iii MAT215–V/3/08–10

APM212–W/3/08–10

Contents

Page
Chapter 12 Single Integrals 1

12.1 Introduction 1
12.2 De¿nition of a Single Integral 1
12.3 Geometric Interpretation of Single Integrals 3
12.4 The Fundamental Theorem of Calculus 4
12.5 Rules of Integration for Single Integrals 5
12.6 Integration by Substitution 5

Chapter 13 Double Integrals 9

13.1 Introduction 9
13.2 De¿nition of a Double Integral 9
13.3 Geometric Interpretation of Double Integrals 11
13.4 Rules of Integration 13
13.5 Evaluation of Double Integrals 14
13.6 Transformation of Coordinates 27
13.7 Calculation of Volumes by means of Double Integrals 39
13.8 Calculation of Areas by means of Double Integrals 42

Chapter 14 Triple Integrals 47

14.1 Introduction 47
14.2 De¿nition of a Triple Integral 47
14.3 Geometric Interpretation of Triple Integrals 48
14.4 Rules of Integration for Triple Integrals 49
14.5 Evaluation of Triple Integrals 49
14.6 Coordinate Transformations 52
14.7 Calculating Volumes by using Triple Integrals 64
 iv

Chapter 15 Line Integrals of Real–valued Functions 69

15.1 Introduction 69
15.2 Line Integrals in the Plane, with respect to Arc Length 70
15.3 Line Integrals in the Plane, with respect to the Coordinate Variables 78
15.4 Line Integrals of Rn  R functions 83
15.5 Line Integrals over Piecewise Smooth Curves 85

Chapter 16 Line Integrals of Vector Fields 89

16.1 Introduction 89
16.2 De¿nition of a Line Integral of a Vector Field 89
5
16.3 Physical Interpretation of F · dr 91
5 C
16.4 Evaluating F · dr 92
C
16.5 The Fundamental Theorem of Line Integrals 96
16.6 Path–independent Vector Fields 97
16.7 Line Integrals over Closed Curves 99

Chapter 17 Surface Integrals 103

17.1 Introduction 103

17.2 De¿nition of a Surface Integral 103
17.3 Geometric Interpretation 105
17.4 Evaluation of Surface Integrals 106

Chapter 18 Flux Integrals 111

18.1 Introduction 111

18.2 Orientable Surfaces 112
18.3 Flux 113
18.4 Evaluation of Flux Integrals 115

Chapter 19 The Theorems of Green, Gauss and Stokes 119

19.1 Green’s Theorem 119

19.2 Stokes’ Theorem 123
19.3 Gauss’ Theorem 128
19.4 Generalizations of the Fundamental Theorem of Calculus 131
Answers and hints to exercises 133
Index 138
Chapter 12

Single Integrals

12.1 INTRODUCTION
5
An integral of the form f (x)dx, where f is an R  R function and I is an interval on the X-
I
axis, is called a single integral. Since single integrals are studied in ¿rst level Calculus, we shall
assume that you have grasped how an integral represents the limit of a sum, that you understand the
connection between single integrals and areas and that you have mastered the standard techniques
of integration.
In this chapter we brieÀy revise some important aspects of single integrals. We begin by revising
the de¿nition of a single integral, since this forms the basis for the de¿nitions of the various types
of integrals that we shall study in later chapters.

12.2 DEFINITION OF A SINGLE INTEGRAL

5b
The steps which lead to the de¿nition of the single integral a f (x)dx over a bounded interval
[a, b] may be summarized as follows:

1. Consider an R  R function f de¿ned on a closed interval [a, b].

2. Form a partition P of the interval [a, b] by dividing it into n subintervals

[ck1 , ck ]; k = 1, ..., n

where
a = c0 < c1 < c2 < · · · < cn = b.

3. Choose a point xk in the kth subinterval of the partition, for k = 1, ..., n.

∆xk

a x1 c1 x2 c2 ck-1 xk ck cn-1 xn b
Fig. 12.1

1
 2

4. Let xk denote the length of the kth subinterval, i.e.

xk = ck  ck1

and let PPP (the norm of the partition) be the length of a longest subinterval, i.e.

PPP = max {xk | k = 1, . . . , n} .

Y
y = f(x)
f(xk)

Ak

a x1 c1 c k-1 x k c k cn-1 b X

Fig. 12.2 The area of Ak is f (xk )xk

5. Form the sum

;
n
f (xk )xk
k=1
and consider the limit of this sum as PPP tends towards zero.

De¿nition 12.2.1 Let f be an R  R function de¿ned on a closed interval [a, b]. The single
integral of f over [a, b] is given by
= b ;
n
f (x)dx = lim f (xk )xk ,
a PP P0
k=1

provided this limit exists.

If this limit exists, we say f is integrable over [a, b].

Remark 12.2.2

If f (x) o 0 for every x + I then, for each k = 1, . . . , n, the value of f (xk )xk gives the area of
the kth rectangle, Ak , shown in in Fig 12.2. Let R be the region below the graph of f, above the X-
3
axis and between the lines x = a and x = b. Then the sum nk=1 f (xk )xk is an approximation
of the area of R. Generally, the smaller the norm of the partition, the better the approximation. If
the limit in De¿nition 12.2.1 exists, then it gives the exact value of the area of R.
 3 MAT215–V/3
APM212–W/3

12.3 GEOMETRIC INTERPRETATION OF SINGLE INTEGRALS

Suppose f is a continuous R  R function de¿ned on a closed interval [a, b]. If R is the region
between the X-axis, the graph of f , and the lines x = a and x = b, then the following hold:

=b
If f (x) o 0 for all x + [a, b], then f (x)d x = area of R. (12.1)
a
=b
If f (x) n 0 for all x + [a, b], then f (x)dx = (area of R). (12.2)
a

y = f(x)
Y

R
a b
a b X X

y = f(x)
5b 5b
(a) a f (x)dx = area of R (b) a f (x)dx =  (area of R)

Y
y = f(x)

R2
a
b X
R1

5b
(c) a f (x)d x =  (area of R1 )+ area of R2

Fig. 12.3
 4

12.4 THE FUNDAMENTAL THEOREM OF CALCULUS

We remind you of the following important theorem.

Theorem 12.4.1 The Fundamental Theorem of Calculus

If f is a smooth R  R function that is de¿ned on a closed interval [a, b], then
=b
f ) (x)d x = f (b)  f (a) .
a

We can interpret f ) (x) as the rate of increase in f at the point x, while f (b) f (a) gives the total
increase in f over the interval [a, b]. Thus the Fundamental Theorem of Calculus tells us that the
integral of the rate of increase in a smooth function over an interval gives the total increase in the
function over that interval. Let us explain this brieÀy.
If f ) is constant over [a, b], then the equation of f throughout [a, b] is of the form f (x) = mx +c,
where m = f ) (x), and the total increase in f over [a, b] equals the slope of f times the length of
the interval [a, b], which equals

m(b  a) = (mb + c)  (ma + c) = f (b)  f (a).

If f ) is not constant over [a, b], then the total increase in f is given by the sum of the increases
)
over the sub-intervals on which f (x) is constant. If f is a smooth function, the total change is
5b )
given by the integral a f (x)d x.
If F and f are R  R functions such that

F(x) = f ) (x) for all x + [a, b]

then f is called an antiderivative of F on [a, b]. We know that if F is continuous on [a, b] then
it has an antiderivative on [a, b]. Theorem 12.4.1 can therefore also be stated as follows:

Theorem 12.4.2 The Fundamental Theorem of Calculus (Alternative form)

If F is an R  R function that is de¿ned and continuous over a closed interval [a, b] and f is any
antiderivative of F on [a, b], then
= b
F(x)d x = f (b)  f (a).
a

Theorem 12.4.2 tells us that, if F is continuous over a closed interval [a, b], we can evaluate the
5b
integral a F(x)dx, by ¿nding an antiderivative of F and evaluating it only at the two boundary
points of the interval.
 5 MAT215–V/3
APM212–W/3

12.5 RULES OF INTEGRATION FOR SINGLE INTEGRALS

We remind you of the following rules:

Theorem 12.5.1 Rules of Integration

If f and g are both R  R functions that are integrable over an interval [a, b], then:
5b 5b 5b
1. a ( f (x) + g(x))d x = a f (x)d x + a g(x)d x.
5b 5b
2. a k f (x)d x = k a f (x)dx for all k + R.

3. If c + [a, b] then

= b = c = b
f (x)d x = f (x)dx + f (x)d x.
a a c

12.6 INTEGRATION BY SUBSTITUTION

Recall the Change of Variable Formula for evaluating a single integral by means of a substitution.

Theorem 12.6.1 Change of Variable Formula

Suppose g is a smooth, invertible R  R function of u that maps the interval [c, d] onto the interval
[a, b]. If f is an R  R function of x that is integrable over [a, b], then
1 (b)
=b g=

f (x)d x = f (g (u)) g ) (u) du.

a g1 (a)

The requirement that g be smooth and invertible ensures that g ) (u) /= 0 for all u + [c, d] and
that g is either monotone increasing (in which case g 1 (a) = c and g 1 (b) = d) or monotone
decreasing (in which case g 1 (a) = d and g 1 (b) = c).
5b
The procedure for evaluating the integral f (x)d x by means of the substitution x = g(u) involves
a
the following four steps:

1. Replace x with g (u) .

2. Replace d x with g ) (u) du.
3. Find the limits of integration for u (i.e. determine g 1 (a) and g 1 (b)).
4. Evaluate the new integral.
 6

Example 12.6.2

Evaluate the integral

=1
x3
T d x.
1  x2
1

Solution.
Put
H H
x = sin u for all u + [ , ].
2 2
(Note that sin is a smooth function that is invertible on [ H2 , H2 ]).
Now
dx
= cos u,
du
so we should replace dx with cos u du.
Now we determine the limits for u:
H
x = 1 " u = sin1 (1) = 
2
and
H
x = 1 " u = sin1 (1) = .
2
Hence
H
=1 =2
x3 sin3 u
T dx = S cos u du
1  x2 1  sin2 u
1  H2
H
=2 K H HL
= sin3 u du since cos u o 0 for u +  ,
2 2
 H2
H
=2
b c
= sin u 1  cos2 u du
 H2
v w H2
1
= cos3 u  cos u
3  H2
= 0.
 7 MAT215–V/3
APM212–W/3

Exercises 12.6.3 Revision of
rst year work

1. Evaluate the following integrals.

51 x
(a) T d x.
0 1  x2
53y2
(b) S dy.
0 9  y2
52 T
(c) x 2 x 3 + 1 dx.
1
5 H4
(d) 0 tan x sec2 x d x.
5H cos x
(e) T dx.
H 4 + 3 sin x
52
ln
(f) xe x dx.
0

2. Use single integrals to calculate the area of the region enclosed by the curves
T
y = x 2  1 and y = 1  x 2 .

3. Use single integrals to calculate the area of the region enclosed by the curves
y = x 2  2 and y = x.

4. Use a single integral to derive a formula for the area of a circle.

8
Chapter 13

Double Integrals

13.1 INTRODUCTION

In this chapter we consider the integral of an R2  R function over a bounded region in R2 .

Such an integral is called a double integral. Double integrals, like single integrals, have important
practical applications in various subjects. In order to model practical problems correctly in terms
of integrals, it is crucial to have an intuitive understanding of the meaning of different types of
integrals. In this chapter we illustrate the meaning of double integrals and also provide techniques
for evaluating such integrals.

13.2 DEFINITION OF A DOUBLE INTEGRAL

Just as the de¿nition of a single integral is motivated by the problem of computing areas, the
de¿nition of a double integral is motivated by the problem of computing volumes.
55
The following steps lead to the de¿nition of the double integral f (x, y)d A, where f is an
R
R2  R function and R is a bounded region in R2 . (The concept of a bounded region is explained
in Section 2.13.)

1. Consider an R2  R function f de¿ned on a bounded region R in the XY-plane.

2. Form a partition of R into subregions by placing over R a grid of lines parallel to the X-and
Y-axes. Consider only the rectangles of the grid that are completely contained in R and
number them R1 , ..., Rn . (In general, the ¿ner the grid, the larger the part of R that will be
covered by these rectangles.)

9
 10

R1
(x k ,yk )

Rk

R X

Rn

Fig. 13.1

3. Choose a point (xk , yk ) in Rk , for k = 1, ..., n.

4. Let Ak denote the area of the rectangle Rk ; k = 1, ..., n and let PPP be the length of the
longest of the diagonals of the rectangles.

z = f(x,y)

f(xk ,yk )

Dk

(xk ,yk )

Y
R

X
Fig. 13.2 The volume of Dk is f (xk , yk )Ak
 11 MAT215–V/3
APM212–W/3

5. Form the sum

;
n
f (xk , yk )Ak
k=1

and consider the limit of this sum as PPP tends towards zero.

De¿nition 13.2.1 Let f be an R2  R function de¿ned over a bounded region R in the XY-plane.
The double integral of f over R is given by
== ;
n
f (x, y)d A = lim f (xk , yk )Ak ,
PP P0
k=1
R

provided this limit exists.

If this limit exists, we say that f is integrable over R.

If f (x, y) o 0 for every (x, y) + R then, for each k = 1, . . . , n the value of f (xk , yk )Ak gives
the volume of the rectangular prism Dk with Rk as base and height f (xk , yk ). Now let D be the
3
region below the graph of f that lies vertically above R. Then the sum nk=1 f (xk , yk )Ak gives
an approximation of the volume of D. Generally, the smaller the norm of the partition, the better
the approximation. If the limit in De¿nition 13.2.1 exists, then it gives the exact volume of D.

We have seen that an R  R function is integrable over a closed interval in R if it is continuous

over that interval. For R2  R functions we have a similar situation.

Theorem 13.2.2 If f is an R2  R function that is de¿ned and continuous over a bounded region
R, whose boundary consists of a ¿nite number of smooth curves, then f is integrable over R.

13.3 GEOMETRIC INTERPRETATION OF DOUBLE INTEGRALS

Interpretation as a volume

The relationship between double integrals and volumes is similar to that between single integrals
and areas.

Let f be an R2  R function that is integrable over a bounded region R in the XY-plane. If D is

the three-dimensional region below the graph of f that lies vertically above the region R, then the
following holds:
 12

==
If f (x, y) o 0 for all (x, y) + R, then f (x, y)d A = volume of D. (13.1)
R
==
If f (x, y) n 0 for all (x, y) + R, then f (x, y)d A =  (volume of D) . (13.2)
R

Z
z = f(x,y)

D Z

R R
Y Y
X
X D

z = f(x,y)

55 55
(a) Volume of D = f (x, y)d A (b)Volume of D =  f (x, y)d A
R R

Fig. 13.3

Interpretation as an area.

We know that the volume of a prism is obtained by multiplying its height with the area of its base.
Thus if the height of a prism is 1 unit, then its volume has the same numerical value as the area of
its base. If we integrate the function f (x, y) = 1 over a bounded region R in the XY-plane, we
therefore obtain the area of R, i.e.
55
1 d A = area of R. (13.3)
R
 13 MAT215–V/3
APM212–W/3

z=1

R Y

X
55
Fig. 13.4 1 d A = area of R
R

13.4 RULES OF INTEGRATION

The Rules of Integration for double integrals are similar to those for single integrals.

Theorem 13.4.1 Rules of Integration

If f and g are both R2 R functions which are integrable over a bounded region R in the XY-plane,
then:
55 55 55
1. [ f (x, y) + g(x, y)]d A = f (x, y)d A + g(x, y)d A.
R R R
55 55
2. k f (x, y)d A = k f (x, y)d A.
R R

3. If R = R1 C R2 , where R1 and R2 are regions with no points other than boundary points in
common, then

== == ==
f (x, y)d A = f (x, y)d A + f (x, y)d A.
R R1 R2

R1 R2

Fig. 13.5 Union of two regions with only boundary points in common
 14

13.5 EVALUATION OF DOUBLE INTEGRALS.

In this section we shall show how to integrate a continuous R2  R function f over a bounded
region in the XY-plane, by making use of the integration techniques for single integrals. We shall
begin with rectangular regions and then proceed to bounded regions of a more general nature.
Let T be the rectangle in the XY-plane de¿ned by

T = {(x, y) | a n x n b and c n y n d} ,

where a, b, c and d are constants.

T
a b X
c

Fig. 13.6 The rectangle T = {(x, y) | a n x n b and c n y n d}

Let f be a continuous R2  R function de¿ned on T.

For each x + [a, b] we now de¿ne a function f x of y alone, by

f x (y) = f (x, y) for all y + [c, d].

Now f x is continuous and hence integrable over [c, d] for every x + [a, b]. Thus, for every
x + [a, b] we can consider the single integral

= d
f x (y)dy.
c

To evaluate this integral we regard x as a constant and integrate f x (y) with respect to y.
We now de¿ne an R  R function F of x by

= d
F(x) = f x (y)dy for every x + [a, b].
c

Note that, after having substituted the limits of integration for y in the integral above, we are left
with a function of x alone. It can be proved that F is integrable over [a, b], so we may consider
5b
the single integral a F(x)d x.
 15 MAT215–V/3
APM212–W/3

Now = = t= u
b b d
F(x)dx = f x (y)dy dx.
a a c

This integral is called a repeated integral (or an iterated integral) of f over T and we also write
it as = = b d
f (x, y)dy dx
a c
In a similar way we de¿ne the repeated integral
= d= b = d t= b u
f (x, y)dx dy = f y (x)dx dy .
c a c a

The following theorem is named after the Italian mathematician Guido Fubini (1879 - 1943), who
proved a more general version of this result.

Theorem 13.5.1 Fubini’s Theorem

Let f be an R2  R function and let T be a rectangle in the domain of f de¿ned by

T = {(x, y) | a n x n b and c n y n d} .
55
If f is continuous over T, then the double integral f (x, y)d A exists, and
T
== = b = d = d = b
f (x, y)d A = f (x, y)dy dx = f (x, y)d x dy.
a c c a
T

We shall not present the precise proof of Theorem 13.5.1 here, but we illustrate the proof by
means of sketches. In Fig. 13.7 the area of the shaded surface Ak in (a) equals
= d
f xk (y)dy.
c

The volume of the shaded solid Dk in (b) equals

t= d u
f xk (y)dy xk
c

and the volume of the solid D in (c) is given by

; n t= d u = b t= d u
lim f xk (y)dy xk = f xk (y)dy dx.
PP P0 c a c
k=1
5b5d
Thus the repeated integral a c f (x, y)dy dx gives the volume of D. But the double integral
55
f (x, y)d A also gives the volume of D, showing that
T
== = b = d
f (x, y)d A = f (x, y)dy d x.
a c
T
 16

Z
surface z = f(x,y)
curve z = f xk(y)

Ak
d
a Y
xk
b y=d
X
5d
(a) Area of Ak is c f xk (y)dy

Dk
d
a Y
∆x k
∆x k
b
X
r5 s
d
(b) Volume of Dk is c f xk (y)dy xk

D d
a Y

b
X
n r
; s
5d 5b5d
(c) Volume of D is limPPP0 c f xk (y)dy xk = a c ( f xk (y)dy)dx
k=1
Fig. 13.7
 17 MAT215–V/3
APM212–W/3

Theorem 13.5.1 allows us to evaluate a double integral over a rectangle by successive evaluation
of single integrals.

Example 13.5.2

55
Evaluate the double integral f (x, y)d A, where f (x, y) = x y 2 and T is the rectangle in the
T
sketch below.

T
-1 2 Y
-1

Fig. 13.8 T = {(x, y) |  1 n x n 2; 1 n y n 3}

Solution. Since f is a continuous function, it follows from Theorem 13.5.1 that

== = 3 = 2
f (x, y)d A = x y 2 dy dx
1 1
T
= 3 w2 v
y3
= x dx
1 3 1
= 3 t u
8 1
= x + dx
1 3 3
= 3
= 3x d x
1
v w3
x2
= 3
2 1
= 12 .
 18

We now consider the case where the region of integration is not necessarily rectangular, but can
be described in the form

G = {(x, y) | a n x n b and g1 (x) n y n g2 (x)},

where a and b are constants , and g1 and g2 are smooth functions of x such that

g1 (x) n g2 (x) for all x + [a, b].

Y
y = g2 (x)

a b X

y = g1(x)

Fig. 13.9 A Type 1 Region

The lower and upper boundaries of the region G are formed by the smooth curves y = g1 (x) and
y = g2 (x), respectively, while the left and the right boundaries of G are formed by the vertical
lines x = a and x = b, respectively. We call such a region a Type 1 Region.
For a Type 1 region the limits for x are constants, but the limits for y may depend on x.
The following generalization of Fubini’s Theorem is useful for evaluating a double integral over a
Type 1 Region.

Theorem 13.5.3 If f is an R2  R function that is de¿ned and continuous on a Type 1 Region G,

de¿ned by
G = {(x, y) | a n x n b and g1 (x) n y n g2 (x)}
55
then the double integral f (x, y)d A exists and
G

== =b g=2 (x)
f (x, y)d A = f (x, y)dy dx.
G
a g1 (x)
 19 MAT215–V/3
APM212–W/3

Remarks 13.5.4

1. The repeated integral in Theorem 13.5.3 is calculated by ¿rst considering x as constant and
integrating with respect to y. After substituting the integration limits g1 (x) and g2 (x) for y,
the integrand becomes a function of x alone and we then integrate normally with respect to
x. Since the limits for x are constants, our ¿nal answer will be independent of x and y.

2. The order of integration “dy dx” indicates that we ¿rst integrate with respect to y, and then
with respect to x.

Example 13.5.5

Evaluate
==
(x + 2y)d A,
G

where G is the region inside the circle x 2 + y 2 = 2, between the y-axis and the line x = 1.
Solution. We ¿rst sketch the region of integration.

Y
y= 2-x2

G 1
0 X

y=- 2-x 2

Fig. 13.10

T
The lower boundary for G is the smooth curve y =  2  x 2 and the upper boundary is
T
the smooth curve y = 2  x 2 . The left boundary is the vertical line x = 0 and the right
boundary is the vertical line x = 1. Thus G is a Type 1 Region, and can be described as
S S
G = {(x, y) | 0 n x n 1 and  2  x 2 n y n 2  x 2 }.
 20

Thus, according to Theorem 13.5.3,

T
== =1 =2x
2

(x + 2 y)d A = (x + 2 y)dy d x
G 0
T
 2x 2

=1 T
d e 2x 2
= x y + y T 2 dx
2
 2x
0
=1 S
= 2x 2  x 2 d x
0
v w1
2
=  (2  x )2 3/2
3 0
2 T
= (2 2  1).
3

Now we consider a region of integration of the form

H = {(x, y) | c n y n d and h 1 (y) n x n h 2 (y)},

wherec and d are constants, and h 1 and h 2 are smooth functions of y such that

h 1 (y) n h 2 (y) for all y + [c, d ].

x = h1(y) H x = h2 (y)

X
c

Fig. 13.11 A Type 2 Region

 21 MAT215–V/3
APM212–W/3

The left and the right boundaries of the of the region H are formed by the smooth curves x =
h 1 (y) and x = h 2 (y), respectively, while the lower and upper boundaries are formed by the
horizontal lines y = c and y = d , respectively. Such a region is called a Type 2 Region. The
following generalization of Fubini’s Theorem is useful for evaluating a double integral over a Type
2 Region.

Theorem 13.5.6 If f is an R2  R function that is de¿ned and continuous on a Type 2 Region H ,

where
H = {(x, y) | c n y n d and h 1 (y) n x n h 2 (y)},

then f is integrable over H and

== =d h=2 (y)
f (x, y)d A = f (x, y)d x dy.
H
c h 1 (y)

Remarks 13.5.7

1. The repeated integral in Theorem 13.5.6 is calculated by ¿rst considering y as constant and
integrating with respect to x . After substituting the limits of integration for x , the integrand
becomes a function of y alone and we then integrate normally with respect to y . Since the
limits for y are constants, our ¿nal answer will be independent of x and y .

2. The order of integration “d x dy ” indicates that we ¿rst integrate with respect to x , and then
with respect to y.

Example 13.5.8

Evaluate the integral

==
(4x  y)d A,
H

where H is the region in the XY-plane that lies inside the circle x 2 + y 2 = 4, above the X-axis
and below the line y = x.
S
Solution. The semicircle x = 4  y 2 and the line x = y intersect where
S
y = 4  y2
T T
i.e. y = 2 and x = 2.
 22

( 2, 2 )
x=y H x = 4 - y2
0 X

Fig. 13.12

The region of integration, H , is a Type 2 Region, since the left boundary is the smooth curve x =
S
y , its right boundary is the smooth curve x = 4  y 2 , and the lower and upper boundaries
T
are the horizontal lines y = 0 and y = 2, respectively.
Thus

T S
H = {(x, y) | 0 n y n 2 and y n x n 4  y 2 }.

Hence, according to Theorem 13.5.6,

T T
== = 2 =4y 2

(4x  y)d A = (4x  y)dx dy

H 0 y
T
=2 T
d 2 e 4y 2
= 2x  yx y dy
0
T
=2 r S s
= 8  3 y 2  y 4  y 2 dy
0
v wT2
1
= 8 y  y + (4  y )
3 2 3/2
3
T 0
20 2  8
= .
3
 23 MAT215–V/3
APM212–W/3

Reversing the Order of Integration

Sometimes it is easier to evaluate a given repeated integral if we reverse the order of integration.

Example 13.5.9

Evaluate the repeated integral

T T
= 2 =4y 2
1
T d x dy .
4  x2
0 y

Solution. The region of integration of this integral is the region H in Fig. 13.12. Since integrating
T 1 with respect to x is laborious, we would prefer to integrate with respect to y ¿rst, i.e. we
4x2

wish to use Theorem 13.5.3, with the order of integration “dy dx ”.

However, H is not a Type 1 region, because its upper boundary is not smooth, since a sudden
T
change occurs where x = 2.
T
To the left of the vertical line x = 2 the upper boundary is formed by the smooth curve y = x .
T T
To the right of the line x = 2 it is formed by the smooth curve y = 4  x 2 .
Thus H is not a Type 1 region, but it is the union of two Type 1 regions, H1 and H2 where

T
H1 = {(x, y) | 0 n x n 2 and 0 n y n x}

and

T S
H2 = {(x, y)| 2 n x n 2 and 0 n y n 4  x 2 }.

y=x y = 4-x 2

H1 H2
0 2 2 X
y=0
Fig. 13.13

Since H = H1 C H2 , Rule 3 of Theorem 13.4.1 implies

== == ==
1 1 1
T dA = T dA + T d A.
4  x2 4  x2 4  x2
H H1 H2

Thus
 24

T T T T
= 2 =4y 2 = 2=x =2 =4x
2

1 1 1
T dx dy = T dy dx + T dy dx
4  x2 4  x2 T 4  x2
0 y 0 0 2 0
T T
= 2v wx =2 v w 4x 2
y y
= T dx + T dx
4  x2 0 T 4  x2 0
0 2
T
=2 =2
x
= T dx + 1 d x
4  x2 T
0 2
K S LT2 T
=  4x 2 + [ x ]2 2
T 0
= 4  2 2.

Remarks 13.5.10 Hints on reversing the order of integration

1. If the region of integration is a rectangle, then the limits of integration are all constants.
In this case the limits of the inner and outer integrals are simply interchanged when the
order of integration is reversed. Thus, if f is de¿ned and continuous on the rectangle
T = {(x, y) | a n x n b and c n y n d}, then

=b =d =d =b
f (x, y)dy d x = f (x, y)dx dy .
a c c a

5b g25(x)
2. When we want to reverse the order of integration of the integral f (x, y)dy dx , where
a g1 (x)
the limits of integration for y are dependent on x , then we cannot simply interchange the
limits of integration. (Remember, the limits of the outer integral must be constants!) In
this case we need to describe the region of integration as a Type 2 region, or as a union of
Type 2 regions. To achieve this we need to sketch the region and rewrite the inequalities
that describe the region in such a way that the limits for x are smooth functions of y , and
the limits for y are constants.
5d h 52 (y)
3. When we want to reverse the order of integration of the integral f (x, y)d x dy , where
c h 1 (y)
the limits of integration for x are dependent on y , we need to describe the region of integra-
tion as a Type 1 region, or as a union of Type 1 regions. To achieve this we need to sketch
the region and rewrite the inequalities that describe the region in such a way that the limits
for y are smooth functions of x , and the limits for x are constants.

4. When we want to describe a Type 1 region it is useful to shade the region by means of
vertical lines in order to note which curves form the lower and upper boundaries for y (in
terms of x).

5. When we want to describe a Type 2 region we shade it by means of horizontal lines in order
to note which curves form the left and right boundaries for x (in terms of y).
 25 MAT215–V/3
APM212–W/3

Example 13.5.11

Write the repeated integral

=1 =x 2
y dy dx
1 0

as a sum of repeated integrals in which the order of integration is reversed.

Solution. We ¿rst sketch the region of integration.

y = x2

-1 1 X
y=0
Fig. 13.14(a)

We can shade the whole region by drawing vertical lines from the line y = 0 up to the curve
y = x 2 , between the vertical lines x = 1 and x = 1.
When we want to reverse the order of integration, the limits of integration for x must be written
in terms of y , and the limits for y must be constants. In order to determine the left and right
boundaries, we shade the region of integration by means of horizontal lines.

x= 1 x= y x= y x=1
H1 H2
0 X
Fig. 13.14(b)

For the region H1 the left boundary is the line x = 1 and the right boundary is the curve
T
x = y.
T
For the region H2 , the left boundary is the curve x = y , and the right boundary is the line x = 1.
Both H1 and H2 lie between the horizontal lines y = 0 and y = 1.
Thus
T
=1 =x 2 =1 
= y = 1 =1
y dy dx = y dx dy + y d x dy.
T
1 0 0 1 0 y
 26

Exercises 13.5.12

1. In each of the following, sketch the region R and calculate the double integral
==
f (x, y)d A,
R

using the order of integration that you ¿nd most convenient.

(a) f (x, y) = 2x + y; R = {(x, y) | 1 n x n 4 and 2 n y n 4}.
S
(b) f (x, y) = x + y; R = {(x, y) | 0 n y n 2 and 0 n x n 4  y 2 }.
(c) f (x, y) = 2(x + y) R = {(x, y) |  H2 n x n H2 and 0 n y n cos x}.
(d) f (x, y) = sin(x + y); R = {(x, y) | x o 0, y o 0 and x + y n H}.
j T k
(e) f (x, y) = ye x R = (x, y) | 0 n y n 1 and y n x n y .

2. Let R be the region in the XY-plane above the curve y = x 2 , below the line y = x + 2.
55
Evaluate the integral (x + y)d A.
R

3. Express each of the following repeated integrals as a repeated integral in which the order of
integration is reversed (or as a sum of such integrals, if necessary). Check your answers by
evaluating the integrals both ways.

51 5x
(a) ydy dx.
0 x2

52 4x
5 2
(b) 1dy d x.
2 0

51 5x 2 b c
(c) x 2 + 2x y dy dx.
0 0
T
51 51x
2

(d) xdy d x.
0 1x 2

52 5y
(e) (x + y)dx dy.
1 y
 27 MAT215–V/3
APM212–W/3

13.6 TRANSFORMATION OF COORDINATES

Consider the transformation

x = T1 (u, v) , y = T2 (u, v) ; (u, v) + Q

and the corresponding R2  R2 function T , de¿ned by

T = (T1 , T2 ) .

We can regard T as transforming the region Q in the UV plane into a region R in the XY-plane. If
T is an invertible function, then the above transformation establishes a one-to-one correspondence
between the points in R and the points in Q . (Study Theorem 1.3.15, Study Guide 1)

V Y

Q T
R
U X

Fig. 13.15

If (x, y) are the rectangular coordinates of a point in the XY-plane, we can think of (u, v) as
specifying the same point, since by knowing (u, v) we can determine (x, y) from the equations
x = T1 (u, v) and y = T2 (u, v) .

The evaluation of a given integral can often be simpli¿ed by a change of variables. For double
integrals we have a change of variables formula which is similar to the one for single integrals
(see Theorem 12.6.1), with the absolute value of the Jacobian taking the place of the ordinary
derivative.

Theorem 13.6.1 Change of Variables Formula

Suppose each of R and Q is a bounded, 2-dimensional region, whose boundary consists of a ¿nite
number of smooth curves, and T is a smooth, invertible R2  R2 function of u and v that maps Q
onto R. If f is an R2  R function of x and y that is integrable over R, then
== ==
b cn n
f (x, y) d A = f T (u, v) ndet JT (u, v)n du dv.
R Q
 28

n n
The factor ndet JT (u, v)n in the Change of Variables Formula compensates for the change in area
brought about by the transformation (x, y) = T (u, v). Let us explain this brieÀy.
55
In de¿ning the double integral f (x, y) d A, a rectangular grid is placed over the region R in the
R
XY-plane (see Fig. 13.1). This grid consists of horizontal lines, called x -lines and vertical lines,
called y-lines. On each x -line y is constant and on each y -line x is constant. We now replace the
rectangular grid with a uv -grid, consisting of u -curves (where v is constant) and v -curves (where u
is constant). Consider the point P = (x, y) and the region with an area of A shown in Fig. 13.16.
Let u and v be the width and length of the corresponding rectangle in the UV-plane.

Y
u - curve

a
∆A
P
b

v - curve

o o
Fig. 13.16 A r oa × bo
r s
"y
At the point P , the vector derivative ""ux , "u is tangent to the u -curve through P and the vector
r s
"y
derivative ""vx , "v is tangent to the v -curve through P . Now let

t u t u
"x "y " x "y
a= , u and b = , v.
"u "u "v "v

Then A is approximately equal to the area of the parallelogram determined by the vectors a and
b. From the geometric interpretation of the cross product (see Section 2.10) we know that the area
o o
of this parallelogram equals oa × bo (see Fig. 2.18). Hence
o o
A r oa × bo .

Now
u
t t u
" x "y " x "y
a×b = , u × , v
"u "u "v "v
t u t u
" x "y " x "y
= , × , uv.
"u "u "v "v
 29 MAT215–V/3
APM212–W/3

But, by De¿nition 2.10.3,

n n
n "x "y n
t u t u n n
" x "y " x "y n "u "u n
n n
, × , = n nk
"u "u "v "v n n
n "x "y n
n n
"v "v

= det JT (u, v) k.

Thus
o o o o
oa × bo = odet JT (u, v) k u v o
n no o
= ndet JT (u, v) u v n ok o by N.3, Section 2.5
n n
= ndet JT (u, v)n u v
o o
since ok o = 1, and u and v are positive. Hence
n n
A r ndet JT (u, v)n u v.

It can be shown that the limit of the sum

; b cn n
f T (u, v) ndet JT (u, v)n u v

taken over the entire region R gives the integral

==
b cn n
f T (u, v) ndet JT (u, v)n du dv.
Q

Remark 13.6.2

det JT (u, v) is called the Jacobian of the transformation (x, y) = T (u, v) . It is also denoted by
" (x, y)
.
" (u, v)

Calculation of an integral by means of a coordinate transformation

55
Evaluating an integral of the form f (x, y) d A by using the transformation
R

(x, y) = (T1 (u), T2 (v))

involves the following four steps.

1. Replace x with T1 (u, v) and y with T2 (u, v) .

n n
2. Replace d A with ndet JT (u, v)n du dv.
3. Describe the region Q in terms of u and v and ¿nd the limits of integration for u and v.
4. Evaluate the new integral.
 30

Polar coordinates

Evaluation of certain double integrals can often be simpli¿ed by changing to polar coordinates.

De¿nition 13.6.3 Polar coordinates represent a point p in R2 by means of an ordered pair (r, A),
where

(1) r is the distance from p to the origin,

(2) A is the angle that the vector p makes with the positive X-axis (measured anticlockwise).

If (x, y) are the rectangular coordinates of p , and (r, A ) the polar coordinates, then it is clear from
Figure 13.17 that

x = r cos A, y = r sin A . (13.4)

Moreover,
S
r= x 2 + y2.

y (x,y) = (rcosθ, rsinθ )

r rsinθ

θ
0 rcosθ x X

Fig. 13.17 Relationship between polar and rectangular coordinates

 31 MAT215–V/3
APM212–W/3

Remarks 13.6.4

1. Note that r is never negative, since it represents a distance.

2. If c is a constant, then the equation r = c represents a circle in the XY-plane.

3. If : is a constant, then the equation A = : de¿nes a ray from the origin in the XY-plane.

θ=α
Y Y

c
r=c
α

-c c X
X

-c

Fig. 13.18 (a) The circle r = c (b) The ray A = :

Thus a rectangular grid in the r A -plane corresponds to a curved grid in the XY- plane, as shown
in Fig. 13.19. The A-curves are circles (since r remains constant on a circle) and the r-curves are
rays drawn from the origin (since A remains constant on such a ray).

θ θ=π
2
Y
2π
θ = 3π
4 θ=π
3π 4
2

π r=3
r=2
π r=1
2 θ=π θ=0
X
0
1 2 3 r

θ = 5π
4 θ = 7π
4

θ = 3π
2

Fig. 13.19 A rectangular grid in the r A plane and the corresponding grid in the XY-plane
 32

Now consider the region R in Fig. 13.20. This region can be described in terms of polar coordinates
as
R = {(r, A) | A 1 n A n A 2 and r1 (A) n r n r2 (A )} ,

where A 1 and A 2 are constants, with

0 n A 1 n A 2 n 2H

and r1 and r2 are both smooth functions of A, with

0 n r1 (A ) n r2 (A) for all A + [A 1 , A 2 ].

Y θ = θ1
r = r2(θ)

R
r = r1(θ)
θ1
θ2
X

θ = θ2

Fig. 13.20 The region R = {(r, A) | A 1 n A n A 2 and r1 (A) n r n r2 (A )}

Remark 13.6.5

Note that neither of the curves bounding the region R in the sketch above is the graph of a function
of x (because on each curve some x -value corresponds to more than one y -value) and hence it is
not easy to describe this region by means of rectangular coordinates. However, r1 and r2 are
functions of A , since each value of A corresponds to only one value of r on each of these curves.

Consider the polar coordinate transformation

(x, y) = P (r, A)

where
P (r, A) = (r cos A, r sin A)

with
r o 0 and A + I,

where I is an interval of length at most 2H. Then P is a smooth R2  R2 function. The restriction
on the length of I ensures that P is one-to-one, and hence invertible. (We usually take I to lie
 33 MAT215–V/3
APM212–W/3

in the interval [0, 2H ], but if negative angles are involved, we can take it to lie in [H, H ], for
example.)
Now
The Jacobian of this transformation is

n n
n "x n"x
n n
n "r n"A
n n
= n n
n n
n " y "y n
n n
n "r "A n
n cos A r sin A n
n n
= n n
n sin A r cos A n
= r cos2 A + r sin2 A
= r.

Hence
n n
ndet J P (r, A)n = r. (13.4)

Let f be an R2  R function which is continuous on R . From Theorem 13.6.1 we obtain the

Transformation Rule for Polar Coordinates:

55 5A 2 r25(A) b c
f (x, y) d A = f P (r, A ) r dr dA . (13.6)
R A 1 r1 (A )

Examples 13.6.6

1. Evaluate the integral

T T
= 2 =4y 2

(4x  y)d x dy.

0 y

Solution.
Note that this is the integral of Example 13.5.8. Since the region of integration is a cir-
cle sector, polar coordinates are very suitable for describing this region, so we proceed as
follows:
Step 1. We put
x = r cos A and y = r sin A.

The integrand then becomes

4x  y = 4r cos A  r sin A.

Step 2. Since | det J P (r, A)| = r , we replace d A with r dr dA . The integral then becomes
== ==
(4r cos A  r sin A) r dr dA = r 2 (4 cos A  sin A) dr dA,
R R

where R denotes the region of integration.

 34

Step 3. In order to describe the region R by means of polar coordinates, we shade the region
by drawing rays from the origin, starting on the X-axis (i.e. where A = 0) and ending on the
line y = x, i.e. where

r sin A = r cos A,
i.e. tan A = 1,
H
i.e. A = .
4
S
The length of each of theses rays is 2 units (since every point on the semicircle x = 4  y2
lies at a distance of 2 units from the origin).

r=2
π
4

2
r=0 X

Fig. 13.21

Thus the region R can be described by

Q H R
(r, A ) | 0 n A n and 0 n r n 2 .
4
Step 4.
T T H
=2 =4y 2 = 4 =2
(4x  y) d x dy = r 2 (4 cos A  sin A) dr dA
0 y 0 0
H
= v 4 w2
r3
= (4 cos A  sin A) dA
3 0
0
H
=4
8
= (4 cos A  sin A) dA
3
0
8 H
= [4 sin A + cos A ]04
3t u
8 4 1
= T + T 1
3 2 2
T
20 2  8
= .
3
 35 MAT215–V/3
APM212–W/3

2. Evaluate the integral

==
(x 2 + y 2 )3/2 d A,
R
T
where R is the region in the XY-plane below the graph of y = 2  x 2 , above the line
T
y = 1 and between the lines y = x and y =  3x.

Solution.

The integrand, (x 2 +y 2 )3/2 , is laborious to integrate with respect to x as well as with respect
to y , but it becomes simpler if we set x = r cos A and y = r sin A, since then x 2 + y 2 = r 2 .
For that reason it seems worthwhile to switch to polar coordinates.

Step 1. We set
x = r cos A, y = r sin A.

Then the integrand becomes

(r 2 cos2 A + r 2 sin2 A)3/2 = r 3 .

n n
Step 2. Since ndet JF (r, A)n = r, the integral becomes
== ==
b 3 c
r r dr dA = r 2 dr dA.
R R

Step 3. Let us sketch the region R.

y = 2 - x2

R
y=1

y= 3x y=x

2 X

Fig.13.22(a)
 36

We now shade the region R by means of rays emanating form the origin.

Y
2π
θ=
3 r= 2 π
2 θ= 4

R
1
r = cosecθ
y = 3x y=x

O X

Fig. 13.22(b)
T
R lies between the ray y = x and the ray y =  3x . Hence the lower limit for A is given
by

y = x,
i.e. r sin A = r cos A
i.e. tan A = 1
H
i.e. A =
4

The upper limit for A is given by

T
y =  3x,
T
i.e. r sin A =  3r cos A
T
i.e. tan A =  3
2H
i.e. A = .
3
T H
This agrees with our sketch, since the lines y = x and y =  3x make angles of 4
and
2H
3
, respectively, with the positive X-axis.

Since rays from the origin enter the region R on the line y = 1, the lower limit for r is given
by

y = 1,
i.e. r sin A = 1,
i.e. r = cosec A.
 37 MAT215–V/3
APM212–W/3

The upper limit for r is given by the curve

x 2 + y 2 = 2,
i.e. r 2 = 2,
T
i.e. r = 2.

The limits for r may also be determined geometrically. Consider a ray that makes an angle
of A units with the positive X-axis, where H4 n A n 23H .Let B be the point where this rays
enters the region R , and C the point where it exits R .

2
C

Α θ Β
1
θ
ec
cos

θ
Ο X

Fig. 13.22(c)

By considering the triangle O AB , we see that

OA 1
sin A = = ;
OB OB
hence
O B = cosec A.
T
Since C lies on a circle with radius 2, centred at the origin,
T
OC = 2.

Thus we can describe the region in terms of polar coordinates as

| }
T H 2H
R = (r, A) | cosec A n r n 2 and n A n .
4 3

Remark. All points on a circle centred at the origin lie at the same distance from the origin,
but that is not the case for points on a straight line. That is why, for the given region R , the
lower limit for r depends on A , while the upper limit is a constant.
 38

Step 4.
2H T
== =3 = 2

(x 2 + y 2 )3/2 d A = r 2 dr dA
R H cosec A
4
2H
=3
d 1 eT2
= r cosec A dA
H
4
2H
=3 t u
1
=  T  sin A dA
H
2
4
v w 23H
A
=  T + cos A
2 H
T 4
T
12  5 2H + 12 2
= .
24

Exercises 13.6.7

1. Describe the following regions by means of polar coordinates.

(a) (b)
Y Y

X -1 1 X
-1

(c) (d)
Y Y
2
1

-1 X 2 X
-1

2. Describe the following regions by means of polar coordinates.

(a) The region R = {(x, y)|1 n y n 2 and y n x n y}.

S
(b) The region enclosed by the graph of x = |y| and the semi-circle x = 1  y2.
 39 MAT215–V/3
APM212–W/3

3. Use polar coordinates to calculate the following double integrals:

T
52 54y 2
2

(a) (x + y 2 )3/2 d x dy.

0 0
50 50
(b) sin(x 2 + y 2 )dy dx.
3
T
 9x 2
T
51 51y (x 2 +y 2 )
2

(c) e dx dy.
1
T
 1y 2
T
51 51y
2
S
(d) x y x 2 + y 2 d x dy.
0 0

13.7 CALCULATION OF VOLUMES BY MEANS OF DOUBLE INTEGRALS

The following examples illustrate how we can calculate the volume of a given 3-dimensional
region by using the formulas (13.1) and (13.2), given in Section 13.3.

Examples 13.7.1

1. Let f be the R2  R function de¿ned by

f (x, y) = 4  x 2  y 2

and let R be the triangle in the XY-plane with vertices (0, 0), (1, 0) and (0, 1). Determine
the volume of the region D under the graph of f that lies vertically above R .
Solution.
Let us sketch the region D.

Z
z = 4 - x2 - y 2
4

1
1 R Y

X
Fig. 13.23
 40

R is a Type 1 region and can be described as

R = {(x, y) | 0 n x n 1 and 0 n y n 1  x} .

By the formula (13.1),

==
volume of D = f (x, y)d A
R
=1 =1x
b c
= 4  x 2  y 2 dy dx
0 0
=1 v w1x
1
= 4y  x y  y3
2
dx
3 0
0
=1 t u
1
= 4(1  x)  x 2 (1  x)  (1  x)3 dx
3
0
v w1
x3 x4 1
= 4x  2x 
2
+ + (1  x)4
3 4 12 0
11
= .
6

2. Let D be the region in R3 inside the cylinder x 2 + y 2 = 2 that lies above the surface
S
z = x 2 + y 2  2 and below the surface z = 4  x 2  y 2 . Find the volume of D by using
double integrals.
Solution.
Let us sketch the region D.

Z
z = 4 - x 2- y 2
2

D2 R
2 Y
D1
z = x2 + y 2- 2
X -2

Fig. 13.24

Let R be the region in the XY-plane bounded by the circle x 2 + y 2 = 2. Let D1 be the region
above the paraboloid z = x 2 + y 2  2 that lies below R, and let D2 be the region below the
S
hemisphere z = 4  x 2  y 2 that lies vertically above R. Then

volume of D = volume of D1 + volume of D2 .

 41 MAT215–V/3
APM212–W/3

By the formula (13.1) ,

==
b c
volume of D1 =  x 2 + y2  2 d A
R

and by the formula (13.2),

== S
volume of D2 = 4  x 2  y 2 d A.
R

Since the area of integration, R, is bounded by a circle, polar coordinates will be most
suitable for evaluating these integrals. Thus we set

x = r cos A and y = r sin A.

Then
x 2 + y2  2 = r 2  2

and S
S
4  x 2  y2 = 4  r 2

As shown in Section 13.6, the Jacobian of the transformation to polar coordinates is

n n
ndet J P (r, A)n = r.

We can describe R as
Q T R
R = (r, A) | 0 n A n 2H and 0 n r n 2 .

Thus
== == S
b c
volume of D =  x + y 2 dA +
2 2
4  x 2  y2 d A
R R
T T
=2H = 2 =2H = 2S
b 2 c
=  r  2 r dr dA + 4  r 2 r dr dA
0 0 0 0
T T
=2H = 2 =2H = 2
S
b 3 c
=  r  2r dr dA + r 4  r 2 dr dA
0 0 0 0
=2H v wT2 =2H v wT
1 4 1b c3 2
=  r  r2 dA +  4r 2 2
dA
4 0 3 0
0 0
=2H =2H
1r T s
= 1 dA + 8  2 2 dA
3
0 0
2H r T s
= 11  2 2 .
3
 42

Exercises 13.7.2

1. In each of the following cases, use double integrals to calculate the volume of the region
below the graph of f that lies above the region R .

(a) f (x, y) = e x cosy and R = {(x, y) + R2 | 0 n x n ln 3 and  H2 n y n H2 }.

(b) f (x, y) = x + 4 and R is the region in R2 enclosed by the line y = 3x and the curve
y = 4  x 2.
(c) f (x, y) = x 2 + y 2 and R is the region in R2 between the Y-axis and the semicircle
S
x = 1  y2.

2. Calculate the volume of the region inside the cylinder x 2 + y 2 = 4, above the XY-plane and
below the paraboloid z = x 2 + y 2 .

3. Calculate the volume of the region enclosed by the R2  R functions f and g given by

f (x, y) = 8  x 2  y 2 and g(x, y) = x 2 + y 2 .

4. Use a double integral and polar coordinates to deduce a formula for the volume of

(a) a cylinder with radius a and height h;

(b) a sphere with radius a .

13.8 CALCULATION OF AREAS BY MEANS OF DOUBLE INTEGRALS

The following examples illustrate how we can calculate the area of a given region in R2 , by using
the formula ==
Area of R = 1 dA
R

which was given in Section 13.3.

Examples 13.8.1

1. Calculate the area of the region R in R2 enclosed by the curves y = x 2  2 and y = x.

Solution.
Let us sketch the region R.
First, we note that the curves that bound the region intersect where

x 2  2 = x,
i.e. where x = 1 or 2.
 43 MAT215–V/3
APM212–W/3

2
y=x

-1
R 2 X

-2
y = x 2- 2

Fig. 13.25

By shading the region R by means of vertical lines, we note that R is a Type 1 region and
can be described as

j k
(x, y) |  1 n x n 2 and x 2  2 n y n x .

By (13.3),

==
area of R = 1d A
R
= 2 =x
= dy dx
1 x 2 2
=2
d ex
= y x 2 2 d x
1
=2 =
b c
= x  x 2 + 2 dx
1
v w2
1 2 1 3
= x  x + 2x
2 3 1
9
= .
2
 44

2. Use a double integral to derive a formula for the area of a circle.

Solution.

Let R be the circle in the XY-plane with radius a , centred at (0, 0).

a X

Fig. 13.26

We can describe R by means of polar coordinates as

R = {(r, A) | 0 n A n 2H and 0 n r n a} .

By (13.3)

==
area of R = 1d A
R
=2H =a
= r dr dA
0 0
=2H v wa
1 2
= r dA
2 0
0
=2H
1 2
= a dA
2
0
v w2H
1 2
= a A
2 0
= Ha 2 .

Hence the area of a circle with radius a is Ha 2 .

 45 MAT215–V/3
APM212–W/3

Exercises 13.8.2

1. In each of the following cases, use a double integral to calculate the area of the region R.

(a) R = {(x, y) | 0 n x n 2, 0 n y n x 2 }.
T T
(b) R = {(x, y) | 0 n x n 2,  4  x 2 n y n 4  x 2 }.
j k
(c) R = { (r, A) |  H3 n A n H3 and 0 n r n 2 .
T
(d) R is the region in R2 enclosed by the curves y = 1  x 2 and y = 1  x 2.
(e) R is the region in R2 enclosed by the curves y = x 2 , x = 1, x = 1 and y = 1.
(f) R is the region in R2 enclosed by the lines y = x, y = 2x, x = 1 and x = 1.
46
Chapter 14

Triple Integrals

14.1 INTRODUCTION

The de¿nition of double integrals of R2  R functions and their calculation by means of repeated
integrals can easily be generalized to triple integrals of R3  R functions, and even to n -fold
integrals of Rn  R functions, for all natural numbers n.
Since the analogy with double integrals is so close, the treatment of triple integrals will be brief.

14.2 DEFINITION OF A TRIPLE INTEGRAL

The following steps lead to the de¿nition of a triple integral:

1. Consider an R3  R function f de¿ned on a bounded region D in R3 .

(x k ,yk ,zk )

Dk D
Y

X
Fig. 14.1

47
 48

2. Form a partition P of D into subregions by means of a 3-dimensional grid of planes parallel

to the XY-plane, the XZ-plane and the YZ-plane, respectively. Consider only those blocks
that lie completely inside D , and number them D1 , D2 , ..., Dn .

3. Choose a point (xk , yk , z k ) in Dk , for k = 1, ..., n.

4. Let Vk represent the volume of the block Dk , for k = 1, ..., n , and let PPP be the length
of the longest diagonal of the Dk .

5. Form the sum

;
n
f (xk , yk , z k )Vk
k=1
and consider the limit of this sum as PPP tends towards zero.

De¿nition 14.2.1 Let f be an R3  R function that is de¿ned on a bounded region D in R3 . The

triple integral of f over D is given by
=== ;
n
f (x, y, z)d V = lim f (xk , yk , z k )Vk ,
PP P0
k=1
D

provided this limit exists.

If the limit exists we say that f is integrable over D.

The graph of a smooth R2  R function f is called a smooth surface. For triple integrals we have
the following analogue of Theorem 13.2.2.

Theorem 14.2.2 If f is an R3  R function that is de¿ned and continuous on a bounded region

D, the boundary of which consists of a ¿nite number of smooth surfaces, then f is integrable over
D.

14.3 GEOMETRIC INTERPRETATION OF TRIPLE INTEGRALS

In the same manner that a single integral of an R  R function represents an area and a double
integral of an R2 R function represents a volume, a triple integral of an R3 R function can be re-
garded as representing a 4-dimensional “volume” (which, of course, has no immediate geometric
meaning).
55
As shown in Section 13.3, the double integral 1 d A gives the area of a bounded region R in
R
R2 . Similarly, if D is a bounded region in R3 , then
555
1 d V = volume of D. (14.1)
D
 49 MAT215–V/3
APM212–W/3

14.4 RULES OF INTEGRATION FOR TRIPLE INTEGRALS

The rules of integration for triple integrals are similar to those for double integrals. Their formu-
lation is left as an exercise for the reader.

14.5 EVALUATION OF TRIPLE INTEGRALS

Consider a Type 1 region G in the XY-plane, given by

G = {(x, y) | a n x n b and g1 (x) n y n g2 (x)}.

Suppose that h 1 and h 2 are smooth functions of x and y such that

h 1 (x, y) n h 2 (x, y) for all (x, y) + G.

Let D be the region in R3 de¿ned by

D = {(x, y, z) | a n x n b, g1 (x) n y n g2 (x) and h 1 (x, y) n z n h 2 (x, y)}.

We call such a region a Type I Region. It is the 3-dimensional analogue of a 2-dimensional Type
1 Region.

z = h 2(x,y)

z = h1(x,y)

a Y
y = g1(x) G y = g2(x)

b
X
Fig. 14.2 A Type I Region.
 50

Theorem 14.5.1 If f is an R3  R function that is continuous over a Type I Region D, then

555 5b g25(x) h 25(x,y)

f (x, y, z) d V = f (x, y, z) dz dy d x.
D a g1 (x) h 1 (x,y)

Remark 14.5.2

In the formula above the limits of integration for z may depend on both x and y ; those for y
may depend only on x , and the limits for x must be constants. The limits of integration for
a particular variable cannot depend on a variable with respect to which integration has
already taken place.

Example 14.5.3

Evaluate ===
x dz dy d x,
D

where D is the tetrahedron enclosed by the plane x + y + z = 2 and the three coordinate planes.
Solution.

2
z=2-x-y

D 2
Y
2 z=0
y=2-x
X
Fig. 14.3

The tetrahedron D is given by

D = {(x, y, z)|0 n x n 2, 0 n y n 2  x, 0 n z n 2  x  y}.

 51 MAT215–V/3
APM212–W/3

Thus

=== =2 =2x 2xy

=
x dz dy dx = x dz dy dx
D 0 0 0
=2 =2x
2xy
= x [z ]0 dy dx
0 0
=2 =2x
= (2x  x 2  x y)dy dx
0 0
=2 v w2x
1
= 2x y  x 2 y  x y 2 dx
2 0
0
=2 t u
1 3
= 2x  2x + x dx
2
2
0
v w2
2 1
= x2  x3 + x4
3 8 0
2
= .
3

Exercises 14.5.4

1. Formulate the analogue of Theorem 14.5.1 for the case where the order of integration
is dy dx dz . (First describe an appropriate region D.)

2. Evaluate the following triple integrals.

51 5x x+y
5
(a) (x + y  z) dz dy dx.
1 x xy

52 5z 5z
(b) 1dy dx dz.
0 0 0

3. Evaluate the triple integral ===

f (x, y, z) d V
D

where
f (x, y, z) = x + y

and
D = {(x, y, z) | x o 0, y o 0, z o 0 and x + y + z n 3} .
 52

14.6 COORDINATE TRANSFORMATIONS

The change of variables formula for a triple integral is similar to the one for a double integral.

Theorem 14.6.1 Suppose each of D and B is a bounded, 3-dimensional region, whose boundary
consists of a ¿nite number of smooth surfaces, and T is a smooth, invertible R3  R3 function of
u, v and w that maps B onto D. If f is an R3  R function of x, y and z that is integrable over
D, then
=== ===
b cn n
f (x, y, z) d V = f T (u, v, w) ndet JT (u, v, w)n du dv dw.
D B

We shall apply this formula to two special cases, namely to change from rectangular to cylindrical
coordinates and to change from rectangular to spherical coordinates. These two coordinate sys-
tems are especially useful for describing a 3-dimensional region that can be obtained by rotating
a 2-dimensional region about the Z-axis. (Many of the graphs considered in Section 3.2 can be
obtained in this manner.)

Cylindrical Coordinates

Cylindrical coordinates are a natural mixture of polar and rectangular coordinates.

De¿nition 14.6.2 Cylindrical coordinates represent a point p in R3 by means of an ordered triple

(r, A, z) where

1. (r, A) are the polar coordinates of the vertical projection of p onto the XY-plane,

2. z is the vertical rectangular coordinate.

 53 MAT215–V/3
APM212–W/3

(x, y, z) = (rcosθ, rsinθ, z)

y
θ r Y
x

X
Fig. 14.4 Relationship between rectangular and cylindrical coordinates.

If (x, y, z) are the rectangular coordinates of p and (r, A , z) its cylindrical coordinates, then

x = r cos A , u = r sin A, and z remains z . (14.2)

1. The term cylindrical coordinates arises from the fact that if c is a constant then, in R3 , the
equation r = c represents a cylinder with radius c.

2. If : is a constant then, in R3 , the equation A = : de¿nes a vertical plane through the origin.

c Y α Y
X X

Fig. 14.5(a) The cylinder r = c. (b) The plane A = :.

 54

Now consider the cylindrical coordinate transformation

(x, y, z) = C(r, A, z),

where
C(r, A, z) = (r cos A, r sin A, z) ,

with
r o 0, A + I, and z + R,

where I is an interval of length at most 2H.

Clearly, C is a one-to-one R3  R3 function. The Jacobian of this transformation is given by
n n
n " x "x " x n
n n
n "r "A "z n
n n
n n
n n
n "y " y "y n
det(JC (r, A, z)) = nn n
n
n "r "A "z n
n n
n n
n "z "z "z n
n n
n "r "A "z n
n n
n cos A r sin A 0 n
n n
n n
= n sin A r cos A 0 n
n n
n 0 0 1 n

= r cos2 A + r sin2 A
= r.

Hence
n n
ndet(JC (r, A, z))n = r. (14.3)

Now let R be the 2-dimensional region de¿ned by

R = {(r, A) | A 1 n A n A 2 , r1 (A) n r n r2 (A)}

where A 1 and A 2 are constants, with

0 n A 2  A 1 n 2H,

and r1 and r2 are smooth functions of A , with

r1 (A) n r n r2 (A).

Let z 1 and z 2 be smooth functions of r and A , with

z 1 (r, A) n z 2 (r, A) for all (r, A) + R.

and let D be the region de¿ned by

D = {(r, A, z) | A 1 n A n A 2 , r1 (A) n r n r2 (A ), z 1 (r, A) n z n z 2 (r, A)}.

 55 MAT215–V/3
APM212–W/3

Note that R is the XY-projection of D .

From Theorem 14.6.1 we obtain the Transformation Rule for Cylindrical Coordinates: If f
is an R3  R function that is continuous on D , then

555 5A 2 r25(A ) z25(r,A) b c

f (x, y, z) dV = f C (r, A, z) r dz dr dA. (14.4)
D A 1 r1 (A) z 1 (r,A)

Example 14.6.3

Evaluate the integral

===
I = (x 2 + y 2 + z 2 )d V,
D

where Q R
S
D = (x, y, z) | x 2 + y 2 n z n 2 .

Solution.
Step 1. Put

x = r cos A , y = r sin A , z = z.

The integrand then becomes

x 2 + y 2 + z 2 = r 2 cos2 A + r 2 sin2 A + z 2
= r 2 + z2

Step 2. Since | det JC (r, A, z)| = r , the integral becomes

===
I = (r 2 + z 2 )r dz dr dA.
D

Step 3. In order to ¿nd the limits of integration, we sketch the region D. We note that it is bounded
S
below by the cone z = x 2 + y 2 , and bounded above by the plane z = 2.

Z
z=2

D
z = x2 + y2 = r

R 2 Y
2
X
Fig. 14.6(a)
 56

S
The cone z = x 2 + y 2 intersects the plane z = 2 where
S
x 2 + y 2 = 2,

i.e. where
x 2 + y 2 = 4.

Thus the projection of the region D onto the XY-plane is the region R inside the circle in the
XY-plane with equation x 2 + y 2 = 4.

2 X

Fig. 14.6(b)

We can describe R by means of polar coordinates as

R = {(r, A ) | 0 n A n 2H and 0 n r n 2} .

The whole region D is obtained by rotating the region D ` in Fig, 14.6(c) about the Z-axis through
an angle of 2H.

Z
z=2

D*
z=r

Y
Fig. 14.6(c)

By shading the region D ` by means of vertical lines (lines parallel to the Z-axis), we note that
the lower limit for z is given by z = r and the upper limit for z is z = 2. (This is also clear from
Fig. 14.6(a).
Thus we can describe the region D in terms of cylindrical coordinates by

D = {(r, A, z) | 0 n A n 2H and 0 n r n 2 and r n z n 2}.

 57 MAT215–V/3
APM212–W/3

Step 4.

=2H =2 =2
I = (r 2 + z 2 )rdz dr dA
0 0 r
=2H =2 =2
= (r 3 + r z 2 )dz dr dA
0 0 r
=2H =2 v w
1 3 2
= r z + r z ) dr dA
3
3 r
0 0
=2H =2 t u
8 4 4
= 2r + r  r dr dA
3
3 3
0 0
=2H v w2
r 4 4r 2 4r 5
= +  dA
2 3 15 0
0
=2H
24
= dA
5
0
24 2H
[A ]0
5
48H
= .
5

Spherical Coordinates

Spherical coordinates locate points in R3 by means of two angles and a distance.

De¿nition 14.6.4 Spherical coordinates represent a point p in R3 by means of an ordered triple

(I, A , M), where

(1) I is the distance between p and the origin,

(2) M is the angle that the vector p makes with the positive Z-axis,

(3) A is the angle from cylindrical coordinates.

 58

Consider a point in R3 with rectangular coordinates (x, y, z), cylindrical coordinates (r, A, z) and
spherical coordinates (I, A, M).

Z
z

(x,y,z) = ( ρ sinφ cosθ, ρ sinφ sinθ, ρ cosφ)

ρ z = ρ cosφ

φ
y
r=
θ ρs Y
sθ
in
φ
o
rc
=
x

x
y = rsinθ
X

Fig. 14.7 Relationship between spherical, cylindrical and rectangular coordinates.

Using trigonometry, we see that

z = I cos M and r = I sin M.

Since x = r cos A, and y = r sin A, we now have

x = I sin M cos A , y = I sin M sin A, z = I cos M, (14.5)

Furthermore,
S
I= x 2 + y2 + z2.

Remark 14.6.5

1. I cannot be negative, since it represents a distance.

2. The term spherical coordinates arises from the fact that, if c is a constant, the equation
I = c de¿nes a sphere in R3 (with radius c).

3. If : is a constant, the equation M = : de¿nes a cone in R3 . This cone is obtained by taking

a ray that makes an angle of magnitude : with the positive Z-axis and rotating it about the
Z-axis through an angle of 2H (i.e. by letting A vary from 0 to 2H ).
 59 MAT215–V/3
APM212–W/3

Z Z

α
c Y
X Y
X

Fig. 14.8 (a) The sphere I = c. (b) The cone M = :.

Now consider the spherical coordinate transformation

(x, y, z) = S(I, A , M)
where
S(I, A, M) = (I sin M cos A, I sin M sin A, I cos M) ,

with

I o 0, A + I and M + [0, H ],
where I is an interval of length at most 2H.
S is clearly a one-to-one R3  R3 function. The Jacobian for this transformation is
n n
n "x " x " x n
n n
n "I "A "M n
n n
n n
n n
n " y "y "y n
det(JS (I, A , M)) = nn n
n
n "I "A "M n
n n
n n
n "z "z "z n
n n
n n
"I "A "M
n n
n sin M cos A I sin M sin A I cos M cos A n
n n
n n
= n sin M sin A I sin M cos A I cos M sin A n
n n
n cos M 0 I sin M n
= I 2 sin M.

The restriction 0 n M n H ensures that sin M o 0. Hence

n n
ndet JS (I, A, M)n = I 2 sin M. (14.6)
 60

Now let D be the region in R3 described by

j k
D = (I, A, M) | A 1 n A n A 2 , M 1 (A) n M n M 2 (A) and I 1 (A, M) n I n I 2 (A, M)

where A 1 and A 2 are constants, with

0 n A 2  A 1 n 2H,

M 1 and M 2 are smooth functions of A, with

0 n M 1 (A) n M 2 (A) n H

and I 1 andI 2 are smooth functions of A and M, with

0 n I 1 (A , M) n I 2 (A, M).

If f is an R3  R function that is continuous on D , then from Theorem 14.6.1 we have the

Transformation Rule for Spherical Coordinates

555 5A 2 M52 (A) I 2 5(A,M) b c

f (x, y, z) dV = f S(I, A, M) I 2 sin M dI dM dA . (14.7)
D A 1 M 1 (A) I 1 (A,M)

Example 14.6.6

Use spherical coordinates to evaluate the integral of Example 14.6.3.

Solution.
Step 1. We put
x = I sin M cos A, y = I sin M sin A, z = I cos M.

The integrand then becomes

x 2 + y 2 + z 2 = I 2 sin2 M cos2 A + I 2 sin2 M sin2 A + I 2 cos2 M

b c
= I 2 sin2 M cos2 A + sin2 A + I 2 cos2 M
b c
= I 2 sin2 M + cos2 M
= I2.

Step 2. Since | det JS (I, A, M)| = I 2 sin M , the integral now becomes
===
I = I 4 sin M dI dM dA.
D
 61 MAT215–V/3
APM212–W/3

Step 3. As noted in Example 14.6.3, the region D can be obtained by rotating the region D ` in
Fig. 14.6(c) about the Z-axis through an angle of 2H
Geometrically it is clear that, in the region D ` , the value of M varies between 0 and H4 . (Remember
that M is measured downward from the positive Z-axis.) We can check the upper limit for M by
noting that, if
S
z = x 2 + y2,

then

I 2 cos2 M = I 2 sin2 M cos2 A + I 2 sin2 M sin2 A

i.e. I 2 cos2 M = I 2 sin2 M
i.e. tan2 M = 1
H
i.e. M = .
4
In order to ¿nd the limits for I we note that the region D ` can be shaded by drawing, for each M
between 0 and H4 , a ray from the origin up to the line z = 2.

φ=0
Z ρ = 2secφ φ = π
4

D*

Y
ρ=0
Fig. 14.9(a)

Thus the lower limit for I is 0, and the upper limit is given by

z = 2
i.e. I cosM = 2
i.e. I = 2 sec M

We can also see this geometrically, by noting that the length of a ray in D ` that makes an angle of
magnitude M with the positive Z-axis, has length 2 sec M.

2
cφ
2 se

Y
Fig. 14.9(b)
 62

Thus the region D can be described in terms of spherical coordinates by

H
D = {(I, A, M) | 0 n A n 2H, 0 n M n and 0 n I n 2 sec M}.
4
Step 4.
===
I = (x 2 + y 2 + z 2 )d x dy dz
D
H
=2H = 4 2=sec M
= I 4 sin M dI dM dA
0 0 0
H
=2H = 4 v w2 sec M
1 5
= I sin M dM dA
5 0
0 0
H
=2H = 4
32
= sin M sec5 M dM dA
5
0 0
H
=2H = 4
32
= sec4 M tan M dM dA
5
0 0
H
=2H = 4
32
= sec3 M (sec M tan M) dM dA
5
0 0
=2H v w H4
32 1
= sec4 M dA
5 4 0
0
=2H
8
= [4  1] dA
5
0
48H
= .
5
 63 MAT215–V/3
APM212–W/3

Exercises 14.6.7

1. In each of the following cases, sketch the region D and describe it in terms of cylindrical
coordinates.

(a) D is the region in R3 inside the cylinder x 2 + y 2 = 2, above the plane z = 1 and
below the plane z = 3.
(b) D is the region inside the sphere x 2 + y 2 + z 2 = 4.
(c) D is the region in R3 enclosed by the surfaces z = 3  x 2  y 2 and
z = x 2 + y 2  3.

2. In each of the following cases, sketch the region D and describe it in terms of spherical
coordinates.

(a) D is the region inside the sphere x 2 + y 2 + z 2 = 4.

S
(b) D is the region inside the sphere x 2 + y 2 + z 2 = 4 above the cone z = x 2 + y2.
S S
3. Let D be the region in R3 enclosed by the surfaces z = x 2 + y 2 and z = 9  x 2  y2.
555
Evaluate the integral x z dV, using
D

(a) cylindrical coordinates.

(b) spherical coordinates.
S
4. Let D be the region in R3 inside the cone z = x 2 + y 2 , below the hemisphere z =
S
4  x 2  y 2 and above the plane z = 1. Evaluate the integral
===
(x 2 + y 2 )d V,
D

using the coordinate system that you ¿nd most convenient.

5. Let D be the region in R3 inside the cylinder x 2 + y 2 = 1, above the cone

S S
z = x 2 + y 2 , below the cone z = 2 x 2 + y 2 . Evaluate the integral
=== r S s
1 + x 2 + y2 d V
D

using the coordinate system that you ¿nd most convenient.

6. Evaluate the integral === S

x 2 + y2 + z2 d V
D

where
j k
D = (x, y, z) | x 2 + y 2 + z 2 n 4 .
 64

14.7 CALCULATING VOLUMES BY USING TRIPLE INTEGRALS

The following examples illustrate how we can calculate the volume of a given 3-dimensional
region by using the formula ===
volume of D = 1 dV
D

which was given in Section 14.3.

Examples 14.7.1

1. Let D be the region in R3 above the plane z = 2 and below the paraboloid z = 3  x 2  y 2 .
Calculate the volume of D.
Solution.
The plane z = 2 and the paraboloid z = 3  x 2  y 2 intersect where

2 = 3  x 2  y2,
i.e. x 2 + y 2 = 1.

Thus the XY-projection of D is the region R in the XY-plane inside the circle x 2 + y 2 = 1.

Z
z = 3 - x 2- y 2
3

z=2
2 D

R 1 Y
1
X
Y Z
z = 3 - x 2- y 2
3
R D*
2
1 X
z=2

1 Y
Fig. 14.10
 65 MAT215–V/3
APM212–W/3

The whole region D can be obtained by rotating the region D ` about the Z-axis through an
angle of 2H.

It will be convenient to use cylindrical coordinates to describe D. (Why?)

We put
x = r cos A , y = r sin A and z = z.

The XY-projection R can be described in terms of polar coordinates as

{(r, A) | 0 n A n 2H and 0 n r n 1} .

The region D is bounded below by

z=2

and bounded above by

z = 3  x 2  y2.
i.e. z = 3  r 2.

Thus
j k
D = (r, A , z) | 0 n A n 2H, 0 n r n 1 and 2 n z n 3  r 2 .

Since det JC (r, A, z) = r, we have

=2H =1 3=r 2
volume of D = rdz dr dA
0 0 2
=2H =1
r [z ]32r dr dA
2
=
0 0
=2H =1
b c
= r  r 3 dr dA
0 0
=2H v w1
1 2 1 4
= r  r dA
2 4 0
0
=2H
1
= dA
4
0
H
= .
2
 66

2. Make use of a triple integral to derive a formula for the volume of a sphere.
Solution.
We consider a sphere in R3 with radius a , centred at the origin. Let D be the region enclosed
by the sphere.

Z Z

ρ=a

D a Y D*
Y
ρ=0

φ=π
Fig. 14.11

The whole region D can be obtained by rotating the region D ` about the Z-axis through an
angle of 2H and can thus be described in terms of spherical coordinates as

D = {(I, A, M) | 0 n A n 2H, 0 n M n H and 0 n I n a} .

We know that det JS (I, A, M) = I 2 sin M.

Thus
===
volume of D = I 2 sin MdI dM dA
D
=2H =H =a
= I 2 sin MdI dM dA
0 0 0
=2H =H
a3
= sin M dM dA
3
0 0
=2H
a3
= [ cos M ]H0 dA
3
0
=2H
a3
= 2 dA
3
0
4Ha 3
= .
3
4Ha 3
Thus the volume of a sphere with radius a equals .
3
 67 MAT215–V/3
APM212–W/3

Exercises 14.7.2

1. Calculate the volumes of the regions in Exercises 13.7.2 by using triple integrals instead of
double integrals.

2. In each of the following, calculate the volume of D , using an appropriate integral.

S
(a) D is the region in R3 inside the sphere x 2 + y 2 +z 2 = 4, below the cone z = x 2 + y2.
(b) D is the region in R3 enclosed by the cylinder x 2 + y 2 = 1 and the planes z = 0 and
z = 2 y + 3.
(c) D is the region inside the paraboloid z = x 2 + y 2 , below the plane z = 4.
(d) D is the region in R3 inside the cylinder x 2 + y 2 = 2, above the plane z = 1 and
below the plane z = 3.
(e) D is the region in R3 enclosed by the surfaces z = 3  x 2  y 2 and z = x 2 + y 2  3.
68
Chapter 15

Line Integrals of Real-valued Functions

15.1 INTRODUCTION

The idea of a single integral of an R  R function over an interval in R can be generalized to an

integral of an Rn  R function over a curve in Rn . An integral over a curve is called a line integral
(although a curve integral would actually be a better description). In this chapter we de¿ne two
types of line integrals, namely line integrals with respect to arc length and line integrals with
respect to the coordinate variables. For the ¿rst type the direction of the curve of integration is
not taken into account, but for the second type the direction of the curve is important.

We need the following terminology regarding curves.

If C is a curve in R2 with parametric equation

(x, y) = r(t); t + [a, b],

the direction of C is taken as the direction of the path described by (x, y) as t increases from a to
b. Thus in this case r(a) is the initial point of C and r(b) its endpoint.

If r(a) = r(b) then C is a closed curve.

If r(t1 ) /= r (t2 ) for all t1 , t2 + (a, b) with t1 /= t2 , then C is a simple curve. Thus, a curve is simple
if it does not cross itself. This terminology carries over in a natural manner to curves in Rn .

69
 70

r(b)
r(a) = r(b)

r(a)

(a) A simple curve (b) A closed curve

(c) A simple, closed curve

Fig. 15.1

We trust that you are familiar with the concept of a smooth curve and a piecewise smooth curve.
(If you are not, please revise Section 6.4).

15.2 LINE INTEGRALS IN THE PLANE, WITH RESPECT TO ARC LENGTH

The following steps lead to the de¿nition of a line integral with respect to arc length of an R2  R
function over a curve in the XY-plane.

1. Consider an R2  R function f and a smooth curve C which lies in D f .

2. Form a partition P of C by dividing C into n sub-curves, as shown in Fig. 15.2.

3. Choose a point (xk , yk ) on the kth sub-curve; k = 1, ..., n.

4. Let sk be the length of the kth sub-curve; k = 1, ..., n and let PPP be the length of a
longest sub-curve. Note that f (xk , yk )sk gives the area of a vertical rectangle with length
sk and height f (xk , yk ), as shown in Fig. 15.3.
 71 MAT215–V/3
APM212–W/3

.
∆s k
(xk ,yk )
C
Df

Fig. 15.2 Partition of the curve C

5. Form the sum

;
n
f (xk , yk )sk
k=1
and consider the limit of this sum as PPP tends towards 0.
Z

f(x k,y k ) z = f(x,y)

f(C)

yk

xk C
sk
Df

X
Fig. 15.3 The kth rectangle, with area f (xk , yk )sk
 72

De¿nition 15.2.1 Let f be an R2  R function de¿ned on a region in the plane that contains a
smooth curve C. The line integral of f over C with respect to arc length is
= ;n
f (x, y)ds = lim f (xk , yk )sk ,
PP P0
k=1
C
provided this limit exists.

Geometric interpretation

Suppose that f is an R2  R function that is de¿ned and continuous over a smooth curve C in the
5
XY-plane. If f (x, y) o 0 for all (x, y) + C, then the line integral f (x, y)ds gives the area of
C
the vertical surface below the graph of f that lies above C.

Z
z = f(x,y)

f(C)

C
Df

X
5
Fig. 15.4 f (x, y)ds = area of B
C

5
Similarly, if f (x, y) n 0 for all (x, y) + C , then  f (x, y)ds gives the area of the vertical
C
surface above the graph of f that lies below C.
 73 MAT215–V/3
APM212–W/3
5
If f (x, y) = 1 for all (x, y) + C, then f (x, y)ds gives the area of a vertical surface with height
C
1 and length equal to the length of C. Hence
5
length of C = 1ds. (15.1)
C

Y
C
X
5
Fig. 15.5 Length of C = 1ds
C
=
Evaluation of f (x, y)ds
C

Suppose C is a smooth curve with parametric equation

(x, y) = (x(t), y(t)); t + [a, b].

In Section 6.3 we saw that

V
t u2 t u2
ds dx dy
= + .
dt dt dt
Thus the relationship between the differentials ds and dt is given by
U r s2
b dx c2
ds = dt
+ dy
dt
dt. (15.2)

If f is an R2  R function that is continuous over C, then

U r s2
5 5b b dx c2
f (x, y) ds = f ((x(t), y(t))) dt
+ dy
dt
dt. (15.3)
C a

Remark 15.2.2

The integral on the right hand side of (15.3) is a single integral, since the integrand is a function
of t alone. Thus the formula (15.3) converts a line integral of an R2  R function over a curve in
R2 to a single integral of an R  R function over an interval in R. This enables us to evaluate a
line integral by applying the techniques for evaluating single integrals.
 74

If r is the R  R2 function de¿ned by

r (t) = (x(t), y(t)) ; t + [a, b]

then t u
) dx dy
r (t) = ,
dt dt
and Vt u2 t u2
o ) o dx dy
or (t)o = + .
dt dt
Thus an alternative form of (15.3) is:

5 5b o o
f (x, y) ds = f (r (t)) or ) (t)o dt. (15.4)
C a

Examples 15.2.3

1. Evaluate =
x y 3 ds
C
where C is the curve with parametric equation
H
(x, y) = (4 cos t, 4 sin t); t + [0, ].
2
Solution.
We have
dx dy
= 4 sin t and = 4 cos t.
dt dt
Thus
ds S
= 16 sin2 t + 16 cos2 t
dt T b c
= 16 sin2 t + cos2 t
T
= 16
= 4,

hence
ds = 4dt.
Now
H
= =2
x y 3 ds = (4 cos t) (4 sin t)3 (4dt)
C 0
H
=2
= 45 sin3 t cos t dt
0
d eH
= 44 sin4 t 02
= 256.
 75 MAT215–V/3
APM212–W/3

Alternative Method.
If we wish to use (15.4) instead of (15.3) we put
H
r(t) = (4 cos t, 4 sin t); t + [0, ] and f (x, y) = x y 3 .
2
Then
H
r ) (t) = (4 sin t, 4 cos t); t + [0, ];
2
hence
o ) o S
or (t)o = 16 sin2 t + 16 cos2 t
= 4.

Furthermore,
b c
f r(t) = f (4 cos t, 4 sin t)
= 44 cos t sin3 t.

By (15.4),
H
= =2
o o
f (x, y) ds = f (r(t)) or ) (t)o dt
C 0
H
=2
= 44 cos t sin3 t (4) dt
0
= 256,

as in the ¿rst method.

2. Let C be the curve in R2 with parametric equation

b c
(x, y) = et + 1, et  1 ; t + [1, 1].

Calculate the length of C.

Solution.
We have
dx dy
= et and = et .
dt dt
Hence
ds S
= e2t + e2t
dt T t
= 2e .

Thus
T t
ds = 2e dt.
 76

Now
=
length of C = 1 ds
C

T = t
1

= 2 e dt
1
T b c
= 2 e  e1 .

If C is de¿ned by an explicit equation of the form

y = g(x); x + [a, b],

then we take x as parameter and obtain the parametric equation

(x, y) = (x, g(x)) ; x + [a, b].

Since
dx dy
= 1 and = g ) (x) ,
dx dx
it follows from (15.2) that T
ds = 1 + (g ) (x))2 d x,

so the formula (15.3) becomes

5 5b T
f (x, y) ds = f (x, g(x)) 1 + (g ) (x))2 dx. (15.5)
C a

Similarly, if C is de¿ned by an explicit equation of the form

x = h (y) , y + [c, d]
5
then we take y as parameter and evaluate the integral f (x, y) ds by means of the formula
C

5 5d T
f (x, y) ds = f (h (y) , y) (h ) (y))2 + 1 dy. (15.6)
C c

Example 15.2.4

Evaluate =
x y ds
C

where C is the curve in R with equation

2

y = 1 + 2x; x + [0, 1].

 77 MAT215–V/3
APM212–W/3

Solution.
We take x as parameter. Then

dx dy
= 1 and = 2.
dx dx
Hence
ds T T
= 1+4= 5
dx
and hence
T
ds = 5 d x.
Thus
= =1 T
x yds = x (1 + 2x) 5 dx
C 0

T = b
1
c
= 5 x + 2x 2 dx
0
T
7 5
= .
6

Remark 15.2.5

As shown in Examples 3.2.13, there are many ways of parametrizing a given curve, and different
5
parametrizations can impose different directions on a curve. However, f (x, y)ds has the same
C
value for all parametrizations of C. The direction of C plays no role, since sk represents a length
only.

Exercises 15.2.6
5
1. Evaluate f (x, y)ds in each of the following cases.
C

(a) f (x, y) = 2x y and C is the curve with parametric equation

H
(x, y) = (5 cos t, 5 sin t) ; t + [0, ].
4
(b) f (x, y) = 3x 2 + 6y 2 and C is the line y = 2x + 1; x + [1, 0].
5
2. Evaluate the line integral x yds, where C is the line with parametric equation
C

(x, y) = (0, 1) + t (1, 2); t + [0, 1].

Compare your answer with the answer of Example 15.2.4. Comment on these answers.

3. Determine the length of that part of the parabola y = x 2 where 0 n x n 1.

 78

15.3 LINE INTEGRALS IN THE PLANE, WITH RESPECT TO THE COORDINATE

VARIABLES

In order to de¿ne a line integral with respect to x, we consider a directed curve C. We then use the
same ¿ve steps that were used for De¿nition 15.2.1, but we replace the arc length sk with xk ,
the difference between the x-coordinates of the endpoint and the initial point of the kth sub-curve.
Similarly, we de¿ne a line integral with respect to y by replacing sk with yk , the difference
between the y-coordinate of the endpoint and the initial point of the kth sub-curve.

∆s k
∆yk
∆x k C
Df

Fig. 15.6 Partition of the directed curve C

De¿nition 15.3.1 Let f be an R2  R function de¿ned on a region in the XY-plane that contains
a smooth, directed curve C.

(i) The line integral of f over C with respect to x is given by

= ;
n
f (x, y)dx = lim f (xk , yk )xk ,
PP P0
k=1
C

provided this limit exists.

(ii) The line integral of f over C with respect to y is given by

= ;
n
f (x, y)dy = lim f (xk , yk )yk ,
PP P0
k=1
C

provided this limit exists.

 79 MAT215–V/3
APM212–W/3
5 5
Evaluation of f (x, y)d x and f (x, y)dy
C C

Let C be a smooth, directed curve with parametrization

(x, y) = (x(t), y(t)); t + [a, b].

Then

dx dy
= x ) (t) and = y ) (t) ;
dt dt
hence
d x = x ) (t) dt and dy = y ) (t) dt.

If f is an R2 R function that is continuous over C, then the line integral of f over C with respect
to x can be converted to a single integral by the formula

5 5b
f (x, y)d x = f (x (t) , y (t)) x ) (t) dt. (15.7)
C a

Similarly, the line integral of f over C with respect to y can be converted to a single integral by
the formula
5 5b
f (x, y)dy = f (x (t) , y (t)) y ) (t) dt. (15.8)
C a

Example 15.3.2

Evaluate the line integrals = =

x y dx and x y dy,
C C
where C is the directed line segment in the sketch below.

Y
3

2
C
1

0 1 X
Fig. 15.7

Solution.
C has initial point (0, 1) and endpoint (1, 3). Thus a vector in the same direction as C is given by

(1, 3)  (0, 1) = (1, 2).

 80

Therefore, by De¿nition 2.11.1, a parametric equation for C is

(x, y) = (0, 1) + t (1, 2); t + [0, 1].

Note that if t = 0, then (x, y) = (0, 1) and if t = 1 then (x, y) = (1, 3), so this parametrization
directs the curve from (0, 1) to (1, 3), as desired.
Now, since
x = t and y = 1 + 2t,

we have
d x = dt and dy = 2dt.

Thus
= =1
xy dx = (t)(1 + 2t)dt
C 0
=1
b c
= t + 2t 2 dt
0
7
=
6
and

= =1
x y dy = (t)(1 + 2t)(2dt)
C 0
=1
b c
= 2 t + 2t 2 dt
0
7
= .
3

5
If C is a simple, smooth, directed curve, then the line integral f (x, y)d x has the same value
C 5
for all parametrizations of C that agree with the direction of C. The same is true for f (x, y)dy.
C
However, if  C is the curve consisting of the same points as C but with the opposite direction,
then
5 5
f (x, y)d x =  f (x, y)d x, (15.9)
C C

and
5 5
f (x, y)dy =  f (x, y)dy. (15.10)
C C
5
As mentioned in Remark 15.2.5, the line integral f (x, y)ds is independent of the direction of
C
the curve, i.e.
5 5
f (x, y)ds = f (x, y)ds. (15.11)
C C
 81 MAT215–V/3
APM212–W/3

This is because the arc length sk is always positive, whereas xk and yk change sign when we
reverse the direction of C.

Example 15.3.3

Let C be the directed line segment in Fig. 15.7. Evaluate the following line integrals:
5
1. xy dx
C
5
2. x y ds.
C

It is clear that  C is the directed line segment with initial point (1, 3) and endpoint (0, 1).

Y
3

2
-C
1

0 1 X
Fig. 15.8

Solution.

1. A parametric equation for  C is

(x, y) = (1, 3)  t (1, 2); t + [0, 1].

Thus
dx = 1dt

and hence
= =1
x y dx = (1  t)(3  2t)(1dt)
C 0
=1
=  (1  t)(3  2t)dt
0
=1
b c
=  3  5t + 2t 2 dt
0
7
=  .
6
 82

Comparing this with our answer to 15.3.2, we note that

= =
x y d x =  x y dx,
C C

as expected.

2. Since
dx = 1dt and dy = 2dt,

we have T T
ds = (1)2 + (2)2 dt = 5 dt.

Thus
= =1 rT s
x yds = (1  t)(3  2t) 5 dt
C 0

T = b
1
c
= 5 3  5t + 2t 2 dt
0
T
7 5
= .
6
Comparing this answer with that of Example 15.2.4, we note that
= =
x y ds = x y ds,
C C

as expected.

Exercises 15.3.4
5 5
1. Evaluate the line integrals f (x, y)d x and f (x, y)dy in each of the following cases.
C C

(a) f (x, y) = 2x y and C is the curve with parametric equation

H
(x, y) = (5 cos t, 5 sin t) ; t + [0, ].
2
(b) f (x, y) = 2x y and C is the curve with parametric equation
H
(x, y) = (5 sin t, 5 cos t) ; t + [0, ].
2

(c) f (x, y) = 3x 2 + 6y 2 and C is the curve y = 2x + 1; x + [1, 0].

5 5
2. Evaluate (x 2 + y 2 )dx + 2x ydy if C is the curve from (0, 0) to (1, 1) with equation
C C

(a) y = x; x + [0, 1].

 83 MAT215–V/3
APM212–W/3

(b) y = x 2 ; x + [0, 1].

(c) y = x 3 ; x + [0, 1].
5T 5
3. Evaluate y dx + (x  y) dy if C is the curve from (0, 0) to (1, 1) with equation
C C

(a) (x, y) = (t, t) t + [0, 1].

b c
(b) (x, y) = t 2 , t t + [0, 1].
b 2c
(c) (x, y) = t, t t + [0, 1].
b 3c
(d) (x, y) = t, t t + [0, 1].

15.4 LINE INTEGRALS OF Rn  R FUNCTIONS

If f is an Rn  R function that is de¿ned and continuous on a smooth curve C in Rn , we de¿ne

the line integral =
f (x1 , ..., xn )ds
C
in the same way as the line integral in De¿nition 15.2.1.
If C is a , smooth, directed curve, we also de¿ne the n line integrals
=
f (x1 , ..., xn )d xi ; i = 1, ..., n
C
in the same way as the line integrals in De¿nition 15.3.1.
In order to calculate such line integrals, we ¿nd a parametric equation

(x1 , ..., xn ) = (x1 (t), ..., xn (t)); t + [a, b]

for C and then use the formulae

5 5b Tb c2 b c2
f (x1 , ..., xn )ds = f (x1 (t) , ..., xn (t)) x1) (t) + · · · + xn) (t) dt (15.12)
C a

and
5 5b
f (x1 , ..., xn )d xi = f (x1 (t) , ..., xn (t)) xi) (t) dt; i = 1, . . . , n. (15.13)
C a

If we put
r(t) = (x1 (t), ..., xn (t)); t + [a, b]

then we can also write (15.12) as

5 5b b co o
f (x1 , ..., xn )ds = f r (t) or ) (t)o dt. (15.14)
C a
 84

Example 15.4.1

Evaluate the line integral

=
2 +y 2 z
ex ds,
C

where C is the circle in Fig. 15.9

-1 1
1 Y
X
Fig. 15.9

Solution.
C is the image of the function

r (t) = (cos t, sin t, 1); t + [0, 2H].

Now
r ) (t) = ( sin t, cos t, 0).

Hence
o ) o S 2
or (t)o = sin t + cos2 t + 0 = 1.

Therefore, by (15.14)

= =2H
x 2 +y 2 z 2 t+sin2 t1
e ds = ecos (1)dt
C 0
=2H
= 1 dt
0
= 2H.
 85 MAT215–V/3
APM212–W/3

Exercises 15.4.2

1. Evaluate the line integral

=
(x y + z) d x,
C

where C is the line segment from (1, 0, 1) to (3, 1, 2) .

2. Evaluate
= = =
yd x + xdy + zdz,
C C C

where C is curve in R3 with parametric equation

(x, y, z) = (2 cos t, 2 sin t, t); t + [0, 2H].

15.5 LINE INTEGRALS OVER PIECEWISE SMOOTH CURVES

In De¿nitions 15.2.1 and 15.3.1 we required that C be a smooth curve. A line integral over a
piecewise smooth curve, consisting of a ¿nite number of smooth curves, is de¿ned as the sum of
the line integrals over the smooth pieces of C.

6
k
De¿nition 15.5.1 Let C be the piecewise smooth curve de¿ned by C = Ci , where Ci is a smooth
i=1
curve in Rn , for i = 1, . . . , k. If f is an Rn  R function that is de¿ned on C and is continuous on
each of the curves C1 , . . . , Ck , then
 
= ;
k =
% &
f (x1 , ..., xn )ds = # f (x1 , ..., xn )ds $
i=1
C Ci

and
 
= ;
k =
% &
f (x1 , ..., xn )dx j = # f (x1 , ..., xn )dx j $ , for j = 1, . . . , n.
i=1
C Ci
 86

Example 15.5.2

Evaluate the line integral =

(3x + 2y + z)d x,
C
where C = C1 C C2 in Fig. 15.10.

(2,0,2) (0,0,1)
C2
C1

(0,1,0) Y

X
Fig. 15.10

Solution.
Both C1 and C2 are piecewise smooth curves; hence C is a smooth curve.
C1 is a line segment with initial point (2, 0, 2) and endpoint (0, 1, 0) . A parametric equation for
C1 is

(x, y, z) = (2, 0, 2) + t (2, 1, 2); t + [0, 1].

Now
d x = 2dt;
hence
=
(3x + 2y + z)dx
C1
=1
= (3 (2  2t) + 2t + (2  2t)) (2dt)
0
=1
= (12t  16) dt
0
= 10.

C2 is a quarter circle in the plane x = 0, with initial point (0, 1, 0) and endpoint (0, 0, 1). A
parametric equation for C2 is
 87 MAT215–V/3
APM212–W/3

H
(x, y, z) = (0, cos t, sin t); t + [0, ].
2
Now

d x = 0 dt, dy =  sin t dt, dz = cos t dt.

Thus

=
(3x + 2y + z)d x
C2
H
=2
= (2 cos t + sin t) (0) dt
0
H
=2
= 0 dt
0
= 0.

By De¿nition 15.5.1,
= = =
(3x + 2y + z)d x = (3x + 2y + z)d x + (3x + 2y + z)d x
C C1 C2
= 10 + 0 = 10.

Exercises 15.5.3
5 5
1. Evaluate the line integrals (3x + 2y + z)dy and (3x + 2y + z)dz, where C is the curve
C C
given in Example 15.5.2.
5 5 5
2. Evaluate y dx + x dy + z dz for each of the directed curves C in the sketches below.
C C C

Z Z

(6,8,5) (6,8,5)

(0,0,0) (0,0,0)
Y Y

(6,8,0) (6,8,0)
X X
(a) (b)
 88

3. Evaluate = =
y dx + x dy
C C

where C is the closed, directed curve in the sketch below.

y = x2

0 2 X
Chapter 16

Line Integrals of Vector Fields

16.1 INTRODUCTION

Line integrals of vector ¿elds have important physical applications. They are used, for example,
to calculate the work done by a variable force along a curve.

We shall show that a line integral of a vector ¿eld can be expressed as a sum of line integrals
of real-valued functions. A line integral of a vector ¿eld can thus be evaluated by means of the
techniques presented in Chapter 15.

In this chapter we also present the Fundamental Theorem of Line Integrals, which is an analogue
of the Fundamental Theorem of Calculus.

16.2 DEFINITION OF A LINE INTEGRAL OF A VECTOR FIELD

5
The line integral of a vector ¿eld F over a directed curve C with respect to r is denoted by F ·dr.
C
The following steps lead to the de¿nition.

1. Consider an n-dimensional vector ¿eld F and a smooth, directed curve C that lies in the
domain of F.

2. Choose m + 1 points r1 , . . . , rm+1 on C, which divide the curve into m sub-curves, taking
r1 as the initial point and rm+1 as the endpoint of C.

89
 90

rm+1

∆r k
rk rk+1

r1

X
Fig. 6.1

3. Let rk be the kth displacement vector, de¿ned by

rk = rk+1  rk , for k = 1, . . . , m.

Let PPP be the length of a longest displacement vector.

b c
4. Form the dot products F rk · rk .

5. Take the sum

;
m
b c
F rk · rk
k=1

and consider the limit of this sum as PPP tends towards 0.

De¿nition 16.2.1 If F is an n-dimensional vector ¿eld and C is a smooth curve in the domain of
F, then the line integral of F over C with respect to r is

5 3
m b c
F · dr = lim F rk · rk ,
C PP P0 k=1

provided this limit exists.

 91 MAT215–V/3
APM212–W/3
5
16.3 PHYSICAL INTERPRETATION OF F · dr
C

If a constant force F acts on an object moving along a straight line from a point a to a point
b, the work done by the force on the object is de¿ned to be the product of the distance moved
and the magnitude of the component of F in the direction of movement. If d = b  a, then the
o o
distance moved equals od o . If A is the angle between the vectors F and d, then the magnitude of
o o
the component of F in the direction of movement is given by o F o cos A.
Hence
o oo o
work done = o F o od o cos A.

F
θ

d
o oo o
Fig. 16.2 Work done = o F o od o cos A = F · d

But, from the formula in Section 2.9, we have

o oo o
o F o od o cos A = F · d .

Hence, if F is a constant force acting on an object moving along a straight line through a dis-
placement d, then the work done by the force on the object is given by the formula

Work done = F · d . (16.1)

In practical situations, forces are not always constant and particles do not always move in straight
5
lines. The line integral F · dr can be used to calculate the work done by a variable force F on
C
an object moving along a curve C. Let us explain this brieÀy:

Let F be a force ¿eld acting on an object moving along a smooth, directed curve C in R2 and
suppose F is continuous on C. Now we divide C as shown in Fig. 16.1. If the sub-curves are very
small, then each piece is very nearly a straight line and the force ¿eld F is very nearly constant on
b c
each piece. The value of F on the kth sub-curve is approximately equal to F rk and hence, by
(16.1), the work done by F on the particle as it moves from rk to rk+1 is approximately equal to

b c
F rk · rk .
 92

F(r k)
rk
rk rk+1

X
b c
Fig. 16.3 Work done r F rk · rk

The total work done by F along the curve C is approximately equal to the sum
3m b c 3
m b c
F rk · rk and exactly equal to lim F rk · rk .
k=1 PP P0 k=1

Thus the work done by a continuous force ¿eld F along a smooth, directed curve C is given by the
formula
5
Work done = F · dr. (16.2)
C

5
16.4 EVALUATING F · dr
C

Let C be a smooth, directed curve in R2 and let F = (M, N ) be a 2-dimensional vector ¿eld that
is continuous on C. Divide C into small sub-curves as shown in Fig. 16.1 and let

rk = (xk , yk ) for k = 1, . . . , m.

Now let

xk = xk+1  xk and yk = yk+1  yk , for k = 1, . . . , m.

Then

rk = (xk , yk ) for k = 1, . . . , m.

Thus

b c
F rk · rk = (M (xk , yk ) , N (xk , yk )) · (xk , yk )
= M (xk , yk ) xk + N (xk , yk ) yk .
 93 MAT215–V/3
APM212–W/3

Then
= ;
m
b c
F · dr = lim F rk · rk
PP P0
k=1
C
;
m
= lim (M (xk , yk ) xk + N (xk , yk ) yk )
PP P0
k=1
;m ;
m
= lim M (xk , yk ) xk + lim N (xk , yk ) yk
PP P0 PP P0
k=1 k=1
= =
= M(x, y)d x + N (x, y)dy.
C C
5
Note that M and N are real-valued functions, and hence the line integrals M(x, y)d x and
5 C
N (x, y)dy may be evaluated by using the techniques presented in Chapter 15. It is customary to
C 5
write the sum of these two integrals as one integral without brackets as M(x, y)dx + N (x, y)dy
5 C
and also to abbreviate the notation further to Mdx + N dy. Thus we write the formula derived
C
above as
5 5
F · dr = Mdx + N dy. (16.3)
C C

If the curve C has parametric equation

(x, y) = r (t) ,

then, by (15.7) and (15.8),

= =
b b c b c c
M(x, y)dx + N (x, y)dy = M r (t) x ) (t) + N r (t) y ) (t) dt
C
=C
b b c b cc b c
= M r (t) , N r (t) · x ) (t) , y ) (t) dt

=C
b c
= F r (t) · r ) (t) dt.
C
5
Thus we can write F · dr as a line integral with respect to the parameter t, as
C

5 5 b c
F · dr = F r (t) · r ) (t) dt. (16.4)
C C

If F = ( f 1 , . . . , f n ) is an n-dimensional vector ¿eld that is continuous on a smooth, directed curve

in Rn , then (16.4) also holds, and, similar to (16.3), we have

5 5
F · dr = f 1 d x1 + · · · + f n d xn . (16.5)
C C
 94

If C is a piecewise smooth curve de¿ned by

>
m
C= Ci ,
i=1

5
where each Ci is a smooth curve, we de¿ne the line integral F · dr by
C

5 m 5
3
F · dr = F · dr. (16.6)
C i =1 Ci

Example 16.4.1

Evaluate the line integral

=
F · dr,
C

where
F(x, y, z) = (y, z, x)

and C is the curve with parametric equation

(x, y, z) = (t, t 2 , t 3 ); t + [0, 1] .

Solution.
By (16.5),
= =
F · dr = y dx + z dy + x dz.
C C

On the curve C we have

x = t, y = t 2, z = t 3.

Thus
dx = dt, dy = 2t dt and dz = 3t 2 dt.

Thus

= =1
F · dr = t 2 dt + t 3 (2t dt) + t (3t 2 dt)
C 0
=1
b2 c
= t + 2t 4 + 3t 3 dt
0
89
= .
60
 95 MAT215–V/3
APM212–W/3

Exercises 16.4.2
5
1. Evaluate F · dr in each of the following cases.
C

(a) F(x, y) = (y 3 , x 2 y) and C is the curve in R2 with parametric equation

b c
(x, y) = e2t , et ; t + [0, ln 2].

(b) F(x, y, z) = (e x , xe x y , x ye x yz ) and C is the image of the R  R3 function r de¿ned

by
b c
r (t) = t, t 2 , t 3 ; t + [0, 1].

2. Let C be the curve in R3 with parametric equation

(x, y, z) = (t, t 2 , t); t + [0, *).

Calculate the work done by the force ¿eld

F(x, y, z) = (x y, y, yz)

on an object moving along the curve C from the point (0, 0, 0) to the point (1, 1, 1).

3. Calculate the work done by the force ¿eld

b c
F(x, y) = x 2 , x y

on an object that moves once around the circle x 2 + y 2 = 1 in an anticlockwise direction.

4. Find the value of =

T
y d x + (x  y)dy
C
over each of the following paths C from (0, 0) to (1, 1).

(a) (x, y) = (t, t).

(b) (x, y) = (t 2 , t).
(c) (x, y) = (t, t 2 ).
Sketch the different paths.

5. Find the value of =

(x 2 + y 2 )dx + 2x y dy.
C
over each of the following paths C from (0, 0) to (1, 1):

(a) y = x.
(b) y = x 2 .
(c) x = y 2 .
5
6. Evaluate F · dr, where F(x, y) = (x 2 + y 2 , 2x y) and C is the path along the parabola
C
y = x 2 from (0, 0) to (1, 1).
 96

16.5 THE FUNDAMENTAL THEOREM OF LINE INTEGRALS

As mentioned in Section 12.4, The Fundamental Theorem of Calculus (Theorem 12.4.1) tells us
that if f is a smooth R  R function, the single integral of the derivative f ) over an interval [a, b]
gives the total increase in f between a and b. The following theorem tells us that if a and b are
points in Rn and f is a smooth Rn  R function, the line integral of grad f over any piecewise
smooth curve from a to b gives us the total increase in f between a and b.

Theorem 16.5.1 The Fundamental Theorem of Line Integrals

Suppose f is a smooth Rn  R function and C is a piecewise smooth, directed curve in the interior
of D F , with initial point a and endpoint b. Then
=
b c
grad f · dr = f (b)  f a .
C

Proof.
First, suppose C is a smooth curve with parametric equation

(x1 , · · · , xn ) = r (t) ; t + [a, b].

Then
r (a) = a and r (b) = b.

We de¿ne an R  R function g by

g(t) = f i r (t) ; t + [a, b].

Then, by the Chain Rule (Theorem 7.6.1),

b c
g ) (t) = grad f r (t) · r ) (t) .

Now, by (16.4),

= =b
b c
grad f · dr = grad f r (t) · r ) (t) dt
C a
=b
= g ) (t)dt
a
= g (b)  g (a) ,

by the Fundamental Theorem of Calculus (Theorem 12.4.1).

But by the de¿nition of g,

g (b)  g (a) = f i r (b)  f i r (a)

b c b c
= f r (b)  f r (a)
= f (b)  f (a).
 97 MAT215–V/3
APM212–W/3

This proves that =

grad f · dr = f (b)  f (a).
C

If C is not smooth, we simply consider each smooth section of C separately, and use De¿nition
15.5.1.

We remind you that if F is an n-dimensional vector ¿eld and f is an Rn  R function such that
grad f = F, then f is called a potential function of F. Theorem 16.5.1 may also be formulated as
follows.

Theorem 16.5.2 Alternative form of the Fundamental Theorem of Line Integrals

Suppose F is a continuous n-dimensional vector ¿eld and C is a piecewise smooth, directed curve
in the interior of D F , with initial point a and endpoint b. Then
=
b c
F · dr = f (b)  f a .
C

Remark 16.5.3
5
Theorem 16.5.2 tells us that if we can ¿nd a potential function of F, then we can compute F ·dr
C
by calculating the value of the potential function only at the initial point and the endpoint of C.

16.6 PATH-INDEPENDENT VECTOR FIELDS

If a and b are two points in Rn , there are many different paths from a to b. (By a path from a to b
we mean a curve with initial point a and endpoint b.)

C1
a

C2

C3

b
Fig. 16.4 Different paths from a to b
 98

5 5
In general, if C1 and C2 are two different paths from a to b, the line integrals F · dr and F · dr
C1 C2
may have different values. However, Theorem 16.5.2 tells us that if F has a potential function
5
and C is a piecewise smooth curve with initial point a and endpoint b, then F · dr depends only
C
on a and b, and not on the path taken to get from a to b.

In many of the natural vector ¿elds studied by physicists, such as gravitational ¿elds and elec-
tric ¿elds, the amount of work done in moving a mass or a charge from one point to another
depends only on the displacement, and not on the path taken. Such vector ¿elds are called path-
5
independent. In view of the physical interpretation of the line integral F · dr, we now de¿ne the
C
concept of path-independence as follows.

De¿nition 16.6.1 A continuous vector ¿eld F is said to be path-independent if, for any two points
5
a and b in the domain of F, the line integral F · dr has the same value for all piecewise smooth
C
curves C from a to b lying in the domain of F.

Theorem 16.5.2 tells us that if a continuous vector ¿eld F has a potential function, then F is
path-independent. It can be proved that the converse of this is also true. Thus we have:

Theorem 16.6.2 A continuous vector ¿eld F is path-independent if and only if F has a potential
function.

Combining Theorem 8.5.6 and 16.6.2, we obtain the following result:

Theorem 16.6.3 If F is a smooth vector ¿eld whose domain is a simply connected region, then F
is path-independent if and only if F is conservative.

The terminology conservative vector ¿eld comes from physics, where it refers to energy potential
¿elds in which the principle of conservation of energy holds.
 99 MAT215–V/3
APM212–W/3

16.7 LINE INTEGRALS OVER CLOSED CURVES

If C is a closed, directed curve, we write the line integral of F over C as

,
F · dr .
C

The following result follows from our de¿nition of path-independence.

Theorem 16.7.1 A continuous vector ¿eld F is path-independent if and only if

,
F · dr = 0
C

for every piecewise smooth, closed, directed curve C in the domain of F.

Proof.
First, suppose F is path-independent and C is a closed, piecewise smooth, directed curve Let a
and b be any two points on C and let C1 be that part of C that starts at a, and ends at b, and C2 the
part that starts at b and ends at a .

C1 C2

b
6
Fig. 16.5 C = C 1 C2

Then, since C1 and  C2 both start at a and end at b, it follows from De¿nition 16.6.1 that

= =
F · dr = F · dr
C1  C2
=
=  F · dr .
C2
 100

Since C = C1 C C2 , it follows from (16.6) that

, = =
F · dr = F · dr + F · dr
C C1 C2
= =
=  F · dr + F · dr
C2 C2
= 0.
,
To prove the converse, suppose F · dr = 0 for every piecewise smooth, closed, directed curve
C
C in D F . Now let a and b be any two points in D F and let D1 and D2 be two paths in D F , both
going from a to b . Put
D = (D1 ) C D2 .

D1 D2

b
Fig. 16.6 D = (D1 ) C D2

Then D is a piecewise smooth, closed curve and hence, by our assumption,

,
F · dr = 0.
D

But
, = =
F · dr = F · dr + F · dr
D D1 D2
= =
=  F · dr + F · dr
D1 D2

hence = =
 F · dr + F · dr = 0,
D1 D2
which proves that = =
F · dr = F · dr .
D1 D2
 101 MAT215–V/3
APM212–W/3

Exercises 16.7.2

1. Let F be a smooth, 2-dimensional or 3-dimensional vector ¿eld whose domain is a simply

connected region and let C be a piecewise smooth, closed, directed curve in the domain of
F. By using theorems and de¿nitions in Chapter 16 and in Section 8.5, prove that
,
F · dr = 0 if and only if curl F = 0 .
C

2. Evaluate the line integral of Exercise 15.5.3(3) by using

(a) the result of 1.

(b) the Fundamental Theorem of Line Integrals.

3. Explain why the answers of 4(a) -(c) of Exercise 16.4.2 are not all the same, while those of
5(a) - (e) are all the same.

4. Let
b c
F (x, y) = x 2 + y 2 , 2x y
5
and let C be the path from (0, 0) to (1, 1) along the parabola y = x 2 . Evaluate F · dr by
C
each of the following methods:

(a) Use the formula (16.3).

(b) Use the formula (16.4).
(c) Find a potential function for F and then use the Fundamental Theorem for Line Inte-
grals.
102
Chapter 17

Surface Integrals

17.1 INTRODUCTION

In the same way that we can generalize a single integral over an interval on the real line to a line
integral over a curve in the plane, we can generalize a double integral over an area in the plane to
a surface integral over a surface in space.

Just as double integrals are useful for calculating areas of Àat surfaces, surface integrals are useful
for calculating areas of curved surfaces.

17.2 DEFINITION OF A SURFACE INTEGRAL

A surface S in R3 that has an equation of the form

z = g(x, y); (x, y) + R,

where R is a region in the XY-plane, is called a smooth surface if g is a smooth function.

The following steps lead to the de¿nition of a surface integral over a smooth surface in R3 .

1. Consider an R3  R function f de¿ned on a region in R3 . Let S be a smooth surface in the

domain of f , with equation

z = g(x, y); (x, y) + R,

where R is a region in the XY-plane.

2. Partition R into n rectangles, as was done for the de¿nition of a double integral. This
determines a corresponding partition of S into n sections, S1 , ..., Sn , as shown in Fig. 17.1.

3. Choose a point (xk , yk , z k ) in Sk ; k = 1, ..., n.

103
 104

(xk ,yk ,zk )

Sk
S2
S1
S

R2 Y
R Rk
1
R
X
Fig. 17.1 A partition of S, corresponding to a partition of R

4. Let Sk be the area of Sk and let PPP be the length of the longest of the diagonals of the
rectangles in R. Let Dk be the solid whose base is Sk and whose perpendicular height at the
point (xk , yk , z k ) is f (xk , yk , z k ). If PPP is very small, then the surface Sk is almost Àat and
the volume of Dk is approximately equal to f (xk , yk , z k )Sk .

f(x k,yk ,zk )

Dk
Sk

Fig. 17.2 Volume of Dk r f (xk , yk , z k )Sk

5. Take the sum

;
n
f (xk , yk , z k )Sk
k=1

and consider the limit of this sum as PPP tends towards zero.
 105 MAT215–V/3
APM212–W/3

De¿nition 17.2.1 Let f be an R3  R function that is de¿ned on a smooth surface S in R3 . The

surface integral of f over S is
== ;
n
f (x, y, z)d S = lim f (xk , yk , z k )Sk ,
PP P0
k=1
S

provided this limit exists.

A surface that is a union of smooth surfaces is called a piecewise smooth surface.

A surface integral over a piecewise smooth surface is de¿ned as the sum of surface integrals over
the smooth sections.

De¿nition 17.2.2 Suppose

>
n
S= Si ,
i=1

where each Si is a smooth surface and Si and S j have only boundary points in common, for i /= j.
If f is an R3  R function that is continuous on each Si , then we de¿ne the surface integral of f
over the piecewise smooth surface S by
== n ==
;
f (x, y, z)d S = f (x, y, z)d S.
i=1
S Si

17.3 GEOMETRIC INTERPRETATION

Just as a double integral can be used to calculate the volume of a solid with a Àat base, a surface
integral can be used to calculate the volume of a solid with a curved base.
55
If f (x, y, z) o 0 for all (x, y, z) + S, then the integral f (x, y, z)d S may be interpreted as the
S
volume of a solid whose base is the surface S and whose perpendicular height at the point (x, y, z)
equals f (x, y, z) .
55
Just as the area of a region R in the XY-plane is given by 1 d A, the area of a surface in 3-space
R
may be computed by the formula
==
area of S = 1 d S. (17.1)
S
 106

17.4 EVALUATION OF SURFACE INTEGRALS

Suppose S is a surface in R3 with equation

z = g(x, y); (x, y) + R,

where R is a region in the XY-plane. Then the differential d S of the surface S is given by
V
t u2 t u2 t u2
"z "z "z
dS = + + dA
"x "y "z

i.e.
U r s2
b "z c2 "z
dS = "x
+ "y
+ 1dA (17.2)

(Compare this with the expression (15.2) for the arc length differential ds.)
55
This relationship allows us to convert the surface integral f (x, y, z)d S to an ordinary double
S
integral:

U
55 55 b c2 r s2
f (x, y, z)d S = f (x, y, g (x, y)) ""zx + ""zy + 1 d A. (17.3)
S R

In the above formula R is the XY-projection of S (see Fig 17.1).

If the equation of the surface S is given in the form y = g(x, z), we let R be the XZ-projection
of S, and we use the formula
Ur s r s2
55 55 2
f (x, y, z)d S = f (x, g (x, z) , z) "" xy + 1 + ""zy d A. (17.4)
S R

If the equation of S is given by x = g(y, z), we let R be the YZ-projection of S, and we use the
formula
U r s2 b c
55 55 2
f (x, y, z)d S = f (g(y, z), y, z) 1 + "" xy + ""zx d A. (17.5)
S R

Examples 17.4.1

1. Evaluate == S
x 2 + y 2 + z 2 d S,
S
S
where S is that part of the cone z = x 2 + y 2 where 1 n z n 2.

Solution.

We ¿rst sketch the given region S and its XY-projection R.

 107 MAT215–V/3
APM212–W/3

S
1

-2 -1 1 2 Y
1
R
2
X
Fig. 17.3
On the surface S,
S
z= x 2 + y2.
Thus
"z x "z y "z
=S , =S , = 1.
"x x 2 + y2 "y x 2 + y2 "z
Hence
V
x2 y2
dS = + + 1dA
x 2 + y2 x 2 + y2
V
x 2 + y2
= + 1dA
x 2 + y2
T
= 2 d A.

Thus
== S == S T
x 2 + y2 + z2 d S = x 2 + y 2 + (x 2 + y 2 ) 2dA
S
=R= S T
= 2(x 2 + y 2 ) 2 d A
R
== S
= 2 (x 2 + y 2 ) d A.
R

We have thus converted the given surface integral to an ordinary double integral. The
region of integration of this double integral can be described in terms of polar coordinates
as
R = {(r, A) | 0 n A n 2H and 1 n r n 2} .
In polar coordinates x 2 + y 2 = r 2 . Hence, by the Transformation Rule for Polar coordinates
(13.6),
 108

== S =2H =2 T
2 (x + y ) d A = 2
2 2 r 2 r dr dA
R 0 1
=2H =2
= 2 r 2 dr dA
0 1
=2H
2 d 3 e2
= r 1 dA
3
0
=2H
14
= dA
3
0
28H
=
3

Thus == S
28H
x 2 + y2 + z2 d S = .
3
S

2. Calculate the area of that part of the parabola z = x 2 + y 2 that lies below the plane z = 4.

Solution.

We ¿rst sketch the given surface S and its XY-projection R.

-2 R 2 Y
2

X
Fig. 17.4
 109 MAT215–V/3
APM212–W/3

S lies on the paraboloid

z = x 2 + y2.

Thus we have
"z "z "z
= 2x, = 2y = 1.
"x "y "z
Thus
S
dS = 4x 2 + 4y 2 + 1 d A.

Now

==
Area of S = dS
=S= S
= 4x 2 + 4y 2 + 1 d A
R
=2H =2 S
= 4r 2 + 1 rdr dA
0 0
=2H v w
1 b 2 c 32 2
= 4r + 1 dA
12 0
0
Hr 3 s
= 17 2  1 .
6

Exercises 17.4.2
55
1. Evaluate the surface integral f (x, y, z)d S in each of the following cases:
S

(a) f (x, y, z) = x yz and S is that part of the plane x + y + z = 1 that lies in the ¿rst
octant (i.e. where x o 0, y o 0 and z o 0).
(b) f (x, y, z) = x + y and S is that part of the plane z = 2x + 3y that lies inside the
cylinder x 2 + y 2 = 9.
(c) f (x, y, z) = z and S is that part of the paraboloid z = x 2 + y 2 that lies inside the
cylinder x 2 + y 2 = 2.

2. Calculate the area of the surface S in R3 de¿ned by

j k
S = (x, y, z) | z = x 2 + y 2 and 2 n z n 6 .

3. Calculate the area of the ellipse that is formed when the plane z = cx cuts the cylinder
x 2 + y 2 = 1.
110
Chapter 18

Flux Integrals

18.1 INTRODUCTION

The velocity ¿eld of a Àuid may be represented by a vector ¿eld. If the Àuid Àows through a
surface S, the volume of Àuid that passes through S per unit time is called the Àux of the Àuid
through S. A Àux integral is a special type of surface integral that is used to compute Àux.

X
Fig. 18.1 A vector ¿eld representing the Àow of a Àuid through a surface S

Before computing the Àux through a surface, we need to decide which direction of Àow we will
regard as the positive direction. This is described as choosing an orientation. In some cases we
shall consider the upward direction as positive, and in other cases the downward direction. In the
case of a closed surface we shall choose either the inward or the outward direction as the positive
direction. Choosing an orientation only makes sense when the surface is orientable, so let us ¿rst
explain the notion of an orientable surface.

111
 112

18.2 ORIENTABLE SURFACES

If S is a smooth surface in R3 and (x, y, z) + S, then any vector that is perpendicular to the
tangent plane to S at (x, y, z) is called a normal of S at (x, y, z). A normal with a length of one
unit is called a unit normal. At every interior point of a smooth surface S in R3 there are two unit
normals (with opposite directions).

S
(x,y,z)

X
Fig. 18.2 The two unit normals of S at (x, y, z)

An orientable surface is a surface that can be oriented by its normals. Some surfaces are not
orientable. Consider, for example, the Möbius band, which is obtained by taking a strip of pa-
per ABCD, giving the end CD a single twist and then joining the two ends together so that A
corresponds with D and B with C.

A C AD
B D BC

finish

start

Fig. 18.3 The Möbius band is not orientable

 113 MAT215–V/3
APM212–W/3

The Möbius band is not orientable because, if we choose a normal at a given point P on the band
and move it once round the band, it returns to P pointing in the opposite direction. To avoid the
dif¿culties caused by this phenomenon, we shall de¿ne Àux integrals for orientable surfaces only.

If S is an orientable surface in R3 then it has two possible orientations. Once we have chosen one
of the two orientations, we refer to S as an oriented surface. The unit normal of S at (x, y, z) that
points in the direction determined by the orientation of S is called the orientation vector of S at
(x, y, z) and is denoted by n (x, y, z) . If S is a smooth, orientable surface, the orientation vectors
determine a continuous vector ¿eld n, called the orientation of S.

18.3 FLUX

Let F be a continuous vector ¿eld that represents the velocity ¿eld of a Àuid Àowing through a
smooth, oriented surface S.
We ¿rst consider the case where S is Àat and the Àow F is constant. In this case the orientation n
of S is also constant, so at every point (x, y, z) on S we can represent the velocity of the Àuid by
the vector F and the orientation by the vector n. Let A be the magnitude of the angle between F
and n.

F
n
θ
S

Fig. 18.4

Then the Àux through S (i.e. the volume of the Àuid passing through S per unit time) equals the
volume of the skew solid in Fig. 18.5, which is given by

volume = perpendicular height × area of base

bo o c
= o F o cos A (area of S)
bo o o o c o o
= o F o on o cos A (area of S) since on o = 1
b c
= F · n (area of S) ,

by the formula in Section 2.9. Thus if F is constant and S is Àat, the Àux of the Àuid through S is
given by the formula
b c
Àux through S = F · n (area of S) . (18.1)
 114

F cosθ F
θ
S
b c
Fig. 18.5 Flux through S = F · n (area of S)

Now we consider the case where S may be curved and the Àow may be variable.
We form a partition as was done for the de¿nition of a surface integral (see Fig. 17.1).
If PPP is very small, then the surface Sk is almost Àat and the Àow through Sk is almost constant
so, using (18.1), we have

b c
Àux through Sk r F(xk , yk , z k ) · n(xk , yk , z k ) Sk .

Hence

;
n
b c
Àux through S r F(xk , yk , z k ) · n(xk , yk , z k ) Sk .
k=1

Thus

;
n
b c
Àux through S = lim F(xk , yk , z k ) · n(xk , yk , z k ) Sk
PP P0
k=1
==
b c
= F · n (x, y, z)d S,
S

where F · n is the continuous R3  R function de¿ned by

b c
F · n (x, y, z) = F(x, y, z) · n(x, y, z) for every (x, y, z) + S.

== ==
b c
It is customary to abbreviate the notation F · n (x, y, z)d S to F · n d S.
S S

Thus if S is a smooth, oriented surface with orientation n, we can calculate the Àux through S by
the formula

==
Flux through S = F · n d S. (18.2)
S
 115 MAT215–V/3
APM212–W/3

If S is a piecewise smooth oriented surface, then the Àux of a Àuid through S can be determined
by calculating the Àux through each smooth section of S and then adding these. Thus if
>
n
S= Si ,
i=1

where each Si is a smooth oriented surface, it is natural to de¿ne the integral

==
F · n d S by
S
55 n 55
3
F · n dS = F · ni d S (18.3)
S i=1 Si

where n i is the orientation of Si .

==
An integral of the form F · n d S, is called a Àux integral.
S
Flux integrals are also used to calculate electrical Àux and magnetic Àux.

18.4 EVALUATION OF FLUX INTEGRALS

==
Evaluation of the Àux integral F · n d S involves the following three steps.
S

1. Determine the orientation n of the oriented surface S.

2. Determine the R3  R function F · n.

3. The Àux integral now becomes an ordinary surface integral. Evaluate this surface integral
by converting it to an ordinary double integral, as described in Section 17.4.

Determining n

There are several methods for determining the orientation, n, of a given oriented surface S. One
method is to ¿rst de¿ne an R3  R function that has S as one of its contours, and then to use the
fact that gradient vectors are perpendicular to contours. Let us describe this method:
Suppose S is a smooth, oriented surface, with equation

z = g(x, y); (x, y) + R,

where R is a region in the XY-plane.

We de¿ne an R3  R function G by

G(x, y, z) = z  g(x, y); (x, y, z) + S.

Now S is the contour C(0) of G, because the equation G(x, y, z) = 0 is equivalent to the equation
z = g(x, y) (study De¿nition 3.2.9). Therefore, by Theorem 7.8.3, the vector grad G(x, y, z) is
perpendicular to S at the point (x, y, z) and hence the unit normal n(x, y, z) has either the same
or the opposite direction as grad G(x, y, z).
 116

Thus

grad G(x, y, z) grad G(x, y, z)

n(x, y, z) = o o or o o,
o o o o
ograd G(x, y, z)o ograd G(x, y, z)o

depending on the orientation of S.

Remark 18.4.1

If S has an upward orientation, then n(x, y, z) must point upward, so its z-coordinate must be
positive.
If S has a downward orientation, then n(x, y, z) must point downward, so its z-coordinate must
be negative.

Example 18.4.2
55
Evaluate the Àux integral F · n d S, where
S

F(x, y, z) = (0, z, z)

and S is the surface de¿ned by

S = {(x, y, z) | z = 6  3x  2y, x o 0, y o 0, z o 0} ,

oriented upward.
Solution.
Let us sketch the situation.

2 3 Y

X
Fig. 18.6

In order to determine n we put

G(x, y, z) = z  (6  3x  2y) .
 117 MAT215–V/3
APM212–W/3

Then
grad G(x, y, z) = (3, 2, 1).

Thus at each point (x, y, z) on S the upward unit normal is

(3, 2, 1)
n(x, y, z) = T .
14
Thus
==
F · n dS
S
==
(0, z, z) · (3, 2, 1)
= T dS
14
S
==
3
= T z d S.
14
S

In order to determine this surface integral we need to ¿nd the relationship between d S and d A.
Since the equation of S is

z = 6  3x  2y,

we have
"z "z "z
= 3, = 2 = 1.
"x "y "z
Thus
T
dS = 14 d A.

The XY-projection of S is the triangle

| }
6  3x
R = (x, y)|0 n x n 2 and 0 n y n .
2
Thus
== ==
3
F · n dS = T z dS
14
S
=S= T
3
= T (6  3x  2y) 14 d A
14
R
63x
=2 = 2

= 3 (6  3x  2y) dy d x
0 0
= 18.
 118

Exercises 18.4.3

1. Consider the vector ¿eld F de¿ned by

F (x, y, z) = (y, x, 1).

55
Evaluate the Àux integral F · n d S in each of the following cases.
S

(a) S is that part of the sphere x 2 + y 2 +z 2 = 1 that lies in the ¿rst octant, oriented upward.
(b) S is the sphere x 2 + y 2 + z 2 = 1, oriented outward.

2. Suppose the velocity ¿eld F of a Àuid is given by

F(x, y, z) = (2x, 2y, 3).

Calculate the upward Àux of the Àuid through the surface

j k
S = (x, y, z) | z = x 2 + y 2 and z n 4 .
Chapter 19

The Theorems of Green, Gauss and

Stokes

This section deals with three theorems that give relationships between various types of integrals.
These theorems have important practical applications and they are also important from the view-
point of pure mathematics, since each of them can be regarded as a generalization of the Funda-
mental Theorem of Calculus.

In this chapter we shall state the three theorems and discuss their applications. Readers who are
interested in their proofs are urged to consult one of the recommended books, listed in Tutorial
Letter 101.

19.1 GREEN’S THEOREM

Green’s Theorem relates a double integral over a bounded region in the plane to a line integral
over the curve that bounds the surface.

R
C

X
Fig. 19.1 A region R bounded by a curve C

119
 120

Theorem 19.1.1 Green’s Theorem

Let C be a piecewise smooth, simple, closed curve in the XY-plane, directed anti-clockwise, and let
R be the region bounded by C. If F = (M, N ) is a smooth 2-dimensional vector ¿eld, then

== t u ,
"N "M
 dA = Md x + N dy. (19.1)
"x "y
R C

Remark 19.1.2

This theorem was proved by George Green (1793 - 1841), an English mathematician and physicist.

r s
"N "M
If F is a 2-dimensional vector ¿eld then, by De¿nition 8.4.4, curl F = "x
 "y
k, where
k = (0, 0, 1). Since k · k = 1, we can write

t u t u
"N "M "N "M
 =  k·k
"x "y "x "y
b c
= curl F · k.

We can also write

, ,
Md x + N dy = F · dr
C C

(see (16.3)). Thus the formula in Green’s Theorem can also be written as

== ,
b c
curl F · k d A = F · dr . (19.2)
R C

,
The evaluation of a line integral of the form F · dr can sometimes be simpli¿ed by converting
C
it to a double integral, by applying Green’s Theorem.
 121 MAT215–V/3
APM212–W/3

Example 19.1.3
,
Evaluate F · dr, where
C b c
F (x, y) = 2xe y , 2x 2 e y

and C is the curve in the sketch below.

y=x

y = x2 - 2

Fig. 19.2(a)

Solution.
C is a piecewise smooth curve consisting of two smooth pieces, so evaluating the given line integral
by using the methods given in Chapter 16 would be rather cumbersome, since we would have to
evaluate the integral over the two smooth pieces separately and then add the results. However,
the region R enclosed by C is a Type 1 region, so if we apply Green’s Theorem, we only need to
evaluate a single double integral.

y=x

-1

2 X
R

y = x2 - 2

Fig. 19.2(b)
 122

Let
M(x, y) = 2xe y and N (x, y) = 2x 2 e y .
Since M and N are products of polynomial and exponential functions, all their partial derivatives
are continuous and hence F is a smooth vector ¿eld. Furthermore, C is a piecewise smooth, simple,
closed curve enclosing the region R shown in the sketch.
Therefore, by Green’s Theorem,
= ==
"N "M
F · dr = (  )d A
"x "y
C R
==
= (4xe y  2xe y )d A
=R=
= 2xe y d A
R
= 2 =x
= 2xe y dy d x
1 x 2 2

=2
2 2
= 2x(e x  e x )d x
1
=2 =2
2 2
= 2xe d x 
x
2xe x dx
1 1
=2 K 2 L2
d e2
= 2xe x 1  2 e x dx  e x 2 (integration by parts)
1
1

= 4e2 + 2e1  2e2 + 2e1  e2 + e1

= e2 + 5e1 .

Exercises 19.1.4

1. Evaluate each of the following integrals, ¿rst by using (15.7) and (15.8) to convert the
line integral to a single integral, and then by applying Green’s Theorem to convert the line
integral to a double integral.
,
(a) (x  y)d x + x y dy where C is the triangle with vertices (0, 0), (1, 0) and (1, 3),
C
directed anticlockwise.
,
(b)  y 2 d x + x 2 dy with C the circle with parametric equation
C

(x, y) = (3 cos t, 3 sin t) , t + [0, 2H],

directed anticlockwise.
 123 MAT215–V/3
APM212–W/3
, r s
b c 3
(c) x 5 + 3y dx + 2x  e y dy, where C is the circle (x  1)2 + (y  5)2 = 4,
C
directed anticlockwise.

2. Consider the force ¿eld F de¿ned by

b c
F (x, y) = x y 2  1, x 2 y + 1 .

Apply Green’s Theorem to determine the work done by this force ¿eld in moving an object
once around the curve C given in Fig. 19.2, in the anticlockwise direction.

3. Let F be the vector ¿eld de¿ned by

F (x, y) = (x 2 , x y).

and let R be the region in the XY-plane de¿ned by

j k
R = (x, y) | x 2 + y 2 n 1 .

Apply Green’s Theorem to evaluate the double integral

==
b c
curl F · k d A.
R

19.2 STOKES’ THEOREM

The boundary of a surface S in space is the curve that runs round the edge of S (like the hem of a
table cloth). If S is a surface bounded by a simple, closed, directed curve C , then the direction of
C imposes an orientation on S according to the Right Hand Rule:
If the ¿ngers of the right hand curl around the surface S in the direction of the curve C, then
the thumb points in the direction of the orientation of S.
If S is oriented upward, the orientation of its boundary is anticlockwise when viewed from above.

Z
n
S

Y
X
Fig. 19.3(a) An upward oriented surface S and its boundary curve C
 124

If S is oriented downward, the orientation of its boundary is anticlockwise when viewed from
below.

Z
C

S
n

Y
X
Fig. 19.3(b) A downward oriented surface S and its boundary curve C

Stokes’ Theorem relates a Àux integral over a surface in space to a line integral over the bound-
ary of the surface.

Theorem 19.2.1 Stokes’ Theorem

Let S be a piecewise smooth, oriented surface, whose oriented boundary C is a piecewise smooth,
simple closed, curve . If F = (M, N, P) is a smooth, 3-dimensional vector ¿eld whose domain is
a region in R3 that contains S, then
== ,
b c
curl F · n d S = F · dr.
S C

Remarks 19.2.2

1. This theorem is named after the Irish mathematician and physicist, Sir George Stokes (1819
- 1903). The theorem was actually discovered by the Scottish physicist Sir William Thom-
son (1824 - 1907), also known as Lord Kelvin. He mentioned this theorem to Stokes in a
letter, and Stokes then asked students to prove it in an examination at Cambridge University
in 1854. It is not known whether any of the students were able to prove it.

2. Stokes’ Theorem may be regarded as the 3-dimensional analogue of Green’s Theorem (com-
pare it with (19.1)). For this reason some text books refer to Green’s Theorem as Green’s
Theorem in the plane and to Stokes’ Theorem as Green’s Theorem in 3-space.
 125 MAT215–V/3
APM212–W/3
,
If C is a closed curve in R3 , the evaluation of a line integral of the form F · dr can sometimes
C
be simpli¿ed by applying Stokes’ Theorem.

Example 19.2.3

5
Evaluate F · dr, where
C
F(x, y, z) = (3z, 5x, 2y)

and C is the curve formed by the intersection of the plane z = y + 3 and the cylinder x 2 + y 2 = 1,
directed as in Fig. 19.4.

Solution.
Let S be that part of the surface z = y + 3 that is enclosed by C and let R be the projection of S
onto the XY-plane.

n
S
3

R 1 Y

X
Fig. 19.4

By Stokes’ theorem

= ==
b c
F · dr = curl F · n d S.
C S

Now
 126

n n
n i j k n
n n
n n
n n
n " " " nn
n
curl F = n n
n "x "y "z n
n n
n n
n n
n 3z 5x 2y n
= (2, 3, 5) .

The surface S lies on the plane with equation

z = y+3

and, by the Right Hand Rule, the curve C imposes an upward orientation on S. In order to deter-
mine n, we put

G(x, y, z) = z  y  3.
Then
grad G(x, y, z) = (0, 1, 1).
Thus (0, 1, 1) is a normal of the surface S. Since its z-coordinate is positive, it points upward
and hence has the same direction as the orientation vector. Hence

(0, 1, 1)
n(x, y, z) = T for every (x, y, z) + S.
2
Thus
= ==
(0, 1, 1)
F · dr = (2, 3, 5) · T dS
2
C S
==
2
= T dS
2
S
T = =
= 2 d S.
S

We have now converted the given line integral to a surface integral. In order to evaluate this
surface integral, we need to convert it to a double integral, as explained in Section 17.4, so we
need to describe the XY-projection R of S. Since R is bounded by the circle x 2 + y 2 = 1, it can
be described in terms of polar coordinates by

R = {(r, A ) | 0 n A n 2H and 0 n r n 1} .

We also need to ¿nd the relationship between d S and d A. Since S lies on the plane

z = y + 3,
we have
"z "z "z
= 0, =1 = 1.
"x "y "z
 127 MAT215–V/3
APM212–W/3

Hence
T
dS = 2 d A.

Thus
= T ==
F · dr = 2 dS
C S
T == T
= 2 2dA
R
=2H=1
= 2 r dr dA
0 0
= 2H.

Exercises 19.2.4

1. Evaluate each of the following line integrals, ¿rst by using the methods of Chapter 16 to
convert it to a single integral, then by using Stokes’ Theorem to convert it to a surface
integral.

,
(a) ydx  5xdy + 3dz, where C is the closed, directed curve in the sketch below.
C

2 3 Y

X
Fig. 19.5

,
(b) F ·dr where F = (5y, 5x, 3z) and C is the curve of intersection of the plane z = 1
C
and the cylinder x 2 + y 2 = 4, whose direction is clockwise when viewed from above.
 128

19.3 GAUSS’ THEOREM

Gauss’ Theorem relates a triple integral over a bounded region in R3 to a Àux integral over the
outer surface of the region.

Z
S n

Y
X
Fig. 19.6 A solid D bounded by a surface S, oriented outward

Theorem 19.3.1 Gauss’ Theorem Let D be a closed, bounded region in R3 with a piecewise
smooth outer surface S, oriented outward. If F = (M, N, P) is a smooth 3-dimensional vector
¿eld de¿ned on S, then
=== ==
div F d V = F · n d S.
S
D

Remark 19.3.2

Theorem 19.3.1 is also called the Divergence Theorem. It is named after the great German math-
ematician Karl Friedrich Gauss (1777 - 1855), who proved this theorem during his investigation
of electrostatics.
 129 MAT215–V/3
APM212–W/3
==
If S is a a closed surface, the evaluation of F · n d S can often be simpli¿ed by applying Gauss’
S
Theorem.

Example 19.3.3

==
Evaluate F · n d S where
S

F(x, y, z) = (x y, y 2 z, z 3 )

and S is the outer surface of the cube

D = {(x, y, z)|0 n x n 1, 0 n y n 1, 0 n z n 1}

oriented outward.
Solution.

1 Y
1
X
Fig. 19.7

S is a piecewise smooth surface, consisting of 6 smooth sections (the 6 sides of the cube). Thus
==
evaluating F · n d S as a Àux integral would involve calculating the sum of six different surface
S
integrals. However, if we apply Gauss’ Theorem, we only need to evaluate one triple integral,
since D is a single Type I region.
 130

Now
t u
" " "
div F = , , · (x y, y 2 z, z 3 )
" x " y "z
= y + 2yz + 3z 2 .

By Gauss’ Theorem

== ===
F · n dS = div F d V
S D
= 1 =1 = 1
= (y + 2yz + 3z 2 )dz dy d x
0 0 0
= 1 =1
= (2y + 1) dy dx
0 0
=1
= 2dx
0
= 2.

Exercises 19.3.4

1. Determine the outward Àux of a Àuid with velocity ¿eld F through a surface S in each of
the following cases, ¿rst using a method from Chapter 17, then using Gauss’ Theorem.

j k
(a) F(x, y, z) = (x 3 , y 3 , z 3 ) and S = (x, y, z) | x 2 + y 2 + z 2 = 1 .

(b) F(x, y, z) = (x, y, z  1) and

Q S R>j k
S = (x, y, z) | z = 4  x 2  y 2 (x, y, z) | x 2 + y 2 n 4 and z = 0 .

(c) F (x, y, z) = (yz, x z, x y) and S is as in Example 19.3.3.

2. Do Exercise 18.4.3 1(b), using Gauss’ Theorem.

 131 MAT215–V/3
APM212–W/3

19.4 GENERALIZATIONS OF THE FUNDAMENTAL THEOREM OF CALCULUS

We have seen that the Fundamental Theorem of Line Integrals is a generalization of Fundamental
Theorem of Calculus. In a similar manner the Theorems of Green, Stokes and Gauss are gener-
alizations of the Fundamental Theorem of Calculus. In order to explain this, let us compare the
formulas of these theorems.

5b
1. Fundamental Theorem of Calculus f ) (x)d x = f (b)  f (a)
5
a
2. Fundamental Theorem of Line Integrals grad f · dr = f (b)  f (a)
C
55 b c +
3. Green’s Theorem curl F · k d A = F · dr
55R b c +C
4. Stokes’ Theorem curl F · n d S = F · dr
555
S
55 C
5. Gauss’ Theorem div F dV = F · n d S
D S

In each of these formulas the integrand on the left contains a derivative. In the Fundamental
Theorem of Calculus we have the ordinary derivative, in the theorems of Green and Stokes we
have the curl and in Gauss’ Theorem the divergence.
Furthermore, in each of these theorems the set of points involved in the right hand side of the
formula is the boundary of the region of integration on the left.

a b C
a
(1) (2) (3)
C
S
S
S
D
C
(4a) (4b) (5)
Fig. 19.8 Sets and their boundaries
 132

Answers of exercises

12.6.3

1. (a) 1 (b) 94 H (Put y = 3 sin A)

T
(c) 6  4 2
9
(d) 12 (e) 0
(f) 2 ln 2  1 (Use integration by parts.)
51 T 51 b 2 c
2. 1  x2 dx  x  1 dx = 12 H + 4
3
1 1
T
52 52 b c 52b 2 c 50
3. x dx  x  2 dx 
2
x  2 dx + x dx = 9
2
T
0 2 1 1

5r T
4. 2 r 2  x 2 d x = Hr 2
r

13.5.12

1. (a) 48 (b) 16
3
(c) H2 (d) H (e)4e + 11

52 x+2
5
2. (x + y) dyd x = 189
20
1 x 2
T
51 5 y
3. (a) yd xdy = 1
15
0 y
T
54 54y
(b) 1d xdy = 32
3
T
0  4y

51 51 b c
(c) x 2 + 2x y d xdy = 11
30
T
0 y
T
51 51y
2

(d) xdxdy = 1
12
T
0 1y

5 52
1 51 52 52 52
(e) (x + y) dyd x + (x + y) dyd x + (x + y) dyd x = 14
3
2 x 1 1 1 x

13.6.7
j H H
k
1. (a) (r, A ) | 0 n r n 1 and  2
nA n 2

(b) {(r, A) | 0 n r n 1 and H n A n 2H}

Q T R
(c) (r, A) | 0 n r n 2 and H2 n A n 3H4
Q T R
(d) (r, A) | 0 n r n 2 and  H4 n A n H4
j k
2. (a) (r, A ) | sin1 A n r n sin2 A and H4 n A n 3H
4
j k
(b) (r, A ) | 0 n r n 1 and  H4 n A n H4
 133 MAT215–V/3
APM212–W/3
H
52 52
3. (a) r 4 dr dA = 16
5
H
0 0
3H
52 53 b c
(b) sin r 2 r dr dA = 14 H (1  cos 9)
H 0

52H 51 b c
er r dr dA = H 1  e1
2
(c)
0 0
H
52 51
(d) r 4 cos A sin A dr dA = 1
10
0 0

13.7.2
H
53
ln 52
1. (a) e x cos y dyd x = 4
0  H2

51 4x
5 2
(b) (x + 4) dydx = 625
12
4 3x
H
52 51 H
(c) r 3 dr dA = 4
 H2 0

52H 52
2. r 3 dr dA = 8H
0 0

52H 52 b c
3. 8  2r 2 r dr dA = 16H
0 0

52H 5a
4. (a) Volume of cylinder = hr dr dA = Ha 2 h
0 0
52H 5a T
(b) Volume of sphere = 2 r a 2  r 2 dr dA = 43 Ha 3
0 0

13.8.2
5 2 5 x2
1. (a) 0 0 1 dy d x = 8
3
H
52 52
(a) r dr dA = 2H
 H2 0
H
53 52
(b) r dr dA = 43 H
 H3 0
T
51 51x
2

(c) 1 dy dx = 12 H  4
3
1 1x 2

51 5x 2
(d) 1 dy dx = 8
3
1 1

50 5x 51 52x
(e) 1 dy d x + 1 dy dx = 1
1 2x 0 x
 134

14.5.4

1. Consider a Type 2 region G in the XY-plane, given by

G = {(x, y) | a n y n b and g1 (y) n x n g2 (y)} .

Suppose h 1 and h 2 are smooth functions of x and y such that

h 1 (x, y) n h 2 (x, y) for all (x, y) + G.

Let D = {(x, y, z) | a n y n b, g1 (y) n x n g2 (y) , h 1 (x, y) n z n h 2 (x, y)}.

If f is an R3  R function that is continuous over D, then

=== =b g=2 (y) h 2=(x,y)

f (x, y, z) dV = f (x, y, z) dz dx dy.
D a g1 (y) h 1 (x,y)

2. (a) 0 (b) 83

53 3x
5 3xy
5
3. (x + y) dz dy d x = 27
4
0 0 0

14.6.8
Q T R
1. (a) D = (r, A, z) | 0 n A n 2H, 0 n r n 2,  1 n z n 3
T T
(b) D = {(r, A, z) | 0 n A n 2H, 0 n r n 2,  4  r 2 n z n 4  r 2 }
Q T R
(c) D = (r, A, z) | 0 n A n 2H, 0 n r n 3, r 2  3 n z n 3  r 2

2. (a) D = {(I, A, M) | 0 n A n 2H, 0 n M n H, 0 n I n 2}

Q T R
(b) D = (I, A, M) | 0 n A n 2H, 0 n M n H4 , 0 n I n 2
T
3
T
52H 52 59r
2

3. (a) r 2 z cos A dz dr dA = 0
0 0 r
H
52H 54 53
(b) I 4 cos A sin3 M cos M dI dM dA = 0
0 0 0

H
52H 54 52 r T s
4. I 4 sin3 M dI dM dA = H 253
30
 16 2
3
0 0 1
cos M

52H 51 52r b c
5. r + r 2 dz dr dA = 76 H
0 0 r

52H 5H 52
6. I 3 sin M dI dM dA = 16H
0 0 0
 135 MAT215–V/3
APM212–W/3

14.7.2
H
53
ln 52 ex 5
cos y
1.(1) (a) dz dy dx = 4
0  H2 0

51 4x
5 2 x+4
5
(b) dz dy dx = 625
12
4 3x 0
H
5 51 5
2 r2
H
(c) r dz dr dA = 4
 H2 0 0

52H 52 5r 2
(2) r dz dr dA = 8H
0 0 0

52H 52 8r
52
(3) r dz dr dA = 16H
0 0 r2

52H 5a 5h 52H 5H 5a
(4a) r dz dr dA = Ha 2 h (4b) I 2 sin M dI dM dA = 43 Ha 3
0 0 0 0 0 0

52H 5H 52 r T s
2. (a) I 2 sin M dI dM dA = 8H
3
2+ 2
0 H 0
4

52H 51 5 A+3
2r sin
(b) r dz dr dA = 3H
0 0 0
52H 52 54
(c) r dz dr dA = 8H
0 0 r2
T
52H 5 2 54
(d) r dz dr dA
0 0 1
T
5 5 3 3r
2H 52
(e) r dz dr dA = 9H
0 0 r 2 3

15.2.6

125
T
1. (a) 2
(b) 3 5
T
2. 7 5
6
.C
is the same curve as in Example 15.2.4. A line integral w.r.t. arc length is independent
of the parametrization of the curve.
51T T rT s
3. 0 4x 2 + 1d x = 12 5  14 ln 52

15.3.4

250 250
1. (a) 3
; 3 (b) 250
3
;  250
3
(c) 3; 6
4 4 4
2. (a) 3
(b) 3
(c) 3

2 19 2 13
3. (a) 3
(b) 30
(c) 3
(d) 20
 136

15.4.2

16
1. 3
2. 2H 2

15.5.3

16H H19
1. 4
; 2

2. (a) 121
2
; (b) 121
2

3. 0

16.4.2

1. (a)  19
8
(b) 13(e1)
6

1
2. 2
3. 0

2 19 2
4. (a) 3
(b) 30
(c) 3

4 4 4
5. (a) 3
(b) 3
(c) 3

4
6. 3

16.7.2

1. Hint: Combine De¿nition 8.5.2, Theorem 16.6.3 and Theorem 16.7.1.

,
2. (a) If F (x, y) = (y, x), then curl F = 0. Hence, by (1), F · dr = 0.
C
(b) Let f be a potential function for F and let p be any point on C. Then
,
F · dr = f ( p)  f ( p) = 0.
C
cbT
3. If F (x, y) = y, x  y , then curl F = / 0, so F is not path-independent.
b 2 c
If F (x, y) = x + y 2 , 2x y , then curl F = 0, so F is path-independent.

5 51 b c 51 b c
4. (a) F · dr = x 2 + y 2 dx + 2x ydy = 5x 4 + x 2 dx = 43 .
C 0 0

5 51 b c 51 b c
(b) F · dr = F t, t 2 · (1, 2t) dt = t 2 + t 4 , 2t 3 · (1, 2t) dt = 43 .
C 0 0

(c) A potential function is f (x, y) = 1 3

3
x + x y2.
5
F · dr = f (1, 1)  f (0, 0) = 43 .
C
 137 MAT215–V/3
APM212–W/3

17.4.2
T
3 149H
1. (a) 120
(b) 0 (c) 30
T
2. 49H
3
3. H c2 + 1

18.4.3

H
1. (a) 4
(b) 0

2. 20H

19.1.4

1. (a) 3 (b) 0 (c) 4H (Using methods from Chapter 16 is dif¿cult here - using
Green’s Theorem is much easier.)

2. 0

3. Let C be the circle x 2 + y 2 = 1, directed anticlockwise. Then

== ,
(curl F) · k d A = x 2 dx + x ydy = 0.
C
R

19.2.4

1. (a) 18 (b) 40H

19.3.4

1. (a) 12H
5
(b) 16H (c) 0, since div F = 0.
 138

Index of De¿nitions
anti-derivative, 4 Type 1 Region, 18
Type 2 Region, 21
closed curve, 69 triple integral, 48
coordinate transformations, 27, 52
cylindrical coordinates, 52 unit normal, 112
double integral, 11
work done, 91, 92
Àux, 114
Àux integral, 115

integration by substitution, 5
iterated integral, 15

line integral w.r.t. arc length, 72

line integral w.r.t. x and y, 78
line integral of a multivariable function, 83
line integral of a vector ¿eld, 90
Möbius band, 112

orientation, 113
orientation vector, 113
oriented surface, 113
orientable surface, 112

path independent vector ¿eld, 98

polar coordinates, 30
repeated integral, 15
reversing the order of integration, 23
rules of integration, 5, 13

simple curve, 69
single integral, 2
spherical coordinates, 57
surface integral, 105