MAT 280: Multivariable Calculus

James V. Lambers
September 15, 2014

Contents
1 Partial Derivatives
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1.1 Partial Differentiation . . . . . . . . . . . . . . . . . .
1.1.2 Multiple Integration . . . . . . . . . . . . . . . . . . .
1.1.3 Vector Calculus . . . . . . . . . . . . . . . . . . . . . .
1.2 Functions of Several Variables . . . . . . . . . . . . . . . . . .
1.2.1 Terminology and Notation . . . . . . . . . . . . . . . .
1.2.2 Visualization Techniques . . . . . . . . . . . . . . . . .
1.3 Limits and Continuity . . . . . . . . . . . . . . . . . . . . . .
1.3.1 Terminology and Notation . . . . . . . . . . . . . . . .
1.3.2 Defining Limits Using Neighborhoods . . . . . . . . .
1.3.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.4 Techniques for Establishing Limits and Continuity . .
1.4 Partial Derivatives . . . . . . . . . . . . . . . . . . . . . . . .
1.4.1 Terminology and Notation . . . . . . . . . . . . . . . .
1.4.2 Clairauts Theorem . . . . . . . . . . . . . . . . . . . .
1.4.3 Techniques . . . . . . . . . . . . . . . . . . . . . . . .
1.5 Tangent Planes, Linear Approximations and Differentiability
1.5.1 Tangent Planes and Linear Approximations . . . . . .
1.5.2 Functions of More than Two Variables . . . . . . . . .
1.5.3 The Gradient Vector . . . . . . . . . . . . . . . . . . .
1.5.4 The Jacobian Matrix . . . . . . . . . . . . . . . . . . .
1.5.5 Differentiability . . . . . . . . . . . . . . . . . . . . . .
1.6 The Chain Rule . . . . . . . . . . . . . . . . . . . . . . . . . .
1.6.1 The Implicit Function Theorem . . . . . . . . . . . . .
1.7 Directional Derivatives and the Gradient Vector . . . . . . . .
1.7.1 The Gradient Vector . . . . . . . . . . . . . . . . . . .
1.7.2 Directional Derivatives . . . . . . . . . . . . . . . . . .
1.7.3 Tangent Planes to Level Surfaces . . . . . . . . . . . .
3

5
5
5
7
7
9
10
12
14
14
16
17
20
22
22
26
27
31
31
33
34
36
38
40
43
47
48
49
51

CONTENTS
1.7.4 Tangent Lines to Level Curves . . . . . . . . .
1.8 Maximum and Minimum Values . . . . . . . . . . . .
1.8.1 Absolute Extrema . . . . . . . . . . . . . . . .
1.8.2 Relation to Taylor Series . . . . . . . . . . . . .
1.8.3 Principal Minors . . . . . . . . . . . . . . . . .
1.9 Constrained Optimization . . . . . . . . . . . . . . . .
1.10 Appendix: Linear Algebra Concepts . . . . . . . . . .
1.10.1 Matrix Multiplication . . . . . . . . . . . . . .
1.10.2 Eigenvalues . . . . . . . . . . . . . . . . . . . .
1.10.3 The Transpose, Inner Product and Null Space

2 Multiple Integrals
2.1 Double Integrals over Rectangles . . . . . . . .
2.2 Double Integrals over More General Regions . .
2.2.1 Changing the Order of Integration . . .
2.2.2 The Mean Value Theorem for Integrals
2.3 Double Integrals in Polar Coordinates . . . . .
2.4 Triple Integrals . . . . . . . . . . . . . . . . . .
2.5 Applications of Double and Triple Integrals . .
2.6 Triple Integrals in Cylindrical Coordinates . . .
2.7 Triple Integrals in Spherical Coordinates . . . .
2.8 Change of Variables in Multiple Integrals . . .

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

54
54
57
59
62
63
68
68
69
70

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

73
73
77
80
82
82
87
92
93
95
98

3 Vector Calculus
3.1 Vector Fields . . . . . . . . . . . . . . . . . . . . .
3.2 Line Integrals . . . . . . . . . . . . . . . . . . . . .
3.3 The Fundamental Theorem for Line Integrals . . .
3.4 Greens Theorem . . . . . . . . . . . . . . . . . . .
3.5 Curl and Divergence . . . . . . . . . . . . . . . . .
3.6 Parametric Surfaces and Their Areas . . . . . . . .
3.7 Surface Integrals . . . . . . . . . . . . . . . . . . .
3.7.1 Surface Integrals of Scalar-Valued Functions
3.7.2 Surface Integrals of Vector Fields . . . . . .
3.8 Stokes Theorem . . . . . . . . . . . . . . . . . . .
3.8.1 A Note About Orientation . . . . . . . . . .
3.9 The Divergence Theorem . . . . . . . . . . . . . .
3.10 Differential Forms . . . . . . . . . . . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.

105
105
108
115
120
125
129
134
134
136
143
146
147
150

.
.
.
.
.
.
.
.
.
.

Chapter 1

Partial Derivatives
1.1

Introduction

This course is the fourth course in the calculus sequence, following MAT
167, MAT 168 and MAT 169. Its purpose is to prepare students for more
advanced mathematics courses, particularly courses in mathematical programming (MAT 419), advanced engineering mathematics (MAT 430), real
analysis (MAT 441), complex analysis (MAT 436), and numerical analysis
(MAT 460 and 461). The course will focus on three main areas, which we
briefly discuss here.

1.1.1

Partial Differentiation

In single-variable calculus, you learned how to compute the derivative of a

function of one variable, y = f (x), with respect to its independent variable x,
denoted by dy/dx. In this course, we consider functions of several variables.
In most cases, the functions we use will depend on two or three variables,
denoted by x, y and z, corresponding to spatial dimensions.
When a function f (x, y, z), for example, depends on several variables, it
is not possible to describe its rate of change with respect to those variables
using a single quantity such as the derivative. Instead, this rate of change is a
vector quantity, called the gradient, denoted by f . Each component of the
gradient is the partial derivative of f with respect to one of its independent
variables, x, y or z. That is,
h
i
f
f
f = f
.
x
y
z
For example, the partial derivative of f with respect to x, denoted by
f /x, describes the instantaneous rate of change of f with respect to x,
5

CHAPTER 1. PARTIAL DERIVATIVES

when y and z are kept constant. Partial derivatives can be computed using
the same differentiation techniques as in single-variable calculus, but one
must be careful, when differentiating with respect to one variable, to treat
all other variables as if they are constant. For example, if f (x, y) = x2 y +y 3 ,
then
f
f
= 2xy,
= x2 + 3y 2 ,
x
y
because the y 3 term does not depend on x, and therefore its partial derivative
with respect to x is zero.
If

F1 (x, y, z)
F(x, y, z) = F2 (x, y, z)
F3 (x, y, z)
is a vector-valued function of three variables, then each of its component
functions F1 , F2 , and F3 has a gradient vector, and the rate of change of
F with respect to x, y and z is described by a matrix, called the Jacobian
matrix

JF (x, y, z) =

F1
x
F2
x
F3
x

F1
y
F2
y
F3
y

F1
z
F2
z
F3
z

where each entry of JF (x, y, z) is a partial derivative of one of the component

functions with respect to one of the independent variables.
We will learn how to generalize various concepts and techniques from
single-variable differential calculus to the multi-variable case. These include:
tangent lines, which become tangent planes for functions of two variables and tangent spaces for functions of three or more variables. These
are used to compute linear approximations similar to those of functions
of a single variable.
The Chain Rule, which generalizes from a product of derivatives to a
product of Jacobian matrices, using standard matrix multiplication.
This allows computing the rate of change of a function as its independent variables change along any direction in space, not just along
any of the coordinate axes, which in turn allows determination of the
direction in which a function increases or decreases most rapidly.
computing maximum and minimum values of functions, which, in the
multi-variable case, requires finding points at which the gradient is
equal to the zero vector (corresponding to finding points at which the

1.1. INTRODUCTION

derivative is equal to zero) and checking whether the matrix of second

partial derivatives is positive definite for a minimum, or negative definite for a maximum (which generalizes the second derivative test from
the single-variable case). We will also learn how to compute maximum and minimum values subject to constraints on the independent
variables, using the method of Lagrange multipliers.

1.1.2

Multiple Integration

Next, we will learn how to compute integrals of functions of several variables

over multi-dimensional domains, generalizing the definite integral of a function f (x) over an interval [a, b]. The integral of a function of two variables
f (x, y) represents the volume under a surface described by the graph of f ,
just as the integral of f (x) is the area under the curve described by the
graph of f .
In some cases, it is more convenient to evaluate an integral by first
performing a change of variables, as in the single-variable case. For example,
when integrating a function of two variables, polar coordinates is useful. For
functions of three variables, cylindrical and spherical coordinates, which are
both generalizations of polar coordinates, are worth considering.
In the general case, evaluating the integral of a function of n variables
by first changing to n different variables requires multiplying the integrand
by the determinant of the Jacobian matrix of the function that maps the
new variables to the old. This is a generalization of the u-substitution from
single-variable calculus, and also relates to formulas for area and volume
from MAT 169 that are defined in terms of determinants, or equivalently, in
terms of the dot product and cross product.

1.1.3

Vector Calculus

In the last part of the course, we will study vector fields, which are functions
that assign a vector to each point in its domain, like the vector-valued function F described above. We will first learn how to compute line integrals,
which are integrals of functions along curves. A line integral can be viewed
as a generalization of the integral of a function on an interval, in that dx
is replaced by ds, an infinitesimal distance between points on the curve. It
can also be viewed as a generalization of an integral that computes the arc
length of a curve, as the line integral of a function that is equal to one yields
the arc length. A line integrals of a vector field is useful for computing the
work done by a force applied to an object to move it along a curved path. To

CHAPTER 1. PARTIAL DERIVATIVES

facilitate the computation of line integrals, a variation of the Fundamental

Theorem of Calculus is introduced.
Next, we generalize the notion of a parametric curve to a parametric
surface, in which the coordinates of points on the surface depend on two
parameters u and v, instead of a single parameter t for a parametric curve.
Using the relation between the cross product and the area of a parallelogram,
we can define the integral of a function over a parametric surface, which is
similar to how a change of variables in a double integral is handled. Then,
we will learn how to integrate vector fields over parametric surfaces, which
is useful for computing the mass of fluid that crosses a surface, given the
rate of flow per unit area.
We conclude with discussion of several fundamental theorems of vector
calculus: Greens Theorem, Stokes Theorem, and the Divergence Theorem.
All of these can be seen to be generalizations of the Fundamental Theorem of
Calculus to higher dimensions, in that they relate the integral of a function
over the interior of a domain to an integral of a related function over its
boundary. These theorems can be conveniently stated using the div and
curl operations on vector fields. Specifically, if F = hP, Q, Ri, then
div F = F =

curl F = F =

P
Q R
+
+
,
x
y
z

R Q P
R Q P

y
z z
x x
y


.

However, using the language of differential forms, we can condense the Fundamental Theorem of Calculus and all four of its variations into one theorem,
known as the General Stokes Theorem. We now state all six results; their
discussion is deferred to Chapter 3.
Fundamental Theorem of Calculus:
Z b
f 0 (x) dx = f (b) f (a)
a

where f is continuously differentiable on [a, b]

Fundamental Theorem of Line Integrals:
Z
a

f (r(t)) r0 (t) dt = f (r(b)) f (r(a))

1.2. FUNCTIONS OF SEVERAL VARIABLES

where r(t) = hx(t), y(t), z(t)i, a t b, is the position function for a curve
C and f (x, y, z) is a continuously differentiable function defined on C
Greens Theorem:
Z 
D

Q P

x
y

Z
dA =

P dx + Q dy
D

where D is a 2-D region with piecewise smooth boundary D and P and Q

are continuously differentiable on D
Stokes Theorem:

Z
curl F n dS =

F T ds

where S is a surface in 3-D with unit normal vector n, and piecewise smooth
boundary S with unit tangent vector T, and F is a continuously differentiable vector field
Divergence Theorem:
Z

Z
F n dS

div F dV =
E

where E is a solid region in 3-D with boundary surface E, which has

outward unit normal vector n, and F is a continuously differentiable vector
field

General Stokes Theorem:

Z

Z
d =

where M is a k-manifold and is a (k 1)-form on M

1.2

Functions of Several Variables

Multi-variable calculus is concerned with the study of quantities that depend

on more than one variable, such as temperature that varies within a threedimensional object, which is a scalar quantity, or the velocity of a flowing

10

CHAPTER 1. PARTIAL DERIVATIVES

liquid, which is a vector quantity. To aid in this study, we first introduce

some important terminology and notation that is useful for working with
functions of more than one variable, and then introduce some techniques for
visualizing such functions.

1.2.1

Terminology and Notation

The following standard notation and terminology is used to define, and discuss, functions of several variables and their visual representations. As they
will be used throughout the course, it is important to become acquainted
with them immediately.
The set R is the set of all real numbers.
If S is a set and x is an element of S, we write x S.

Example 2 R, and R, but i

/ R, where i = 1 is an imaginary
number.
If S is a set and T is a subset of S (that is, every element of T is in
S), we write T S.
Example The set of natural numbers, N, and the set of integers, Z,
satisfy N Z. Furthermore, Z R.
The set Rn is the set of all ordered n-tuples x = (x1 , x2 , . . . , xn ) of real
numbers. Each real number xi is called a coordinate of the point x.
Example Each point (x, y) R2 has coordinates x and y. Each point
(x, y, z) R3 has an x-, y- and z-coordinate.
A function f with domain D Rn and range R Rm is a set of
ordered pairs of the form {(x, y)}, where x D and y R, such that
each element x D is mapped to only one element of R. That is, there
is only one ordered pair in f such that x is its first element. We write
f : D R to indicate that f maps elements of D to elements of R.
We also say that f maps D into R.
Example Let R+ denote the set of non-negative real numbers. The
function f (x, y) = x2 + y 2 maps R2 into R+ , and we can write f :
R2 R + .
Let f : D R, and let D Rn and R Rm . If m = 1, we say that
f is a scalar-valued function, and if m > 1, we say that f is a vectorvalued function. If n = 1, we say that f is a function of one variable,

1.2. FUNCTIONS OF SEVERAL VARIABLES

11

and if n > 1, we say that f is a function of several variables. For each

x = (x1 , x2 , . . . , xn ) D, the coordinates x1 , x2 , . . . , xn of x are called
the independent variables of f , and for each y = (y1 , y2 , . . . , ym ) R,
the coordinates y1 , y2 , . . . , ym of y are called the dependent variables
of f .
Example The function z = x2 + y 2 is a scalar-valued function of
several variables. The independent variables are x and y, and the
dependent variable is z. The function r(t) = hx(t), y(t), z(t)i, where
x(t) = t cos t, y(t) = t sin t, and z(t) = et , is a vector-valued function
with independent variable t and dependent variables x, y and z.
Let f : D Rn R. The graph of f is the subset of Rn+1 consisting of
the points (x1 , x2 , . . . , xn , f (x1 , x2 , . . . , xn )), where (x1 , x2 , . . . , xn )
D.
Solution The graph of the function z = x2 + y 2 is a parabola of
revolution obtained by revolving the parabola z = x2 around the zaxis. The graph of the function z = x + y 1 is a plane in 3-D space
that passes through the points (0, 0, 1) and (1, 1, 1).
A function f : Rn R is a linear function if f has the form
f (x1 , x2 , . . . , xn ) = a1 x1 + a2 x2 + + an xn + b,
where a1 , a2 , . . . , an and b are constants.
Example The function y = mx + b is a linear function of the single
independent variable x. Its graph is a line contained within the xyplane, with slope m, passing through the point (0, b). The function
z = ax + by + c is a linear function of the two independent variables
x and y. Its graph is a plane in 3-D space that passes through the
points (0, 0, c) and (1, 1, a + b + c).
Let f : D Rn R. We say that a set L is a level set of f if L D
and f is equal to a constant value k on L; that is, f (x) = k if x L.
If n = 2, we say that L is a level curve or level contour; if n = 3, we
say that L is a level surface.
2
2
Example A level surface of the function f (x, y, z) = x2 +y
+z , where
f (x, y, z) = k for a constant k, is a sphere of radius k. The level
curves of the function z = x2 + y 2 are circles of radius k with center
(0, 0, k), situated in the plane z = k, for each nonnegative number k.

12

CHAPTER 1. PARTIAL DERIVATIVES

1.2.2

Visualization Techniques

While it is always possible to obtain the graph of a function f (x, y), for
example, by substituting various values for its independent variables and
plotting the corresponding points from the graph, this approach is not necessarily helpful for understanding the graph as a whole. Knowing the extent
of the possible values of a functions independent and dependent variables
(the domain and range, respectively), along with the behavior of a few select
curves that are contained within the functions graph, can be more helpful.
To that end, we mention the following useful techniques for acquiring this
information.

Figure 1.1: Level curves of the function z = x2 + y 4

To find the domain and range of a function f , it is often necessary to
account for the domains of functions that are included in the definition
of f . For example, if there is a square root, it is necessary to avoid
taking the square root of a negative number.
Example Let f (x, y) = ln(x2 y 2 ). Since ln |x| is only defined for
x > 0, we must have x2 > y 2 , which, upon taking the square root
of both sides, yields |x| > |y|. Therefore, this inequality defines the
domain of f . The range of f is the range of ln, which is R.
To find the level set of a function f (x1 , x2 , . . . , xn ), solve the equation
f (x1 , x2 , . . . , xn ) = k, where k is a constant. This equation will implicitly define the level set. In some cases, it can be solved for one of

1.2. FUNCTIONS OF SEVERAL VARIABLES

13

Figure 1.2: Sections of the function z = x2 + y 4

the independent variables to obtain an explicit function that describes
the level set.
Example Let z = 2x + y be a function of two variables. The graph
of this function is a plane. Each level set of this function is described
by an equation of the form 2x + y = k, where k is a constant. Since
z = k as well, the graph of this level set is the line with equation
y = 2x + k, contained within the plane z = k.
Example Let z = ln y x. Each level set of this function is described
by an equation of the form ln y x = k, where k is a constant. Exponentiating both sides, we obtain y = ex+k . It follows that the graph of
this level set is that of the exponential function, contained within the
plane z = k, and shifted k units to the left (that is, along the x-axis).
To help visualize a function of two variables z = f (x, y), it can be
helpful to use the method of sections. This involves viewing the func-

14

CHAPTER 1. PARTIAL DERIVATIVES

tions when restricted to vertical planes, such as the xz-plane and
the yz-plane. To take these two sections, first set y = 0 to obtain z as
a function of x, and then graph that function in the xz-plane. Then,
set x = 0 to obtain z as a function of y, and graph that function in
the yz-plane. Using these graphs as guides, in conjunction with level
curves, it is then easier to visualize what the rest of the graph of f
looks like.
Example Let z = x2 + y 4 . Setting y = 0 yields z = x2 , the graph of
which is a parabola in the xz-plane. Setting x = 0 yields z = y 4 , which
has a graph that is a parabola-like curve, where z increases much more
rapidly. Combining these graphs
with selected level curves,
which are
4
2
described by the equations y = k x , where |x| k for k 0,
allows us to visualize the graph of this function. Level curves and
sections are shown in Figures 1.1 and 1.2, respectively.

1.3

Limits and Continuity

Recall that in single-variable calculus, the fundamental concept of a limit

was used to define derivatives and integrals of functions, as well as the notion
of continuity of a function. We now generalize limits and continuity to the
case of functions of several variables.

1.3.1

Terminology and Notation

Let f : D R R, and a D. We say f (x) approaches L as x

approaches a, and write
lim f (x) = L

xa

if, for any  > 0, there exists a > 0 such that if 0 < |x a| < , then
|f (x) L| < .
If x = (x1 , x2 , . . . , xn ) is a point in Rn , or, equivalently, if x =
hx1 , x2 , . . . , xn i is a position vector in Rn , then the magnitude, or
length, of x, denoted by kxk, is defined by
q
kxk = x21 + x22 + + x2n =

n
X

!1/2
x2i

i=1

Note that if n = 1, then x is actually a scalar x, and kxk = |x|.

1.3. LIMITS AND CONTINUITY

Example
If x = (3, 1, 4) R3 , then kxk =

26. 2

15
p

32 + (1)2 + 42 =

Let f : D Rn Rm , and a D. We say f (x) approaches b as x

approaches a, and write
lim f (x) = b,

xa

if, for any  > 0, no matter how small, there exists a > 0 such that
for any x such that 0 < kx ak < , kf (x) bk < . This definition
is illustrated in Figure 1.3. Note that the condition kx ak > 0
specifically excludes consideration of x = a, because limits are used to
understand the behavior of a function near a point, not at a point.

Figure 1.3: Illustration of the limit, as x approaches a (left plot), of f (x)

being equal to b (right plot). For any ball around the point b of radius 
(right plot), no matter how small, there exists a ball around the point a, of
radius (left plot), such that every point in the ball around a is mapped by
f to a point in the ball around b.
Example Consider the function
xy
f (x, y) = p
.
x2 + y 2
We will use the definition of a limit to show that as (x, y) (0, 0),
f (x, y) 0. Let  > 0. We need to show that there exists some > 0

16

CHAPTER 1. PARTIAL DERIVATIVES

such that if 0 < k(x, y) (0, 0)k =
| < . First, we note that
| xy
2
2

p
x2 + y 2 < , then |f (x, y) 0| =

x +y


 


  y   y 
y

 

p
 <  p  =   = 1.
 x2 + y 2   y 2   |y| 
Therefore, if we set = , we obtain








p

y
xy




 = |x|  p
 < |x| = x2 < x2 + y 2 < = ,
p
 x2 + y 2 
 x2 + y 2 
from which it follows that the limit exists and is equal to zero. 2
Let f : D Rn Rm , and let a D. We say that f is continuous at
a if limxa f (x) = f (a).
Let f : D Rn R. We say that f is a polynomial if, for each
x = (x1 , x2 , . . . , xn ) D, f (x) is equal to a sum of terms of the form
xp11 xp22 xpnn , where p1 , p2 , . . . , pn are nonnegative integers.
Example The functions f (x) = x3 +3x2 +2x+1, g(x, y) = x2 y 3 +3xy+
x2 + 1, and h(x, y, z) = 4xy 2 z 3 + 8yz 2 are all examples of polynomials.
2
Let f, p, q : D Rn R, and let q(x) 6= 0 on D. We say that
f is a rational function if p and q are both polynomials and f (x) =
p(x)/q(x).
Example The functions f (x) = 1/(x+1), g(x, y) = xy 2 /(x2 +y 3 ), and
h(x, y, z) = (xy 2 + z 3 )/(x2 z + xyz 2 + yz 3 ) are all examples of rational
functions. 2
An algebraic function is a function that satisfies a polynomial equation
whose coefficients are themselves polynomials.

Example The square root function y = x is an algebraic function,

because it satisfies the equation y 2 x = 0. The function y = x5/2 +
x3/2 is also an algebraic function, because it satisfies the equation
x5 + 2x4 + x3 y 2 = 0. 2

1.3.2

Defining Limits Using Neighborhoods

An alternative approach to defining limits involves the concept of a neighborhood, which generalizes open intervals on the real number line.

1.3. LIMITS AND CONTINUITY

17

Let x0 Rn and let r > 0. We define the ball centered at x0 of radius

r, denoted by Dr (x0 ), to be the set of all points x Rn such that
kx x0 k < r.
Example In 1-D, the open interval (0, 1) is also the ball centered at
x0 = 1/2 of radius r = 1/2. In 3-D, the inside of the sphere with
center (0, 0, 0) and radius 2, {(x, y, z)|x2 + y 2 + z 2 < 4}, is also the
ball D2 ((0, 0, 0)). 2
We say that a set U Rn is open if, for any point x0 U , there exists
an r > 0 such that Dr (x0 ) U .
Example In 1-D, any open set is an open interval, such as (1, 1),
or a union of open intervals. In 2-D, the interior of the ellipse defined
by the equation 4x2 + 9y 2 = 1 is an open set; the ellipse itself is not
included. 2
Let x0 Rn . We say that N is a neighborhood of x0 if N is an open
set that contains x0 .
Let A Rn be an open set. We say that x0 Rn is a boundary point
of A if every neighborhood of x0 contains at least one point in A and
one point not in A.
Example Let D = {(x, y)|x2 + y 2 < 1}, which is often called the unit
ball in R2 . This set consists of all points inside the unit circle with
center (0, 0)and
radius 1, not including the circle itself. The point
(x0 , y0 ) = ( 2/2, 2/2), which is on the circle, is a boundary point of
D because, as illustrated in Figure 1.4, any neighborhood of (x0 , y0 )
must contain points inside the circle, and points that are outside. 2
Let f : D Rn Rm , and let a D or let a be a boundary point of
D. We say that a D. We say that f (x) approaches b as x approaches
a, and write
lim f (x) = b,
xa

if, for any neighborhood N of b, there exists a neighborhood U of a

such that if x U , then f (x) N .

1.3.3

Results

In the statement of the following results concerning limits and continuity,

f , g : D Rn Rm , a D or a is a boundary point of D, b, b1 , b2 Rm ,
and c R.

18

CHAPTER 1. PARTIAL DERIVATIVES

Figure 1.4: Boundary point (x0 , y0 ) of the set D = {(x, y)|x2 + y 2 < 1}. The
neighborhood of (x0 , y0 ) shown, Dr ((x0 , y0 )) = {(x, y)|(xx0 )2 +(y y0 )2 <
0.1}, contains points that are in D and points that are not in D.
,
The limit of a function f (x) as x approaches a, if it exists, is unique.
That is, if
lim f (x) = b1 and lim f (x) = b2 ,
xa

xa

then b1 = b2 . It follows that if f (x) approaches two distinct values as

x approaches a along two distinct paths, then the limit as x approaches
a does not exist.
If limxa f (x) = b, then limxa cf (x) = cb.
If limxa f (x) = b1 and limxa g(x) = b2 , then
lim (f + g)(x) = b1 + b2 .

xa

1.3. LIMITS AND CONTINUITY

19

Furthermore, if m = 1, then
lim (f g)(x) = b1 b2 .

xa

If m = 1 and limxa f (x) = b 6= 0, and f (x) 6= 0 in a neighborhood

of a, then
1
1
lim
= .
xa f (x)
b
If f (x) = (f1 (x), f2 (x), . . . , fm (x)), where f1 , f2 , . . . , fm are the component functions of f , and b = (b1 , b2 , . . . , bm ), then limxa f (x) = b
if and only if limxa fi (x) = bi for i = 1, 2, . . . , m.
If f and g are continuous at a, then so is cf and f + g. If, in addition,
m = 1, then f g is continuous at a. Furthermore, if m = 1 and if f is
nonzero in a neighborhood of a, then 1/f is continuous at a.
If f (x) = (f1 (x), f2 (x), . . . , fm (x)), where f1 , f2 , . . . , fm are the component functions of f , then f is continuous at a if and only if fi is
continuous at a, for i = 1, 2, . . . , m.
Any polynomial function f : Rn R is continuous on all of Rn .
Any rational function f : D Rn R is continuous wherever it is
defined.
Example The function f (x, y) = 2x/(x2 y 2 ) is defined on all of R2
except where x2 y 2 = 0; that is, where |x| = |y|. Therfore, f is
continuous at all such points. 2
Let f : D Rn Rm , and let g : U Rp D. If the composition
(f g)(x) = f (g(x)) defined on U , then f g is continuous at a U if
g is continuous at a and f is continuous at g(a).
Example The function g(x, y) = x2 + y 2 , being a polynomial, is continuous on all of R2 . The function f (z) = sin z is continuous on all of
R. Therefore, the composition (f g)(x, y) = f (g(x, y)) = sin(x2 + y 2 )
is continuous on all of R2 . 2
Algebraic functions, such as xr where r is any rational number (for

example, f (x) = x) and trigonometric functions, such as sin x or

tan x, are continuous wherever they are defined.

20

CHAPTER 1. PARTIAL DERIVATIVES

1.3.4

Techniques for Establishing Limits and Continuity

We now discuss some techniques for computing limits of functions of several

variables, or determining that they do not exist. We also demonstrate how
to determine if a function is continuous at a point.
To show that the limit of a function f : D Rn R as x a does not
exist, try letting x approach a along different paths to see if different values
are approached. If they are, then the limit does not exist.
For example, let n = 2 and let x = (x, y) and a = (a1 , a2 ). Then, try
setting x = a1 in the formula for f (x, y) and letting y approach a2 , or vice
versa. Other possible paths include, for example, setting x = cy, where
c = a1 /a2 , if a2 6= 0, and letting y approach a2 , or considering the cases of
x < a1 and x > a1 , or y < a2 and y > a2 , separately.
Example Let f (x, y) = x3 y/(x4 +y 4 ). If we let (x, y) (0, 0) by first setting
y = 0 and then letting x 0, we observe that f (x, 0) = x3 (0)/(x4 + 0) = 0
for all x 6= 0. This suggests that f (x, y) 0 as (x, y) (0, 0). However, if
we set x = y and let x, y 0 together, we note that f (x, x) = x3 x/(x4 +
x4 ) = x4 /(2x4 ) = 1/2, which suggests that the limit is equal to 1/2. We
conclude that the limit does not exist. 2
To show that the limit of a function f : D Rn R as x a does
exist and is equal to b, use the definition of a limit by first assuming  > 0,
and then trying to find a so that |f (x) b| <  whenever 0 < kx ak < .
To that end, try to find an upper bound on |f (x) b| in terms of kx ak.
Specifically, if it can be shown that |f (x) b| < g(kx ak), where g is an
invertible, increasing function, then a suitable choice for is = g 1 ().
Then, if kx ak < = g 1 (), then
|f (x) b| < g(kx ak) < g(g 1 ()) = .

Example Let f (x, y) = (x3 y 3 )/(x2 + y 2 ). Letting (x, y) (0, 0) along

various paths, it appears that f (x, y) 0 as (x, y) (0, 0).
p To confirm
this, we assume  > 0 and try to find > 0 such that if 0 < x2 + y 2 < ,
then |(x3 y 3 )/(x2 + y 2 )| < .
Factoring the numerator of f (x, y), we obtain
 3
 
 


 x y 3   (x y)(x2 + xy + y 2 )  


=
 = (x y) 1 + xy
.
 x2 + y 2  


2
2
2
2
x +y
x +y 

1.3. LIMITS AND CONTINUITY

21

p
p

Using |x| = x2 x2 + y 2 , and similarly |y| x2 + y 2 , yields


!
!

 3


 x y3 
x
y

 = |x y| 1 + p

p

 x2 + y 2 

x2 + y 2
x2 + y 2 
2|x y|
2(|x| + |y|)
p
4 x2 + y 2 .
p
Therefore, if we let = /4, it follows that when x2 + y 2 < , then
|f (x, y)| < 4 = 4(/4) = , and therefore the limit exists and is equal to
zero. 2
Sometimes, it is helpful to simplify the formula for a function before
attempting to determine its limit.
Example Consider the function
f (x, y) =

(x + y)2 (x y)2
,
xy

x, y 6= 0.

Expanding the numerator yields, for x, y 6= 0,

f (x, y) =

4xy
(x2 + 2xy + y 2 ) (x2 2xy + y 2 )
=
= 4.
xy
xy

Therefore, even though f (x, y) is not defined at (0, 0), its limit as (x, y)
(0, 0) exists, and is equal to 4. This example demonstrates that a limit
depends only on the behavior of a function near a particular point; what
happens at that point is irrelevant. 2
In many cases, determining whether a function f : D Rn R is
continuous can be accomplished by applying the various properties of continuous functions stated above, and using the fact that various types of
functions, such as polynomial and rational functions, are known to be continuous wherever they are defined.
Example Let c = h2, 1, 3i, and let f : R3 R3 be defined by f (x) = cx,
the cross product of the vector c and the vector x = hx1 , x2 , x3 i. This
function is continuous on all of R3 , because
f (x) = c x = hx3 3x2 , 3x1 2x3 , 2x2 + x1i,
and each component function of f can be seen to be not only a polynomial,
but a linear function. 2

22

CHAPTER 1. PARTIAL DERIVATIVES

However, in cases where a function is defined in a piecewise manner,

continuity at boundaries between pieces must be determined by applying
the definition of continuity directly, which requires computing limits.
Example Consider the function
(
f (x, y) =

xy 2
2x2 +y 2

(x, y) 6= (0, 0)
.
(x, y) = (0, 0)

For (x, y) 6= (0, 0), f is continuous at (x, y) because it is a rational function

that is defined. As (x, y) (0, 0), f (x, y) 0, as can be shown by applying
the definition of a limit with = . Because this limit is equal to f (0, 0) = 0,
we conclude that f is continuous at (0, 0) as well. 2

1.4

Partial Derivatives

Now that we have become acquainted with functions of several variables,

and what it means for such functions to have limits and be continuous, we
are ready to analyze their behavior by computing their instantaneous rates
of change, as we know how to do for functions of a single variable. However,
in contrast to the single-variable case, the instantaneous rate of change of
a function of several variables cannot be described by a single number that
represents the slope of a tangent line. Instead, such a slope can only describe
how one of the functions dependent variables (outputs) varies as one of its
independent variables (inputs) changes. This leads to the concept of what
is known as a partial derivative.

1.4.1

Terminology and Notation

Let f : D R R be a scalar-valued function of a single variable. Recall

that the derivative of f (x) with respect to x at x0 is defined to be
df
f (x0 + h) f (x0 )
(x0 ) = f 0 (x0 ) = lim
.
h0
dx
h
Now, let f : D R2 R be a scalar-valued function of two variables,
and let (x0 , y0 ) D. The partial derivative of f (x, y) with respect to x at
(x0 , y0 ) is defined to be
f (x0 + h, y0 ) f (x0 , y0 )
f
(x0 , y0 ) = fx (x0 , y0 ) = lim
.
h0
x
h

1.4. PARTIAL DERIVATIVES

23

Note that only values of f (x, y) for which y = y0 influence the value of
the partial derivative with respect to x. Similarly, the partial derivative of
f (x, y) with respect to y at (x0 , y0 ) is defined to be
f
f (x0 , y0 + h) f (x0 , y0 )
(x0 , y0 ) = fy (x0 , y0 ) = lim
.
h0
y
h
Note the two methods of denoting partial derivatives used above: f /x or
fx for the partial derivative with respect to x. There are other notations,
but these are the ones that we will use.
Example Let f (x, y) = x2 y, and let (x0 , y0 ) = (2, 1). Then
fx (2, 1) =
=
=
=
fy (2, 1) =
=
=
=

(2 + h)2 (1) 22 (1)

h0
h
(4 + 4h + h2 ) + 4
lim
h0
h
4h h2
lim
h0
h
4,
22 (1 + h) 22 (1)
lim
h0
h
4(h 1) + 4
lim
h0
h
4h
lim
h0 h
4.
lim

2
In the preceding example, the value fx (2, 1) = 4 can be interpreted
as the slope of the line that is tangent to the graph of f (x, 1) = x2
at x = 2. That is, we consider the restriction of f to the portion of its
domain where y = 1, and thus obtain a function of the single variable x,
g(x) = f (x, 1) = x2 . Note that if we apply differentiation rules from
single-variable calculus to g, we obtain g 0 (x) = 2x, and g 0 (2) = 4, which
is the value we obtained for fx (2, 1).
Similarly, if we consider fy (2, 1) = 4, this can be interpreted as the
slope of a line that is tangent to the graph of p(y) = f (2, y) = 4y at y = 1.
Note that if we differentiate p, we obtain p0 (y) = 4, which, again, shows that
the partial derivative of a function of several variables can be obtained by
freezing the values of all variables except the one with respect to which we

24

CHAPTER 1. PARTIAL DERIVATIVES

are differentiating, and then applying differentiation rules to the resulting

function of one variable.
It follows from this relationship between partial derivatives of a function
of several variables and the derivative of a function of a single variable
that other interpretations of the derivative are also applicable to partial
derivatives. In particular, if fx (x0 , y0 ) > 0, which is equivalent to g 0 (x0 ) > 0
where g(x) = f (x, y0 ), we can conclude that f is increasing as x varies from
x0 , along the line y = y0 . Similarly, if fy (x0 , y0 ) < 0, which is equivalent to
p0 (y0 ) < 0 where p(y) = f (x0 , y), we can conclude that f is decreasing as y
varies from y0 along the line x = x0 .
We now define partial derivatives for a function of n variables. Let the
vectors e1 , e2 , . . . , en Rn be defined as follows: for each i = 1, 2, . . . , n, ei
has components that are all equal to zero, except the ith component, which
is equal to 1. Then these vectors are called the standard basis vectors of Rn .
Example If n = 3, then

1

e1 = 0 ,
0

0
e2 = 1 ,
0

0
e3 = 0 .
1

We also have that e1 = i, e2 = j and e3 = k. 2

Let f : D Rn R be a scalar-valued function of n variables x1 ,x2 ,. . .,xn .
Then, the partial derivative of f with respect to xi at x0 Rn , where
1 i n, is defined to be
f
(x0 ) = fxi (x0 )
xi
f (x0 + hei ) f (x0 )
= lim
h0
h
f (x1 , . . . , xi + h, . . . , xn ) f (x1 , . . . , xn )
= lim
.
h0
h
Example Let f : R4 R be defined by f (x) = (c x)2 , where c R4 is the
vector c = h4, 3, 2, 1i. Let x0 R4 be the point x0 = h1, 3, 2, 4i. Then,
the partial derivative of f with respect to x2 at x0 is given by
f (x0 + he2 ) f (x0 )
h0
h
(c (x0 + he2 ))2 (c x0 )2
= lim
h0
h

fx2 (x0 ) =

lim

1.4. PARTIAL DERIVATIVES

25

(c x0 + hc e2 )2 (c x0 )2
h0
h
2
(c x0 ) + 2(c x0 )(hc e2 ) + (hc e2 )2 (c x0 )2
= lim
h0
h
2h(c x0 )(c e2 ) + h2 (c e2 )2
= lim
h0
h
= lim 2(c x0 )(c e2 ) + h(c e2 )2
=

lim

h0

= 2(c x0 )(c e2 )
= 2(h4, 3, 2, 1i h1, 3, 2, 4i)(h4, 3, 2, 1i h0, 1, 0, 0i)
= 2[4(1) 3(3) + 2(2) 1(4)](3)
= 2(5)(3)
= 30.
This shows that f is increasing sharply as a function of x2 at the point x0 .
Note that the same result can be obtained by defining
g(x2 ) = f (1, x2 , 2, 4)
= (c h1, x2 , 2, 4i)2
= (h4, 3, 2, 1i h1, x2 , 2, 4i)2
= (4 3x2 )2 ,
differentiating this function of x2 to obtain g 0 (x2 ) = 2(43x2 )(3), and then
evaluating this derivative at x2 = 3 to obtain g 0 (3) = 2(4 3(3))(3) = 30.
2
Just as functions of a single variable can have second derivatives, third
derivatives, or derivatives of any order, functions of several variables can
have higher-order partial derivatives. To that end, let f : D Rn R be a
scalar-valued function of n variables x1 , x2 , . . . , xn . Then, the second partial
derivative of f with respect to xi and xj at x0 D is defined to be
2f
(x0 ) = fxi xj (x0 )
xi xj



f
=
(x0 )
xi xj
fx (x0 + hi ei ) fxj (x0 )
= lim j
hi 0
hi
1
=
lim
[f (x0 + hi ei + hj ej ) f (x0 + hi ei )
(hi ,hj )(0,0) hi hj

26

CHAPTER 1. PARTIAL DERIVATIVES

f (x0 + hj ej ) + f (x0 )].

The second line of the above definition is the most helpful, in terms of
describing how to compute a second partial derivative with respect to xi
and xj : first, compute the partial derivative with respect to xj . Then,
compute the partial derivative of the result with respect to xi , and finally,
evaluate at the point x0 . That is, the second partial derivative, or a partial
derivative of higher order, can be viewed as an iterated partial derivative.
A commonly used method of indicating that a function is evaluated at
a given point, especially if the formula for the function is complicated or
otherwise does not lend itself naturally to the usual notation for evaluation
at a point, is to follow the function with a vertical bar, and indicate the
evaluation point as a subscript to the bar. For example, given a function
f (x), we can write


df 
df 
0
f (4) =
=
dx 
dx 
x=4

or, given a function f (x, y), we can write



f 
f 
=
.
fx (2, 3) =
x x=2,y=3
x (2,3)
This notation is similar to the use of the vertical bar in the evaluation of
definite integrals, to indicate that an antiderivative is to be evaluated at the
limits of integration.

1.4.2

Clairauts Theorem

The following theorem is very useful for reducing the amount of work necessary to compute all of the higher-order partial derivatives of a function.
Theorem (Clairauts Theorem): Let f : D R2 R, and let x0 D.
If the second partial derivatives fxy and fyx are continuous on D, then they
are equal:
fxy (x0 ) = fyx (x0 ).
Example Let f (x, y) = sin(2x) cos2 (4y). Then
fx = 2 cos2 (4y) cos(2x),

fy = 8 sin(2x) cos(4y) sin(4y),

which yields
fxy = (2 cos2 (4y) cos(2x))y = 16 cos(2x) cos(4y) sin(4y)

1.4. PARTIAL DERIVATIVES

27

and
fyx = (8 sin(2x) cos(4y) sin(4y))x = 16 cos(2x) cos(4y) sin(4y),
and we conclude that these mixed partial derivatives are equal. 2

1.4.3

Techniques

We now describe the most practical techniques for computing partial derivatives. As mentioned previously, computing the partial derivative of a function with respect to a given variable, at a given point, is equivalent to freezing the values of all other variables at that point, and then computing the
derivative of the resulting function of one variable at that point.
However, generally it is most practical to compute the partial derivative
as a function of all of the independent variables, which can then be evaluated
at any point at which we wish to know the value of the partial derivative,
just as when we have a function f (x), we normally compute its derivative
as a function f 0 (x), and then evaluate that function at any point x0 where
we want to know the rate of change.
Therefore, the most practical approach to computing a partial derivative
of a function f with respect to xi is to apply differentiation rules from
single-variable calculus to differentiate f with respect to xi , while treating
all other variables as constants. The result of this process is a function that
represents f /xi (x1 , x2 , . . . , xn ), and then values can be substituted for the
independent variables x1 , x2 , . . . , xn .
2

Example To compute fx (/2, ) of f (x, y) = e(x +y ) sin 3x cos 4y, we

treat y as a constant, since we are differentiating with respect to x. Using
the Product Rule and the Chain Rule from single-variable calculus, as well
as the rules for differentiating exponential and trigonometric functions, we
obtain


(x2 +y2 )
fx (/2, ) =
[e
sin 3x cos 4y]
x
x=/2,y=



2
2
= cos 4y [e(x +y ) sin 3x]
x
x=/2,y=



(x2 +y2 )
(x2 +y 2 )
]+e
= cos 4y sin 3x [e
[sin 3x] 
x
x
x=/2,y=


2
2
= cos 4y e(x +y ) sin 3x [(x2 + y 2 )]+
x

28

CHAPTER 1. PARTIAL DERIVATIVES

3e(x

2 +y 2 )

i

cos 3x 

x=/2,y=

h
i
2
2
2
2

= cos 4y 2xe(x +y ) sin 3x + 3e(x +y ) cos 3x 
x=/2,y=
h
((/2)2 + 2 )
= cos 4 2(/2)e
sin(3/2)+
i
2
2
3e((/2) + ) cos 3(/2)
= e5

2 /4

Similarly, to compute fy (/2, ), we treat x as a constant, and apply these

differentiation rules to differentiate with respect to y. Finally, we substitute
x = /2 and y = into the resulting derivative. 2
This approach to differentiation can also be applied to compute higher-order
partial derivatives, as long as any substitution of values for the variables is
deferred to the end.
Example To evaluate the second partial derivatives of f (x, y) = ln |x + y 2 |
at x = 1, y = 2, we first compute the first partial derivatives of f :
fx =

1
[x + y 2 ] =
,
2
x + y x
x + y2

fy =

2y
1
[x + y 2 ] =
.
2
x + y y
x + y2

Next, we differentiate each of these partial derivatives with respect to both

x and y to obtain
fxx = (fx )x


1
=
x + y2 x
1

[x + y 2 ]
=
2
2
(x + y ) x
1
=
,
(x + y 2 )2
fxy = (fx )y


1
=
x + y2 y
=

[x + y 2 ]
2
2
(x + y ) y

1.4. PARTIAL DERIVATIVES

=

29

2y
,
(x + y 2 )2

fyx = fxy
2y
,
(x + y 2 )2
= (fy )y


2y
=
x + y2 y
=

fyy

=
=
=

(x + y 2 )(2y)y 2y(x + y 2 )y
(x + y 2 )2
2
2(x + y ) 4y 2
(x + y 2 )2
2(x y 2 )
.
(x + y 2 )2

Finally, we can evaluate these second partial derivatives at x = 1 and y = 2

to obtain
fxx (1, 2) =

1
,
25

fxy (1, 2) = fyx (1, 2) =

4
,
25

fyy (1, 2) =

6
.
25

2
Example Let f (x, y, z) = x2 y 4 z 3 . We will compute the second partial
derivatives of this function at the point (x0 , y0 , z0 ) = (1, 2, 3) by repeated
computation of first partial derivatives. First, we compute
fx = (x2 y 4 z 3 )x = (x2 )x y 4 z 3 = 2xy 4 z 3 ,
by treating y and z as constants, then
fy = (x2 y 4 z 3 )y = (y 4 )y x2 z 3 = 4x2 y 3 z 3 ,
by treating x and z as constants, and then
fz = (x2 y 4 z 3 )z = x2 y 4 (z 3 )z = 3x2 y 4 z 2 .
We then differentiate each of these with respect to x, y and z to obtain the
second partial derivatives:
fxx = (fx )x = (2xy 4 z 3 )x = 2y 4 z 3 ,
fxy = (fx )y = (2xy 4 z 3 )y = (2x)(4y 3 )(z 3 ) = 8xy 3 z 3 ,

30

CHAPTER 1. PARTIAL DERIVATIVES

fxz = (fx )z = (2xy 4 z 3 )z = (2xy 4 )(3z 2 ) = 6xy 4 z 2 ,
fyx = (fy )x = (4x2 y 3 z 3 )x = 8xy 3 z 3 ,
fyy = (fy )y = (4x2 y 3 z 3 )y = (4x2 )(3y 2 )(z 3 ) = 12x2 y 2 z 3 ,
fyz = (fy )z = (4x2 y 3 z 3 )z = (4x2 y 3 )(3z 2 ) = 12x2 y 3 z 2 ,
fzx = (fz )x = (3x2 y 4 z 2 )x = 6xy 4 z 2 ,
fzy = (fz )y = (3x2 y 4 z 2 )y = (3x2 )(4y 3 )(z 2 ) = 12x2 y 3 z 2 ,
fzz = (fz )z = (3x2 y 4 z 2 )z = (3x2 y 4 )(2z) = 6x2 y 4 z.

Then, these can be evaluated at (x0 , y0 , z0 ) by substituting x = 1, y = 2,

and z = 3 to obtain
fxx (1, 2, 3) = 864,

fxy (1, 2, 3) = 1728,

fyx (1, 2, 3) = 1728,

fyy (1, 2, 3) = 1296,

fzx (1, 2, 3) = 864,

fzy (1, 2, 3) = 864,

fxz (1, 2, 3) = 864,

fyz (1, 2, 3) = 864,
fzz (1, 2, 3) = 288.

Note that the order in which partial differentiation operations occur does
not appear to matter; that is, fxy = fyx , for example. That is, Clairauts
Theorem applies for any number of variables. It also applies to any order of
partial derivative. For example,
fxyy = (fxy )y = (8xy 3 z 3 )y = 24xy 2 z 3 ,
fyyx = (fyy )x = (12x2 y 2 z 3 )x = 24xy 2 z 3 .
2
In single-variable calculus, implicit differentiation is applied to an equation that implicitly describes y as a function of x, in order to compute dy/dx.
The same approach can be applied to an equation that implicitly describes
any number of dependent variables in terms of any number of independent
variables. The approach is the same as in the single-variable case: differentiate both sides of the equation with respect to the independent variable,
leaving derivatives of dependent variables in the equation as unknowns. The
resulting equation can then be solved for the unknown partial derivatives.
Example Consider the equation
x2 z + y 2 z + z 2 = 1.
If we view this equation as one that implicitly describes z as a function of x
and y, we can compute zx and zy using implicit differentiation with respect
to x and y, respectively. Applying the Product Rule yields the equations
2xz + x2 zx + y 2 zx + 2zzx = 0,

1.5. TANGENT PLANES, LINEAR APPROXIMATIONS AND DIFFERENTIABILITY31

x2 zy + 2yz + y 2 zy + 2zzy = 0,
which can then be solved for the partial derivatives to obtain
zx =

x2

2xz
,
+ y 2 + 2z

zy =

x2

2yz
.
+ y 2 + 2z

1.5

Tangent Planes, Linear Approximations and

Differentiability

Now that we have learned how to compute partial derivatives of functions of

several independent variables, in order to measure their instantaneous rates
of change with respect to these variables, we will discuss another essential
application of derivatives: the approximation of functions by linear functions. Linear functions are the simplest to work with, and for this reason,
there are many instances in which functions are replaced by a linear approximation in the context of solving a problem such as solving a differential
equation.

1.5.1

Tangent Planes and Linear Approximations

In single-variable calculus, we learned that the graph of a function f (x) can

be approximated near a point x0 by its tangent line, which has the equation
y = f (x0 ) + f 0 (x0 )(x x0 ).
For this reason, the function Lf (x) = f (x0 ) + f 0 (x0 )(x x0 ) is also referred
to as the linearization, or linear approximation, of f (x) at x0 .
Now, suppose that we have a function of two variables, f : D R2
R, and a point (x0 , y0 ) D. Furthermore, suppose that the first partial
derivatives of f , fx and fy , exist at (x0 , y0 ). Because the graph of this
function is a surface, it follows that a linear function that approximates f
near (x0 , y0 ) would have a graph that is a plane.
Just as the tangent line of f (x) at x0 passes through the point (x0 , f (x0 )),
and has a slope that is equal to f 0 (x0 ), the instantaneous rate of change of
f (x) with respect to x at x0 , a plane that best approximates f (x, y) at
(x0 , y0 ) must pass through the point (x0 , y0 , f (x0 , y0 )), and the slope of the
plane in the x- and y-directions, respectively, should be equal to the values
of fx (x0 , y0 ) and fy (x0 , y0 ).

32

CHAPTER 1. PARTIAL DERIVATIVES

Since a general linear function of two variables can be described by the

formula
Lf (x, y) = A(x x0 ) + B(y y0 ) + C,
so that Lf (x0 , y0 ) = C, and a simple differentiation yields
Lf
= A,
x

Lf
= B,
y

we conclude that the linear function that best approximates f (x, y) near
(x0 , y0 ) is the linear approximation
Lf (x, y) = f (x0 , y0 ) +

f
f
(x0 , y0 )(x x0 ) +
(x0 , y0 )(y y0 ).
x
y

Furthermore, the graph of this function is called the tangent plane of f (x, y)
at (x0 , y0 ). Its equation is
z z0 =

f
f
(x0 , y0 )(x x0 ) +
(x0 , y0 )(y y0 ).
x
y

Example Let f (x, y) = 2x2 y+3y 2 , and let (x0 , y0 ) = (1, 1). Then f (x0 , y0 ) =
5, and the first partial derivatives at (x0 , y0 ) are
fx (1, 1) = 4xy|x=1,y=1 = 4,

fy (1, 1) = 2x2 + 6y|x=1,y=1 = 8.

It follows that the tangent plane at (1, 1) has the equation

z 5 = 4(x 1) + 8(y 1),
and the linearization of f at (1, 1) is
Lf (x, y) = 5 + 4(x 1) + 8(y 1).
Let (x, y) = (1.1, 1.1). Then f (x, y) = 6.292, while Lf (x, y) = 6.2, for an
error of 6.292 6.2 = 0.092. However, if (x, y) = (1.01, 1.01), then f (x, y) =
5.120902, while Lf (x, y) = 5.12, for an error of 5.120902 5.12 = 0.000902.
That is, moving 10 times as close to (1, 1) decreased the error by a factor of
over 100. 2
Another useful application of a linear approximation is to estimate the
error in the value of a function, given estimates of error in its inputs. Given
a function z = f (x, y), and its linearization Lf (x, y) around a point (x0 , y0 ),
if x0 and y0 are measured values and dx = x x0 and dz = y y0 are

1.5. TANGENT PLANES, LINEAR APPROXIMATIONS AND DIFFERENTIABILITY33

regarded as errors in x0 and y0 , then the error in z can be estimated by
computing
dz = z z0 = Lf (x, y) f (x0 , y0 )
= [f (x0 , y0 ) + fx (x0 , y0 )(x x0 ) + fy (x0 , y0 )(y y0 )] f (x0 , y0 )
= fx (x0 , y0 ) dx + fy (x0 , y0 ) dy.
The variables dx and dy are called differentials, and dz is called the total
differential, as it depends on the values of dx and dy. The total differential
dz is only an estimate of the error in z; the actual error is given by z =
f (x, y) f (x0 , y0 ), when the actual errors in x and y, x = x x0 and
y = y y0 , are known. Since this is rarely the case in practice, one instead
estimates the error in z from estimates dx and dy of the errors in x and y.
Example Recall that the volume of a cylinder with radius r and height h
is V = r2 h. Suppose that r = 5 cm and h = 10 cm. Then the volume is
V = 250 cm3 . If the measurement error in r and h is at most 0.1 cm, then,
to estimate the error in the computed volume, we first compute
Vr = 2rh = 100,

Vh = r2 = 25.

It follows that the error in V is approximately

dV = Vr dr + Vh dh = 0.1(100 + 25) = 12.5 cm3 .
If we specify r = 0.1 and h = 0.1, and compute the actual volume using
radius r + r = 5.1 and height h + h = 10.1, we obtain
V + V = (5.1)2 (10.1) = 262.701 cm3 ,
which yields the actual error
V = 262.701 250 = 12.701 cm3 .
Therefore, the estimate of the error, dV , is quite accurate. 2

1.5.2

Functions of More than Two Variables

The concepts of a tangent plane and linear approximation generalize to

more than two variables in a straightforward manner. Specifically, given
(0) (0)
(0)
f : D Rn R and p0 = (x1 , x2 , . . . , xn ) D, we define the tangent

34

CHAPTER 1. PARTIAL DERIVATIVES

space of f (x1 , x2 , . . . , xn ) at p0 to be the n-dimensional hyperplane in Rn+1

whose points (x1 , x2 , . . . , xn , y) satisfy the equation
y y0 =

f
f
f
(0)
(0)
(p0 )(x1 x1 ) +
(p0 )(x2 x2 ) + +
(p0 )(xn x(0)
n ),
x1
x2
xn

where y0 = f (p0 ). Similarly, the linearization of f at p0 is the function

Lf (x1 , x2 , . . . , xn ) defined by
f
f
(0)
(0)
(p0 )(x1 x1 ) +
(p0 )(x2 x2 ) +
x1
x2
f
+
(p0 )(xn x(0)
n ).
xn

Lf (x1 , x2 , . . . , xn ) = y0 +

1.5.3

The Gradient Vector

It can be seen from the above definitions that writing formulas that involve
the partial derivatives of functions of n variables can be cumbersome. This
can be addressed by expressing collections of partial derivatives of functions
of several variables using vectors and matrices, especially for vector-valued
functions of several variables.
(0) (0)
(0)
By convention, a point p0 = (x1 , x2 , . . . , xn ), which can be identified
(0) (0)
(0)
with the position vector p0 = hx1 , x2 , . . . , xn i, is considered to be a
column vector
(0)
x
1(0)
x2

p0 =
.. .
.

(0)
xn
Also, by convention, given a function of n variables, f : D Rn R,
the collection of its partial derivatives with respect to all of its variables is
written as a row vector
i
h
f
f
f
.
f (p0 ) = x
(p
)
(p
)

(p
)
0
0
0
x2
xn
1
This vector is called the gradient of f at p0 .
Viewing the partial derivatives of f as a vector allows us to use vector
operations to describe, much more concisely, the linearization of f . Specifically, the linearization of f at p0 , evaluated at a point p = (x1 , x2 , . . . , xn ),
can be written as
Lf (p) = f (p0 ) +

f
f
(0)
(0)
(p0 )(x1 x1 ) +
(p0 )(x2 x2 ) +
x1
x2

1.5. TANGENT PLANES, LINEAR APPROXIMATIONS AND DIFFERENTIABILITY35

f
(p0 )(xn x(0)
n )
xn
n
X
f
(0)
(p0 )(xi xi )
= f (p0 ) +
xi
+

i=1

= f (p0 ) + f (p0 ) (p p0 ),
where f (p0 ) (p p0 ) is the dot product, also known as the inner product,
of the vectors f (p0 ) and p p0 . Recall that given two vectors u =
hu1 , u2 , . . . , un i and v = hv1 , v2 , . . . , vn i, the dot product of u and v, denoted
by u v, is defined by
uv =

n
X

ui vi = u1 v1 + u2 v2 + + un vn = kukkvk cos ,

i=1

where is the angle between u and v.

Example Let f : R3 R be defined by
f (x, y, z) = 3x2 y 3 z 4 .
Then
f (x, y, z) =

fx fy fz

6xy 3 z 4 9x2 y 2 z 4 12x2 y 3 z 3

Let (x0 , y0 , z0 ) = (1, 2, 1). Then

f (x0 , y0 , z0 ) = f (1, 2, 1)


fx (1, 2, 1) fy (1, 2, 1) fz (1, 2, 1)
=


48 36 96 .
=
It follows that the linearization of f at (x0 , y0 , z0 ) is
Lf (x, y, z) = f (1, 2, 1) + f (1, 2, 1) hx 1, y 2, z + 1i
= 24 + h48, 36, 96i hx 1, y 2, z + 1i
= 24 + 48(x 1) + 36(y 2) 96(z + 1)
= 48x + 36y 96z 192.
At the point (1.1, 1.9, 1.1), we have f (1.1, 1.9, 1.1) 36.5, while Lf (1.1, 1.9, 1.1) =
34.8. Because f is changing rapidly in all coordinate directions at (1, 2, 1),
it is not surprising that the linearization of f at this point is not highly
accurate. 2

36

1.5.4

CHAPTER 1. PARTIAL DERIVATIVES

The Jacobian Matrix

Now, let f : D Rn Rm be a vector-valued function of n variables, with

component functions

f1 (p)
f2 (p)

f (p) =
,
..

.
fm (p)
where each fi : D Rm . Combining the two conventions described above,
the partial derivatives of these component functions at a point p0 D are
arranged in an m n matrix
f

f1
f1
1
x1 (p0 )
x2 (p0 ) xn (p0 )
f2

f2
f2
(p0 )
(p0 ) x
x1 (p0 ) x
n
2
.
Jf (p0 ) =
..
..

fm
x1 (p0 )

fm
x2 (p0 )

fm
xn (p0 )

This matrix is called the Jacobian matrix of f at p0 . It is also referred to

as the derivative of f at x0 , since it reduces to the scalar f 0 (x0 ) when f is a
scalar-valued function of one variable. Note that rows of Jf (p0 ) correspond
to component functions, and columns correspond to independent variables.
This allows us to view Jf (p0 ) as the following collections of rows or columns:

f1 (p0 )
f2 (p0 ) h
i

f
f
f
(p
)
(p
)

(p
)
Jf (p0 ) =
=
.

0
0
0
..
x1
x2
xn

.
fm (p0 )
The Jacobian matrix provides a concise way of describing the linearization of a vector-valued function, just the gradient does for a scalar-valued
function. The linearization of f at p0 is the function Lf (p), defined by

f1
f1 (p0 )
x1 (p0 )

f2
f2 (p0 )

x1 (p0 )

(x1 x(0) ) +
Lf (p) =
+
..
.
1

.
.

fm
fm (p0 )
x1 (p0 )
f

1
(p0 )
n

x
f2
xn (p0 )
(xn x(0)

+
..
n )

fm
xn (p0 )

1.5. TANGENT PLANES, LINEAR APPROXIMATIONS AND DIFFERENTIABILITY37

= f (p0 ) +

n
X
f
(0)
(p0 )(xj xj )
xj
j=1

= f (p0 ) + Jf (p0 )(p p0 ),

where the expression Jf (p0 )(p p0 ) involves matrix multiplication of the
matrix Jf (p0 ) and the vector p p0 . Note the similarity between this
definition, and the definition of the linearization of a function of a single
variable.
In general, given a m n matrix A; that is, a matrix A with m rows
and n columns, and an n p matrix B, the product AB is the m p matrix
C, where the entry in row i and column j of C is obtained by computing
the dot product of row i of A and column j of B. When computing the
linearization of a vector-valued function f at the point p0 in its domain, the
ith component function of the linearization is obtained by adding the value
of the ith component function at p0 , fi (p0 ), to the dot product of fi (p0 )
and the vector p p0 , where p is the vector at which the linearization is to
be evaluated.
Example Let f : R2 R2 be defined by

  x

e cos y
f1 (x, y)
.
=
f (x, y) =
e2x sin y
f2 (x, y)
Then the Jacobian matrix, or derivative, of f is the 2 2 matrix
 


 
(f1 )x (f1 )y
f1 (x, y)
ex cos y
ex sin y
=
.
Jf (x, y) =
=
f2 (x, y)
(f2 )x (f2 )y
2e2x sin y e2x cos y
Let (x0 , y0 ) = (0, /4). Then we have
"
Jf (x0 , y0 ) =

2
2

22
2
2

#
,

and the linearization of f at (x0 , y0 ) is





f1 (x0 , y0 )
x x0
Lf (x, y) =
+ Jf (x0 , y0 )
f2 (x0 , y0 )
y y0
" # "
#

2
2
2

0
2
2
2

=
+
2
2
y 4
2
2
2
"

#
2
2
2

+
x

2
2
2
4

=

.
2
2

2 2x + 2 y 4

38

CHAPTER 1. PARTIAL DERIVATIVES

At the point (x1 , y1 ) = (0.1, 0.8), we have





0.76998
0.76749
f (x1 , y1 )
, Lf (x1 , y1 )
.
0.58732
0.57601
Because of the relatively small partial derivatives at (x0 , y0 ), the linearization at this point yields a fairly accurate approximation at (x1 , y1 ). 2

1.5.5

Differentiability

Before using a linearization to approximate a function near a point p0 , it is

helpful to know whether this linearization is actually an accurate approximation of the function in the first place. That is, we need to know if the
function is differentiable at p0 , which, informally, means that its instantaneous rate of change at p0 is well-defined. In the single-variable case, a
function f (x) is differentiable at x0 if f 0 (x0 ) exists; that is, if the limit
f 0 (x0 ) = lim

xx0

f (x) f (x0 )
x x0

exists. In other words, we must have

lim

xx0

f (x) f (x0 ) f 0 (x0 )(x x0 )

= 0.
x x0

But f (x0 ) + f 0 (x0 )(x x0 ) is just the linearization of f at x0 , so we can

say that f is differentiable at x0 if and only if
lim

xx0

f (x) Lf (x)
= 0.
x x0

Note that this is a stronger statement than simply requiring that

lim f (x) Lf (x) = 0,

xx0

because as x approaches x0 , |1/(x x0 )| approaches , so the difference

f (x)Lf (x) must approach zero particularly rapidly in order for the fraction
[f (x) Lf (x)]/(x x0 ) to approach zero. That is, the linearization must be
a sufficiently accurate approximation of f near x0 for this to be the case, in
order for f to be differentiable at x0 .
This notion of differentiability is readily generalized to functions of several variables. Given f : D Rn Rm , and p0 D, we say that f is
differentiable at p0 if
lim

pp0

kf (p) Lf (p)k
= 0,
kp p0 k

1.5. TANGENT PLANES, LINEAR APPROXIMATIONS AND DIFFERENTIABILITY39

where Lf (p) is the linearization of f at p0 .
Example Let f (x, y) = x2 y. To verify that this function is differentiable at
(x0 , y0 ) = (1, 1), we first compute fx = 2xy and fy = x2 . It follows that the
linearization of f at (1, 1) is
Lf (x, y) = f (1, 1) + fx (1, 1)(x 1) + fy (1, 1)(y 1)
= 1 + 2(x 1) + (y 1) = 2x + y 2.
Therefore, f is differentiable at (1, 1) if
|x2 y (2x + y 2)|
|x2 y (2x + y 2)|
p
=
lim
= 0.
k(x, y) (1, 1)k
(x,y)(1,1)
(x,y)(1,1)
(x 1)2 + (y 1)2
lim

By rewriting this expression as

|x2 y (2x + y 2)|
|x 1||y(x + 1) 2|
p
=p
,
2
2
(x 1) + (y 1)
(x 1)2 + (y 1)2
and noting that
lim
(x,y)(1,1)

|y(x + 1) 2| = 0,

|x 1|
1,
0 p
(x 1)2 + (y 1)2

we conclude that the limit actually is zero, and therefore f is differentiable.

2
There are three important conclusions that we can make regarding differentiable functions:
If all partial derivatives of f at p0 exist, and are continuous, then f is
differentiable at p0 .
Furthermore, if f is differentiable at p0 , then it is continuous at p0 .
Note that the converse is not true; for example, f (x) = |x| is continuous at x = 0, but it is not differentiable there, because f 0 (x) does not
exist there.
If f is differentiable at p0 , then its first partial derivatives exist at
p0 . This statement might seem redundant, because the first partial
derivatives are used in the definition of the linearization, but it is
important nonetheless, because the converse of this statement is not
true. That is, if a functions first partial derivatives exist at a point,
it is not necessarily differentiable at that point.

40

CHAPTER 1. PARTIAL DERIVATIVES

The notion of differentiability is related to not only partial derivatives, which

only describe how a function changes as one of its variables changes, but also
the instantaneous rate of change of a function as its variables change along
any direction. If a function is differentiable at a point, that means its rate of
change along any direction is well-defined. We will explore this idea further
later in this chapter.

1.6

The Chain Rule

Recall from single-variable calculus that if a function g(x) is differentiable at

x0 , and f (x) is differentiable at g(x0 ), then the derivative of the composition
(f g)(x) = f (g(x)) is given by the Chain Rule
(f g)0 (x0 ) = f 0 (g(x0 ))g 0 (x0 ).
We now generalize the Chain Rule to functions of several variables. Let
f : D Rn Rm , and let g : U Rp D. That is, the range of g is the
domain of f .
Assume that g is differentiable at a point p0 U , and that f is differentiable at the point q0 = g(p0 ). Then, f has a Jacobian matrix Jf (q0 ), and
g has a Jacobian matrix Jg (p0 ). These matrices contain the first partial
derivatives of f and g evaluated at q0 and p0 , respectively.
Then, the Chain Rule states that the derivative of the composition (f
g) : U Rm , defined by (f g)(x) = f (g(x)), at p0 , is given by the Jacobian
matrix
Jf g (p0 ) = Jf (g(p0 ))Jg (p0 ).
That is, the derivative of f g at p0 is the product, in the sense of matrix
multiplication, of the derivative of f at g(p0 ) and the derivative of g at p0 .
This is entirely analogous to the Chain Rule from single-variable calculus,
in which the derivative of f g at x0 is the product of the derivative of f at
g(x0 ) and the derivative of g at x0 .
It follows from the rules of matrix multiplication that the partial derivative of the ith component function of f g with respect to the variable xj ,
an independent variable of g, is given by the dot product of the gradient of
the ith component function of f with the vector that contains the partial
derivatives of the component functions of g with respect to xj . We now
illustrate the application of this general Chain Rule with some examples.
Example Let f : R3 R be defined by
f (x, y, z) = ez cos 2x sin 3y,

1.6. THE CHAIN RULE

41

and let g : R R3 be a vector-valued function of one variable defined by

g(t) = hx(t), y(t), z(t)i = h2t, t2 , t3 i.
Then, f g is a scalar-valued function of t,
3

(f g)(t) = ez(t) cos 2x(t) sin 3y(t) = et cos 4t sin 3t2 .

To compute its derivative with respect to t, we first compute
 


f = fx fy fz = 2ez sin 2x sin 3y 3ez cos 2x cos 3y ez cos 2x sin 3y ,
and
g0 (t) = hx0 (t), y 0 (t), z 0 (t)i = h2, 2t, 3t2 i,
and then apply the Chain Rule to obtain
df
dt

= f (x(t), y(t), z(t)) g0 (t)

fx (x(t), y(t), z(t)) fy (x(t), y(t), z(t)) fz (x(t), y(t), z(t))

dx
dt
dy
dt
dz
dt

dx
dy
dz
+ fy (x(t), y(t), z(t)) + fz (x(t), y(t), z(t))
dt
dt
dt
z(t)
z(t)
= (2e sin 2x(t) sin 3y(t))(2) + (3e cos 2x(t) cos 3y(t))(2t) +
= fx (x(t), y(t), z(t))

(ez(t) cos 2x(t) sin 3y(t))(3t2 )

3

= 4et sin 4t sin 3t2 + 6tet cos 4t cos 3t2 + 3t2 et cos 4t sin 3t2 .
2
Example Let f : R2 R be defined by
f (x, y) = x2 y + xy 2 ,
and let g : R2 R2 be defined by

 

x(s, t)
2s + t
g(s, t) =
=
.
y(s, t)
s 2t
Then, f g is a scalar-valued function of s and t,
(f g)(s, t) = x(s, t)2 y(s, t)+x(s, t)y(s, t)2 = (2s+t)2 (s2t)+(2s+t)(s2t)2 .

42

CHAPTER 1. PARTIAL DERIVATIVES

To compute its gradient, which includes its partial derivatives with respect
to s and t, we first compute
 


f = fx fy = 2xy + y 2 x2 + 2xy ,
and


Jg (s, t) =

xs xt
ys yt


=

2 1
1 2


,

and then apply the Chain Rule to obtain

(f g)(s, t) = f (x(s, t), y(s, t))Jg (s, t)



 xs xt
fx (x(s, t), y(s, t)) fy (x(s, t), y(s, t))
=
ys yt

fx (x(s, t), y(s, t))xs + fy (x(s, t), y(s, t))ys
=

fx (x(s, t), y(s, t))xt + fy (x(s, t), y(s, t))yt

[2x(t)y(t) + y(t)2 ](2) + [x(t)2 + 2x(t)y(t)](1)
=

[2x(t)y(t) + y(t)2 ](1) + [x(t)2 + 2x(t)y(t)](2)

4(2s + t)(s 2t) + 2(s 2t)2 + (2s + t)2 + 2(2s + t)(s 2t)
=
2(2s + t)(s 2t) + (s 2t)2 2(2s + t)2 4(2s + t)(s 2t)
2
Example Let f : R R be defined by
f (x) = x3 + 2x2 ,
and let g : R2 R be defined by
g(u, v) = sin u cos v.
Then f g is a scalar-valued function of u and v,
(f g)(u, v) = (sin u cos v)3 + 2(sin u cos v)2 .
To compute its gradient, which includes partial derivatives with respect to
u and v, we first compute
f 0 (x) = 3x2 + 4x,
and
g =

gu gv

cos u cos v sin u sin v

1.6. THE CHAIN RULE

43

and then use the Chain Rule to obtain

(f g)(u, v) = f 0 (g(u, v))g(u, v)
= [3(g(u, v))2 + 4g(u, v)]


cos u cos v sin u sin v


= [3 sin2 u cos2 v + 4 sin u cos v] cos u cos v sin u sin v


(3 sin2 u cos2 v + 4 sin u cos v) cos u cos v (3 sin2 u cos2 v + 4 sin u cos v) sin u sin v .
=
2
Example Let f : R2 R2 be defined by

  2 
f1 (x, y)
x y
f (x, y) =
=
,
f2 (x, y)
xy 2
and let g : R R2 be defined by
g(t) = hx(t), y(t)i = hcos t, sin ti.
Then f g is a vector-valued function of t,
f (t) = hcos2 t sin t, cos t sin2 ti.
To compute its derivative with respect to t, we first compute


 
(f1 )x (f1 )y
2xy x2
Jf (x, y) =
,
=
(f2 )x (f2 )y
y 2 2xy
and g0 (t) = h sin t, cos ti, and then use the Chain Rule to obtain

 0

x (t)
(f1 )x (x(t), y(t)) (f1 )y (x(t), y(t))
0
0
(f g) (t) = Jf (x(t), y(t))g (t) =
y 0 (t)
(f2 )x (x(t), y(t)) (f2 )y (x(t), y(t))


= (f1 )x (x(t), y(t))x0 (t) + (f1 )y (x(t), y(t))y 0 (t), (f2 )x (x(t), y(t))x0 (t) + (f2 )y (x(t), y(t))y 0 (t)
= h2 cos t sin t( sin t) + cos2 t(cos t), sin2 t( sin t) + 2 cos t sin t(cos t)i
= h2 cos t sin2 t + cos3 t, sin3 t + 2 cos2 t sin ti.
2

1.6.1

The Implicit Function Theorem

The Chain Rule can also be used to compute partial derivatives of implicitly defined functions in a more convenient way than is provided by implicit
differentiation. Let the equation F (x, y) = 0 implicitly define y as a differentiable function of x. That is, y = f (x) where F (x, f (x)) = 0 for x in

44

CHAPTER 1. PARTIAL DERIVATIVES

the domain of f . If F is differentiable, then, by the Chain Rule, we can

differentiate the equation F (x, y(x)) = 0 with respect to x and obtain
Fx + Fy

dy
= 0,
dx

which yields
dy
Fx
= .
dx
Fy
By the Implicit Function Theorem, the equation F (x, y) = 0 defines y implicitly as a function of x near (x0 , y0 ), where F (x0 , y0 ) = 0, provided that
Fy (x0 , y0 ) 6= 0 and Fx and Fy are continuous near (x0 , y0 ). Under these
conditions, we see that dy/dx is defined at (x0 , y0 ) as well.
Example Let F : R2 R be defined by
F (x, y) = x2 + y 2 4.
The equation F (x, y) = 0 defines y implicitly as a function of x, provided
that F satisfies the conditions of the Implicit Function Theorem.
We have
Fx = 2x, Fy = 2y.
Since both of these partial derivatives are polynomials, and therefore are
continuous on all of R2 , it follows that if Fy 6= 0, then y can be implicitly
defined as a function of x at a point where F (x, y, z) = 0, and
dy
Fx
x
=
= .
dx
Fy
y
For example, at the point (x, y) = (0, 2), F (x, y) = 0, and Fy = 4. Therefore,
y can be implicitly defined as a function of x near this point, and at x = 0,
we have dy/dx = 0. 2
(0)

(0)

(0)

More generally, let F : D Rn+1 R, and let p0 = (x1 , x2 , . . . , xn , y (0) )

(0) (0)
(0)
D be such that F (x1 , x2 , . . . , xn , y (0) ) = 0. In this case, the Implicit
Function Theorem states that if Fy 6= 0 near p0 , and all first partial derivatives of F are continuous near p0 , then this equation defines y as a function
of x1 , x2 , . . . , xn , and
Fx
y
= i,
xi
Fy

i = 1, 2, . . . , n.

1.6. THE CHAIN RULE

45
(0)

(0)

(0)

To see this, we differentiate the equation F (x1 , x2 , . . . , xn , y (0) ) = 0 with

respect to xi to obtain the equation
Fxi + Fy

y
= 0,
xi

where all partial derivatives are evaluated at p0 , and solve for y/xi at p0 .
Example Let F : R3 R be defined by
F (x, y, z) = x2 z + z 2 y 2xyz + 1.
The equation F (x, y, z) = 0 defines z implicitly as a function of x and y,
provided that F satisfies the conditions of the Implicit Function Theorem.
We have
Fy = z 2 2xz,

Fx = 2xz 2yz,

Fz = x2 + 2yz 2xy.

Since all of these partial derivatives are polynomials, and therefore are continuous on all of R3 , it follows that if Fz 6= 0, then z can be implicitly defined
as a function of x and y at a point where F (x, y, z) = 0, and
zx =

Fx
2yz 2xz
= 2
,
Fz
x + 2yz 2xy

zy =

Fy
2xz z 2
= 2
.
Fz
x + 2yz 2xy

For example, at the point (x, y, z) = (1, 0, 1), F (x, y, z) = 0, and Fz = 1.

Therefore, z can be implicitly defined as a function of x and y near this
point, and at (x, y) = (1, 0), we have zx = 2 and zy = 3. 2
We now consider the most general case: let F : D Rn+m Rm , and
let
(0)

(0)

(0)

(0)

(0)
p0 = (x1 , x2 , . . . , x(0)
n , y1 , y2 , . . . , ym ) D

be such that
(0)

(0)

(0)

(0)

(0)
F(x1 , x2 , . . . , x(0)
n , y1 , y2 , . . . , ym ) = 0.

If we differentiate this system of equations with respect to xi , we obtain the

systems of linear equations
Fxi + Fy1

y1
y2
ym
+ F y2
+ + F ym
= 0,
xi
xi
xi

where all partial derivatives are evaluated at p0 .

i = 1, 2, . . . , n,

46

CHAPTER 1. PARTIAL DERIVATIVES

To examine the solvability of these systems of equations, we first define

(0) (0)
(0)
x0 = (x1 , x2 , . . . , xn ), and denote the component functions of the vectorvalued function F by F = hF1 , F2 , . . . , Fm i. We then define the Jacobian
matrices
F1

F1
F1
x1 (p0 )
x2 (p0 )
xn (p0 )
F2 (p0 ) F2 (p0 ) F2 (p0 )
x1

x2
xn
Jx,F (p0 ) =
,
.
.
..
..

Jy,F (p0 ) =

and

Jy (x0 ) =

Fm
x1 (p0 )

Fm
x2 (p0 )

Fm
xn (p0 )

F1
y1 (p0 )
F2
y1 (p0 )

F1
y2 (p0 )
F2
y2 (p0 )

F1
ym (p0 )
F2
ym (p0 )

..
.
Fm
y1 (p0 )

Fm
y2 (p0 )

y1
x1 (x0 )
y2
x1 (x0 )

y1
x2 (x0 )
y2
x2 (x0 )

..
.
ym
x1 (x0 )

ym
(x
)

0
x2

..
.
Fm
ym (p0 )
y1
xn (x0 )
y1
xn (x0 )

..
.
y1
xn (x0 )

Then, from our previous differentiation with respect to xi , for each i =

1, 2, . . . , n, we can concisely express our systems of equations as a single
system
Jx,F (p0 ) + Jy,F (p0 )Jy (x0 ) = 0.
If the matrix Jy,F (p0 ) is invertible (also nonsingular), which is the case
if and only if its determinant is nonzero, and if all first partial derivatives
of F are continuous near p0 , then the equation F(p) = 0 implicitly defines
y1 , y2 , . . . , ym as a function of x1 , x2 , . . . , xn , and
Jy (x0 ) = [Jy,F (p0 )]1 Jx,F (p0 ),
where [Jy,F (p0 )]1 is the inverse of the matrix Jy,F (p0 ).
Example Let F : R4 R2 by defined by

 

F1 (x, y, u, v)
xu + y 2 v
F(x, y, u, v) =
=
.
F2 (x, y, u, v)
x2 v + yu + 1
Then the vector equation F(x, y, u, v) = 0 implicitly defines (u, v) as a
function of (x, y), provided that F satisifes the conditions of the Implicit

1.7. DIRECTIONAL DERIVATIVES AND THE GRADIENT VECTOR47

Function Theorem. We will compute the partial derivatives of u and v with
respect to x and y, at a point that satisfies this equation.
We have
"
# 

F1
F1
u 2yv
x
y
J(x,y),F (x, y, u, v) = F2 F2 =
,
2xv u
x
y
 F1 F1  

2
x
y
u
v
J(u,v),F (x, y, u, v) = F
=
.
F2
2
y x2
u
v
From the formula for the inverse of a 2 2 matrix,



1
1
d b
a b
,
=
c d
ad bc c a
we obtain

J(u,v) (x, y) =

ux uy
vx vy

= [J(u,v),F (x, y, u, v)]1 J(x,y),F (x, y, u, v)



 2
1
u 2yv
x y 2
= 3
2xv u
x
x y 3 y

 2
1
x u 2xy 2 v 2x2 yv y 2 u
=
.
2x2 v yu
xu 2y 2 v
y 3 x3
These partial derivatives can then be evaluated at any point (x, y, u, v) such
that F(x, y, u, v) = 0, such as (x, y, u, v) = (0, 1, 0, 1). Note that the
matrix J(u,v),F (x, y, u, v) is not invertible (that is, singular) if its determinant
x3 y 3 = 0; that is, if x = y. When this is the case, (u, v) can not be
implicitly defined as a function of (x, y). 2

1.7

Directional Derivatives and the Gradient Vector

Previously, we defined the gradient as the vector of all of the first partial
derivatives of a scalar-valued function of several variables. Now, we will
learn about how to use the gradient to measure the rate of change of the
function with respect to a change of its variables in any direction, as opposed
to a change in a single variable. This is extremely useful in applications in
which the minimum or maxmium value of a function is sought. We will
also learn how the gradient can be used to easily describe tangent planes to
level surfaces, thus providing an alternative to implicit differentiation or the
Chain Rule.

48

CHAPTER 1. PARTIAL DERIVATIVES

1.7.1

The Gradient Vector

Let f : D Rn R be a scalar-valued function of n variables x1 , x2 , . . . , xn .

Recall that the vector of its first partial derivatives,
f =

fx1

fx2

fxn

is called the gradient of f .

Example Let f (x, y, z) = e(x
f =

2 +y 2 )

2
2
2xe(x +y ) cos z

cos z. Then

2
2
2ye(x +y ) cos z

2
2
e(x +y ) sin z

Therefore, at the point (x0 , y0 , z0 ) = (1, 2, /3), the gradient is the vector
f (x0 , y0 , z0 ) =

fx (1, 2, /3) fy (1, 2, /3) fz (1, 2, /3)

*
+

3
=
e5 , 2e5 ,
e5 .
2

2
It should be noted that various differentiation rules from single-variable
calculus have direct generalizations to the gradient. Let u and v be differentiable functions defined on Rn . Then, we have:
Linearity:
(au + bv) = au + bv
where a and b are constants
Product Rule:
(uv) = uv + vu
Quotient Rule:

u
v

vu uv
v2

Power Rule:
un = nun1 u

1.7. DIRECTIONAL DERIVATIVES AND THE GRADIENT VECTOR49

1.7.2

Directional Derivatives

The components of the gradient vector f represent the instantaneous rates

of change of the function f with respect to any one of its independent variables. However, in many applications, it is useful to know how f changes as
its variables change along any path from a given point. To that end, given
f : D R2 R, and a unit vector u = ha, bi R2 , we define the directional
derivative of f at (x0 , y0 ) D in the direction of u to be
f (x0 + ah, y0 + bh) f (x0 , y0 )
.
h0
h

Du f (x0 , y0 ) = lim

When u = i = h1, 0i, then Du f = fx , and when u = j = h0, 1i, then

Du f = fy . For general u, Du f (x0 , y0 ) represents the instantaneous rate of
change of f as (x, y) change in the direction of u from the point (x0 , y0 ).
Because it is cumbersome to compute a directional derivative using the
definition directly, it is desirable to be able to relate the directional derivative
to the partial derivatives, which can be computed easily using differentiation
rules. We have
Du f (x0 , y0 ) =
=
=

f (x0 + ah, y0 + bh) f (x0 , y0 )

h
f (x0 + ah, y0 + bh) f (x0 , y0 + bh) + f (x0 , y0 + bh) f (x0 , y0 )
lim
h0
h
f (x0 + ah, y0 + bh) f (x0 , y0 + bh)
lim
+
h0
h
f (x0 , y0 + bh) f (x0 , y0 )
h
f (x0 + ah, y0 + bh) f (x0 , y0 + bh)
a+
lim
h0
ah
f (x0 , y0 + bh) f (x0 , y0 )
b
bh
fx (x0 , y0 )a + fy (x0 , y0 )b
lim

h0

= f (x0 , y0 ) u.
That is, the directional derivative in the direction of u is the dot product of
the gradient with u. It can be shown that this is the case for any number of
variables: given f : D Rn R, and a unit vector u Rn , the directional
derivative of f at x0 Rn in the direction of u is given by
Du f (x0 ) = f (x0 ) u.

50

CHAPTER 1. PARTIAL DERIVATIVES

Because the dot product a b can also be defined as
a b = kakkbk cos ,

where is the angle between a and b, the directional derivative can be used
to determine the direction along which f increases most rapidly, decreases
most rapidly, or does not change at all.
We first note that if is the angle between f (x0 ) and u, then
Du f (x0 ) = f (x0 ) u = kf (x0 )k cos .
Then we have the following:
When = 0, cos = 1, so Du f is maximized, and its value is
kf (x0 )k. In this case,
u=

f (x0 )
,
kf (x0 )k

and this is called the direction of steepest ascent.

When = , cos = 1, so Du f is minimized, and its value is
kf (x0 )k. In this case,
u=

f (x0 )
,
kf (x0 )k

and this is called the direction of steepest descent.

When = /2, cos = 0, so Du = 0. In this case, u is a unit vector
that is orthogonal (perpendicular) to f (x0 ). Since f is not changing
at all along this direction, it follows that u indicates the direction of
a level set of f , on which f (x) = f (x0 ).
The direction of steepest descent is of particular interest in applications in
which the goal is to find the minimum value of f . From a starting point
x0 , one can choose a new point x1 = x0 + u, where u = f (x0 ) is the
direction of steepest descent, by choosing so as to minimize f (x1 ). Then,
this process can be repeated using the direction of steepest descent at x1 ,
which is f (x1 ), to compute a new point x2 , and so on, until a minimum
is found. This process is called the method of steepest descent.
While not used very often in practice, it serves as a useful building
block for some of the most powerful methods that are used in practice for
minimizing functions.

1.7. DIRECTIONAL DERIVATIVES AND THE GRADIENT VECTOR51

Example Let f (x, y) = x2 y + y 3 , and let (x0 , y0 ) = (2, 2). Then
f (x, y) =

fx (x, y) fy (x, y)

2xy x2 + 3y 2

which yields f (x0 , y0 ) = hfx (2, 2), fy (2, 2)i = h8, 16i. It follows that
the direction of steepest ascent is


h8, 16i
h8, 16i
h8, 16i
f (2, 2)
1 2

=
u=
=p
=
= ,
.
kf (2, 2)k
320
8 5
5 5
(8)2 + 162

For this u, we have Du f (2, 2) = kf (2, 2)k = 8 5.

Furthermore, the direction of steepest descent is


1
2
u = ,
,
5
5

and along this direction, we have Du f (2, 2) = kf (2, 2)k = 8 5.

Finally, the directions along which f does not change at all are those that
are orthogonal to the directions of steepest ascent and descent,


2 1
u= ,
.
5 5
The level curve defined by the equation f (x, y) = f (2, 2) = 16 proceeds
along these directions from the point (2, 2). 2

1.7.3

Tangent Planes to Level Surfaces

Let F : D R3 R be a function of three variables x, y and z that

implicitly defines a surface through the equation F (x, y, z) = 0, and let
(x0 , y0 , z0 ) be a point on that surface. If F satisfies the conditions of the
Implicit Function Theorem at (x0 , y0 , z0 ), then the equation of the plane
that is tangent to the surface at this point can be obtained using the fact
that z is implicitly defined as a function of x and y near this point. It then
follows that the equation of the tangent plane is
z z0 = zx (x0 , y0 )(x x0 ) + zy (x0 , y0 )(y y0 ),
where, by the Chain Rule,
zx (x0 , y0 ) =

Fx (x0 , y0 , z0 )
,
Fz (x0 , y0 , z0 )

zy (x0 , y0 ) =

Fy (x0 , y0 , z0 )
.
Fz (x0 , y0 , z0 )

52

CHAPTER 1. PARTIAL DERIVATIVES

This is not possible if Fz (x0 , y0 , z0 ) = 0, because then the Implicit Function

Theorem does not apply.
It would be desirable to be able to obtain the equation of the tangent
plane even if Fz (x0 , y0 , z0 ) = 0, because the level surface still has a tangent
plane at that point even if z cannot be implicitly defined as a function of x
and y. To that end, we note that any direction u within the tangent plane
is parallel to the tangent vector of some curve that lies within the surface
and passes through (x0 , y0 , z0 ). Because F (x, y, z) = 0 on this surface, it
follows that Du F (x0 , y0 , z0 ) = 0. However, this implies that F (x0 , y0 , z0 )
must be orthogonal to u, in view of
Du F (x0 , y0 , z0 ) = F (x0 , y0 , z0 ) u = 0.
Since this is the case for any direction u within the tangent plane, we conclude that F (x0 , y0 , z0 ) is normal to the tangent plane, and therefore the
equation of this plane is
Fx (x0 , y0 , z0 )(x x0 ) + Fy (x0 , y0 , z0 )(y y0 ) + Fz (x0 , y0 , z0 )(z z0 ) = 0.
Note that this equation is equivalent to that obtained using the Chain Rule,
when Fz (x0 , y0 , z0 ) 6= 0.
The gradient not only provides the normal vector to the tangent plane,
but also the direction numbers of the normal line to the surface at (x0 , y0 , z0 ),
which is the line that passes through the surface at this point and is perpendicular to the tangent plane. The equation of this line, in parametric form,
is
x = x0 + tFx (x0 , y0 , z0 ),

y = y0 + tFy (x0 , y0 , z0 ),

z = z0 + tFz (x0 , y0 , z0 ).

Example Let F (x, y, z) = x2 + y 2 + z 2 2x 4y 4. Then the equation

F (x, y, z) = 0 defines a sphere of radius 3 centered at (1, 2, 0). At the point
(x0 , y0 , z0 ) = (3, 3, 2), we have
F (x0 , y0 , z0 ) =

Fx (x0 , y0 , z0 ) Fy (x0 , y0 , z0 ) Fz (x0 , y0 , z0 )



2x0 2 2y0 4 2z0
=

= h4, 2, 4i.
It follows that the equation of the plane that is tangent to the sphere at
(3, 3, 2) is
4(x x0 ) + 2(y y0 ) + 4(z z0 ) = 0,

1.7. DIRECTIONAL DERIVATIVES AND THE GRADIENT VECTOR53

and the equation of the normal line, in parametric form, is
x = x0 + tFx (x0 , y0 , z0 ) = 3 + 4t,
y = y0 + tFy (x0 , y0 , z0 ) = 3 + 2t,
z = z0 + tFz (x0 , y0 , z0 ) = 2 + 4t.
Equivalently, we can describe the normal line using its symmetric equations,
x3
y3
z2
=
=
.
4
2
4
2
When Fz 6= 0 at (x0 , y0 , z0 ), two linearly independent vectors within
a tangent plane can easily be obtained by first setting x x0 = 1 and
y y0 = 0 in the equation of the tangent plane to obtain the first vector,
and then x x0 = 0 and y y0 = 1 to obtain the second vector. For both
vectors, the initial point is the point of tangency (x0 , y0 , z0 ).
Recall the equation of the tangent plane, in its most general form,
Fx (x x0 ) + Fy (y y0 ) + Fz (z z0 ) = 0,
where, for brevity, we assume that all partial derivatives are evaluated at
(x0 , y0 , z0 ). Setting x x0 = 1 and y y0 = 0 yields
Fx + Fz (z z0 ) = 0,
which can be rearranged to obtain
z z0 =

Fx
= zx .
Fz

It follows that the vector h1, 0, zx i is a vector that lies within the tangent
plane.
Similarly, by setting x x0 = 0 and y y0 = 1, it follows that the vector
h0, 1, zy i is also within the tangent plane. Because these vectors cannot be
scalar multiples of each other, they must be linearly independent. It can
be verified that both of these vectors are orthogonal to the normal to the
tangent plane, F (x0 , y0 , z0 ). For example, we have
h1, 0, zx i hFx , Fy , Fz i = Fx + Fz zx = Fx + Fz (Fx /Fz ) = 0.

54

CHAPTER 1. PARTIAL DERIVATIVES

1.7.4

Tangent Lines to Level Curves

The preceding discussion about tangent planes to level surfaces can be scaled
down to two dimensions, yielding equations of tangent lines to level curves.
Consider a curve defined by the equation F (x, y) = 0. At a point (x0 , y0 )
on this curve, the tangent vector is pointing in a direction u such that
Du F (x0 , y0 ) = F (x0 , y0 ) u = 0,
since F is equal to the constant function 0 along the curve. That is,
F (x0 , y0 ) is orthogonal to the tangent line to the curve at (x0 , y0 ). Therefore, for any point (x, y) on the tangent line, we have
Fx (x0 , y0 )(x x0 ) + Fy (x0 , y0 )(y y0 ) = 0.
Thus we have an equation for the tangent line that is valid even when
dy/dx = Fx /Fy is not defined due to Fy (x0 , y0 ) = 0, which is analogous to
the most general form of the equation of a tangent plane that is valid even
when Fz (x0 , y0 , z0 ) = 0.
When Fy (x0 , y0 ) 6= 0, a vector contained within this tangent line can be
obtained by setting x x0 = 1 in the above equation of the tangent plane,
which yields the vector
hx x0 , y y0 i = h1, Fx /Fy i = h1, dy/dxi
with initial point (x0 , y0 ). It can then be verified that this vector is orthogonal to F (x0 , y0 ). We have
F h1, dy/dxi = hFx , Fy i h1, Fx /Fy i = Fx + Fy (Fx /Fy ) = 0,
where it is assumed that all derivatives are evaluated at (x0 , y0 ).

1.8

Maximum and Minimum Values

In single-variable calculus, one learns how to compute maximum and minimum values of a function. We first recall these methods, and then we will
learn how to generalize them to functions of several variables.
Let f : D Rn R. A local maximum of a function f is a point
a D such that f (x) f (a) for x near a. The value f (a) is called a local
maximum value. Similarly, f has a local minimum at a if f (x) f (a) for x
near a, and the value f (a) is called a local minimum value.

1.8. MAXIMUM AND MINIMUM VALUES

55

When a function of a single variable, f (x), has a local maximum or

minimum at x = a, then a must be a critical point of f , which means that
f 0 (c) = 0, or f 0 does not exist at a (which is the case if, for example, the
graph of f has a sharp corner at a). In general, if f is differentiable at a
point a, then in order for a to be a local maximum or minimum of f , the
rate of change of f , as its independent variables change in any direction u,
must be zero. That is, we must have f (a) u = 0 for any unit vector u.
The only way to ensure this is to require that f (a) = 0. Therefore, we say
that a is a critical point if f (a) = 0 or if any partial derivative of f does
not exist at a.
Once we have found the critical points of a function, we must determine
whether they correspond to local maxima or minima. In the single-variable
case, we can use the Second Derivative Test, which states that if a is a critical
point of f , and f 00 (a) > 0, then a is a local minimum, while if f 00 (a) < 0, a
is a local maximum, and if f 00 (a) = 0, the test is inconclusive.
This test is generalized to the multivariable case as follows: first, we
form the Hessian, which is the matrix of second partial derivatives at a. If
f is a function of n variables, then the Hessian is an n n matrix H, and
the entry in row i, column j of H is defined by
Hij =

2f
(a).
xi xj

Because mixed second partial derivatives are equal if they are continuous,
it follows that H is a symmetric matrix, meaning that Hij = Hji .
We can now state the Second Derivatives Test. If a is a critical point of
f , and the Hessian, H, is positive definite at a, then a is a local minimum
of f . The notion of a matrix being positive definite is the generalization to
matrices of the notion of a positive number. When a matrix H is symmetric
and positive definite, the following statements are all true:
xT Hx > 0, where x is a nonzero column vector of real numbers, and
xT is the transpose of x, which is a row vector. Note that xT Hx is
the same as x (Hx).
The eigenvalues of H are positive.
The determinant of H is positive.
The diagonal entries of H, Hii for i = 1, 2, . . . , n, are positive.
If a curve contained within the graph of f passes through a critical point a
with unit tangent vector u, then the concavity of the curve at a is measured

56

CHAPTER 1. PARTIAL DERIVATIVES

by uT Hu). If H is positive definite, then this concavity is positive for any

unit tangent vector u, which means that f has a local minimum at a.
On the other hand, if H is negative definite, then f has a local maximum
at a. This means that xT Hx < 0 for any nonzero real vector x, and that
the eigenvalues and diagonal entries of H are negative. However, the determinant is not necessarily negative. Because it is equal to the product of
the eigenvalues, the determinant is positive of n is even, and negative if n is
odd. When H is negative definite at a, then any curve contained within the
graph of f that passes through a is concave down at a. That is, uT Hu < 0,
where u is the unit tangent vector of the curve at a.
If H is indefinite, which is the case if it is neither positive definite nor
negative definite, and therefore has both positive and negative eigenvalues,
then we say that f has a saddle point at a. This means that the graph of f
crosses its tangent plane at a, and the term saddle point arises from the
fact that curves contained within the graph of f that pass through a can
be concave up along some directions and concave down along orthogonal
directions, thus giving the graph the appearance of a saddle.
Finally, if H is a singular matrix, meaning that one of its eigenvalues, and
therefore its determinant, is equal to zero, the test is inconclusive. Therefore, a could be a local minimum, local maximum, saddle point, or none of
the above. One must instead use other information about f , such as its directional derivatives, to determine if f has a maximum, minimum or saddle
point at a.
Example Let f : R2 R be defined by
f (x, y) = 6x2 + 4xy + 8y 2 x 3y.
We wish to find any local minima or maxima of this function. First, we
compute its gradient,
f =

12x + 4y 1 4x + 16y 3

To determine where f = 0, we must solve the system of linear equations

12x + 4y = 1,
4x + 16y = 3.
Using the second equation to obtain x = (3 16y)/4 and substituting this
into the first equation, we obtain y = 2/11 and x = 1/44. Since the solution
of this system is unique, it follows that this is the only critical point of f .

1.8. MAXIMUM AND MINIMUM VALUES

57

To determine whether this critical point corresponds to a maximum or

minimum, we must compute the Hessian H, whose entries are the second
partial derivatives of f at (1/44, 2/11). We have

 

fxx fxy
12 4
H=
=
.
fyx fyy
4 16
To determine whether this matrix is positive definite, we first compute its
determinant,
2
det(H) = fxx fyy fxy
= 12(16) 4(4) = 176.

Since the determinant, which is the product of Hs two eigenvalues, is positive, it follows that they must both be the same sign.
To determine that sign, we check the diagonal entries of H. If H is positive definite, then the diagonal entries fxx and fyy would both be positive.
Therefore, it is sufficient to check fxx ; because det(H) > 0, both diagonal
entries must the same sign. We have fxx = 12, so we conclude that H is
positive definite, and that the critical point (2/11, 1/44) is a local minimum
of f . 2
The preceding example describes how the Second Derivatives Test can be
performed for a function of two variables:
2 > 0, and f
If det(H) = fxx fyy fxy
xx > 0, then the critical point is a
minimum.

If det(H) > 0 and fxx < 0, then the critical point is a maximum.
If det(H) < 0, then the critical point is a saddle point.
If det(H) = 0, then the test is inconclusive.

1.8.1

Absolute Extrema

In many applications, it is desirable to know where a function assumes its

largest or smallest values, not just among nearby points, but within its entire
domain. We say that a function f : D Rn R has an absolute maximum
at a if f (a) f (x) for x D, and that f has an absolute minimum at a if
f (a) f (x) for x D.
In the single-variable case, it is known, by the Extreme Value Theorem,
that if f is continuous on a closed interval [a, b], then it has has an absolute
maximum and an absolute minimum on [a, b]. To find them, it is necessary

58

CHAPTER 1. PARTIAL DERIVATIVES

to check all critical points in [a, b], and the endpoints a and b, as the absolute
maximum and absolute minimum must each occur at one of these points.
The generalization of a closed interval to the multivariable case is the
notion of a compact set. Previously, we defined an open set, and a boundary
point. A closed set is a set that contains all of its boundary points. A
bounded set is a set that is contained entirely within a ball Dr (x0 ) for some
choice of r and x0 . Finally, a set is compact if it is closed and bounded.
We can now state the generalization of the Extreme Value Theorem to
the multivariable case. It states that a continuous function on a compact
set has an absolute minimum and an absolute maximum. Therefore, given
such a compact set D, to find the absolute maximum and minimum, it is
sufficient to check the critical points of f in D, and to find the extreme
(maximum and minimum) values of f on the boundary. The largest of all of
these values is the absolute maximum value, and the smallest is the absolute
minimum value.
It should be noted that in cases where D has a simple shape, such as
a rectangle, triangle or cube, it is possible to check boundary points by
characterizing them using one or more equations, using these equations to
eliminate a variable, and then substituting for the eliminated variable in f
to obtain a function of one less variable. Then, it is possible to find extreme
values on the boundary by solving a maximization or minimization problem
in one less dimension.
Example Consider the function f (x, y) = x2 + 3y 2 4x 6y. We will find
the absolute maximum and minimum values of this function on the triangle
with vertices (0, 0), (4, 0) and (0, 3).
First, we look for critical points. We have


f = 2x 4 6y 6 .
We see that there is only one critical point, at (x0 , y0 ) = (2, 1). Because
the triangle includes points that satisfy the inequalities x 0, y 0 and
y 3 3x/4, and the point (2, 1) satisfies all of these inequalities, we
conclude that this point lies within the triangle. It is therefore a candidate
for an absolute maximum or minimum.
We now check the boundary, by examining each edge of the triangle
individually. On the edge between (0, 0) and (0, 3), we have x = 0, which
yields f (0, y) = 3y 2 6y. We then have fy (0, y) = 6y 6, which has a
critical point at y = 1. Therefore, (0, 1) is also a candidate for an absolute
extremum. Similarly, along the edge between (0, 0) and (4, 0), we have y = 0,
which yields f (x, 0) = x2 4x. We then have fx (x, 0) = 2x 4, which has

1.8. MAXIMUM AND MINIMUM VALUES

59

a critical point at x = 2. Therefore, (2, 0) is a candidate for an absolute

extremum.
We then check the edge between (0, 3) and (4, 0), along which y = 3
3x/4. Substituting this into f (x, y) yields the function


3x
43
g(x) = f x, 3
= x2 + 9 13x.
4
16
To determine the critical points of this function, we solve g 0 (x) = 0, which
yields x = 104/43. Since y = 33x/4 along this edge, the point (104/43, 51/43)
is a candidate for an absolute extremum.
Finally, we must include the vertices of the triangle, because they too
are boundary points of the triangle, as well as boundary points of the edges
along which we attempted to find extrema of single-variable functions. In
all, we have seven candidates: the critical point of f , (2, 1), the three critical
points found along the edges, (0, 1), (2, 0) and (104/43, 51/43), and the three
vertices, (0, 0), (4, 0) and (0, 3). Evaluating f (x, y) at all of these points, we
obtain
x
2
0
2
104/43
0
4
0

y
1
1
0
51/43
0
0
3

f(x,y)
7
3
4
289/43
0
0
9

We conclude that the absolute minimum is at (2, 1), and the absolute maximum is at (0, 3). The function is shown on Figure 1.5. 2

1.8.2

Relation to Taylor Series

Previously, we learned that when seeking a local minimum or maximum

of a function of variables, the Second Derivative Test from single-variable
calculus, in which the sign of the second derivative indicated whether a
local extremum was a maximum or minimum, generalizes to the Second
Derivatives Test, which indicates that a local extremum x0 is a minimum if
the Hessian, the matrix of second partial derivatives, is positive definite at
x0 .
We will now use Taylor series to explain why this test is effective. Recall
that in single-variable calculus, Taylors Theorem states that a function f (x)

60

CHAPTER 1. PARTIAL DERIVATIVES

Figure 1.5: The function f (x, y) = x2 + 3y 2 4x 6y on the triangle with

vertices (0, 0), (4, 0) and (0, 3).

with at least three continuous derivatives at x0 can be written as

1
1
f (x) = f (x0 ) + f 0 (x0 )(x x0 ) + f 00 (x0 )(x x0 )2 + f 000 ()(x x0 )3 ,
2
6
where is between x and x0 . In the multivariable case, Taylors Theorem
states that if f : D Rn R has continuous third partial derivatives at
x0 D, then
f (x) = f (x0 ) + f (x0 ) (x x0 ) + (x x0 ) Hf (x0 )(x x0 ) + R2 (x0 , x),
where Hf (x0 ) is the Hessian, the matrix of second partial derivatives at x0 ,

1.8. MAXIMUM AND MINIMUM VALUES

61

defined by
2f
(x0 )
x21
2
f
x2 x1 (x0 )

Hf (x0 ) =

..
.
2f
xn x1 (x0 )

2f
x1 x2 (x0 )
2f
(x0 )
x22

2f
x1 xn (x0 )
2f
x2 xn (x0 )

2f
xn x2 (x0 )

..
.
2f
(x0 )
x2

and R2 (x0 , x) is the Taylor remainder, which satisfies

lim

xx0
(0)

(0)

R2 (x0 , x)
= 0.
kx x0 k2

(0)

If we let x0 = (x1 , x2 , . . . , xn ), then Taylors Theorem can be rewritten

using summations:
n
n
X
X
f
2f
(0)
(0)
(0)
f (x) = f (x0 ) +
(x0 )(xi xi ) +
(x0 )(xi xi )(xj xj ) +
xi
xi xj
i=1

i,j=1

R2 (x0 , x).
Example Let f (x, y) = x2 y 3 + xy 4 , and let (x0 , y0 ) = (1, 2). Then, from
partial differentiation of f , we obtain its gradient

 

f = fx fy = 2xy 3 + y 4 3x2 y 2 + 4xy 3 ,
and its Hessian,

Hf (x, y) =

fxx fxy
fyx fyy


=

2y 3
6xy 2 + 4y 3
6xy 2 + 4y 3 6x2 y + 12xy 2


.

Therefore
f (1, 2) =

0 20


,

Hf (1, 2) =

16 8
8 36


,

and the Taylor expansion of f around (1, 2) is

f (x, y) = f (x0 , y0 ) + f (x0 , y0 ) hx x0 , y y0 i +
1
hx x0 , y y0 i Hf (x0 , y0 )hx x0 , y y0 i + R2 ((x0 , y0 ), (x, y))
2



 x1
= 8 + 0 20
+
y+2

62

CHAPTER 1. PARTIAL DERIVATIVES



16 8
8 36
R2 ((1, 2), (x, y))
hx 1, y + 2i



x1
y+2


+

= 8 20(y + 2) 16(x 1)2 16(x 1)(y + 2) + 36(y + 2)2 +

R2 ((1, 2), (x, y)).
The first three terms represent an approximation of f (x, y) by a quadratic
function that is valid near the point (1, 2). 2
Now, suppose that x0 is a critical point of x. If this point is to be
a local minimum, then we must have f (x) f (x0 ) for x near x0 . Since
f (x0 ) = 0, it follows that we must have
(x x0 ) [Hf (x0 )(x x0 )] 0.
However, if the Hessian Hf (x0 ) is a positive definite matrix, then, by definition, this expression is actually strictly greater than zero. Therefore, we are
assured that x0 is a local minimum. In fact, x0 is a strict local minimum,
since we can conclude that f (x) > f (x0 ) for all x sufficiently near x0 .

1.8.3

Principal Minors

As discussed previously, there are various properties possessed by symmetric

positive definite matrices. One other, which provides a relatively straightforward method of checking whether a matrix is positive definite, is to check
whether the determinants of its principal submatrices, known as principal
minors, are positive. Given an n n matrix A, its principal submatrices are
the submatrices consisting of its first k rows and columns, for k = 1, 2, . . . , n.
Note that checking these determinants, the principal minors, is equivalent
to the test that we have previously described for determining whether a 22
matrix is positive definite.
Example Let f (x, y, z) = x2 + y 2 + z 2 + xy. To find any local maxima or
minima of this function, we compute its gradient, which is


f (x, y, z) = 2x + y 2y + x 2z .
It follows that the only critical point is at (x0 , y0 , z0 ) = (0, 0, 0). To perform
the Second Derivatives Test, we compute the Hessian of f , which is

fxx fxy fxz

2 1 0
Hf (x, y, z) = fyx fyy fyz = 1 2 0 .
fzx fzy fzz
0 0 2

1.9. CONSTRAINED OPTIMIZATION

63

To determine whether this matrix is positive definite, we can compute the

determinants of the principal submatrices of Hf (0, 0, 0), which are
[Hf (0, 0, 0)]11 = 2,

2
[Hf (0, 0, 0)]1:2,1:2 =
1

2
[Hf (0, 0, 0)]1:3,1:3 = 1
0

1
2


,

1 0
2 0 .
0 2

For the principal minors, we have

det([Hf (0, 0, 0)]11 ) = 2,

det([Hf (0, 0, 0)]1:2,1:2 ) = 2(2) 1(1) = 3,

det([Hf (0, 0, 0)]1:3,1:3 ) = 2 det([Hf (0, 0, 0)]1:2,1:2 ) = 6.

Since all of the principal minors are positive, we conclude that Hf (0, 0, 0) is
positive definite, and therefore the critical point is a minimum of f . 2

1.9

Constrained Optimization

Now, we consider the problem of finding the maximum or minimum value of

a function f (x), except that the independent variables x = (x1 , x2 , . . . , xn )
are subject to one or more constraints. These constraints prevent us from
using the standard approach for finding extrema, but the ideas behind the
standard approach are still useful for developing an approach to the constrained problem.
We assume that the constraints are equations of the form
gi (x) = 0,

i = 1, 2, . . . , m

for given functions gi (x). That is, we may only consider x = (x1 , x2 , . . . , xn )
that belong to the intersection of the hypersurfaces (surfaces, when n = 3,
or curves, when n = 2) defined by the gi , when computing a maximum
or minimum value of f . For conciseness, we rewrite these constraints as a
vector equation g(x) = 0, where g : Rn Rm is a vector-valued function
with component functions gi , for i = 1, 2, . . . , m.
By Taylors theorem, we have, for x0 Rn at which g is differentiable,
g(x) = g(x0 ) + Jg (x0 )(x x0 ) + R1 (x0 , x),

64

CHAPTER 1. PARTIAL DERIVATIVES

where Jg (x0 ) is the Jacobian matrix of g at x0 , consisting of the first partial

derivatives of the gi evaluated at x0 , and R1 (x0 , x) is the Taylor remainder,
which satisfies
R1 (x0 , x)
lim
= 0.
xx0 kx x0 k
It follows that if u is a vector belonging to all of the tangent spaces of the
hypersurfaces defined by the gi , then, because each gi must remain constant
as x deviates from x0 in the direction of u, we must have Jg (x0 )u = 0. In
other words, gi (x0 ) u = 0 for i = 1, 2, . . . , m.
Now, suppose that x0 is a local minimum of f (x), subject to the constraints g(x0 ) = 0. Then, x0 may not necessarily be a critical point of f , but
f may not change along any direction from x0 that satisfies the constraints.
Therefore, we must have f (x0 ) u = 0 for any vector u in the intersection
of tangent spaces, at x0 , of the hypersurfaces defined by the constraints.
It follows that if u is any such vector in this tangent plane, and there
exist constants 1 , 2 , . . . , m such that
f (x0 ) = 1 g1 (x0 ) + 2 g2 (x0 ) + + m gm (x0 ),
then the requirement f (x0 ) u = 0 follows directly from the fact that
gi (x0 ) u = 0, and therefore x0 must be a constrained critical point of f .
The constants 1 , 2 , . . . , m are called Lagrange multipliers.
Example When m = 1; that is, when there is only one constraint, the
problem of finding a constrained minimum or maximum reduces to finding
a point x0 in the domain of f such that
f (x0 ) = g(x0 ),
for a single Lagrange multiplier .
Let f (x, y) = 4x2 + 9y 2 . The minimum value of this function is at 0,
which is attained at x = y = 0, but we wish to find the minimum of f (x, y)
subject to the constraint x2 + y 2 2x 2y = 2. That is, we must have
g(x, y) = 0 where g(x, y) = x2 + y 2 2x 2y 2. To find any points that
are candidates for the constrained minimum, we compute the gradients of f
and g, which are


f = 8x 18y ,


g = 2x 2 2y 2 .
In order for the equation f (x, y) = g(x, y) to be satisfied, we must
have, for some choice of , x and y,
8x = (2x 2),

18y = (2y 2).

1.9. CONSTRAINED OPTIMIZATION

65

From these equations, we obtain

x=

,
4

y=

.
9

Substituting these into the constraint x2 + y 2 2x 2y 2 = 0 yields the

fourth-degree equation
44 1043 + 8672 2808 + 2592 = 0.
This equation has two real solutions,
3
1 = ,
2

2 13.6.

Substituting these values into the above equations for x and y yield the
critical points
3
1
3
x1 = , y1 = , 1 = ,
5
5
2
x2 1.416626,

y2 2.956124,

2 13.6.

Substituting the x and y values into f (x, y) yields the minimum value of 9/5
at (x1 , y1 ) and the maximum value of approximately 86.675 at (x2 , y2 ). 2
Example Let f (x, y, z) = x + y + z. We wish to find the extremea of this
function subject to the constraints x2 + y 2 = 1 and 2x + z = 1. That is, we
must have g1 (x, y, z) = g2 (x, y, z) = 0, where g1 (x, y, z) = x2 + y 2 1 and
g2 (x, y, z) = 2x + z 1. We must find 1 and 2 such that
f = 1 g1 + 2 g2 ,
or


1 1 1

= 1

2x 2y 0

+ 2

2 0 1

This equation, together with the constraints, yields the system of equations
1 = 2x1 + 22
1 = 2y1
1 = 2
1 = x2 + y 2
1 = 2x + z.

66

CHAPTER 1. PARTIAL DERIVATIVES

From the third equation, 2 = 1, which, by the first equation, yields 2x1 =
1. It follows from the second equation that x =y. This,
in conjunc2,
1/
2) or (x, y) =
tion with
the
fourth
equation,
yields
(x,
y)
=
(1/

(1/ 2, 1/ 2). From the fifth equation, we obtain the two critical points





1
1
1 1
(x1 , y1 , z1 ) = , , 1 2 , (x2 , y2 , y2 ) = , , 1 + 2 .
2
2
2 2

Substituting
these points into f yields f (x1 , y1 , z1 ) = 1 2 and f (x2 , y2 , z2 ) =

1+ 2, so we conclude that (x1 , y1 , z1 ) is a local minimum of f and (x2 , y2 , z2 )

is a local maximum of f , subject to the constraints g1 (x, y, z) = g2 (x, y, z) =
0. 2
The method of Lagrange multipliers can be used in conjunction with the
method of finding unconstrained local maxima and minima in order to find
the absolute maximum and minimum of a function on a compact (closed
and bounded) set. The basic idea is as follows:
Find the (unconstrained) critical points of the function, and exclude
those that do not belong to the interior of the set.
Use the method of Lagrange multipliers to find the constrained critical points that lie on the boundary of the set, using equations that
characterize the boundary points as constraints. Also, include corners
of the boundary, as they represent critical points due to the function,
restricted to the boundary, not being differentiable.
Evaluate the function at all of the constrained and unconstrained critical points. The largest value is the absolute maximum value on the
set, and the smallest value is the absolute minimum value on the set.
From a linear algebra point of view, f (x0 ) must be orthogonal to any
vector u in the null space of Jg (x0 ) (that is, the set consisting of any vector
v such that Jg (x0 )v = 0), and therefore it must lie in the range of Jg (x0 )T ,
the transpose of Jg (x0 ). That is, f (x0 ) = Jg (x0 )T u for some vector u,
meaning that f (x0 ) must be a linear combination of the rows of Jg (x0 ) (the
columns of Jg (x0 )T ), which are the gradients of the component functions of
g at x0 .
Another way to view the method of Lagrange multipliers is as a modified
unconstrained optimization problem. If we define the function h(x, ) by
h(x, ) = f (x) g(x) = f (x)

m
X
i=1

i gi (x),

1.9. CONSTRAINED OPTIMIZATION

67

then we can find constrained extrema of f by finding unconstrained extrema

of h, for


h(x, ) = f (x) Jg (x) g(x) .
Because all components of the gradient must be equal to zero at a critical point (when the gradient exists), the constraints must be satisfied at a
critical point of h, and f must be a linear combination of the gi , so f
is only changing along directions that violate the constraints. Therefore, a
critical point is a candidate for a constrained maximum or minimum. By
the Second Derivatives Test, we can then use the Hessian of h to determine
if any constrained extremum is a maximum or minimum.

68

CHAPTER 1. PARTIAL DERIVATIVES

1.10

Appendix: Linear Algebra Concepts

1.10.1

Matrix Multiplication

As we work with Jacobian matrices for vector-valued functions of several

variables, matrix multiplication is a highly relevant operation in multivariable calculus. We have previously defined the product of an m n matrix
A (that is, A has m rows and n columns) and an n p matrix B as the
m p matrix C = AB, where the entry in row i and column j of C is the
dot product of row i of A and column j of B. This can be written using
sigma notation as
cij =

n
X

aik bkj ,

i = 1, 2, . . . , m,

j = 1, 2, . . . , p.

k=1

Note that the number of columns in A must equal the number of rows in B,
or the product AB is undefined. Furthermore, in general, even if A and B
can be multiplied in either order (that is, if they are square matrices of the
same size), AB does not necessarily equal BA. In the special case where the
matrix B is actually a column vector x with n components (that is, p = 1),
it is useful to be able to recognize the summation
yi =

n
X

aij xj

j=1

as the formula for the ith component of the vector y = Ax.

Example Let A a 3 2 matrix, and B be a 2 2 matrix, whose entries are
given by



1 2
7
8
A = 3 4 , B =
.
9 10
5 6
Then, because the number of columns in A is equal to the number of rows
in B, the product C = AB is defined, and equal to the 3 2

1(7) + (2)9 1(8) + (2)(10)

25 28
C = (3)(7) + 4(9) (3)(8) + 4(10) = 57 64 .
5(7) + (6)9 5(8) + (6)(10)
89 100
Because the number of columns in B is not the same as the number of rows
in A, it does not make sense to compute the product BA. 2

1.10. APPENDIX: LINEAR ALGEBRA CONCEPTS

69

In multivariable calculus, matrix multiplication most commonly arises

when applying the Chain Rule, because the Jacobian matrix of the composition f g at point x0 in the domain of g is the product of the Jacobian
matrix of f , evaluated at g(x0 ), and the Jacobian matrix of g evaluated at
x0 . It follows that the Chain Rule only makes sense when composing functions f and g such that the number of dependent variables of g (that is, the
number of rows in its Jacobian matrix) equals the number of independent
variables of f (that is, the number of columns in its Jacobian matrix).
Matrix multiplication also arises in Taylor series expansions of multivariable functions, because if f : D Rn R, then the Taylor expansion
of f around x0 D involves the dot product of f (x0 ) with the vector
x x0 , which is a multiplication of a 1 n matrix with an n 1 matrix
to produce a scalar (by convention, the gradient is written as a row vector,
while points are written as column vectors). Also, such an expansion involves the dot product of x x0 with the product of the Hessian matrix, the
matrix of second partial derivatives at x0 , and the vector x x0 . Finally, if
g : U Rn Rm is a vector-valued function of n variables, then the second
term in its Taylor expansion around x0 U is the product of the Jacobian
matrix of g at x0 and the vector x x0 .

1.10.2

Eigenvalues

Previously, it was mentioned that the eigenvalues of a matrix that is both

symmetric, and positive definite, are positive. A scalar , which can be real
or complex, is an eigenvalue of an n n matrix A (that is, A has n rows
and n columns) if there exists a nonzero vector x such that
Ax = x.
That is, matrix-vector multiplication of A and x reduces to a simple scaling
of x by . The vector x is called an eigenvector of A corresponding to .
The eigenvalues of A are roots of the characteristic polynomial det(A
I), which is a polynomial of degree n in the variable . Therefore, an n n
matrix A has n eigenvalues, which may repeat. Although the eigenvalues
of a matrix may be real or complex, even when the matrix is real, the
eigenvalues of a real, symmetric matrix, such as the Hessian of any function
with continuous second partial derivatives, are real.
For a general matrix A, det(A), the determinant of A, is the product
of all of the eigenvalues of A. The trace of A, denoted by tr(A), which is
defined to be the sum of the diagonal entries of A, is also the sum of the
eigenvalues of A. It follows that when A is a 2 2 symmetric matrix, the

70

CHAPTER 1. PARTIAL DERIVATIVES

determinant and trace can be used to easily confirm that the eigenvalues of
A are either both positive, both negative, or of opposite signs. This is the
basis for the Second Derivatives Test for functions of two variables.
Example Let A be a symmetric 2 2 matrix defined by


4 6
A=
.
6 10
Then
tr(A) = 4 + 10 = 14,

det(A) = 4(10) (6)(6) = 4.

It follows that the product and the sum of As two eigenvalues are both
positive. Because A is symmetric, its eigenvalues are also real. Therefore,
they must both also be positive, and we can conclude that A is positive
definite.
To actually compute the eigenvalues, we can compute its characteristic
polynomial, which is


4
6
det(A I) = det
6 10
= (4 )(10 ) (6)(6)
= 2 14 + 4.
Note that
det(A I) = 2 tr(A) + det(A),
which is true for 2 2 matrices in general. To compute the eigenvalues,
we use the quadratic formula to compute the roots of this polynomial, and
obtain
p

14 142 4(4)(1)
=
= 7 3 5 13.708, 0.292.
2(1)
If A represented the Hessian of a function f (x, y) at a point (x0 , y0 ), and
f (x0 , y0 ) = 0, then f would have a local minimum at (x0 , y0 ). 2

1.10.3

The Transpose, Inner Product and Null Space

The dot product of two vectors u and v, denoted by uv, can also be written
as uT v, where u and v are both column vectors, and uT is the transpose of
u, which converts u into a row vector. In general, the transpose of a matrix
A is the matrix AT whose entries are defined by [AT ]ij = [A]ji . That is, in
the transpose, the sense of rows and columns are reversed. The dot product

1.10. APPENDIX: LINEAR ALGEBRA CONCEPTS

71

is also known as an inner product; the outer product of two column vectors
u and v is uvT , which is a matrix, whereas the inner product is a scalar.
Given an m n matrix A, the null space of A is the set N (A) of all
n-vectors such that if x N (A), then Ax = 0. If x is such a vector, then
for any m-vector v, vT (Ax) = vT 0 = 0. However, because of two properites
of the transpose, (AT )T = A and (AB)T = B T AT , this inner product can
be rewritten as vT Ax = vT (AT )T x = (AT v)T x. It follows that any vector
in N (A) is orthogonal to any vector in the range of AT , denoted by R(AT ),
which is the set of all n-vectors of the form AT v, where v is an m-vector.
This is the basis for the condition f = JgT in the method of Lagrange
multipliers when there are multiple constraints.
Example Let

1 2 4
A = 1 3 6 .
1 5 10
Then

1
1
1
AT = 2 3 5 .
4 6 10

The null space of A, N (A), consists of all vectors that are multiples of the
vector

0

v = 2 ,
1
as it can be verified by matrix-vector multiplication that Av = 0. Now, if we
let w be any vector in R3 , and we compute u = AT w, then v u = vT u = 0,
because
vT u = vT AT w = (Av)T w = 0T w = 0.
For example, it can be confirmed directly that v is orthogonal to any of the
columns of AT . 2

72

CHAPTER 1. PARTIAL DERIVATIVES

Chapter 2

Multiple Integrals
2.1

Double Integrals over Rectangles

In single-variable calculus, the definite integral of a function f (x) over an

interval [a, b] was defined to be
n
X

f (x) dx = lim

f (xi )x,

i=1

where x = (b a)/n, and, for each i, xi1 xi xi , where xi = a + ix.

The purpose of the definite integral is to compute the area of a region
with a curved boundary, using the formula for the area of a rectangle. The
summation used to define the integral is the sum of the areas of n rectangles,
each with width x, and height f (xi ), for i = 1, 2, . . . , n. By taking the
limit as n, the number of rectangles, tends to infinity, we obtain the sum of
the areas of infinitely many rectangles of infinitely small width. We define
the area of the region bounded by the lines x = a, y = 0, x = b, and the
curve y = f (x), to be this limit, if it exists.
Unfortunately, it is too tedious to compute definite integrals using this
definition. However, if we define the function F (x) as the definite integral
x

Z
F (x) =

f (s) ds,
a

then we have
1
F (x) = lim
h0 h
0

Z

x+h

Z
f (s) ds

73


Z
1 x+h
f (s) ds =
f (s) ds.
h x

74

CHAPTER 2. MULTIPLE INTEGRALS

Intuitively, as h 0, this expression converges to the area of a rectangle of

width h and height f (x), divided by the width, which is simply the height,
f (x). That is, F 0 (x) = f (x). This leads to the Fundamental Theorem of
Calculus, which states that
Z

f (x) dx = F (b) F (a),

a

where F is an antiderivative of f ; that is, F 0 = f . Therefore, definite

integrals are typically evaluated by attempting to undo the differentiation
process to find an antiderivative of the integrand f (x), and then evaluating
this antiderivative at a and b, the limits of the integral.
Now, let f (x, y) be a function of two variables. We consider the problem
of computing the volume of the solid in 3-D space bounded by the surface
z = f (x, y), and the planes x = a, x = b, y = c, y = d, and z = 0, where
a, b, c and d are constants. As before, we divide the interval [a, b] into n
subintervals of width x = (b a)/n, and we similarly divide the interval
[c, d] into m subintervals of width y = (d c)/m. For convenience, we also
define xi = a + ix, and yj = c + jy.
Then, we can approximate the volume V of this solid by the sum of the
volumes of mn boxes. The base of each box is a rectangle with dimensions
x and y, and the height is given by f (xi , yj ), where, for each i and j,
xi1 xi xi and yj1 yj yj . That is,
V

n X
m
X

f (xi , yj ) y x.

i=1 j=1

We then obtain the exact volume of this solid by letting the number of
subintervals, n, tend to infinity. The result is the double integral of f (x, y)
over the rectangle R = {(x, y) | a x b, c y d}, which is also written
as R = [a, b] [c, d]. The double integral is defined to be
Z Z
V =

f (x, y) dA =
R

lim

m,n

n X
m
X

f (xi , yj ) y x,

i=1 j=1

which is equal to the volume of the given solid. The dA corresponds to the
quantity A = xy, and emphasizes the fact that the integral is defined
to be the limit of the sum of volumes of boxes, each with a base of area A.
To evaluate double integrals of this form, we can proceed as in the singlevariable case, by noting that if f (x0 , y), a function of y, is integrable on [c, d]

2.1. DOUBLE INTEGRALS OVER RECTANGLES

75

for each x0 [a, b], then we have

Z Z
n X
m
X
f (x, y) dA =
lim
f (xi , yj ) y x
m,n

i=1 j=1

n
X

lim

lim

i=1

n Z d
X

lim

f (xi , yj )y x

j=1

f (xi , y) dy


x

i=1
bZ d

m
X

f (x, y) dy dx.
a

Similarly, if f (x, y0 ), a function of x, is integrable on [a, b] for each y0 [c, d],

we also have
Z Z
Z Z
d

f (x, y) dA =
R

f (x, y) dy dx.
c

This result is known as Fubinis Theorem, which states that a double integral of a function f (x, y) can be evaluated as two iterated single integrals,
provided that f is integrable as a function of either variable when the other
variable is held fixed. This is guaranteed if, for instance, f (x, y) is continuous
on the entire rectangle R.
That is, we can evaluate a double integral by performing partial integration with respect to either variable, x or y, which entails applying the
Fundamental Theorem of Calculus to integrate f (x, y) with respect to only
that variable, while treating the other variable as a constant. The result will
be a function of only the other variable, to which the Fundamental Theorem
of Calculus can be applied a second time to complete the evaluation of the
double integral.
Example Let R = [0, 1] [0, 2], and let f (x, y) = x2 y + xy 3 . We will use
Fubinis Theorem to evaluate
Z Z
f (x, y) dy dx.
R

We have
Z Z

1Z 2

x2 y + xy 3 dy dx
0
0

Z 1 Z 2
2
3
x y + xy dy dx
=

f (x, y) dy dx =
R

76

CHAPTER 2. MULTIPLE INTEGRALS

1 Z 2

2
3

xy dy dx
x y dy +
0

Z 2
Z 1 Z 2
y 3 dy dx
y dy + x
x2
=
0
0
0
2
2 #
Z 1"
2
4


y 
y
x2
=
+ x  dx

2 0
4 0
0
Z 1
2x2 + 4x dx
=
=


=
=

1

2x3
+ 2x2 
3
0

8
.
3

2
In view of Fubinis Theorem, a double integral is often written as
Z Z
Z Z
Z Z
f (x, y) dA =
f (x, y) dy dx =
f (x, y) dx dy.
R

Example We wish to compute the volume V of the solid bounded by the

planes x = 1, x = 4, y = 0, y = 2, z = 0, and x + y + z = 8. The
plane that defines the top of this solid is also the graph of the function
z = f (x, y) = 8 x y. It follows that the volume of the solid is given by
the double integral
Z Z
V =
8 x y dA, R = [1, 4] [0, 2].
R

Using Fubinis Theorem, we obtain

Z Z
V =
8 x y dA
R

Z 4 Z 2
=
8 x y dy dx
1

0
4

=
1

Z
=

2
y 2 
8y xy
dx
2 0

14 2x dx
4
(14x x2 )
1

2.2. DOUBLE INTEGRALS OVER MORE GENERAL REGIONS

77

= (56 16) (14 1)

= 27.
2
We conclude by noting some useful properties of the double integral, that
are direct generalizations of corresponding properties for single integrals:
Linearity: If f (x, y) and g(x, y) are both integrable over R, then
Z Z
Z Z
Z Z
g(x, y) dA
f (x, y) dA +
[f (x, y) + g(x, y)] dA =
R

Homogeneity: If c is a constant, then

Z Z
Z Z
cf (x, y) dA = c
f (x, y) dA
R

Monotonicity: If f (x, y) 0 on R, then

Z Z
f (x, y) dA 0.
R

Additivity: If R1 and R2 are disjoint rectangles and Q = R1 R2 is a

rectangle, then
Z Z
Z Z
Z Z
f (x, y) dA =
f (x, y) dA +
f (x, y) dA.
Q

2.2

R1

R2

Double Integrals over More General Regions

We have learned how to integrate a function f (x, y) of two variables over a

rectangle R. However, it is important to be able to integrate such functions
over more general regions, in order to be able to compute the volume of a
wider variety of solids.
To that end, given a region D R2 , contained within a rectangle R, we
define the double integral of f (x, y) over D by
Z Z
Z Z
f (x, y) dA =
F (x, y) dA
D

where


F (x, y) =

f (x, y) (x, y) D
.
0
(x, y) R,
/D

78

CHAPTER 2. MULTIPLE INTEGRALS

It is possible to use Fubinis Theorem to compute integrals over certain

types of general regions. We say that a region D is of type I if it lies between
the graphs of two continuous functions of x, and is also bounded by two
vertical lines. Specifically,
D = {(x, y) | a x b, g1 (x) y g2 (x)}.
To integrate f (x, y) over such a region, we can apply Fubinis Theorem. We
let R = [a, b] [c, d] be a rectangle that contains D. Then we have
Z Z
Z Z
F (x, y) dA
f (x, y) dA =
R
D
Z bZ d
=
F (x, y) dy dx
a

Z bZ

g2 (x)

F (x, y) dy dx
a

g1 (x)
b Z g2 (x)

f (x, y) dy dx.
a

g1 (x)

This is valid because F (x, y) = 0 when y < g1 (x) or y > g2 (x), because
in these cases, (x, y) lies outside of D. The resulting iterated integral can
be evaluated in the same way as iterated integrals over rectangles; the only
difference is that when the limits of the inner integral are substituted for y
in the antiderivative of f (x, y) with respect to y, the limits are functions of
x, rather than constants.
A similar approach can be applied to a region of type II, which is bounded
on the left and right by continuous functions of y, and bounded above and
below by vertical lines. Specifically, D is a region of type II if
D = {(x, y)|h1 (y) x h2 (y),
Using Fubinis Theorem, we obtain
Z Z
Z
f (x, y) dA =
D

c y d}.

d Z h2 (y)

f (x, y) dx dy.
h1 (y)

Example We wish to compute the volume of the solid under the plane
x + y + z = 8, and bounded by the surfaces y = x and y = x2 . These
surfaces intersect along the lines x = 0, y = 0 and x = 1, y = 1. It follows
that the volume V of the solid is given by the double integral
Z 1Z x
8 x y dy dx.
0

x2

2.2. DOUBLE INTEGRALS OVER MORE GENERAL REGIONS

79

Note that g2 (x) = x is the upper limit of integration, because x2 x when

0 x 1. We have
Z 1Z x
8 x y dy dx
V =
0
x2
x
Z 1
y 2 
=
8y xy
dx
2 x2
0
 

Z 1
x4
x2
2
3
2
8x x
dx
=
8x x
2
2
0
Z 1 4
x
19x2
=
+ x3
+ 8x dx
2
0 2
1
 5

x4 19x3
x
2 
+

+ 4x 
=
10
4
6
0

=
=

1 19
1
+
+4
10 4
6
71
.
60

2
Note that it is sometimes necessary to determine the intersections of surfaces
that define a solid, in order to obtain the limits of integration.
To compute the volume of a solid that is bounded above and below
(along the z-direction) by two different surfaces, we can add the volume of
the solid bounded by the top surface and the plane z = 0 to the volume of
the solid bounded above by z = 0 and below by the lower surface, which
is equivalent to subtracting the volume of the solid bounded above by the
lower surface and below by z = 0.
Example We will compute the volume V of the solid in the first octant
bounded by the planes z = 10 + x + y, z = 2 x y, and x = 0, as well as
the surfaces
y = sin x and y = cos x. As these surfaces intersect along the
line y = 2/2, x = /4, this volume is given by the double integral
Z
V

/4 Z cos x

(10 + x + y) (2 x y) dy dx

=
0

sin x
/4 Z cos x

8 + 2x + 2y dy dx
0

Z
=
0

sin x
/4

cos x
8y + 2xy + y 2 sin x dx

80

CHAPTER 2. MULTIPLE INTEGRALS

/4

(2x + 8)(cos x sin x) + cos2 x sin2 x dx

=
0
/4

(2x + 8)(cos x sin x) + cos 2x dx

=
0

/4

1
2x sin x + 2x cos x + 6 sin x + 10 cos x + sin 2x 
2
0

19
2
+8 2 .
2
2


=
=

The final anti-differentiation requires integration by parts,

Z
Z
u dv = uv v du,
with u = x and dv = (cos x sin x) dx. The function z = 10 + x + y is the
top plane because for 0 x /4, sin x y cosx, 10+x+y 2xy.
2
By setting the integrand f (x, y) 1 on a region D, and integrating over
D, we can obtain A(D), the area of D.
Example We will compute the area of a half-circle by integrating f (x, y) 1
over a region D
that is bounded by the planes z = 0, z = 1, and y = 0, and
the surface y = 1 x2 . This surface intersects the plane y = 0 along the
lines y = 0, x = 1 and y = 0, x = 1. Therefore the area is given by
Z

A(D) =

1x2

1 dy dx =
1

y|0 1x dx =

p
1 x2 dx.

To evaluate this integral, we use the trigonometric substitution x = sin ,

for which dx = cos d, which yields
Z

/2

A(D) =

/2

cos d =
/2

/2

1 + cos 2
d =
2


sin 2 /2

= .
+

2
4
2
/2

2.2.1

Changing the Order of Integration

In some cases, a region can be classified as being of either type I or type II,
and therefore a function can be integrated over the region in two different
ways. However, one approach or the other may be impractical, due to the

2.2. DOUBLE INTEGRALS OVER MORE GENERAL REGIONS

81

complexity, or even impossibility, of carrying out the anti-differentiation.

Therefore, it is important to be able to change the order of integration if
necessary.
Example Consider the double integral
Z Z

ey dA

where D = {(x, y) | 0 x 1, x y 1}. This region is defined as a

region of type I, so it is natural to attempt to evaluate the iterated integral
Z

1Z 1

ey dy dx.

x
3

Unfortunately, it is impossible to anti-differentiate ey with respect to y.

However, the region D is also a region of type II, as it can be redefined as
D = {(x, y) | 0 y 1, 0 x y 2 }.
We then have
Z Z

ey dA =

1 Z y2

Z
0

0
1

 2
3 y
xey  dy

0
1

ey dx dy

y 2 ey dy

=
0

=
=
=

1
3

eu du,

u = y3,


1 u 1
e
3 0
1
(e 1).
3

It should be noted that usually, when changing the order of integration, it is

necessary to use the inverse functions of the functions that define the curved
portions of the boundary, in order to obtain the limits of the integration of
the new inner integral.

82

2.2.2

CHAPTER 2. MULTIPLE INTEGRALS

The Mean Value Theorem for Integrals

It is important to note that all of the properties of double integrals that have
been previously discussed, including linearity, homogeneity, monotonicity,
and additivity, apply to double integrals over non-rectangular regions as
well. One additional property, that is a consequence of monotonicity, is that
if f (x, y) m on a region D, and f (x, y) M on D, then
Z Z
f (x, y) dA M A(D),
mA(D)
D

where, as before, A(D) is the area of D. Furthermore, if f is continuous on

D, then, by the Mean Value Theorem for Double Integrals, we have
Z Z
f (x, y) dA = f (x0 , y0 )A(D),
D

where (x0 , y0 ) is some point in D. This is a generalization of the Mean Value

Theorem for Integrals, which is closely related to the Mean Value Theorem
for derivatives.
Example Consider the double integral
Z Z
ey dA
D

where D is the triangle defined by D = {(x, y) | 0 x 1, 0 y 4x}.

The area of this triangle is given by A(D) = 12 bh, where b, the base, is 1
and h, the height, is 4, which yields A(D) = 2. Because 1 ey e4 when
0 y 4, it follows that
Z Z
2
ey dA 2e4 109.2.
D

The exact value is 14 (e4 5) 12.4, which is between the above lower and
upper bounds. 2

2.3

Double Integrals in Polar Coordinates

We have learned how to integrate functions of two variables, x and y, over

various regions that have a simple form. The variables x and y correspond
to Cartesian coordinates that are normally used to describe points in 2D space. However, a region that may not be of type I or type II, when

2.3. DOUBLE INTEGRALS IN POLAR COORDINATES

83

described using Cartesian coordinates, may be of one of these types if it is

instead described using polar coordinates r and .
We recall that polar coordinates are related to Cartesian coordinates by
the equations
x = r cos , y = r sin ,
or, alternatively,

y
.
x
In order to integrate a function over a region defined using polar coordinates,
we must derive the double integral in these coordinates, as was previously
done in Cartesian coordinates.
Let a solid be bounded by the surface z = f (r, ), as well as the surfaces
r = a, r = b, = and = , which define a polar rectangle. To compute
the volume of this solid, we can approximate it by several solids for which
the volume can easily be computed. This is accomplished by dividing the
polar rectangle into several smaller polar rectangles of dimensions r and
. The height of each solid is obtained from the value of the function at a
point in the polar rectangle.
Specifically, we divide the interval [a, b] into n subintervals of width r =
(b a)/n. Each subinterval is of the form [ri1 , ri ], where ri = a + ir, for
i = 1, 2, . . . , n. Similarly, [, ] is divided into m subintervals of width =
( )/m, and each subinterval is of the form [j1 , j ], where j = +j.
Then, the volume V of the solid is approximated by
r 2 = x2 + y 2 ,

n X
m
X
1
i=1 j=1

tan =

2
f (ri , j )(ri2 ri1
),

where, for each i, ri1 ri ri , and for each j, j1 j j .

The quantity 21 r2 is the area of a polar rectangle, for it is not truly
a rectangle, but rather the difference between two circular sectors with angle
and radii ri1 and ri . However, from
1 2
1
1
2
(ri ri1
) = (ri1 + ri )(ri ri1 ) = (ri1 + ri )r,
2
2
2
we see that as m, n , this approximation of the volume converges to
the exact volume, which is given by the double integral
Z

V =

f (r, ) r dr d.

84

CHAPTER 2. MULTIPLE INTEGRALS

Note the extra factor of r in the integrand, which is the limit as n of

(ri1 + ri )/2.
If the base of the solid can be represented by a polar region of type I,
D = {(r, ) | , h1 () r h2 ()},
then the volume V of the solid defined by the surface z = f (r, ) and the
surfaces that define D is given by the iterated integral
Z Z h2 ()
f (r, ) r dr d.
V =

h1 ()

As before, if f (r, ) 1, then this integral yields A(D), the area of D.

Example To evaluate the double integral
Z Z
x + y dA,
D

x2

y2

where D = {(x, y) | 1
+
4, x 0}, we convert the integrand, and
the description of D, to polar coordinates. We then have
Z Z
r cos + r sin dA
D

where D = {(r, ) | 1 r 2, /2 3/2}. This simplifies the integral

considerably, because D can be described as a polar rectangle. We then
have
Z Z
Z 3/2 Z 2
x + y dA =
(r cos + r sin )r dr d
D

/2

3/2 Z 2

=
/2

3/2

=
/2

r2 (cos + sin ) dr d

2
r3 
(cos + sin )  d
3 1

Z
7 3/2
(cos + sin ) d
3 /2
7
3/2
=
(sin cos )|/2
3
7
[(1 0) (1 0)]
=
3
14
= .
3
=

2.3. DOUBLE INTEGRALS IN POLAR COORDINATES

85

2
Example To computep
the volume of the solid in the first octant bounded
below by the cone z = x2 + y 2 , and above by the sphere x2 + y 2 + z 2 = 8,
as well as the planes y = x and y = 0, we first rewrite the equations of the
bounding surfaces in polar coordinates. The solid is bounded below by the
cone z = r, above by the sphere r2 + z 2 = 8, and the surfaces = 0 and
= /4, since the solid lies in the first octant. The surfaces that bound the
solid above and below intersect when 2r2 = 8, or r = 2. It follows that the
volume is given by
/4 Z 2

Z
V

=
0

p
[ 8 r2 r]r dr d

0
/4 Z 2

=
0

=
=
=
=

1
2
Z

Z
p
8 r2 dr d

0
/4 Z 4

0
/4

u1/2 du d

/4 Z 2

0
/4

r2 dr d

0
2
3
r 

 d
3 0


Z /4
8
2 3/2 8
u  d
d
3
3
0
0
4
Z
1 /4
2
[16 2 8] d
3 0
3
4
[ 2 1].
3
1
2

In the third step, the substitution u = 8 r2 is used. Then, the limits of

integration are interchanged in order to reverse the sign of the integral. 2
Example The double integral
Z

1x2

f (x, y) dy dx
1

can be converted to polar coordinates by converting the equation

that de
scribes the top boundary of the domain of integration, y = 1 x2 , into a
polar equation. We substitute x = r cos and y = r sin into this equation
to obtain
p
r sin = 1 cos2 .
Squaring both sides yields r2 sin2 = 1 cos2 , and, in view of the identity
cos2 + sin2 = 1, we obtain the polar equation r = 1. Because the bottom

86

CHAPTER 2. MULTIPLE INTEGRALS

boundary, y = 0, corresponds to the rays = 0 and = , the integral can

be expressed in polar coordinates as
Z Z 1
f (r cos , r sin ) r dr d.
0

2
Example We evaluate the double integral
Z

2Z

2xx2

p
x2 + y 2 dy dx

by converting to polar coordinates By completing the square, we obtain

2x x2 = 1 (x 1)2 . It follows that the region D over which the integral
is to be evaluated,
p
D = {(x, y) | 0 x 2, 0 y 2x x2 },

has its top boundary defined by the equation y = 2x x2 , or

(x 1)2 + y 2 = 1.
That is, the top boundary is the upper half of the circle with radius 1 and
center (1, 0). In polar coordinates, the equation of the top boundary becomes
(r cos 1)2 + r2 sin2 = 1,
or, upon expanding and simplifying,
r = 2 cos .
The region D is contained between the rays = 0 and = /2. It follows
that in polar coordinates, D is defined by
D = {(r, ) | 0 /2, 0 r 2 cos }.
The lower limit r = 0 is obtained from the fact that D contains the origin.
We thus obtain the integral
Z
0

2Z
0

2xx2

x2

y 2 dy dx

/2 Z 2 cos

=
0

r2 dr d.

The integrand of the original integral is r, but the additional factor of r

required by the change to polar coordinates yields an integrand of r2 .

2.4. TRIPLE INTEGRALS

87

Evaluating this integral, we obtain

Z 2 Z 2xx2 p
Z /2 Z
2
2
x + y dy dx =
0

Z
=

=
=
=
=
=

r2 dr d

0

/2 3 2 cos
r


3 0

2 cos

8
3

8
3

8
3

/2

cos3 d

0
/2

cos2 cos d

0
/2

(1 sin2 ) cos d

Z
Z
8 /2
8 /2 2
cos d
sin cos d
3 0
3 0
Z
8
8 1 2
/2
sin |0
u du
3
3 0
1
8
8 u3 
(1)
3
3 3
0

16
.
9

2.4

Triple Integrals

The integral of a function of three variables over a region D R3 can be

defined in a similar way as the double integral. Let D be the box defined by
D = {(x, y, z) | a x b, c y d, r z s}.
Then, as with the double integral, we divide [a, b] into n subintervals of width
x = (b a)/n, with endpoints [xi1 , xi ], for i = 1, 2, . . . , n. Similarly, we
divide [c, d] into m subintervals of width y = (d c)/m, with endpoints
[yj1 , yj ], for j = 1, 2, . . . , m, and divide [r, s] into ` subintervals of width
z = (s r)/`, with endpoints [zk1 , zk ] for k = 1, 2, . . . , `.
Then, we can define the triple integral of a function f (x, y, z) over D by
Z Z Z
f (x, y, z) dV =
D

lim

m,n,`

n X
m X
`
X
i=1 j=1 k=1

f (xi , yj , zk ) V,

88

CHAPTER 2. MULTIPLE INTEGRALS

where V = xyz. As with double integrals, the practical method of

evaluating a triple integral is as an iterated integral, such as
Z Z Z
Z sZ dZ b
f (x, y, z) dV =
f (x, y, z) dx dy dz.
D

By Fubinis Theorem, which generalizes to three dimensions or more, the

order of integration can be rearranged when f is continuous on D.
A triple integral over a more general region can be defined in the same
way as with double integrals. If E is a bounded subset of R3 , that is contained within a box B, then we can define
Z Z Z
Z Z Z
f (x, y, z) dV =
F (x, y, z) dV,
E

where


F (x, y, z) =

f (x, y, z) (x, y, z) E,
.
0
(x, y, z)
/E

All of the properties previously associated with the double integral, such as
linearity and additivity, generalize to the triple integral as well.
Just as regions were classified as type I or type II for double integrals,
they can be classified for the purpose of setting up triple integrals. A solid
region E is said to be of type 1 if it lies between the graphs of two continuous functions of x and y that are defined on a two-dimensional region D.
Specifically,
E = {(x, y, z) | (x, y) D, u1 (x, y) z u2 (x, y)}.
Then, an integral of a function f (x, y, z) over E can be evaluated as
Z Z Z

Z Z Z

u2 (x,y)

f (x, y, z) dV =

f (x, y, z) dz dA,

u1 (x,y)

where the double integral over D can be evaluated in a manner that is

appropriate for the type of D.
For example, if D is of type I, then
E = {(x, y, z) | a x b, g1 (x) y g2 (x), u1 (x, y) z u2 (x, y)},
and therefore
Z Z Z

Z bZ

g2 (x) Z u2 (x,y)

f (x, y, z) dV =
E

f (x, y, z) dz dy dx.
a

g1 (x)

u1 (x,y)

2.4. TRIPLE INTEGRALS

89

On the other hand, if E is of type 2, then it has a definition of the form

E = {(x, y, z) | (y, z) D, u1 (y, z) x u2 (y, z)}.
That is, E lies between the graphs of two continuous functions of y and z
that are defined on a two-dimensional region D. Finally, if E is a region of
type 3, then it lies between the graphs of two continuous functions of x and
z. That is,
E = {(x, y, z) | (x, z) D, u1 (y, z) y u2 (y, z)}.
If more than one type applies to a given region E, then the order of evaluation can be determined by which ordering leads to the integrands that are
most easily anti-differentiated within each single integral that arises.
Example Let E be a solid tetrahedron bounded by the planes x = 0, y = 0,
z = 0 and x + y + z = 1. We wish to integrate the function f (x, y, z) = xz
over this tetrahedron. From the given bounding planes, we see that the
tetrahedron is bounded below by the plane z = 0 and above by the plane
z = 1 x y. Therefore, we surmise that E can be viewed as a solid of type
1. This requires finding a region D in the xy-plane such that E is bounded
by z = 0 and z = 1 x y on D.
We first note that these planes intersect along the line x + y = 1. It
follows that the base of E is a 2-D region D that can be described by the
inequalities x 0, y 0, and x + y 1. This region is of type I or type II,
so we choose type I and obtain the description
D = {(x, y) | 0 x 1, 0 y 1 x}.
Therefore, we can integrate f (x, y, z) over E as follows:
Z Z Z
Z 1 Z 1x Z 1xy
xz dV =
xz dz dy dx
E
0
0
0
Z 1 Z 1x Z 1xy
=
x
z dz dy dz
0

0
1

Z
=

x
0

=
=

1x


2 0

1
2

1
2

Z
x

0
1xy
z2 

dy dx

1x

(1 x y)2 dy dx

0
1


1x
(1 x y)3 
x
dx

3
0

90

CHAPTER 2. MULTIPLE INTEGRALS

=
=
=
=
=
=

Z
1 1
x(1 x)3 dx
6 0
Z
1 1
(1 u)u3 du,
6 0
Z
1 1 3
u u4 du
6 0

1
1 u4 u5 

6 4
5 

 0
1 1 1

6 4 5
1
.
120

u=1x

Example We will compute the volume of the solid E bounded by the surfaces y = x, y = x2 , z = x, and z = 0. Because E is bounded by two
surfaces that define z as a function of x and y, we view E as a solid of type
1. It is bounded by the graphs of the functions z = 0 and z = x that are
defined on a region D in the xy-plane. This region is bounded by the curves
y = x and y = x2 . Because these curves intersect when x = 0 and x = 1,
we can describe D as a region of type I:
D = {(x, y) | 0 x 1, x2 y x}.
It follows that the volume of E is given by the iterated integral
Z 1Z xZ x
Z Z Z
1 dV =
1 dz dy dx
0
x2 0
E
Z 1Z x
=
x dy dx
0
x2
Z 1 Z x
=
x
1 dy dx
0
x2
Z 1
=
x(x x2 ) dx
0
Z 1
=
x2 x3 dx
0


=
=

1
x3 x4 

3
4 0

1
.
12

2.4. TRIPLE INTEGRALS

91

2
Example We evaluate the triple integral
Z Z Z
x dV
E

where E is the solid bounded by the paraboloid x = 4y 2 + 4z 2 and the

plane x = 4. The paraboloid and the plane intersect when y 2 + z 2 = 1. It
follows that the right boundary of the solid E is the unit disk y 2 + z 2 1,
contained within the plane x = 4. Because the paraboloid serves as the
left boundary of E, we can define E by the inequalities
E = {(x, y, z) | y 2 + z 2 1, 4(y 2 + z 2 ) x 4}.
Therefore, the triple integral can be written as an iterated integral
#
Z Z Z
Z Z "Z 4
x dx dA,
x dV =
E

4(y 2 +z 2 )

where D is the unit disk in the yz-plane, y 2 + z 2 = 1. If we convert y and z

to polar coordinates y = r cos and z = r sin , we can rewrite this integral
as
Z Z Z
Z
Z Z
2

x dV =
E

xr dx dr d.
0

4r2

Evaluating this integral, we obtain

Z 2 Z
Z Z Z
x dV =
0

xr dx dr d
4r2
4
x2 

0
2

1Z 4

=
0

r(1 r4 ) dr d

= 8
0

0
2

= 8
0

1
d
3

= 8
0

r r5 dr d

1
r2 r6 
 d
2
6 0

= 8


dr d
2 4r2

16
.
3

92

2.5

CHAPTER 2. MULTIPLE INTEGRALS

Applications of Double and Triple Integrals

We now explore various applications of double and triple integrals arising

from physics. When an object has constant density , then it is known that
its mass m is equal to V , where V is its volume. Now, suppose that a
flat plate, also known as a lamina, has a non-uniform density (x, y), for
(x, y) D, where D defines the shape of the lamina. Then, its mass is given
by
Z Z
m=

(x, y) dA.
D

Similarly, if E is a solid region in 3-D space, and (x, y, z) is the density of

the solid at the point (x, y, z) E, then the mass of the solid is given by
Z Z Z
m=
(x, y, z) dV.
E

We see that just as the integral allows simple product formulas for area and
volume to be applied to more general problems, it allows similar formulas
for quantities such as mass to be generalized as well.
The center of mass, also known as the center of gravity, of an object is
the point at which the object behaves as if its entire mass is concentrated
at that point. If the object is one- or two-dimensional, the center of mass is
the point at which the object can be balanced horizontally (like a see-saw
with riders at either end, in the one-dimensional case).
For a lamina with its shape defined by a bounded region D R2 , and
with density given by(x, y), its center of mass (
x, y) is located at
x
=

My
,
m

y =

Mx
,
m

where Mx and My are the moments of the lamina about the x-axis and
y-axis, respectively. These are given by
Z Z
Z Z
Mx =
y(x, y) dA, My =
x(x, y) dA.
D

These integrals are obtained from the formula for the moment of a point
mass about an axis, which is given by the product of the mass and the
distance from the axis.
Similarly, the moments about the xy-, yz- and xz-planes, Mxy , Myz , and
Mxz , of a solid E R3 with density (x, y, z) are given by
Z Z Z
Mxy =
z(x, y, z) dV,
E

2.6. TRIPLE INTEGRALS IN CYLINDRICAL COORDINATES

93

Z Z Z
x(x, y, z) dV,

Myz =
Z Z

ZE
y(x, y, z) dV.

Mxz =
E

It follows that its center of mass (

x, y, z) is located at
x
=

Myz
,
m

y =

Mxz
,
m

z =

Mxy
.
m

As in the 2-D case, each moment is defined using the distance of each point
of E from the coordinate plane about which the moment is being computed.
The moment of interia, or second moment, of an object about an axis
gives an indication of the objects tendency to rotate about that axis. For
a lamina defined by a region D R2 with density function (x, y), its
moments of inertia about the x-axis and y-axis, Ix and Iy respectively, are
given by
Z Z
Z Z
2
Ix =
y (x, y) dA, Iy =
x2 (x, y) dA.
D

On the other hand, for a solid defined by a region E R3 with density

(x, y, z), its moments of inertia about the coordinate axes are defined by
Z Z Z

Ix =

Z Z Z

(y + z )(x, y, z) dV,
E

(x2 + z 2 )(x, y, z) dV,

Iy =
E

Z Z Z

(x2 + y 2 )(x, y, z) dV.

Iz =
E

The moment Iz is also called the polar moment of interia, or the moment of
interia about the origin, when E reduces to a lamina with density (x, y).

2.6

Triple Integrals in Cylindrical Coordinates

We have seen that in some cases, it is convenient to evaluate double integrals

by converting Cartesian coordinates (x, y) to polar coordinates (r, ). The
same is true of triple integrals. When this is the case, Cartesian coordinates
(x, y, z) are converted to cylindrical coordinates (r, , z).
The relationships between (x, y) and (r, ) are exactly the same as in
polar coordinates, and the z coordinate is unchanged.

94

CHAPTER 2. MULTIPLE INTEGRALS

Example The point (x, y, z) = (3, 3, 4) can be converted to cylindrical

coordinates (r, , z) using the relationships from polar coordinates,
p
y
r = x2 + y 2 , tan = .
x
These relationships yield
p

r = 32 + (3)2 = 18 = 3 2,

tan = 1.

Since x = 3 < 0, we have = tan1 (1) + = 3/4. We conclude that

the cylindrical coordinates of the point (3, 3, 4) are (3 2, 3/4, 4). 2
Furthermore, just as conversion to polar coordinates in double integrals
introduces a factor of r in the integrand, conversion to cylindrical coordinates
in triple integrals also introduces a factor of r.
Example We evaluate the triple integral
Z Z Z
f (x, y, z) dV,
E

where E is the solid bounded below by the paraboloid z = x2 + y 2 , above

by the plane z = 4, and the planes y = 0 and y = 2. This integral can be
evaluated as an iterated integral
Z

4x2

f (x, y, z) dz dy dx,
2

x2 +y 2

but if we instead describe the region using cylindrical coordinates, we find

that the solid is bounded below by the paraboloid z = r2 , above by the
plane z = 4, and contained within the polar box 0 r 2, 0 .
We can therefore evaluate the iterated integral
Z 2Z Z 4
f (r cos , r sin , z) r dz d dr,
0

r2

that has much simpler limits. 2

Example We use cylindrical coordinates to evaluate the triple integral
Z Z Z
x dV
E

where E is the solid bounded by the planes z = 0 and z = x + y + 5, and the

cylindrical shells x2 + y 2 = 4 and x2 + y 2 = 9. In cylindrical coordinates, E

2.7. TRIPLE INTEGRALS IN SPHERICAL COORDINATES

95

is bounded by the planes z = 0 and z = r(cos + sin ) + 5, and the cylinders

r = 2 and r = 3. It follows that the integral can be written as the iterated
integral
Z

Z Z Z

3 Z r(cos +sin )+5

(r cos )r dz dr d.

x dV =
0

Evaluating this integral, we obtain

Z Z Z

Z
x dV

cos
0

r
Z

r(cos +sin )+5

dz dr d
0

2
2

[r3 (cos + sin ) + 5r2 ] dr d

Z 2
Z 3
Z 2
Z
=
cos (cos + sin )
r3 dr d +
cos
cos

Z
=

cos (cos + sin )

2
3
r4 

0
2

5r2 dr d

2
3
3

r

 d + 5
cos  d
4 2
3 2
0
Z 2
Z 2
95
65
cos2 + sin cos d +
cos d
4 0
3 0
Z
65 2 1
1
95
(1 + cos 2) + sin 2 d +
sin |2
0
4 0 2
2
3

2

65 1
1
1
+ sin 2) cos 2  d
4 2
2
4
0

=
=
=

65
.
4

2.7

Triple Integrals in Spherical Coordinates

Another approach to evaluating triple integrals, that is especially useful

when integrating over regions that are at least partially defined using spheres,
is to use spherical coordinates. Consider a point (x, y, z) that lies on a sphere
of radius . Then we know that x2 + y 2 + z 2 = 2 . Furthermore, the points
(0, 0, 0), (0, 0,
pz) and (x, y, z) form a right triangle with hypotenuse and
legs |z| and 2 z 2 .
If we denote by the angle adjacent to the leg of length |z|, then can be
interpreted as an angle of inclination of the point (x, y, z). The angle = 0
corresonds to the north pole of the sphere, while = /2 corresponds to

96

CHAPTER 2. MULTIPLE INTEGRALS

the equator, and = corresponds to the south pole. By right triangle

trigonometry, we have
z = cos .
It follows that x2 + y 2 = 2 sin2 . If we define the angle to have the same
meaning as in polar coordinates, then we have
x = sin cos ,

y = sin sin .

We define the spherical coordinates of (x, y, z) to be (, , ).

Example To convert the point (x, y, z) = (1, 3, 4) to spherical coordinates, we first compute
q
p

2
2
2
= x + y + z = 12 + ( 3)2 + (4)2 = 20 = 2 5.
Next, we use the relation tan = y/x, and the fact that x = 1 > 0, to obtain
= tan1

= tan1 3 = .
x
3

Finally, to obtain , we use the relation z = cos , which yields



4
1 z
1
= cos
= cos

2.6779 radians.

2 5
2
To evaluate integrals in spherical coordinates, it is important to note
that the volume of a spherical box of dimensions r, and , as
, , 0, converges to the infinitesimal
2 sin dr d d,
where (, , ) denotes the location of the box in the limit. Therefore, the
integral of a function f (x, y, z) over a solid E, when evaluated in spherical
coordinates, becomes
Z Z Z
Z Z Z
f (x, y, z) dV =
f ( sin cos , sin sin , cos ) 2 sin d d d.
E

Example We wish to compute the volume of the solid E in the first octant
bounded below by the plane z = 0 and the hemisphere x2 + y 2 + z 2 = 9,
bounded above by the hemisphere x2 + y 2 + z 2 = 16, and the planes y = 0
and y = x. This would be highly inconvenient to attempt to evaluate in

2.7. TRIPLE INTEGRALS IN SPHERICAL COORDINATES

97

Cartesian coordinates; determining the limits in z alone requires breaking

up the integral with respect to z. However, in spherical coordinates, the
solid E is determined by the inequalities
3 4,

,
4

.
2

That is, the solid is actually a spherical rectangle. It follows that the
volume V is given by the iterated integral
Z /2 Z /4 Z 4
2 sin d d d
V =
0

=
=
=

0
3
/2 Z 4
2

sin d d d

3
/2

2 d d d

sin
0
/2

sin
0

3
4
3 

 d d
3 3

Z
37 /2
=
sin d d
4 3 0
37
/2
=
cos |0
4 3
37
.
=
12
2
Example We use spherical coordinates to evaluate the triple integral
Z Z Z
(x2 + y 2 ) dV,
H

where H is the solid that is bounded below by the xy-plane, and bounded
above by the sphere x2 + y 2 + z 2 = 1. In spherical coordinates, H is defined
by the inequalities
H = {(, , ) | 0 1, 0 2, 0 /2}.
As the integrand x2 +y 2 is equal to ( cos sin )2 +( sin sin )2 = 2 sin2
in spherical coordinates, we have
Z Z Z
Z 2 Z /2 Z 1
2
2
(x + y ) dV =
(2 sin2 )2 sin d d d.
H

98

CHAPTER 2. MULTIPLE INTEGRALS

Evaluating this integral, we obtain

Z
Z 2 Z /2
Z Z Z
sin3
(x2 + y 2 ) dV =
0

0
2

/2

sin3

=
0

1
5

1
5

Z
Z

1
5

=
=
=
=
=

0
2 Z /2

4 d d d

0
1
5 

 d d
5 0

sin3 d d

0
2

/2

sin2 sin d d

0
2

/2

(1 cos2 ) sin d d

Z
Z
Z
Z
1 2 /2
1 2 /2
sin d d
cos2 sin d d
5 0
5
0
0
0
Z
Z
Z
1 2
1 2 1 2
/2
( cos )|0 d
u du d
5 0
5 0
0
1
Z
Z
1 2
1 2 u3 
1 d
d
5 0
5 0
3 0


2
1
2
5
3
4
.
15

2.8

Change of Variables in Multiple Integrals

Recall that in single-variable calculus, if the integral

Z b
f (u) du
a

is evaluated by making a change of variable u = g(x), such that the interval

x is mapped by g to the interval a u b, then
Z b
Z
f (u) du =
f (g(x))g 0 (x) dx.
a

g 0 (x)

The appearance of the factor

in the integrand is due to the fact
that if we divide [a, b] into n subintervals [ui1 , ui ] of equal width u =

2.8. CHANGE OF VARIABLES IN MULTIPLE INTEGRALS

99

(b a)/n, and if we divide [, ] into n subintervals [xi1 , xi ] of equal width

x = ( )/n, then
u = ui ui1 = g(xi ) g(xi1 ) = g 0 (xi )x,
where xi1 xi xi . We will now generalize this change of variable to
multiple integrals.
For simplicity, suppose that we wish to evaluate the double integral
Z Z
f (x, y) dA
D

by making a change of variable

x = g(u, v),

y = h(u, v),

a u b,

c v d.

We divide the interval [a, b] into n subintervals [ui1 , ui ] of equal width u =

(b a)/n, and we divide [c, d] into m subintervals [vi1 , vi ] of equal width
v = (d c)/m. Then, the rectangle [ui1 , ui ] [vi1 , vi ] is approximately
mapped by g and h into a parallelogram with adjacent sides
ru = hg(ui , vi1 ) g(ui1 , vi1 ), h(ui , vi1 ) h(ui1 , vi1 )i,
rv = hg(ui1 , vi ) g(ui1 , vi1 ), h(ui1 , vi ) h(ui1 , vi1 )i.
By the Mean Value Theorem, we have
ru hgu (ui1 , vi1 ), hu (ui1 , vi1 )iu,
rv hgv (ui1 , vi1 ), hv (ui1 , vi1 )iv.
The area of this parallelogram is given by


 g h g h 
 uv.

|ru rv | = 
u v
v u 
It follows that


Z Z
Z Z
 (x, y) 

 du dv,
f (x, y) dx dy =
f (g(u, v), h(u, v)) 
(u, v) 

D
D
= [a, b] [c, d] is the domain of g and h, and
where D
 x x 
 x y x y
(x, y)  u
v  =
=  y y
 u v v u
(u, v)
u
v

100

CHAPTER 2. MULTIPLE INTEGRALS

is the Jacobian of the transformation from (u, v) to (x, y). It is also the
determinant of the Jacobian matrix of the vector-valued function that maps
(u, v) to (x, y).
Example Let D be the parallelogram with vertices (0, 0), (2, 4), (6, 1), and
(8, 5). To integrate a function f (x, y) over D, we can use a change of variable
(x, y) = (g(u, v), h(u, v)) that maps a rectangle to this parallelogram, and
then integrate over the rectangle.
Using the vertices, we find that the equations of the edges are
x + 6y = 0,

x + 6y = 22,

2x y = 0,

2x y = 11.

Therefore, if we define the new variables u and v by the equations

u = x + 6y,

v = 2x y,

then, for (x, y) D, we have (u, v) belonging to the rectangle 0 u 22,

0 v 11.
To rewrite an integral over D in terms of u and v, it is much easier to
express the original variables in terms of the new variables than the other
way around. Therefore, we need to solve the equations defining u and v for
x and y. From the equation for u, we have x = 6y u. Substituting into
the equation for v, we obtain v = 2(6y u) y, which yields y = h(u, v) =
1
11 (2u + v). Subtituting this into the equation for u yields x = g(u, v) =
1
11 (u + 6v).
The Jacobian of this transformation is
 x x 
 x y x y
(x, y)  u
1
1
v  =
=  y y
 u v v u = 112 [1(1) 6(2)] = 11 .
(u, v)
u
v
We conclude that
Z Z
Z Z
1
f (x, y) dx dy =
f (g(u, v), h(u, v)) du dv.
11

D
D
2
In general, when integrating a function f (x1 , x2 , . . . , xn ) over a region
D Rn , if the integral is evaluated using a change of variable (x1 , x2 , . . . , xn ) =
g(u1 , u2 , . . . , un ) that maps a region E Rn to D, then
Z
Z
f (x1 , . . . , xn ) dx1 dxn = (f g)(u1 , . . . , un )| det(Jg (u1 , . . . , un ))| du1 dun ,
D

2.8. CHANGE OF VARIABLES IN MULTIPLE INTEGRALS

101

where

Jg (u1 , u2 , . . . , un ) =

x1
u1
x2
u1

x1
u2
x2
u2

xn
u1

xn
u2

..
.

..
.

x1
un
x1
un

..
.

xn
un

is the Jacobian matrix of g and det(Jg (u1 , u2 , . . . , un )) is its determinant,

which is simply referred to as the Jacobian of the transformation g.
Example Consider the transformation from spherical to Cartesian coordinates,
x = sin cos , y = sin sin , z = cos .
Then, the Jacobian matrix of this transformation is

sin cos sin sin cos cos

= sin sin sin cos cos sin .
cos
0
sin

It follows that the Jacobian of this transformation is given by the determinant of this matrix,

 sin cos sin sin cos cos

 sin sin sin cos cos sin

 cos
0
sin








 = cos  sin sin cos cos 

 sin cos cos sin 



 sin cos sin sin 

sin 
sin sin sin cos 
= cos [2 sin cos sin2 2 sin cos cos2 ]
sin [ sin2 cos2 + sin2 sin2 ]
= 2 cos2 sin 2 sin2 sin
= 2 sin .

The absolute value of the Jacobian is the factor that must be included in
the integrand when converting a triple integral from Cartesian to spherical
coordinates. 2
Example We evaluate the double integral
Z Z
(x2 xy + y 2 ) dA,
R

102

CHAPTER 2. MULTIPLE INTEGRALS

where R is the region bounded by the ellipse x2 xy + y 2 = 2, using the

change of variables
p
p

x = 2u 2/3v, y = 2u + 2/3v.
First, we compute the Jacobian of the change of variables,
p


 x x 
p
p
(x, y)
4
2

2/3
u
v
p

= det
= 2 2/3+ 2 2/3 = .
= det
y
y
2
2/3
(u, v)
3
u
v
Next, we need to define the region R in terms of u and v. Rewriting the
equation x2 xy +y 2 = 2 in terms of u and v yields the equation 2u2 +2v 2 =
to R, where R

2. It follows that the change of variables maps the region R

is the unit disk. If we then use polar coordinates u = r cos and v = r sin ,
we have
Z Z
Z Z
Z 2 Z 1
4
4
(x2 xy+y 2 ) dA =
(2u2 +2v 2 ) du dv =
(2r2 )r dr d.

3
3
R
R
0
0
Evaluating this integral, we obtain
Z Z

(x xy + y ) dA =
R

=
=
=

Z
0

r3 dr d

0

2 4 1
r

Z
8
 d

4 0
3 0
Z 2
2

1 d
3 0
4
.
3

2
Example We wish to use an appropriate change of variable to evaluate the
double integral
Z Z
2
2
(x + y)ex y dA,
R

where R is the rectangle enclosed by the lines x y = 0, x y = 2, x + y = 0

and x + y = 3. If we define u = x + y and v = x y, then R is mapped by
this change of variables to the rectangle
= {(u, v) | 0 u 3, 0 v 2}.
R

2.8. CHANGE OF VARIABLES IN MULTIPLE INTEGRALS

103

Solving for x and y in terms of u and v, we obtain

1
x = (u + v),
2

1
y = (u v).
2

It follows that
x y x y
1
(x, y)
=

=
(u, v)
u v
v u
2

 
1
11
1

=
2
22
2

and the integral becomes

Z Z
Z Z
Z
x2 y 2
(x+y)(xy)
(x+y)e
dA =
(x+y)e
dA =
R

Evaluating this integral, we obtain

Z Z
2
2
(x + y)ex y dA =
R

=
=
=
=
=
2

3Z 2

ue
0

Z Z
1 3 2 uv
ue dv du
2 0 0
Z
1 3 uv 2
e |0 du
2 0
Z
1 3 2u
[e 1] du
2 0

3

1 e2u
u 
2 2
0


1
1 e6
3
2 2
2
1 6
(e 7).
4

uv

 
 1
  dv du.
 2

104

CHAPTER 2. MULTIPLE INTEGRALS

Chapter 3

Vector Calculus
3.1

Vector Fields

To this point, we have mostly worked with scalar-valued functions of several

variables, in the interest of computing quantities such as the maximum or
minimum value of a function, or the volume or center of mass of a solid.
Now, we will study applications involving vector-valued functions of several
variables. The difficulty of visualizing such functions leads to the notion of
a vector field.
A function F : U Rn Rn is a function that assigns to each point
x U a vector
F(x) = hF1 (x), F2 (x), . . . , Fn (x)i
in Rn . The functions F1 , F2 , . . . , Fn are the component functions, or component scalar fields, of F. For our purposes, n = 2 or 3. To visualize a
vector field, one can plot the vector F (x) at any given point x, using the
component functions to obtain the components of the vector to be plotted
at each point.
The following are certain vector fields of interest in applications:
Given a fluid, for example, a velocity field is a vector field V(x, y, z)
that indicates the velocity of the fluid at each point (x, y, z). When
plotting a velocity field, the speed of the fluid at each point is indicated
by the length of the vector plotted at that point, and the direction of
the fluid at that point is indicated by the direction of the vector.
A curve c(t) is said to be a flow line, or streamline, of a velocity field
V if, for each value of the parameter t,
c0 (t) = V(c(t)).
105

106

CHAPTER 3. VECTOR CALCULUS

Figure 3.1: The vector field V(x, y) = hy, xi

That is, at each point along the curve, its tangent vector coincides with
V. A flow line can be approximated by first choosing an initial point
x0 = c(t0 ), then using the value of V at that point to approximate a
second point x1 = c(t1 ) as follows:
x 1 x0
c(t1 ) c(t0 )
=
V(c(t0 ))
t1 t0
t1 t0

x1 x0 +(t1 t0 )V(x0 ).

This can be continued to obtain the locations of any number of points

along the flow line. The closer the times t0 , t1 , . . . are to one another,
the more accurate the approximate flow line will be.
Consider two objects with mass m and M , with the object of mass M
located at the origin, and the vector field F defined by
F(r) =

mM G
r,
krk3

3.1. VECTOR FIELDS

107

where r is a position vector of the object of mass m, and G is the

gravitational constant. This vector field indicates the gravitational
force exerted by the object at the origin on the object at position r,
and is therefore an example of a gravitational field.
Suppose an electric charge Q is located at the origin, and a charge q
is located at the point with position vector x. Then the electric force
exerted by the first charge on the second is given by the vector field
F (x) =

qQ
x,
kxk3

where is a constant. This field, and the gravitational field described

above, are both examples of force fields.

Figure 3.2: The conservative vector field F(x, y) = hy, xi

A vector field F is said to be conservative if F = f for some function
f . We also say that F is a gradient field, and f is a potential function for

108

CHAPTER 3. VECTOR CALCULUS

F. When we discuss line integrals, we will learn the physical meaning
of a conservative vector field.

In upcoming sections we will learn how to integrate vector fields, as well

as the physical interpretations of such integrals.
Example Consider the velocity field V(x, y) = hy, xi. It is shown in Figure
3.1. It can be seen from the figure that the flow lines of this velocity field
are circles centered at the origin. 2
Example The vector field F(x, y) = hy, xi is conservative, because F = f ,
where f (x, y) = xy. The field is shown in Figure 3.2. It should be noted that
conservative vector fields are also called irrotational; a fluid whose velocity
field is conservative has no vorticity. 2

3.2

Line Integrals

Recall from single-variable calclus that if a constant force F is applied to

an object to move it along a straight line from x = a to x = b, then the
amount of work done is the force times the distance, W = F (b a). More
generally, if the force is not constant, but is instead dependent on x so that
the amount of force applied when the object is at the point x is given by
F (x), then the work done is given by the integral
Z b
W =
F (x) dx.
a

This result is obtained by applying the basic formula for work along each
of n subintervals of width x = (b a)/n, and taking the limit as x 0.
Now, suppose that a force is applied to an object to move it along a path
traced by a curve C, instead of moving it along a straight line. If the amount
of force that is being applied to the object at any point p on the curve C is
given by the value of a function F (p), then the work can be approximated by,
as before, applying the basic formula for work to each of n line segments
that approximate the curve and have lengths s1 , s2 , . . . , sn . The work
done on the ith segment is approximately F (pi )si , where pi is any point
on the segment. By taking the limit as max si 0, we obtain the line
integral
Z
n
X
W =
F (p) ds =
lim
F (pi ) si ,
C

provided that this limit exists.

max si 0

i=1

3.2. LINE INTEGRALS

109

In order to actually evaluate a line integral, it is necessary to express

the curve C in terms of parametric equations. For concreteness, we assume
that C is a plane curve defined by the parametric equations
x = x(t),

y = y(t),

a t b.

Then, if we divide [a, b] into subintervals of width t = (b a)/n, with endpoints [ti1 , ti ] where ti = a+it, we can approximate C by n line segments
with endpoints (x(ti1 ), y(ti1 )) and (x(ti ), y(ti )), for i = 1, 2, . . . , n. From
the Pythagorean Theorem, it follows that the ith segment has length
s



q
xi 2
yi 2
2
2
+
t,
si = xi + yi =
t
t
where xi = x(ti ) x(ti1 ) and yi = y(ti ) y(ti1 ). Letting t 0, we
obtain
s 
 2
Z
Z b
dx 2
dy
F (p) ds =
F (x(t), y(t))
+
dt.
dt
dt
C
a
We recall that if F (x, y) 1, then this integral yields the arc length of the
curve C.
Example (Stewart, Section 13.2, Exercise 8) To evaluate the line integral
Z
x2 z ds
C

where C is the line segment from (0, 6, 1) to (4, 1, 5), we first need parametric equations for the line segment. Using the vector between the endpoints,
v = h4 0, 1 6, 5 (1)i = h4, 5, 6i,
we obtain the parametric equations
x = 4t,

y = 6 5t,

z = 1 + 6t,

0 t 1.

It follows that
Z
Z 1
p
2
x z ds =
(x(t))2 z(t) [x0 (t)]2 + [y 0 (t)]2 + [z 0 (t)]2 dt
C
0
Z 1
p
=
(4t)2 (6t 1) 42 + (5)2 + 62 dt
0
Z 1

=
16t2 (6t 1) 77 dt
0

110

CHAPTER 3. VECTOR CALCULUS

6t3 t2 dt

= 16 77
0

1
 4
t
t3 
= 16 77 6
4
3 0



3 1
= 16 77

2 3

56 77
=
.
3

2
Example (Stewart, Section 13.2, Exercise 10) We evaluate the line integral
Z
(2x + 9z) ds
C

where C is defined by the parametric equations

x = t,

y = t2 ,

z = t3 ,

0 t 1.

We have
Z
Z 1
p
(2x + 9z) ds =
(2x(t) + 9z(t)) [x0 (t)]2 + [y 0 (t)]2 + [z 0 (t)]2 dt
C
0
Z 1
p
=
(2t + 9t3 ) 12 + (2t)2 + (3t2 )2 dt
0
Z 1
p
=
(2t + 9t3 ) 1 + 4t2 + 9t4 dt
0
Z
1 14 1/2
=
u du, u = 1 + 4t2 + 9t4
4 1

1 2 3/2 14
=
u 
4 3
1
1 3/2
=
(14 1).
6
2
Although we have introduced line integrals in the context of computing
work, this approach can be used to integrate any function along a curve. For
example, to compute the mass of a wire that is shaped like a plane curve C,
where the density of the wire is given by a function (x, y) defined at each

3.2. LINE INTEGRALS

111

point (x, y) on C, we can evaluate the line integral

Z
(x, y) ds.
m=
C

It follows that the center of mass of the wire is the point (

x, y) where
Z
Z
1
1
x
=
x(x, y) ds, y =
y(x, y) ds.
m C
m C
Now, suppose that a vector-valued force F is applied to an object to
move it along the path traced by a plane curve C. If we approximate the
curve by line segments, as before, the work done along the ith segment is
approximately given by
Wi = F(pi ) [T(pi )si ]
where pi is a point on the segment, and T(pi ) is the unit tangent vector
to the curve at this point. That is, F T = kFk cos is the amount of force
that is applied to the object at each point on the curve, where is the angle
between F and the direction of the curve, which is indicated by T. In the
limit as max si 0, we obtain the line integral of F along C,
Z
F T ds.
C

If the curve C is parametrized by the the vector equation r(t) = hx(t), y(t)i,
where a t b, then the tangent vector is parametrized by
T(t) = r0 (t)/kr0 (t)k,
and, as before, ds =
Z

Z
F T ds =

p
[x0 (t)]2 + [y 0 (t)]2 dt = kr0 (t)k dt. It follows that

F(r(t))
a

r0 (t)
kr0 (t)k dt =
kr0 (t)k

Z
a

F(r(t)) r0 (t) dt =

Z
F dr.
C

The last form of the line integral is merely an abbreviation that is used for
convenience. As with line integrals of scalar-valued functions, the parametric representation of the curve is necessary for actual evaluation of a line
integral.
Example (Stewart, Section 13.2, Exercise 20) We evaluate the line integral
Z
F dr
C

112

CHAPTER 3. VECTOR CALCULUS

where F(x, y, z) = hz, y, xi and C is the curve defined by the parametric

vector equation
r(t) = hx(t), y(t), z(t)i = ht, sin t, cos ti,

0 t .

We have
Z

F dr =
C

F(r(t)) r0 (t) dt

hz(t), y(t), x(t)i hx0 (t), y 0 (t), z 0 (t)i dt

Z0
hcos t, sin t, ti h1, cos t, sin ti dt

=
0

[cos t + sin t cos t + t sin t] dt

Z
Z
=
cos t dt +
sin t cos t dt +
t sin t dt
0
0
0

Z

1
= sin t|0 + sin2 t t cos t|0 +
cos t dt
2
0
0
= .
Z0

2
If we write F(x, y) = hP (x, y), Q(x, y)i, where P and Q are the component functions of F, then we have
Z
Z b
F dr =
F(r(t)) r0 (t) dt
C
a
Z b
=
hP (x(t), y(t)), Q(x(t), y(t))i hx0 (t), y 0 (t)i dt
a
Z b
Z b
0
=
P (x(t), y(t))x (t) dt +
Q(x(t), y(t))y 0 (t) dt.
a

When the curve is approximated by n line segments, as before, the difference

in the x-coordinates of each segment is, by the Mean Value Theorem,
xi = x(ti ) x(ti1 ) x0 (ti ) t,
where ti1 ti ti . For this reason, we write
Z

P (x(t), y(t))x (t) dt =

a

P dx,
C

3.2. LINE INTEGRALS

Z

113

Q dy,

Q(x(t), y(t))y (t) dt =

C

and conclude

Z
F dr =

P dx + Q dy.
C

These line integrals of scalar-valued functions can be evaluated individually to obtain the line integral of the vector field F over C. However, it is
important to note that unlike line integrals with respect to the arc length s,
the value of line integrals with respect to x or y (or z, in 3-D) depends on the
orientation of C. If the curve is traced in reverse (that is, from the terminal
point to the initial point), then the sign of the line integral is reversed as
well. We denote by C the curve C with its orientation reversed. We then
have
Z
Z
F dr =
F dr,
C

and

Z
P dx =

Z
Q dy =

P dx,
C

Z
Q dy.
C

All of this discussion generalizes to space curves (that is, curves in 3-D) in
a straightforward manner, as illustrated in the examples.
Example (Stewart, Section 13.2, Exercise 6) Let F(x, y) = hsin x, cos yi and
let C be the curve that is the top half of the circle x2 + y 2 = 1, traversed
counterclockwise from (1, 0) to (1, 0), and the line segment from (1, 0) to
(2, 3). To evaluate the line integral
Z
Z
F T ds =
sin x dx + cos y dy,
C

we consider the integrals over the semicircle, denoted by C1 , and the line
segment, denoted by C2 , separately. We then have
Z
Z
Z
sin x dx + cos y dy =
sin x dx + cos y dy +
sin x dx + cos y dy.
C

C1

C2

For the semicircle, we use the parametric equations

x = cos t,

y = sin t,

This yields
Z
Z
sin x dx + cos y dy =
C1

0 t .

sin(cos t)( sin t) dt + cos(sin t) cos t dt

114

CHAPTER 3. VECTOR CALCULUS

cos(cos t)|0 + sin(sin t)|0

= cos(1) + cos(1)
= 0.
For the line segment, we use the parametric equations
x = 1 t,

y = 3t,

0 t 1.

This yields
Z

sin(1 t)(1) dt + cos(3t)(3) dt

sin x dx + cos y dy =
C2

cos(1 t)|10 + sin(3t)|10

= cos(2) + cos(1) + sin(3) sin(0)

= cos(2) + cos(1) + sin(3).
We conclude
Z
sin x dx + cos y dy = cos(1) cos(2) + sin(3).
C

In evaluating these integrals, we have taken advantage of the rule

Z

f 0 (g(t))g 0 (t) dt = f (g(b)) f (g(a)),

from the Fundamental Theorem of Calculus and the Chain Rule. However,
this shortcut can only be applied when an integral involves only one of the
independent variables. 2
Example (Stewart, Section 13.2, Exercise 12) We evaluate the line integral
Z
F dr
C

where
F(x, y, z) = hP (x, y, z), Q(x, y, z), R(x, y, z)i = hz, x, yi,
and C is defined by the parametric equations
x = t2 ,

y = t3 ,

z = t2 ,

0 t 1.

3.3. THE FUNDAMENTAL THEOREM FOR LINE INTEGRALS

115

We have
Z

Z
F dr =

P dx + Q dy + R dz

C
Z 1

z(t)x0 (t) dt + x(t)y 0 (t) dt + y(t)z 0 (t) dt

0
1

t2 (2t) dt + t2 (3t2 ) dt + t3 (2t) dt

=
0
1

2t3 dt + 3t4 dt + 2t4 dt

=
0
1

(5t4 + 2t3 ) dt

=
0


=

1
t5
t4 
5 +2
5
4 0

3
.
2

=
2

3.3

The Fundamental Theorem for Line Integrals

We have learned that the line integral of a vector field F over a curve
piecewise smooth C, that is parameterized by a vector-valued function r(t),
a t b, is given by
Z b
Z
F dr =
F(r(t)) r0 (t) dt.
C

Now, suppose that F continuous, and is a conservative vector field; that is,
F = f for some scalar-valued function f . Then, by the Chain Rule, we
have
Z
Z b
Z b
d
F dr =
f (r(t))r0 (t) dt =
[(f r)(t)] dt = (f r)(t)|ba = f (r(b))f (r(a)).
dt
C
a
a
This is the Fundamental Theorem of Line Integrals, which is a generalization
of the Fundamental Theorem of Calculus.
If the curve C is a closed curve; that is, the initial and terminal points
of C are the same, then r(b) = r(a), which yields
Z
F dr = f (r(b)) f (r(a)) = 0.
C

116

CHAPTER 3. VECTOR CALCULUS

If we decompose C into two curves C1 and C2 , and use the fact that the sign
of the line integral of a vector field over a curve depends on the orientation
of the curve, then we have
Z
Z
Z
Z
Z
F dr = 0.
F dr
F dr =
F dr +
F dr =
C

C1

C2

C1

C2

That is,
Z

F dr.

F dr =
C1

C2

However, C1 and C2 have the same initial and terminal points. It follows
that if F is conservative within an open, connected domain D (so that any
two points in D can be connected by a path that lies within D), then the
line integral of F is independent of path in D; that is, the value of the line
integral of F over a path C depends only on its initial and terminal points.
The converse of this statement is also true: if the line integral of a
vector field F is independent of path within an open, connected domain
D, then F is a conservative vector field on D. To see this, we consider
the two-variable case and let D be a region in R2 . Furthermore, we let
F(x, y) = hP (x, y), Q(x, y)i. We choose an arbitrary point (a, b) D, and
define
Z
(x,y)

F dr.

f (x, y) =
(a,b)

Since this line integral is independent of path, we can define f (x, y) using
any path between (a, b) and (x, y) that we choose, knowing that its value at
(x, y) will be the same in any case.
By choosing a path that ends with a horizontal line segment from (x1 , y)
to (x, y) contained entirely in D, parametrized by x = t, y = y, for x1 t
x, we can show that
"Z
#
"Z
#
(x1 ,y)
(x,y)

(x, y) =
F dr +
F dr
x
x (a,b)
x (x1 ,y)
Z x


0
0
= 0+
P (x(t), y)x (t) dt + Q(x(t), y)y (t) dt
x x1
Z x


=
P (t, y) dt + 0
x x1
= P (x, y).
Using a similar argument, we can show that f /y = Q. We have thus
shown that F is conservative, and conclude that F is a conservative vector
field if and only if its line integral is independent of path.

3.3. THE FUNDAMENTAL THEOREM FOR LINE INTEGRALS

117

However, in order to use the Fundamental Theorem of Line Integrals to

evaluate the line integral of a conservative vector field, it is necessary to
obtain the function f such that f = F . Furthermore, the theorem cannot
be applied to a vector field that is not conservative, so we need to be able
to confirm that a given vector field is conservative before we proceed.
Continuing to restrict ourselves to the two-variable case, suppose that
F = hP, Qi is a conservative vector field defined on a domain D, and that
P and Q have continuous first partial derivatives. Then, we have
f
= P,
x

f
= Q,
y

for some function f . It follows that

P
2f
=
,
y
yx

Q
2f
=
.
x
xy

However, by Clairauts Theorem, these mixed second partial derivatives of

f are equal, so it follows that
P
Q
=
y
x
if F = hP, Qi is conservative.
If the domain D is simply connected, meaning that any region enclosed
by a closed curve in D contains only points in D (informally, D has no
holes), then the converse is true: if
P
Q
=
y
x
in D, then F = hP, Qi is a conservative vector field. Similarly, if F =
hP, Q, Ri is a vector field defined on a simply connected domain D R3 ,
and
Q
P
R
Q
R
P
=
,
=
,
=
,
y
x
z
x
z
y
then F is conservative.
It remains to be able to find the function f such that f = F for a
given vector field F = hP, Qi that is known to be conservative. The general
technique is as follows:
Integrate P (x, y) with respect to x to obtain
f (x, y) = f1 (x, y) + g(y),

118

CHAPTER 3. VECTOR CALCULUS

where f1 (x, y) is obtained by anti-differentiation of P (x, y), and g(y) is
an unknown function that plays the role of the constant of integration,
since f (x, y) is obtained by anti-differentiating with respect to x.
Differentiate f with respect to y to obtain

[f1 (x, y)] + g 0 (y) = Q(x, y),

y
and solve for g 0 (y).
Integrate g 0 (y) with respect to y to complete the definition of f (x, y),
up to a constant of integration.

A similar procedure can be used for a vector field defined on R3 , except that
the function g depends on both y and z, and differentiation with respect to
both y and z is needed to completely define the function f (x, y, z) such that
f = F.
Example (Stewart, Section 13.3, Exercise 14) Let
F(x, y, z) = hP (x, y, z), Q(x, y, z), R(x, y, z)i = h2xz + y 2 , 2xy, x2 + 3z 2 i.
To confirm that F is conservative, we check the appropriate first partial
derivatives of P , Q and R:
Py = 2y = Qx ,

Pz = 2x = Rx ,

Qz = 0 = Ry .

Now, to find a function f (x, y, z) such that f = F, which must satisfy

fx = P , we integrate P (x, y, z) with respect to x and obtain
f (x, y, z) = x2 z + y 2 x + g(y, z).
Differentiating with respect to y and z yields the equations
fy (x, y, z) = 2xy + gy (y, z) = Q(x, y, z) = 2xy,
fz (x, y, z) = x2 + gz (y, z) = R(x, y, z) = x2 + 3z 2 .
It follows that
gy (y, z) = 0,

gz (y, z) = 3z 2 ,

which yields
g(y, z) = z 3 + K

3.3. THE FUNDAMENTAL THEOREM FOR LINE INTEGRALS

119

for some constant K. We conclude that F = f where

f (x, y, z) = x2 z + y 2 x + z 3 + K
where K is an arbitrary constant.
To evaluate the line integral of F over the curve C parametrized by
x = t2 ,

y = t + 1,

z = 2t 1,

0 t 1,

we apply the Fundamental Theorem of Line Integrals and obtain

Z
F dr = f (x(1), y(1), z(1)) f (x(0), y(0), z(0))
C

= f (1, 2, 1) f (0, 1, 1)
= 12 (1) + 22 (1) + 13 + K (02 (1) + 12 (0) + (1)3 + K)
= 1 + 4 + 1 + K (0 + 0 1 + K)
= 7.
2
Let F represent a force field. Then, recall that the work done by the
force field to move an object along a path r(t), a t b, is given by the
line integral
Z
Z b
W =
F dr =
F(r(t)) r0 (t) dt.
C

From Newtons Second Law of Motion, we have

F(r(t)) = mr00 (t),
where m is the mass of the object, and r00 (t) = a(t) is its acceleration. We
then have
Z b
W =
mr00 (t) r0 (t) dt
a
Z b
1
d 0
=
m
[r (t) r0 (t)] dt
2
a dt
Z b
1
d
=
m
[kr0 (t)k2 ] dt
2
dt
a
1
1
=
mkv(b)k2 mkv(a)k2
2
2
where v(t) = r0 (t) is the velocity of the object.

120

CHAPTER 3. VECTOR CALCULUS

It follows that
W = K(B) K(A),

where A = r(a) and B = r(b) are the initial and terminal points, respectively, and
1
K(P ) = mkv(t)k, r(t) = P,
2
is the kinetic energy of the object at the point P . That is, the work done
by the force field along C is the change in the kinetic energy from point A
to point B.
If F is also a conservative force field, then F = P , where P is the
potential energy. It follows from the Fundamental Theorem of Line Integrals
that
Z
Z
W =
F dr =
P dr = [P (B) P (A)].
C

We conclude that
P (A) + K(A) = P (B) + K(B).
That is, when an object is moved by a conservative force field, then its
total energy remains constant. This is known as the Law of Conservation
of Energy.

3.4

Greens Theorem

We have learned that if a vector field is conservative, then its line integral
over a closed curve C is equal to zero. However, if this is not the case, then
evaluation of a line integral using the formula
Z

Z
F dr =

F(r(t)) r0 (t) dt,

where r(t) is a parameterization of C, can be very difficult, even if C is a

relatively simple plane curve. Fortunately, in this case, there is an alternative
approach, using a result known as Greens Theorem.
We assume that F = hP, Qi, and consider the case where C encloses a
region D that can be viewed as a region of either type I or type II. That is,
D has the definitions
D = {(x, y) | a x b, g1 (x) y g2 (x)}

3.4. GREENS THEOREM

121

and
D = {(x, y) | c y d, h1 (y) x h2 (y)}.
Using the first definition, we have C = C1 C2 (C3 ) (C4 ), where:
C1 is the curve with parameterization x = t, y = g1 (t), for a t b
C2 is the vertical line segment with parameterization x = b, y = t, for
g1 (b) t g2 (b)
C3 is the curve with parameterization x = t, y = g2 (t), for a t b
C4 is the vertical line segment with parameterization x = a, y = t, for
g1 (a) t g2 (a)
We use positive orientation to describe the curve C, which means that the
curve is traversed counterclockwise. This means that as the curve is traversed, the region D is on the left.
In view of
Z
Z
F dr =
P dx + Q dy,
C

we have
Z
Z
Z
Z
Z
P dx =
P dx +
P dx +
P dx +
P dx
C
ZC1
ZC2
ZC3
Z C4
=
P dx +
P dx
P dx
P dx
C1
b

Z
=

C2

C3

P (x(t), y(t))x0 (t) dt +

P (x(t), y(t))x0 (t) dt

a
b

Z
a

Z
=

C4

g2 (b)

g1 (b)
g2 (a)

P (b, t)(0) dt
g1 (b)

g2 (a)

P (t, g2 (t))(1) dt
a

P (a, t)(0) dt
g1 (a)

[P (t, g1 (t)) P (t, g2 (t))] dt

P (x(t), y(t))x0 (t) dt

g1 (a)
g2 (b)

P (t, g1 (t))(1) dt +
a

P (x(t), y(t))x0 (t) dt

Z bZ

g2 (t)

Py (t, y) dy dt
a

g1 (t)

Z Z
=
D

P
dA.
y

122

CHAPTER 3. VECTOR CALCULUS

Using a similar approach in which D is viewed as a region of type II, we

obtain
Z
Z Z
Q
Q dy =
dA.
C
D x
Putting these results together, we obtain Greens Theorem, which states
that if C is a positively oriented, piecewise smooth, simple (that is, not
self-intersecting) closed curve that encloses a region D, and P and Q are
functions that have continuous first partial derivatives on D, then

Z Z 
Z
Q P
dA.
P dx + Q dy =

x
y
D
C
Another common statement of the theorem is

Z Z 
Z
Q P

dA =
P dx + Q dy,
x
y
D
D
where D denotes the positively oriented boundary of D.
This theorem can be used to find a simpler approach to evaluating a
line integral of the vector field hP, Qi over C by converting the integral to
a double integral over D, or it can be used to find a simpler approach to
evaluating a double integral over a region D by converting it into an integral
over its boundary.
To show that Greens Theorem applies for more general regions than
those that are of both type I and type II, we consider a region D that is
the union of two regions D1 and D2 that are of both type I and type II. Let
C be the positively oriented boundary of D, let D1 have positively oriented
boundary C1 C3 , and let D2 have positively oriented boundary C2 (C3 ),
where C3 is the boundary between D1 and D2 . Then, C = C1 C2 . It follows
that for functions P and Q that satisfy the assumptions of Greens Theorem
on D, we can apply the theorem to D1 and D2 individually to obtain


Z Z 
Z Z 
Q P
Q P

dA =

dA +
x
y
x
y
D
D1


Z Z
Q P

dA
x
y
D2
Z
Z
=
P dx + Q dy +
P dx + Q dy
C1 C3
C2 (C3 )
Z
Z
=
P dx + Q dy +
P dx + Q dy +
C1
C3
Z
Z
P dx + Q dy +
P dx + Q dy
C2

C3

3.4. GREENS THEOREM

123
Z

P dx + Q dy +

P dx + Q dy +

C2

C1

Z
P dx + Q dy
=

P dx + Q dy
ZC3

ZC3
P dx + Q dy +
C1

P dx + Q dy
C2

Z
=

P dx + Q dy
ZC1 C2
P dx + Q dy.
=
C

We conclude that Greens Theorem holds on D1 D2 . The same argument

can be used to easily show that Greens Theorem applies on any finite union
of simple regions, which are regions of both type I and type II.
Greens Theorem can also be applied to regions with holes, that is,
regions that are not simply connected. To see this, let D be a region enclosed
by two curves C1 and C2 that are both positively oriented with respect to D
(that is, D is on the left as either C1 or C2 is traversed). Let C2 be contained
within the region enclosed by C1 ; that is, let C2 be the boundary of the
hole in D. Then, we can decompose D into two simply connected regions
D0 and D00 by connecting C2 to C1 along two separate curves that lie within
D. Applying Greens Theorem to D0 and D00 individually, we find that the
line integrals along the common boundaries of D0 and D00 cancel, because
they have opposite orientations with respect to these regions. Therefore, we
have


Z Z 
Z Z 
Q P
Q P

dA =

dA +
x
y
x
y
D
D0

Z Z 
Q P

dA
x
y
D00
Z
Z
=
P dx + Q dy +
P dx + Q dy
C2
ZC1
=
P dx + Q dy.
C1 C2

Therefore, Greens Theorem applies to D as well.

Example The vector field

F(x, y) = hP (x, y), Q(x, y)i =

y
x
2
, 2
2
x + y x + y2

124

CHAPTER 3. VECTOR CALCULUS

is conservative on all of R2 except at the origin, because it is not defined

there. Specifically, F = f where
y
f (x, y) = tan1 .
x
Now, consider a region D that is enclosed by a positively oriented, piecewise
smooth, simple closed curve C, and also has a hole that is a disk of
radius a, centered at the origin, and contained entirely within C. Let C 0 be
the positively oriented boundary of this disk. Then, the boundary of D is
C (C 0 ), because, as a portion of the boundary of D, rather than the disk,
it is necessary for C 0 to switch orientation. Applying Greens Theorem to
compute the line integral of F over the boundary of D yields

Z
Z
Z Z 
Q P
P dx + Q dy +
P dx + Q dy =

dA = 0,
x
y
C
C 0
D
since F is conservative on D. It follows that
Z
Z
Z
F dr =
F dr =
C 0

F dr,

C0

so we can compute the line integral of F over C, which we have not specified,
by computing the line integral over the circle C 0 , which can be parameterized
by x = a cos t, y = a sin t, for 0 t 2. This yields
Z
Z 2
F dr =
P (x(t), y(t))x0 (t) dt + Q(x(t), y(t))y 0 (t) dt
0
C
0

Z 2 
a sin t
=

(a sin t) dt +
(a cos t)2 + (a sin t)2
0


a cos t
(a cos t) dt
(a cos t)2 + (a sin t)2
Z 2
a2 sin2 t
a2 cos2 t
=
dt
+
dt
a2 cos2 +a2 sin2 t
a2 cos2 t + a2 sin2 t
0
Z 2
=
1 dt
0

= 2.
We conclude that the line integral of F over any positively oriented, piecewise
smooth, simple closed curve that encloses the origin is equal to 2. 2
Example Consider a n-sided polygon P with vertices (x1 , y1 ), (x2 , y2 ), . . .,
(xn , yn ). The area A of the polygon is given by the double integral
Z Z
A=
1 dA.
P

3.5. CURL AND DIVERGENCE

125

Let P (x, y) = y/2 and Q(x, y) = x/2. Then


 
 
Q P
1
1
=
= 1.

x
y
2
2
It follows from Greens Theorem that if P is positively oriented, then
Z
Z
1
A=
Q dy + P dx =
x dy y dx.
2 P
P
To evaluate this line integral, we consider each edge of P individually. Let
C be the line segment from (x1 , y1 ) to (x2 , y2 ), and assume, for convenience,
that C is not vertical. Then C can be parameterized by x = t, y = mx + b,
for x1 x x2 , where
m=

y2 y1
,
x2 x1

b = y1 mx1 .

We then have
Z

x2

x dy y dx =
C

mt dt (mt + b) dt
x1

x2

b dt
x1

= b(x1 x2 )
= y1 (x1 x2 ) mx1 (x1 x2 )
= y1 (x1 x2 ) + (y2 y1 )x1
= x1 y2 x2 y1 .
We conclude that
A =

1
[(x1 y2 x2 y1 ) + (x2 y3 x3 y2 ) + + (xn1 yn xn yn1 )+
2
(xn y1 x1 yn )] .

3.5

Curl and Divergence

We have seen two theorems in vector calculus, the Fundamental Theorem of

Line Integrals and Greens Theorem, that relate the integral of a set to an
integral over its boundary. Before establishing similar results that apply to

126

CHAPTER 3. VECTOR CALCULUS

surfaces and solids, it is helpful to introduce new operations on vector fields

that will simplify exposition.
We have previously learned that a vector field F = hP, Q, Ri defined on
R3 is conservative if
Ry Qz = 0,

Pz Rx = 0,

Qx Py = 0.

These equations are equivalent to the statement




hP, Q, Ri = h0, 0, 0i.
, ,
x y z
Therefore, we define the curl of a vector field F = hP, Q, Ri by
curl F = F,
where


=

, ,
x y z


.

From the definition of a conservative vector field, it follows that curl F = 0 if

F = f where f has continuous second partial derivatives, due to Clairauts
Theorem. That is, the curl of a gradient is zero.
This is equivalent to the statement that the curl of a conservative vector
field is zero. The converse, that a vector field F for which curl F = 0 is
conservative, is also true if F has continuous first partial derivatives and
curl F = 0 within a simply connected domain. That is, the domain must not
have holes.
When F represents the velocity field of a fluid, the fluid tends to rotate
around the axis that is aligned with curl F, and the magnitude of curl F
indicates the speed of rotation. Therefore, when curl F = 0, we say that F
is irrotational, which is a term that has previously been associated with the
equivalent condition of F being conservative.
Another operation that is useful for discussing properties of vector fields
is the divergence of a vector field F, denoted by div F. It is defined by
div F = F.
For example, if F = hP, Q, Ri, then



div F =
, ,
hP, Q, Ri = Px + Qy + Rz .
x y z

3.5. CURL AND DIVERGENCE

127

Unlike the curl, the divergence is defined for vector fields with any number of
variables, as long as the number of independent and the number of dependent
variables are the same.
It can be verified directly that if F is the curl of a vector field G, then
div F = 0. That is, the divergence of any curl is zero, as long as G has
continuous second partial derivatives. This is useful for determining whether
a given vector field F is the curl of any other vector field G, for if it is, its
divergence must be zero.
Example (Stewart, Section 13.5, Exercise 18) The vector field F(x, y, z) =
hyz, xyz, xyi is not the curl of any vector field G, because
div F = (yz)x + (xyz)y + (xy)z = 0 + xz + 0 = xz,
whereas if F = curl G, then
div F = div curl G = 0.
2
If F represents the velocity field of a fluid, then, at each point within the
fluid, div F measures the tendency of the fluid to diverge away from that
point. Specifically, the divergence is the rate of change, with respect to time,
of the density of the fluid. Therefore, if div F = 0, then we say that F, and
therefore the fluid as well, is incompressible.
The divergence of a gradient is

 


f f f
2f
2f
2f
div(f ) = f =
, ,

,
,
=
+
+
.
x y z
x y z
x2
y 2
z 2
We denote this expression f by 2 f , or f , which is called the Laplacian of f . The operator 2 is called the Laplace operator. Its name comes
from Laplaces equation
f = 0.
The curl and divergence can be used to restate Greens Theorem in
forms that are more directly generalizable to surfaces and solids in R3 . Let
F = hP, Q, 0i, the embedding of a two-dimensional vector field in R3 . Then


Q P
curl F =

k,
x
y
where, as before, k = h0, 0, 1i. It follows that




Q P
Q P
curl F k =

kk=

.
x
y
x
y

128

CHAPTER 3. VECTOR CALCULUS

This expression is called the scalar curl of the two-dimensional vector field
hP, Qi. We conclude that Greens Theorem can be rewritten as
Z Z

(curl F) k dA.

F dr =
D

Another useful form of Greens Theorem involves the divergence. Let

F = hP, Qi have continuous first partial derivatives in a domain D with
a positively oriented, piecewise smooth boundary C that has parametrization r(t) = hx(t), y(t)i, for a t b. Using the original form of Greens
Theorem, we have
Z Z 

Z Z
div F dA =
D

P
Q
+
x
y


dA

Z
P dy Q dx

=
C
Z b

P (x(t), y(t))y 0 (t) dt Q(x(t), y(t))x0 (t) dt


Z b
x0 (t)
y 0 (t)
+ Q(x(t), y(t)) 0
kr0 (t)k dt
=
P (x(t), y(t)) 0
kr
(t)k
kr
(t)k
a
Z b
=
(F n)(t)kr0 (t)k dt
Za
=
F n ds

where
n(t) =

1
kr0 (t)k

hy 0 (t), x0 (t)i

is the outward unit normal vector to the curve C. Note that nT = 0, where
T is the unit tangent vector
T(t) =

1
kr0 (t)k

hx0 (t), y 0 (t)i.

We have established a third form of Greens Theorem,

Z Z

Z
F n ds =
C

div F dA.
D

3.6. PARAMETRIC SURFACES AND THEIR AREAS

3.6

129

Parametric Surfaces and Their Areas

We have learned that Greens Theorem can be used to relate a line integral
of a two-dimensional vector field F over a closed plane curve C to a double
integral of a component of curl F over the region D that is enclosed by
C. Our goal is to generalize this result in such a way as to relate the line
integral of a three-dimensional vector field F over a closed space curve C to
the integral of a component of curl F over a surface enclosed by C.
We have also learned that Greens Theorem relates the integral of the
normal component of a two-dimensional vector field over a closed curve C to
the double integral of div F over the region D that it encloses. We wish to
generalize this result in order to relate the integral of the normal component
of a three-dimensional vector field F over a closed surface S to the triple
integral of div F over the solid E contained within S.
In order to realize either of these generalizations, we need to be able to integrate functions over piecewise smooth surfaces, just as we now know how to
integrate functions over piecewise smooth curves. Whereas a smooth curve
C, being a curved one-dimensional entity, is most conveniently described by
a parameterization r(t), where a t b and r(t) is a differentiable function
of one variable, a smooth surface S, being a curved two-dimensional entity,
is most conveniently described by a parametrization r(u, v), where (u, v) lies
within a 2-D region, and r(u, v) = hx(u, v), y(y, v), z(u, v)i is a differentiable
function of two variables. We say that S is a parametric surface, and
x = x(u, v),

y = y(u, v),

z = z(u, v)

are the parametric equations of S.

Example The M
obius strip is a surface that is famous for being a nonorientable surface; that is, it has only one side. It can be parameterized
by
v
u
cos
cos u,
2
2

v
u
y(u, v) =
1 + cos
sin u,
2
2
v
u
z(u, v) =
sin ,
2
2

x(u, v) =

1+

where 0 u 2 and 1 v 1. It is shown in Figure 3.3. 2

Example The paraboloid defined by the equation x = 4y 2 + 4z 2 , 0 x 4,

130

CHAPTER 3. VECTOR CALCULUS

Figure 3.3: The Mobius strip

can also be defined by the parametric equations

x
x
x = x, y =
cos , z =
sin ,
2
2
where 0 2, since for each x, a point (x,
py, z) on the paraboloid must
lie on a circle centered at (x, 0, 0) with radius x/4, parallel to the yz-plane.
This is an example of a surface of revolution, since the surface is obtained
by revolving the curve y = f (x) around the x-axis. 2
Let P0 = (x0 , y0 , z0 ) = r(u0 , v0 ) be a point on a parametric surface S. A
curve defined by g(v) = r(u0 , v) that lies within S and passes through P0
has the tangent vector


x
y
z
0
rv = g (v) =
(u0 , v0 ), (u0 , v0 ), (u0 , v0 )
v
v
v
at P0 . Similarly, the tangent vector at P0 of the curve h(u) = r(u, v0 ), that
also lies within S and passes through P0 , is


x
y
z
0
ru = h (u) =
(u0 , v0 ),
(u0 , v0 ),
(u0 , v0 ) .
u
u
u

3.6. PARAMETRIC SURFACES AND THEIR AREAS

131

If these vectors are not parallel, then together they define the tangent plane
of S at P0 . Its normal vector is
ru rv = ha, b, ci
which yields the equation
a(x x0 ) + b(y y0 ) + c(z z0 ) = 0
of the tangent plane.
Example (Stewart, Section 13.6, Exercise 30) Consider the surface defined
by the parametric equations
x = u2 ,

y = v2,

z = uv,

0 u, v 10.

At the point (x0 , y0 , z0 ) = (1, 1, 1), which corresponds to u0 = 1, v0 = 1,

the equation of the tangent plane can be obtained by first computing the
partial derivatives of the coordinate functions. We have
xu = 2u,
xv = 0,

yu = 0,

zu = v,

yv = 2v,

zv = u.

Evaluating at (u0 , v0 ) yields

ru = hxu , yu , zu i = h2, 0, 1i,

rv = hxv , yv , zv i = h0, 2, 1i.

It follows that the normal to the tangent plane is

n = ru rv = h2, 0, 1i h0, 2, 1i = h2, 2, 4i.
We conclude that the equation of the tangent plane is
2(x 1) 2(y 1) + 4(z 1) = 0.
2
The vectors ru and rv are helpful for computing the area of a smooth
surface S. For simplicity, we assume that S is parametrized by a function
r(u, v) with domain D, where D = [a, b][c, d] is a rectangle in the uv-plane.
We divide [a, b] into n subintervals [ui1 , ui ] of width u = (b a)/n, and
divide [c, d] into m subintervals [vj1 , vj ] of width v = (d c)/m.

132

CHAPTER 3. VECTOR CALCULUS

Then, r approximately maps the rectangle Rij with lower left corner
(ui1 , vj1 ) into a parallelogram with adjacent edges defined by the vectors
r(ui , vj1 ) r(ui1 , vj1 ) ru u
and
r(ui1 , vj ) r(ui1 , vj1 ) rv v.
The area of this parallelogram is
Aij = kru rv kuv.
Adding all of these areas approximates the area of S, which we denote by
A(S). If we let m, n , we obtain
A(S) =

lim

m,n

n X
m
X
i=1 j=1

Z Z
kru rv k dA.

Aij =
D

Example (Stewart, Section 13.6, Exercise 34) We wish to find the area of
the surface S that is the part of the plane 2x + 5y + z = 10 that lies inside
the cylinder x2 + y 2 = 9. First, we must find parametric equations for this
surface. Because x and y are restricted to the circle of radius 3 centered at
the origin, it makes sense to use polar coordinates for x and y. We then
have the parametric equations
x = u cos v,

y = u sin v,

z = 10 u(2 cos v + 5 sin v),

where 0 u 3 and 0 v 2. We then have

ru = hxu , yu , zu i = hcos v, sin v, 2 cos v 5 sin vi,
rv = hxv , yv , zv i = hu sin v, u cos v, u(2 sin v 5 cos v)i.
We then have

kru rv k = kh2u, 5u, uik = |u| 30.

It follows that
Z

3 Z 2

A(S) =
0

u 30 du dv = 2 30

u du = 9 30.

It should be noted that it is to be expected that the direction of ru rv is

parallel to the normal vector of the plane 2x + 5y + z = 10, since it is normal
to the surface at every point. 2

3.6. PARAMETRIC SURFACES AND THEIR AREAS

133

Often, a surface is defined to be the graph of a function z = f (x, y).

Such a surface can be parametrized by
x = u,

y = v,

z = f (u, v),

(u, v) D.

It follows that
ru = h1, 0, fu i,

rv = h0, 1, fv i.

We then have rv ru = hfu , fv , 1i, which yields the equation of the tangent
plane
f
f
(u0 , v0 )(x x0 ) +
(u0 , v0 )(y y0 ) = z z0 ,
u
v
which, using the relations x = u and y = v, can be rewritten as
f
f
(x0 , y0 )(x x0 ) +
(x0 , y0 )(y y0 ) = z z0 .
x
y
Recall that this is the equation of the tangent plane of a surface defined by
an equation of the form z = f (x, y) that had been previously defined. It
follows that the area of such a surface is given by the double integral
s
 2  2
Z Z
df
df
A(S) =
+
dA.
1+
dx
dy
D
Example (Stewart, Section 13.6, Exercise 38) To find the area A(S) of the
surface z = 1 + 3x + 2y 2 that lies above the triangle with vertices (0, 0),
(0, 1) and (2, 1), we compute
z
= 3,
x

z
= 4y,
y

and then evaluate the double integral

s
 2  2
Z 1 Z 2y
z
z
A(S) =
1+
+
dx dy
x
y
0
0
Z 1 Z 2y p
=
10 + 16y 2 dx dy
0
0
Z 1 p
=
2y 10 + 16y 2 dy
0
Z 26
1
=
u1/2 du
16 10

134

CHAPTER 3. VECTOR CALCULUS


1 3/2 26
=
u 
24
10
1
3/2
(26 103/2 )
=
24
4.206.

2
A surface of revolution S that is obtained by revolving the curve y =
f (x), a x b, around the x-axis has parametric equations
x = u,

y = f (u) cos v,

z = f (u) sin v,

where a u b and 0 v 2. From these equations, we obtain

p
kru rv k = |f (u)| 1 + [f 0 (u)]2 ,
which yields
Z

A(S) = 2

p
|f (u)| 1 + [f 0 (u)]2 du.

If y = f (x) is revolved around the y-axis instead, then the area is

Z b p
A(S) = 2
|u| 1 + [f 0 (u)]2 du,
a

which can be obtained by considering the case of revolving x = f 1 (y)

around the y-axis and proceeding with a parametrization similar to the case
of revolving around the x-axis.

3.7
3.7.1

Surface Integrals
Surface Integrals of Scalar-Valued Functions

Previously, we have learned how to integrate functions along curves. If a

smooth space curve C is parameterized by a function r(t) = hx(t), y(t), z(t)i,
a t b, then the arc length L of C is given by the integral
Z b
kr0 (t)k dt.
a

Similarly, the integral of a scalar-valued function f (x, y, z) along C is given

by
Z
Z b
f ds =
f (x(t), y(t), z(t))kr0 (t)k dt.
C

3.7. SURFACE INTEGRALS

135

It follows that the integral of f (x, y, z) 1 along C is equal to the arc length
of C.
We now define integrals of scalar-valued functions on surfaces in an analogous manner. Recall that the area of a smooth surface S, parametrized by
r(u, v) = hx(u, v), y(u, v), z(u, v)i for (u, v) D, is given by the integral
Z Z
A(S) =
kru rv k du dv.
D

To integrate a scalar-valued function f (x, y, z) over S, we assume for simplicitly that D is a rectangle, and divide it into sub-rectangles {Rij } of
dimension u and v, as we did when we derived the formula for A(S).
Then, the function r maps each sub-rectangle Rij into a surface patch Sij
that has area Sij . This area is then multiplied by f (Pij ), where Pij is any
point on Sij .
Letting u, v 0, we obtain the surface integral of f over S to be
n X
m
X

Z Z
f (x, y, z) dS =
S

lim

u,v0

f (Pij ) Sij

i=1 j=1

Z Z
f (r(u, v))kru rv k du dv,

=
D

since, in the limit as u, v 0, we have

Sij kru rv k u v.
Note that if f (x, y, z) 1, then the surface integral of f over S yields the
area of S, A(S).
Example (Stewart, Section 13.7, Exercise 6) Let S be the helicoid with
parameterization
r(u, v) = hu cos v, u sin v, vi,

0 u 1,

0 v .

Then we have
ru = hcos v, sin v, 0i,

rv = hu sin v, u cos v, 1i,

which yields
kru rv k = khsin v, cos v, uik =

sin2 v + cos2 v + u2 =

p
1 + u2 .

136

CHAPTER 3. VECTOR CALCULUS

It follows that
Z 1Z p
Z Z p
2
2
1 + x + y dS =
1 + (u cos v)2 + (u sin v)2 kru rv k dv du
0
0
S
Z 1Z p
p
=
1 + u2 1 + u2 dv du
0
0
Z 1Z
1 + u2 dv du
=
0
0
Z 1
1 + u2 du
=
0


1
u3 
= u+
3 0
4
=
.
3
2
The surface integral of a scalar-valued function is useful for computing
the mass and center of mass of a thin sheet. If the sheet is shaped like a
surface S, and it has density (x, y, z), then the mass is given by the surface
integral
Z Z
(x, y, z) dS,

m=
S

and the center of mass is the point

Z
1
x
=
m
Z
1
y =
m
Z
1
z =
m

3.7.2

(
x, y, z), where
Z
x(x, y, z) dS,
ZS
y(x, y, z) dS,
ZS
z(x, y, z) dS.
S

Surface Integrals of Vector Fields

Let v be a vector field defined on R3 that represents the velocity field of a

fluid, and let be the density of the fluid. Then, the rate of flow of the
fluid, which is defined to be the rate of change with respect to time of the
amount of fluid (mass), per unit area, is given by v.
To determine the total amount of fluid that is crossing S per unit of
time, called the flux across S, we divide S into several small patches Sij ,

3.7. SURFACE INTEGRALS

137

as we did when we defined the surface integral of a scalar-valued function.

Since each patch Sij is approximately planar (that is, parallel to a plane),
we can approximate the flux across Sij by
(v n)A(Sij ),
where n is a unit vector that is normal (perpendicular) to Sij . This is
because if is the angle between Sij and the direction of v, then the fluid
directed at Sij is effectively passing through a region of area A(Sij )| cos |.
If we sum the flux over each patch, and let the areas of the patches
approach zero, then we obtain the total flux across S,
Z Z
(x, y, z)v(x, y, z) n(x, y, z) dS,
S

where n(x, y, z) is a continuous function that describes a unit normal vector

at each point (x, y, z) on S. For a general vector field F, we define the
surface integral of F over S by
Z Z
Z Z
F dS =
F n dS.
S

When F represents an electric field, we call the surface integral of F over S

the electric flux of F through S. Alternatively, if F = Ku, where u is
a function that represents temperature and K is a constant that represents
thermal conductivity, then the surface integral of F over a surface S is called
the heat flow or heat flux across S.
If S is parameterized by a function r(u, v), where (u, v) D, then
n=

ru rv
,
kru rv k

and we then have

Z Z
Z Z
ru rv
F dS =
F
dS
kru rv k
S
Z ZS
ru rv
=
F(r(u, v))
kru rv k dA
kru rv k
D
Z Z
=
F(r(u, v)) (ru rv ) dA.
D

This is analogous to the definition of the line integral of a vector field over
a curve C,
Z
Z
Z b
F dr =
F T ds =
F(r(t)) r0 (t) dt.
C

138

CHAPTER 3. VECTOR CALCULUS

Just as the orientation of a curve was relevant to the line integral of

a vector field over a curve, the orientation of a surface is relevant to the
surface integral of a vector field. We say that a surface S is orientable, or
oriented, if, at each point (x, y, z) in S, it is possible to choose a unique
vector n(x, y, z) that is normal to the tangent plane of S at (x, y, z), in such
a way that n(x, y, z) varies continuously over S. The particular choice of n
is called an orientation.
An orientable surface has two orientations, or, informally, two sides,
with normal vectors n and n. This definition of orientability excludes the
M
obius strip, because for this surface, it is possible for a continuous variation
of (x, y, z) to yield two distinct normal vectors at every point of the surface,
that are negatives of one another. Geometrically, the Mobius strip can be
said to have only one side, because negating any choice of continuously
varying n yields the same normal vectors.
For a surface that is the graph of a function z = g(x, y), if we choose the
parametrization
x = u, y = v, z = g(u, v),
then from
ru = h1, 0, gu i,

rv = h0, 1, gv i,

we obtain
ru rv = hgu , gv , 1i = hgx , gy , 1i
which yields
n=

hgx , gy , 1i
ru rv
=q
.
kru rv k
1 + gx2 + gy2

Because the z-component of this vector is positive, we call this choice of n

an upward orientation of the surface, while n is a downward orientation.
Example
(Stewart, Section 13.7, Exercise 22) Let S be the part of the cone
p
z = x2 + y 2 that lies beneath the plane z = 1, with downward orientation.
We wish to evaluate the surface integral
Z Z
F dS
S

where F = hx, y, z 4 i.
First, we must compute the unit normal vector for S. Using cylindrical
coordinates yields the parameterization
x = u cos v,

y = u sin v,

z = u,

0 u 1,

0 v 2.

3.7. SURFACE INTEGRALS

139

We then have
ru = hcos v, sin v, 1i,

rv = hu sin v, u cos v, 0i,

which yields
ru rv = hu cos v, u sin v, u cos2 v + u sin2 vi = uh cos v, sin v, 1i.
Because we assume downward orientation, we must have the z-component
of the normal vector be negative. Therefore, ru rv must be negated, which
yields
Z Z
Z Z
F dS =
F(x(u, v), y(u, v), z(u, v)) (ru rv ) dA,
S

where D is the domain of the parameters u and v, the rectangle [0, 1][0, 2].
Evaluating this integral, we obtain
Z Z
Z Z
F dS =
hu cos v, u sin v, u4 i uh cos v, sin v, 1i dA
S
D
Z 2 Z 1
=
(u cos2 v u sin2 v + u4 )u du dv
0
0
Z 2 Z 1
=
(u2 u5 ) du dv
0
0
Z 1
= 2
(u2 u5 ) du
0


= 2
=

.
3

1
u3 u6 

3
6 0

An alternative approach is to retain Cartesian coordinates, and then use the

formula for the unit normal for a downward orientation of a surface that is
the graph of a function z = g(x, y),
*
+
hgx , gy , 1i
x
1
y
p
=
n = q
,p
, 1 .
2
x2 + y 2
x2 + y 2
g2 + g2 + 1
x

This approach stil requires a conversion to polar coordinates to integrate

over the unit disk in the xy-plane. 2

140

CHAPTER 3. VECTOR CALCULUS

For a closed surface S, which is the boundary of a solid region E, we

define the positive orientaion of S to be the choice of n that consistently
point outward from E, while the inward-pointing normals define the negative
orientation.
Example (Stewart, Section 13.7, Exercise 26) To evaluate the surface integral
Z Z
F dS
S

where F(x, y, z) = hy, z y, xi and S is the surface of the tetrahedron with

vertices (0, 0, 0), (1, 0, 0), (0, 1, 0), and (0, 0, 1), we must evaluate surface
integrals over each of the four faces of the tetrahedron separately. We assume
positive (outward) orientation.
For the first side, S1 , with vertices (0, 0, 0), (1, 0, 0) and (0, 0, 1), we first
parameterize the side using
x = u,

y = 0,

0 u 1,

z = v,

0 v 1 u.

Then, from
ru = h1, 0, 0i,

rv = h0, 0, 1i,

we obtain
ru rv = h0, 1, 0i.
This vector is pointing outside the tetrahedron, so it is the outward normal
vector that we wish to use. Therefore, the surface integral of F over S1 is
Z Z

1 Z 1u

Z
F dS =
S1

h0, v 0, ui h0, 1, 0i dv du
0

0
1 Z 1u

Z
=

v dv du
0

Z
=

0

1 2 1u
v


2 0

1
2

du

(1 u)2 du

1
1 (1 u)3 
=

2
3
0
1
= .
6

3.7. SURFACE INTEGRALS

141

For the second side, S2 , with vertices (0, 0, 0), (0, 1, 0) and (0, 0, 1), we
parameterize using
x = 0,

y = u,

0 u 1,

z = v,

0 v 1 u.

Then, from
ru = h0, 1, 0i,

rv = h0, 0, 1i,

we obtain
ru rv = h1, 0, 0i.
This vector is pointing inside the tetrahedron, so we must negate it to obtain
the outward normal vector. Therefore, the surface integral of F over S2 is
Z Z

1 Z 1u

Z
F dS =
S2

hu, v u, 0i h1, 0, 0i dv du
0

0
1 Z 1u

=
u dv du
0
0
Z 1
=
u(u 1) du
0

1
u3 u2 
=

3
2 0
1
= .
6


For the base S3 , with vertices (0, 0, 0), (1, 0, 0) and (0, 1, 0), we parametrize
using
x = u, y = v, z = 0, 0 u 1, 0 v 1 u.
Then, from
ru = h1, 0, 0i,

rv = h0, 1, 0i,

we obtain
ru rv = h0, 0, 1i.
This vector is pointing inside the tetrahedron, so we must negate it to obtain
the outward normal vector. Therefore, the surface integral of F over S3 is
Z Z

1 Z 1u

F dS =
S3

hv, 0 v, ui h0, 0, 1i dv du
0

0
1 Z 1u

u dv du
0

142

CHAPTER 3. VECTOR CALCULUS

1

u(u 1) du

=
0

1
u3 u2 
=

3
2 0
1
= .
6


Finally, for the top face S4 , with vertices (1, 0, 0), (0, 1, 0) and (0, 0, 1),
we parametrize using
x = u,

y = v,

z = 1 u v,

0 u 1,

0 v 1 u,

since the equation of the plane containing this face is x + y + z 1 = 0. This

can be determined by using the three vertices to obtain two vectors within
the plane, and then computing their cross product to obtain the planes
normal vector.
Then, from
ru = h1, 0, 1i, rv = h0, 1, 1i,
we obtain
ru rv = h1, 1, 1i.
This vector is pointing outside the tetrahedron, so it is the outward normal
vector that we wish to use. Therefore, the surface integral of F over S4 is
Z Z

1 Z 1u

F dS =
S4

hv, 1 u 2v, ui h1, 1, 1i dv du

0

0
1 Z 1u

1 v dv du

=
0

0
1

=
0

1u

=
0

Z
=
=


1u
v 2 
du
v
2 0
(1 u)2
du
2

1 1 2
u du
2
0 2

1
u u3 

2
6 
0

1
.
3

3.8. STOKES THEOREM

143

Adding the four integrals together yields

Z Z
1 1 1 1
1
F dS = + = .
6 6 6 3
6
S
2

3.8

Stokes Theorem

Let C be a simple, closed, positively oriented, piecewise smooth plane curve,

and let D be the region that it encloses. According to one of the forms of
Greens Theorem, for a vector field F with continuous first partial derivatives
on D, we have
Z
Z Z
F dr =
C

(curl F) k dA,
D

where k = h0, 0, 1i.

By noting that k is normal to the region D when it is embedded in 3D space, we can generalize this form of Greens Theorem to more general
surfaces that are enclosed by simple, closed, piecewise smooth, positively
oriented space curves. Let S be an oriented, piecewise smooth surface that
is enclosed by a such a curve C. If we divide S into several small patches
Sij , then these patches are approximately planar. We can apply Greens
Theorem, approximately, to each patch by rotating it in space so that its
unit normal vector is k, and using the fact that rotating two vectors u and
v in space does not change the value of u v.
Most of the line integrals along the boundary curves of each path cancel
with one another due to the positive orientation of all such boundary curves,
and we are left with the line integral over C, the boundary of S. If we take
the limit as the size of the patches approches zero, we then obtain
Z
Z Z
Z Z
F dr =
curl F dS =
curl F n dS,
C

where n is the unit normal vector of S. This result is known as Stokes

Theorem.
Stokes Theorem can be used either to evaluate an surface integral or an
integral over the curve that encloses it, whichever is easier.
Example (Stewart, Section 13.8, Exercise 2) Let F(x, y, z) = hyz, xz, xyi
and let S be the part of the paraboloid z = 9 x2 y 2 that lies above the
plane z = 5, with upward orientation. By Stokes Theorem,
Z Z
Z
curl F dS =
F dr
S

144

CHAPTER 3. VECTOR CALCULUS

where C is the boundary curve of S, which is a circle of radius 2 centered at

(0, 0, 5), and parallel to the xy-plane. It can therefore be parameterized by
x = 2 cos t,

y = 2 sin t,

z = 5,

0 t 2.

Its tangent vector is then

r0 (t) = h2 sin t, 2 cos t, 0i.
We then have
Z Z
Z 2
curl F dS =
F(r(t)) r0 (t) dt
S
0
Z 2
=
h10 sin t, 10 cos t, 4 cos t sin ti h2 sin t, 2 cos t, 0i dt
0
Z 2
=
20 sin2 t + 20 cos2 t dt
0
Z 2
= 20
cos 2t dt
0

10 sin 2t|2
0

= 0.
This result can also be obtained by noting that because F = f , where
f (x, y, z) = xyz, it follows that curl F = 0. 2
Example (Stewart, Section 13.8, Exercise 8) We wish to evaluate the line
integral of F(x, y, z) = hxy, 2z, 3yi over the curve C that is the intersection
of the cylinder x2 + y 2 = 9 with the plane x + z = 5.
To describe the surface S enclosed by C, we use the parameterization
x = u cos v,

y = u sin v,

z = 5 u cos v,

0 u 3,

0 v 2.

Using
ru = hcos v, sin v, cos vi,

rv = hu sin v, u cos v, u sin vi,

we obtain
ru rv = hu, 0, ui.
We then compute

curl F =

, ,
x y z


hxy, 2z, 3yi = h1, 0, xi.

3.8. STOKES THEOREM

145

Let D be the domain of the parameters,

D = {(u, v) | 0 u 3,

0 v 2.

We then apply Stokes Theorem and obtain

Z Z
Z
curl F dS
F dr =
S
C
Z Z
curl F(r(u, v)) (ru rv ) dA
=
D
Z 3 Z 2
=
h1, 0, u cos vi hu, 0, ui dA
0
0
Z 3 Z 2
=
u u2 cos v dv du
0
0
Z 3
2
=
(uv u2 sin v)0 dv du
0
Z 3
= 2
u du du
= 2

0
3
u2 


2 0

= 9.
2
Stokes Theorem can also be used to provide insight into the physical
interpretation of the curl of a vector field. Let Sa be a disk of radius a
centered at a point P0 , and let Ca be its boundary. Furthermore, let v be a
velocity field for a fluid. Then the line integral
Z
Z
v dr =
v T ds,
Ca

Ca

where T is the unit tangent vector of Ca , measures the tendency of the fluid
to move around Ca . This is because this measure, called the circulation of v
around Ca , is greatest when the fluid velocity vector is consistently parallel
to the unit tangent vector. That is, the circulation around Ca is maximized
when the fluid follows the path of Ca .
Now, by Stokes Theorem,
Z
Z Z
v dr =
curl v dS
Ca

Sa

146

CHAPTER 3. VECTOR CALCULUS

Z Z
curl v n dS

=
Sa

Z Z
curl V(P0 ) n(P0 )

1 dS
Sa

a2 curl v(P0 ) n(P0 ).

As a 0, and Sa collapses to the point P0 , this approximation improves,
and we obtain
Z
1
curl v(P0 ) n(P0 ) = lim
v dr.
a0 a2 Ca
This shows that circulation is maximized when the axis around which the
fluid is circulating, n(P0 ), is parallel to curl v. That is, the direction of
curl v indicates the axis around which the greatest circulation occurs.

3.8.1

A Note About Orientation

Recall Stokes Theorem,

Z

Z Z
F dr =

curl F dS,
S

where C is a simple, closed, positively oriented, piecewise smooth curve and

S is a oriented surface enclosed by C. If C is parameterized by a function
r(t), where a t b, and S is parameterized by a function g(u, v), where
(u, v) D, then Stokes Theorem becomes
Z

Z Z

F(r(t)) r (t) dt =
a

curl F(g(u, v)) (gu gv ) du dv.

D

It is important that the parameterizations r and g have the proper orientation for Stokes Theorem to apply. This is why it is required that C
have positive orientation. It means, informally, that if one were to walk
along C, in such a way that n, the unit normal vector of S, can be viewed,
then S should always be on the left relative to the path traced along C.
It follows that the parameterizations of C and S must be consistent with
one another, to ensure that they are oriented properly. Otherwise, one of the
parameterizations must be reversed, so that the sign of the corresponding
integral is corrected. The orientation of a curve can be reversed by changing
the parameter to s = a + b t. The orientation of a surface can be reversed
by interchanging the variables u and v.

3.9. THE DIVERGENCE THEOREM

3.9

147

The Divergence Theorem

Let F be a vector field with continuous first partial derivatives. Recall a

statement of Greens Theorem,
Z Z
Z
div F dA,
F n ds =
D

where n is the outward unit normal vector of D. Now, let E be a threedimensional solid whose boundary, denoted by E, is a closed surface S
with positive orientation. Then, if we consider two-dimensional slices of E,
each one being parallel to the xy-plane, then each slice is a region D with
positively oriented boundary C, to which Greens Theorem applies. If we
multiply the integrals on both sides of Greens Theorem, as applied to each
slice, by dz, the infinitesimal thickness of each slice, then we obtain
Z Z
Z Z Z
F n dS =
div F dV,
S

or, equivalently,
Z Z

Z Z Z
F dS =

F dV.
E

This result is known as the Gauss Divergence Theorem, or simply the Divergence Theorem.
As the Divergence Theorem relates the surface integral of a vector field,
known as the flux of the vector field through the surface, to an integral of its
divergence over a solid, it is quite useful for converting potentially difficult
double integrals into triple integrals that may be much easier to evaluate,
as the following example demonstrates.
Example (Stewart, Section 13.9, Exercise 6) Let S be the surface of the
box with vertices (1, 2, 3), and let F(x, y, z) = hx2 z 3 , 2xyz 3 , xz 4 i. To
compute the surface integral of F over S directly is quite tedious, because
S has six faces that must be handled separately. Instead, we apply the
Divergence Theorem to integrate div F over E, the interior of the box. We
then have
Z Z
Z Z Z
F dS =
div F dV
S
E
Z 1Z 2Z 3
=
(x2 z 3 )x + (2xyz 3 )y + (xz 4 )z dz dy dz
1 2 3
1 Z 2 Z 3

Z
=

2xz 3 + 2xz 3 + 4xz 3 dz dy dx

148

CHAPTER 3. VECTOR CALCULUS

1

=
1

2
1

8xz 3 dz dy dx

3
3

z 3 dz dx
3
1
Z 1 " 4 3 #
z 
= 32
x
dx
4 3
1
= 32

= 0.
2
The Divergence Theorem can also be used to convert a difficult surface
integral into an easier one.
Example (Stewart, Section 13.9, Exercise 17) Let F(x, y, z) = hz 2 x, 31 y 3 +
tan z, x2 z + y 2 i. Let S be the top half of the sphere x2 + y 2 + z 2 = 1. To
evaluate the surface integral of F over S, we note that if we combine S
with S1 , the disk x2 + y 2 1, z = 0, with downward orientation, we then
obtain a new surface S2 that is the boundary of the top half of the ball
x2 + y 2 + z 2 1, which we denote by E. By the Divergence Theorem,
Z Z
Z Z
Z Z
Z Z Z
F dS +
F dS =
F dS =
div F dV.
S

S1

S2

We parameterize S1 by
x = u sin v,

y = u cos v,

z = 0,

0 u 1,

0 v 2.

This parameterization is used instead of the usual one arising from polar
coordinates, due to the downward orientation. It follows from
ru = hsin u, cos u, 0i,

rv = hu cos v, u sin v, 0i

that
ru rv = h0, 0, u sin2 v u cos2 vi = uh0, 0, 1i,
which points downward, as desired. From


1 3
2
div F(x, y, z) = (z x)x +
y + tan z + (x2 z + y 2 )z = x2 + y 2 + z 2 ,
3
y
which suggests the use of spherical coordinates for the integral over E, we
obtain
Z Z
Z Z Z
Z Z
F dS =
div F dV
F dS
S

S1

3.9. THE DIVERGENCE THEOREM

Z Z Z

149

(x2 + y 2 + z 2 ) dV

=
E
1 Z 2

F(x(u, v), y(u, v), z(u, v)) uh0, 0, 1i dv du

0

0
1 Z 2

/2

0
1

0
4

= 2

u3

u(u2 cos2 v) dv du

0
0
Z 2

cos2 v dv du

2
1 + cos 2v
u3
dv du
2
0
0
0

Z 1
Z 1 
sin 2v 2
4
3 v
+
2
d +
u
 du
2
4
0
0
0
1
Z 1
5


2  +
u3 du
5 0
0
1
4
u 
2
+
5
4 0
2
+
5
4
13
.
20

= 2

Z
sin d d +

/2

1 Z 2

sin d d d +

2 2

/2

4 cos |0

2
Suppose that F is a vector field that, at any point, represents the flow
rate of heat energy, which is the rate of change, with respect to time, of
the amount of heat energy flowing through that point. By Fouriers Law,
F = KT , where K is a constant called thermal conductivity, and T is a
function that indicates temperature.
Now, let E be a three-dimensional solid enclosed by a closed, positively
oriented, surface S with outward unit normal vector n. Then, by the law
of conservation of energy, the rate of change, with respect to time, of the
amount of heat energy inside E is equal to the flow rate, or flux, or heat
into E through S. That is, if (x, y, z) is the density of heat energy, then

Z Z Z

Z Z
F (n) dS,

dV =
E

where we use n because n is the outward unit normal vector, but we need
to express the flux into E through S.

150

CHAPTER 3. VECTOR CALCULUS

From the definition of F, and the fact that = c0 T , where c is the

specific heat and 0 is the mass density, which, for simplicity, we assume to
be constant, we have
Z Z
Z Z Z

KT n dS.
c0 T dV =
t
S
E
Next, we note that because c, 0 , and E do not depend on time, we can
write
Z Z
Z Z Z
T
c0
KT dS.
dV =
t
E
S
Now, we apply the Divergence Theorem, and obtain
Z Z Z
Z Z Z
Z Z Z
T
c0
dV =
K div T dV =
K2 T dV.
t
E
E
E
That is,

Z Z Z 
T
2
c0
K T dV = 0.
t
E
Since the solid E is arbitrary, it follows that
T
K 2
=
T.
t
c0
This is known as the heat equation, which is one of the most important
partial differential equations in all of applied mathematics.

3.10

Differential Forms

To date, we have learned the following theorems concerning the evalution of

integrals of derivatives:
The Fundamental Theorem of Calculus:
Z b
f 0 (x) dx = f (b) f (a)
a

The Fundamental Theorem of Line Integrals:

Z
a

f (r(t)) r0 (t) dt = f (r(b)) f (r(a))

3.10. DIFFERENTIAL FORMS

Greens Theorem:
Z Z

151

Z
(Qx Py ) dA =

P dx + Q dy

Z Z

Stokes Theorem:
F dr

curl F dS =
S

Gauss Divergence Theorem:

Z Z Z
Z Z
div F dV =
F dS
E

All of these theorems relate the integral of the derivative or gradient of a

function, or partial derivatives of components of a vector field, over a higherdimensional region to the integral or sum of the function or vector field over a
lower-dimensional region. Now, we will see how the notation of differential
forms can be used to combine all of these theorems into one. It is this
notation, as opposed to vectors and operations such as the divergence and
curl, that allows the Fundamental Theorem of Calculus to be generalized to
functions of several variables.
A differential form is an expression consisting of a scalar-valued function
f : K Rn R and zero or more infinitesimals of the form dx1 , dx2 , . . . , dxn ,
where x1 , x2 , . . . , xn are the independent variables of f . The order of a differential form is defined to be the number of infinitesimals that it includes.
For simplicity, we set n = 3 of three variables. With that in mind, a
0-form, or a differential form of order zero, is simply a scalar-valued function
f (x, y, z). A 1-form is a function f (x, y, z) together with one of the expressions dx, dy or dz. A 2-form is a function f (x, y, z) together with a pair of
distinct infinitesimals, which can be either dx dy, dy dz or dz dx. Finally, a
3-form is an expression of the form f (x, y, z) dx dy dz.
Example The function f (x, y, z) = x2 y + y 3 z is a 0-form on R3 , while
f dx = (x2 y + y 3 z) dx and f dy = (x2 y + y 3 z) dy are both examples of a
1-form on R3 . 2
Example Let f (x, y, z) = 1/(x2 + y 2 + z 2 ). Then f dx dy is a 2-form on
R3 {(0, 0, 0}, while f dx dy dz is a 3-form on the same domain. 2
Forms of the same order can be added and scaled by functions, as the following examples show.

152

CHAPTER 3. VECTOR CALCULUS

Example Let f (x, y, z) = exy sin z and let g(x, y, z) = (x2 + y 2 + z 2 )3/2 .
Then f , g and f + g are all 0-forms on R3 , and
f + g = exy sin z + (x2 + y 2 + z 2 )3/2 .
That is, addition of 0-forms is identical to addition of functions.
If we define 1 = f dx and 2 = g dy, then 1 and 2 are both 1-forms
on R3 , and so is = 1 + 2 , where
= f dx + g dy = exy sin z dx + (x2 + y 2 + z 2 )3/2 dy.
Furthermore, if h(x, y, z) = xy 2 z 3 , and
1 = f dx dy,

2 = g dz dx

are 2-forms on R3 , then

= h1 + 2 = xy 2 z 3 exy sin z dx dy + (x2 + y 2 + z 3 )3/2 dz dx
is also a 2-form on R3 . 2
Example Let f (x, y, z) = cos x, g(x, y, z) = ey and h(x, y, z) = xyz 2 . Then,
1 = f dx dy dz and 2 = g dx dy dz are 3-forms on R3 , and so is
= 1 + h2 = (cos x + xyz 2 ey ) dx dy dz.
2
It should be noted that like addition of functions, addition of differential
forms is both commutative, associative, and distributive. Also, there is
never any need to add forms of different order, such as adding a 0-form to
a 1-form.
We now define two essential operations on differential forms. The first
is called the wedge product, a multiplication operation for differential forms.
Given a k-form and an l-form , where 0 k + l 3, the wedge product
of and , denoted by , is a (k + l)-form. It satisfies the following laws:
1. For each k there is a k-form 0 such that 0 = 0 = 0 for any
l-form .
2. Distributitivy: If f is a 0-form, then
(f 1 + 2 ) = f (1 ) + (2 ).

3.10. DIFFERENTIAL FORMS

153

3. Anticommutativity:
= (1)kl ( ).
4. Associativity:
1 (2 3 ) = (1 2 ) 3
5. Homogeneity: If f is a 0-form, then
(f ) = (f ) = f ( ).
6. If dxi is a basic 1-form, then dxi dxi = 0.
7. If f is a 0-form, then f = f .
Example Let = f dx and = g dy be 1-forms. Then
= (f dx g dy) = f g(dx dy) = f g dx dy,
by homogeneity, while
= (1)1(1) ( ) = f g dx dy.
On the other hand, if = h dy dz is a 2-form, then
= f h(dy dz dx) = f h dy dz dx = f h dy dx dz = f h dx dy dz
by homogeneity and anticommutativity, while
= f h(dy dz dy) = f h dy dz dy = f h dy dy dz = 0.
2
Note that if any 3-form on R3 is multiplied by a k-form, where k > 0, then
the result is zero, because there cannot be distinct basic 1-forms in the wedge
product of such forms.
Example Let = x dx y dy, and = z dy dz x dz dx. Then
= (x dx y dy) (z dy dz x dz dx)
= (x dx z dy dz) (y dy z dy dz) (x dx x dz dz) +
(y dy x dz dx)
= xz dx dy dz yz dy dy dz x2 dx dz dx + xy dy dz dx
= xz dx dy dz yz dy dy dz + x2 dx dx dz + xy dy dz dx
= xz dx dy dz 0 0 xy dy dx dz
= (xz + xy) dx dy dz.

154

CHAPTER 3. VECTOR CALCULUS

2
The second operation is differentiation. Given a k-form , where k < 3,
the derivative of , denoted by d, is a (k +1)-form. It satisfies the following
laws:
1. If f is a 0-form, then
df = fx dx + fy dy + fz dz
2. Linearity: If 1 and 2 are k-forms, then
d(1 + 2 ) = d1 + d2
3. Product Rule: If is a k-form and is an l-form, then
d( ) = (d ) + (1)k ( d)
4. The second derivative of a form is zero; that is, for any k-form ,
d(d) = 0.
We now illustrate the use of these differentiation rules.
Example Let = x2 y 3 z 4 dx dy be a 2-form. Then, by Linearity and the
Product Rule,
d = [d(x2 y 3 z 4 ) dx dy] + (1)0 [x2 y 3 z 4 d(dx dy)]
 2 3 4



=
(x y z )x dx + (x2 y 3 z 4 )y dy + (x2 y 3 z 4 )z dz dx dy +
 2 3 4

x y z {(d(dx) dy) + (1)1 (dx d(dy)}




=
2xy 3 z 4 dx + 3x2 y 2 z 4 dy + 4x2 y 3 z 3 dz dx dy +
 2 3 4

x y z {(0 dy) (dx 0)}
= 2xy 3 z 4 dx dx dy + 3x2 y 2 z 4 dy dx dy + 4x2 y 3 z 3 dz dx dy + 0
= 4x2 y 3 z 3 dx dz dy
= 4x2 y 3 z 3 dx dy dz.
In general, differentiating a k-form , when k > 0, only requires differentiating the coefficient function with respect to the variables that are not among
any basic 1-forms that are included in . In this example, since = f dx dy,
we obtain d = fz dz dx dy = fz dx dy dz. 2

We now consider the kind of differential forms that appear in the theorems of vector calculus.

3.10. DIFFERENTIAL FORMS

155

Let = f (x, y, z) be a 0-form. Then, by the first law of differentiation,

d = f hdx, dy, dzi.
If C is a smooth curve with parameterization r(t) = hx(t), y(t), z(t)i,
a t b, then
Z b
Z b
f (r(t)) hx0 (t), y 0 (t), z 0 (t)i dt
f (r(t)) r0 (t) dt =
a
a
Z b
d(r(t))
=
a
Z
d.
=
C

It follows from the Fundamental Theorem of Line Integrals that

Z
d = (r(b)) (r(a)).
C

The boundary of C, C, consists of its initial point A and terminal point B. If we define the integral of a 0-form over this 0dimensional region by
Z
= (B) (A),
C

which makes sense considering that, intuitively, the numbers 1 and 1

serve as an appropriate outward unit normal vector at the terminal
and initial points, respectively, then we have
Z
Z
d =
.
C

Let = P (x, y) dx + Q(x, y) dy be a 1-form. Then

d = d[P (x, y) dx] + d[Q(x, y) dy]
= dP (x, y) dx P (x, y) d(dx) + dQ(x, y) dy
Q(x, y) d(dy)
= (Px dx + Py dy) dx 0 + (Qx dx + Qy dy) dy 0
= Px dx dx + Py dy dx + Qx dx dy + Qy dy dy
= (Qy Px ) dx dy.
It follows from Greens Theorem that
Z
Z Z
=
d.
C

156

CHAPTER 3. VECTOR CALCULUS

If we proceed similarly with a 1-form
= F hdx, dy, dzi = P (x, y, z) dx + Q(x, y, z) dy + R(x, y, z) dz,
then we obtain
d = curl F hdy dz, dz dx, dx dyi
= (Ry Qz ) dy dz + (Pz Rx ) dz dx + (Qy Px ) dx dy.
Let S be a smooth surface parameterized by
r(u, v) = hx(u, v), y(u, v), z(u, v)i,

(u, v) D.

Then the (unnormalized) normal vector ru rv is given by

ru rv = hxu , yu , zu i hxv , yv , zv i
= hyu zv zu yv , zu xv xu zv , xu yv yu xv i


(y, z) (z, x) (x, y)
,
,
.
=
(u, v) (u, v) (u, v)
We then have
Z Z
Z Z
curl F dS =
curl F n dS
S
Z ZS
=
curl F(r(u, v)) (ru rv ) du dv
Z ZD 
(y, z)
+
=
[Ry (r(u, v)) Qz (r(u, v))]
(u, v)
D
(z, x)
[Pz (r(u, v)) Rx (r(u, v))]
+
(u, v)

(x, y)
[Qx (r(u, v)) Py (r(u, v))]
du dv
(u, v)
Z Z
(Ry Qz ) dy dz + (Pz Rx ) dz dx +
=
S

(Qy Px ) dx dy
Z Z
=
d.
S

If C is the boundary curve of S, and C is parameterized by r(t) =

hx(t), y(t), z(t)i, a t b, then
Z
Z b
F dr =
F(r(t)) r0 (t) dt
C

3.10. DIFFERENTIAL FORMS

Z

157

hP (r(t)), Q(r(t)), R(r(t)) hx0 (t), y 0 (t), z 0 (t)i dt

(r(t)) dt
Za

.
C

It follows from Stokes Theorem that

Z Z
Z
d.
=
S

Let F = hP, Q, Ri. Let be the 2-form

= P dy dz + Q dz dx + R dx dy.
Then
d = dP dy dz + dQ dz dx + dR dx dy
= [Px dx + Py dy + Pz dz] dy dz + [Qx dx + Qy dy + Qz dz] dz dx +
[Rx dx + Ry dy + Rz dz] dx dy
= Px dx dy dz + Qy dy dz dx + Rz dz dx dy
= Px dx dy dz Qy dy dx dz Rz dx dz dy
= Px dx dy dz + Qy dx dy dz + Rz dx dy dz
= div F dx dy dz.
Let E be a solid enclosed by a smooth surface S with positive orientation, and let S be parameterized by
r(u, v) = hx(u, v), y(u, v), z(u, v)i,
We then have
Z Z
Z
F dS =
S
Z
=
Z
=
Z
=

(u, v) D.

Z
F n dS
ZS
F(r(u, v)) (ru rv ) du dv
D

Z
hP (r(u, v)), Q(r(u, v)), R(r(u, v))i
ZD
P (r(u, v))
D

(y, z) (z, x) (x, y)

,
,
(u, v) (u, v) (u, v)

(y, z)
(z, x)
+ Q(r(u, v))
+
(u, v)
(u, v)


du dv

158

CHAPTER 3. VECTOR CALCULUS

(x, y)
R(r(u, v))
du dv
(u, v)
Z Z
P dy dz + Q dz dx + R dx dy
=
Z ZS
.
=
S

It follows from the Divergence Theorem that

Z Z Z
Z Z
d.
=
E

Putting all of these results together, we obtain the following combined

theorem, that is known as the General Stokes Theorem:
If M is an oriented k-manifold with boundary M , and is a
(k 1)-form defined on an open set containing M , then
Z
Z
=
d.
M

The importance of this unified theorem is that, unlike the previously stated
theorems of vector calculus, this theorem, through the language of differential forms, can be generalized to functions of any number of variables. This
is because operations on differential forms are not defined in terms of other
operations, such as the cross product, that are limited to three variables.
For example, given a 3-form = f (x, y, z, w) dx dy dw, its integral over a
3-dimensional, closed, positively oriented hypersurface S embedded in R4 is
equal to the integral of d over the 4-dimensional solid E that is enclosed by
S, where d is computed using the previously stated rules for differentiation
and multiplication of differential forms.