is a unit vector so 'Y is unit-speed;
II dt/ds II = II i II /(1 - AK,) = K,/(l - AK,). i = (~(l+W'/2, ~(1- W ' / ,0), so K = II i II = 1/V8(1- t 2 ); n = ~i =
2

2.8 The circle passes through 'Y(s) because II < - 'Y II = II ••'n • II = l/IK,I, },«1- t)'/2, (1 + t)1/2,0); b = t x n = (-~(1 +t)'/2, ~(1- t)'/2, -},);
which is the radius of the circle. It is tangent to 'Y at this point because
b = (_~(I+W'/2,-W - W ' / 2,0) so the torsion 7 = I/V8(I-t 2).
< - 'Y = -L n , is perpendicular to the tangent t of 'Y. The curvature of the The equation iJ. = -K.t + Tn is easily checked.
'.
circle is the reciprocal of its radius, Le. IK,I, which is also the curvature (ii) t = (-t sin t, - cos t, ~ sin t) is a unit vector so 'Y is unit-speed; i =
of 'Y.
2.9 T he tangent vector of < is t + -L(-K,t) - ~n, = -~n, so its arc-length
(-t t
cos t, sin t, ~ cos t), so K ~ II i II = 1; n = ~i = (- cos t, sin t, ~ cos t);
n. K. K. b = t X n= (-~,O,-t), so b=O and 7 =0.
is u = J II E II ds = J ~ds = Uo - ,L,
where Uo is a constant. Hence, 2.15 Show that 7 = 0 or observe that x = '1", y = t + 1, z = It' satisfy
the unit tangent vector of < is -n, and its signed unit normal is t. Since x - y - z = O.
-dn,/du = K,t/(du/ds) = r,-t, • ' the signed curvature 0 f < IS. K,3/'K,. 2.16 By Proposition 2.5, 'Y is a circle of radius I/K = 1 with centre 'Y +
Denoting d/dt by a dash, 'Y' = a(l - cost,sint) so the arc-length ~n = (0,1,0) in the plane passing through (0,1,0) perpendicular to
s of 'Y is given by ds/dt = 2asin(t/2) and its unit tangent vector b = (-~, 0, -t), i.e. the plane 3x + 4z = O.
is t = 'Y = (sin(t/2), cos(t/2)). So n, = (- cos(t/2), sin(t/2)) and 2.17 Let a = K/(K 2 + 7 2 ), b = 7/(K 2 + 7 2 ). By Examples 2.1 and 2.4, the
i = (dt/dt)/(ds/dt) = 4a'i~{t)2) (cos(t/2), - sin(t/2)) = -1/4asin(t/2)n" circular helix with parameters a and b has curvature a/Ca' + b') = K and
so the signed curvature of'Y is -1/4asin(t/2). Its evolute is therefore torsion b/(a 2 + b2 ) = 7. By Theorem 2.3, every curve with curvature K
< = a(t-sin t, I-cos t) -4a sin(t/2)(- cos(t/2), sin(t/2)) = a(t+sin t, -1+ and torsion 7 is obtained by applying a rigid motion to this helix.
cost). Reparametrising by t = 71'H, we get a(t-sint, 1-cost)+a( -71', -2), 2.18 This follows from Proposition 2.3 since the numerator and denominator
which is obtained from a reparametrisation of'Y by translating by the vec- of the expression in (11) are smooth functions of t.
tor a(-1r, -2). 2.19 Let a dot denote d/dt. Then, 6.=;y = Kn, so the unit tangent vector of 6 is
2.10 The free part of the string is tangent to 'Y at 'Y(s) and has length t-s, hence T = n and its arc-length s satisfies ds/dt = K. Now dT Ids = ti./(ds/dt) =
the stated formula for t(s). The tangent vector of t is 'Y - 'Y + (t - s)'Y K-l(_~t + 7b) = -t + ~b. Hence, the curvature of 6 is II -t + ~b II =
= K,(t - s)n, (a dot denotes d/ds). The arc-length v of t is given by (1 + ;;'-)1/2 = 1', say. The principal normal of 6 is N = 1'-1 (-t + ~b)
dv/ds = K,(t - s) so its unit tangent vector is n, and its signed unit and its binormal is B = T x N =.= IJ-l(b + ~t). The torsion T of 6 is
normal is -to Now dn,/dv = •• d_,jti., = f-:';t, so the signed curvature given by dB/ds = -TN, Le. K- 1B = -TN. Computing the derivatives
of tis l/(t - s). and equating coefficients of b gives T = (KT - 71<)/K(K 2 + 7 2 ).
2.11 (i) With the notation in Exercise 2.9, the involute of f is 2.20 Differentiating t.a (= constant) gives n.a = 0; since t, n, b are an or-
dE 1 thonormal basis of R 3, a = t cos 8 + IJb for some scalar 1'; since a is a unit
t(u) =<+(t-u)- ='Y+ -n, - (t-u)n, ='Y- (t-uo)n"
du "'s vector, I' = ± sin 8; differentiating a = t cos 8 ± b sin 8 gives 7 = K cot 8.
since u = tLo - ...!.., so " is the parallel curve -y-(l-uo). Conversely, if 7 = AK, there exists 8 with A = cot 8; differentiating shows
"
(ii) Using the results of Exercise 2.10, the evolute of t is that a = t cos 8 + b sin 8 is a constant vector and t.a = cos 8 shows that
8 is the angle between t and a.
t+ (t - s)(-t) = 'Y+ (t - s)t - (t - 8)t = 'Y.
2.21 Differentiating ('Y - a).('Y - a) = r 2 repeatedly gives t.('Y - a) = 0; t.t +
2.12 If we take the fixed vector in Proposition 2.2 to be parallel to the line of Kn.('Y-a) = 0, and so n.('Y-a) = -l/K; n.t+(-Kt+7b).('Y- a ) = 1</",2,
reflection, the effect of the reflection is to change 'P to -'P, and hence to and so b.('Y - a) = I</7K 2; and finally b.t - m.('Y - a) = (K/7K 2)" and so
2
change 1'\,8 to -"'s' 7 / K = (K/n )'.
2.13 If two unit-speed curves have the same non-zero curvature, their signed Conversely, if Eq. (20) holds, then p = -u(pu)', so (p' + (pU)2)' = 2pp +
curvatures are either the same or differ in sign. In the first case the curves 2(pu)(pu)' = 0, hence p2 + (pU)2 is a constant, say r 2 (where r > 0). Let
differ by a rigid motion by Theorem 2.1; in the latter case, applying a a = 'Y+pn+pub; then Ii = t+pn+p(-Kt+7b)+(pu)'b+(pu)(-m) = 0
reflection to one curve gives two curves with the same signed curvature using Eq. (20); so a is a constant vector and II 'Y-a 11 2 = p2 + (pU)2 = r 2,
by Exercise 2.12, and these curves then differ by a rigid motion.
 ¥ Solutions 289
288 Elementary Differential Geometry

hence 'Y is contained in the sphere with centre a and radius r. of the ellipse because
2.22 Let Aij = Vi.Vj. The vectors VI, V2, V3 are orthonormal if and only if (tXI + (1 - t)X2)2 (tYI + (1 - t)Y2)2
if
Aij = (iij (= 1 if i = j and = 0 i '" j). So it is enough to prove that a2 + b2
Aij = Jij for all values of s given/that it holds for s = So. Differentiating
Vi.Vj gives
=t2 (Xia + Yi)
2 b
+(1_t)2 (x~ + y~) +2t(1-t) (XIX2 + YIY2)
2 a b 2 a 2 b 2 2

3 x2 y2 x2 y2)
< t 2 + (1 - t)2 + t(l - t) ( ---.!.2 + -.!2 + --1.2 + -1.2
Aij = E(ai.A,j + aj.Ai.). a b a b
k=l
< t' + (1 - t)2 + 2t(1 - t) = 1.
Now Aij = Jij is a solution of this system of differential equations because
a,; + aj' = O. But the theory of ordinary differential equations tells us 3.7 If 'Y(t) is a reparametrisation of 1(t), the vertices of 1 (resp. 'Y) are given by
that there is a unique solution with given values when s = So· dl<,/dt = 0 (resp. dl<,/dt = 0). Since dl<,/dt = (dl<,/dt)(dt/dt) and since
dt/dt is never zero, 1 and 'Y have the same vertices.
3.8 "Y = (- sin t - 2 sin 2t, cos t + 2 cos 2t) and II "Y II = v'r.5-+~4-c-os---'t, so the angle
<p between "Y and the x-axis is given by
Chapter 3 cost,p =
-sint - 2sin2t
1
.
sm<p =
cost + 2 cos 2t
.
';5 + 4cost ';5 + 4cost
3.1 If 'Y is obtained from 1 by a translation, then.y = "y so by Eqs. (1) and (3)
Differentiating the second equation gives <p cos <p = - sin t(24 cos2 t +
the length and area of 'Yare the same as those of 1. If 'Y is obtained from 1
42cost + 9)/(5 + 4cost)3/2, so <p = sint(24cos2 t + 42cost'+ 9)/(5 +
by a rotation by an angle 0 about the origin, then x = x cos 0 - Y sin 0, ii =
4cost)(sint + 2sin2t) = (9 + 6cost)/(5 + 4cost). Hence, if s is the arc-
xsinO + ycosO. This implies that l + = x 2 +y2 and Ii' xii -
iii = xy - yx, length of 1, 1<, = d<p/ds = (d<p/dt)/(ds/dt) = (9 + 6C08t)/(5 + 4cost)3/2,
so Eqs. (1) and (3) again show that the length and area are unchanged. so 1£, = 12 sin t(2 + cos t)/(5 + 4 cos t)5/2. This vanishes at only two points
3.2 1(t') = 1(t) { = } cost' = cost and sint' = sint { = } t' - t is a multiple of of the curve, where t = 0 and t = 11'.
211', so 1 is simple closed with period 211'. Taking x = acost,y = bsint in
!
Eq. (3) gives the area as J;' ab dt = 1I'ab.
3.3 "y = (- sin t - 2 sin 2t, cos t + 2 cos 2t) so II "Y II = ';5 + 4 cos t, which is never
zero, so 1 is regular. 1(t + 211') = 1(t) is obvious. If 1 is simple closed with Chapter 4
period a, then 1(0) =1(a) implies (1 + 2 cos a) cos a = 3 so cos a = 1 and
a = 211'. But 1(211'/3) = 1(411'/3) and 411'/3 - 211'/3 is not a multiple of 211'. 4.1 LetUbeanopendiscinR2andS={(x,y,z)ER31 (x,y)EU, z=O}.
3.4 Differentiating 1(t + a) = 1(t) gives t(t + a) = t(a); rotating anti-clockwise If W = {(x, Y, z) E R 3 I (x, y) E U}, then W is an open subset of R 3,
by 11' /2 gives n, (Ha) = n,(t); differentiating again gives 1<,(Ha)n,(Ha) = and S n W is homeomorphic to U by (x, Y, 0) >-+ (x, y). So S is a surface.
I<,(t)n,(t), so I<,(t + a) = I<,(t). 4.2 Let U = {(u,v) E R 2 10 < u 2 +v 2 < 11'2}, let r = ';U 2 +V 2, and define
3.5 By Exercise 3.2 and the isoperimetric inequality, the length l of the ellipse t7: U -t R 3 by t7(u, v) = (~, ~,tan(r - ~)).
satisfies l > ';411' x 1I'ab = 211'vIab, with equality if and only if the ellipse 4.3 The image of t7l is the intersection of the sphere with the open set ±x > 0
is a circle, Le. a = b. But parametrising the ellipse as in Exercise 3.2, its in R 3, and its inverse is the projection (x, Y, z) >-+ (Y, z). Similarly for t7~
length is and t7±. A point of the sphere not in the image of any of the six patches
{2' {2' would have to have x, y and z all zero, which is impossible.
J II "Y II dt = J v' a2 sin 2 t + b2 cos2 t dt. 4.4 Multiplying the two equations gives (X 2_Z 2) sin 0 cos 0 = (1_ y2) sin 0 cos 0,
o o
so x 2 +y2 - z2 = 1 unless cosO = 0 or sinO = 0; if cosO = 0, then x = -z
3.6 Let (XI,YI) and (X2,Y2) be points in the interior of the ellipse, so that
and y = 1 and if sin 0 = 0 then x = z and y = -1, and both of these lines
4 ¥t
+ < 1 for i = 1,2. A point of the line segment joining the two points are also contained in the surface.
i~ (tx: + (1- t)X2, tYI + (1- t)Y2) for some 0:0; t :0; 1. This is in the interior
 4
290 Elementary Differential Geometry Solutions 291

The given line L. p<)Sses through (sin 20, - cos 20, 0) and is parallel to the the open set U = {(u,v) E R 2 I u 2 _v 2 > O}. This is a reparametrisation
vector (cos 21J, sin 21J, 1); it follows that we get all of the lines by taking of tT via the repararnetrisation map (u, v) >-t (r, 0), where r = v'u 2 - v 2 ,
o < 0 < 7r. Let (x, y, z) be a poin~ of the surface; if x l' z, let 0 be such 0= cosh-l(ulv'u 2 v 2) (this is smooth because u 2 - v 2 > 0).
th;;:t cotlJ = (1 - y)/(x - z); ther! (x,y,z) is on L.; similarly if xl' -z. For the part with z < 0, we can take IT(r,lJ) = (rsinhlJ,rcoshO,-r2 )
The only remaining cases are the points (0,0, ±1), which lie on the lines defined on the open set R; and a(u, v) = (u, v, u 2 - v 2 ) defined on the
L~/2 and L o. open set V = {(u,v) E R 2 I u 2 - v 2 < O}.
3
With the notation of Exercise 4.2, define tT : U .... R by tT(u, v) = 4.9 This is similar to Example 4.5, but using the 'latitude longitude' patch
(sin 21J, - cos 21J, 0) + t(cos 21J, sin 21J, 1), where t = tan(r - ;;), cos IJ = ulr 1T(0,<p) = (acoslJcos<p,bcosOsin<p,csinO). Alternatively, one can use
and sin IJ = vir. This is a surface patch which, by the preceding paragraph, Theorem 4.1, noting that if I(x, y, z) = ~+ ~ +~ -1, then (I., I y , Iz) =
covers the whole surface. (2xla 2, 2yIb 2, 2zlc2) vanishes only at the origin, and hence at no point of
Let M ~ be the line the ellipsoid.
(x-z)cos<p=(I+y)sin<p, (x +z)sin<p = (l-y)cos<p. 4.10 A typical point on the circle C has coordinates (a + bcoslJ,O,bsinlJ); ro-
tating this about the z-axis through an angle <p gives the point tT(O, <p);
By the same argument as above, M~ is contained in the surface and the whole of the torus is covered by the four patches obtained by tak-
every point of the surface lies on some M~ with 0 :5 <p < 7r. If 0 + ing (0, <p) to lie in one of the following open sets: (i) 0 < 0 < 27r,0 <
<p is not a multiple of 1r, the lines L. and M~ intersect in the point 'I' < 27r, (ii) 0 < 0 < 27r, -7r < <p < 1r, (iii) -7r < 0 < 7r,0 < <p < 27r,
c", .-~ .in .-~ co, 9+ ). for each IJ with 0 < IJ < 7r, there is exactly
( sin(8+r.p 1 sin 8+11' 'sin(8+lf' J - • (iv) -7r < 0 < 1r, -1r < <p < 7r. Each patch is regular because
one <p with 0 :5 <p < 1r such that 0 + <p is a multiple of 7r, and the hnes L. o
IT. x IT~ = -b(a + bcoslJ)(cos cos <p, cos IJ sin <p, 1) is never zero (since
and M~ do not intersect. If (x, y, z) lies on both L. and L~, with 0 l' <p, a + bcoslJ 2:: a - b > 0).
then (1 - y) tanO = (1 - y) tan<p and (1 + y) cotO = (1 - y) cot <p, which Let 1T(0,<p) = (x,y,z). Then,x 2 +y2 +z2+ a2 _b 2 = 2a(a+bcoslJ), so
gives both y = 1 and y = -1 (the case in which 0 = 0 and <p = 7r 12, or (x 2+y2+ z 2+ a2_b2)2 = 4a2(a+bcoslJ)2 = 4a 2(x 2+y2). Let I(x,y,z) be
vice versa, has to be treated separately, but the conclusion is the same). the left-hand side minus the right-hand side; then, I. = 4x(x 2 + y2 + z2_
This shows that L. and L~ do not intersect; similarly, M. and M ~ do not a2 _ b2), I y = 4y(x 2 + y2 + Z2 _ a2 - b2), Iz = 4z(x 2 + y2 + z2 +a2 - b2); if
intersect. 3 Iz = 0 then z = 0 since x 2 + y2 + z' + a' - b2 > 0 everywhere on the torus;
4.5 If the sphere S could be covered by a single surface patch tT : U -+ R , then if I. = I y = 0 too, then since the origin is not on the torus, we must have
2
S would be homeomorphic to the open subset U of R . As S is a closed x 2 + y2 = a2 + b2, but then substituting into the equation of the torus
3
and bounded subset of R , it is compact. Hence, U would be compact, gives (2a 2)2 = 4a2(a 2+b2), a contradiction. So (I., I y , Iz) is nowhere zero
and hence closed. But, since R 2 is connected, the only non-empty subset on the torus, which is therefore a smooth surface by Theorem 4.l.
of R 2 that is both open and closed is R 2 itself, and this is not compact 4.11 If S is covered by a single surface patch tT : U .... R 3 , then I : S .... R is
as it is not bounded. smooth if and only if I 0 tT : U -+ R is smooth. We must check that, if
4.6 tT is obviously smooth and tTu X tTv = (- lu, - Iv, 1) is nowhere zero, so tT a : fj .... R 3 is another patch covering S, then loa is smooth if and only
is regular. if I 0 tT is smooth. This is true because loa = (10 tTl o!P, where !P is the
2 2
4.7 tT~ is a special case of Exercise 4.6, with I = ±v'1 - u 2 - v 2 (v'1 u v transition map from tT to a, and both !P and !p-l are smooth.
is smooth because 1 - u 2 - v 2 > 0 if (u, v) E U); similarly for the other 4.12 If iT = (1 + 8, where a is a constant vector 1 then is is smooth if IT is smooth,
patches. The transition map from tT'i- to tT~, for example, is !p(u, v) ~ and iTu = t1 u ,av = (1v, so i1 is regular if tr is regular.
(u,v), where tT~(u,v) = tT'i-(u,v); so u = v'1- u 2 - v 2, v = v, and thIS IS If a = A 0 IT, where A is a linear transformation of R" then a is smooth
smooth since 1 - u2 - v2 > 0 if (u, v) E U.
2
if tT is smooth, and au = A(lTu),a v = A(lT v ), so, if A is invertible, au and
4.8 tT is a smooth map on the open set R = {(r,lJ) E R I r > O}, and a v are linearly independent if lTu and tTv are linearly independent.
2 2
its image is contained in the surface because cosh 0 - sinh 0 = 1; and 4.13 (i) At (1,1,0), lT u = (1,0,2), lTv = (0,1, -2) so tT u x lTv = (-2,2,1) and
tT r X tT. = (-2r 2 coshO,2r 2 sinhO,r), which is never zero on R. the tangent plane is - 2x + 2y + z = O.
Exercise 4.6 gives the pararnetrisation a(u, v) = (u, v, u 2 - v 2 ), defined on (ii) At (1,0,1), wherer = 1,1J = O,tTr = (1,0,2),1T. = (0,1,2) so tT r XtT. =
292 Elementary Differential Geometry 4 Solutions 293

(-2, -2, 1) and the equation of the tangent plane is -2x - 2y + z = 0. dv/du = iJ/u = ±tana, and so is a straight line v = ±utana + c, where
4.14 Suppose the centre of the propeller is initially at the origin. At time t, the c is a constant.
centre is at (0,0, at) where a is the speed of the aeroplane. If the propeller 4.21 The point at a distance v from the z-axis on the ruling through (0,0, u)
is initially along the x-axis, the point initially at (v, 0, 0) is therefore at has position vector given by a(u,v) = (O,O,u) +v(cosO(u),sinO(u),O) =
the point (v cos wt, v sin wt, at) at'time t, where w is the angular velocity (vcosO(u),vsinO(u),u); a. x a v = (-sinO,-cosO,-vdO/du) is clearly
never zero, so (1 is regular.
of the propeller. Let u = wt, A = a/w.
Uu = (-vsinu,vcosu,'x), (Tv = (cosu,sinu,O), so the standard unit nor- °
4.22 (i) ";'a = so i is contained in the plane perpendicular to a and passing
through the origin; (ii) simple algebra; (iii) ii is clearly a smooth function
mal is N = (A' +v')-l/'(-Asinu,AcosU,-v). If 0 is the angle between
N and the z-axis, cosO =
-V/(A' +v')'/' and hence cotO ±V/A, while
the distance from the z-axis is v.
= of (u, v) and the jacobian matrix of the map (u, v) >-t (u, ii) is ("I1a ~) ,
where a dot denotes d/du; this matrix is invertible so i is a reparametri-
4.15 Let iT(fj, ii) he a reparametrisation of a. Then,
salion of 'Y.
{)fj _
au = au au
{)ii _
+ au tl v , fl v
{)fj _ {)ii _
= av Uti + av (Iv,
so Uu and (Tv are linear combinations of iT u and lTv> Hence, any linear
combination of (11). and U v is a linear combination of Uu and avo The
(iii) (u,v'±V1 + *
4.23 (i) (acosucosv,bcosusinv,csinu) (ef. Exercise 4.9); (ii) see Exercise 4.4;
+ ~); (iv), (v) see Exercise 4.6; (vi) see Example 4.3;
(vii) (pcosu,qcosu,v); (viii) (±pcoshu,qsinhu,v); (ix) (U,u 2/p2,v); (x)
(0, u, v); (xi) (±p, u, v).
converse is also true since tT is a reparametrisation of iT.
4.24 (a) Types (vii)-(xi); (b) type (vi); (c) types (ii) (see Exercise 4.4), (v) (see
4.16 If'Y(t) = (x(t), y(t), zit)) is a curve in S, differentiating f(x(t), y(t), zit)) =
° gives f.x+ fyy+ f.ft = 0, i.e. V f."I = 0, showing that V f is perpendicular
to the tangent vector of every curve in S, and hence to the tangent plane
Exercise 4.25) and (vi)-(x); (d) type (i) if p2, q2 and T 2 are not distinct,
types (ii), (iii), (iv), (vi) and (vii) if p2 = q2, and type (x).
of S. Since S has a canonical (smooth) choice of unit normal V f / II V f II 4.25 z = (~- ~) (~ + ~) = uv, x = ~p(u + v), Y = ~q(v - u), so
at each point, it is orientable. a parametrisation is a(u,v) = (~p(u + v), ~q(v - u),uv); a. x a v =
4.17 See the proof of Proposition 11.1 for the first part. By the argument in (-~q(u +v), ~p(v - u),pq) is never zero so a is regular. For a fixed value
Exercise 4.16, V F."I(O) = !,F('Y(t))lt=o, so V sF - V F is perpendicular to of u, a( u, v) = (~pu, -~qu, 0) + v(~p, ~q, u) is a straight line; similarly for
the tangent plane of S at P. This proves that V sF is the perpendicular a fixed value of v; hence the hyperbolic paraboloid is the union of each of
projection of V F onto the tangent plane. If F has a local maximum or two families of straight lines.
minimum at P, then t >-t F('Y(t)) has a local maximum or minimum at 4.26 Substituting the components (x, y, z) of 'Y(t) = a + ht into the equation
°
t = for all curves '1 on S with '1(0) = P; hence, V sF = 0 at P, so V F
is perpendicular to the tangent plane of S at P, and hence is parallel to
of the quadric gives a quadratic equation for t; if the quadric contains
three points on the line, this quadratic equation has three roots, hence is
Vf by Exercise 4.16. SO VF = AVf for some scalar A. identically zero, so the quadric contains the whole line.
4.18 From Example 4.13, the surface can be parametrised by a(u,v) = Take three points on each of the given lines; substituting the coordinates
(coshucosv,coshusinv,u), with u E R and -7r < V < 11'" or 0 < v < 27r. of these nine points into the equation of the quadric gives a system of
4.19 II a(u, v) 11'= secb2 u(cos 2 v+sin2 v)+tanh 2 u = sech2 u+ tanh 2 u = 1, soa nine homogeneous linear equations for the ten coefficients al, . .. , c of the
parametrises part of the unit sphere; t1 is clearly smoothj and tl u X U v = quadric; such a system always has a non-trivial solution. By the first part,
-sech 2 uO'(u,v) is never zero, so 0' is regular. Meridians correspond to the the resulting quadric contains all three lines.
parameter curves v = constant, and parallels to the curves u = constant. 4.27 Let L L 2 , L 3 be three lines from the first family; by Exercise 4.26, there
4.20 The vector 0' u is tangent to the meridians, so a unit-speed curve 'Y is a loxo- "
is a quadric Q containing all three lines; all but finitely many lines of
drome if "I.a./ II a. II = cos a, which gives u = cosh u cos a; since 'Y is unit- the second family intersect each of the three lines; if L' is such a line, Q
speed, 'Y = (-u sech u tanh u cos v - v sech u sin v, -it sech u tanh u sin v + contains three points of L', and hence the whole of L' by Exercise 4.26;
vsechucosv,itsech2 u) is a unit vector; this gives il. 2 + iJ2 = cosh 2 u, so so Q contains all but finitely many lines of the second family; since any
V = ± cosh u sin Ct. The correponding curve in the uv-plane is given by quadric is a closed subset of R 3 , Q must contain all the lines of the second
 4
294 Elementary Differential Geometry Solutions 295

family, and hence must contain S. Chapter 5

4.28 Both parts are geometrically obvious.
4.29 Let (a, b, c) be a point of R 3 with a and b non-zero. Then Ft(a, b, c) --+ 00 5.1 (i) Quadric cone; we have tT u = (cosh u sinh v, cosh u cosh v, cosh u), tTv =
as t --+ 00 and as t approaches p2/and q2 from the left; and Ft (a, b, c) --+ (sinhucoshv,sinhu sinhv, OL and so we get II (lu W= 2cosh2 u, tTu.u" ==
-00 as t --+ -00 and as t approaches p2 and q2 from the right. From 2 sinh u cosh u sinh v cosh v, II U" 11 2 = sinh 2 u, and the first fundamental
this and the fact that Ft(a, b, c) = 0 is equivalent to a cubic equation form is 2 cosh 2 U du 2 + 4 sinh u cosh u sinh v cosh v dudv + sinh 2 u dv 2 •
for t, it follows that there exist unique numbers u, v, w with u < p2, (ii) Paraboloid of revolution; (2 + 4u 2 ) du 2 + 8uv dudv + (2 + 4v 2 ) dv 2 .
p2 < V < q2 and q2 < w such that Ft(a, b, c) = 0 when t = u, v or w. The (iii) Hyperbolic cylinder; (cosh 2 u + sinh 2 u) du 2 + dv 2 •
surfaces Fu(x,y,z) = 0 and Fw(x,y,z) = 0 are elliptic paraboloids and (iv) Paraboloid of revolution; (1 + 4u 2 ) du 2 + 8uv dudv + (1 + 4v 2 ) dv 2 ,
Fv(x,y,z) = 0 is a hyperbolic paraboloid, and we have shown that there 5.2 The first fundamental form is 2 du 2 + u 2 dv 2 , so the length of the curve is
is one surface of each type passing through each point (a, b, c). f o" (2it 2 + u 2 V2 )1/2 dt = fo" (2,X2 e 2At + e 2At )I/2 dt = (2),Z"tl)'/2 (e>.t - 1).
To parametrise these surfaces, write Ft(x,y,z) = 0 as the cubic equation
X2(q2 _t)+y2(p2 -t) -2Z(p2 -t)(q2 -t) +t(p2 _t)(q2 -t) = 0, and note that
the left-hand side must be equal to (t - u)(t - v)(t - w)j putting t = p2, q2
2
and then equating coefficients of t (say) gives x = ±J(pZ-u)(::~;i(pZ-w),
y = ±J(92-u)~:~;J(qLW), z = ~(u + v +W _ p2 _ q2).
4.30 Let F : W --+ V be the smooth bijective map with smooth inverse F- 1 :
V --+ W constructed in the proof of Proposition 4.1. Then, (u(t), vet»~ =
F- 1 ('Y(t» is smooth.
4.31 Suppose, for example, that fy f:. 0 at (xo, Yo). Let F(x, y) = (x, f(x, y»;
then F is smooth and its jacobian matrix (~ ~:) is invertible at
(xo,Yo)· By the inverse function theorem, F has a smooth inverse G de-
5.3 Applying a translation to a surface patch U does not change U11. or U v ' If
fined on an open subset ofR2 containing F(xo, Yo) = (xo,O), and G must
A is a rotation about the origin, (Au)" = A(uu), (Au)" = A(u,,), and A
be of the form G (x, z) = (x, g( x, z» for some smooth function g, Then
'Y(t) = (t, get, 0» is a parametrisation of the level curve f(x, y) = 0 con-
preserves dot products (A(p).A(q) =
p.q for all vectors p,q E a 3 ).
5 ,4 By the ch am· rue,1 (1il
- -- uU8u 8u +u"8iil
8" -
(1" -- (lu8;,
811. 8" wh'ICh gIves
+u"8;,' .
taining (xo,Yo).
E = O'u.O'u = E (~~)2 +2F~ ~~ +G (~~( Similar expressions for P and
The matrix (Ixgx fy
gy
fz) has rank 2 everywhere; suppose that, at some
9z G can be found; multiplying out the matrices shows that these formulas
are equivalent to the matrix equation in the question.
point (xo,yo,zo) on the level curve, the 2 x 2 submatrix (f v fz) is
gy gz 5.5 Themapis(1(u,v)f--t (UV2cos~,uV2sin~,O)= O'(u,v),say, Theim-

:;~::~~::e:.:~' '(lE<r'r)":::::::,::' :::o~:. :0:°:

age ofthis map is the sector of the xy-plane whose polar coordinates (r,O)
satisfy 0 < () < 7fV2. The first fundamental form of u is 2 du 2 +u 2 dv 2 (Ex-
ercise 5.2); Uu = (V2cos ~, V2sin ~,o), 0'" = (-usin ~, ucos ~,o),
gx 9y 9z
G(x,u,v) = (x,<p(x,u,v),1J}(x,u,v» be the smooth inverse of F defined so 11 O'u 11 = 2, (1u.U" == 0, II (1" 11 = u and the first fundamental form of
2 2 2

near (xo, 0, 0). Then 'Y(t) = (t, <pet, 0, 0), '¢(t, 0,0» is a parametrisation of U is also 2 du 2 + u 2 dv 2 •
the level curve f(x,y,z) = g(x,y,z) = 0 containing (xo,yo,zo). 5.6 No: the part of the ruling (t, 0, t) with 1 ~ t ~ 2 (say) has length V2 and
is mapped to the straight line segment (t, 0, 0) with 1 :s t :s 2, which has
length 1.
5.7 For the generalised cylinder, Example 5.3 shows that the first fundamental
 ¥
296 Elementary Differential Geometry Solutions 297

form is du 2 + dv 2 ,
so a(u, v) H (u, v, 0) is an isometry from the cylinder This is less than the area 211" of the hemisphere.
to part of the xv-plane. For the generalised cone, Example 5.4 shows that 5.16 Parametrize the surface by u(u, v) '" (p(u) cosv, p(u) sin v, u(u)), where
the first fundamental form is v 2 rtu 2 + dv 2 • This is the same as the first 'Y(u) = (p(u), 0, u(u)). By Example 6.2, the first fundamental form is
fundamental form of it (;:7;, uv'Z), where it is as in Exercise 5.5. Since it du 2 + p(U)2 dv 2 , so the area is 11 p(u) dudv = 2rr 1 p(u) duo
parametrises part of the plane, the generalised cone is isometric to part (i) Take p(u) = casu, u(u) = sinu, with -rr/2 S; u S; rr/2; so
of the plane. 2rr I~~~2 cos u du = 4rr is the area.
5.8 A straightforward calculation shows that the first fundamental form of u' (ii) For the torus, the profile curve is 'Y(O) = (a + bcosO,O,bsinO),
is cosh2 u(du 2 +dv 2); in particular, it is independent of t. Hence, u(u, v) >4 but this is not. unit-speed; a unit-speed reparametrisation is i( u) =
2
u'(u, v) is an isometry for all t. Taking t = rr/2 gives the isometry from (a + bcos *,O,bsin *) with 0 S; u S; 2rrb. So 2rr 10 ., (a + bcos *) du =
the catenoid to the helicoid; under this map, the parallels u = constant 4rr 2 ab is the area.
on the catenoid go to circular helices on the helicoid, and the meridians 5.17 u is the tube swept out by a circle of radius a in a plane perpendicular to
v = constant go to the rulings of the helicoid. 'Yas its centre moves along 'Y. u, = (1- I<acosO)t - rasinOn+racosOb,
5.9 If the first fundamental forms of two surfaces are equal, they are certainly Ue = -asinOn+acosOb,givingu,xuo = -a(l-I<acosO)(cosOn+sinOb);
proportional, so any isometry is a conformal map. Stereogtaphic projec- this is never zero since IW < 1 implies that 1 - I<a cos 0 > 0 for all O.
tion is a conformal map from the unit sphere to the plane, but it is not The first fundamental form is ((1 - I<acosO)2 + r 2a 2) ds 2 + 2ra' dsdO +
an isometry since ,\ # 1 (see Example 5.7). 1::
a' d0 2 , so the area is I;~ a(l- I<a cos O)dsdO = ha(Sl - so).
5.10 The vector u. is tangent to the rulings of the cone, so the angle 0 between 5.18 If E , = E"F, = F 2 ,G , = G 2 , then E , G , - Ff = E 2G 2 - Fl, so any
the curve and the rulings is given by cosO", ,,(.uu/ll "( IIII UU II· i.From isometry is equiareal. The map I in Archimedes's theorem is equiareal
'Y(t) = (e"'cost,e"'sint,e At ) and Uu '" (cost,sint,l) at 'Y(t), we get but not an isometry (as E , # E 2 , for example).
cos 0 = .)2,\2/(2'\' + 1), which is independent of t. 5.19 If E , '" '\E2 ,F, = ,\F2 ,G , '" ,\G2, and if E , G , - Ff = E 2G, - Fl, then
5.11 The first fundamental form is sech2u(du' + dv 2). ,\2 = 1 so ,\ = 1 (since ,\ > 0).
5.12 The first fundamental form is (1 + P)du 2 + u 2 dv 2 , where a dot denotes 5.20 By Theorem 5.5, the sum of the angles of the triangle is rr + AI R 2 ,
dldu. So u is conformal if and only if j = ±vu2 - 1, I.e. if and only if where A is its area and R is the radius of the earth, and so is ~
!
I(u) = ±(!UVU2 - 1- cosh-' u) + c, where c is a constant. rr + (7500000)/(6500)2 = rr + 136°9 radians. Hence, at least one angle
5.13 The first fundamental form is (1 + 2v"(.6 + v'6.6)du2 + 2"(.6 dudv + dv 2. So of the triangle must be at least one third of this, i.e. rr + "6°9 radians.
u is conformal if and only if 1 + 2v"(.6 + v 26.6 = 1 and "(.6 = 0 for all u, v; 5.21 3F '" 2E because every face has three edges and every edge is an edge
the first condition gives 6 = 0, so 6 is constant, and the second condition of exactly two faces. The sum of the angles around any vertex is 2rr, so
then says that 'Y.6 is constant, say equal to d. Thus, u is conformal if and the sum of the angles of all the triangles is 2rr V; on the other hand, by
only if 6 is constant and 'Y is contained in a plane r.6 = d. In this case, u Theorem 5.5) the sum of the angles of any triangle is 11" plus its area, so
is a generalised cylinder. since there are F triangles and the sum of all their areas is 4rr (the area
5.14 u is conformal if and only if I~ + g~ = f; + g; and lui. + gug. = O. Let of the sphere), the sum of all the angles is rrF + 4rr. Hence, 2V = F + 4.
z = fu +igU1 W = Iv +igv ; then tT is conformal if and only if zz = ww and Then, V - E + F = 2 + !F - E + F = 2 + !(3F - 2E) = 2.
ziiJ + zw =0, where the bar denotes complex conjugate; if z 0, then= 5.22 Let u : U -t R 3 . Then, f is equiareal if and only if
w = 0 and ail four equations are certainly satisfied; if z # 0, the equations
give Z2 = -W 2 , SO Z = ±iwj these are easily seen to be equivalent to
J£ (E, G, - Ff)'/2 dudv = J£ (E 2G 2 - FiJ1/2 dudv
the Cauchy-Riemann equations if the sign is +, and to the 'anti-Cauchy- for all regions R <; U. This holds if and only if the two integrands are
Riemann' equations if the sign is -. equal everywhere, i.e. if and only if E , G , - Ff = E 2 G 2 - Fl·
5.15 Parametrise the paraboloid by u(u, v) = (u, v, u' +v 2), its first fundamen-
tal form is (1 + 4u2)du 2 + 8uv dudv + (1 + 4v 2)dv 2. Hence, the required
area is 11
.)1 + 4(u2 + v 2) dudv, taken over the disc u 2 + v 2 < 1. Let
u = r sinO, v = rcosO; then the area is 2rr 1;
vI + 4r 2rdr = H5 3 / 2 - 1).
 Solutions 299
298 Elementary Differential Geometry

The angle 1jJ in Eq. (8) is therefore equal to 8 or 1f - 8 so I<g = ± ~ sin 8 =

Chapter 6 ±* tan8. Note that this is zero if aod only if the parallel is a great circle.
6.9 The unit normal is N = (-gcosv, -g sinv, j), where a dot denotes d/du.
6.1 tTu = (1,0, 2u), tT v = (0,1, 2v), sc:i N = A(-2u, -2v, 1), where A = (1 +
On a meridian v = constant, we can use u as the parameter; since t1 u =
4u' +4v')-I/'; tT uu = (0,0, 2), tTu~ = 0, tT vv = (0,0,2), so L = 2A, M = 0,
ejcosv,jsinv,iJ) is a unit vector, u is a unit-speed parameter on the
N = 2A, and the second fundamental form is 2A(du' + dv').
meridiao aod
6.2 tTu.N u = -tTuu.N (since tTu.N = 0), so Nu.tTu = 0; similarly, Nu.tT v =
Nv.u u = Nv'u v = 0; hence, N u and N v are perpendicular to both Uu jcosv jsinv 9
and tT v , and so are parallel to N. On the other hand, N u and N v are I<g=<1 uu .(NxtTu )= -gcosv -gsinv j =0.
perpendicular to N since N is a unit vector. Thus, N u = Nt) = 0, and jcosv jsinv 9
hence N is constaot. Then, (tT.N)u = tTu.N = 0, aod similarly (tT.N)v = 0, On a parallel u = constant, we can use vasa parameter, but iT v =
so tT.N is constaot, say equal to d, aod then tT is part of the plaoe r.N = d. (-f sin v, f cos v, 0) is not a unit vector; the arc-length s is given by
6.3 From Section 4.3, N = ±N, the sign being that of det(J). From = u" ds/dv = II tT v II = f(u), so s = f(u)v (we can take the arbitrary con-
8u+
tT u8u
8v-_ 8u+
C1 v 8fi 1 D'fj - t1 u8v
8v
f1'v 8ii 1 we ge
t stant to be zero). Then,
- fcosv - f sin v 0
f(~),tTvv. (N x
_ cPu a'v (au)' auav (av)' 1 j
tTiiu = U u [Jij,2 + Cl v [JfL2 + tl uu ail + 2uuv au au + D'vv 8u 1<9 = f!utv) -iJ cos v -iJ sin v j = f
f(u)3
- f sin v fcosv
So
6.10 Itl = ItN1.n, 1t2 = K.N 2 .n, so °
_
L=± ((au)' auav +N (av)')
L au +2M auau au I<IN, - I<,N I = I<((NI.n)N, - (N,.n)N I ) = I«N I x N,) x n.
Taking the squared length of each side, we get
since tTu.N = tTv.N = O. This, together with similar formulas for M aod
N, are equivalent to the matrix equation in the question. I<i + I<~ - 21<11<,N I .N, = 1<' II (N I x N,) x n 11 2 .

6.4 Applying a traoslation to a surface patch tT does not chaoge tTu aod tT v , aod Now, NI.N, = cosa; 'I is perpendicular to N I and N" so N I X N, is
hence does not change N'O'UU1D'utJ or tTVV1 and hence does not change the
parallel tot, hence perpendicular to n; hence,
second fundamental form. A rotation A about the origin has the following
effect: tTu --t A(tTu ), tTu --t A(tT v ) and hence N --t A(N), tT uu --t A(tTuu ), II (N I x N,) x n11=11 N I x N,IIII n II = sina.
tT uv --t A(tT uv ), tT vv --t A(tT vv ); since A(p).A(q) = p.q for aoy vectors 6.11 N.n = cos1jJ, N.t = 0, so N.b = sin1jJ; hence, N = ncos1jJ + bsin1jJ and
p, q E R 3 , the second fundamental form is again unchaoged. B = t x (n cos 1jJ + b sin 1jJ) = bcos1jJ - nsin1jJ. Hence,
6.5 By Exercise 6.1, the second fundamental form of the paraboloid is 2(du' +
dv')/";1 + 4u' + 4v'; so N= ncos1jJ + bsin 1jJ +,j,( -nsin 1jJ + b cos1jJ)
= (-I<t + Tb) cos1jJ - nTsin1jJ + ,j,(-n sin 1jJ + bcos1jJ)
= 2«( - sin t)' + cos' t)/Vl + 4 cos' t + 4 sin' t = 2/..;5.
I<n
= -I< cos 1jJt + (T + ,j,)(bcos1jJ - nsin1jJ)
6.6 1<' = I<~ + 1<; = 0, so I< = 0 and "( is part of a straight line.
= -I<nt + TgB.
6.7 Let"( be a unit-speed curve on the sphere of centre a and radius r. Then,
("( - a).("( - a) = r'; differentiating gives '1.("( - a) = 0, so 'Y-("( - a) = The formula for B is proved similarly. Since {t,N,B} is a right-handed
-'1.'1 = -1. At the point "(t), the unit normal of the sphere is N = orthonormal basis of R', Exercise 2.22 shows that the matrix expressing
±~("((t) - a), so I<n = 'Y.N = ±h·("( - a) = 'F~. i,N,B in terms oft,N,B is skew-symmetric, hence the formula for i.
6.8 If the sphere has radius R, the parallel with latitude 8 has radius r = 6.12 A straight line has a unit-speed parametrisation "(t) = p + qt (with q a
R cos 8; if P is a point of this circle, its principal normal at P is parallel to unit vector), so 'Y = 0 aod hence I<n = 'Y.N = O.
the line through P perpendicular to the z-axis, while the unit normal to In general, I<n = 0 {=> 'Y is perpendicular to N {=> N is perpendicular
the sphere is parallel to the line through P aod the centre of the sphere. to n {=> N is parallel to b (since N is perpendicular to t).
 ¥
Elementary Differential Geometry Solutions 301
300

6.13 The second fundamental form is (-du 2 +u 2 dv 2 )/uv'l+u , so a curve

2 6.18 ~ = uO'u + VtT v is a principal vector corresponding to the principal curva-
-y(t) = 2
D"(u(t),v(t)) is asymptotic if and only if _u + u iJ2
2
0, i.e. = ture K ~ (FII - KFI) (~) (~) ~ F I l FlI (:) = K, (~) ~
=
= vlu = ±l!u, so In u J ±( v + c), where c is a constant.
dv jdu
6.14 By Exercise 6.12, b is parallel to N, so b == ±N; then, B = t x N = Tn. (
ab c) (u)v
d = -K, (u)
V ~ au.,
+ cv = . andb' + d'v = . But,
-KU U -K,V.
Hence,:B = Til = T(-ld+7b) = ±~t-7N; comparing with the formula
N = iJ.Nu+vN v = u(aO'u +bO'v) +v(CO'u+dO'v) = (au+cv)O'u+(biJ.+dv)O'v.
for B in Exercise 6.11 shows that 79 = T (and K,n = ±K,).
Hence, ~ is principal '¢=> N = -K,(iJ.tT u + VtTv) = -K,~.
6.15 For the helicoid a(u,v) = (vcosu,vsinu,).u), the first and second funda-
From Example 6.2, the first and second fundamental forms of a surface
mental forms are ().2 + v 2)du2 + dv 2 and 2>"dudv/v'>..2 + v 2 , respectively.
of revolution are du 2 + J(u)2dv 2 and (jg - iiJ)du2 + fiJdv 2 , respectively.
Hence, the principal curvatures are the roots of Since the terms dudv are absent, the vectors tTu and O'v are principal; but
v'>,,!~v21 = 0,
2
\ _K,(>..2>.. + v ) these are tangent to the meridians and parallels, respectively.
6.19 By Exercise 6.11, N = -Knt + TgB, so by Exercise 6.18 'Y is a line of
~ r curvature if and only if 7 g = 0 (in which case A = K,n).
i.e. ±)./(>..2 + v ). 2

For the catenoid O'(u,v) = (coshucosv,coshusinv,u), the first and sec- 6.20 Let N l and N 2 be unit normals of the two surfaces; if "I is a unit-speed
ond fundamental forms are cosh 2 u(du 2 +dv 2) and -du 2+dv , respectively.
2 parametrisation of C, then N1 = - Al ~ for some scalar >'1 by Exercise
6.18. If C is a line of curvature of S2, then N 2 = -A2~ for some scalar ).2,
Hence, the principal curvatures are the roots of
2
and then (N l .N 2 )" = ->'l'Y.N2 - ).2~.Nl = 0, so N l .N2 is constant along
-1 - K, cosh U 0 \ "I, showing that the angle between 51 and S2 is constant. Conversely, if
\ o 2
1 - K, cosh U '
N 1 .N 2 is constant> then N l .N 2 = 0 since N1 .N2 = ->"l'Y.N2 = 0; thus,
2
i.e. K, = ±sech u. N2 is perpendicular to N l, and is also perpendicular to N 2 as N 2 is a
6.16 Let s be arc-length along "I, and denote d/ds by a dash. Then, by Propo- unit vector; but ~ is also perpendicular to N l and N 2; hence, N2 must
sition 6.1, be parallel to .y, so there is a scalar >'2 (say) such that N 2 = ->'21'.
6.21 (i) Differentiate the three equations in (21) with respect to w, u and v,
2 I I ,2 Lu 2 + 2Muv + NiP LiJ. 2 + 2Muv + NiP
K,n=::.Lu' + 2Mu v + Nv = (ds/dt)2 - EiJ. 2 + 2FiJ.v + Gv 2 . respectively; this gives

6.17 By Exercises 5.4 and 6.3, we have (in an obvious notation),:if = PFf J ,
PII = ±Jt FIIJ, where the sign is that of det(J). The principal curvatures Subtracting the second equation from the sum of the other two gives
t
of ir are the roots of det(:iII -K:if ) = 0, i.e. det(±Jt F1 d -KJ FIIJ) = 0, (fu'O'vw = 0, and similarly (fv.(Tuw = tTw·(Tuv = O.
which (since J is invertible) are the same as the roots of det(±,r1 [-K,r[) = (ii) Since O'v.(fw = 0, it follows that the matrix F[ for the u = Uo surface
O. Hence, the principal curvatures of ir are ± those of (f. is diagonal (and similarly for the others). Let N be the unit normal of the
Let tiT,. + ijir v be a principal vector for iT corresponding to the principal u = Uo surface; N is parallel to 0'" XO'w by definition, and hence to O'u since
curvature K. Then, O'u, O'v and O'w are perpendicular; by (i), O'vw.O'u = 0, hence O'vw. N = 0,
proving that the matrix F[ [ for the u = Uo surface is diagonal.
(iii) By part (ii), the parameter curves of each surface U =Uo are lines
of curvature. But the parameter curve v = va, say, on this surface is the
so if J (~) = (~). then ~(fu +71O'v is a principal vectorfor 0' correspond- curve of intersection of the u = Uo surface with the v = Va surface.
. h" 1 t B t' t au E+ fJu - - av i + av -
mg to t e prmclpa ~r:,!a&~re K,. 8~' smce_'" ~t'&u '" B~ 71, 71 ~ ~5L_'" ~~~'
we have ~(fu + 7JO'v - ~ (8ii(fu + fJuO'v) + 71 (~(fu + 8ii(fV) - ~O'u + 7JO'v,
which shows that tir" + fjir ii is a principal vector for (f corresponding to
the principal curvature K,.
The second part also follows from Corollary 6.2.
 ¥
302 Elementary Differential Geometry Solutions 303

6.22 We have N" :::: aa" + OO v , N v :::: CtT" + du v, so Chapter 7

FIll:::: (Nu.Nu N".Nvc)·- 7.1 a" :::: (1,I,v), a v :::: (l,-l,u), au" :::: a vv :::: 0, a uv :::: (0,0,1). When
Nu.N vNv·Nv
u = v :::: 1, we find from this that E :::: 3, F :::: 1, G :::: 3 and L ::::
_ ( Ea + 2Fab + Gb 2
2
Eae + F(ad + be) + Gbd)
N :::: a, M :::: -1/0. Hence, K :::: (LN '- M 2 )/(EG - P2) :::: -1/16,
- Eae + F(ad + be) + Gbd Ec2 + 2Fed + G~
H =(LG - 2MF + NG)/2(EG - F ) = 1/8.;2.
2

::::(~ ~)(; ~)(: ~)::::(-FilFIl)tFI(-Fi1FIl) 7.2 v) = (v

For the helicoid i'7(u, u' v cos sin u, AU), au = (-v sin u, v cos u, A),
O'v :::: (cosu,sinu,O), N :::: (A 2 + v 2)-1/2(-Asinu,Acosu,-v), O'uu =
::::FlIFi1 JIFil FIl ::::FlIFil FlI· (-vcosu,-vsinu,O), iTv.v = (-sinu,cosu,O), O'Vl1 = O. This gives
6.23 By Example 6.2, the principal curvatures are jg - j'iJ and iJl/. If g :::: 0, E = A +v ,F = O,G = land L:::: N = O,M = A/\/A 2 +v 2 • Hence,
2 2

the surface is part of a plane z :::: constant and no point is parabolic. Thus, K = (LN.;... M 2 )j(EG - P2) = _A2 /(A 2 + v 2 ?
9 i- a, and every point is parabolic if and only if jy - j'9 :::: O. Multiplying For the catenoid O'(u, v) :::: (coshucosv,coshusinv,u), we have O'u ::::
through by g and using j2 + g2 :::: 1 (which implies that j j + iJy :::: 0), we (sinh u cosv, sinh u sin v, 1), a v :::: (- cosh u sin v, cosh u cosv, 0),
get j' = O. Hence, /(u) = au + b, where a and b are constants. If a :::: 0, N :::: sech uC.... cosv, - sin v, sinh u), Crv.v. = (cosh U cos v, cosh u sin v, 0),
we have a circular cylinder; if a "I- 0, we have a circular cone. a uv :::: (- sinh U sin v, sinh u cosv, 0), (rvv = (- cosh '!.teas v, - cosh u sin v, 0).
6.24 On the part of the ellipsoid with z i- 0, we can use the parametrisation This gives E:::: G =cosh2 u,F = 0 and L = -I,M:::: O,N:::: 1. Hence,
K :::: (LN - M 2 )j(EG - F 2 ) = -sech 4 u.
O'(x, y) :::: (x, y, z), where z :::: ±rJ1 - ~ - ~. The first and second
Alternatively, use the results of Exercise 6.15.
fundamental forms are (1+z;)dx 2 +2zzzlldxdy+(1+z~)dy2 and (zxx dx2 + 7.3 Parametrise the surface by iT(u,v) = (u,v,/(u,v)). Then,"a u :::: (1,0,/,,),
2ZXydxdY+Zyydy2)1 J1 + z; + z~, respectively. By Proposition 5.3(ii), the O'v = (O,l,/v), N = (1 + /~ + f;)-1/2(_/u'-/1J,1), (Tv.u :::: (O,O,/uu),
condition for an umbilic is that FIl :::: K,Fl for some scalar K,. This leads O'"v:::: (O,O,/uv), O'vv =(O,O,/vv)· This gives E = 1 + /~,F:::: /u/v,c::::
to the equations 1 + /; and L :::: (1 + /~ + 1;)-1/2/uu , M = (1 +1~ + /;)-1/2/uv ,N ::::
(1 + /~ + {;)-1/21vv. Hence,
z"'x = A(l + z;), Z",y:::: AZ",Zy, zYII:::: A(l + z~),
K = /uu/vv - /~v
where A :::: K,J1 + z; + z~. If x and yare both non-zero, the middle equa- (1 +f~ + f';)2'
tion gives A :::: II z, and substituting into the first equation gives the con-
(i) From Example 7.3, !< = 0 ¢=>&.N =0 ¢=> &.(t + v6) x 6) :::: 0 ¢=>
tradiction p2 : : r
2 . Hence, either x:: : °or y :::: 0. If x:: : 0, the equations 7.4
= =
6.(t x 6) O. If 6 =n, 6 -K,t + Tb, t x 6 :::: h, so K =0 ¢=> T =0 ¢=>
have the four solutions "lis planar (by Proposition 2.4). If 6 = h, 6:::: -rn, t x 6:::: -n, so again

x~ 0, F ±qJ":-P:,
-rq
z~ ±rj': -~ r -q
°
K::::: ¢=> T = 0.
°
(ii) Let N the a unit 'normal of S. Then, K = ¢=> N 1. (t x N 1) = O. Since
N1 is perpendicular to Nl and N 1 is perpendicular to t, this condition
Similarly, one finds the following eight other candidates for umbilics:
holds ¢=> Nt is parallel to t, i.e. ¢=> N1 = -Ai' for some scalar A_ Now

XdPJ~-q:,
p -r
F±r/':-q:,
y=O,
r -p
use Exercise 6.18.
7.5 Using the parametrisation iT in Exercise 4.10, we find that E = b , F =
2

O,G:::: (a+ bCOSU)2 and L = b,M = O,N = (a + bcosu) cosu. This gives
x~ ±pJ'!:,-r:, q22
y = ±qV -r:, z 0. = K ::::Gosulb(a + bcosu), dAD- = (EG - F 2 )1/2dudv = b(a + bcosu)dudv.
lr-q q-p Hence,
Of these 12 points, exactly 4 are real, depending on the relative sizes of
p2 , q2 and r 2 . !! K dAD-:::: 1121r 21r

cos u dudv = 0.
304 ¥
Elementary Differential Geometry Solutions 305

7.6 The first part follows from Exercises 5.3 and 6.4. The dilation multiplies (v, w) is a reparametrisation map. The first fundamental form in terms of
E,F,G by a 2 and L,M,N by a, hence H by a-I and K bya- 2 (using v, w is given by
Proposition 7.1).
7.7 Since tT is smooth and (fu X (fu is never zero, N = (fu X uu/ II UU x UU II (~~)=Jt(; ~)J
is smooth. Hence, E,F,G,L,M and N are smooth. Since EG - F2 > °
(by the remark following Proposition 5.2), the formulas in Proposition
7.1(i) and (ii) show that H and K are smooth. By Proposition 7.1(iii),
= (_Oi ~) (~ f(~)2) (~ -o!) = (* ?), ~

the principal curvatures are smooth provided H 2 > K, i.e. provided there so the first fundamental form is (dv 2 + dw 2 )/w 2 .
are no umbilics. (ii) We find that the matrix
7.8 At a point P of an asymptotic curve, the normal curvature is zero. By
Corollary 6.2, one principal curvature "'1 ~ 0 and the other "'2 ~ O. Hence
j= (g~ ~) = ( v(w+l) Hv 2 -(W+l)2))
g~ &~ _~(v2 - (w + 1)2) v(w + 1) ,
K = "'1"'2 ~ O. On a ruled surface, there is an asymptotic curve, namely
so the first fundamental form in terms of U and V is given by
a straight line, passing through every point (see Exercise 6.12).
7.9 By Exercise 6.22, FIll = FIIFi 1 FIl. Multiplying on the left by :Fi 1 ,
the given equation is equivalent to
jt (~
o
0) j = (v
~
2
+ (w + 1)2)2 1=
4w 2
1 I
(1 - U2 _ V2)2 '

A2 +2HA+KI=0, after some tedious algebra.

In (i), u < 0 and -tr < v < rr corresponds to -rr < v < rr and w > 1, a
where A = -Fi 1 FIl = (~ ~). By the remarks following Definition semi-infinite rectangle in the upper half of the vw-plane.
6.1, the principal curvatures are the eigenvalues of -A. Hence, 2H =
sum of eigenvalues of -A = -(a + d), K = product of eigenvalues of
-A = ad - be. Now use the fact stated in the question.
7.10 By Eq. (9) in Chapter 6, -y/y = TtFlT; by Eq. (10) in Chapter 6,
N.-y = -N..:y (since N.-Y = 0) = -"'n = -TtFIlT. Now, N.N = (uN u +
=
vNu).(uNu+vN v ) (Nu.Nu)U2+2(Nu.Nv)uir+(Nv.Nu)V2 TtFIIlT. =
Hence, multiplying the equation in Exercise 7.9 on the left by Tt and on
the right by T gives N.N+2HN.-Y+K-Y.-Y = 0; J(.")' is an asymptotic curve,
"'n= 0 so N.-y = O. So, assuming that 'Y is unit-speed, we get N.N = -K.
But Exercise 6.12 gives N = ±b, so N = =f1'n and N.N = 1'2.
7.11 The parametrisation is tT(u, v) = (f(u) cosv, f(u) sin v, g(u», f(u) =e",
g(u) = v'1 - e 2u - cosh-\e- U ), -00 < u < O. To find the corresponding region in (ii), it is convenient to introduce the
(i) A parallel u = constant is a circle of radius f(u) = eU , so has length complex numbers z = v + iw, Z = U + iV. Then, the equations in (ii)
2rre" . are equivalent to Z = ~+~, z = i(~!ll)' The line v = rr in the vw-plane
(ii) From Example 7.2, E = 1, F = 0, G = f(U)2, so dAo- = f(u)dudv and correspond~ to z + Z = 2rr (the bar denoting complex conjugate), i.e.
the area isI:'" I~= e"dudv = 2rr. i(~!?l) - i(1!.11 ) = 2rr, which simplifies to IZ - (1- ~)12 = ~; so v = 7T
(iii) From Example 7.2, the principal curvatures are "'I
= jg - jt; corresponds to the circle in the UV -plane with centre 1 - i1f and radius .!.
1f •

-jig = _(e- 2u _1)-1/2, ~2 = fiJIP = g/f = (e- 2u _1)1/2. Similarly, v = -rr corresponds to the circle with centre 1 + i1f and radius
(iv) ~1 < 0, "'2 > O. ~. Finally, w = 1 corresponds to z - z = 2i, i.e. i(~!?1) + i(~!.11) = 2i.
7.12 (i) Setting ii. = v,v = w = e- u , we have u = -lnv,v = it so, in the This simplifies to IZ - tl 2 = !; so w = 1 corresponds to the circle with
notation of Exercise 5.4, J = (~ -}i). Since J is invertible, (u, v) I-t
centre 1/2 and radil.;ls 1/2 in the UV-plane. The required region in the
UV -plane is that bounded by these three circles:
 Elementary DifFerential Geometry
• Solutions 307
306
v
Proposition 7.4 gives the result.
7,15 See the proof of Proposition 9.5 for the first part.
The first fundamental form ofthe given surface patch is (1+u 2 +v 2)2 (du 2+
dv 2 ), so it is conformal, and U uu + t1 1111 = (-2u, 2v, 2) + (2u, -2v, -2) = O.
7.16 Parametrize the surface by t1(u, v) = (u, v, feu, v)). By Exercise 7.3, E =
1 + f~,F == fufll,G = 1 + f;,L == (1 + f~ + f;)-1/2 fuu,M == (1 + f~ +
!J)-1/2 fuv, N = (1 + f~ + f~)-1/2 11111' Hence,

H = LG - 2M F +NE = fuu(l + /~) - 2fullfufll + fllll{1 + f~)

2(EG - F2) 2(1 + I~ + J;)3/2
Taking f(u,v) = In (~~::) gives
2
H = sec2 u{l + tan2 v) - sec 2 v{1 + tan u) = 0
2(1 + tan 2 u + tan 2 v)3/2 .
7.17 E u = t1 u + wN u , E v = t1 11 + wN ll > E w == N. Eu·E w = since
t1 u .N = Nu.N = 0, and similarly E ll .Ew = O. Finally, 11".Ell = t1 U 'U ll +
°
2
7.13 Let "Y(u) == (f(u), O,g(u)) and denote d/du by a dot; by Eq. (2), I+K 1 == w(O'".N ll + D'l1. N U) +w 2 N".N ll = F - 2wM + w2 N u .N v = w N u .N ll ; by
O. If K < 0, the general solution is f = ae-.;=1?u + be.;=Ku where a, b Proposition 6.4, N u = -ft1 u , N ll ::::: -~t111' so N u .N ll = ~~F = O.
°
are constants; the condition f(7t /2) = f{ -7t /2) = forces a := b == 0,
so "Y coincides with the z-axis, contradicting the assumptions. If K = 0,
Every surface u = Uo (a constant) is ruled as it is the union of the straight
lines given by v = constant; by Exercise 7.4(ii), this surface is flat provided
f == a + bu and again a = b == 0 is forced. So we must have K > 0 and the curve "Y(v) = t1(uo, v) is a line of curvature of S, Le. if 0'11 is a prin-
f = acosVKu+bsinVKu. This time, f(7t/2) == f(-7t/2) == 0 and a,b cipal vector; but this is true since the matrices :F1 and :FII are diagonal.
not both zero implies that the determinant Similarly for the surfaces v = constant.
7.18 By Eq. (15), the area of t1(R) is
cosVK7t/2 sin../K7t/2 \=0.
\ cos../K7t/2 - sin ../K7t/2
This gives sin.JR7t = 0, so K = n for some integer n :j=. O. If n = 2k
2
/1 11 N u x N ll II dudv = / 11K!11 t1 u x U ll 11 dudv = /l1K1dAo-.
7.18 From the formula for K in the solution of Exercise 7.5, it follows that S+
is even, f = b sin 2ku, but then f{O) = 0, contradicting the assumptions.
and S- are the annular regions on the torus given by -7t /2 ::; u ::; 7t/2
If n = 2k + 1 is odd, f = acos(2k + l)u and f(7t/2(2k + 1)) = 0, which
contradicts the assumptions unless k = 0 or -1, Le. unless K = (2k+l)2 == and 7t /2 ::::; u ::::; 37t/2, respectively.

1. Thus, 1 == a cos u, 9 == J1 - j2 = vii - a2sin 2

(j, 0, g) is
u. Now, "y =
perpendicular to the z-axis {=}!J :::: 0. So the assumptions give ~ ==
0, i.e. a:::: ±1. Then, "Y(u) == (± cosu, 0, ± sinu) (up to a translation along
the z-axis) and S is the unit sphere.
7.14 Let 0'(1£, v) be a patch of S containing P = O'(uo, vo). The gaussian curva-
ture K of S is < 0 at P; since K is a smooth function of (u, v) (Exercise
7.7), K(1£, v) < 0 for (u, v) in some open set U containing (Ua, vo); then
every point of O'(U) is hyperbolic. Let 1\;1, 1'02 be the principal curvatures
v'
of 0', let 0 < () < 7t /2 be such that tan e = -1\;1/ 1"2, and let e1 and e2 be
the unit tangent vectors of 0' making angles () and -0, respectively, with S+ S-
the principal vector corresponding to K1 (see Corollary 6.1). Applying
 4 solutions 309
308 Elementary Differential Geometry

It is clear that as a point P moves over S+ (resp. S-), the unit normal at P 1<9t - T 9 N = ±b = 'fTn = -TN. Hence, T g = T (and 1<9 = 0, which we
covers the whole of the unit sphere. Hence, JJs + IKldA = JJs - IKldA = knew already).
411' by Exercise 7.18; since IKI = ±K on S±, this gives the result. 8.5 ;Y is parallel to II since 'Y lies in II, and ;Y is parallel to N since 'Y is a
geodesic; so N is parallel to II. It follows that N is also parallel to II.
Since N is perpendicular to N, and 'Y is also perpendicular to N and
parallel to II, N is parallel to 'Y. By Exercise 6.18, 'Y is a line of curvature
Chapter 8 of S.
8.6 If P and Q lie On the same parallel of the cylinder, there are exactly
8.1 By Exercise 4.4, there are two straight lines on the hyperboloid passing two geodesics joining them, namely the two circular arcs of the parallel
through (1,0,0); by Proposition 8.3, they are geodesics. The circle given of which P and Q are the endpoints. If P and Q are not On the same
by z = 0, x 2 + y2 = 1 and the hyperbola given by y = 0, x 2 - Z2 = 1 parallel, there are infinitely many circular helices joining P and Q (see
are both normal sections, hence geodesics by Proposition 8.4 (see also Example 8.7).
Proposition 8.5). 8.7 Take the COne to be O'(u,v) = (ucosv,usinv,u). By Exercise 5.5, 0' is
8.2 Let II, be the plane through 'Y(s) perpendicular to t(s); the parameter isometric to part ofthe xv-plane by O'(u, v) >-l (uv'2 cos ;.-, uv'2 sin ;'-,0).
curve s = constant is the intersection of the surface with IIs . From the By Corollary 8.2, the geodesics on the COne correspond to the straight lines
solution to Exercise 5.17, the standard unit normal of 0' is N = - (cos 8 n + in the xv-plane. Any such line, other than the axes x = 0 and y = 0, has
sin 0 b). Since this is perpendicular to t, the circles in question are normal equation ax + by = 1, where a J b are constants; this line corresponds to the
sections.
8.3 Take the ellipsoid to be
2
?'
2 2
f.,
+ ~ + ;.- = 1; the vector (-to, ;"0-) is normal
. curve v H- (cosv sinv 1).
'\I2(acos ·~72+b8in~) J V2(acos ~+bsin~)' J2'(acos ~+bsin./2) ,

to the ellipsoid by Exercise 4.16. If 'Y(t) = (f(t), g(t), h(t)) is a curve On the x and y-axes correspond to straight lines On the COne.
8.8 From Example 5.3, O'(u, v) = 'Y(u) + va, where 'Y is unit-speed, II a 11= 1,
the ellipsoid, R = (f; + ~ + ~)-1/2, S = (~+ ~ + ~)-1/2. Now, 'Y is a
and 'Y is contained in a plane perpendicular to a; the map O'(u, v) H
geodesic <=>;Y is parallel to the normal <=> (j,y,h) = A(f,., #t, f,-) for (u, v, 0) is an isometry from the cylinder to the xv-plane. A curve t H
some scalar A(t). From ~ +~ +~ = 1 we get ~ + ,. + ~ = 0, hence O'(u(t), v(t)) is a geodesic on the cylinder <=> t >-l (u(t), v(t), 0) is a
f;+ ~+~+ f;?+" +~ = 0, i.e. f;+~+~ +A (~+ ~ +~) = 0,
constant-speed parametrisation of a straight line in the plane <=> it
2 2 and iJ are constant <=> iJ is constant (since it 2 + iJ2 is constant). Since
which gives A = _5 / R • The curvature of'Y is
a.ftO'(u(t),v(t)) = iJ, iJ = constant <=> the tangent vector ftO'(u,v)

II ;Y II = (p + y2 + h 2)'/2 = IAI (~: + :: + ~: t 2

I~I = ~2' makes a fixed angle with the unit vector a parallel to the axis of the
cylinder.
8.9 Take the cylinder tobe O'(u,v) = (cosu,sinu,v). Then,E = G = 1,F = 0,
Finally,
so the geodesic equations are il = ij = O. Hence, u = a + bt, v = c + dt,
1 d (
"2 dt
1
R2 5 2
) = fj+ gg.
- - + hi.)
-
( p4 q4 r 4
(P
-+-+-
p' q2 r2
2
g2 i. ) where a, b, c, d are constants. If b = 0 this is a straight line on the cylinder;
otherwise, it is a circular helix.
2 8.10 E = 1, F = 0, G = 1 + u 2, so 'Y is unit-speed <=> ,,2 + (1 + U 2)iJ2 = 1.
+(1' +g2 +h ) (i/+gy+i.h) The second equation in (2) gives ft((l + u 2)iJ) = 0, i.e. iJ = 1:." where
p4 q4 r4 p2 q2 r2 ,
a is a constant. So ,,2 = 1 - (1~U') and, along the geodesic,
1 (fj gg hi.) A (fj gg hi.)
= R2 p4 + q4 + f4 + 52 p4 + q4 + f4 = 0, dv=~=± a
du" y'(1 - a2 + u 2)(1 + u 2)
since A = _5 2/ R 2 • Hence, R5 is constant.
If a = 0, then v = constant and we have a ruling. If a = 1, then dv/du =
8.4 If 'Y is a geodesic, ;Y = I<n is parallel to N (in the usual notation), so
±l/uv'l + u 2, which can be integrated to give v = Vo 'f sinh-1~, where
n = ±N. In the notation of Exercise 6.II, B = t x N = ±b, so II =
310 Elementary Differential Geometry
• Solutions 311

Vo is a constant. (ii) f preserves angles and takes meridians to meridians, so must take
8.11 We have parallels to parallels. Hence, w does not depend on v.
(au x av) x au (au.au)a" - (a".av)a u Ea v - Fat< (iii) The parallel w :::: constant has length 2rr/w by Exercise 7.11(i) (w ::::
N x a,,:::: II a" x lTv II:::: . vEG _ F2 :::: vEG _ F2' e- U ). As f preserves lengths, part (ii) implies that 2rr /w :::: 2rr /w, so
w::::w.
and similarly for N x a v - Now, i':::: ua" + va", so (iv) We now know that f(tT(u,v)) :::: u(F(v),w) for some smooth function
. u(Ea" - FiT,,) + v(F'CT v - Gau ) F(v). The first fundamental form of a(F(v), w) is w- 2 (( ~~) 2 dv 2 + dw 2 );
N x "" - --'--------=.--~::;;;=~;-'---------';:.:..
,- JEG-F2 '
since f is an isometry, this is equal to w- 2 (dv 2 +dw 2 ), hence dF/dv :::: ±l,
;Y:::: iUr" + va" + u a uu + 2uvauv + iJ2 avv .
2
so f(v) :::: ±v + n, where 0: is a constant. If the sign is +, f is rotation by
Hence, "'g :::: ;Y.(N xi') :::: (uv-vii)VEG- F2+Au 3 +Bu 2 iJ+Cuv 2 +Dil 3 , 0: around the z-axis; if the sign is -, f is reflection in the plane containing
where A :::: auu.(Eav - FiT.. ) :::: E«uu.iTv)" - au.auv ) - ~F(au.au)u :::: the z-axis making an angle 0:/2 with the xz-plane.
E(Fu - ~Ev) - ~FEu, with similar expressions for B,C,D. 8.17 From the solution to Exercise 7.12, Z :::: U + iV :::: ~+L where z ::::; v +
8.12 We have iw. Since the geodesics on the pseudosphere correspond to straight lines
and circles in the vw-plane which are perpendicular to the v-axis, they
(Eu 2 +2Fuv + 01/)'
correspond in the UV -plane to straight lines and circles perpendicular to
:::: (Euu + Evv) u 2 + 2(Fu u + Fvv)uiJ + (Guu + Gvv)v 2 the image of the v-axis under the transformation z I-t ~+:, i.e. the unit
+ 2Euii + 2F(uii + iiiJ) + 2GiJii circle U 2 + V 2 ::::; l.
:::: u{Eu u 2 + 2Fu uv + Guv2 ) + v(Evu 2 + 2Fv uiJ +G v v 2 ) 8.18 (i) Let the spheroid be obtained by rotating the ellipse ~ + ~ ::::; 1 around
the z-axis, where a, b > O. Then, a is the maximum distance of a point of
+ 2Euu + 2F(uv + ilv) + 2Gvv
the spheroid from the z-axis, ~o the angular momentum n of a geodesic
:::: 2(Eu u + FiJ)'iL + 2(Fu + Gv)'iI must be :::; a (we can assume that n ~ 0). If n :::: 0, the geodesic is a
+ 2(Eu + Fv)il + 2(Fu + Gv)ii (by the geodesic equations) meridian. If 0 < n < a, the geodesic is confined to the annular region
:::: 2[(EiL + FiJ)u]' + 2[(Fu + GiJ)v]' on the spheroid contained between the circles z :::: ±bJl - q;.,
and the
:::: 2(Eu 2 + 2Fuv + GiJ 2 )'. discussion in Example 8.9 shows that the geodesic 'bounces' between these
two circles:
Hence, (EiL 2 +2Fuv + Gv 2 y:::: 0 and so II i'1I 2 :::: EiL 2 + 2FiLv + Gv 2 is
constant.
8.13 They are normal sections.
8.14 Every parallel is a geodesic ¢=::} every value of u is a stationary point
of feu) (in the notation of Proposition 8.5) ¢=::} f :::: constant ¢=::} the
surface isa circular cylinder.
8.15 The two solutions of Eq. (13) are v :::: VQ ± Jb - w 2 , so the condition
for a self-intersection is that, for some w > 1, 2Jb - w 2 :::: 2krr for some
integer k > O. This holds <::=} 2Jb - 1 > 2rr,Le. n < (1 + rr 2)-1/2. In
this case, there are k self-intersections, where k is the largest integer such
that 2krr < 2J b - 1.
8.16 (i) If'Y(t) is a geodesic, so is f('Y(t)), and if 'Y is defined for all-oo < t < 00,
so is f(-y(t)). So f takes meridians to meridians, i.e. if v is constant, so is
v. Hence, ii does not depend on w.
312 Elementary Differential Geometry Soiubons 3d

If n= a, Eq. (10) shows that the geodesic must be the parallel z = O. 8.19 From Exercise 5.5, the cone is isometric to the 'sector'S ofthe plane with
(ii) Let the torus be as in Exercise 4.10. If [l = 0, the geodesic is a meridian vertex at the origin and angle 1l'V2:
(a circle). If 0 < n < a - b, the g~odesic spirals around the torus:

Geodesics on the cone correspond to possibly broken line segments in S: if

If n = a - b, the geodesic is either the parallel of radius a - b or spi- a line segment meets the boundary of S at a point A, say, it may continue
rals around the torus approaching this parallel asymptotically (but never from the point B on the other boundary line at the same distance as A
crossing it): from the origin and with the indicated angles being equal:

o B

If a - b < n < a + b, the geodesic is confined to the annular region

consisting of the part of the torus a distance :2: {l from the axis, and (i) TRUE: if two points P and Q can be joined by a line segment in S
bounces between the two parallels which bound this region: there is no problem; otherwise, P and Q can be joined by a broken line
segment satisfying the conditions above:

q,

o s

If {l = a + b, the geodesic must be the parallel of radius a + b.

 Elementary Differential Geometry Solubons 315
314

To see that this is always possible, let PI, P2, q, and q, be the indicated
distances, and let R and S be the points on the boundary of the sector at
a distance (p,q, + P1q2)/(P, + q,) from the origin. Then, the broken line
segment PR followed by SQ isthe desired geodesic.
(ii) FALSE:

a
° °
8.20 (i) This is obvious if n ~ since e- 1 / t ' --+ as t --+ 0. We prove that

p
°° °
t- n e- 1/ t ' --+ as t --+ by induction on n > 0. We know the result if
n = 0, and if n > we can apply L'Hopital's rule:
. t- n . nt- n - 1 . n t-(n-2)
hm -1 -
t-.o e / t2
= lun
t--+O 2 e1/t2
= hm -~~-
t--+o 2 e 1/ t2 ,
".
which vanishes by the induction hypothesis.
(ii) We prove by induction on n that 0 is n-times differentiable with
(iii) FALSE: many meet in two points, such as the two geodesics joining
dnO = { ~,~t) e- 1/ t ' if t ¥- 0,
P and Q in the diagram in (ii).
(iv) TRUE: the meridians do not intersect (remember that the vertex of dtn ° if t = 0,
the cone has been removed), and parallel straight lines that are entirely where Pn is a polynomial in t. For n = 0, the assertion holds with Po = l.
contained in S do not intersect. Assuming the result for some n ~ 0,
(v) TRUE: since (broken) line segments in S can clearly be continued ~+!O _ (-3nPn P~ 2Pn) -l/t'
indefinitely in both directions. dtn+1 - t3n+1 + t 3n + t 3n + 3 e
(vi) TRUE: a situation of the form
if t ¥- 0, so we take Pn+ 1 = (2 - 3nt2 )Pn + t 3P~. If t = 0,
1t
dn+!O _. Pn(t) -l/t' _
dtn+! -1~ t3n+! e
. e- / ' _
- Pn(o) 1~ t 3n +! - °
by part (i).
Parts (iii) and (iv) are obvious.
8.21 Since "Ie is unit-speed, Ur.U r = 1, so foR Ur.U r dr = R. Differentiating
a with respect to 0 gives foR Ur.Ure dr = 0, and then integrating by parts
gives
R
r=R
O'e.D'r[r=O -
lo
0 U()·f1rr dr = O.

Now U(O, 0) = P for all 0, so Ue = 0 when r = 0. So we must show that

in which the indicated angles are equal is clearly impossible. But the the integral in the last equation vanishes. But, U rr = t e , the dot denoting
answer to this part of the question depends on the angle of the cone: if the derivative with respect to the parameter r of the geodesic "re, so tT rr
the angle is a, instead of rr/4, lines can self-intersect if a < rr/6, for then is parallel to the unit normal N of U; since ue·N = 0, it follows that
US.a rr = O.
the corresponding sector in the plane has angle < rr:
The first fundamental form is as indicated since ur.u r = 1 and ur.ue = 0.
 316 Elementary Differential Geometry Solutions 317

Chapter 9 which integrates to give f == -~ In cos a(x + b), where b is a constant;

we can assume that b = 0 by translating the surface parallel to the x-
9.1 This follows from Exercise 7.6. axis. Similarly, 9 == ~ In cos ay, after translating the surface parallel to the
9.2 This follows from Exercise 7.3. y-axis. So, up to a translation, we have
9.3 1t1 + 1t2 = 0 and 1t1 = 1t2 ==> 1t1 = 1t2 = O.
y
9.4 1t1 + 1t2 = 0 ==> 1t2 = -It,
==> K = 1t11t2 = -It;
:s O. K = 0 ¢==} = It; z--n
a
- - ),
_ 11 (cosa
cos ax
o ¢==} 1t1 = 1t2 = 0 ¢==} the surface is part of a plane (by Proposition
6.5). which is obtained from Scherk's surface by the dilation (x, y, z) >-+
9.5 By Proposition 7.6, a compact minimal surface would have K > 0 at some a(x, y, z). If a = 0, then / = g" = 0 so f = b + cx, 9 = d + ey, for
point, contradicting Exercise 9.4. some constants b, c, d, e, and we have the plane z = b + d + ex + ey.
9.6 By the solution of Exercise 7.2, the helicoid /7(u,v) = (vcosu,vsinu,Au) 9.10 The first fundamental form is (cosh v + l)(cosh v - cos u) (du 2 + dv 2 ), so /7
is conformal. By Exercise 7.15, to show that is minimal we must show
has E = A2 + v 2,F = O,G = I,L = O,M = A/(A 2 +V 2)1/2,N = 0, so
iT
• I that U uu + tT vv == 0; but this is so, since
- LG-2MF+NG_
H - 2(EG _ F2) - 0. tT uu = (sin u cosh v, cos u cosh v, sin ~ sinh ~),
9.7 A straightforward calculation shows that the first and second fundamental /7vv = (- sin u cosh v, - cosu cosh v, - sin ~ sinh ~).
forms of /7' are cosh 2 u(du 2 +dv 2) and - cos t du 2 - 2 sin tdudv + cos t dv 2,
respectively, so (i) /7(0, v) = (0,1 - cosh v, 0), which is the y-axis. Any straight line is a
2 2
geodesic.
H = - cos t cosh u + costcosh u = O.
(ii) /7(1T, v) = (1T, 1 + cosh v, -4 sinh ¥), which is a curve in the plane x = 1T
2 cosh4 u such that
9.8 From Example 4.10, the cylinder can be parametrised by /7(u, v) = ')'(u) +
va, where')' is unit-speed, II a 11= 1 and ')' is contained in a plane II Z2 = 16sinh2 ~ = 8(coshv -1) = 8(y - 2),
perpendicular to a. We have /7 u = t = t (a dot denoting d/du) , /7 v = a, i.e. a parabola. The geodesic equations are
so E == 1, F == 0, G == 1; N == t x a, O'uu == t == Kn, tT uv == t1 vv == 0, so
L = ItD.(t x a), M = N = O. Now t x a is a unit vector parallel to II ~ (Eu) = ~Eu(U2 + v 2 ), ~ (Ev) = ~Ev(u2 + v 2),
and perpendicular to t, hence parallel to Dj so L = ±K and H = ±K/2.
SO H = 0 ¢==} K = 0 ¢==} ')' is part of a straight line ¢==} the cylinder is where a dot denotes the derivative with respect to the parameter t of the
part of a plane. geodesic and E = (cosh v + 1) (cosh v - cos u). When u = 1T, the unit-speed
9.9 Using Exercise 9.2, the surface is minimal ¢==} condition is Ev 2 = 1, so v = 1/(cosh v + 1). Hence, the first geodesic
equation is 0 = ~EuV2, which holds because E u = sin u(cosh v + 1) = 0
(1 + gI 2)/ + (1 + p)g" = 0, when u == 1I"j and the second geodesic equation is
where a dot denotes d/ dx and a dash denotes d/ dy; hence the stated d
equation. Since the left-hand side of this equation depends only on x and
dt (cosh v + 1) = (cosh v + 1) sinh v v2 = sinh v v,
the right-hand side only on y, we must have which obviously holds.
(iii) /7(u, 0) = (u-sin u, I-casu, 0), which is the cycloid of Exercise 1. 7 (in
/ g"
1 + j2 = a, 1 + g12 = -a, the xy-plane, with a = 1 and with t replaced by u). The second geodesic
equation is satisfied because E v = sinhv(2coshv + 1- cosu) = 0 when
for some constant a. Suppose that a i O. Let r = j; then / = rdr / df and v = O. The unit-speed condition is 2(I-cosu)u 2 = 1, so u = 1/2 sin ¥. The
the first equation is rdr/df = a(1 + r 2), which can be integrated to give first geodesic equation is 1t(4sin 2 ~u) = sinuu 2 , Le.1t(2sin ¥) = cos ¥u,
af = ~ In(1 + r 2 ), up to adding an arbitary constant (which corresponds which obviously holds.
to translating the surface parallel to the z-axis). So df / dx = ±v' e2af - 1, 9.11 A = a2 + be = -(ad - be) (since d = -a) = - det W = -K.
318 Elementary Differential Geometry f Soiubons 319

9.12 (i) From Example 9.1, N = (-sechucosv,-sechusinv,tanhu). Hence, 9.17 '{J = (~(1- (-4)(1_ (2), ~(1- (-4)(1 + (2),((1 _ (-4)), so
if N(u,v) = N(u',v'), then u = u l since u 14 tanhu is injective, so 3 2
cas v = cas Vi and sinv = sinv', hence v = Vi; thus, N is injective. If N = u='.Ro(~((_(3 _C'+C3) ~((+(3 +C 1 +C ) (2 +C )
2 3 3'2 3 3'2 2
(x, y, z), then x' + y2 = sech 2 u i' 0, so the image of N does not contain
the poles. Given a point (x, y, z) on the unit sphere with x' + y' i' 0, let = '.Ro (-~(( - C ' )3, ~((+ C ' )3, ~((+ (-1)2) ,
u = ±sech -1 v' x' + y', the sign being that of z, and let v be such that
cosv = -xl .jx2 + y2, sin v = -y/ v'x 2 + y2; then, N(u, v) = (x, y, z). up to a translation. Put (= e<, C= u+iii. Then, u(u,v) = o7(u,ii), where
(ii) By the solution of Exercise 7.2, N = (A 2 +V')-'/'( -A sin u, Acos u, -v).
Since N(u,v) = N(u+2krr,v) for all integers k, the infinitely many points o7(u, ii) = '.Ro (- ~ sinh 3 C, ;i cosh3 C, 2 cosh' C)
u(u+2k7l",v) = u(u,v)+(0,0,2k7l") of the helicoid all have the same image
under the Gauss map. If N = (x, y, z), then x' + y' = A' /(A 2 + v') i' 0,
so the image of N does not contain the poles. If (x, y, z) is on the unit
= (4 sinhu cos ii(cosh 2 usin 2 ii - ~sinh2ucos2ii),
sphere and x' + y' i' 0, let v = - AZ/ .jx' + y' and let u be such that 1
4 sinh fisin v( 3" sinh 2 usin 2 v - cosh 2 ucos 2 v),
sin u = -xl v'x' + y', cosu = -y/ .jx' + y'; then N(u, v) = (x, y, z).
9.13 The plane can be parametrised by u(u, v) = ub + ve, where {a, b, e} is a 2(cosh2 U cos' ii - sinh' u sin 2 ii)) .
right-handed orthonormal basis of R 3. Then, '(J = uu - iuv = b - ie. The
conjugate surface corresponds to if{) = c + ib; since {8, C, - b} is also a 9.18 By Example 5.7, rr(x,y,z) = (u,v,O), where x = 2u/(u 2 + v' + 1),y =
right-handed orthonormal basis of R 3, the plane is self-conjugate (up to 2 2
2v/(u + v + 1),z = (u 2 + v 2 - 1)/(u' + v 2 + 1). If ( = u + iv, then
a translation). x+iy = ~(/(I(12+l),z = (1(1 2_1)/(1(1 2+1). This gives ( = (x+iy)/(I-z).
9.14 '{J = (~f(l- g'), ~f(1 + g2),fg) ==} i'{J = (~if(l- g'), fif(1 + g2),ifg), Hence, identifying (u, v, 0) with u+iv, we have rr(x, y, z) = (x+iy)/(I-z).
which corresponds to the pair if and g. From Eq. (25), 9(() = IgIJ+ 1 (g + g, -i(g - g), Igl 2 - 1), so
9.15 By Example 9.6, '(J(() = (sinh(, -icosh(, 1). From the proof of Proposi-
g+g+g-g
tion 9.7, f = 'P1 - i'P2 = sinh ( - cosh ( = -c', 9 = 'P3/f = -e'. rr(9(()) = Igl2 + 1 _ (lg12 - 1) = g(().
9.16 '{J = u. - iu v = (1 - u 2 + v 2 - 2iuv,2uv - i(1 - v 2 + u'), 2u + 2iv) =
(1 - (2, -i(1 + ('), 2(). So the conjugate surface is

o7(u,v) = '.Ro /(i(1- (2), 1 + (2, 2i() d( Chapter 10

= '.Ro (i (( - ~) , (+ ~ ,;(') (up to a translation) 10.1 By Corollary 1O.2(i),

= (-v + u'v _ v;, u + ~3 _ uv' , - 2UV) .

K _ 1 ({)
- - 2e' {)u
((e').)
ex- + {)v{) ((e')v))
ex- 1
= - 2e' (Au. + Avv ).
Let U = (u - v)/V2, V = (v + v)/V2, o7(U, V) = u(u, v); then, 10.2 By Exercise 5.4,
V)
(FE aF) = Jt (EF GF) J, where J = (8rQUa;;8r) = ~» ~~

o7(U,v) = (~ (U - V + UV 2
- U'V + ~V 3
- ~U3) , (
u
!v ~ . By Exercise 8.21, E = 1, F = 0, and we get the stated
7!" T2"
.2-. (U+ V+UV' +U2V _ ~V3 _ ~U3) U' - V'). formulas 2for~ E,F,a. From E- 1 = ~ (g. -1) a-I = T ~ (g.T-1)'
V2 3 3 ' - T_ r _. '
we get u (E - 1) = v 2 (G - 1). Since E and G are smooth functions of
Applying the rr /4 rotation (x, y, z)>-+( },(x + y), :7,
(y - x), z) to o7(U, V) (u, v), they have Taylor expansions E = Ei+j,;2 eijUiV j + orr'),
k k
= a
2 2 2 Ei+j,;2 gijU'V' +o(r'), where o(r ) denotes terms such that o(rk)/r --t
then gives (U - tU3 + UV , V - tV3 + U V,U - V'), which is En-
neper's surface again.
o as r --t O. Equating coefficients on both sides of u 2 (E - 1) = v'(a - 1)
320 Elementary Differential Geometry Solutions 321

shows that all the e's and g's are zero except e02 = g20 = k, say. Then, By Exercise 10.2, .;a = r + 0(1') so avo/ar = 1 at l' = O. Hence,
E = 1 + kv 2 + o(r 2 ), which implies that G = r 2 + kr4 + 0(r 4 ).
By Corollary 1O.2(ii), K = - Jc
a~'fP. From the first part, = r+ va j'rJ
ABC
KdAa = 1jJ(LA} -1/;{O) + LA = LC - (rr - LB) + LA
~kr3 + 0(1'3), hence K = -3k + 0(1). Taking r = 0 gives K(P) = -3k. = LA + LB + LC - rr.
10.3 (i) C R = t"II U8 II dO = J;" .j(j de = J;" (R - i K (P)R 3 + o(R3 ») dO .p(LA)
= 2rr (R - 0!K(P)R 3 + 0(R 3 »).
6 R 2" tri
(ii) Since dAa = VG drdO, the area AR = fo fo v G drdO is equal to
2rr foR (1' - ~K(P)r3 + 0(1'3)) dr = rrR 2 Ki{)R2 + 0(R 2 (1- »).
Let S be the unit sphere, let P be the north pole, and let (0, I.p) be
the usual latitude longitude parametrisation of S. Then, the geodesic .p(O)

circle with centre P and radius R is the parallel 0 = I - R, which is an

B
ordinary circle of radius sin R, so that C R = 2rr sin R; since 2rr sin R =
2rr(R - *R 3 + O(R3» and K = 1, this agrees with the formula in (i).
Similarly, 10.5 By Exercise 5.7, generalised cylinders and cones are isometric to a plane.

AR= r r 211" ,,/2 (R R+o(R

cosOd()dl.p=2rr(1-cosR)=2rr 2-
2 4
24
4 )
) ,
Hence, if the sphere were isometric to a generalised cylinder or cone, the
sphere would be isometric to a plane, since composites of isometries are
1 1"/2-R
0
isometries. This contradicts Proposition 10.1.
in agreement with the formula in (ii). 10.6 With the notation of Example 4.9 we have, on the median circle t = 0,
10.4 (i) Let s be the arc-length of "Y, so that ds/dO = A, and denote d/.ds by Ut = (- sin ~ cosO, - sin ~ sin 0, cos ~), (79 = (- sin8,cos8, 0), hence E ==
a dot. The first of the geodesic equations «2) in Chapter 8) apphed to 1, F = 0, G == 1 and N = (- cos! cos 8, - sin ~ sin (J, - sin ~) j 6 tt = 0,
"Y gives f = ~GTB2. Since r = f(8), this gives Ut8 == (_lcos~cosO+sin~sin8 _!cos~sin8-sin~cos8 -!sin~)
2 2 2' 2 2 2' 2 2'
1(1 /)/ = 12Gr .
X Xf 2A
from which L == O,M == t.
Hence, K == (LN - M 2 )/(EG - F 2 ) = -1/4.
Since K 1:- 0, the Theorema Egregium implies that the Mobius band is
not isometric to a plane.
This simplifies to give the stated equation. 10.7 The first fundamental forms ofa and iT are found to be (1+~ )du 2+u 2dv 2
(ii) Since U T and -y are unit vectors, cos'l/J = 6 r .-Y = t6r.(f/6r + iTo) = and du 2 + (u 2 + 1)dv 2, respectively. Since these are different, the map
I'/A. Also, U r x -y = t(iT r X (8) = so sin'l/J xfN, = VG/>". Hence, u(u,v) I---t iT(u, v) is not an isometry. Nevertheless, by Corollary 1O.2(ii),
the gaussian curvatures are equal:
(1')' VG
"I = -'l/J/ sin'l/J = - ---y::'l/J',
1 ( 1'>"') = - 2VG 1 aG oVG
K=K = -2~ (:u (~)) = (1+1U2)2'
1jJ' = - va !" ~ T or = -~. If u(u, v) iT( ii, ii) is an isometry, the Theorema Egregium tells us
I---t

(iii) Using the formula for K in Corollary 1O.2(ii) and the expression for that -1/{1 + U 2)2 = -1/(1 + U2)2, so ii = ±u; let ii = f(u,v). The
the first fundamental form of u in Exercise 8.21, we get first fundamental form of u{±u,f(u,v) is (1 + (I + u2)f~)du2 + 2(1 +

LA 1'(9) 1 a2 va v'G drd(J

u2)fufvdudv+(1+u2)f;dv2; this is equal to the first fundamental form of
u(u,v) ¢::=} 1+(1+u2)f~ = 1+1/u2, fufv = 0 and (1+u 2}f; = u 2. The
If ABC
K dAa =
l 0 0
-
vG ar
rr; - 2 -
middle equation gives fu = 0 or fv = 0, but these are both impossible

= -
l
LA
0..;7]
-
ar
-I
T

=/(8
l
dB - _l 0
LA
(/
'l/J ~ I
+ oVG
vr
) dO. by the other two equations. Hence, the isometry does not exist.
10.8 For the catenoid u(u,v) = (coshucosv,coshusinv,u), the first funda-
o r=O r=O 2
mental form is cosh u(du 2 + dv 2 ) (Exercise 5.8) and the gaussian curva-
4
ture is K = -sech u (Exercise 7.2). If 6(U, v) I---t 6(U, ii) is an isometry,
322 Elementary Differential Geometry Solutions 323

then sech4 u = sech4 u, so U reflecting in the plane z = a changes

= ±Uj Chapter 11
u to -u, so assume that the sign is +. Let Ii = f(u, v); the first funda-
mental form of a(±u, f(u,v)) is (cbsh2 U+ f~)du2 + 2fufv cosh 2 ududv + 11.1 If 'Y is a simple closed geod~sic, Theorem 11.1 gives JJ.nt('Y) KdA = 2rr;
{; cosh udv 2; hence, cosh 2 u = cosh U + f~, = 0 and f; cosh 2
2 2
fufv u= smce K :5 0, thiS 's Impossible. The parallels of a cylinder are not the
cosh2 u, so fu = 0, Iv = ±l. Thus, f = ±v + 0, where 0' is a constant; if images under a surface patch tr : U --+ R 3 of a simple closed curve 7r in
the sign is +, we have a rotation by a about the z-axis; if the sign is - the plane such that int(1l") is contained in U. Note that the whole cylinder
we have a reflection in the plane containing the z-axis, making an angle can actually be covered by a single patch (see Exercise 4.2) in which U
a/2 with the xz-plane. is an annulus, but that the parallels correspond to circles going 'around
-I
2
(cos v 0)-1 (-COS
2
V 0) = I, so N u = au, the hole' in the annulus.
10.9 W = -F[ F II = - 0 1 0 -1 11.2 By Proposition 2.2, K, = d<p/ds, where <p is the angle through which a
N v = a v • Thus, N = a-a, where a is a constant vector. Hence, II a-a II fixed unit vector must be rotated anti-clockwise to bring it into coinci-
= 1, showing that the surface is part of the sphere of radius 1 and centre dence with the unit tangent vector of 'Y. So
a. The standard latitude longitude parametrisation of the unit sphere has '('Y) l'('Y) d<p
first and second fundamental forms both given by du 2 + cos2 vdv 2 , so the Its ds = -ds = 21r,
lo 0 ds
parametrisation a(v, u) has the given first and second fundamental forms
by the Umlaufsatz.
(the second fundamental form changes sign because a v X t1 u = -t1 u X t1 v ).
11.3 By Corollary 11.1, the interior angles al, ... , an of the polygon satisfy
10.10 ri2 = sin u cosu and the other Christoffel symbols are zero; the second

La, = (n-2)rr + 11.

n
Codazzi-Mainardi equation is not satisfied.
10.11 The Christoffel symbols are rll = E u/2E, rrl = -Ev/2G, rl2 = KdAu·
i=l mt("Y)
E v /2E, rr2 = G u/2G,ri2 = -G u/2E, ri2 = G v /2G. The first Codazzi-
Mainardi equation is Since 0 < a, < 2rr for all i, the left-hand side is > 0; since K < 0, we have
(n - 2)rr > 0 and hence n ~ 3; and ifn = 3, then JJ.nt('Y) (-K)dAu < rr
Lv = ~~v _N (~~v ) = ~Ev (~ + ~) , so

and similarly for the other equation. Finally, 11. int('Y)

dAa < 11. int('Y)
(-K)dAa < rr.
E v (L
(I<rJv = 2E E + N)
G -
LEv Ev
E2 = 2E
(NG - L) Ev ( )
E = 2E 1<2 - 1<1 ,
11.4 The parallel u = UI is the curve 'YI(V) = (f(urJcosv,j(urJsinv,g(urJ);
if s is the arc-length of 'YI, ds/dv = f(uI). Denote d/ds by a dot and
and similarly for (K2)u. d/du by a dash. Then, 'y = (-sinv,cosv,O), 'Y = -"'!'-'-(cosv sinv 0)
J \ Ul' ' "
10.12 Arguing as in the proof of Theorem 10.4, we suppose that J attains and the unit normal of the surface is N = (- g' cos v, - g' sin v, f'). This
its maximum value > 0 at some point P of S contained in a patch gives the geodesic curvature of 'Y as Kg = 'Y.(N x 'y) = ~i::I. Since
t1 of S. We can assume that the principal curvatures KI and K2 of t1
('YI) = 2rrf(ur), J~('Y,) Kg ds = 2rrf'(urJ. Similarly for 'Y2' By Example
satisfy KI > K2 > 0 everywhere. Since H = tiKI + K2), KI > H and 7.2, K = -I"/f, so
J = 4(KI - H)2. Thus, J increases with KI when KI > H, so KI must
have a maximum at P, and then K2 = 2H - KI has a minimum there.
.{ {2' (u, I"
/ J R KdAa = Jo J , - / fdudv = 2rrU'(ur)- f'(U2))'
By Lemma 10.2, K :5 0 at P, contradicting the assumption that K > 0 u
everywhere. Hence,

J
{'('Y,)
Kgds- J
{'('Y,)
Kgds= JR KdAa .
I{
o o
This equation is the result of applying Theorem 11.2 to the following
curvilinear polygon,
324 ¥
Elementary Differential Geometry Solutions 325

11.7 If such curves exist they would give a triangulation of the sphere with 5
vertices and 5 x 4/2 = 10 edges, hence 2+ 10 - 5 = 7 polygons. Since each
edge is an edge of two polygons and each polygon has at least 3 edges,
3F ~ 2Ej but 3 x 7 > 2 x 10. If curves satisfying the same conditions
exist in the plane, applying the inverse of the stereographic projection
map of Example 5.7 would give curves satisfying the conditions on the
sphere, which we have shown is impossible.
n.S Such a collection of curves would give a triangulation of the sphere with
V = 6, E = 9, and hence F = 5. The total number of edges of all the
polygons in the triangulation is 2E = 18. Since exactly 3 edges meet at
each vertex, going around each polygon once counts each edge 3 times,
so there should be 18/3 = 6 polygons, not 5.
where AB is part of the meridian v = 0 (or v = 211"); the integrals of n.9 By Corollary 11.3, lIs KdA = 411"(1 - g), and by Theorem 11.6, 9 = 1
"'g along AB and along BA cancel out. (Strictly speaking, this curve since S is diffeomorphic to T1 . By Proposition 7.6, K > 0 at some point
is not a curvilinear polygon in the sense of Definition 11.2 - condition of S.
(i) is violated - but this difficulty can be circumvented by replacing the 11.10 The ellipsoid is diffeomorphic to the unit sphere by the map (x,y,x) f-t
double path AB and BA by two meridians v = E and v = 211" - f and (x/a, yla, z/b), so the genus of the ellipsoid is zero. Hence, Corollary 11.3
then letting E tend to zero.) gives ffs KdA = 411"(1- 0) = 411".
n.5 3F = 2E because each face has 3 edges and each edge is an edge of 2 Parametrising the ellipsoid by d( 0, cp) = (a cos () cos cp, a cos 0 sin cp, b sin ())
faces. From X = V -E+F, we get X = V -E+iE, so E = 3(V -X). Since (cf. the latitude longitude parametrisation of the sphere), the first and
each edge has 2 vertices and two edges cannot intersect in more than one second fundamental forms of t1 are (a 2 sin 2 0 + b2 cos 2 8)dfj2 + a2 cos 2 (}dcp2
!
vertex, E ~ !V(V -1); hence, 3(V - X) ~ V(V -I), which is equivalent and v'(&2 ain 2 ~~b2 cos 2 6 (d(}2 + cos 2 ()dcp2), respectively. This gives K =
t
to V 2 -7V +6X ~ O. The roots of the quadratic are (7 ± J49 - 7X), so 2
b2 /(a 2 sin ()+b 2 cos 2 (})2, dAu = a cos ()J a 2 sin 2 8 + b2 cos2 ()d()dcp; hence,
! !
V ~ (7 - .;49 - 7X) or V ~ (7 + .;49 - 7X). Since X = 2,0, -2, ...,
r K dAu = 1 2
r 2
the first condition gives V ~ 3, which would allow only one triangle;
hence, the second condition must hold.
11.6 The n triangles have 3n vertices, but each vertex is counted r times as
J1
s 0
11" /2
1-11"/2
ab cos 8 dfJdcp
2
(a 2 sin fJ+b 2 cos2 (})3/2'
11.11 K > 0 =::} IIs KdA > O. By Corollary 11.3, 9 < 1; since 9 is a non-
it is a vertex of r triangles, so V = 3n/rj similarly, E = 3n/2. Then,
negative integer, 9 = 0 so S is diffeomorphic to a sphere by Theorem
V - E + F = 2 gives 61r - 4/n = 1. This implies 6/r > 1, so r < 6. 11.6. The converse is false: for example, a 'cigar tube' is diffeomorphic
'Ifiangulations for r = 3,4 and 5 can be obtained by 'inflating' a regular to a sphere but K = 0 on the cylindrical part.
tetrahedron, octahedron and icosahedron, respectively.

tetrahedron octahedron icosahedron

 $
326 Elementary Differential Geometry Solutions 327

11.12 Both surfaces are closed subsets of R 3 , as they are of the form 1- 1 (0), rl{'y) ~ds = 0. Using Green's theorem, this integral is equal to
Jo
where I : R 3 --+ R is a continuous function (equal to x 2 - y2 + Z4 - 1 yl_p2

and x 2 + y2 + z4 - 1 in the two cas'es). The surface in (i) is not bounded,

and hence not compact, since it contains the point (I, aZ , a) for all real
{ Pudu + Pvdv { a( Pv ) a( Pu )
J'Jr ~ = Jint('Jr) 8u ~ - 8v ~ ,
numbers aj that in (ii) is bounded, and hence compact, since x Z+yZ+z4 =
1 ==> -1 ~ x,y,z ~ 1. where'Jr is the curve in U such that -yes) = O'('II'(s))j and this line integral
The surface in (ii) is obtained by rotating the curve x 2 + Z4 = 1 in the vanishes because
xz-plane around the z-axis:
8 ( Pv ) 8 ( Pu )
8u ~ =8v ~
11.16 Let F : S --+ R be a smooth function on a surface S, let P be a point of S,
let (1 and f1 be patches of S containing P, say O'(uo, vo) = u(Uo"vo) = P,
&v-
and let I = - - - lIu
F 0 0' and 1 = F 0 0'. Then, fu ~ Iv.~ii. + 11l7ffJ.' I" =
lu ~ + Ill~' so if lu = Iv = 0 at (uo, vo), then Iii. = Iv = 0 at (uo, vo).
Since 11.1. = Iv = 0 at P, we have

-
fuii. = luu (8U)Z
8u
8u 8v
+ 2/uv 8ft 8ft + Ivv
(8v)2
8ft '
It is clearly diffeomorphic to the sphere, so X = 2.
11.13 Take the reference tangent vector field to be ( = (1,0), and take the with similar expressions for fuv and ivv. This gives, in an obvious no-
simple closed curve 'Y(s) = (coss,sins), At 'Y(s), we have V = (a,I1),
where
tation, il = J t 1lJ, where J :::: (i~
7ffJ.
i!)
~_
is the jacobian matrix ·of
the reparametrisation map (ft,ii) I-t (u,v). Since J is invertible, 11. is
OR _ {(COsS+isinS)k if k > 0,
+ 1/-' invertible if 11. is invertible.
a - ( . ) Ie
cos S - 1.sm s- 1'f k <.0
Since the matrix 11. is real and symmetric, it has eigenveGtors,vl,v2,
By de Moivre's theorem, a = cos ks, (3 = sin ks in both cases. Hence, with eigenvalues AI, AZ, say, such that vfvj = 1 if i = j and 0 ifi¥- j.
the angle 1/J between V and ( is equal to ks, and Definition 11.6 shows Then, if v = al VI + a2V2 is any vector, where 01, a2 are scalars, v t 1iv =
that the multiplicity is k. AlaI + >'za~; hence, v t 1lv > 0 (resp. < 0) for all v ;l '0 <=> A!1and A2
11.14 If O'(u,v) = u(u,v), where (u,~) I-t (u,v) is a:. repar?ID~trisati~m map, are both> 0 (resp. both < 0) {::::::} P is a local minimum '(r~. ilaca:J.
then V = aO'u + /-'O'v
R - -
= O!O'ij + /-'O'iJ
R-
==> a- -- O!&u &v /-'R -- a &u + (3 &v'
&u + /-'R &u' &v
8v maximum); and hence P is a saddle point {::::::} v t 1ivcan be br:Jth > Il!)
Hence, it and i3 are smooth if a and /3 are smooth. and < 0, depending on the choice of v. Since J is invertible,a vecter
Since the components of the vectors 0' u and 0' v are smooth, if V is smooth v ;l 0 {::::::} v = Jv ;l OJ and vtili) = iJt J t 1lJv = v t1lv. ~heasserti<ms in
so are its components. If the components of V = aO'v + /3O'v are smooth, the last sentence of the exercise follow from this.
then V.O'u and V.O'v are smooth functions, hence 11.17 (i) Ix = 2x ~ 2y, 111 = -2x +8y, so Ix = 111 = Oat the 0r.igin. ixx =

a=
G(V.O'u) - F(V.O'v)
EG-F2
(3 = B(V'O'Il) - F(V.O'u)
EG-F2
2 , 1X1/ = -2 , 11 1=8
11'
so 1l = -2 8 . '1J" ,( 2 -2)
n IS mvertL·'bl'eso 11'e
. h :rn:Igm
...

are smooth functions, so V is smooth. is non-degenerate; and the eigenvalues 5 ± VI3 of 1l are ibdtih > (@, 'so;it
11.15 If;j, is the angle between V and (, we have -if; - t/J = 8 (up to multiples is a local minimum.
of 271" ) j so we must show that Jot (-y) iJ ds = 0 (a dot denotes d/ ds) ° This is (ii) Ix = 11/ = 0 and 1l = (~ ~) at the origin; det1i = -16< i(iJ, ~S0
not obvious since 8 is not a well defined smooth function of s (although the eigenvalues of 1l are of opposite sign and the origin is a saddle \point.
d9/ ds is well defined). However, p = cos 9 is well defined and smooth,
since p = e.V II ( 1111 ( II· Now, p = -iJsin9, so we must prove that critical point.
°
(iii) Ix = 11/ = 0 and 11. : : : at the origin, which is therefol'e;a·degenerate
328 Elementary Differential Geometry
T---------------------
11.18 Using the parametrisation iT in Exercise 4.10 (with a = 2, b = 1) gives
1(0,1{J) = F(iT(O,I{J)) = (2 + cosO)COSI{J + 3. Then, 19 = -sinOcosl{J,
Index
f", = -(2 + cosO)sin<p; since 2 +"cosO > 0, f", = 0 => <p = 0 or
IT, and then 19 = 0 ==} 0 = 0 or IT; So there are four critical points,
P = (3,0,0), Q = (1,0,0), R = (-1,0,0) and S = (-3,0,0). Next, 11. =
- cosOCOSI{J
( sinOsin<p
sin 0 sin I{J
-(2 + cosO) cos <p
) = (-I 0)
0 -3
at P =
'
(1 0)
0-1

at Q, = (~I ~) at R, and = (~ ~) at S; hence, P is a local

maximum, Q and R are saddle points, and S is a local minimum (all of
which is geometrically obvious).

angle, 106 Codazzi-Mainardi equations, 241 , 244 ,

angular momentum, 185 322
anti-holomorphic, 112, 219 compact , 164
Archimedes's Theorem, 117, 238, 297 conformal, 107-8, 111, 190, 203, 216-7,
arc-length, 7-8, 11, 97 219-21, 296, 307
area, 50, 113-5, 202 connected, 18, 72
- of a geodesic triangle, 119, 237, continuous, 60
256-7, 320-1 convex, 55, 289
- signed, 168 Cornu's spiral, 33
astroid, 3, 6, 281, 285 critical point, 275-8, 280, 327-8
asymptotic curve, 129-30, 150-1, 161, - non-degenerate, 276-8, 280
304 curvature
atlas, 60 - constant, 31-2, 41, 44
- maximal 68 - of a curve, 24, 25
bifurcation, 271 - gaussian, 147, 150, 161-2, 164, 166,
binormal, 36 168-9, 207, 219, 225-6, 229, 233, 239,
Catalan'8 surface, 216, 317 245, 248, 268
catenary, 9, 283 - geodesic, 127, 248, 310, 323
- mean, 147, 150, 161-2, 201, 307
catenoid, 83, 106, 140, 150, 203, 216,
219, 223, 227, 240, 296, 300, 303, 318 - normal, 127, 130, 137-8, 140, 299-300
321 ' - principal, 132-3, 137-8, 140, 142-3,
148, 150, 162, 164, 244-4, 300, 304,
Cauchy-Riemann equations, 112, 220, 306
296
centre of curvature, 35
- signed, 28-30, 34-6, 51, 165, 242, 252
curve
Christoffel symbols, 241, 322 - asymptotic, 129-30, 150-1, 161, 304
circular cone, 63, 72, 103, 105, 180, - level, 2, 16 , 19, 96
196, 313-5 - parallel, 35, 285
circular cylinder, 65,103,136,179-80, - parameter, 107,161,175
252, 289 - parametrised, 2, 16, 19, 21
circular helix, 26-7,39, 104, 115, 287, - plane, 40, 150, 242
296 - smooth, 4, 74, 96
cissoid, 15 - spherical, 6, 45, 287, 298
Clairaut's theorem , 183 , 185 curvilinear polygon, 252
closed set , 195 - positively-oriented, 252
329
 x
33.0~ --=.E::le:.::m::en:.::t::a:2ry...:D::.i:.::ff.:er::e:.::nt:.::ia::I...:G:::e::o:.::m::et:::2.ry

geodesic equations, 176-7, 181 , 309, meridian, 81, 129, 140, 182, 189, 292 ruling, 80, 293, 296
cycloid, 6, 35, 282-3
317, 320 Meusnier's Theorem, 131 saddle point, 277-8, 280, 327-8
diffeomorphic, 71
Mobius band, 76-7,239-40, 321 Scherk's surface, 164, 214, 216, 317
diffeomorphism, 71 geodesic patch, 199
"geodesic polar patch, 199 Mobius transformation, 190 second fundamental form, 125-6, 242
- conformal, 107-8, 111, 121, 190
geodesic torsion, 129, 141, 309 monkey saddle, 143 signed unit normal, 28, 51
- equiareal, 116, 121, 297
multiplicity, 269, 274-5, 277, 326 simple closed cu~ve, 47, 247-8, 288
dilation, 207, 304 gradient, 275
graph, 72, 163 non-euclidean geometry, 154 - area of, 50-1
disc model, 155, 190
great circle, 174, 179, 299 normal - exterior of, 48
double torus, 274
Green's Theorem, 50 - to a curve, 17 - interior of, 48, 248
doubly ruled, 66, 90, 290, 293
helicoid, 77, 106, 140, 150, 181, 210-1, - to a surface, 75 - length of, 49, 51
Dupin's Theorem, 141
216, 219, 223, 227, 239, 292, 296, 300, normal curvature, 127, 130, 137-8, 140, - period of 47, 247
edge, 252
303, 318 299-300 - positively-oriented, 49, 248
ellipsoid, 73, 85, 91, 145, 175, 268, 302,
Henneberg's surface, 227 normal section, 128, 173 simply-connected, 221
308
holomorphic, 112, 219-20, 222 open ball, 60 sink, 271, 273
elliptic cylinder, 86
homeomorphism, 60 open disc, 60, 289 smooth
elliptic paraboloid, 85, 93, 126, 129,
Hope's Umlaufsatz, 250-2, 270, 323 open interval, 2, 60 - curve, 4, 74, 96
143, 294
hyperbolic cylinder, 86 open set, 59 - function, 73, 78
elliptic point, 143, 164
Enneper's surface, 163, 214, 227 hyperbolic paraboloid, 72-3, 86, 89, 93, osculating circle, 35, 285 - map, 69
143, 293-4 parabolic cylinder, 87 - surface, 67, 71
epicycloid, 6, 282
hyperbolic point, 143, 161, 164 parabolic point, 143, 164 soap films, 202-3
equiareal, 116, 121, 297
Euler number, 260-2, 264-5, 267, 269, byperboloid parallel, 81, 129, 140, 182, 189, 252,
source, 271, 273
- of one sheet, 66,85,91, 175, 187, 292
271, 324 speed, 9, 172
Euler's Theorem, 137 289, 308 parallel curve, 35, 285
spheroid, 190, 311-2
- of two sbeets, 85, 91 parallel surface, 161-4
evolute, 35-6, 285 standard unit normal, 75, 139
first fundamental form, 98, 100, 102, hypocycloid, 6, 282 parameter curve, 107, 161, 175
stationary point, 269, 271, 274-5
108, 242 integral curve, 269 parametrisation
- of a curve, 2 stereographic projection, 110, 238, 296
Four Vertex Theorem, 56 Inverse Function Theorem, 93-4 surface, 60
Frenet-Serret equations, 41 involute, 35-6, 285 - of a surface, 60
area of, 113-5 .-',",
Fresnel '8 integrals, 33 isometric, 10l, 235, 296 parametrised curve, 2, 16, 19, 21
period, 47,247 - compact, 164, 207, 244, 246, 258,
Gauss-Bonnet Theorem, 248, 253, 255, - deformation, 105-6, 216, 296 260, 268, 271
260 isometry, 101,111,121,179,190,229, planar point, 143, 164
Plateau's problem, 201 - conjugate, 223, 226-7, 318
Gauss equations, 240 238-40, 297, 321-2 - fiat, 155, 164
Gauss map, 166, 169, 217, 228, 318-9 isoperimetric inequality, 51, 288 principal curvature, 132-3, 137-8, 140,
142-3, 148, 150, 162, 164, 244-5, 304, - level, 71
gaussian curvature, 147, 150, 161-2, jacobian matrix, 93, 100, 126, 160 - minimal, 202, 207-8, 211, 214-7,
164, 166, 168-9, 207, 219, 225-6, 229, Jordan Curve Theorem, 48, 252 306
principal normal, 36, 175, 298 219-21
233, 239, 245, 248, 268 Lagrange's Method of Undetermined
principal patch, 156 - of revolution, 81, 89, 125, 129, 140,
- constant, 162, 235, 244 Multipliers, 78, 292 145, 149, 182-3, 189, 208, 234, 257
Gauss's Lemma, 200, 315 latitude, 61 principal vector, 133, 137-8, 140, 162,
300,306 - orientable, 75, 78, 166
general helix, 44-5 level curve, 2, 16, 19, 96 - parallel, 161-4
generalised cone, 79, 89, 99, 105, 157-8, level surface, 71 profile curve, 81
pseudosphere, 153-5, 185-6, 190, 235-6, - ruled, 80, 89, 111, 149-50, 156, 164,
239, 296 lim~on, 20,51, 57, 289
257, 304-6, 310-1 211, 304
generalised cylinder, 78, 84, 89, 99, line of curvature, 140-1, 150, 301, 307,
quadric, 84-7, 89, 293 - smooth 67, 71
105, 157-8, 168, 174, 180, 216, 239, 309
quadric cone, 86, 295 - translation, 216, 316
295-6, 309 local maximum, 277-8, 280, 292, 327-8
local minimum, 277-8, 280, 292, 327-8 regular, 11, 15, 21 surface patch, 60
genus, 258, 265, 325
regular point, 11, 16 - allowable 68-70
geodesic, 171,175,177-83,187-96,217, logarithmic spiral, 7,8, 14, 34-5, 285
308-15, 317 longitude, 61 reparametrisation, 10, 69, 100, 126 - regular 67-8
- incomplete, 186, 196 loxodrome, 83 - unit-speed, 12, 13, 172 surface variation, 201-2, 206
- simple closed, 252, 323 lune, 118 reparametrisation map, 10, 69 tangent developable, 103-4, 157-8
geodesic circle, 199-200, 237, 320 mean curvature, 147, 150, 161-2, 201, right conoid, 83 tangent plane, 75, 292
geodesic coordina.tes, 199, 235 307 rigid motion, 3D, 42, 100, 150, 207, tangent space, 74
geodesic curvature, 127,172,181,248, - constant, 203, 246 242, 285-6 tangent vector, 4, 5, 51
310, 323 Mercator's projection, 83, 111, 238 Rodrigues' formula, 140 tangent vector field, 269, 274, 326
Theorema Egregium, 229, 238, 240, unit sphere, 61, 65, 67-8, 71-2, 135,
242, 321 152, 168, 177
third fundamental form, 141, 150, 302 upper half-plane model, 155
torsion, 37-8, 40-1, 44, 151, 175 " Utilities problem, 268
torus, 73, 145, 150, 169, 190, 258, 274, vertex
291, 297, 312 - of a curve, 55, 289
tractrix, 153
- of a curvilinear polygon, 252
transition map, 65, 68, 72, 291
triangulation, 259-60, 324-5 Viviani's curve, 6, 45
triply orthogonal system, 90, 92-3, 141, vortex, 271
164, 294 Weierstrass's representation, 224,
tube, 115, 175, 297 226-8
twisted cubic, 14 Weingarten matrix, 139
umbilic, 133, 144-5, 155, 207, 245, 302 Wirtinger's inequality, 52, 54
unit-speed, 9 witch of Agnesi, 6